recommendations for adverse event monitoring programs for

Transcription

LSRO
RECOMMENDATIONS FOR
ADVERSE EVENT MONITORING PROGRAMS
FOR DIETARY SUPPLEMENTS
June 2004
Life Sciences Research Office
9650 Rockville Pike
Bethesda, Maryland 20814
www.LSRO.org
The distribution or electronic posting of this PDF file is strictly prohibited without
the written permission of the Life Sciences Research Office.
RECOMMENDATIONS FOR
ADVERSE EVENT MONITORING PROGRAMS
June 2, 2004
Editor:
Catherine J. Klein, Ph.D., R.D.
Prepared for Metabolife International, Inc., 5643 Copley Drive, San Diego, CA 92111,
through its counsel, Patton Boggs, L.L.P., 2550 M Street, N.W., Washington, DC 20037-1350
9650 Rockville Pike
Bethesda, Maryland 20814
www.LSRO.org
LSRO Report: Adverse Event Monitoring Programs for Dietary Supplements
 2004 Life Sciences Research Office
No part of this document may be reproduced by any mechanical, photographic, or electronic process, or in form of a
phonographic recording, nor may it be stored in a retrieval system, transmitted, or otherwise copied for public or
private use, without written permission from the publisher, except for the purposes of official use by the U.S.
Government.
Copies of this publication may be obtained from Life Sciences Research Office. Orders and inquiries may be
directed to LSRO, 9650 Rockville Pike, Bethesda, MD 20814. Phone: 301-634-7030. Fax: 301-634-7876.
Website: www.LSRO.org.
ISBN: 0975316710
Library of Congress Catalog Number: 2004103755
PREFACE
The Life Sciences Research Office (LSRO) provides scientific assessments of topics in the biomedical sciences.
LSRO reports are based on comprehensive literature reviews and the scientific opinions of knowledgeable
investigators, who are engaged in work in relevant areas of biology and medicine.
This LSRO study, Recommendations for Adverse Event Monitoring Programs for Dietary Supplements, was
initiated and funded by Metabolife International, Inc., 5643 Copley Drive, San Diego, CA 92111, through its
counsel, Patton Boggs, L.L.P., 2550 M Street, N.W., Washington, DC 20037-1350, in accordance with a contract
between LSRO and Patton Boggs, in its capacity as counsel to Metabolife.
This LSRO report was based on discussions of, and materials evaluated by, an ad hoc Expert Committee convened
by LSRO. LSRO independently appointed members of the Expert Committee, who were chosen for their
qualifications, experience, and judgment, with due considerations for balance and breadth in the appropriate
professional disciplines. Members of the Expert Committee and others who assisted in the preparation of the report
are identified in Appendix A. The Expert Committee convened six times (four full meetings and two conference
calls) to assess the available data and make recommendations.
The study described by this report represents a two-phase project. The objective of Phase I was to review and
compare individual data records associated with the use of dietary supplements and evaluate their usefulness for
generating signals of potential product problems. The objectives of Phase II were to review postmarketing
surveillance programs described in the scientific literature and make recommendations for the design and
implementation of an effective system for monitoring and responding to complaints of health-related adverse events
associated with the use of dietary supplements. Currently there is no federal guidance or mandate for the collection,
documentation, or evaluation of health-related consumer complaints associated with the use of dietary supplements.
The Committee directed the selection and review of individual data records associated with the use of ephedrine
alkaloid-containing dietary supplements from industry (Metabolife) and federal (U.S. Food and Drug
Administration) data sets. The Committee also heard from outside speakers and interacted with experts who
summarized sample cases. LSRO staff, special consultants, and members of the Expert Committee considered all
available information when drafting the report, incorporated reviewers’ comments, and provided additional
documentation and viewpoints for incorporation into the final report. This detailed LSRO report, which reviews
examples of surveillance programs and highlights positive characteristics of such systems, is intended for industrywide use.
The final report was reviewed and approved by the Expert Committee and the LSRO Board of Directors. On
completion of these review procedures, the Executive Director of LSRO approved and transmitted the report to
Metabolife through its counsel.
LSRO is solely responsible for the content of the report. The report was developed independently, and the
conclusions drawn therein do not necessarily reflect the views or policies of Metabolife, its counsel, or of any of
their employees. Although LSRO accepts full responsibility for the study conclusions and accuracy of the report,
the report does not necessarily represent the opinion of the LSRO Board of Directors. The mention of trade names,
commercial products, or organizations does not imply endorsement by LSRO.
Michael Falk, PhD.
Executive Director,
June 2, 2004
TABLE OF CONTENTS
EXECUTIVE SUMMARY
-1-
Setting
-1-
Study objectives and scope
-1-
Surveillance programs
-2-
Study methods
-2-
Findings
-4-
Summary of recommendations
-5-
I. INTRODUCTION
-7-
I.1 Dietary supplements and key regulations
-7-
I.1.1 Dietary supplements
I.1.2 Product users
I.1.3 Key product regulations
-7-7-8-
I.2 Product risk management and postmarketing surveillance programs
I.2.1 Product risk management
I.2.2 Postmarketing surveillance
II. GENERAL SURVEILLANCE PROGRAMS
II.1 Postmarketing risk management
II.1.1 Postmarketing sources of risk
II.1.2 Individual data record-based postmarketing surveillance
II.2 General features of surveillance programs
II.2.1 Recognition and reporting of adverse events in individual product users
II.2.2 Creation of individual data records by surveillance programs
II.2.3 Building a relational database of individual data records
II.2.4 Signaling from databases of individual data records
II.2.5 Signaling from other information sources
II.2.6 Issues resolution and formulation of risk management interventions
II.3 Summary of general surveillance programs
III. RETROSPECTIVE REVIEW OF SAMPLE INDIVIDUAL DATA RECORDS
-8-8-9- 11 - 11 - 11 - 13 - 14 - 14 - 16 - 17 - 18 - 20 - 21 - 22 - 22 -
III.1 Introduction
- 22 -
III.2 Sample individual data records
- 23 -
III.2.1 The industry (MET) data set
III.2.2 The federal (FDA) data set
III.2.3 Selection of sample individual data records
- 23 - 24 - 26 -
III.3 Analyses of sample sets
III.3.1 Information of value for signal detection
III.3.2 Information of value for signal prioritization
III.3.3 Information of value for signal evaluation
III.4 Results and discussion
III.4.1 Information of value for data intake and creation of individual data records
III.4.5 Issues resolution and formulation of risk management interventions
III.5 Summary of findings related to methods used to collect adverse event data and create
individual data records
III.5.1 Standardized data forms
III.5.2 Follow-up data collection
III.6 General conclusions related to the value of data from federal (FDA) and industry (MET)
data sets for signaling
IV. POSTMARKETING SURVEILLANCE PROGRAMS FOR DIETARY SUPPLEMENTS
IV.1 Considerations for the design of a postmarketing surveillance program for dietary
supplements
IV.1.1 Well-designed postmarketing surveillance programs enable detection and evaluation of
signals arising from problems introduced during manufacturing
IV.1.2 Well-designed postmarketing surveillance programs enable detection and evaluation of
signals arising from non-manufacturing factors
IV.1.3 Effective surveillance programs detect relevant new patterns and outlier experiences
worthy of follow up
IV.1.4 Summary of considerations for the design of a postmarketing surveillance program for
dietary supplements
IV.2 Desired features of a postmarketing surveillance program for dietary supplements
IV.2.1 Overview of recommendations
IV.2.2 General recommendation
IV.2.3 Recommendations for reporting and intake processes
IV.2.4 Recommendations for creation of individual data records
IV.2.5 Recommendations for building a relational database of individual data records
IV.2.6 Recommendations for signaling from databases of individual data records
IV.2.7 Recommendations for signaling from other information sources
IV.2.8 Recommendations for issues resolution and risk management
IV.2.9 Recommendations for quality assurance and oversight
IV.3 Discussion and summary
IV.3.1 Summary of considerations for the design of a postmarketing surveillance program for
dietary supplements
IV.3.2 Summary of desired features and recommended enhancements of postmarketing
surveillance programs for dietary supplements
IV.3.3 Opportunities for collaboration
IV.3.4 Summary of postmarketing surveillance programs for dietary supplements
- 27 - 27 - 28 - 29 - 30 - 30 - 30 - 31 - 35 - 39 - 39 - 39 - 41 - 41 - 43 - 43 - 43 - 44 - 45 - 45 - 45 - 46 - 49 - 49 - 50 - 52 - 53 - 54 - 55 - 55 - 59 - 59 - 60 - 60 - 61 -
V. CONCLUSIONS
- 61 -
V.1 Summary of findings
- 61 -
V.1.1 Identification of the necessary data for signal generation
V.1.2 Principal findings of the LSRO Phase I review of individual data records
V.1.3 Principal findings of the LSRO Phase II review of existing postmarketing surveillance
programs
- 61 - 62 - 62 -
V.2 Supplemental remarks
- 63 -
V.3 Summary of recommendations
- 64 -
LITERATURE CITATIONS
- 64 LIST OF TABLES
Table 2-1. General functional steps in postmarketing surveillance programs
- 13 -
Table 3-1. Descriptive characteristics of individual data records associated with dietary supplements
containing ephedrine alkaloids from two sets of sample records
- 32 -
Table 3-2. Comparison of the percentage of individual data records that met minimal criteria for signal
detection in two sets of sample records
- 33 -
Table 3-3. The types of adverse events reported to a federal surveillance program (having a relatively
large percentage of health care professional reporters) compared with the frequency of those
events reported via a manufacturer’s customer service telephone line
- 34 -
Table 3-4. Comparison of the percentage of individual data records that met criteria for signal evaluation
in two sets of sample records
- 36 -
Table 3-5. General conclusions of a retrospective review of sample individual data records and their
usefulness for signal generation
- 42 -
Table 4-1. Summary of primary recommendations for surveillance programs for dietary supplements
- 48 -
Table 4-2. Functional needs of an intake form useful for recording adverse events associated with the
use of dietary supplements
- 56 -
LIST OF FIGURES
Figure 1-1. Theoretical risk management cycle for dietary supplement products
-9-
Figure 2-1. Sources of risk from products regulated by the Food and Drug Administration
- 11 -
Figure 2-2. Comparison of the distributions of adverse events, grouped by system organ class, for the
prescription drug metoprolol reported during premarket studies before 1975 and during
postmarketing surveillance from 1975 to 1984
- 12 -
Figure 2-3. Vaccine related events and coincidental events contribute to adverse events occurring
after vaccination
- 15 -
Figure 2-4. Processing of an individual product user experience using the paper method versus new
technology
- 16 -
Figure 2-5. Steps in the spontaneous event signaling process from the initial report of data to the
formation of a spontaneous signaling argument
- 19 -
Figure 3-1. Example of the “Nurses Database-Caller Info” form used by Metabolife International, Inc.
(San Diego, CA) to document health-related complaints received on their customer service
telephone line
- 25 -
Figure 4-1. Flow diagram of operational components of a surveillance program for dietary supplements
- 46-
APPENDICES
Table of contents for appendices
- 69 -
Appendix A: Individuals and organizations
- 72 -
Appendix B: Phase I letter
- 77 -
Appendix C: Dietary supplements
- 79 -
Appendix D: Key regulations for dietary supplements and related activities
- 87 -
Appendix E: Example surveillance programs for food additives
- 93 -
Appendix F: Existing surveillance programs for dietary supplements
-100 -
Appendix G: Ancillary issues
- 122 -
Appendix H: Acronyms and glossary
- 133 -
EXECUTIVE SUMMARY
SETTING
Dietary supplements are used by an estimated one-third of adults in the United States on a daily or routine basis
(Commission on Dietary Supplement Labels, 1997). The term “dietary supplements,” as defined in the Dietary
Supplement Health and Education Act (DSHEA) (U.S. Congress, 1994), includes vitamins, minerals, amino acids,
enzymes, herbs, and other botanicals. Dietary supplements are used to confer health benefits. For example, it is
recognized that nutrients such as calcium and folic acid, which are available as supplements, help build strong bones
(i.e., calcium) and help reduce the risk of neural tube defects (i.e., folic acid) (U.S. Congress, 1994).
Similar to food additives and medications (prescription and nonprescription), health problems from dietary
supplements can arise from improper use, product tampering, and product defects that alter product identity, quality,
purity, strength, and/or composition. In addition, safety concerns over certain dietary supplements have been
expressed by various sources, e.g., the lay-press, medical journals, the U.S. Food and Drug Administration (FDA).
Several studies have reviewed safety and regulatory issues related to dietary supplements (Shekelle et al., 2003; U.S.
Food and Drug Administration, 1994; U.S. General Accounting Office, 2003a, 2003b, 2003c).
FDA regulates dietary supplements by enforcing the Federal Food, Drug, and Cosmetic Act as amended by DSHEA,
in that dietary supplements are regulated as a subcategory of food (U.S. Congress, 1994). Hence, the regulatory
framework for product safety is primarily a “postmarket” program, where there is no requirement for manufacturers
to provide evidence of product safety prior to marketing products. However, the manufacturer is required to notify
FDA at least 75 days in advance of introducing a product in the market that contains a “new” dietary ingredient that
has never been present in the food supply. Most other food products are regulated similarly to dietary supplements,
with some exceptions (e.g., infant formulas, new food additives). In contrast, before marketing an infant formula,
the manufacturer must provide FDA with assurance of the nutritional quality of the infant formulation. Food
additives are also subject to premarket approval by FDA unless their use is generally recognized as safe (GRAS) by
qualified experts. Under DSHEA, FDA has the authority to determine that a dietary supplement may not be sold if it
"presents a significant or unreasonable risk of injury."
Because problems in manufacturing of food and medications may lead to adulterated or misbranded products that
have the potential to cause adverse health effects, FDA promulgates “current good manufacturing practice” to
reduce the risk of adulterated or misbranded products. Recently, FDA proposed “Current Good Manufacturing
Practice in Manufacturing, Packing, or Holding Dietary Ingredients and Dietary Supplements” (U.S. Food and Drug
Administration, 2003a). Currently, there is no requirement for a manufacturer to monitor adverse events related to
the use of their dietary supplement products. Hence, it is anticipated that dietary supplement manufacturers are
unfamiliar with terminology and processes that have been used by other industries to monitor adverse events. An
adverse event is any undesirable health-related sign or symptom that is detected in an exposed individual after use of
a product.
STUDY OBJECTIVES AND SCOPE
The Life Sciences Research Office (LSRO) undertook a study of adverse event monitoring programs for dietary
supplements. Metabolife International, Inc. (MET) through its counsel, the law firm Patton Boggs, initiated and
funded the study.1 Concerned that the current system of documenting individual data records (IDRs) was not
intended as an adverse event reporting system and may not support recognition of product-related effects, MET
sought scientific advice on how such a system could be tailored to the task of monitoring the safety of dietary
supplements.
The specific objective of the study is to make recommendations for the design and development of adverse event
monitoring programs for dietary supplements. Currently there is no federal guidance or mandate for the collection,
documentation, or evaluation of health-related consumer complaints associated with the use of dietary supplements.
1
Although providing financial resources, MET and its counsel, Patton Boggs, were not responsible for the content of
this LSRO report nor were they asked to endorse this report.
-1The distribution or electronic posting of this PDF file is strictly prohibited without
SURVEILLANCE PROGRAMS
Over the last 40 years, both voluntary and mandatory postmarketing risk management programs have been
established by various industries to gather and integrate product-specific, health related data. These programs,
referred to as “postmarketing surveillance,” have proven to be effective and relatively inexpensive to operate.
Postmarketing surveillance programs exist for household goods, foods, and products produced by the chemical
industry. Products monitored by surveillance programs, such as nonprescription medications, are generally labeled
with toll-free numbers to report adverse events. Because voluntary postmarketing surveillance is uncommon in the
dietary supplement industry, there are no accepted surveillance terminologies or standards. Hence, for the purposes
of this report, LSRO has drawn upon terminology utilized by other industries. A glossary and list of acronyms used
in this LSRO report are provided in Appendix H. Some definitions, adopted for the purposes of this report, may
differ from definitions presented by others (e.g., the term “serious”).
Postmarketing surveillance programs are similar across diverse product categories because they follow similar
information processing principles. An overview of postmarketing surveillance programs is provided in Section II.
In general, surveillance programs establish standard operating procedures by which the receipt, collection, analysis,
interpretation, and documentation of IDRs can be consistently achieved. In this LSRO report, the term “surveillance
program” is used to refer to all personnel, processes, and computer resources that carry out surveillance. The
primary purpose of a surveillance program is to identify potential product-related health problems that are
appropriate for rapid dissemination of information or other actions. In an operational sense, a surveillance program
allows the adverse events that are submitted in connection with a particular product to be scrutinized for
disproportions or unusual cases that might represent problems, a process known as signaling. Surveillance systems
generate signals that require further quantification and evaluation to determine whether the signals represent
coincidence, artifact, or a genuine problem in toxicity that might lead to changes in labeling and/or restrictions in
use. Thus, in other industries, such as the pharmaceutical industry, adverse event data are required to be collected
and evaluated without a requirement for evidence of causality.
A surveillance system consists of the following functional steps:
1.
2.
3.
4.
5.
6.
Recognition and reporting of adverse events in individual product users2
Creation of IDRs
Building an integrated, relational database of IDRs
Signaling from databases of IDRs
Signaling from other information sources
Issues resolution and formulation of risk management interventions
In terms of signaling (Step 4), there are three sequential stages by which product-adverse event signals are generated
and assessed: (a) signal detection, (b) signal prioritization, and (c) signal evaluation. These stages provide additional
data that assist in evaluating whether further investigation of the product-event association is warranted. Signals that
have been identified from an adverse event monitoring program are fully assessed in the issues resolution phase of
the process by using all pertinent data that is available.
IDR-based postmarketing surveillance programs are intended to identify adverse events early during the marketing
of a health care product and to provide additional information relevant to the adverse event profile of a product.
These surveillance activities are particularly valuable when identified problems can be addressed by a targeted risk
management program and lead to prevention of similar events in the future.
STUDY METHODS
LSRO reviewed IDRs and postmarketing surveillance for dietary supplements and other consumer products to meet
the objectives of the study in two Phases:
• Phase I. A case study of Metabolife 356® and other ephedrine alkaloid-containing dietary supplements
was used to make conclusions about the usefulness of the data generally as signals for product safety
(Section III).3
2
For the purposes of this LSRO publication, the term “reporting” includes the action of submitting a voluntary
complaint/report of a consumer’s adverse event to industry.
•
Phase II. Several noteworthy surveillance programs described in the scientific literature (Appendices E and
F) were reviewed and input from speakers and agency representatives familiar with approaches to
postmarketing surveillance of dietary supplements, food additives, and prescription and nonprescription
medication was evaluated to identify procedural problems with these systems, and make recommendations
about establishing a useful, practical system to monitor and respond to adverse event reports, including
criteria for documenting and reporting health complaints.
In Phase I, LSRO examined sample ephedrine-related IDRs that were collected by FDA through a surveillance
program and IDRs collected by MET through a customer service hotline. Although the manufacturer did not intend
the collection and documentation of calls to a hotline as “postmarketing surveillance,” this system and the resultant
data were reviewed by the Committee4 to gain insight into the design of possible surveillance systems for dietary
supplements.
The sample IDRs were drawn from the same data sets that were reviewed in other studies, such as the review
conducted by the U.S. General Accounting Office (2003b). LSRO did not attempt to examine, describe, or tabulate
events for the entire pool of available IDRs related to ephedrine alkaloid-containing dietary supplements available
from MET or from FDA (Docket No. 95N-0304), but selected 200 IDRs from each data set. LSRO did not evaluate
causality or use aggregate data in any way to assess the probability of a causal link between a product and an
adverse event.5
LSRO built its analysis of sample FDA and MET IDRs around the specific steps of signal detection, signal
prioritization, and signal evaluation. Signal detection is a process of sorting and identifying IDRs in order to find
interesting supplement-adverse event relationships that merit further attention. Evidence of causality is not a
requirement for signal detection, and the detection of a signal is not equivalent to determining causality (U.S. Food
and Drug Administration, 2003d).
The Committee determined that a minimum of five criteria must be met in order for an IDR to support signal
detection:
1. An identifier for the product user (e.g., initials, age, gender) was obtained
2. The complaint was health-related
3. A dietary supplement product was consumed prior to the onset of the health-related complaint
4. The IDR was dated (day/month/year)
5. An identifier for the reporter was obtained
For the purpose of the case review, only the first four of these five criteria were used to assess sample sets for signal
detection because the identifiable reporter was redacted from all sample IDRs obtained by LSRO.
During the signal prioritization process, weighting factors are applied to select/rank the output from signal detection
in preparation for signal evaluation. The Committee determined that two types of data were needed for signal
prioritization: (1) data that rank the reliability and accuracy of the record, and (2) data used to rank the seriousness
of the health-related event or outcome.
The purpose of signal evaluation is to form and characterize a case series to determine whether further investigation
of the product-event relationship is indicated. The Committee identified 11 types of data that are useful for signal
evaluation. They are listed below:
1. Gender
2. Age at the time of the event
3
Effective November 18, 2003, MET suspended the sale of their ephedra-containing products. On December 30,
2003, FDA announced plans to prohibit sales of dietary supplements containing ephedra (U.S. Food and Drug
Administration, 2003e). The final rule prohibiting the sale of ephedrine alkaloid-containing dietary supplements
was issued February 11, 2004 (U.S. Food and Drug Administration, 2004a).
4
An ad hoc expert Committee was assembled by LSRO to guide the review (Appendix A).
5
Although LSRO had access to specific databases of IDRs related to ephedrine alkaloid-containing dietary
supplements, neither the Committee nor LSRO made any evaluation or conclusion of the safety or efficacy of
ephedra products. No attempts were made to identify specific signals. Findings and conclusions on the safety of
dietary supplement products are outside the scope of work described in this report. This report does not support any
conclusions on the safety of dietary supplement products.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Information on pre-existing medical condition
Information on concomitant exposure to prescription and/or nonprescription medications
Information on concomitant exposure to other dietary supplements
Information on why the product was used
Batch/lot number
Daily dose of bioactive ingredients
Time to onset from initial use to the adverse event
Rechallenge information
Geographic location (e.g., Zip Code)
FINDINGS
Principal findings of the LSRO Phase I review of individual data records
It should be noted that the MET records represented information collected, in most cases, during a single telephone
contact for the purpose of customer service and these records were not intended to support an adverse event
reporting system. In contrast, FDA records represented information collected in an initial contact and in follow up
by FDA field representatives for the purpose of surveillance.
Overall, the records collected by MET were qualitatively less informative than those collected by FDA and therefore
pose a greater challenge for application in public health-related analyses (Appendix B):
• Signal detection. Both MET and FDA sample sets of IDRs had sufficient information to permit signal
detection.
• Signal prioritization. A substantially greater number of IDRs in the FDA data set had sufficient
information to permit signal prioritization because a relatively high percentage of reporters were health care
professionals and because most MET IDRs lacked follow up and physician evaluation.
• Signal evaluation. Similarly, a greater number of IDRs in the FDA sample set supplied information useful
for signal evaluation than did the MET set. This information was likely to have been available at the time
of the initial call for many MET cases.
Although FDA records contained more information useful for signal generation than did MET records, the FDA set
could have been even more useful had a greater portion of FDA IDRs included important information to calculate
exposure, noted batch numbers, and recorded use of concomitant dietary supplements.
The limitations in signal generation for these sample sets (e.g., lack of batch numbers) stemmed in part from
systematic problems in the methods used for initial data collection. For the MET set, which as previously noted,
was not intended to be a postmarketing surveillance system, this was coupled with a lack of follow up to obtain
outcome information. Despite these limitations, it is likely that signals worthy of further investigation would arise
from such data sets if they were developed in the context of a continuing surveillance program that has the requisite
functional components detailed in Section II.
Principal findings of the LSRO Phase II review of existing postmarketing surveillance programs:
•
•
•
•
•
Although there is no statutory regulation requiring postmarketing surveillance for dietary supplements, a
properly designed and implemented surveillance program could be an important component of a proactive
risk management plan.
An adverse event monitoring program for dietary supplements will share many of the challenges of other
such surveillance systems in terms of specificity, amount of data, completeness, limitations with analysis,
and interpretation of findings.
Current technologies make it feasible to implement quality systems, but manufacturers may need qualified
operators and further guidance to ensure they are using the best possible techniques for analysis.
When signals arise, especially if at-risk populations have not been adequately studied, additional
information beyond routine postmarket reporting may be needed to determine if the signal is real and to
determine the appropriate actions to be taken.
LSRO concludes that it is not appropriate to recommend adoption of the methods of surveillance developed
for medications (prescription and nonprescription) for surveillance of dietary supplements. A surveillance
system for dietary supplements must take into consideration (1) the absence of a statutory requirement to
collect and report these data, and (2) less reliance on reporting by medical professionals.
•
Surveillance programs in the private sector will complement federal efforts to monitor adverse events
associated with dietary supplements. No one system will suffice for postmarketing surveillance of dietary
supplements. All likely avenues that product users and health care professionals use to request or convey
information should be viewed as complementary. All should be cultivated and combined for maximal
detection and analysis of product-event relationships with the goal of preventing future problems and
protecting the public health.
Summary of principal findings
The Committee found that properly designed and implemented postmarketing surveillance programs for dietary
supplements could contribute to public health through the detection and evaluation of potential supplement-event
relationships. A postmarketing surveillance program can complement quality control in manufacturing and extend
the tools available for effective risk management. Potential product problems, such as a defective product lot, that
are detected through a postmarketing surveillance program can be rapidly investigated and addressed. In this way,
product quality is controlled. Postmarketing surveillance programs can increase the confidence of the public that
health-related problems associated with the use of these products will be identified quickly and addressed
effectively. Consumer awareness of the manufacturer’s commitment to product quality is thus enhanced. In
addition, such a system can help protect reputable products from false allegations.
SUMMARY OF RECOMMENDATIONS
LSRO’s recommendations are intended to be broadly applicable to dietary supplements and are not tailored to any
one particular type of supplement or manufacturer.
General recommendation for postmarketing surveillance programs
In general, LSRO recommends that surveillance programs for dietary supplements should be:
• Practical. To be practical, the postmarketing surveillance program should be manageable, professional,
efficient, and cost effective.
• Active. To be active, the program should be utilized by those with unexpected and/or serious adverse
events, have the capacity to produce timely signals, and be responsive to product users and issues in need
of resolution.
• Auditable. To be auditable, the program should have adequate standard operating procedures, retain
documented accounts, and have measures of performance that are examined and verified through quality
assurance and oversight.
• Connectable. To be connectable, the surveillance program should be able to move confidential IDR
information securely between designated databases.
Minimal elements for a postmarketing surveillance program
The following are intended as minimal elements in the design of a postmarketing surveillance program tailored to
the dietary supplement industry. Should an individual company or the industry as a whole decide to voluntarily
implement such a program, LSRO recommends these minimal elements for an effective program.
Reporting and intake
In postmarketing surveillance systems, the term “adverse event” is used whether or not the reported event can be
attributed to the use of a product. Hence a causal relationship does not have to be established between the product
and the event for the reported data to be useful for signal detection and analysis. The surveillance program should
have the flexibility to collect incoming data from different input mechanisms. Processes for submission of adverse
event experiences should be user-friendly, simple, and concise. The individual submitting an adverse event should
be directed to an appropriate resource that is available 24 hours a day, 365 days a year. Company representatives
should be trained to ensure that they collect consistent and complete information during each encounter and
maintain the confidentiality of the reporter and product user.
Creation of individual data records and databases
A standardized questionnaire should be developed that facilitates the collection of data that are useful for detecting,
prioritizing, and evaluating adverse event signals. The data collection procedures should be designed to collect as
much information as possible during the first encounter. Follow up is conducted to obtain additional information
relevant for signal generation. Qualified medical professionals should supervise the medical evaluation of IDRs. A
standardized data dictionary should be used to facilitate consistent and accurate coding and data entry. The system
should process incoming data and IDRs in a consistent and accurate manner
The status of the IDR with regard to its completeness during processing and ongoing activities should be readily
evident (e.g., physician to return call and supply diagnosis). Source materials (e.g., paper materials, computer files)
should be securely stored and readily retrieved as needed. Information imported and exported between databases of
IDRs should be accurate and kept confidential.
Signaling
Signaling procedures should be designed and managed by professionals possessing the requisite technical
knowledge and experience to effectively perform the operational tasks and provide intelligent analysis. The review
and analysis of IDRs should be timely so that notices of potential signals are rapidly disseminated to those who can
minimize product risk, especially when serious risks are identified. Only those IDRs with product-events or
outcomes that are judged to be serious and/or unexpected should be considered in signal evaluation.
Appropriate statistical methods should be applied that include analysis of changes in the pattern of reporting rates.
Once a potential signal surfaces from disproportional type analyses, the product-event combination should be
examined in the context of the quality, quantity, and dispersion of previous IDRs (i.e., case series). The need for
additional studies to investigate potential signals should be determined, and sufficient resources should be allocated
to ensure adequate investigation.
Issues resolution and risk management
Issues resolution requires that a company have a forum whereby professionals (e.g., experts in toxicology, product
development, and surveillance) can come together to integrate and discuss all existing relevant information that
pertain to a safety signal. Hence, companies should provide an organizational procedure by which signals as well as
other relevant information (e.g., studies, medical literature) can be formulated into specific issues. Each issue that is
identified should then be formally assessed in terms of its potential public health impact.
We recommend in systems that are developed that company executives responsible for safety oversight be informed
of new, serious adverse events related to their product and take appropriate action to protect public health. Although
there is no legal requirement for industry to notify FDA, in a system going forward, LSRO advises that FDA should
also be informed of new, serious adverse events related to the product. Outreach should be adequate to inform
product users and others about new health-related findings associated with product use.
Quality assurance
A quality assurance program should be developed to ensure that the surveillance program is functioning to obtain
high quality data and that data are processed in a timely manner and without error, distortion, or disclosure of
confidential data. Objective performance measures should accompany the surveillance program.
Recommended enhancements
In addition to the minimal elements, the evidence suggests an optimal postmarketing surveillance program for
dietary supplements will have some or all of the following enhancements:
• Technological infrastructure. Surveillance programs will benefit from direct, immediate, and electronic
data entry with built-in logic and automated signal generation. To implement this enhancement, the dietary
supplement industry should collaborate with experts in key emerging technologies and statistics related to
automated surveillance systems.
• Inter-product databases of IDRs. A voluntary, industry-wide, and multi-product system would
strengthen statistical power to detect adverse events, especially those associated with a specific ingredient
•
•
•
or involving at-risk populations. This type of system would be particularly beneficial for manufacturers
that produce products that are marketed under several different labels. An inter-product system can be
constructed to ensure the protection of proprietary information and provide an individual company with
autonomy over their customer service and resolution of issues.
Reference databases. Product identification and coding will be improved by establishing a product label
database containing up-to-date, accurate, and complete dietary supplement information. This could be
made even more effective if employed in conjunction with ingredient reference databases (i.e., contact
information for sources, monographs on pharmacologically active substances, methods for chemical
analysis).
Oversight. Postmarketing surveillance programs that are monitored by an independent body of experts will
have greater credibility than programs that do not have such oversight. An independent body of experts
can ensure quality, review outcome measures of success, and identify areas for improvement.
Research. An effective program will identify and minimize the barriers to submitting an account of an
adverse event associated with use of dietary supplements. A thorough review of the literature should be
conducted to help investigate potential signals.
I. INTRODUCTION
Dietary supplements are used in the United States (and other countries) with increasing frequency as individuals
seek to improve their nutrition and health. Supplements include substances such as vitamins, minerals, botanicals,
and amino acids. Although dietary supplements may provide health benefits to product users, they may also pose
health risks. Because the U.S. Food and Drug Administration (FDA) does not have authority to require that dietary
supplements undergo premarket approval for safety and efficacy, adverse events associated with dietary
supplements may be detected only after a product has entered the marketplace. In the case of dietary supplements
and other food-related products that do not require premarket approval, an effective system to collect and analyze
information on adverse events (including adverse reactions) can contribute to the protection of public health.
I.1 DIETARY SUPPLEMENTS AND KEY REGULATIONS
I.1.1 Dietary supplements
Dietary supplements are used to confer health benefits (Appendix C). Historically, dietary supplements were used
in medicine for providing protein, calories, vitamins, and minerals to the ill and malnourished to prevent or correct
nutritional deficiencies. Examples of nutrients supplemented to produce recognized benefits are calcium (to
increase bone density) and folic acid (to reduce the risk of neural tube defects) (U.S. Congress, 1994).
There may be 1500 or more manufacturing and repacking facilities for dietary supplements in the United States
(U.S. General Accounting Office, 2001). They produce approximately 29,000 unique formulations that are packed
into more than 75,000 distinctly labeled products in the market (U.S. Food and Drug Administration, 2001). Dietary
supplements are readily available via the Internet, television infomercials, mail order, multilevel marketing firms,
supermarkets, and other retail outlets. U.S. sales of all dietary supplements reached $18 billion in 2002 (National
Nutritional Foods Association, 2003). FDA expects the dietary supplement industry to grow at a rate of 12% to 14%
per year (U.S. Food and Drug Administration, 2001).
I.1.2 Product users
Dietary supplements are used by a broad spectrum of the U.S. population (Appendix C). By some estimates, onethird of adults take supplements every day or nearly every day (Commission on Dietary Supplement Labels, 1997).
Approximately 40% of 33,905 participants in a nationwide survey reported taking dietary supplements (Ervin et al.,
1999). Among women, 76% of those who were pregnant or lactating reported taking a dietary supplement, typically
a prenatal vitamin and/or iron supplement (Ervin et al., 1999). The prevalence of dietary supplement use among
children by age group ranged from 25% (12-19 years) to 48% (3-5 years) (Ervin et al., 1999). For both children and
adults, the most commonly consumed dietary supplement was a multivitamin with or without minerals (Ervin et al.,
1999; Stang et al., 2000).
Use of dietary supplements tends to increase with age, years of education, and income (Ervin et al., 1999; Foote et
al., 2003). Use of dietary supplements varies by gender; women are more likely to consume supplements than men
(Ervin et al., 1999). Healthy lifestyle characteristics are also associated with the use of dietary supplements (Foote
et al., 2003; Greger, 2001; Hoggatt et al., 2002; Lyle et al., 1998; Neuhouser, 2003). However, a substantial portion
of chronically ill adults (Patterson et al., 2002) and children (Neuhouser et al., 2001) use dietary supplements as
adjuncts (not substitutes) for conventional medical therapy.
I.1.3 Key product regulations
I.1.3.1 Dietary Supplement Health and Education Act
FDA regulates dietary supplements by enforcing the Federal Food, Drug, and Cosmetic Act as amended by the
Dietary Supplement Health and Education Act (DSHEA), in that dietary supplements are regulated as a subcategory
of food (U.S. Congress, 1994). According to DSHEA, it is the responsibility of the manufacturer6 to make sure their
products in the market are safe for the product’s intended use. Under this law, FDA is responsible for taking action
against any dietary supplement product in the market if the product presents an unacceptable health risk or has false
or misleading labeling (Crawford, 2002).
I.1.3.2 Product labeling
FDA requires that certain information be present on a dietary supplement label, including a descriptive name of the
product, a statement indicating that the product is a dietary supplement, and the name and place of business of the
manufacturer, packer, or distributor (Appendix D). In addition, each dietary supplement must have a "Supplement
Facts" panel (U.S. Food and Drug Administration, 1999b), which identifies each dietary ingredient contained in the
product. The label must also contain information on the suggested serving size (U.S. Food and Drug
Administration, 1999b). The product is considered misbranded if the label bears false or misleading information or
does not bear the required information (U.S. Congress, 1994).
I.1.3.3 Good manufacturing practice
Similar to medications and other food products, dietary supplements are prone to quality control problems, which
can pose potential health risks. One way FDA has addressed quality control problems is to promulgate “good
manufacturing practice” specific to each product category that FDA regulates. Recently, FDA proposed “Current
Good Manufacturing Practice in Manufacturing, Packing, or Holding Dietary Ingredients and Dietary Supplements”
(CGMP) (U.S. Food and Drug Administration, 2003a).
The proposed CGMP rule is intended to introduce widely accepted standards into the dietary supplement industry.
Specifically, CGMP is intended to help prevent unsanitary manufacturing, which predisposes products to
contamination, and inconsistent product batches that differ from labeled specifications. It is generally believed that
the use of CGMP will lead to consistent identification, purity, quality, strength, and composition of finished
products for sale, and will ensure that the composition is accurately represented on the product label.
Adverse events associated with the use of dietary supplements can occur despite adherence to CGMP. In some
cases, adverse events happen for reasons that are unrelated to use of the product. Other health problems come about
for a multitude of reasons including, but not limited to, excessive intake, drug-supplement interactions,
hypersensitivity, and errors or accidents that lead to contamination, spoilage, or enhanced potency. Postmarketing
surveillance programs assist manufacturers in recognizing and addressing signals of potential problems. However,
there is no federal guidance or mandate for manufacturers and producers of foods or dietary supplements to collect,
document, or evaluate health-related consumer complaints associated with the use of their products.
I.2 PRODUCT RISK MANAGEMENT AND POSTMARKETING SURVEILLANCE PROGRAMS
I.2.1 Product risk management
Protection of product user health and well-being is of utmost importance not only to public health agencies like FDA
but also to manufacturers and retailers of products for human consumption. Manufacturers that identify potential
adverse events associated with their products and assess the risk to product users may be better able to reduce risk
6
The term “manufacturer” is used throughout this report to indicate the company that produces, distributes and/or
markets dietary supplement products, and is not meant to be exclusionary.
associated with these products. These activities are components of a process referred to as “product risk
management.” Product risk management encompasses all safety activities from product development to
manufacture to postmarket. FDA views risk management as a continuous process of: (1) learning about and
interpreting a product’s benefits and risks; (2) designing and implementing interventions to minimize a product’s
risks; (3) evaluating interventions in light of new knowledge that is acquired over time; and (4) revising
interventions when needed (U.S. Food and Drug Administration, 2003c).
A theoretical risk management cycle for dietary supplements is presented in Figure 1-1. Risk management begins in
the premarket phase with the initial product conception and continues through research and development to
production and test marketing. As the figure demonstrates, risk management continues throughout marketing, and
often must consider the disposition of wastes, residues, or by-products from making and/or using the product. A
component of risk management is the collection and evaluation of various types of feedback that are sent to the
manufacturer during the postmarket phase. The collection of these unsolicited, spontaneous complaints and
feedback is termed “passive surveillance.” Manufacturers may also choose to conduct active surveillance to
research and gather additional information about their product. The findings of postmarketing surveillance are
evaluated in order for the manufacturer to decide whether an intervention is needed to correct product defects or
misuse, or other problems. Interventions may include product withdrawal from the market, labeling changes, and
other means of declaring the risk of adverse events and identifying any specific conditions associated with that risk.
Purchasing and Preparation (suppliers)
Production and Packaging (manufacturers, processors, packers)
Storage and Distribution (distributors)
Sales and Advertising (brokers and retailers)
Complaints and Other Feedback (product users, HCP, others)
Passive Surveillance (manufacturer)
Non-health
Related Complaints
Postmarket
Regulatory Authorities
e.g., DSHEA, CGMP, other
Health Related Events
Active Surveillance (manufacturer)
e.g., literature review, epidemiological
studies, survey, registry, clinical trials
Intervention (manufacturer)
e.g., labeling changes, product withdrawal, other
Premarket
Review and Evaluation
(manufacturer)
Figure 1-1. Theoretical risk management cycle for dietary supplement products. CGMP: Current Good
Manufacturing Practice in Manufacturing, Packing, or Holding Dietary Ingredients and Dietary Supplements;
DSHEA: Dietary Supplement Health and Education Act; HCP: Health care professional.
I.2.2 Postmarketing surveillance
The focus of this Life Sciences Research Office (LSRO) report is the postmarketing component of risk management.
Specifically, recommendations are made for the desired features of an adverse event monitoring program for passive
surveillance of dietary supplements already in the market.
I.2.2.1 Adverse event
Due to the lack of standard surveillance terminology in the dietary supplement industry, many readers may not be
familiar with terminology used in postmarketing surveillance. The following section describes surveillance
programs used by other industries and organizations and defines related terms. In addition, a glossary and list of
acronyms used in this report are provided in Appendix H. The definitions provided were adopted for the purposes
of this report and may differ from definitions presented by others (e.g., the term “serious”).
Because the process of postmarketing surveillance begins with recognizing and submitting an adverse event, we
need to clarify what is meant by “adverse event.” An adverse event is any undesirable health-related sign or
symptom that is detected in an exposed individual after use of a product. Accounts of adverse events are to be
distinguished from other types of complaints that refer to products (e.g., cap not sealed, contaminant in bottle) or
other complaints (lack of anticipated effect). The description of a single adverse event may vary widely depending
on the time-course of the event, the stage at which the event was submitted, and the type of individual describing the
event. For example, a product user might complain of experiencing sweating and an upset stomach early in the
time-course of the event history. In contrast, a physician might report the culmination of several health-related signs
and symptoms as a final diagnosis (e.g., myocardial infarction).
Although this has not been a standard in the dietary supplement industry, other postmarketing systems consider that
an adverse event may or may not be an adverse reaction to the product; adverse reactions imply causality. Actual
adverse reactions can be caused by product defects that might affect product quality, purity, strength, and/or
composition. Reactions might also be caused by misuse of product (e.g., dose related), result from interactions of
the product with other products (e.g., medications, foods or other dietary supplements), or be related to specific
environmental, physical, or medical conditions. Allergic or idiosyncratic reactions occur in only a minority of
product users and are typically unpredictable (Meyboom et al., 1997). In postmarketing surveillance systems, the
term “adverse event” is used whether or not the reported event can be attributed to the use of a product (Clark et al.,
2001). Hence a causal relationship does not have to be established between the product and the event for the
reported data to be useful for signal generation and analysis.
I.2.2.2 Surveillance programs for dietary supplements
This LSRO report provides background information and guidance for the effective use of postmarketing
surveillance. Unique to this report is the discussion of the design features of adverse event monitoring programs for
dietary supplements. If implemented, such programs should better inform risk management choices by
manufacturers and by product users, and thereby reduce potential health risks from dietary supplements. For those
companies already conducting postmarketing surveillance, specific recommendations for improving the
effectiveness of these programs are also provided.
In this report, several examples of postmarketing surveillance of foods, medications, and other products are
reviewed (Appendices E and F). The positive characteristics of such systems are highlighted to provide guidance
for effective adverse event monitoring programs for dietary supplements. This report makes many references to
FDA documents on food and drug products. These references to laws and documents present overarching
experiences that are relevant to surveillance programs in general, not necessarily specifying dietary supplements,
and are not intended to imply that similar laws should pertain to dietary supplements.
By way of an in-depth example, accounts of adverse events associated with ephedrine alkaloid-containing dietary
supplements that were collected by Metabolife International, Inc. and FDA were reviewed and qualitatively
compared (Section III). These records were reviewed to determine whether they were useful for generating signals
of potential product problems. Neither the Committee7 nor LSRO made any evaluation or conclusion concerning
the safety or efficacy of ephedra products. Findings on the safety of dietary supplement products are outside the
scope of work described in this report.
LSRO endorses the use of a postmarketing surveillance program, but not as the primary means of safety evaluations
for dietary supplements. The biopharmaceutical industry has long appreciated that postmarketing reporting systems
are a starting place, not an ending point for the surveillance process. The Committee observed that even in the case
of optimal data collection and analysis, adverse event monitoring programs are generally not sufficient to
demonstrate causality. In selecting the desired features of a future adverse event monitoring program for dietary
supplements, LSRO recognizes that the program should be capable of identifying signals of potential product
problems and guiding the selection of signals for which further research is warranted. Further research might
consist of controlled biomedical laboratory studies, clinical trials, epidemiological studies, and/or other types of
7
An ad hoc expert Committee was assembled by LSRO to guide the review (Appendix A).
- 10 The distribution or electronic posting of this PDF file is strictly prohibited without
postmarketing studies that provide the best means for evaluating and reducing the risks associated with the use of
dietary supplements.
II. GENERAL SURVEILLANCE PROGRAMS
This section provides an overview of postmarketing risk management and general features of surveillance programs
used in food and drug industries.
II.1 POSTMARKETING RISK MANAGEMENT
The postmarketing risk profile of a product extends to all aspects of its manufacturing, distribution, and use that
could pose a threat to health. The management of postmarketing risk involves the establishment of processes aimed
at comprehensive data gathering, scientific assessment, and dissemination of information pertinent to adverse events
and product-related problems. The systematic capture and assessment of such information establishes a logical
sequence of activities that allow monitoring systems to link with corrective actions (i.e, detection, assessment,
understanding, and prevention of future problems) (U.S. Food and Drug Administration, 2003d).
Postmarketing risk management programs are based on gathering and integrating product-specific health-related
data (Hennessy, 1998). These information-oriented tasks are generally referred to as “postmarketing surveillance.”
Typical information sources that contribute to postmarketing surveillance are unpublished anecdotal complaints of
health-related events, case reports from the medical-scientific literature, and human, animal, and in vitro studies.
Known Side Effects
Medication Errors
Unavoidable
Product Quality Defects
Avoidable
Preventable
Adverse Events
Injury or Death
Remaining Uncertainties
• Unexpected side effects
• Unstudied uses
• Unstudied populations
Figure 2-1. Sources of risk from products regulated by the Food and Drug Administration. Reprinted from U.S.
Food and Drug Administration (1999a).
II.1.1 Postmarketing sources of risk
For virtually all health care products, some degree of postmarketing risk cannot be avoided. Once a dietary
supplement, food, or medication is marketed for use in the general population, it is inevitable that product exposure
will occur in individuals with acute and chronic diseases and in individuals concomitantly using other health care
products. Hence, with sufficiently wide use, adverse events that are associated with a product can arise as a
consequence of product dynamics in particular population subgroups. Problems during manufacturing and
distribution can also lead to contamination, spoilage, or mislabeling, or to changes in product purity, strength, or
composition. Figure 2-1 provides an illustration of the various sources of risk that can arise in the postmarketing
environment (U.S. Food and Drug Administration, 1999a). Preventable adverse effects include those aspects of a
product’s side effect profile that are amenable to intervention, as well as errors in product use and product failure
(Leape, 2002). Postmarketing surveillance is an approach through which rapid detection of, and reaction to, these
different kinds of preventable problems can be accomplished.
Figure 2-2. Comparison of the distributions of adverse events, grouped by system organ class, for the prescription
drug metoprolol reported during premarket studies before 1975 and during postmarketing surveillance from 1975 to
1984. Modified with permission of Adis International from Wallander (1993).
Premarket knowledge is rarely sufficient for manufacturers, regulatory authorities, and others to recognize the full
spectrum of adverse events that are attributable to a product. Product use in large populations provides increased
opportunity to identify interactions between various agents (Avorn, 1990). This is particularly true for patient
subpopulations who are at higher risk for adverse events, such as the elderly and patients with chronic diseases
(Butchko & Stargel, 2001; Faich et al., 1988). Hence, once a new dietary supplement, food, or medication is
introduced into the marketplace, adverse events may emerge that were not previously reported.
An example of the difference in the distribution of adverse events reported during premarket clinical trial studies
compared with events reported during postmarketing surveillance is illustrated in Figure 2-2 (Wallander, 1993).
Unlike premarket programs, postmarketing surveillance derives adverse event data from diverse populations that are
receiving new products in “real world” consumption scenarios and extreme conditions of use. A major advantage of
such passive surveillance when conducted nationwide is the detection of infrequent, idiosyncratic responses that
may not be predictable from premarket experience (Butchko et al., 1996).
II.1.2 Individual data record-based postmarketing surveillance
For most health care products, postmarketing surveillance programs are founded on the systematic evaluation of
individual product user experiences that become available to product managers. Under the international guidance
documents that govern much of health care product surveillance, this kind of information is structured into standard
displays, which are referenced in this Life Sciences Research Office (LSRO) report as “individual data records”
(IDRs). The formatting of IDRs varies across different health care environments, depending on the functional
objectives and data availability that apply to that sector. The Committee used the abbreviation “IDR” whenever
individual product user experiences were being discussed in order to distinguish this kind of data from single or
aggregate records of any other kind, such as single complaints involving product defects, or aggregate product user
information that gives rise to summary findings.
Table 2-1. General functional steps in postmarketing surveillance programs
Recognition and reporting of adverse events in individual product users
• Recognition of adverse events by reporters
• Forwarding of adverse events by reporters to surveillance programs
• Intake of adverse events
Creation of individual data records by surveillance programs
• Identification of data elements for collection and data entry
• Data entry and coding of individual data records
• Follow up of individual data records
• Archiving of source material
Building a relational database of individual data records
• Incorporation of individual data records into databases
• Exchange of information between databases used for individual data records
and databases of product-focused complaints (e.g., defects)
• Creation of inter-product databases of individual data records
Signaling from databases of individual data records
• Signal detection
• Signal prioritization
• Signal evaluation
Signaling from other information sources (published or unpublished)
• In vitro and animal studies
• Human studies and other human study environments
• Information based on product-class effects and analogous products
• Other data sources
• Response to reporters
• Labeling changes and other product information changes
• Other risk management activities
IDR-based postmarketing surveillance systems first came into use in the early 1960s, and have been an important
constituent of postmarketing risk management programs since. The concepts of IDR postmarketing databases and
signaling are usually credited to Professor David Finney of the University of Edinburgh, who first proposed their
implementation in 1963 in response to the thalidomide tragedy (Clark et al., 2001; U.S. Food and Drug
Administration, 1991b). Health care surveillance programs continue to emphasize this kind of early warning
system, in large part because this approach has been shown to consistently produce useful material and is cost
effective. Modern adverse event monitoring programs embed the receipt, retention, retrieval, and review of
postmarketing IDRs in processes and computing systems that promote efficiency, and that provide scientific and
analytic perspective (Allan, 1992a).
In order to derive a general description of IDR-based surveillance, the Committee closely examined surveillance
problems and issues in several areas (i.e., biopharmaceuticals, foods, over-the-counter products, poison control).
For a descriptive summary of these examples, the reader is referred to Appendices E and F. These examples,
together with experience set forth elsewhere in this report, allowed the Committee to propose an outline of
information processing steps for IDRs that should be applicable to IDR-based postmarketing surveillance schemes
(Table 2-1). Because the Committee believes that this outline of surveillance steps simplifies the Committee’s
recommendations, it was used to organize sections and appendices wherever IDR surveillance measures were the
primary subject of discussion.
II.2 GENERAL FEATURES OF SURVEILLANCE PROGRAMS
In this LSRO report, the term “surveillance program” refers to all personnel, processes, and computer resources that
are needed to carry out IDR-based postmarketing surveillance. Surveillance programs establish standard operating
procedures by which the collection, receipt, analysis, interpretation, and documentation of IDRs can be consistently
achieved. The primary purpose of a surveillance program is to identify potential product-related health problems
that necessitate rapid dissemination of information or other actions. In an operational sense, a surveillance program
allows the adverse events that are reported in connection with a particular product to be scrutinized for
disproportions or unusual cases that might represent problems, a process known as signaling. Allan (1992c)
summarized four traditional elements of postmarketing surveillance (i.e., receipt, retention, retrieval, and review).
II.2.1 Recognition and reporting of adverse events in individual product users8
Adverse events can be either causal occurrences or unrelated events that are recognized and submitted by chance
because of a purely temporal association. Typically, only a fraction of the adverse events that occur following the
use of a health care product are forwarded to a surveillance program. Siev (1999) diagrammed this process in the
context of vaccine products with respect to both under- and over-reporting (Figure 2-3). In the Siev model, underreporting occurs for two major reasons: (1) there is a lack of a perceived association between the product and the
adverse event, or (2) there is a failure to report an association that has been recognized by the reporter.
Consumers of dietary supplements may be less likely to reveal their use of dietary supplements to their health care
professional than their use of medications, especially regarding the more unconventional types of supplements
(Walker, 2000). This represents a unique reason for under-reporting among users of dietary supplements.
Siev (1999) noted that over-reporting also occurs in surveillance programs because spurious reports of adverse
events that did not actually occur are included as a portion of the IDRs that are forwarded for any given product.
II.2.1.1 Recognition of adverse events by reporters
In order for adverse events to be forwarded, potential reporters must associate a product with an event. Recognition
of this association can be challenging because health-related effects are usually caused by a variety of factors such
as concomitant exposure to foods and drugs, pre-existing diseases, and environmental conditions. Thus a health
event may be attributed to another factor and not submitted, even if it was due fully or in part to exposure to a health
care product (U.S. Food and Drug Administration, 1998). Hence, the product user’s (and/or observer’s, e.g.,
physician’s) knowledge and bias have a great deal to do with the recognition of signs and symptoms and the
subsequent association of these events with a suspect exposure.
8
For the purposes of this LSRO publication, the term “reporting” includes the action of submitting a voluntary
consumer health-related complaint to industry.
There are several ways in which the recognition of potentially attributable events can be improved. Product labeling
of adverse events or publicity of important health-related events can inform product users and health care
professionals of possible symptoms, thereby improving the recognition process. If the time to onset between
product exposure and an event is relatively rapid (e.g., symptoms occur within an hour or two after exposure) the
product user may be more likely to attribute the symptoms to the product than if the time to onset is delayed.
Additionally, effective communication channels between health care professionals and their patients improve the
chances that a practitioner will forward adverse experiences to the manufacturer or regulatory authorities.
Figure 2-3. Vaccine related events and coincidental events contribute to adverse events occurring after
vaccination. Only a fraction of adverse events are observed, and of those only a fraction are reported. AER:
Individual data records. Reprinted from Siev (1999).
II.2.1.2 Forwarding of adverse events by reporters to surveillance programs
Once a product-health event is recognized, the potential reporter must be motivated to forward the information to a
surveillance program. Individuals with symptoms that are severe, dramatic in nature, or cause great inconvenience
may be particularly motivated to provide such experiences to surveillance specialists (Walker, 2000).
An adverse event experience will not be submitted if the potential reporter decides that the event was not worthy of
further action or if the reporter has concerns over the possible disclosure of confidential information. Also, a
product user experience may not be submitted if there is a lack of understanding of how to submit the event or an
inability to access the monitoring system. Hence, the knowledge of a mechanism for submission of a user
experience and the ease of submission will influence whether events are reported.
Effective surveillance programs have well-defined information portals through which relevant data are collected.
Electronic communication tools, such as email and web pages can facilitate communication that might not be
received by telephone because of foreign language, hearing impairment, or other barriers.
In manufacturer-based programs, allowance should be made for incidental adverse event information that comes to
attention when product users or health care professionals contact the manufacturer for other reasons, such as to
obtain product information.
II.2.1.3 Intake of adverse events
In the postmarketing environment, patient-level complaints are typically forwarded to manufacturers and regulatory
authorities by product users, the relatives of product users, health care professionals, or other kinds of professionals
(e.g., legal reports, reports from health care administrators). Manufacturers receive this feedback through a variety
of communication channels, such as telephone, e-mail, mail, facsimile, and Internet web sites. The traditional
process for the acquisition of new IDRs is to document the receipt of the complaint in a log, write up the information
in a standardized format, code key data fields, enter information into a database, and verify accuracy. This
information then undergoes professional review (Figure 2-4) (Faich, 2003). Further investigation of the product user
experience may be initiated if warranted.
For those events that are submitted, the goal of data receipt is to obtain the maximal amount of information in the
shortest possible time and be as accurate, relevant, and complete as possible (Allan, 1992c). Because the intake
process involves data that is heterogeneous and inconsistent, standardized instruments facilitate data collection.
Intake forms and procedures have two primary functional attributes: (1) rapid intake of the most important
information during telephone interactions, and (2) the structuring of collected data for subsequent data entry.
II.2.2 Creation of individual data records by surveillance programs
An IDR is a standardized presentation of one or more adverse events that have been experienced by a single product
user, along with associated patient data and product information. The IDR is the basic unit of information upon
which surveillance programs are built.
Paper Based Process
Incoming Call,
Facsimile, Mail,
or Electronic
Form
Write Up
Report
Log
Enter Data
Code
Electronic Process
Incoming Call,
Facsimile,
Mail, or
Electronic
Form
Automated logging
and notification of
receipt
Computer-assisted
write-up, coding, and
data entry
Review
Verify
Verify
Review
Figure 2-4. Processing of an individual product user experience using the paper method versus new technology.
Modified and reprinted with permission from Faich (2003).
II.2.2.1 Identification of data elements for collection and data entry
IDRs are of most value to the surveillance program when the reporter provides all relevant information needed for
signaling. Information appropriate to obtain for an IDR would logically include the adverse event, administrative
data (e.g., date of IDR), demographic data (e.g., age, gender), and product information (e.g., dose, route of
administration)( U.S. Food and Drug Administration, 2004c). The importance of quality data collection has been
emphasized by Nelson et al. (2002), who suggested that under-reporting of spontaneous events is not a limitation to
signal generation as long as a sufficient amount of high quality IDRs are submitted.
The Committee studied those aspects of IDRs that it believes are necessary for effective surveillance programs.
These findings are discussed in greater detail in Section IV.
IDRs can contain information that reflects a causal relationship between a product and adverse events. However,
based on IDRs alone, it is unusual to know with certainty whether a given event resulted from use of a specific
product (U.S. Food and Drug Administration, 2003b). Supportive causal evidence that can be obtained through
surveillance programs as part of an individual product user experience includes (U.S. Food and Drug
Administration, 2003b):
• A temporal course in which the time to onset of the adverse event is biologically plausible
• Consistency of the adverse event type with known biological activity and/or other characteristics of the
product (e.g., the kinetics and pharmacodynamics of caffeine could logically contribute to dehydration)
• Recurrence of an event following a second exposure or “challenge” to the product (also called “positive
rechallenge”)
• Evidence that minimizes potential confounding factors
II.2.2.2 Data entry and coding of individual data records
Forms and structured methods are utilized to facilitate standardized and complete acquisition of relevant information
during collection of adverse event data. At the time of data entry, the use of such instruments allows the collected
data to be rapidly entered into a database. Certain key data fields in IDRs are further reduced to codes using a
standard coding language such as the Medical Dictionary for Regulatory Activities (MedDRA®) to assist the
retrieval of cases and to support signaling (International Conference on Harmonisation, 1998).9 The data fields that
are selected for coding may vary by company or product, but usually include adverse events and the reason for
product use. Data entry and coding are structured as a controlled process, which emphasizes quality control
measures to assure database accuracy. Fully entered IDRs are then reviewed by medically qualified individuals.
II.2.2.3 Follow up of individual data records
After intake procedures have been carried out, incomplete IDRs and IDRs of questionable accuracy are further
investigated as needed. For example, follow up may be conducted to: (a) determine whether or not a patient’s
symptoms resolved, (b) document the ultimate outcome of the episode, (c) obtain consent to review medical records,
and (d) verify clinical details. Typically, only a portion of follow-up efforts is successful in obtaining additional,
relevant information (Appendix E and F).
II.2.2.4 Archiving of source material
Source materials should be cataloged and securely archived for future reference as needed (Allan, 1992c). An
efficient archiving system allows the rapid retrieval of case data for in-depth examination.
II.2.3 Building a relational database of individual data records
II.2.3.1 Incorporation of individual data records into databases
Data entry and coding of IDRs are supported by relational database technology that has been adapted to certain
standards. In any given area, there tend to be a small amount of specialized manufacturers of hardware and software
that are used for relational databases of IDRs. Software vendors in the biopharmaceutical industry are currently
adapting their solutions to the International Conference on Harmonisation (ICH) E2B standard, which defines the
data fields for an IDR in the biopharmaceutical industry (International Conference on Harmonisation Expert
Working Group, 2001; U.S. Food and Drug Administration, 2004b). Ultimately, the ICH E2B standard will allow
the transmission of IDRs among interested parties as encrypted electronic files. The implementation and
management of such automated systems requires dedicated administrators and information technology expertise,
which can be partially or completely outsourced.
II.2.3.2 Exchange of information between databases used for individual data records and product-focused
complaints
Manufacturers receive individual product user information that involves two kinds of problems, adverse events and
product-focused complaints (e.g., container defects). These two variants of individual product user records are
9
Many international documents use the British English spelling rather than the American spelling of
“harmonization.”
usually organized into two separate databases. Because adverse event information could be submitted as a productfocused complaint, or vice versa, it is prudent to screen both databases periodically for material that should be
entered into the other database.
II.2.3.3 Creation of inter-product databases of individual data records
Databases that contain multiple types and brands of products have been shown to be useful sources for safety
signals. Multi-product databases allow inter-product analyses to be performed on merged product data, which
allows the refinement of potential signals and the detection of signals that might not otherwise have been identified.
There are several challenges to constructing and utilizing inter-product databases. For instance, depending on how
and when data are transmitted into the system, inter-product databases may be less up-to-date and may yield results
later than intra-product (within a single product) manufacturer administered systems. In addition, methods of
evaluation may need to differ for different products and/or product categories and computer program settings may
need adjustment depending on the product of interest (Bright & Nelson, 2002). Furthermore, if the database is
managed by an organization external to the manufacturer, proprietary and confidential information about products
needs to be protected from disclosure.
II.2.4 Signaling from databases of individual data records
Once databases of IDRs are established, they can be systematically examined for signals. A signal is a potential
relationship between a product and an adverse event that is deserving of further consideration (Clark et al., 2001;
U.S. Food and Drug Administration, 2003d). It is well known that signals of product-event relationships can arise
as a result of chance alone. Because adverse event signals are speculative, and represent clues pointing to a potential
hazard, a product-adverse event relationship should never be assumed to be causal at the outset (Bortnichak et al.,
2001).
There are three sequential stages by which product-adverse event signals are generated and assessed: (1) signal
detection, (2) signal prioritization, and (3) signal evaluation. These stages provide evidence that may or may not
suggest the need for further investigation (Clark et al., 2001). In a general sense, signal detection corresponds to
hypothesis generation from a database, while signal evaluation corresponds to hypothesis strengthening by a
surveillance analyst. Well-designed signaling processes can generate useful signaling arguments that link products
to adverse events. A flow diagram of the steps leading to signal evaluation and the development of a signaling
argument is presented in Figure 2-5.
Side effects that are known to sometimes occur as the result of a product and that are included in the product
label/insert are considered “expected” (International Conference on Harmonisation, 1996). Adverse events that are
not “labeled” are therefore “unexpected.” Signals arising from databases of IDRs can provide perspective on the
following kinds of issues (Clark et al., 2001; U.S. Food and Drug Administration, 2003b):
• New, unexpected adverse events
• An observed increase in the severity or specificity of a labeled event
• An increase in the reporting rate of a labeled event
• New supplement-supplement, supplement-food, or supplement-medication interactions
• Clues to the mechanism of action leading to the adverse event
• Confusion of product users over a product’s name, labeling, packaging, actual use, or potential use
• Factors associated with reporting behavior
II.2.4.1 Signal detection
Signal detection refers to the application of methods to a database of IDRs for the purpose of finding “interesting”
product-adverse event relationships. “Interesting” could mean that a particular type of adverse event is present
disproportionately (quantitative signal detection), or could mean that one or a few cases of an adverse event type
contain potentially important causal evidence (qualitative signal detection). Quantitative signaling methods, which
are based on sets of IDRs, are methods of pattern recognition rather than causality judgments (Kahn, 2003).
Because signals may be detected from any number of product or comparator IDRs, including single, welldocumented IDRs describing an unusual clinical feature or a positive rechallenge time sequence (U.S. Food and
Drug Administration, 2003d), surveillance programs must be able to manage and analyze data both in aggregate and
on a case-by-case basis.
SPONTANEOUS DATA
SORTING
REPORTBASED
STEPS
IDENTIFICATION
CASEBASED
STEPS
Places reports into meaningful
subsets of workable size, based
on content and/or number
Uses an explicit or implicit
statistical model to identify
product-AE pairs of interest
CASE SERIES FORMATION
Applies a case definition and
case series methods to create
an argument in favor of association
CASE SERIES
CHARACTERIZATION
Describes the characteristics and
patterns of product-AE case series,
focusing on reporting, demographics,
risk factors, and possible mechanisms
of action
SPONTANEOUS SIGNALING ARGUMENT
Figure 2-5. Steps in the spontaneous event signaling process from the initial report of data to the formation of a
spontaneous signaling argument. AE: adverse event. Copyright 2001 from Epidemiologic Reviews, by Clark et al.
Reprinted with permission of Oxford University Press.
The simplest method for detecting signals of product-events is crude inspection of lists of spontaneously reported
product-event combinations (Hauben & Zhou, 2003). However, this method is unlikely to detect multiple-variable
cases, and becomes less useful as the number of accumulating IDRs increases and the familiarization of staff with
each IDR declines. Signal detection is database dependent and, to be maximally effective, should be automated
whenever possible.
Many of the quantitative approaches used to generate signals are abstract, theoretical constructs that are difficult to
understand and visualize (U.S. Food and Drug Administration, 2003d; U.S. Food and Drug Administration, 2004c).
They include methods in which the spectrum of adverse events for a product of interest is compared with the
spectrum of adverse events for comparator products (empirical Bayesian techniques and proportional reporting
ratios). They also include methods in which the reported rate of incoming adverse events for a product is compared
with its baseline rate to assess trends (Appendix F).
Early quantitative approaches to signal detection emphasized the value of trend analysis because trend monitoring
was initially thought to be a superior signaling methodology (Wallander, 1993). However, reporting rates are not
true measures of risk because they do not represent actual numbers of product exposures and adverse events, and
may therefore be associated with a high false positive rate in certain circumstances. Consequently, more recent
research in signal detection has focused on inter-product comparisons. Nevertheless, methods that detect increases
in serial reporting rate continue to be employed in health care product surveillance programs because they are useful
in particular problem areas, such as manufacturing errors, and can also signal other phenomena, such as negative
media coverage. According to Meyboom et al. (1997), national spontaneous adverse event reporting systems for
prescription drugs were able to detect events occurring as infrequently as 1/10,000 to 1/50,000 patients.
II.2.4.2 Signal prioritization
The second stage in the signaling process is to sort those IDRs with non-serious, expected product-event
combinations for which no further action is needed from the IDRs with serious or unexpected product-event
combinations. This step may also be automated.
Automated or semi-automated methods during signal detection lead to the generation of a list of many potential
signals, a number of which are likely to be false positive signals. A false positive signal means that a product-event
relationship was identified, but was not found on further scrutiny to be sufficiently “interesting” to warrant an
assessment. Signal prioritization facilitates the signaling process by attempting to screen out the majority of false
positive signals. To do so, it focuses on eliminating product-adverse event relationships that, regardless of signal
strength (e.g., a large disproportion) have minimal medical impact. This can be accomplished by assigning
weighting factors to impact criteria in the IDRs that comprise a safety signal. For example, it would be important to
closely examine a signal that involved multiple fatalities and other serious outcomes, whereas even strong signals
that were entirely non-serious may not merit further consideration.
II.2.4.3 Signal evaluation
Signal evaluation involves the creation of a series of two or more product-exposed cases that associate a similar
event with a product or ingredient (U.S. Food and Drug Administration, 2003b). The foundation for building such a
case series is the collection of quality data that permit confirmation as a case (i.e., what is recorded in the IDR is the
actual diagnosis).
The kind of suspect cases that constitute a detected signal should be clearly and consistently defined and
characterized to facilitate the retrieval of clinically relevant IDRs (U.S. Food and Drug Administration, 2003d).
Hence, a case series defined as aplastic anemia would exclude cases of other types of anemia. Case definitions
specify the minimum quality and diagnostic criteria that are needed to designate IDRs as cases (Allan, 1992c; Clark
et al., 2001). For example, a case definition for acute renal failure might be based on abnormal tests (i.e. creatinine
clearance, elevated blood urea nitrogen), decreased urine production, and other clinical details. The use of
standardized terminology (i.e., MedDRA®) facilitates the proper identification of cases for inclusion in a case series
(U.S. Food and Drug Administration, 2003b).
The formation of a case series permits the simultaneous evaluation of four kinds of information: (1) the number of
cases, (2) the aggregate diagnostic features, (3) the degree of clinical similarity among cases in the series, and (4) the
determination that similar adverse experience has been forwarded from different reporter sources (i.e., the cases
demonstrate independence and repeatability).
Product-adverse event case series are then assessed in terms of patterns and concomitant characteristics. For most
signals, this includes a careful examination for adverse event types and subtypes, clinical features that suggest
product causality (e.g., latency of onset), demographics (e.g., patient subgroups such as the elderly), reporter profile,
dose, and type of product (Clark et al., 2001; Kahn, 2003; U.S. Food and Drug Administration, 1998). An attempt is
also made to assess the case series for the presence of confounding bias (other possible contributors to the adverse
event, such as pre-existing medical conditions). While the presence of confounders in a signal undergoing
evaluation may argue against a monitored product cause, it does not exclude the product from having a contributory
effect (U.S. Food and Drug Administration, 2003d; U.S. Food and Drug Administration, 2004c).
Using case series methodology, adverse events can sometimes be related to important explanatory factors (U.S.
Food and Drug Administration, 2003b):
• Demographics (e.g., by age, gender, or ethnicity)
• Dose effects (e.g., use of greater than labeled doses)
• Duration of exposure
• Concomitant exposure to medications or other supplements
• Other medical conditions
• Changes over the product life-cycle (e.g., lot-to-lot variation or reformulation)
• Indications for use [i.e., events occur related to one indication for use (weight loss) but not another (body
building)]
II.2.5 Signaling from other information sources
Signaling from databases of IDRs obtained through passive surveillance (spontaneous reporting) is useful to identify
signals of potential association, which can then be examined in controlled studies or by other means to further
investigate if there might be a causal link to the product. For some products, it is appropriate to initiate both active
and passive surveillance at the time the product is released into the market, where active surveillance involves the
generation of product user experiences from a structured environment (Appendix E). Additionally, information
derived primarily from non-IDR environments can also provide important signals. These include in vitro studies
(cell culture, cytotoxicity, genotoxicity, tissue models), animal studies (animal toxicology, animal pharmacology),
various human study designs (epidemiological studies, pharmacokinetic studies, clinical trial results, registries,
compassionate use programs, disease management programs), and information based on product-class events and
analogous products (U.S. Food and Drug Administration, 2004c).
II.2.6 Issues resolution and formulation of risk management interventions
During issues resolution, signals that have been identified from an adverse event monitoring program are fully
evaluated using all pertinent data that is available. This can include other sources of data that are consulted and/or
developed, including reviews of the published medical and scientific literature, reports of injury in the lay press,
epidemiological studies, investigations of biologic effects in kinetic studies, the results of controlled clinical trials,
and investigations of the general market experience of other similar products (U.S. Food and Drug Administration,
2003b).
II.2.6.1 Response to reporters
The large majority of reporters that forward experience to surveillance programs for prescription medications are
health care professionals involved with management of an adverse event that occurred in a product user (Appendix
F). Therefore, traditional surveillance programs collect information but tend not to dispense medical advice. As
part of customer service, manufacturers sometimes dispense product information in response to a request.
Furthermore, some manufacturers employ medically trained professionals to respond to current health-events
experienced by product users.
II.2.6.2 Labeling changes and other product information changes
The most common outcome of surveillance programs is an updated revision of the product label that reflects the
evidence (Allan, 1992b). For example, in the case of a dietary supplement containing an ingredient discovered to
interfere with the bioavailability of a drug, the product label might be altered to bear a warning of contraindication
of use with that medication.
II.2.6.3 Other risk management activities
The issues resolution process may identify needed research (e.g., may conclude that a comparison to a reference
population in a formal epidemiological study is necessary) (Avorn, 1990). Such studies, in which the numbers of
individuals exposed to the product and experiencing an adverse event are known, provide more reliable measures of
risk and can place postmarketing IDRs in perspective (Miwa et al., 1997). Further investigation might reveal that a
safety problem was likely to be unrelated to the product (false positive signal) or that the potential product-adverse
event signal was related to a particular product characteristic (e.g., a batch problem or an ingredient) (U.S. Food and
Drug Administration, 2003d). Most signaling programs couple a surveillance program with internal or external
epidemiological expertise in the event that additional study in a controlled environment becomes necessary (U.S.
Food and Drug Administration, 1998).
If issues resolution has concluded that an association between a product and an adverse event is probable, the
manufacturer of the product will need to consider interventions that minimize future risk. This could involve risk
communication (e.g., dissemination of information about new restrictions to distributors, retail outlets, and, through
mass media, to the general public), or, in extreme instances, could involve restrictions on product distribution or
withdrawal from the market (Ajayi et al., 2000). In general, manufacturers with existing postmarketing surveillance
systems can monitor reporting rates or another appropriate measure after intervention for indications that
improvement or resolution of the problem might have occurred. However, a decrease in reporting rate does not
constitute an assurance that a safety issue was resolved.
For issues of adverse events or product defects that involve manufacturing and labeling, FDA’s recall system can be
an effective remedy. From 1990 through 1999, FDA received reports of at least 112 (Class I and Class II) recalls of
dietary supplements (U.S. Food and Drug Administration, 2003a). The majority of these were due to problems
arising from contamination of the product. Other problems included undeclared contents (e.g., sulfites) and toxic
symptoms (e.g., iron poisoning; U.S. Food and Drug Administration, 2003a). For most products regulated by FDA,
product recalls are conducted voluntarily by the manufacturer and distributor after discovering that the product is
defective. In some instances, FDA informs a company of its findings that a product is defective and suggests or
requests a recall (U.S. Food and Drug Administration, 2002). The Federal Food, Drug and Cosmetic Act (U.S.
Congress, 2000) specifically authorizes FDA to initiate recalls as needed for medical devices, human tissue
products, and infant formulas, but does not generally authorize recalls for foods, cosmetics, or dietary supplements.
However, FDA can seek legal action to have these types of products seized and recalled by court order if industry’s
response to the problem is believed to be insufficient to protect the public health.
II.3 SUMMARY OF GENERAL SURVEILLANCE PROGRAMS
IDR-based postmarketing surveillance programs are intended to identify adverse events early during the marketing
of a health care product, and to provide additional information relevant to the adverse event profile of a product.
These surveillance activities are particularly valuable when preventable problems are identified and corrected
through a targeted risk management program. Many of the advantages and challenges of adverse event monitoring
programs have been discussed in this section. These topics will be revisited in greater detail in Section IV in the
context of the design of a surveillance program that is oriented to the dietary supplement industry.
III. RETROSPECTIVE REVIEW OF SAMPLE INDIVIDUAL DATA RECORDS
III.1 INTRODUCTION
The objective of this section is to describe the Life Sciences Research Office (LSRO)’s qualitative comparison of
two sets of individual data records (IDRs) associated with the use of dietary supplements. One set was derived from
IDRs collected by Metabolife International, Inc. (MET; San Diego, CA) on their ephedra-containing products for
purposes of customer service and the other was derived from IDRs collected by the U.S. Food and Drug
Administration (FDA) on MET products and similar ephedrine alkaloid-containing products for purposes of adverse
event monitoring.10 Although manufacturers did not intend the collection and documentation of calls to a hotline as
“postmarketing surveillance,” such a procedure is equated to the preliminary functional steps of postmarketing
surveillance so that the Committee could gain additional insight into the design of possible surveillance systems for
No attempt was made, nor was it the purpose of this review, to identify representative data sets for the broad
category of dietary supplements. Thus, the Committee did not include example IDRs from other product types
regulated by FDA and therefore, no conclusions can or should be drawn from this case study analysis concerning the
characteristics of surveillance programs within the dietary supplement industry or any other industry regulated by
FDA.
The Committee reviewed the type and completeness of the data collected, but did not attempt to interpret the data
for assessing or establishing causality. Neither the Committee nor LSRO made any evaluation or conclusion of the
safety or efficacy of ephedra products. The Committee did not examine whether the use of ephedra caused any
particular adverse event. Therefore, this report does not support any conclusions on the safety of dietary
supplements.
As noted in the preceding section, a surveillance system consists of the following functional steps:
1.
2.
3.
4.
5.
6.
Recognition and reporting of adverse events in individual product users
Creation of IDRs
Building an integrated, relational database of IDRs
10
Effective November 18, 2003, MET suspended the sale of their ephedra-containing products. On December 30,
2003, FDA announced plans to prohibit sales of dietary supplements containing ephedra (U.S. Food and Drug
Administration, 2003e). The final rule prohibiting the sale of ephedrine-alkaloid containing dietary supplements
was issued February 11, 2004 (U.S. Food and Drug Administration, 2004a).
The discussion in this section is organized according to these functional steps, with special emphasis on the fourth
step: signaling. The three sequential stages by which product-adverse event signals are generated and assessed—
signal detection, signal prioritization, and signal evaluation—were each evaluated individually.
III.2 SAMPLE INDIVIDUAL DATA RECORDS
III.2.1 The industry (MET) data set
III.2.1.1 Recognition and reporting of adverse events in individual product users
MET voluntarily instituted a customer service telephone line in 1996 as a way to provide information to interested
customers regarding appropriate usage of its products. Since 1998, product users were directed on the Metabolife
356® label to call a toll free number to seek help to discuss weight loss or related health questions. Generally,
registered nurses answered the telephone line between 7:30 am and 5:00 pm. After hours, callers were instructed to
leave a message to which the company responded the next business day. As noted previously, the manufacturer did
not intend the collection and documentation of calls to a hotline as “postmarketing surveillance.”
The customer service line (i.e., MET Health Information Line) has received over 400,000 calls since its inception.
The majority of calls related to weight loss aspects of Metabolife 356® consumption. Product complaints were also
received, including health-related complaints. In addition to complaints received by telephone, complaints were
received by electronic mail and through the postal service.
III.2.1.2 Creation of individual data records
Initially, documentation of product user complaints was not standardized, but it became more systematic over time.
Because early documentation lacked a structured format, the type of information collected was inconsistent. Also,
early documentation was hand-written, and a portion of it was difficult to read.
The standardized form went through a series of modifications, each increasingly explicit for recording descriptive
data, such as name, age, and height, and information about the health-related complaint. However, even after
standardized forms were utilized, data fields were not consistently filled in (e.g., height was often not obtained).
Most incoming calls in later years used the same one-page form to record product user information (Figure 3-1).
III.2.1.3 Industry (MET) database of individual data records
In 2002, MET shared IDRs for Metabolife 356® with LSRO and FDA (U.S. General Accounting Office, 2003b).
Approximately 14,684 health-related complaints of varying severity were identified among transmissions to the toll
free hotline and/or MET offices (U.S. General Accounting Office, 2003b). A count of unique IDRs would be less
than 14,684 because some transmissions were inadvertently duplicated during documentation (U.S. General
Accounting Office, 2003b). The remaining calls, letters, and emails included testimonials, questions about usage,
and questions about dieting.
Records were released by MET on compact disks and organized by page number in files of assorted information
containing as many as 500 pages each. Specifically, a compact disk dated 7-2-02 contained 25 files and a compact
disk dated 11-7-02 contained 4 files. Approximately 16,000 pages contained in these files included IDRs that were
collected between May 1997 and July 2002 (U.S. General Accounting Office, 2003b). Additional documentation in
the database included in-house call logs tabulated by day and by employee, customer-interaction logs, and requests
for refunds.
The rationale for how page numbers were assigned by MET was not readily apparent and did not necessarily
correspond to the date the complaint/call was received or to the type of document. MET had assigned a unique
identifier number to each page of documentation, but not to each unique complaint/call (IDR). Therefore, one
documented page might contain several different IDRs that shared the same identifier number, or one IDR might
span several pages corresponding to several different identifier numbers.
Information considered identifiable (i.e., name, address, phone number) was redacted by MET to protect the privacy
of the reporter and product user prior to transmission of the records to LSRO and FDA. It was possible that during
redaction, some additional data were inadvertently removed (U.S. General Accounting Office, 2003b).
III.2.2 The federal (FDA) data set
An overview of the development and features of FDA’s postmarketing surveillance program is provided in
Appendix F.
III.2.2.1 Recognition and reporting of adverse events in individual product users
The federal surveillance program for dietary supplements, medications, and other FDA regulated products relies on
the reporting of adverse events filed through MedWatch and through facsimiles, letters, and calls to various FDA
district offices and to FDA’s Center for Food Safety and Nutrition (CFSAN).11
III.2.2.2 Creation of individual data records
To document incoming calls, FDA used one of three forms: (1) FDA MedWatch form 3500 for voluntary reporting,
(2) FDA “Consumer Complaint/Injury Report” or “Complaint/Injury Report” form 2516, or (3) FDA “Adverse
Reaction Report” form 1639 (predecessor to Form 3500). Reports received by mail or facsimile tended to be on
MedWatch forms (see Figure F-2; Appendix F). In general, follow-up investigations by FDA were documented
using the FDA “Adverse Event Questionnaire.”
Consumer Safety Officers answered health-related calls to FDA. Some Officers were experienced in handling
complaints of adverse events (i.e., the Complaints Coordinator for the District Office). In general, Consumer Safety
Officers were required to have a minimum of 30 college credits of science (e.g., chemistry, biology), but were not
required to have any medical training.
FDA assigned a unique 5-digit identifier number to each IDR. All information collected pertaining to an IDR,
beginning with the initial report, was coded with the same identifier number and included in the IDR. The IDR
often contained information obtained during follow up by FDA field representatives. FDA’s IDRs can consist of up
to nine divisions as follows:
1.
2.
3.
4.
5.
6.
7.
8.
9.
Section 0
Section 1
Section 2
Section 3
Section 4
Section 5
Section 6
Section 7
Section 8
Initial incoming report
Affidavits
Label sample
Outpatient medical records
Emergency care records
Hospital record summaries
Hospital physician notes and other medical notes
Hospital surgical operations and procedures forms
Other (e.g., death certificate, autopsy report)
Hence, an FDA IDR might consist of a single page form or multiple pages.
In some instances, two IDRs had the exact same incoming record (Section 0) because the initial record involved two
product users who both experienced a health-related problem associated with use of the same product (e.g., mother
and son; husband and wife). In these cases, FDA assigned each individual a unique identifier number so that a
separate IDR was established for each individual.
11
At the time of this LSRO report, CFSAN directed reporters of adverse events related to the use of dietary
supplements to FDA’s MedWatch system.
III.2.2.3 Federal (FDA) database of individual data records
As of September 27, 2002, FDA had received approximately 1800 IDRs (Docket No. 00N-1200, volumes 2-58)
related to dietary supplements containing ephedrine alkaloids (U.S. General Accounting Office, 2003b). Of these,
322 (18%) complaints reported the consumption of Metabolife 356® (U.S. General Accounting Office, 2003b).
FDA used a subset of the 1800 IDRs to prepare the CFSAN document “Assessment of Public Health Risks
Associated with the Use of Ephedrine Alkaloid-Containing Dietary Supplements” (Docket No. 95N-0304). It is this
subset of ephedrine alkaloid-containing dietary supplements obtained from FDA pursuant to the Freedom of
Information Act that was available to LSRO for review. All identifier information (e.g., names, addresses) was
redacted. FDA used the date of the initial IDR to divide the subset (n=273) into three groups:
• Pre Case Series (n=14) collected prior to June 1, 1997
• New Case Series (n=140) collected June 1, 1997 through March 31, 1999
• Post Case Series (n=119) collected April 1, 1999 through February 15, 2000
III.2.3 Selection of sample individual data records
IDRs, 200 each from the MET and FDA databases, were selected by LSRO for evaluation. These two sets of IDRs
are referred to as “sample sets.” When a duplicate IDR was identified in the sample set, the duplicate was excluded
from study and another unique IDR was substituted.
III.2.3.1 Selection of industry (MET) records
MET records were selected from those in the database that utilized the most current IDR form entitled “Nurses
Database – Caller Info” (Figure 3-1). This form contained the following data fields:
• Names of caller (reporter)/product user and MET representative receiving the call
• Date and time of call
• Product user descriptors (i.e., age, weight, height)
• Exposure (i.e., serving size, servings per day, duration of use)
• Health-related complaints (text box for free-style comments as well as a check list of 76 possible healthrelated complaints)
• Medical history (i.e., medications, past diagnoses)
• Other considerations (i.e., water intake, caffeine intake, current diet)
• Response to caller (e.g., reviewed usage guidelines, recommended increasing water consumption)
The MET form did not include specific data fields to prompt collection of the following information:
• Gender
• Relationship of the reporter (caller) to the product user (e.g. self, spouse, health care professional)
• Confirmation that the product consumed was Metabolife 356®
• Batch/lot number
• Indication of why the individual was using the product
• Other dietary supplements consumed
• Whether there had been a dechallenge/rechallenge test of the supplement
• Plan for follow up with caller and/or plan for further action by the company
• Whether the product user had filed a previous complaint
LSRO chose to use a random number generator to select 200 IDRs from the thousands of IDRs available in the MET
database as follows: 10 to 14 IDRs were randomly selected from each of 15 files containing a large preponderance
of IDRs in the latest format, and two to six IDRs were randomly selected from three files containing a small number
of IDRs in the latest format. Testimonials of weight loss that did not contain any health-related complaints were not
IDRs and therefore were excluded.
III.2.3.2 Selection of federal (FDA) records
In contrast to the method for selection of MET IDRs, FDA IDRs were selected by the latest date in the Post Case
Series, which would be closest in time to the selected MET IDRs, working backwards from most recent to least
recent until 200 FDA IDRs had been selected. However, FDA Case ID 13085 was excluded because it was a
complaint about a product called H.E.L.P. Although included in the FDA data set for ephedrine alkaloid-containing
supplements, this product apparently did not contain any ephedrine alkaloids per labeled ingredients.
III.3 ANALYSES OF SAMPLE SETS
Sample sets were qualitatively assessed using criteria that the Committee determined would be appropriate
components of an effective surveillance program.
When retrieving data from IDRs, blank data fields were assumed to mean that the information was not collected. It
was possible that blank data fields meant that the receiver: (a) failed to ask the caller for the information; (b) the
information was not known to the caller/reporter; (c) the caller/reporter refused to answer; (d) the caller provided a
response that was not recorded.
Signal detection is a process of sorting and identifying IDRs in order to find interesting product-adverse event
relationships that merit further attention. The International Conference on Harmonisation (ICH) Expert Working
Group (2001) recommends that an IDR include one identifiable patient, one identifiable reporter, one reaction/event
and one suspect product.12 In addition, to properly process the data the ICH Expert Working Group (2001)
considers that: (a) the record should be assigned a unique identifier number, (b) the receipt of the most recent
information should be dated, and (c) if the information was forwarded by a third party (i.e. manufacturer), that the
sender should be identified. The LSRO Committee adopted the following five minimal elements of IDRs in order
for the IDRs to be of value for signal detection:
1.
2.
3.
4.
5.
An identifier for the product user was obtained [e.g., name, initials, age, gender, or an externally applied
unique number (i.e., hospital chart number)].
The complaint was health-related (is an adverse event versus another type of complaint).
A dietary supplement product was consumed prior to the onset of the health-related complaint.
The IDR was dated (day/month/year).
An identifier for the reporter was obtained (e.g., name and contact information).
For the purpose of this review, only the first four of these five criteria were used to assess sample sets for signal
detection because the identifiable reporter would have been redacted for all IDRs in the sample sets.
III.3.1.1 Identifier for the product user
There was no data field to record gender on the MET form or on the “Consumer Complaint/Injury Report” or
“Complaint/Injury Report” FDA form 2516. However, LSRO obtained the gender when possible from the type of
complaint (e.g., menstrual irregularity), medications (e.g., birth control pills), medical history (e.g., hysterectomy),
or from records of follow-up investigations, or from text comments (e.g., “he was hoping it would work for him”).
There were 65 MET IDRs that did not include age, gender, or an externally applied unique number and for which
the entire data field for the product user name was redacted. In order to determine whether these IDRs met criteria
for signal detection, MET permitted one supervised viewing of these 65 IDRs in the unredacted state after a
confidentiality agreement had been signed to protect product user privacy. It was determined whether or not a
name or initials had been entered into the data field, but actual names/initials were not recorded or otherwise
retained by LSRO.
III.3.1.2 Coding of adverse events
Health complaints were captured by MET as free text and/or as a selection from the preprinted list of medical
conditions on the MET form (Figure 3-1). Apparently, the listed medical conditions were not used with specific or
standardized definitions, which hindered retrieval of information. For example, if it was documented under
“Comments” that the product user complained of a rapid heartbeat, one of several conditions was selected (i.e.,
irregular heartbeat, palpitations, or tachycardia). Non-specific terms such as “felt sick” were sometimes used and the
severity of the event was not always noted.
12
The term “suspect” was meant to include a temporal association (product exposure prior to the event).
Signs and symptoms of adverse events in the FDA and MET samples were converted from verbatim IDR terms into
standardized terms using the Medical Dictionary for Regulatory Activities (MedDRA®) (International Conference
on Harmonisation, 1998). The coding of adverse events was carried out using a coding tool (dsNavigator®; Galt
Associates, Sterling, VA) that supports the automated assignment of MedDRA® dictionary terms into electronic
files. IDRs that included diagnostic tests (e.g., testing for myocardial infarction) added certainty to the
designation/coding of the adverse event, but unlike causal testing, cannot suggest a specific etiology for the event.
No attempt was made to judge whether or not the supplement caused the recorded adverse event.
III.3.1.3 Dietary supplement product
If the IDRs in the FDA sample set did not include ingredients and labeled amounts to verify that the supplement
contained ephedrine alkaloids, an attempt was made to obtain this information from the manufacturer via the
Internet or by telephone.
III.3.1.4 Date of record
All but one MET record had sufficient information to determine whether a date was documented. A data field for
date was visible for all FDA IDRs.
in preparation for signal evaluation. The Committee determined that two types of data were needed for signal
prioritization: (1) data used to rank the reliability and accuracy of the IDR; and (2) data used to rank the
seriousness of the adverse event or outcome.
III.3.2.1 Reliability and accuracy
It is not known whether MET considered callers to the hotline number as customers in the broadest meaning of the
term, regardless of whether the caller was the one consuming the product or not. Because the MET data form did
not contain a data field to record the relationship of the caller to the product user, it was not readily evident whether
a product user or family member or health care professional called the hotline. Hence, the type of reporter could be
categorized only for those IDRs that contained a description of the call in which the receiver documented that the
caller stated or claimed they themselves experienced the problem (e.g., “Customer states she passed a kidney stone”)
or documented the relationship of the reporter to the product user.
The type of reporter is used as an indication of the reliability and accuracy of the IDR. Those IDRs filed by health
care professionals are given more credence for the terminology of the health-related complaints than IDRs filed by
product users or family members because health care professionals have been trained in medical terminology.
III.3.2.2 Seriousness
LSRO adopted the point of view that the seriousness of an adverse event is determined by its outcome (International
Conference on Harmonisation Expert Working Group, 1994). For prescription drugs, determination that an adverse
event was serious would require that the event had one or more of the following consequences:
• Resulted in a transient but serious disability (e.g., temporary loss of eyesight, hallucinations)
• Resulted in a lasting disability
• Required hospitalization (Note that being observed, evaluated, and/or treated in an emergency room and
released is not considered hospitalization)
• Extended a hospital stay if already hospitalized at the time of the event
• Was life threatening (as judged by a physician)
• Resulted in a congenital anomaly or genetic disease
• Was fatal to product user or fetus
• Was an important medical event, as noted by a physician, that jeopardized the patient in that it required
medical intervention to prevent permanent impairment or damage even though it may not have been
immediately life-threatening or fatal or required hospitalization
For the purposes of the LSRO evaluation, these same criteria were used to determine whether the IDR had sufficient
information to permit the evaluation of whether an adverse event was serious. The IDR had to document that one of
these serious consequences had occurred, or that the signs and symptoms had resolved without serious consequence.
In the case of pregnancy, physician judgment of a healthy baby at birth was considered to indicate the lack of a
congenital anomaly. Some genetic diseases might not be evident at birth and may require longer follow up than the
one-to-two month postpartum investigations included in the sample set. In one FDA case, a woman who became
pregnant while using the dietary supplement decided to terminate her pregnancy out of concern that a supplement
may cause birth defects. The termination of pregnancy was not reported as a fatal event.
The FDA MedWatch form was the only form used in the sample sets that contained data fields to specify whether
the adverse event was life threatening, required intervention to prevent permanent impairment, resulted in
hospitalization, disability or congenital anomaly, and/or was fatal. In the case of hospitalization, sufficient detail
had to be present to distinguish between an emergency room visit (not a criterion for serious) and admission to
hospital (met criterion for serious). Furthermore, check marks selecting life threatening events and intervention to
prevent permanent impairment and disability were considered only if the reporter was a health care professional.
is indicated. The Committee identified eleven types of data that are useful for signal evaluation:
1.
2.
3.
4.
5.
6.
7.
8.
Gender. As noted above, if gender was not specifically identified, other information indicating gender was
considered sufficient. The provision of gender information helps to describe the population of users overall,
and is important to discern whether the product may have a gender-specific effect.
Age at the time of the event. If age was not specified, it could be calculated using the birth date and the
date of the event if this data were available. Information on age is important to determine if events not
typical for a particular age group are recorded. This information can also indicate misuse by minors.
Pre-existing medical condition. If no chronic medical conditions were noted and it was not confirmed that
there were no past medical problems, the data was considered to be unknown. Medical history helps
identify subpopulations of individuals (e.g., individuals with asthma) who may be more susceptible to
adverse events resulting from the use of particular dietary supplements or specific ingredients.
Concomitant exposure to prescription and/or nonprescription medications. If no medications were
noted and it was not confirmed that there were no medications used, the data was considered to be
unknown. Documenting concomitant exposure to medications helps to identify potential confounding
factors leading to adverse events and/or helps identify potential supplement-drug interactions.
Concomitant exposure to other dietary supplements. If no other supplements were noted and it was not
confirmed that there were no other supplements used, the data was considered to be unknown. The
provision of information regarding concomitant exposure to food and other dietary supplements helps to
identify potential confounding factors contributing to adverse events and/or helps identify an additional
source of a bioactive ingredient or a supplement-supplement interaction.
Information on why the product was used. Documenting the reason the individual used the product, in
addition to the general intended use as marketed by the manufacturer, could help identify “off label” uses
that lead to adverse events in a subpopulation of product users. The form developed by MET to record
adverse events did not contain a data field to document why the individual was using the product.
According to the Metabolife 356® label and product information from the manufacturer (Metabolife
International, Inc., 2003), the product would “raise the body’s metabolism,” which “burns fat,” and the
product “reduces appetite” and achieves “higher energy levels” than would occur from solely reducing
caloric intake. It was marketed to those who sought to lose weight while maintaining or increasing their
energy level. Therefore, if it was noted on the IDR that the product user restricted their diet (restricted
energy consumption) and/or had noted recent weight loss, the reason for taking the supplement was
surmised to be for weight loss.
Batch/lot number. If not documented, batch/lot numbers can sometimes be retrieved from photocopies
made of the product label. Obtaining accurate batch/lot numbers assists a manufacturer in identifying
production errors and associated adulterated, misbranded, or spoiled products.
Daily dose of bioactive ingredients. The dose of bioactive ingredients (e.g., mg/day) could be calculated
if the product identity was known, the amount of bioactive ingredients per tablet was known, and the
amount of tablets consumed per day was also known (i.e., ephedrine alkaloid amount per tablet and/or per
serving and number of tablets and/or servings per day). Dose information can be used to determine if
intake exceeded the recommended limit.
9. Time from initial use to onset of the adverse event. Duration of use reported at the time of the event can
be used as an estimate of time to onset. Time to onset can help to determine biological plausibility.
10. Rechallenge information. A rechallenge is sometimes conducted to determine whether or not the adverse
event will reoccur after a second exposure to the dietary supplement. An account of a positive rechallenge
(following the initial challenge and dechallenge) supports the association of product use and the adverse
event occurrence. However, other than a product user’s self-imposed rechallenge, a rechallenge is not
conducted unless, in the judgment of the treating health care professional, it is of clinical value to the
patient.
11. Geographic location (e.g., Zip Code). A cluster of adverse events in a particular region might correspond
with a particular batch/lot or result from product tampering; it might also identify populations that have
different risk factors associated with use. The geographic location of the product user would have been
redacted and therefore this parameter was not included in the analysis.
III.4 RESULTS AND DISCUSSION
III.4.1 Information of value for data intake and creation of individual data records
III.4.1.1 Route of intake of initial complaint/report
Because of the inclusion criterion set by the Committee to select only from MET IDRs documented on the latest
version of the data form, all MET sample IDRs were derived from incoming telephone calls.
In contrast, IDRs in the FDA sample set were selected consecutively by date and so they represented a variety of
formats and modes of transmission (Table 3-1). The majority (68%) of initial FDA records were transmitted on
standardized forms by mail or facsimile and approximately one-fifth (21%) of records to FDA were received by
telephone.
III.4.1.2 Data obtained during follow up
The MET records represented information collected during a single telephone contact, with the exception of eight
cases that documented an improvement or resolution of signs and symptoms. Five reporters stated they did not want
further follow up. In four other cases, a follow up was attempted but not completed. (This data includes
information provided by MET during final proof). Hence, data recorded during follow up in the MET sample set
were only helpful for signal prioritization in a limited number of cases.
In contrast, 135 of 200 (67.5%) IDRs in the FDA sample set included information in addition to the initial data form
(e.g., follow-up forms, medical records). An advantage to FDA’s inclusion of follow-up investigations was that
multiple accounts of the same event were often recorded. This provided for more opportunities to confirm the
accuracy of the information already collected and to gather additional information to make the IDR more complete.
However, multiple component questions with a short space for answers on the follow up form often resulted in only
part of the question being answered. Follow-up investigations conducted by FDA did not appear to be designed to
identify and retrieve specific information that was missing from the initial data form. The portion of data of value
for signal detection, prioritization and/or evaluation that was obtained from the initial FDA data form compared with
the portion obtained from any FDA follow-up records for each of the 200 FDA IDRs was not determined.
As indicated in Table 3-2, even though MET data were not intended to be used for postmarketing surveillance, most
IDRs in MET and FDA sample sets had the minimal information necessary to permit signal detection that would
warrant further attention, i.e., progress to signal prioritization and evaluation. The U.S. General Accounting Office
(GAO; 2003b) emphasized that IDRs should optimally include demographic data such as gender and/or age because
this information is useful for determining whether or not the events were uncommon, particularly for certain
subpopulations (e.g., heart attacks in young adults). Although both sample sets had a high percentage of IDRs
containing an identifier for the product user, the FDA set had more IDRs with age or gender as an identifier than did
the MET set making the FDA data more useful for signal detection and evaluation (evaluation data follows).
III.4.3.1 Reliability and accuracy
The person who files an IDR can be the product user, a member of their family, a friend, co-worker, health care
professional, product distributor, attorney, or other concerned associate. Nearly one-half (48.5%) of IDRs in the
FDA sample set were reported by a health care professional and another one-third (32.5%) of FDA IDRs were selfreported (Table 3-1). The large portion of FDA IDRs filed by health care professionals is similar to patterns for
postmarketing reporting to FDA for other FDA-regulated products (e.g., medications) (Appendix F).
In contrast, the relationship of the product user to the person submitting the adverse event was unclear for most
(85%) IDRs in the MET sample set. Based on comments typed in the IDR, it is likely that most MET IDR reporters
were non-health care professionals, as is the case for IDRs for other food product complaints called in to the
manufacturer (Allgood et al., 2001). Hence, because the type of reporter (i.e., health care professional, product user)
was known for 98% of FDA IDRs compared with only 15% of MET IDRs, the FDA sample set was more useful for
signal prioritization, with regard to determining reliability and accuracy.
The preceding findings are consistent with characteristics of other product surveillance systems. In general, FDA
tends to receive IDRs from health care professionals (Appendix F), who are more likely to report unexpected,
serious events than events that were easily resolved. In contrast, manufacturers of food products tend to receive
complaints directly from product users, who often seek advice and/or may be motivated to obtain a refund
(Appendix E). Hence, manufacturers of food products may receive a preponderance of complaints of adverse events
that, for the most part, have not advanced to the stage where the product user has sought professional health care.
The FDA sample set may be more reflective of a different set of events— those that developed past the stage of
product user self care and, for the most part, required medical intervention.
Table 3-1. Descriptive characteristics of individual data records associated with dietary supplements containing
ephedrine alkaloids from two sets of sample recordsa
Descriptive parameters
Initial record
Product
Transmission
Incoming call:
Form by mail:
Form by facsimile:
Letter/memo:
Email:
N/A:
FDA IDRsb
(n=200)
MET IDRs
(n=200)
42 (21%) records
91 (45.5%) records
45 (22.5%) records
11 (5.5%) records
9 (4.5%) records
2 (1%) records
200 (100%) records
Reporter (relationship to product user)
Self:
Family member:
Health care professionalc:
Other:
N/A:
Product types
Total:
Use of Metabolife 356®:
Specific product not reported:
65 (32.5%) records
26 (13%) records
97 (48.5%) records
8 (4%) records
4 (2%) records
29 (14.5%) records
1 (0.5%) records
55 different products
88 (44%) records
2 (1%) records
1 productd
200 (100%) recordse
Ephedrine alkaloid content per tablet
Range:
Mean (SD):
N/A:
3-32 mg/tablet
14.6 (5.1) mg/tablet
26 (13%) records
12 mg/tablet
12 mg/tablet
Ephedrine alkaloid dose per day
Range:
Mean (SD):
N/A:
10-250 mg/d
54.9 (45.2) mg/d
69 (34.5%) records
6-108 mg/d
39.7 (20.1) mg/d
6 (3%) records
170 (85%) records
Duration of use
1-1096 days
1-240 days
Range:
75.1 (162.6) days
30.0 (42.2) days
Mean (SD):
52 (26%) records
15 (7.5%) records
N/A:
Product user
Age
Range:
15-73 y
16-82 y
Mean (SD):
36.4 (14.1) y
33.0 (12.4) y
N/A:
24 (12%) records
95 (47.5%) records
Genderf
Male/Female:
63/133
3/73
Ratio:
1:2
1:24
4 (2%) records
124 (62%) records
N/A:
a
Data represent a subset of IDRs, selected systematically from much larger databases (see Section III.2.3).
Although the manufacturer did not intend the collection and documentation of calls to a hotline as
“postmarketing surveillance,” this system and the resultant data were reviewed by the Committee to gain
insight into the design of possible surveillance systems for dietary supplements.
b
FDA: U.S. Food and Drug Administration; IDR: individual data record; MET: Metabolife International
Inc; N/A: information could not be determined from IDR.
c
Health care professional includes physicians, nurses, pharmacists, chiropractors, etc.
d
One IDR mentioned use of an additional dietary supplement: Metab-O-LITETM. Some Metab-O-LITETM
products contained ephedra.
e
If not identified, the product in all IDRs was assumed to be Metabolife 356®.
f
More data were available to categorize gender for women than for men.
Table 3-2. Comparison of the percentage of individual data records that met minimal criteria for signal detection
in two sets of sample recordsa
Criteria for signal detectionb
FDA IDRsc
MET IDRs
(n=200)
(n=200)
Identifier for the product user was obtained [i.e., name, initials, age, gender,
or an externally applied unique number (e.g., hospital chart number)]
198 (99.0%)
194 (97.0%)
Can determine whether or not the complaint was health-related (is an adverse
event versus another type of complaint)
200 (100%)
200 (100%)
Can determine whether or not a dietary supplement was consumed prior to
the onset of the health-related complaint
198 (99.0%)
198 (99.0%)
There must be a full precision date of the record (day/month/year)
199 (99.5%)
199 (99.5%)d
There must be an identifiable reporter (i.e., name and contact information)
N/A
N/A
Overall: IDRs meeting minimal criteria for signal detectione
195 (97.5%)
192 (96.0%)
a
Although the manufacturer did not intend the collection and documentation of calls to a hotline as “postmarketing
surveillance,” this system and the resultant data were reviewed by the Committee to gain insight into the design
of possible surveillance systems for dietary supplements.
b
To meet criteria, the information had to be documented or discoverable in the IDR. For example, if it was noted
that the individual had undergone a hysterectomy, that IDR met criteria for the product user identifier because
gender was discoverable (female).
c
FDA: U.S. Food and Drug Administration; IDR: individual data record; MET: Metabolife International Inc;
N/A: Information could not be determined from the IDR.
d
In many instances the last digits of year were not included on the printout because the data field window was too
narrow to display the full field. However, a company spokesperson said that this information was stored and
retrievable.
e
Percentage of IDRs that met the first four criteria listed.
III.4.3.2 Types of adverse events reported to a federal (FDA) surveillance program compared with the
frequency of those events reported via a manufacturer’s (MET’s) customer service telephone line
Coding produced 312 unique terms to describe the types of adverse events in the FDA sample set, with some events
reported rarely and others reported more frequently (Table 3-3). There was an average of approximately five types
of events (five unique terms) per FDA IDR and a sum total of 1038 events (multiples of terms) in the sample set
(n=200 FDA IDRs). In contrast, coding produced 74 unique adverse event terms in the MET sample set. There was
an average of approximately one type of event per MET IDR for a sum total of 279 events in the MET sample set
(n=200 MET IDRs).
The most frequent adverse event terms in the FDA sample set (occurring in 6% or more of IDRs) are listed in Table
3-3. Their frequencies in the MET sample set are also listed for comparison. The purpose of this table is to
highlight differences between the adverse events collected by the federal surveillance system (i.e., FDA), which
receives a large portion of reports from health care professionals, and the events collected by organizations receiving
primarily consumer accounts during customer service (e.g., MET).
Table 3-3 is a compilation of records and does not indicate whether any of the events would qualify as a signal or
whether the respective IDR met criteria for seriousness. As discussed in Section II (and Appendices E and F), many
additional pieces of information and ongoing analyses are needed before an association between a product and event
can be established. Furthermore, association is not synonymous with causation and even more scientific evidence
would be required before reaching a conclusion of causation.
Dizziness and nausea ranked as the top two most submitted events in both sample sets. For nearly every type of
event listed in Table 3-3, the FDA sample set had greater numbers of events reported. The notable exception was
tremor, for which the frequency was similar for both sample sets.
Of the top 40 most frequent adverse events recorded in the FDA sample set, 17 were not recorded in the MET
sample set. These included some of the more critical adverse events (e.g., convulsions, cerebrovascular accident,
and circulatory collapse) and some of the more medically technical terms (e.g., asthenia). However, some clinically
important adverse events (e.g., dyspnea, chest pain) were present in the MET sample set. In general, the most
frequently described adverse events in the MET sample set were all present in multiple reports of the FDA sample
set. However, there were 11 (5.5%) records of rash, 9 (4.5%) records of muscle cramps, and 13 (6.5%) records of
constipation in the MET sample set; which were two, three, and four times more frequent, respectively, than in the
FDA sample set. There were also five instances of alopecia in the MET sample set, yet this condition was not
evident in any IDR in the FDA sample set.
Table 3-3. The types of adverse events reported to a federal surveillance program (having a relatively large
percentage of health care professional reporters) compared with the frequency of those events reported via a
manufacturer’s customer service telephone linea
Type of adverse eventb
FDA IDRs
MET IDRs
(n=18)
(n=200)
(n=200)
Dizziness
40 (20.0%)c
17 (8.5%)
Nausea
32 (16.0%)
17 (8.5%)
Palpitations
31 (15.5%)
9 (4.5%)
Chest pain
30 (15.0%)
6 (3.0%)
Headache
30 (15.0%)
15 (7.5%)
Dyspnea NOS
26 (13.0%)
4 (2.0%)
Insomnia
25 (12.5%)
11 (5.5%)
Convulsions NOS
22 (11.0%)
0 (0.0%)
Fatigue
21 (10.5%)
6 (3.0%)
Vomiting NOS
18 (9.0%)
6 (3.0%)
Asthenia
17 (8.5%)
0 (0.0%)
Hypertension NOS
17 (8.5%)
3 (1.5%)
Tremor
16 (8.0%)
15 (7.5%)
Nervousness
15 (7.5%)
6 (3.0%)
Memory impairment
14 (7.0%)
0 (0.0%)
Tachycardia NOS
13 (6.5%)
3 (1.5%)
Cerebrovascular accident
12 (6.0%)
0 (0.0%)
Loss of consciousness
12 (6.0%)
1 (0.5%)
a
Data represent a subset of IDRs of ephedrine alkaloid-containing dietary supplements (FDA: 55 products; MET:
Metabolife 356®) selected systematically from much larger databases (see Section III.2.3). Although the
manufacturer did not intend the collection and documentation of calls to a hotline as “postmarketing
of possible surveillance systems for dietary supplements. IDRs were initially submitted by telephone for 21% and
100% of FDA and MET IDRs, respectively. Follow up was conducted in 67.5% and 2.5% of FDA and MET
IDRs, respectively. This table is a compilation of records and does not indicate whether any of the listed events
would qualify as a signal or whether the respective IDR met criteria for seriousness. The reader is referred to
Section II for further clarification of signal generation. FDA: U.S. Food and Drug Administration; IDR:
individual data record; MET: Metabolife International, Inc.; NOS: Not otherwise specified.
b
Adverse event: an adverse health-related event is not necessarily a reaction/effect of a product. Verbatim terms
for adverse events as reported were coded using the Medical Dictionary for Regulatory Activities (MedDRA®).
See glossary in Appendix H for definitions.
c
Percentages of total IDRs are not cumulative. Each IDR may have documented more than one type of adverse
event.
Hence, the FDA sample set (a surveillance system receiving a large portion of events reported by health care
professionals) was more complex, owing in part to:
• Additional data collected by FDA during follow-up investigations
• Greater number of total adverse events recorded (1038 FDA vs. 279 MET)
• Difference in types of adverse events recorded, in that the FDA records included more medically critical
events
• Differences in the products used, in that only a fraction of the total complaints to FDA was related to
individuals using Metabolife 356® and many of the supplements included in FDA’s database had higher
doses of active ingredients when compared with Metabolife 356®
• Differences in the types of reporters (health care professional vs. layperson)
III.4.3.3 Seriousness
Although the FDA and MET IDRs were collected for different purposes, the Committee evaluated them in parallel
and applied the terms adopted in the study uniformly. Specifically, the Committee adopted a set of criteria to
classify an adverse event as serious as noted in section III.3.2.2. Generally, the kinds of information needed to
document a serious event would come from a health care professional. In fact, several of the criteria actually require
a physician’s evaluation. Thus, it is not surprising that only 14 of 200 (7%) IDRs in the MET sample set contained
enough information useful for determining the seriousness of the IDRs for signal prioritization. Many of the 14
MET IDRs that met criteria for determination of seriousness had a resolution of symptoms without serious
consequence after the product was discontinued (dechallenge data). The number of IDRs with dechallenge
information may have been greater had all callers been asked for the time elapsed since their last dose and last
symptom. Follow up data was only available in five of the MET sample IDRs. Therefore, there was little evidence
of the duration of the adverse event and final outcomes for the MET sample set.
In contrast, 157 of 200 (78.5%) IDRs in the FDA sample set had sufficient information to determine the seriousness
of the IDR. Reports of fatality occurred in the FDA sample set, 16 of 200 (8%), but not in the 200 IDRs from MET.
Some FDA IDRs did not document outcome (e.g., fatality, permanent disability) even though the patient’s condition
was critical at the time of the initial report. The reasons noted for lack of outcome data were: (a) no follow up was
conducted, or (b) the family or physician was unavailable or unwilling to provide further information at the time of
the follow-up investigation.
III.4.3.4 Summary of data for signal prioritization
The primary conclusion drawn by the Committee is that the established mechanisms of FDA for data reporting,
intake, and follow up are superior and vastly more useful for signal prioritization in an adverse event reporting
system than the method used by MET to document customer service, which was not designed for surveillance and
had less reporting by medical professionals. Hence, one challenge for the dietary supplement industry will be the
preponderance of IDRs generated from consumer calls, which may trigger follow-up communication with the
involved health care professionals, rather than the receipt of initial reports directly from health care professionals.
A comparison of IDRs for the MET and FDA sample sets that met criteria for signal evaluation is presented in Table
3-4.
III.4.4.1 Gender
Provision of gender information represented the largest difference between the two sample sets for information
collected for signal evaluation. More than 60% of MET IDRs did not include mention of gender. In contrast,
gender was determined for nearly all FDA IDRs. Although one of the initial reporting forms (Form FDA 2516) did
not contain a specific data field for gender, the FDA form for follow-up investigations did include such a data field.
For IDRs in which gender was discoverable, the MET sample set was shifted toward a larger female proportion of
product users. This is a typical pattern for self-referred calls to the manufacturer, particularly for products used with
calorie restricted diets [e.g., aspartame (Butchko & Stargel, 2001), olestra (Allgood et al., 2001)]. However,
because gender was not systematically collected in the MET data set, the proportion of female to male product users
may have been skewed by the health-related complaint, medical history, or recorded concomitant medication for
which gender was more easily discoverable for females than males (e.g., menstrual irregularity, hysterectomy,
hormone replacement therapy).
Table 3-4. Comparison of the percentage of individual data records that met criteria for signal
sets of sample recordsa
Criteria for signal evaluationb
FDA IDRsc
(n=200)
Gender was specified or could be determined
196 (98%)
evaluation in two
MET IDRs
(n=200)
76 (38%)
Age at the time of the event was specified or could be calculated
175 (87.5%)
106 (53%)
Specified whether or not there were pre-existing medical conditions
156 (78%)
123 (61.5%)
Specified whether or not there were concomitant exposures to drugs and/or
nonprescription medications
140 (70%)
144 (72%)
Specified whether or not there were concomitant exposures to other dietary
supplements
94 (47%)
18 (9%)
Reported the reason why the product was used
147 (73.5%)
106 (53%)
Recorded batch/lot number
64 (32%)
1 (0.5%)
Daily dose of bioactive ingredients was specified or could be calculated
133 (66.5%)
194 (97%)
Time to onset from initial use to the adverse event was specified or could
be calculated
148 (74%)
183 (91.5%)
Reported the results of a rechallenge
11 (5.5%)
4 (2%)
Specified the geographic location (e.g., Zip Code)d
N/A
N/A
a
of possible surveillance systems for dietary supplements.
b
To meet criteria, the information had to be documented or discoverable in the IDR. For example, if the health
care professional noted that the product user had previously been healthy with no relevant past medical history,
this met criteria for pre-existing medical condition.
c
FDA: U.S. Food and Drug Administration; IDR: individual data record; MET: Metabolife International Inc;
N/A: Could not be determined from IDR.
d
Locator information was redacted from all IDRs and therefore was not included in the comparison.
III.4.4.2 Age
The FDA sample set included values for age for approximately 87% of IDRs, and this percentage of records may
have increased if one of the initial data forms (Form FDA 2516) included a specific data field for age. However, in
those cases where a follow-up investigation was conducted, the FDA follow-up form did include a data field for age.
In contrast, only approximately 50% of MET IDRs included values for age despite having a data field for age.
III.4.4.3 Pre-existing medical condition
Although one of FDA’s initial data forms (Form FDA 2516) did not include a specific data field for medical history,
the FDA sample set had more IDRs containing information about medical history than did the MET sample set.
Because the FDA sample set often included copies of medical records, the details and diagnoses of prior medical
conditions were typically more descriptive and possibly more complete than those provided in the MET sample set,
which relied, in most cases, solely on information obtained during the initial product-user complaint. The portion of
blank (unanswered) data fields, which in the sample review could have represented either missing information or
“no relevant history,” might be reduced if standardized forms included a structured data field that contained specific
categories of response.
III.4.4.4 Concomitant exposure to drugs and/or nonprescription medications
More than two-thirds of IDRs in each data set provided information on whether or not the individual was taking
medications. One of the initial data forms used in the FDA sample set (Form FDA 2516) did not include a specific
data field for medications.
III.4.4.5 Concomitant exposure to food and other dietary supplements
The FDA and MET sample sets differed substantially in their collection of information on the use of other dietary
supplements (supplements used in addition to the ephedrine alkaloid-containing supplements), with 47% and 9% of
IDRs, respectively, documenting this information. None of the initial data forms used in the FDA and MET sample
sets included a data field specifically for collecting information about concomitant exposure to other dietary
supplements. In contrast, the FDA follow-up form instructed the reporter to list all medications and dietary
supplements as well as foods and other products used at the time of the adverse event. The MET form did include
data fields for water intake, caffeine intake, and other related information that was used to document assessments of
intake, such as the level of fluid, protein or calcium intake.
Data on the use of other dietary supplements could be under-reported in manufacturer databases. Reporters might
withhold information about concomitant exposure to other supplements if they perceive that the manufacturer may
discourage them from using competitive products and/or that disclosure might alter the manufacturer’s response.
III.4.4.6 Indications for use
The FDA sample set had a greater portion of IDRs documenting the indication for using the dietary supplement than
did the MET sample set. The FDA form used in follow-up investigations asked specifically for indications for use
as listed on the product label. Product labels for supplements in the sample sets (n=55 products) claimed the
supplement would do one or more of the following: burn fat (n=26), provide energy (n=26), stimulate metabolism
(n=19), suppress appetite (n=16), increase mental alertness (n=6), build muscle (n=6), generally enhance the diet
(n=9), and/or exert a variety of functional effects (e.g., improve circulation) and/or maintain blood sugar (n=14).
This information, however, does not necessarily reflect the reason why the individual used the product.
III.4.4.7 Batch/lot number
Only one IDR in the MET sample set recorded a batch number. Obtaining batch/lot numbers assists the
manufacturer in investigating production errors and in removing adulterated, misbranded, or spoiled products from
the shelves as soon as possible.
If reporters calling the 1-800 number provided on the product label had the Metabolife 356® bottle at the time of
their call to the manufacturer, it was likely that they could have readily provided the batch number if instructed
where to look on the label. It is likely that if the reporter had been prompted to supply this information at the time of
their call, the MET sample set would have had more IDRs meeting the criterion for batch numbers than the FDA
sample set.
The IDRs in the FDA sample also often failed to include the batch numbers that were available. Specifically, 26
FDA IDRs noted that the product bottle/label had been available yet did not document the batch/lot number. For
example, in two cases a photocopy of the product label was obtained to verify the type of product consumed, but the
batch/lot number was only partially legible on the photocopy included in the IDR. The percentage of IDRs in the
FDA sample set having batch/lot numbers could have increased another 13% had FDA investigators documented
available information. In at least two instances, the product container did not have a batch number.
III.4.4.8 Dose and time to onset
Identification of product and bioactive ingredients. Only one individual in the MET sample set mentioned use of
an additional supplement that may have contained ephedrine alkaloid (i.e., Metab-O-LITETM). In contrast, the FDA
sample set was more heterogeneous with respect to the types of ephedrine alkaloid supplements consumed overall
and per individual. Of the 200 IDRs in the FDA sample set, Metabolife 356® was recorded as the associated
product in 88 (44%) IDRs. The remaining FDA IDRs were associated with a wide spectrum of products.
Specifically, there were 54 other ephedrine alkaloid-containing products that were each mentioned in 1-to-13 IDRs.
Eight IDRs in the FDA sample set recorded the use of two different ephedrine alkaloid-containing products. One
individual in the FDA sample set consumed three different ephedrine alkaloid-containing products. For two of the
FDA IDRs, the dietary supplement was reported to contain ma huang (ephedra), but no further description or
identification of the product was provided. Several individuals in the FDA sample set were adhering to dietary
programs that instructed them to consume numerous other types of dietary supplements in addition to an ephedrine
alkaloid-containing dietary supplement.
Many of the ephedrine alkaloid-containing products in the sample sets contained other bioactive substances as well.
For example, according to product information, Metabolife 356® contained 40 mg/tablet of caffeine alkaloids from
guarana. Of the 55 known products in the FDA sample set, 28 contained caffeine (1-200 mg/unit of serving), 11 did
not contain caffeine, and the caffeine content of the remaining 16 products was not reported. Of IDRs providing
dose information for products whose caffeine content was recorded, the intake of caffeine from supplements ranged
from 20 to 360 mg/d and averaged (SD) 132.4 (66.8) mg/d for MET IDRs (n=194), and ranged from 0 to 3000
mg/d and averaged (SD) 276.3 (452.0) mg/d for FDA IDRs (n=119). Moreover, caffeine is readily available in the
food supply (i.e., approximately 103 mg per 6 oz coffee, 36 mg per 6 oz tea, and 37 mg per 12 oz soda) (U.S.
Department of Agriculture, 2003). Total caffeine consumption varied greatly (and generally in undetermined ways)
among individuals. Several individuals in the MET and FDA sample sets reported drinking four or more cups of
coffee per day. An individual drinking four cups of coffee per day would consume 412 mg or more of caffeine per
day in addition to the amount consumed from dietary supplements and nonprescription medications.
One product user in the MET sample set complained that a urine sample of theirs tested positive for marijuana,
which they denied using. Similarly, one FDA IDR documented a urinalysis testing positive for marijuana and
amphetamines for a user of the supplement Midnight Ecstasy, and three other FDA IDRs noted urine test results that
were positive for amphetamines, two of which were associated with the product Metaboloss® and one with
Metabolife 356®. If these product users were not otherwise exposed to marijuana and/or amphetamines, product
samples should be tested to determine whether the product contained constituents that were not disclosed on the
product label or whether labeled substances might interfere with urinalyses to yield inaccurate test results.
Therefore, it is possible that some product users included in the sample sets may have been exposed to bioactive
substances that exerted pharmacological effects, in addition to ephedrine alkaloids.
Determination of daily intake. One strength of the MET sample set for signal evaluation was the ability to
calculate daily exposure of ephedra from dietary supplements for nearly all IDRs (Table 3-4). This was possible
because the total tablets per serving and tablets or servings per day were recorded and because the specifications for
the ephedra content of Metabolife 356® was known. The MET sample set also provided information that could be
used to estimate the time to onset of symptoms from the first use of the supplement because the duration of product
use was documented in 91.5% of IDRs. This information was found in 74% of the FDA IDRs. The data fields for
recording dose and duration of use were featured prominently at the top of the MET data collection form. Given
that more than 90% of MET IDRs contained information on dose and duration, it is likely that staff personnel
systematically questioned callers to obtain this information.
Because the FDA sample set included more than 50 different ephedrine alkaloid-containing products, it was not
surprising that the range of daily ephedrine alkaloid intake varied more for the FDA IDRs than for the MET IDRs
(Table 3-1).
III.4.4.9 Rechallenge
A rechallenge might be conducted to determine whether or not the adverse event recurs after a second exposure to
the dietary supplement. To conduct a rechallenge, the use of a dietary supplement is halted. Once symptoms
subside, the product is reintroduced to the product user, who is monitored for health-related signs and symptoms. A
rechallenge is sometimes self-imposed by the product user or conducted under medical supervision. Intentional
rechallenge could be carried out if the treating health care professional judges it would likely be of clinical benefit to
the patient. A rechallenge is not conducted if a serious and/or life-threatening reaction is anticipated.
In one case of a self-imposed rechallenge in the MET sample set, a 140 lb female took Metabolife 356® for
approximately two months without any health-related problems. Before the supplement bottle she was using was
emptied, she opened a new bottle of the same supplement and began using its tablets. She experienced abdominal
cramps and diarrhea so she discontinued using tablets from the second bottle (dechallenge). Then, she restarted
taking tablets from the first bottle without incident. She switched back to the second bottle again (rechallenge) and
the symptoms returned, indicating that the contents of the two bottles might be different and something about the
tablets in the second bottle might be causing her gastrointestinal problems.
Although the FDA sample set included more records of rechallenge, it is likely that the MET sample set could have
included even more rechallenge findings had they followed up the cases in which they advised a temporary
suspension of product use.
III.4.4.10 Summary of data for signal evaluation
The FDA sample set had more IDRs with information useful for signal evaluation than the MET sample set (Table
3-4). For seven of the ten criteria selected for comparison, the FDA sample set had a greater percentage of IDRs
with the necessary information compared with the MET sample set. The FDA sample set was strongest for the
recording of demographic data (i.e., gender and age). The MET sample set was strongest for the recording of dose
and duration of use. Fewer than 50% of IDRs in both sample sets met the criterion for batch/lot numbers or
documenting concomitant use of other dietary supplements. In each sample set, only a small amount of IDRs
documented a rechallenge exposure.
III.4.5 Issues resolution and formulation of risk management interventions
The FDA and MET sample sets differed greatly in the interactions between reporters and the organizations to whom
they reported. The FDA IDRs did not systematically document any suggestions or advice by FDA officials to the
product user or reporter. In contrast, IDRs in the MET sample set systematically documented consultations with the
reporter. Only one MET IDR lacked any documented response to the caller. User guidelines were reviewed with
some callers and refund information was often provided. Product users of Metabolife 356® were provided with one
or more suggestions in response to the information they had supplied about their diet and health-related complaints.
Dietary changes were often suggested (e.g., increasing water, protein, and fiber intake, and/or decreasing caffeine
and alcohol intake). Instructions to decrease the dose of the product or to take the product nearer to mealtime were
given to 39 (19.5%) callers in the MET sample set. Approximately one-half (n=107) of MET callers were told to
temporarily suspend use of the product until their symptoms resolved or to discontinue the product immediately and
consult a health care professional. In three instances the caller was transferred to a MET supervisor and in one
instance a plan was made for a company representative to correspond with the product user’s physician by facsimile.
III.5 SUMMARY OF FINDINGS RELATED TO METHODS USED TO COLLECT ADVERSE EVENT
DATA AND CREATE INDIVIDUAL DATA RECORDS
LSRO did not fully evaluate industry’s customer service systems or the federal postmarketing surveillance system
for dietary supplements. The primary purpose of the Committee’s review was to assess the efficacy of the two
systems for collecting data that would support signaling (results are summarized in overall conclusions section,
which follows). However, based on a limited comparison of MET and FDA ephedrine-related IDRs, many
additional observations were made related to the first two functional steps of a surveillance system: (1) recognition
and reporting of adverse events in individual product users, and (2) creation of IDRs. Specifically, the strengths and
limitations of standardized forms and follow up investigations are presented. These findings contributed to the
Committee’s deliberations of desired features of surveillance programs.
III.5.1 Standardized data forms
In general, data were presented concisely on the initial data forms. However, two limitations were common in both
sample sets: (1) most IDRs did not have information on batch and lot number, even in some instances when the
product package label had been available and photocopied (this data is useful for signal evaluation); (2) often, there
was no reason specified for why a data field was left blank. In particular, it was unclear whether a blank data field
for medical history meant that information was not collected or that it was collected but not documented because the
information was judged not to be relevant (e.g., the patient was previously healthy).
III.5.1.1 Use of standardized data forms by industry (MET)
•
•
Strengths
o The MET form included specific data fields to prompt collection of: (1) dose and duration of use,
which were featured prominently at the top of the form; this data is useful for signal evaluation,
(2) intake of water, caffeine, and other information related to current diet; this data is useful for
signal evaluation.
o Assessments of the current diet (i.e., level of protein or calcium intake) were often made and
documented; this data is useful for signal evaluation and issue resolution.
o The total number of tablets per serving and number of tablets or servings per day were often
recorded, and because the specifications for the ephedra content of Metabolife 356® were known,
dose could usually be determined; this data is useful for signal evaluation.
o The MET form included both structured and free-text data fields to document the response to the
caller; this data is useful for issue resolution.
o All text was typed and easy to read (note that MET sample IDRs were selected to exclude
handwritten forms).
Limitations
o The MET form lacked specific data fields to prompt collection of:
(1) Relationship of the reporter (caller) to the product user (e.g., self, spouse, physician); this data
is useful for signal prioritization.
(2) Gender; useful for signal detection and to determine if the events are gender-related for signal
evaluation.
(3) Confirmation that the product consumed was Metabolife 356®; this data is useful for signal
detection.
(4) Batch/lot number; this data is useful for signal evaluation.
(5) Indication of why the individual was using the product; this data is useful for signal
evaluation.
(6) Other dietary supplements consumed; this data is useful for signal evaluation.
(7) Affirmation or denial that there had been a dechallenge/rechallenge test of the supplement;
this data is useful for signal evaluation.
o Some data fields on the MET form were not large enough to transmit pertinent information
without truncation or loss.
o Some IDRs had dose units expressed using unconventional terms.
o There were no specific or standardized definitions for health-related complaints.
o Only approximately 50% of MET IDRs included values for age despite having a data field for age;
this data is useful for signal detection and evaluation.
III.5.1.2 Use of standardized data forms by regulatory authorities (FDA)
•
•
Strengths
o The MedWatch form included specific data fields to prompt collection of the seriousness of the
event; this data is useful for signal prioritization.
o The FDA “Adverse Event Questionnaire” follow-up form instructed the reporter to list all
medications and dietary supplements as well as foods and other products used at the time of the
adverse event; this data is useful for signal evaluation.
Limitations
o The MedWatch form lacked a data field for emergency room visit; the data field for
“hospitalization” was sometimes marked “yes” because there was an emergency room visit, which
does not meet the criterion for hospitalization; data to distinguish an emergency room visit from
hospitalization is useful for signal prioritization.
o FDA form 2516 did not include a specific data field for medical history.
o The FDA “Adverse Event Questionnaire” follow-up form: (1) lacked specific data fields to
identify and prompt collection of the information that was missing from the initial record, and (2)
lacked sufficient space to answer questions.
o For some products, the type and amount of ingredients were not included in the IDR or were not
otherwise available. Hence, bioactive substances and the dose exposure could not be calculated in
those cases; this data is useful for signal evaluation.
o
o
o
Individuals receiving complaints/reports may or may not have had medical training or be
experienced in receiving health-related complaints.
Handwritten text was often difficult to read and interpret.
Multiple component questions often resulted in an incomplete answer.
III.5.2 Follow-up data collection
III.5.2.1 Follow up by industry (MET)
•
•
The MET sample set, in most cases, lacked follow-up information. Follow up may be needed to obtain
data for signal detection, prioritization and evaluation. In general, there was no plan documented for follow
up with the reporter. Assessment of further action by the company could have been useful to obtain:
o Outcome; this data is useful for signal prioritization.
o Dechallenge/rechallenge information, especially when a temporary suspension of use had been
advised; this data is useful for signal evaluation.
o Additional information from medical records to address confounding factors (e.g., drug screening
results); this data is useful for signal evaluation.
There were at least three instances in which a caller was transferred to a supervisor. Any additional
information gained by this interaction between the reporter and supervisor, including information that
would have been beneficial for signal detection, prioritization and/or evaluation, was not included with the
original IDR.
III.5.2.2 Follow up by regulatory authorities (FDA)
•
•
The FDA sample set included copies of medical records, providing the details and diagnoses of past and/or
chronic medical problems; this data is useful for signal prioritization and evaluation.
Data for signal detection, prioritization, and evaluation were not always presented together and distinct
from the other, sometimes voluminous, data collected. Although follow up provided much useful
information for surveillance, this data often had to be located and retrieved from medical records and
testimonials included in the total IDR materials. Therefore, retrieval of the pertinent data was time
consuming.
III.6 GENERAL CONCLUSIONS RELATED TO THE VALUE OF DATA FROM FEDERAL (FDA) AND
INDUSTRY (MET) DATA SETS FOR SIGNALING
This LSRO review examines sample ephedrine-related IDRs from the same pool of IDRs that others have reviewed
(e.g., U.S. General Accounting Office, 2003b). LSRO did not attempt, as others have, to examine, describe or
tabulate events for the entire pool of available MET IDRs (U.S. General Accounting Office, 2003b) or FDA IDRs
[Assessment of Public Health Risks Associated with the Use of Ephedrine Alkaloid-Containing Dietary
Supplements, (Docket No. 95N-0304)]. Two additional clarifications are noted:
• The example IDRs, the associated products, and the respective customer service and surveillance systems
from which the IDRs were obtained may not have been fully representative of other industry and regulatory
authorities’ IDRs, surveillance systems, or of other types of dietary supplements.
• Neither the Committee nor LSRO makes any evaluation or conclusion of the safety or efficacy of ephedra
products. Findings and conclusions on the safety of dietary supplement products are outside the scope of
the work described in this report. Specifically, LSRO did not perform any causality assessments of
individual IDRs or use aggregate data in any way to assess the probability of a causal link between a
product and a health-event.
General conclusions of the retrospective review of sample IDRs and their usefulness for signal generation are
summarized in Table 3-5.
• Signal detection. Both MET and FDA sample sets of IDRs had sufficient information to permit signal
detection.
• Signal prioritization. In contrast to signal detection, fewer than 10% of MET sample IDRs had
information useful for determining the seriousness of the IDR. Even less of these would be weighted as
reliable or accurate given that the medical expertise of the reporter was not known and medical records
were not obtained. In comparison, health care professionals submitted 48.5% of IDRs in the FDA sample
set. Moreover, 78.5% of FDA sample IDRs had sufficient information to determine the seriousness of the
IDR. Hence, a substantial portion of the FDA sample would be useful for signal prioritization.
Signal evaluation. More IDRs in the FDA sample set supplied information useful for signal evaluation
than did the MET sample set. The FDA sample was particularly successful in collecting demographic data
[i.e., age (87.5%), gender (98%)]. The MET sample was successful in recording dose (97%) and obtaining
information useful for estimating the time from initial product exposure to the onset of the health-related
events (91.5%). Neither organization systematically recorded batch/lot numbers, a limitation that narrowed
the potential usefulness of their programs.
•
Table 3-5. General conclusions of a retrospective review of sample individual data records and their usefulness for
signal generationa
Sequential stages of signal generation
FDA IDRsb
MET IDRs
(n=200)
(n=200)
Signal detection
•
Met all criteria for signal
detection:
195 (98%)
192 (96%)
196 (98%)
30 (15%)
Signal prioritization
•
Had enough information to
categorize the reliability and
accuracy of the recorded event:
[97 (49%) of initial reporters
were health care professionals]
•
Had enough information to
categorize the seriousness of the
recorded event:
157 (79%)
14 (7%)
Signal evaluation
•
≥ 70% IDR met criterion of
each:
Age
Gender
Drugs – concomitant
Past medical history
Reason for use
Time to onset
Daily dose
Drugs – concomitant
Time to onset
•
< 70% IDR met criterion of
each:
Batch/lot number
Daily dose
Rechallenge
Supplement – concomitant
Age
Batch/lot number
Gender
Past medical history
Reason for use
Rechallenge
Supplement – concomitant
a
Data represent a subset of IDRs selected systematically from much larger databases (see Section III.2.3).
of possible surveillance systems for dietary supplements. Telephone transmission accounted for 21% and 100%
of initial incoming FDA and MET IDRs, respectively. Follow up was conducted in 67.5% and 2.5% of FDA and
MET IDRs, respectively.
b
FDA: U.S. Food and Drug Administration; IDR: individual data record; MET: Metabolife International, Inc.
In reviewing the MET and FDA IDRs, LSRO found overall that information for surveillance of dietary supplements
was limited. The limitations in signal generation for these sample sets stemmed mostly from systematic problems in
the methods used for initial data collection, coupled with insufficient follow up. Despite these limitations, it is likely
that signals (hypotheses) worthy of further investigation would arise from such data sets if they were developed in
the context of a continuing surveillance program having the requisite functional components detailed in Section II.
LSRO concluded that the records collected by MET were qualitatively less informative than those collected by FDA
and therefore pose a greater challenge for application in public health related analyses (Appendix B). It should be
noted that the MET records represented information collected, in most cases, during a single telephone contact for
the purpose of customer service; FDA records represented information collected in an initial contact and in follow
up by FDA field representatives for the purpose of surveillance.
This review of sample IDRs underscored LSRO’s understanding that an effective surveillance program for dietary
supplements will share many functional features of programs already in place for other types of health care products
(but must be tailored to the unique demands of dietary supplement products). For example, accurate, relevant, and
complete data collection during intake of the initial complaint will be of key importance for signal generation for
dietary supplements. Specific desired features for postmarketing surveillance programs for dietary supplements are
discussed in detail in the following section.
IV. POSTMARKETING SURVEILLANCE PROGRAMS
This section presents considerations for designing, operating, and evaluating an adverse event monitoring program
for dietary supplements. Recommendations are provided for the type of information to gather, the infrastructure
necessary to generate signals of potential product problems, the actions that should be taken if a signal of a potential
problem should arise, and the types of further testing that may be necessary. These recommendations are intended
to maximize the potential benefits of surveillance programs and minimize the limitations. Practical issues associated
with implementation are also addressed.
IV.1 CONSIDERATIONS FOR THE DESIGN OF A POSTMARKETING SURVEILLANCE PROGRAM
Use of most dietary supplements does not appear to be associated with any serious adverse events. However,
episodes of safety concerns over dietary supplements have emerged from several sources, e.g., the lay-press,
medical journals, the U.S. Food and Drug Administration (FDA). Problems with dietary supplements can arise from
improper use, product tampering, and product defects that alter product identity, quality, purity, strength, and/or
composition. Surveillance programs should have the capacity to detect and evaluate signals arising from these
types of problems.
IV.1.1 Well-designed postmarketing surveillance programs enable detection and evaluation of signals arising
from problems introduced during manufacturing
IV.1.1.1 Examples of adverse events resulting from problems introduced during manufacturing
Many dietary supplements are distinguished by the concentration of diverse, biologically active substances in a
single preparation. Some authorities believe that this characteristic more than any other differentiates herbal
supplements from foods. Such complex mixtures might increase risk for a variety of reasons, including greater
opportunity for manufacturing error and greater likelihood of batch-to-batch variability in active ingredients.
There are several examples of important adverse events that occurred after specific dietary supplements were
mislabeled (Appendix F). For example, batches of one product were labeled as containing “plantain,” yet instead,
were contaminated with digitalis lanata, a cardiac glycoside (Letkeman, 2003). The use of products mistakenly
containing digitalis was associated with serious adverse events that were consistent with effects of digitalis (i.e.,
heart block) and biomarkers of digitalis were detected in blood samples (Letkeman, 2003; U.S. Food and Drug
Administration, 1997). The contaminated products had been distributed to at least 175 retail outlets (U.S. Food and
Drug Administration, 1997). A second example of adverse events involved the inadvertent mislabeling of zinc
polynicotinate as the ingredient zinc picolinate (U.S. Food and Drug Administration, 1991a, 1991b). After product
users complained of adverse events similar to reactions that are associated with niacin supplements, the problem was
identified and the mislabeled product was withdrawn from the market. Thus, like most other health care products,
dietary supplements could occasionally be associated with health risks that could best be identified through the use
of standard surveillance measures.
IV.1.1.2 Postmarketing surveillance extends risk management tools beyond good manufacturing practices
Application of current good manufacturing practices (CGMPs) by industry would serve to help prevent problems in
production that lead to adulterated or misbranded products, which could cause adverse effects in product users (U.S.
Food and Drug Administration, 2003a). As a component of CGMP, FDA proposed procedures for recording,
reviewing, and investigating consumer complaints associated with dietary supplements (U.S. Food and Drug
Administration, 2003a). Product user feedback can help the manufacturer to recognize deviations that cause
adulteration. Note, however, that complaints specifically related to the pharmacological activity of an intended
dietary ingredient are outside the scope of the proposed regulations (U.S. Food and Drug Administration, 2003a).
FDA proposed that reports of serious illness or injury that are attributable to a product defect should be evaluated by
a competent medical authority to determine if follow-up action was necessary to protect the public health. However,
the extent and type of analyses and evaluations with regard to aggregate records, i.e., individual data records (IDRs)
were not specified. In contrast, a postmarketing surveillance program would require the management and analysis
of data in aggregate for pattern recognition, rather than reliance solely on a case-by-case assessment of individual
reports (Kahn, 2003).
The postmarketing surveillance examples from the FDA, poison control centers, and the World Health Organization
reviewed by the Committee (Appendix F) indicated clearly that in-house quality control, even if in compliance with
good manufacturing practice, is subject to various pathways by which contaminants can be inadvertently introduced
into final products. Given the complexity of worldwide materials sourcing in modern manufacturing, contamination
of final products can be difficult or impossible to detect by typical quality feedback steps such as physical or
chemical assays. Thus, the Life Sciences Research Office (LSRO) concludes that quality defects in the manufacture
of dietary supplements will inevitably occur and may occasionally pose threats to product users. In addition, the
effects of improper formulation, packaging, labeling, storage, and distribution are extremely difficult to predict, and
will also present likely potential hazards (U.S. Food and Drug Administration, 2003a; Walker, 2000).
LSRO concludes that a well-designed surveillance program can be an important supplement to in-house quality
control measures by extending a manufacturer’s risk management tools beyond CGMPs.
IV.1.2 Well-designed postmarketing surveillance programs enable detection and evaluation of signals arising
from non-manufacturing factors
Because FDA does not have the authority to require that dietary supplements undergo premarket approval for safety
and efficacy, some adverse events associated with dietary supplements may be recognized only after a product has
entered the marketplace. Factors other than manufacturing error that could contribute to adverse effects in product
users include:
• The discovery of previously unknown adverse events. After a new or reformulated product is introduced
into a marketplace, previously unknown adverse events might arise that after further study are found to
result from supplement interactions with foods, other dietary supplements, drugs, diseases, or lifestyle
factors. These interactions are more likely to be observed after large numbers of individuals use a product
(Section II and Appendix F).
• The intentional or unintentional misuse of supplements by a population. For example, dietary
supplement users may misinterpret directions for use, or may intentionally use the product excessively or
inappropriately. In particular, Ness et al. (2003) raised concerns about polyherbacy (i.e., the excessive or
inappropriate ingestion of multiple herbs for the treatment or prevention of disease). Polyherbacy by older
patients, who often take prescription medications for chronic disease, is of particular concern because of the
unpredictable number and range of potential interactions with conventional medications and underlying
diseases. Knowledge of misinterpretation and misuse that leads to adverse events is critical to an effective
response by the company. If there were sufficient concern, modification of the product label might be
necessary, perhaps supported by public notification.
• The purposeful contamination of supplement products (e.g., bioterrorism). While instances of willful
contamination of health care products have been rare, the Committee noted that such activities, whether
carried out with criminal intent or for purposes of bioterrorism, could be handled expeditiously if a proper
surveillance mechanism had been put in place.
IV.1.3 Effective surveillance programs detect relevant new patterns and outlier experiences worthy of follow up
Adverse event experiences that are temporally associated with the consumption of dietary supplements may be
submitted for many reasons. Not all adverse events, however, are actually related to the suspect product.
Generally, it is not possible to rule out the likelihood that the suspect product may have contributed to the adverse
experience, except in rare cases where the event occurred prior to exposure to the product (Hostelley, 2003).
Therefore, most adverse events at the level of a singular IDR end up with a possible association to the product. As
Kahn (2003) pointed out, there are two different causality questions that arise: (1) does the product cause the
reported adverse event; and, (2) did the product cause the reported adverse event in this product user? According to
Kahn (2003), the former, broader question of whether or not there is a manufacturing defect or product user
characteristic or other reason why the product might trigger the adverse event will not be answered by an individual
case causality assessment; it requires investigation through a population approach. Therefore, as discussed earlier,
in reviewing established postmarketing surveillance systems in use outside of the dietary supplement industry,
LSRO concluded that a routine one-by-one causality assessment of each IDR does not usually play a role in
signaling.
Moreover, based on the postmarketing surveillance programs of the FDA, poison control centers, and the World
Health Organization that the Committee reviewed, LSRO determined that an effective surveillance program does not
require full substantiation of all recorded adverse events. Rather, the ability to detect relevant new patterns and
outlier experiences that deserve interpretation and reaction by the manufacturer is critical (Appendix F). Optimally,
for protection of public health, full proof of causation should not be required before a preventative action is taken.
IV.1.4 Summary of considerations for the design of a postmarketing surveillance program for dietary
supplements
Surveillance programs can be designed to provide timely alerts of product-related safety concerns without causing
undue focus on trivial and/or unrelated adverse events. To be useful, postmarketing surveillance programs need to
collect data that will trigger signals derived from adverse events caused by:
• Alterations in product identity, quality, purity, strength, and/or composition introduced during
manufacturing
• Misuse of product
• Product tampering
• Currently unknown factors (e.g., supplement-medication interaction)
Evidence from surveillance systems in use outside the dietary supplement industry indicates that full proof that a
product caused a specific event is not required for postmarketing surveillance programs to be effective in identifying
important signals of product problems.
Implementation of surveillance programs can be complex, would involve specific areas of expertise, and would rely
upon medical judgment. These considerations are addressed by LSRO in the following design of an adverse event
monitoring program for dietary supplements.
IV.2 DESIRED FEATURES OF A POSTMARKETING SURVEILLANCE PROGRAM FOR DIETARY
SUPPLEMENTS
In the case of dietary supplements and other food-related products that do not require premarket approval, an
effective system can be designed to collect and analyze information on adverse events (including adverse effects).
However, surveillance cannot achieve its purpose if it is developed in an ad hoc fashion. In this section, LSRO
recommends the desired features of an adverse event monitoring program for dietary supplements.
IV.2.1 Overview of recommendations
For a postmarketing surveillance program for dietary supplements to be developed and utilized, it will need an
identifiable organization to take responsibility for the program, qualified operators, and identifiable reporters who
supply data of adverse events. Moreover, as discussed in Section II, there are many operational issues with the
collection, management, and analysis of IDRs. Data collection and creation of an IDR require an avenue through
which product user experiences are transmitted, a reporting instrument, and data entry and coding. Furthermore,
data management requires systems to store, retrieve, analyze, and report the results of signaling evaluations. A
comprehensive system for surveillance of adverse events would also include issues resolution and intervention,
outreach, and system oversight. Recommendations in this LSRO report address these activities.
Product
Users
Product
Initial
Complaints
Data Collection (intake) and Creation of Individual Data Records
• Reporting instrument (standardized)
• Reporting routes (e.g., Internet, telephone, facsimile, mail)
• Customer assistance (trained personnel)
• Data entry and coding
• Follow up (e.g., product user, health care professional, medical records)
•Quality assurance
Individual Data Records
Data Management and Signaling Using Databases of Individual Data Records
• Document storage and retrieval
• Data storage and retrieval (i.e., relational database)
• Data analysis for signal generation and development of case-series
• Monitoring and evaluation of potential signal
• Quality assurance
Potential Signals
Research and Development for Signaling from Other Information Sources
• Literature search
• Epidemiological studies
• Experimental studies
• Other investigations (e.g., focus groups, production line reports, chemical analysis)
Strengthened Signals
Issues Resolution and Formulation of Risk Management Interventions
• Correction of identifiable problems
• Preventative measures
• Communication and summary reports
Improved Product Quality
Product Literature Education
Figure 4-1. Flow diagram of operational components of a surveillance program for dietary supplements.
Third-party Oversight of Quality and Use of Surveillance Program
The necessary operational components of a surveillance program for dietary supplements are presented in Figure
4-1. Embedded in these key components are clear standard operating procedures (SOP), safeguards to protect
individual privacy, quality assurance measures, and oversight. The surveillance program for dietary supplements
should be validated and permit the manufacturer to collect and analyze product user complaints in a timely manner
for rapid identification of potential product problems.
Tailoring the various database management methods applied in prescription drug surveillance for the dietary
supplement industry would be a practical first step in creating a surveillance program for supplements.
Modifications would be needed for application to dietary supplements in that distinct differences exist in the reporter
base, which is likely to consist of product users rather than health care professionals, and differences in the type of
product.
In the discussion that follows, LSRO presents one general recommendation for postmarketing surveillance programs
along with specific minimal elements for each of the functional system components (Table 4-1). In addition, nine
system enhancements are recommended. The phrasing of recommendations includes the term “should” to convey
the features that effective future systems for dietary supplements would possess and should not be construed as a
value judgment of past procedures used by the dietary supplement industry. These recommendations are applicable
to company-specific as well as industry-wide programs. Some issues unique to industry-wide surveillance programs
are also considered.
Table 4-1. Summary of primary recommendations for surveillance programs for dietary supplements
SYSTEM COMPONENT AND MINIMAL ELEMENTS
ENHANCEMENTS
Reporting and intake processes
1. The system should have the flexibility to accommodate collection • Research should be conducted
and processing of data sent through various common
to determine the barriers to
transmissions.
submitting accounts of adverse
2. Processes for submission of adverse event experiences should be
events for dietary supplements
user-friendly, simple, and concise for reporters in all geographic
and provide solutions.
areas where the product is sold.a
• Ideally, the system should use
3. Confidentiality of data must be safeguarded.
direct, immediate, and
4. Customer service representatives should be trained to ensure that
electronic data entry with
they respond adequately to reporters and that information
built-in logic.
necessary for signal generation is received.
5. The individual submitting an adverse event should be directed to
an appropriate resource that is available 24 hours a day, 365 days
a year for medical-related inquiries.
Creation of IDRs
1. A standardized questionnaire form for recording adverse events
• Up-to-date, accurate, and
should be used, developed with input from those with medical
complete dietary supplement
expertise, as well as those knowledgeable of information
information should be
technology and how the information on the form will be
maintained in a product
processed.
database for assistance in
2. The system should focus on collecting data that are useful for
concomitant product
detecting, prioritizing, and evaluating signals (See Table 4-2).
identification and coding.
3. The system should process incoming data and IDRs in a consistent • Ingredient reference databases
and accurate manner.
should be developed (i.e.,
4. Follow up should be conducted to obtain information relevant for
contact information for
signal generation.
sources, monographs on
5. Company representatives should be knowledgeable and skilled in
pharmacologically active
data coding and entry. They should have the coding tools needed.
substances, methods for
6. The status of IDRs during processing and ongoing related
chemical analysis).
activities should be readily evident.
7. Qualified medical professionals should supervise the medical
evaluation of IDRs.
8. Source materials (e.g., paper materials, computer files) should be
securely stored and readily retrievable.
Building a relational database of IDRs
1. Data should be able to be securely exchanged between internal
• An industry-wide database
and external databases of IDRs.
should collect IDRs to support
inter-product data signaling.
1. Only those IDRs with product-events or outcomes that are judged
• The system should have
to be serious and/or unexpected should be considered in signal
automated signal generation
evaluation.
capabilities.
2. Signaling should be designed and managed by professionals
• The dietary supplement
possessing the requisite technical knowledge and experience to
industry should collaborate
perform operational tasks effectively and provide intelligent
with experts in key areas of
analysis.
emerging technologies and
3. Appropriate statistical methods should be applied to examine
statistics related to automated
changes in the pattern of reporting rates.
surveillance systems.
4. Once a potential signal surfaces from disproportional type
analyses, the product-event combination should be examined in
the context of the quality, quantity, and dispersion of previous
IDRs (i.e., case series).
5. The review and analysis of IDRs should be timely, so that notices
of potential signals are rapidly disseminated to those who can
minimize product risk, especially when serious risks are
identified.
SYSTEM COMPONENT AND MINIMAL ELEMENTS
1. Active postmarketing surveillance should be undertaken to
investigate potential signals; resources should be allocated to
ensure adequate investigation of signals.
Issues resolution and risk management
1. Companies should provide an organizational procedure by which
signals, as well as other relevant information (e.g., studies,
medical literature) are formulated into specific issues. Each issue
that is identified should be formally assessed in terms of its
potential public health impact.
2. Company executives responsible for safety oversight should be
informed of new, serious adverse events related to their product
and they should take appropriate action to protect public health.
3. The U.S. Food and Drug Administration should be informed of
new, serious adverse events related to the product.
4. Outreach should be adequate to inform product users and others
about new health-related findings associated with product use.
Quality assurance and oversight
1. A quality assurance program should be developed to ensure that
the surveillance program is functioning to obtain high quality data,
that data are processed in a timely manner, without error or
distortion, and that confidential data are not disclosed.
2. Objective performance measures of the surveillance program
should be evaluated.
ENHANCEMENTS
A thorough review of the literature
should be conducted to help
investigate potential signals.
•
An independent body of
experts should provide
oversight for the surveillance
program to ensure quality,
review outcome measures of
success, and identify areas for
improvement.
a
The term “reporters” refers to the individuals who submit health-related complaints. This generally includes
product users, health care professionals, relatives, and lay advocates. IDR: individual data record.
IV.2.2 General recommendation
Surveillance programs for dietary supplements should be practical, active, auditable, and
connectable.
• To be practical, the postmarketing surveillance program should be manageable, professional, efficient,
and cost effective.
• To be active, the program should be utilized by those with unexpected and/or serious adverse events,
have the capacity to produce timely signals, and be responsive to product users and issues in need of
resolution.
• To be auditable, the program should have adequate SOP, retain documented accounts, and have
measures of performance that are examined and verified through quality assurance and oversight.
• To be connectable, the surveillance program should be able to move confidential IDR information
securely between designated databases.
IV.2.3 Recommendations for reporting and intake processes
IV.2.3.1 Minimal elements for reporting and intake processes
1. The system should have the flexibility to accommodate collection and processing of data sent through
various common transmissions.
• Multiple entry points into the system should be provided. At a minimum, systems should accept data
by telephone (e.g., direct person-to-person, voice mail, automated audio prompts), and mail (i.e.,
standardized form and freestyle note).
• A procedure should be in place to ensure daily collection and screening of incoming product user
experiences that arrive by all avenues (e.g., voice mail, company website, facsimile).
2.
Processes for submission of adverse event experiences should be user-friendly, simple, and concise
for reporters in all geographic areas where the product is sold.
• Multiple options should be available for individuals to submit information. These options should be
widely publicized (e.g., a toll-free number on the package label and product website).
• Submission procedures should be understandable and usable by individuals with limited educational
background or disabilities, and by individuals of different cultures.
3.
Confidentiality of data must be safeguarded.
• Policies on confidentiality should be clearly communicated to all parties and should be consistent with
established federally mandated requirements.
• Informed consent must be obtained before medical records are accessed.
• Methods to protect patient and reporter confidentiality should not, if possible, obstruct the ability to
conduct follow-up investigations (Allan, 1992c).
4.
Customer service representatives should be trained to ensure that they respond adequately to
reporters and that information necessary for signal generation is received.
• Company representatives should be trained to collect relevant, complete information with exceptional
tact and confidentiality for both the product user and the reporter. They should facilitate the exchange
of information— maximizing the quality of the record by actively listening to the reporter and
knowing when to probe to elicit key elements. Consideration should be given to providing staff with
medical training (e.g., nurses) to manage intake. Cases should be based on first-hand information
rather than hearsay or rumor.
• Evidence of causality is not a requirement for reporting an adverse event.
5.
The individual submitting an adverse event should be directed to an appropriate resource that is
available 24 hours a day, 365 days a year for medical-related inquiries.
IV.2.3.2 Enhancements for reporting and intake processes
1.
Research should be conducted to determine the barriers to submitting accounts of adverse events for
dietary supplements and provide solutions.
2.
Ideally, the system should use direct, immediate, and electronic data entry with built-in logic.
• Technology for electronic transmission does exist, and some have proposed it as the single most
important way to improve adverse prescription drug IDR quality (U.S. Food and Drug Administration,
2003d). An electronic transmission presumably could accomplish the same for dietary supplement
IDR quality. Resources for information on transmission systems are provided in Appendix G.
• Such a system could accomplish data receipt, coding, and review more rapidly and with better record
completion than paper-based systems. Automated data editing can prevent common errors, such as
misspellings or illogical entries that make accurate data retrieval difficult (Mangano, 2002).
• The extent of the data collected may be guided by a decision-tree format during interactive data
collection via the telephone or Internet. Depending on the answers to pivotal questions about the type
of product or about the type of complaint, the appearance of one line of questioning may be triggered
versus another line of questioning. For example, the complaint of rash or wheezing might trigger a
question about known drug allergies, other allergies, other skin problems, or related health problems
(e.g., asthma).
IV.2.4 Recommendations for creation of individual data records
IV.2.4.1 Minimal elements for creation of individual data records
1.
A standardized questionnaire form for recording adverse events should be used, developed with
input from those with medical expertise as well as those knowledgeable of information technology
and how the information on the form will be processed.
• Overall format. The form should be organized in a logical order that facilitates retrieval of critical
information. Data needed for signal generation should be prominent and adjoined on intake forms.
•
•
Data fields. All data fields should be large enough to collect and transmit pertinent information
without truncation or loss. Each question should have its own data field. Avoid pairing multiple
questions with one data field because middle and latter questions tend not to be answered.
Response options. Capture signs and symptoms of the event in the words of the reporter. Limit the
amount of open-ended questions. Default values should only be used with caution because they can
convey misleading information if the question is not answered. Discourage data fields from being left
unanswered (blank) by including options so that unknowns are logged.
2.
The system should focus on collecting data that are useful for detecting, prioritizing, and evaluating
signals.
• Certain data elements that the Committee selected as important for signal generation are presented in
Sections II and III. Examples of the kinds of information that would be helpful for signal generation
are presented in Table 4-2. This should not be considered a comprehensive list, but one that can be
tailored to the needs of the particular product.
3.
The system should process incoming data and IDRs in a consistent and accurate manner.
• The goal of data entry and coding is to facilitate entry of data in an objective standardized format
without distortion or misinterpretation (Allan, 1992c).
• IDRs should be reviewed for completeness and accuracy.
• Coding procedures and training should be developed and periodically modified to include examples of
non-routine IDRs.
4.
Follow up should be conducted to obtain information relevant for signal generation.
• SOP should include criteria for requesting, retrieving, storing, and investigating any unused portion of
product. Retrieval and testing of product associated with an adverse event may help to determine any
product defects in identity, quality, purity, strength, or composition. Product examination may be
especially relevant for product descriptions such as, “off-color,” “odd taste,” “crumbled tablets,” or
“strange odor” (Allan, 1992c).
• When the event was not expected and/or was serious, an effort should be made to obtain medical
records and to contact the product user’s health care professional, who may provide a more accurate
depiction of the adverse event and yield additional information of value for signal generation. Follow
up may also be of value to determine the outcome for individuals who were advised to suspend
consumption.
5.
Company representatives should be knowledgeable and skilled in data coding and entry. They
should have the coding tools needed.
• Company representatives who enter and code data should be able to extract salient characteristics of a
case, particularly for medically complex or unusual cases, that will have bearing on signal generation.
• Standardized dictionaries for uniform coding of health-related terms assist in mapping event terms
[e.g. the Medical Dictionary for Regulatory Activities (Appendix G)] (International Conference on
Harmonisation, 1998). Moreover, using standardized terms supports relational databases.
• The adverse event should be precisely described in terms of the diagnosis, signs, symptoms, and
laboratory tests (Allan, 1992c). The system should have the capacity to permit the assessor to easily
select variables to add to the case definition using standard coding terminology and then promptly
retrieve those IDRs that meet the case definition to build a case series.
• Conflicting information provided by different sources should be mentioned and the sources
documented (Council for International Organization of Medical Sciences, 2001).
6.
The status of IDRs during processing and ongoing related activities should be readily evident.
• All communication and attempts at communication with the reporter and other individuals involved
with the IDR should be documented.
• Referrals should be documented using standard forms and placed with the original file.
• Plans for future action by the manufacturer (e.g., follow up with product user, chemical analyses of
samples) should be documented.
7.
Qualified medical professionals should supervise the medical evaluation of IDRs.
8.
Source materials (e.g., paper materials, computer files) should be securely stored and readily
retrievable.
• Repeat calls/IDRs potentially providing additional information on an affected individual associated
with an earlier IDR should be identified as such and incorporated into the records of the developing
IDR.
IV.2.4.2 Enhancements for creation of individual data records
1.
Up-to-date, accurate, and complete dietary supplement information should be maintained in a
product database for assistance in concomitant product identification and coding.
• Product and ingredient information should be considered in evaluating whether an adverse event is
expected.
• The Supplement Facts information from the product label (Appendix D) (U.S. Food and Drug
Administration, 1999b) should be included in the database.
• Changes in the product label information including ingredients and any indications, contraindications,
warnings or precautions should be added to the database as the label information is updated. Previous
labeling information should be retrievable for reference.
• Product label information in the database should be checked against the actual label on a routine basis.
• Overall, the system must be flexible and adaptable to ever-changing product reformulations and new
product lines.
2.
Ingredient reference databases should be developed.
• Up-to-date, accurate, and complete sources of ingredients should be maintained in a database for
assistance in tracing/tracking ingredient-related problems.
• Monographs on pharmacologically active ingredients used in the manufacture of dietary supplements
should be prepared.
• Methods for chemical analysis of bioactive and key marker ingredients should be developed and made
available.
IV.2.5 Recommendations for building a relational database of individual data records
IV.2.5.1 Minimal elements for building a relational database of individual data records
1.
Data should be able to be securely exchanged between internal and external databases of IDRs.
IV.2.5.2 Enhancement for building a relational database of individual data records
1.
IDRs should be collected by an industry-wide database to support signaling from inter-product data.
• In the absence of data concerning prevalence rates of adverse events in the general (non-supplementtaking) population, comparisons between supplements and various types of adverse events are
important. For example, in DuMouchel's (1999) signaling methodology, the expected numbers of
adverse events would be calculated by examining the marginal probabilities of each type of adverse
event and the marginal number of adverse events per supplement. With data from only one type of
supplement, this and related approaches to signal detection are simply not viable.
• Comparisons of adverse events should be made across related products (after accounting for market
penetration) to assess the significance of events. For example, by examining adverse event rates for
one class of dietary supplements (e.g., non-caffeine-containing weight-loss supplements) to adverse
event rates for a complementary class (e.g., caffeine-containing weight-loss supplements), much
stronger signal detection might be accomplished. Presumably, the demographics of groups who use
related products are more similar than the demographics of those who use unrelated products.
• Comparisons of adverse events within the same product categories should be made (after accounting
for market penetration) to detect adverse events that might be attributed to product impurities.
Furthermore, to the extent that dosages of active ingredients and suggested number of doses vary
across products within the same category (e.g., all caffeine-containing weight-loss supplements), it
might become feasible to enhance signal detection through the examination of dose-response relations
(i.e., whether higher dosages are accompanied by greater adverse event rates).
IV.2.6 Recommendations for signaling from databases of individual data records
IV.2.6.1 Minimal elements for signaling from databases of individual data records
1.
Only those IDRs with product-events or outcomes that are judged to be serious and/or unexpected
should be considered in signal evaluation.
• The surveillance program should have unambiguous criteria that define expected events and serious
events in order to identify and prioritize those product-event combinations or outcomes that merit
further investigation. Only serious, unexpected events are of concern for signal detection and
evaluation of the product and its ingredients.
2.
Signaling should be designed and managed by professionals possessing the requisite technical
knowledge and experience to perform operational tasks effectively and provide intelligent analysis.
• Even with modern data management capabilities, surveillance programs are far from perfect. Thus,
designers and administrators of surveillance programs for dietary supplements should be prepared for
the eventuality of false positive signals, and should be prepared to adjust critical aspects of their
surveillance systems in order to achieve verifiable results.
3.
Appropriate statistical methods should be applied to examine changes in the pattern of reporting
rates.
• The three statistical methods that are most widely used and have the greatest evidentiary support in
spontaneous surveillance programs for detecting changes in reporting rates are: (1) the proportional
reporting ratios, (2) the Bayesian Confidence Propagation Neural Network, and (3) empirical Bayes
screening (Hauben & Zhou, 2003). The empirical Bayes screening method, used by FDA, is less prone
to false positive signals from confounding variables than the other two numerator-based methods
because it stratifies data by key variables such as age and gender (Hauben & Zhou, 2003).
• The surveillance program should be able to detect (Bright & Nelson, 2002):
A product-event combination that appears sporadically over a long period of time
A gradual increase in the amount of IDRs for a product-event combination
A sudden increase in the amount of IDRs for a product-event combination
A rare outcome among a particular population subgroup because of some shared
characteristic such as a concomitant exposure to medication or other therapy
• Reporting rates should not be used as a surrogate for a measure of incidence, but can be used as
evidence that a particular product-event combination could occur.
4.
Once a potential signal surfaces from disproportional type analyses, the product-event combination
should be examined in the context of the quality, quantity, and dispersion of previous IDRs.
• The database should be searched to locate a series of cases that might strengthen the hypothesis
concerning the association between an adverse experience and product exposure. Once the case series
is assembled, IDRs should be individually reviewed in an attempt to understand as much of the clinical
picture as possible (Faich, 2003). The case series should be viewed only as preliminary supporting
evidence, to be used in assisting in the interpretation of patterns of reporting and to refine important
signals. It does not represent definitive conclusions (Faich, 2003).
• The balance of evidence for the signal must be weighed to determine whether more data are needed or
whether an intervention is warranted. Several types of evidence should be considered (Meyboom et
al., 2002), including:
The volume of IDRs and statistical disproportionality
Characteristics of the product-event relationship (e.g., timing, dose)
Biological plausibility
Experimental evidence from the IDRs (e.g., rechallenge, biochemical tests)
Characteristics of the data quality (e.g., objectivity, completeness)
• Signals become useful when they identify a real risk, such as product defects or conditions that
increase an individual’s vulnerability. Once a signal is evident, then there needs to be a decision of
whether the phenomenon is clinically significant, or is just an anomaly (U.S. Food and Drug
Administration, 2003d). When evaluating potential signals, assessors must strike a balance between
competing objectives. Specifically, assessors balance the desire to avoid misappropriated blame for a
•
5.
product (having an uncertain causal relationship) with the need to protect public health by advancing
true signals.
For potential signals that do not have sufficient supporting cases or other supporting data, monitoring
protocols should be established to periodically search the database for trends in the product-event
combination.
The review and analysis of IDRs should be timely, so that notices of potential signals are rapidly
disseminated to those who can minimize product risk, especially when serious risks are identified.
IV.2.6.2 Enhancements for signaling from databases of individual data records
1.
The system should have automated signal generation capabilities.
• Specialized software (e.g., from a third-party technology vendor) can be utilized for signaling and is
implemented on top of the underlying database architecture. Off-the-shelf systems are available that
are compliant with international standards (i.e., E2B) (International Conference on Harmonisation
Expert Working Group, 2001; U.S. Food and Drug Administration, 2004b). The identification of
specific software and hardware is beyond the scope of this LSRO report. Resources that may be of
further assistance are identified in Appendix G.
2.
The dietary supplement industry should collaborate with experts in key areas of emerging
technologies and statistics related to automated surveillance systems.
IV.2.7 Recommendations for signaling from other information sources
IV.2.7.1 Minimal elements for signaling from other information sources
1.
Active postmarketing surveillance should be undertaken to investigate potential signals; resources
should be allocated to ensure adequate investigation of signals.
• Signals arising from the passive, spontaneous adverse event monitoring program are only suspicions
that a product-event combination might exist. Further studies may be needed to explore whether a
safety risk might indeed exist (i.e., to eliminate false positive signals) and/or to better understand the
product-event association.
• When deciding whether to conduct active postmarketing surveillance studies, individuals with
biostatistical and epidemiological expertise should be consulted for the type of additional information
needed and to evaluate the feasibility of the proposed studies.
• It may be of interest to determine whether the product-event occurs commonly or rarely and whether
there might be a specific population at risk, (e.g., athletes, women of child bearing age, elderly on
certain medication). This information may be useful to understand whether the risks are preventable.
IV.2.7.2 Enhancement for signaling from other information sources
1.
A thorough review of the literature should be conducted to help investigate potential signals.
• The medical literature has a long history of being a valued source of first reports of adverse events
related to drugs and of supportive case reports of adverse events (Rossi & Knapp, 1984). Furthermore,
published literature of human and animal studies may provide clues to biological plausibility for
ingredients and suggest confounding variables, such as supplement-drug interactions.
The
biopharmaceutical companies are required by law to conduct literature searches every two weeks to
look for case reports in any well-read journal around the world. Manufacturers of dietary supplements
engaging in postmarketing surveillance should periodically conduct thorough, up-to-date reviews of
relevant literature.
• The annual reports of the American Association of Poison Control Centers Toxic Exposure
Surveillance System (Litovitz et al., 2002) may be useful to identify supplement-related IDRs
collected by poison control centers in the United States (Appendix F).
IV.2.8 Recommendations for issues resolution and risk management
IV.2.8.1 Minimal elements for issues resolution and risk management
1.
Companies should provide an organizational procedure by which signals, as well as other relevant
information (e.g., studies, medical literature) are formulated into specific issues. Each issue that is
identified should be formally assessed in terms of its potential public health impact.
2.
Company executives responsible for safety oversight should be informed of new, serious adverse
events related to their product and they should take appropriate action to protect public health.
• The level of intervention should reflect the severity, frequency, or duration of the product’s risks. In
general, there will not likely be any definitive proof that a dietary supplement caused an adverse event.
Yet, to protect the public, a high degree of suspicion based on all the available evidence may be all that
is necessary to prompt a manufacturer to withdraw their product’s current formulation or a particular
lot of product from the market or to take other actions. Optimally these decisions should be within a
framework of standards of action developed in advance either by the company, the industry as a whole,
or by a regulatory authority.
• Although finding the root cause of a product quality problem is necessary, the goal is to prevent similar
problems from occurring in the future. Therefore, corrective action should be focused on fixing
systematic production problems.
3.
FDA should be informed of new, serious adverse events related to the product.
• Obviously, concern for public safety should prompt manufacturers to report findings of product-event
associations to FDA as expeditiously as possible. External reporting to federal and other industry
surveillance programs alerts others to new potential signals, so that the experience of a few can be
shared and lead to corrective and preventative measures by all.
• FDA should be informed of proposed company responses. Further analyses by FDA that aggregates
and analyses data from multiple sources may strengthen or lesson support for the potential signal.
4.
Outreach should be adequate to inform product users and others about new health-related findings
associated with product use.
• Official product information (i.e., labels) should convey the known health-risks associated with use of
the product and contraindications of use. This assists product users and their health care professionals
in recognizing known interactions and other special circumstances resulting in adverse events.
• Training materials on safe and responsible product use should be accurate and not misleading.
• The manufacturer can directly notify retailers by letter about newly discovered contraindications for
use, product label changes and/or reformulations made in response to product-event signals.
IV.2.9 Recommendations for quality assurance and oversight
IV.2.9.1 Minimal elements for quality assurance and oversight
1.
A quality assurance program should be developed to ensure that the surveillance program is
functioning to obtain high quality data, that data are processed in a timely manner, without error or
distortion, and that confidential data are not disclosed.
• The quality of the data collected should be evaluated. Quality measures might include parameters that
describe how well the system processes clinically significant IDRs.
2.
Objective performance measures of the surveillance program should be evaluated.
• Metrics of how widely and frequently the system is used among the product’s users may highlight
barriers to submitting accounts of adverse events.
• Performance measures might include parameters that describe how well the system minimizes false
negative signals and avoids the generation of false positive signals.
• Trend data in the system should be examined for continuing arrival of IDRs of a product-event
combination, which could herald a persistent problem. However, because a decrease in reporting rate
does not constitute an assurance that a safety issue was resolved, metrics of outcome performance
should not be based on a reduction of reporting rates (U.S. Food and Drug Administration, 2003c).
•
There should be evidence that product problems are investigated and resolved. Measures of the
efficiency and effectiveness of the resulting interventions should be evaluated. The outcome of the
surveillance program should be actions that negate any immediate health-risks related to a specific
instance and proactive changes that focus on corrective improvements to systems, processes, and/or
products.
IV.2.9.2 Enhancement for quality assurance and oversight
1.
An independent body of experts should provide oversight for the surveillance program to ensure
quality, review outcome measures of success, and identify areas for improvement.
• Specifically, a group of experts, independent of the dietary supplement industry, should be charged
with periodically reviewing the overall system function, including the integrity of the data, and the
value of the signals.
• Oversight for quality assurance should make certain that key components of the system are validated
and updated in-step with technological advances.
Table 4-2. Functional needs of an intake form useful for recording adverse events associated with the use of dietary
supplements
CATEGORY
RATIONALE FOR INCLUSION
Purpose of complaint/report
Minimal data needed for signal detectiona
• Reporter’s purpose for complaint/report Directs incoming communication to customer service
(e.g., to notify the manufacturer, and/or representatives and/or the surveillance program as
appropriate; helps to establish the reporter’s expectations
request advice, and/or obtain a refund)
and directs the type of response.
Dates and times
• Date of receipt of initial complaint/report An identifier for the individual data record; also, can be
used to calculate reporting rates.
(i.e., day/month/year)
•
Whether product was consumed prior to
the event (e.g., dates and times of healthrelated event and last exposure)
Additional data of value for signal prioritization
and evaluation
• Date of first use of product
Identifiers of “case” and source information
• Reporter name and contact information
•
Product user identifier (e.g., initials, date of
birth or age, gender, and/or any externally
applied unique number, such as a hospital
chart number)
and evaluation
• Reporter relationship to affected individual
(i.e., self, relative, health care professional
•
Product user’s contact information if this
information can be kept confidential
Establishes the temporal association of product exposure
prior to onset of the event.
Used to calculate the time to onset from initial exposure
until the event.
Used to obtain follow-up information and/or for
verification; helps to authenticate the complaint/report.
Helps to confirm that there is a singular, authentic “case”
and to avoid duplication of cases. (On rare occasions,
fictitious cases are submitted to damage the reputation of a
manufacturer or to take advantage of an opportunity that
may yield private knowledge about the company
contacted.)
Some data may be more appropriately provided by a health
care professional to increase reliability and accuracy of data
(i.e., event was life-threatening).b
May need to follow up with product user to obtain missing
information and/or obtain informed consent.
CATEGORY
Product user’s health care provider if
treatment was sought (i.e., physician,
facility)
Follow-up health care information and/or verification are
often needed.
•
Contact information for additional sources
of data
May need to contact other sources to complete/correct data
supplied by reporter.
•
Product user’s reproductive status (i.e.,
pregnant, lactating)
Indicates potential additional affected individual, the infant
whose parent consumed the supplement; if yes, events in
the child need to be considered.
•
Geographic location
Data of value for signal prioritization and
evaluation
• Zip Code of product user
Product
• Product identification (i.e., name, brand)
and evaluation
• Manufacturer of product (i.e., name,
telephone number, address)
Helps to identify geographic clusters of records that may
indicate a common cause (e.g., product tampering,
spoilage).
Establishes correct product identification. [Data elements
that can verify product identity (e.g., brand, manufacturer,
lot number) should be incorporated into the data collection
in multiple independent data fields. This assists the
identification of the specific product brand as opposed to a
generic category.
Correct product identification is
important for the integrity of both intra-product and interproduct databases.]
Helps establish correct product identification; provides
contact information to retrieve additional product
information.
•
Batch and/or lot number
These numbers verify product identity and assist in
identifying the specific formulation if product underwent
reformulation; also useful to identify the age of the product
and which product may be defective because these numbers
directly link to specific manufacturing dates and
circumstances.
•
Description of product (e.g., pink tablets,
green powder) and description of a possible
quality defect
Helps establish correct product identification because a
unique trade dress (appearance of the product, e.g., oval
green tablets stamped with brand initials) can distinguish
one brand from another brand, especially if the product was
separated from the bottle; helps identify a possible quality
defect.
•
Product ingredients and labeled amount of
ingredients per serving
Identifies possible pharmacologically active substances
associated with the event and amounts per unit serving.
(Continued next page)
CATEGORY
Indication for use
evaluation
• Labeled indications for use
•
Indications for why the consumer used the
product
Product use
evaluation
• Labeled instructions for use (e.g., serving
size, number of servings per day) and
labeled warnings, precautions, and
contraindications for use
•
Product user’s last dose amount, typical
dose, doses per day, duration of use, and
whether use was continuous
Event and outcomes
• Reporter’s verbatim description of healthrelated event (e.g., onset, signs and
symptoms, medical diagnoses)
and evaluation
• Clinical course and whether event is
ongoing or has fully resolved
Health care
evaluation
• Did the product user visit a health care
professional? Did the event require an
emergency room visit?
Helps to define “off label” uses.
Identifies intended and unintended (off label) use; may
provide clues to the product-event relationship. [For
example, a product used to treat a terminal illness (an off
label use) might likely be linked to fatal cases.]
Provides boundaries for compliance with “used as directed”
criteria. (Labeled warnings establish expected events.)
Identifies problems that may be associated with a dose
response (e.g., overdose); establishes whether product was
used as directed; identifies possible dechallenge and/or
rechallenge cases.
Classifies the event as a health-event rather than some other
type of communication (e.g., testimonial, product defect
without a health-related event); enables categorization of
event into standardized medical terminology; enables
formation of product-event combination(s); can evaluate
whether symptoms were consistent with known biological
effects of an ingredient or contaminant.
Provides indication of the stage of the event and whether
follow up may be needed; data helps in judging the
seriousness of the event and outcome (e.g., life-threatening,
fatal, disabling).
Indicates whether medical records might be available.
•
Did a health care professional diagnose
problem and what diagnostic tests were
conducted? What treatment was provided?
Helps to correctly specify the health-related event.
•
Was the product user admitted to a
hospital?
Hospitalization is a criterion for indicating the seriousness
of the event.
CATEGORY
Demographic data and potential confounding factors
evaluation
• Demographic data (i.e., age, gender, height,
weight); health problems prior to initiating
product (e.g., heart disease, high blood
pressure, allergies); medications, food,
alcohol, other dietary supplements taken at
the time of using the product (e.g., dose
and duration of use); dietary restrictions
(e.g., weight loss diet); reproductive status
(i.e., pregnant, lactating); physical factors
(e.g., activity level, dehydration);
environmental and other factors.
Etiological data
and evaluation
• Did symptoms abate when substance was
removed (dechallenge)? Did symptoms
return when substance was reintroduced
(rechallenge)? Was there a prior reaction to
the same ingredients or substance?
Helps to identify subpopulations of product users that may
be susceptible to events. (When possible, document the
true absence of any known disease prior to the event and an
absence of other confounding factors. For example, the
possibility of supplement/drug and supplement/supplement
interactions is eliminated or raised.)
A positive rechallenge and other etiological data support
the association of the product and the event.
•
Other causal testing evidence indicative of
a particular toxicity, such as biochemical
tests (e.g., high serum levels of ingredient,
specific antibody tests to an ingested
substance, concentration of substance in
biopsy specimen), and certain characteristic
physical findings.
a
An individual data record that meets the minimal criteria for data collection will inform a manufacturer that a
health-related event has occurred but does not necessarily mean the information is sufficient to adequately address
the case or add knowledge useful to the safety profile of a product (Council for International Organization of
Medical Sciences, 2001).
b
When the event is serious or is not labeled (not expected) an effort should be made to obtain consent and contact
the product user’s physician for follow-up information (Council for International Organization of Medical
Sciences, 2001).
IV.3 DISCUSSION AND SUMMARY
IV.3.1 Summary of considerations for the design of a postmarketing surveillance program for dietary
supplements
The use and availability of dietary supplements is widespread among all segments of the U.S. population, from
children to the elderly, from the ill to the athlete. Although manufacturers are responsible to assure that their
products on the market are safe, defective products and unknown product-product interactions are possible. A welldesigned postmarketing surveillance program provides arrangements for dealing with abnormal occurrences at the
outset of introducing a dietary supplement product into the market.
The accumulation of well-documented IDRs can increase general knowledge of the product in question, perhaps by
identifying those product users at greatest risk. This is accomplished if spontaneous reporting schemes are designed
to signal new, rare, unusual, and serious adverse events.
IV.3.2 Summary of desired features and recommended enhancements of postmarketing surveillance programs
for dietary supplements
The minimal elements and suggested enhancements of adverse event monitoring programs for the dietary
supplement industry are summarized in Table 4-1. Resources to assist the development of such systems are
presented in Appendix G.
Quality assurance and oversight should play a major role in the early development and testing of the necessary
processes, procedures, and technical infrastructure of the surveillance system before a full-scale system is
implemented. Operations at each intermediary step from the submission of a product user experience through
detection, prioritization, and evaluation of signals should be assessed. After implementation, periodic evaluation for
quality may lead to alterations or redesign to better achieve the overall goal of having a pragmatic and effective
system in place to protect public health. It is expected that a well-run surveillance program will become a routine
operational activity. Then, if a high-profile emergent situation arises, data and past analyses in the adverse event
monitoring program provide a foundation for understanding and responding to the current situation.
In addition to postmarketing surveillance of spontaneous IDRs, periodic reviews of the literature and observational
and clinical studies all contribute data to help understand the impact of a product on potential health effects.
Judgments are needed to determine whether the new data reflect a meaningful change in the product’s safety profile
and whether some form of intervention is necessary, such as updating product labels to reflect new knowledge.
Construction and utilization of a surveillance program can present sizable technical, financial, and organizational
challenges. It is recognized that small companies initiating a surveillance program will require time to effectively
implement the full range of proposed activities, because some techniques require special expertise that is costly and
not widely available. Expecting the smaller companies to allocate resources to their own system does not seem
feasible. Industry-wide (preferred) or regulatory authority sponsorship of a surveillance program for dietary
supplements could aid construction and utilization of such programs.
IV.3.3 Opportunities for collaboration
In its report to the President, the Commission on Dietary Supplement Labels (1997) urged FDA, the industry, the
scientific community, and consumer groups to work together voluntarily to improve passive postmarketing
surveillance systems to ensure that any safety problems that may arise from use of dietary supplements are identified
and corrected promptly. Similarly, Clark et al. (2001) emphasized that constructive cooperation among multiple
stakeholders including manufacturers, regulators, academicians, health professionals, and product users was
essential for effective postmarketing surveillance. By example, the collaboration of those involved with the
postmarketing surveillance of drugs led to international standards for uniform definitions, reporting procedures, and
a single reporting form, which with time and experience have been further modified (Faich et al., 1990).
LSRO also encourages collaboration by those involved with manufacturing and regulation of dietary supplements to
work toward achieving components of surveillance that may be shared. Some targeted activities for development
that would support the implementation and utilization of surveillance programs for dietary supplements would
include:
• Coordinating efforts among those in the private sector and regulatory authorities where the possibility
exists to create common electronic gateways for shared data entry and for the development of needed
software and analytical techniques.
• Organizing a standard-setting group from industry, regulatory authorities, and academia to convene and
apply for grants to develop SOP.
• Improving reference databases that might be shared (i.e., product label databases, ingredient monographs,
methods of chemical analysis of bioactive and key marker ingredients).
• Conducting studies that further the knowledge of background incidence rates of health problems in the
population that are independent of the use of dietary supplements.
IV.3.4 Summary of postmarketing surveillance programs for dietary supplements
The strongest surveillance programs will be derived from inter-product IDR databases (i.e., contains product data
from different manufacturers). Therefore, LSRO advises that an inter-product surveillance program would be the
best type for industry to develop. In the absence of participation in such a system, LSRO recommends that
manufacturers of dietary supplements consider voluntary intra-product surveillance programs. A rudimentary
program with certain basic elements can be improved upon later. Surveillance programs could be developed in tiers,
the first tier consisting of manufacturer brand-specific programs, the second tier consisting of larger, more complex,
outsourced programs, and the third tier consisting of industry-wide, inter-product systems.
No one system will suffice for postmarketing surveillance of dietary supplements. All likely avenues that product
users and health care professionals will use to request or submit information should be viewed as complementary
and should be cultivated and combined for maximal detection and analysis of product-event relationships with the
goal of preventing future problems and protecting the public health.
V. CONCLUSIONS
This Life Sciences Research Office (LSRO) report is intended to facilitate the design and development of adverse
event monitoring programs for dietary supplements. If implemented, such systems better inform risk management
choices by manufacturers, regulators, and by product users, and thereby reduce potential health risks from dietary
supplements.
In order to meet the objectives of the study, LSRO reviewed postmarketing surveillance for dietary supplements and
other consumer products through several efforts including: (1) the use of a case study for Metabolife 356® and other
ephedrine alkaloid-containing dietary supplements; (2) a review of several noteworthy surveillance programs
described in the scientific literature; and (3) evaluation of input from speakers and agency representatives familiar
with approaches to postmarketing surveillance in the areas of dietary supplements, food additives, and prescription
and nonprescription medication.
Findings and conclusions on the safety of dietary supplement products are outside the scope of work described in
this report. Although LSRO had access to specific databases of adverse event records related to ephedrine alkaloidcontaining dietary supplements, neither the Committee nor LSRO made any evaluation or conclusion of the safety or
efficacy of ephedra products. Specific signals were not identified. This report does not support any conclusion on
the safety of dietary supplement products.
V.1 SUMMARY OF FINDINGS
V.1.1 Identification of the necessary data for signal generation
LSRO built its analysis around the specific steps of signal detection, signal prioritization and signal evaluation.
Signal detection is a process of sorting and identifying individual data records (IDRs) in order to find interesting
dietary supplement-adverse event relationships that merit further attention. The Committee determined that, at
minimum, five criteria must be met by an IDR in order to include that data in signal detection:
1.
2.
3.
4.
5.
An identifier for the consumer was obtained
The complaint was health-related
A dietary supplement product was consumed prior to the onset of the health-related complaint
The IDR was dated (day/month/year)
An identifier for the reporter was obtained
in preparation for signal evaluation. The Committee determined that two types of data are needed for signal
prioritization: (1) data that rank the reliability and accuracy of the record; and (2) data used to rank the seriousness
of the adverse event or outcome.
is indicated. The Committee identified eleven types of data that are useful for signal evaluation:
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Gender
Age at the time of the event
Pre-existing medical condition
Information on concomitant exposure to prescription and/or nonprescription medications
Information on concomitant exposure to other dietary supplements
Information on why the product was used
Batch/lot number
Daily dose of bioactive ingredients
Time from initial use to onset of the adverse event
Rechallenge information
Geographic location (e.g., Zip Code)
V.1.2 Principal findings of the LSRO Phase I review of individual data records
It should be noted that the records from Metabolife International, Inc. (MET) represented information collected, in
most cases, during a single telephone contact for purposes of customer service rather than postmarketing
surveillance. In contrast, records from the U.S. Food and Drug Administration (FDA) represented information
collected in an initial contact and in follow up by FDA field representatives for purposes of postmarketing
surveillance.
In terms of the LSRO review of the sample sets of IDRs, the Committee observed the following:
• LSRO concluded that the records collected by MET were qualitatively less informative than those collected
by the FDA and therefore pose a greater challenge for application in public health related analyses
(Appendix B).
• The limitations in signal generation for both sample sets stemmed mostly from systematic problems in the
methods used for initial data collection coupled with insufficient follow up.
• Despite these limitations, it is likely that signals (hypotheses) worthy of further investigation would arise
from such data sets if they were developed in the context of a continuing surveillance program having the
requisite functional components detailed in Section II.
V.1.3 Principal findings of the LSRO Phase II review of existing postmarketing surveillance programs
The principal findings of the LSRO review of postmarketing surveillance programs reported in the literature for
food additives, medications, and other products include the following:
• A properly designed and implemented surveillance program for dietary supplements can be an important
component of a proactive risk management plan.
• Evidence of causality is not a requirement for signal detection.
• An adverse event monitoring program for dietary supplements will share many of the challenges of other
such reporting systems in terms of specificity, amount of data, completeness, limitations with analysis, and
interpretation of findings.
• Current technologies make it feasible to implement quality systems but manufacturers may need qualified
operators and further guidance to ensure they are using the best possible review practices.
• For a product in which signals arise, especially if at risk populations have not been adequately studied, a
collection of additional information beyond routine postmarket reporting will usually be needed to
determine if the signal is real and to determine the appropriate actions to be taken, if any.
• Surveillance programs in the private sector will complement federal efforts to monitor adverse events
associated with dietary supplements. No one system will suffice for postmarketing surveillance of dietary
supplements. All likely avenues that product users and health care professionals will use to request or
submit information should be viewed as complementary and should be cultivated and combined for
maximal detection and analysis of product-event relationships with the goal of preventing future problems
and protecting the public health.
V.2 SUPPLEMENTAL REMARKS
LSRO was not charged with assessing the value of postmarketing surveillance programs. Thus, the Committee did
not investigate the legal and ethical implications of a manufacturer establishing such a program, and did not
thoroughly evaluate the broader societal or competitive business implications. Nevertheless, during the course of
their review of the properties of surveillance programs, the Committee found sufficient and compelling scientific
evidence to suggest that an effective postmarketing surveillance program for dietary supplements can be of great
value to manage potential risks associated with these products. Hence, even though there are no explicit
requirements for dietary supplement manufacturers to submit IDRs to FDA or to analyze IDRs in aggregate, LSRO
proposes that the dietary supplement industry (broadly defined) consider establishing a postmarketing surveillance
program either individually or collectively and not rely solely on federal programs. LSRO advises maximizing as
far as possible the value of the effort sooner rather than later.
Specifically, LSRO suggests that manufacturers of dietary supplements that do not already have a surveillance
program in place consider developing, implementing, and managing such a system for their products.
Recommendations and additional information provided throughout the body of this report may also be useful to
enhance the speed and quality of information of existing postmarketing surveillance programs.
LSRO recognizes that tracking the short-term and long-term health consequences associated with use of dietary
supplements is a substantial undertaking. Depending on the size of the company and the existing internal system for
adverse event monitoring, companies could assemble the LSRO recommendations from this report into tiers and
implement the program in phases. These systems need not be technically sophisticated, so long as they are
organized and systematic in their application to take in reliable information and to produce credible results. Once
implemented, a basic system of surveillance with certain key elements can be improved upon later.
The numbers of adverse events associated with any particular product are expected to be relatively low. Thus, the
ability to detect adverse events is severely limited if small companies, each with limited market share, track products
separately from other similar products. Legal, ethical, and competitive business issues aside, LSRO believes there
would be a substantial benefit in being able to pool data across manufacturers and supplements. Having a
centralized system would greatly increase the probability that important supplement induced adverse events will be
detected because it would permit comparisons of adverse event rates across the product user population. The
development of an industry-wide, inter-product database of IDRs for signal generation should be seriously
considered for postmarketing surveillance.
Many of the limitations common to postmarketing surveillance systems used by the U.S. Food and Drug
Administration, the Poison Control Centers, and the World Health Organization were identified in this report and
include expense, the lack of proof of causation, possibly misleading data, and the need for expert operators – yet we
still see value in such systems. LSRO believes that properly designed and implemented postmarketing surveillance
programs tailored for dietary supplements will enhance the likelihood that true adverse effects, unknown at the time
of marketing, are identified. Even a single case of a suspected serious adverse event could potentially add to general
medical knowledge. Moreover, potential product problems that are detected by surveillance programs, such as a
defective product lot, can be rapidly investigated and addressed. In this way, product quality is controlled.
Postmarketing surveillance provides an opportunity to ensure that the adverse experiences of a few product users are
properly evaluated to reduce the risk of similar events for future product users.
Coupled with good manufacturing practices, comprehensive postmarketing surveillance can be part of a dietary
supplement industry program to assure customers that dietary supplement products are without unexpected harmful
health effects. An effective surveillance program will enable the manufacturer to be responsive to reporters of
adverse events, regulatory authorities, and others in industry. Thus, an effective adverse event surveillance system
is an extremely valuable tool for risk management.
In its final analysis, LSRO notes that the proven value of existing surveillance systems is an extremely strong
argument in favor of a proactive approach.
V.3 SUMMARY OF RECOMMENDATIONS
In general, LSRO recommends that surveillance programs for dietary supplements should be:
• Practical. To be practical, the postmarketing surveillance program should be manageable, professional,
efficient, and cost effective.
• Active. To be active, the program should be utilized by those with unexpected and/or serious adverse
events, have the capacity to produce timely signals, and be responsive to product users and issues in need
of resolution.
• Auditable. To be auditable, the program should have adequate standard operating procedures, retain
documented accounts, and have measures of performance that are examined and verified through quality
assurance and oversight.
• Connectable. To be connectable, the surveillance program should be able to move confidential IDR
information securely between designated databases.
LSRO recommendations are summarized in Table 4-1 (Section IV). LSRO characterizes its recommendations as
falling into two broad categories—minimal elements and recommended enhancements. LSRO strongly encourages
industry to consider the recommended enhancements, especially to consider an industry-wide surveillance program
that would support signaling from inter-product data. This type of system coupled with independent, objective
review and oversight would greatly enhance the dietary supplement industry’s ability to detect and correct true
adverse effects of its products and provide excellent risk management.
LITERATURE CITATIONS
Ajayi, F. O., Sun, H. & Perry, J. (2000) Adverse drug reactions: a review of relevant factors. J. Clin. Pharmacol. 40:
1093-1101.
Allan, M. C. (1992a) A primer of drug safety surveillance: an industry perspective. Part I: Information flow, new
drug development, and federal regulations. J. Pharm. Technol. 8: 162-167.
Allan, M. C. (1992b) A primer of drug safety surveillance: an industry perspective. Part II: Product labeling and
product knowledge. J. Pharm. Technol. 8: 198-202.
Allan, M. C. (1992c) A primer of drug safety surveillance: an industry perspective. Part III: Managing adverse-event
data. J. Pharm. Technol. 8: 259-273.
Allgood, G. S., Kuter, D. J., Roll, K. T., Taylor, S. L. & Zorich, N. L. (2001) Postmarketing surveillance of new
food ingredients: results from the program with the fat replacer olestra. Regul. Toxicol. Pharmacol. 33: 224-233.
Avorn, J. (1990) Reporting drug side effects: signals and noise. JAMA 263: 1823.
Bortnichak, E. A., Wise, R. P., Salive, M. E. & Tilson, H. H. (2001) Proactive safety surveillance.
Pharmacoepidemiol. Drug Saf. 10: 191-196.
Bright, R. A. & Nelson, R. C. (2002) Automated support for pharmacovigilance: a proposed system.
Butchko, H. H. & Stargel, W. W. (2001) Aspartame: scientific evaluation in the postmarketing period. Regul.
Toxicol. Pharmacol. 34: 221-233.
Butchko, H. H., Tschanz, C. & Kotsonis, F. N. (1996) Postmarketing surveillance of anecdotal medical complaints.
In: The Clinical Evaluation of a Food Additive: Assessment of Aspartame. (Tschanz, C., ed.). CRC Press, Boca
Raton, FL. pp. 183-193.
Clark, J. A., Klincewicz, S. L. & Stang, P. E. (2001) Spontaneous adverse event signaling methods: Classification
and use with health care treatment products. Epidemiol. Rev. 23: 191-210.
Commission on Dietary Supplement Labels. (1997) Report of the Commission on Dietary Supplement Labels.
Available at http://www.health.gov/dietsupp/final.pdf.
Council for International Organization of Medical Sciences (2001) Current Challenges in Pharmacovigilance:
Pragmatic Approaches. Council for International Organization of Medical Sciences, Geneva, pp. 1-381.
Crawford, L. M. (5-29-2002) Dietary supplement strategic plan cost out. (Report to Congress). [House Committee
(Report No. 106-619)]. Available at http://www.cfsan.fda.gov/~dms/ds-stra2.html. Accessed 8-1-2002.
DuMouchel, W. (1999) Bayesian data mining in large frequency tables, with an application to the FDA spontaneous
reporting system. Am. Stat. 53: 177-190.
Ervin, R. B., Wright, J. D. & Kennedy-Stephenson, J. (1999) Use of dietary supplements in the United States, 198894. Vital Health Stat. 11 244: 1-14.
Faich, G. (2003) Comments on FDA's pharmacovigilance and pharmacoepidemiologic assessment concepts.
Available at http://www.fda.gov/cder/meeting/RMtranscript3.doc. Accessed 7-18-2003.
Faich, G. A., Castle, W. & Bankowski, Z. (1990) International reporting on adverse drug reactions: the CIOMS
project. CIOMS ADR Working Group. Int. J. Clin. Pharmacol. Ther. Toxicol. 28: 133-138.
Faich, G. A., Dreis, M. & Tomita, D. (1988) National adverse drug reaction surveillance. 1986. Arch. Intern. Med.
148: 785-787.
Foote, J. A., Murphy, S. P., Wilkens, L. R., Hankin, J. H., Henderson, B. E. & Kolonel, L. N. (2003) Factors
associated with dietary supplement use among healthy adults of five ethnicities: the Multiethnic Cohort Study. Am.
J. Epidemiol. 157: 888-897.
Greger, J. L. (2001) Dietary supplement use: consumer characteristics and interests. J. Nutr. 131: 1339S-1343S.
Hauben, M. & Zhou, X. (2003) Quantitative methods in pharmacovigilance: focus on signal detection. Drug Saf. 26:
159-186.
Hennessy, S. (1998) Postmarketing drug surveillance: an epidemiologic approach. Clin. Ther. 20 Suppl C: C32-C39.
Hoggatt, K. J., Bernstein, L., Reynolds, P., Anton-Culver, H., Deapen, D., Peel, D., Pinder, R., Ross, R. K., West, D.
W., Wright, W., Ziogas, A. & Horn-Ross, P. L. (2002) Correlates of vitamin supplement use in the United States:
data from the California Teachers Study cohort. Cancer Causes Control 13: 735-740.
Hostelley, L. S. (2003) Comments Paper #3: Risk Assessment of Observational Data. Available at
http://www.fda.gov/cder/meeting/RMtranscript3.doc. Accessed 7-18-2003.
International Conference on Harmonisation. (1996) ICH Harmonised Tripartite Guideline. Clinical safety data
management: periodic safety update reports for marketed drugs. Available at http://www.ich.org/pdfifpma/e2c.pdf.
Accessed 7-25-2003.
International Conference on Harmonisation. (1998) MedDRA. Available at http://www.ich.org/ichMedDRA.html.
Accessed 9-29-2003.
International Conference on Harmonisation Expert Working Group (1994) ICH Harmonised Tripartite Guideline.
Clinical Safety Data Management: Definitions and Standards for Expedited Reporting.
International Conference on Harmonisation Expert Working Group (2001) ICH Harmonised Tripartite Guideline.
Maintenance of the ICH Guideline on Clinical Safety Data Management: Data Elements for Transmission of
Individual Case Safety Reports.
Kahn, S. N. (2003) Comments on FDA Concept Paper. Risk Assessment of Observational Data. Available at
Leape, L. L. (2002) Reporting of adverse events. N. Engl. J. Med. 347: 1633-1638.
Letkeman, C. (2003) (Third) Declaration of Caroline Letkeman. Available at http://xenu.ca/ex/cl-declaration3.html. Accessed 5-20-2003.
Litovitz, T. L., Klein-Schwartz, W., Rodgers, G. C., Jr., Cobaugh, D. J., Youniss, J., Omslaer, J. C., May, M. E.,
Woolf, A. D. & Benson, B. E. (2002) 2001 Annual report of the American Association of Poison Control Centers
Toxic Exposure Surveillance System. Am. J. Emerg. Med. 20: 391-452.
Lyle, B. J., Mares-Perlman, J. A., Klein, B. E., Klein, R. & Greger, J. L. (1998) Supplement users differ from
nonusers in demographic, lifestyle, dietary and health characteristics. J. Nutr. 128: 2355-2362.
Mangano, M. F. (2002) Testimony of Michael F. Mangano, Principal Inspector General, Department of Health and
Human Services. Available at http://www.senate.gov/~gov_affairs/073102mangano.htm. Accessed 8-29-2002.
Metabolife International, Inc. (2003) Metabolife 356® Available at
http://www.metabolife.com/products/356_detail.htm. Accessed 2-28-2003.
Meyboom, R. H., Egberts, A. C., Edwards, I. R., Hekster, Y. A., de Koning, F. H. & Gribnau, F. W. (1997)
Principles of signal detection in pharmacovigilance. Drug Saf. 16: 355-365.
Meyboom, R. H., Lindquist, M., Egberts, A. C. & Edwards, I. R. (2002) Signal selection and follow-up in
pharmacovigilance. Drug Saf. 25: 459-465.
Miwa, L. J., Jones, J. K., Pathiyal, A. & Hatoum, H. (1997) Value of epidemiologic studies in determining the true
incidence of adverse events. The nonsteroidal anti-inflammatory drug story. Arch. Intern. Med. 157: 2129-2136.
National Nutritional Foods Association. (2003) Facts about the nutrition industry. Available at
http://www.nnfa.org/facts/index.htm. Accessed 8-25-2003.
Nelson, R. C., Palsulich, B. & Gogolak, V. (2002) Good pharmacovigilance practices: technology enabled. Drug
Saf. 25: 407-414.
Ness, J., Johnson, D. & Nisly, N. (2003) "Polyherbacy": herbal supplements as a form of polypharmacy in older
adults. J. Gerontol. A Biol. Sci. Med. Sci. 58: M478.
Neuhouser, M. L. (2003) Dietary supplement use by American women: challenges in assessing patterns of use,
motives and costs. J. Nutr. 133: 1992S-1996S.
Neuhouser, M. L., Patterson, R. E., Schwartz, S. M., Hedderson, M. M., Bowen, D. J. & Standish, L. J. (2001) Use
of alternative medicine by children with cancer in Washington state. Prev. Med. 33: 347-354.
Patterson, R. E., Neuhouser, M. L., Hedderson, M. M., Schwartz, S. M., Standish, L. J., Bowen, D. J. & Marshall, L.
M. (2002) Types of alternative medicine used by patients with breast, colon, or prostate cancer: predictors, motives,
and costs. J. Altern. Complement. Med. 8: 477-485.
Rossi, A. C. & Knapp, D. E. (1984) Discovery of new adverse drug reactions. A review of the Food and Drug
Administration's spontaneous reporting system. JAMA 252: 1030-1033.
Shekelle, P., Hardy, M. L., Maglione, M. & Morton, S. C. (2003) Ephedra and Ephedrine for Weight Loss and
Athletic Performance Enhancement: Clinical Efficacy and Side Effects. Evidence Report/Technology Assessment
Number 76 (Prepared by Southern California Evidence-based Practice Center, RAND, under Contract No 290-970001, Task Order No. 9). [Final Report]. Report No. AHRQ Publication No. 03-E022. Agency for Healthcare
Research and Quality, Rockville, MD.
Siev, D. (1999) An introduction to analytical methods for the postmarketing surveillance of veterinary vaccines.
Adv. Vet. Med. 41: 749-774.
Stang, J., Story, M. T., Harnack, L. & Neumark-Sztainer, D. (2000) Relationships between vitamin and mineral
supplement use, dietary intake, and dietary adequacy among adolescents. J. Am. Diet. Assoc. 100: 905-910.
U.S. Congress. (1994) Dietary Supplement Health and Education Act of 1994. 103-417.
U.S. Congress. (2000) Federal Food Drug and Cosmetic Act of 1938 [Chp 675, 52 stat 1040, 1938. Amended in
Dec. 2000]. Available at http://www.senate.gov/~agriculture/Legislation/Agricultural%20Law/FDA_001.pdf.
Accessed 7-25-2003.
U.S. Department of Agriculture. (2003) Carbonated beverage, cola contains caffeine. Available at
http://www.nal.usda.gov.fnic/cgi-bin/list_nut.pl. Accessed 3-14-2003.
U.S. Department of Health and Human Services. (2002) MEDWATCH Form 3500. [For voluntary reporting of
adverse events and product protection]. Available at
https://www.accessdata.fda.gov/scripts/medwatch/pdf/medwatch.pdf. Accessed 7-17-2003.
U.S. Food and Drug Administration. (1991a) FDA Enforcement Report #F-717-0. Available at
http://www.fda.gov/bbs/topics/ENFORCE/ENF00031.html. Accessed 7-17-2003a.
U.S. Food and Drug Administration. (1991b) FDA Enforcement Report #F-717-0 [Correction]. Available at
http://www.fda.gov/bbs/topics/ENFORCE/ENF00032.html. Accessed 7-17-2003b.
U.S. Food and Drug Administration. (1994) Adverse events with ephedra and other botanical dietary supplements.
Available at http://vm.cfsan.fda.gov/~dms/ds-ephe2.html. Accessed 6-20-2003.
U.S. Food and Drug Administration. (6-12-1997) FDA warns consumers against dietary supplement products that
may contain digitalis mislabeled as "plantain". Available at
http://www.fda.gov/bbs/topics/NEWS/NEW00570.html. Accessed 7-30-2003.
U.S. Food and Drug Administration. (6-15-1998) Food Advisory Committee on Olestra. Volume 1 [Transcript].
Available at http://www.fda.gov/ohrms/dockets/98/transcpt/3485t1.pdf. Accessed 7-8-2003.
U.S. Food and Drug Administration. (1999a) Managing the risks from medical product use. Creating a risk
management framework. Report to the FDA Commissioner from The Task Force on Risk Management. pp. 1-164.
U.S. Food and Drug Administration. (1999b) Summary of nutrition labeling rules for dietary supplements:
Supplement Facts panel. Available at http://www.cfsan.fda.gov/~acrobat/hhssupp2.pdf. Accessed 8-8-2003b.
U.S. Food and Drug Administration. (2001) Dietary Supplements. In: U.S. FDA Performance Plan and Summary.
Available at http://www.fda.gov/ope/fy02plan/dietsupp.html. Accessed 1-31-2003.
U.S. Food and Drug Administration. (2002) FDA recall policies. Available at
http://www.cfsan.fda.gov/~lrd/recall2.html. Accessed 7-8-2003.
U.S. Food and Drug Administration (2003a) Current good manufacturing practice in manufacturing, packing, or
holding dietary ingredients and dietary supplements. [Proposed Rule]. Fed. Reg. 68: 12157-12263.
U.S. Food and Drug Administration. (2003b) Risk assessment of observational data: Good pharmacovigilance
practices and pharmacoepidemiologic assessment [Concept paper for comment purposes only]. Available at
http://www.fda.gov/cder/meeting/groupIIIfinal.pdf. Accessed 4-8-2003b.
U.S. Food and Drug Administration. (2003c) Risk management programs. [Concept paper for discussion purposes
only]. Available at http://www.fda.gov/cder/meeting/groupIIfinal.pdf. Accessed 4-8-2003c.
U.S. Food and Drug Administration. (2003d) Risk Management Public Workshop. Pharmacovigilance Practices and
Pharmacoepidemiologic Assessment. April 11, 2003. [Transcript]. Available at
http://www.fda.gov/cder/meeting/RMtranscript3.doc. Accessed 6-2-2003d.
U.S. Food and Drug Administration. (2003e) FDA announces plans to prohibit sales of dietary supplements
containing ephedra. [News Release] Available at http://www.fda.gov/oc/initiatives/ephedra/december2003/.
Accessed 12-30-03.
U.S. Food and Drug Administration. (2004a) Final rule declaring dietary supplements containing
ephedrine alkaloids adulterated because they present unreasonable risk. Fed. Reg. 69: 6788-6854.
U.S. Food and Drug Administration. (2004b) Guidance for Industry. E2B(M): Data elements for
transmission of individual case safety reports. Questions and answers. Available at
http://www.fda.gov/cder/guidance/6199fnl.pdf. Accessed 5-6-2004.
U.S. Food and Drug Administration. (2004c) Guidance for Industry. Good pharmacovigilance practices and
pharmacoepidemiologic assessment. [Draft Guidance]. Available at http://www.fda.gov/cder/guidance/5767dft.pdf.
Accessed 5-5-2004.
U.S. General Accounting Office. (2001) Health Products for Seniors. "Anti-Aging" Products Pose Potential for
Physical and Economic Harm. Report No. GAO-01-1129. U.S. General Accounting Office, Washington, DC. pp.
1-43.
U.S. General Accounting Office. (2003a) Dietary Supplements Containing Ephedra. Health Risks and FDA's
Oversight. [GAO-03-1042T]. Subcommittee on Oversight and Investigations, Committee Energy and Commerce,
House of Representatives
U.S. General Accounting Office. (2003b) Dietary Supplements. Review of Health-Related Call Records for Users of
Metabolife 356. Report No. GAO-03-494. Government Printing Office, Washington, D.C. pp. 1-30.
U.S. General Accounting Office. (2003c) Dietary Supplements. Uncertainties in Analyses Underlying FDA's
Proposed Rule on Ephedrine Alkaloids. Report No. GGD-99-90. Government Printing Office, Washington, D.C.
pp. 1-83.
Vincent, C. (2003) Understanding and responding to adverse effects. N. Engl. J. Med. 348: 1051-1056.
Walker, A. M. (2000) The relationship between voluntary notification and material risk in dietary supplement safety.
Harvard School of Public Health, Cambridge, MA.
Wallander, M. A. (1993) The way towards adverse event monitoring in clinical trials. Drug Saf. 8: 251-262.
LSRO Report: Adverse Event Monitoring Programs for Dietary Supplements—Appendices
APPENDICES: TABLE OF CONTENTS
APPENDIX A INDIVIDUALS AND ORGANIZATIONS
- 72 -
A.1 Life Sciences Research Office Board of Directors (2003)
- 72 -
A.2 Ad hoc expert committee members
- 72 -
A.3 Life Sciences Research Office staff
- 74 -
A.4 Invited presenters
- 75 -
APPENDIX B PHASE I LETTER
- 77 -
APPENDIX C DIETARY SUPPLEMENTS
- 79 -
C.1 Ingredients, classifications, and forms of supplements
C.1.1 Nutritional supplements
C.1.2 Botanical supplements
C.1.3 Other types of supplements and miscellaneous ingredients
C.1.4 Forms of supplements
- 79 - 79 - 80 - 80 - 80 -
C.2 Manufacturing and sales
- 80 -
C.3 Characteristics of dietary supplement use
- 81 -
C.3.1 Pediatric use
C.3.2 Adult use
C.3.3 Factors associated with use
C.4 Literature citations for dietary supplements
APPENDIX D KEY REGULATIONS FOR DIETARY SUPPLEMENTS AND
RELATED ACTIVITIES
- 82 - 83 - 84 - 84 - 87 -
D.1 Dietary Supplement Health and Education Act
- 87 -
D.2 Product labeling, claims, and advertisement
- 87 -
D.2.1 Supplement facts
D.2.2 Label claims
D.2.3 Product advertisement
- 88 - 90 - 90 -
D.3 Current good manufacturing practice
D.3.1 Proposed good manufacturing practice for the review of complaints
D.4 Literature citations for key regulations
APPENDIX E EXAMPLE SURVEILLANCE PROGRAMS FOR FOOD ADDITIVES
E.1 Surveillance program for aspartame
E.1.1 Recognition and reporting of adverse events in individual product users
E.1.2 Creation of individual data records
E.1.3 Building a relational database of individual data records
E.1.4 Signaling from databases of individual data records
E.1.5 Signaling from other information sources
E.1.6 Issues resolution and formulation of risk management interventions
- 91 - 91 - 92 - 93 - 93 - 93 - 94 - 95 - 95 - 95 - 95 - 69 -
E.2 Surveillance program for Olestra
E.3 Literature citations for example surveillance programs for food additives
APPENDIX F EXISTING SURVEILLANCE PROGRAMS FOR DIETARY SUPPLEMENTS
F.1 Surveillance programs at the Food and Drug Administration
F.1.1 Development of the federal surveillance program
F.1.2 Recognition and reporting of adverse events in individual product users
F.1.3 Creation of individual data records
F.1.4 Signaling from the MedWatch database
F.1.5 Issues resolution and formulation of risk management interventions
F.2 American Association of Poison Control Centers Toxic Exposure Surveillance System
F.2.3 Dietary supplement exposures reported to the Toxic Exposure Surveillance System
F.2.4 Signaling from the poison control database
F.3 Postmarketing surveillance by the World Health Organization
F.3.1 Definitions
F.3.4 The Collaborating Center for International Drug Monitoring database of
individual data records
F.3.5 Signaling from the Collaborating Center for International Drug Monitoring database
F.3.6 Signaling from other information sources
F.4 Literature citations for existing surveillance programs for dietary supplements
APPENDIX G ANCILLARY ISSUES
G.1 Resources
G.1.1 Resources for recognition and reporting of adverse events in individual product users
G.1.2 Resources for creation of individual data records
G.1.3 Resources for building a relational database of individual data records and
support for signaling
G.1.4 Resources for signaling from other information sources
G.2 Future research
- 95 - 95 - 97 - 97 - 97 - 97 - 98 - 98 - 100 - 100 - 100 - 101 - 102 - 102 - 104 - 105 - 106 - 106 - 107 - 108 - 108 - 109 - 109 - 109 - 109 - 111 - 112 - 117 - 117 - 117 - 122 - 122 - 122 - 122 - 125 - 125 - 128 -
G.2.1 Research to improve recognition and reporting of adverse events in individual
product users
- 128 G.2.2 Research for creation of individual data records
- 128 G.2.3 Research for building a relational database and to improve signaling from databases
of individual data records
- 128 -
G.3 Summary of resources and future research
G.4 Literature citations for ancillary issues
APPENDIX H ACRONYMS AND GLOSSARY
- 129 - 129 - 133 -
H.1 Table of acronyms
- 133 -
H.2 Glossary table
- 133 -
H.3 Literature citations for acronyms and glossary
- 136 -
APPENDIX A
INDIVIDUALS AND ORGANIZATIONS
A.1 LIFE SCIENCES RESEARCH OFFICE BOARD OF DIRECTORS (2003)
Chair:
William Heird, M.D.
Baylor College of Medicine
Houston, TX
Vice Chair:
Salvatore J. Enna, Ph.D.
University of Kansas Medical Center
Kansas City, KS
Treasurer:
Arnold Kahn, Ph.D.
University of California-San Francisco
San Francisco, CA
Past Chair:
Donald Beitz, Ph.D.
Iowa State University
Ames, IA
Directors:
Gilbert Leveille, Ph.D.
Cargill Health & Food Technologies
Wayzata, MN
Robert Newburgh, Ph.D.
The Protein Society
Bethesda, MD
Terry Quill, M.S., J.D.
International Society for Regulatory Toxicology and Pharmacology
Washington, DC
Robert Russell, M.D.
United States Department of Agriculture
Tufts University
Boston, MA
Jacob J. Steinberg, Ph.D.
Albert Einstein College of Medicine
Montefiore Medical Center
Bronx, NY
Taffy J. Williams, Ph.D.
Photogen Technologies, Inc.
New Hope, PA
Secretary:
Michael C. Falk
Bethesda, MD
A.2 AD HOC EXPERT COMMITTEE MEMBERS
John A. Clark, M.D., MSPH, holds a B.S. in pharmacy from the University of Wyoming, an M.D. degree from
University of Colorado, and a M.S. in public health from the University of Utah. Currently, Dr. Clark leads the
Surveillance Services Division of Galt Associates, Inc., Sterling, VA, which offers outsourced pharmacovigilance
services, including signal detection and evaluation services, Periodic Safety Update Report services, and adverse
event coding services.13 Dr. Clark has more than 15 years of experience in the fields of pharmaceutical product
safety and development. In 1987, he joined the U.S. Food and Drug Administration as a medical epidemiologist and
safety analyst. Since 1990 he has worked for several pharmaceutical companies on projects pertaining to product
safety, marketing, drug development, and managed care. In addition to several chapters in books, Dr. Clark has
written articles, abstracts, letters, and editorials for medical and industry journals, including the Annals of Internal
Medicine, American Journal of Medicine, New England Journal of Medicine, and Epidemiologic Reviews.
Barbara Insley Crouch, Pharm.D., MSPH, received her B.S. in pharmacy magna cum laude from Philadelphia
College of Pharmacy and Science and received her Doctor of Pharmacy degree jointly administered by the
University of Texas at Austin and the University of Texas Health Science Center in San Antonio. She completed a
Clinical Toxicology Fellowship at the Maryland Poison Center, University of Maryland School of Pharmacy, and
then went on to pursue a M.S. in public health from the University of Utah. Since 1992, Dr. Crouch has served as
the Director of the Utah Poison Control Center in Salt Lake City, UT. In addition, since 1996 she has had an
academic appointment as associate professor (clinical) for the Department of Pharmacy Practice at the University of
Utah College of Pharmacy. Dr. Crouch has published book chapters and reports in peer-reviewed journals. She is
affiliated with the Utah Society of Health Systems Pharmacists, the American Board of Applied Toxicology, the
American Academy of Clinical Toxicology, the American Association of Poison Control Centers, and the American
Society of Health Systems Pharmacists.
Janet Greger, Ph.D., earned her B.S. at the University of Illinois, and her M.S. and Ph.D., both in human nutrition,
from Cornell University. She was a faculty member at Purdue University, then moved to the University of
Wisconsin in Madison where she was the Associate Dean of the graduate school from 1990-1996, Associate Dean of
the medical school from 1996-1998, and Special Assistant to the Provost from 1998-2000. Since 1992, Dr. Greger
has been on the Board of Directors of the Association for Assessment and Accreditation of Laboratory Animal Care.
In 1999-2002, Dr. Greger was a member of a National Institutes of Health panel on regulatory burden, which
addressed issues such as animal care, the use of human subjects for research, and research integrity. She chaired its
animal care and use committee in 2001. In June 2002, Dr. Greger moved to the University of Connecticut where she
became the Vice Provost for Research and Graduate Education and Dean of the Graduate School. She has also
served on several committees with the National Academy of Sciences and is currently a member of a committee
examining mineral toxicities. As an expert in the metabolism of minerals (aluminum, manganese and calcium), she
has received funding from the National Institutes of Health, U.S. Department of Agriculture, and a variety of
industry and foundation sources. She has published more than 150 scientific papers, and her work has been the basis
of several toxicological and nutritional standards by the U.S. Environmental Protection Agency. More than 40
students have completed graduate degrees in her laboratory.
Robert Hoerr, M.D., Ph.D., received both his B.A and his M.D. from Indiana University. He received his Ph.D. in
nutritional biochemistry and metabolism from Massachusetts Institute of Technology in 1987. After serving for
three years as Director of Medical Affairs for Sandoz Nutrition Corporation (now Novartis Nutrition), Dr. Hoerr
joined GalaGen Inc., in Minnetonka, MN, in 1993 and gained increasing responsibility within the company, last
serving as Chairman and Chief Technology Officer in the period of January 2000 to February 2002. At present he is
Chairman, President and Chief Executive Officer for Nanocopoeia, Inc., in St. Paul, MN, a start-up drug formulation
and delivery company, which uses nanoparticle technology invented at the University of Minnesota. He also serves
as President of Nutramentals, Inc., a company that provides scientific and regulatory advisory services to companies
that develop products for the dietary supplement and nutrition markets.14
Valen Johnson, Ph.D., received his B.S. in mathematics from Rensselaer Polytechnic Institute in Troy, NY, his
M.S. in applied mathematics from the University of Texas at Austin and his Ph.D. in statistics from the University of
Chicago. Dr. Johnson was a professor of statistics and decision sciences at Duke University. Currently, he is a
professor in the Department of Biostatistics at the University of Michigan School of Public Health in Ann Arbor.
He is a fellow of the American Statistical Association, past Treasurer of the International Society of Bayesian
Analysis, has served as an Associate Editor for the Journal of the American Statistical Association and Institute of
Electrical and Electronics Engineers Transactions for Medical Imaging. His research interests include ordinal and
rank data modeling, Bayesian image analysis, Bayesian reliability modeling, convergence diagnostics for Markov
chain Monte Carlo algorithms, Bayesian goodness-of-fit diagnostics, and educational assessment. Dr. Johnson won
a Leonard J. Savage Award from the American Statistical Association Section on Bayesian Statistical Science.
13
14
The Life Sciences Research Office makes no recommendation or endorsement of any commercial product.
During the course of this study, Dr. Hoerr met with counsel at Patton Boggs on an unrelated business matter.
D. Warner North, Ph.D., received his B.S. in physics from Yale University and his Ph.D. in operations research
from Stanford University. Dr. North is president and principal scientist of the consulting firm, NorthWorks, Inc., in
Belmont, CA, and consulting professor in the Department of Management Science and Engineering at Stanford
University, CA. Over the past 30 years Dr. North has carried out applications of decision analysis and risk analysis
for electric utilities in the U.S and Mexico for the petroleum and chemical industries and for government agencies
with responsibility for energy and environmental protection. He has served as a member and consultant to the
Science Advisory Board of the U.S. Environmental Protection Agency since 1978 and as a presidentially appointed
member of the U.S. Nuclear Waste Technical Review Board (1989-1994). Dr. North is a co-author of many reports
dealing with environmental risk for the National Research Council of the National Academy of Sciences, including
“Risk Assessment in the Federal Government: Managing the Process” (1983), “Improving Risk Communication”
(1989), “Science and Judgment in Risk Assessment” (1994), and “Understanding Risk: Informing Decisions in a
Democratic Society” (1996). Among other distinguished achievements, Dr. North was a recipient of the Frank P.
Ramsey Medal from the Decision Analysis Society in 1997 for lifetime contributions to the field of decision
analysis, and was the 1999 recipient of the Outstanding Risk Practitioner Award from the Society for Risk Analysis.
Catherine St. Hilaire, Ph.D., holds a B.S. with high honors in biology from West Virginia University (1971) and a
Ph.D. in microbiology from the Pennsylvania State University College of Medicine (1977). Dr. St. Hilaire is Vice
President of Sciences International, Inc., in Alexandria, VA. Prior to joining Sciences International, Dr. St. Hilaire
was Director of Regulatory Affairs for Hershey Foods Corporation in Hershey, PA. She is an internationally
recognized expert with over 20 years of experience in human health risk assessment and risk management. She was
instrumental in the development of the standard four-step risk assessment paradigm. This concept along with the
delineation of a “policy” component, embedded within the risk assessment process (first articulated in the 1983
report, “Risk Assessment in the Federal Government: Managing the Process”), contributed to the widespread
acceptance of public health risk assessment methodologies. She was a consultant to the U.S. Environmental
Protection Agency (EPA)’s Carcinogen Assessment Group during the development of the 1986 Risk Assessment
Guidelines for Carcinogens and was part of the team that developed the first set of site assessment procedures for
EPA Superfund sites. She served as Director of the International Life Sciences Institute (ILSI) Risk Science
Institute, which identified gaps in the analytic basis for risk assessment and sponsored research to develop critical
data needed to improve the accuracy and reliability of these analyses. She co-chaired a series of Risk Science
Symposia held at the Brookings Institution in Washington, D.C. In 1989, she was honored as a Fellow of the Society
for Risk Analysis. She has served on advisory and peer review committees to various organizations including the
National Academy of Sciences, the Life Sciences Research Office, and the Pennsylvania Department of
Environmental Protection. She lectures on risk science policy issues, is the author of numerous reports evaluating
the toxicity and potential health risks of a wide variety of environmental agents, and has taught university-based
courses on risk assessment and product risk management.
A.3 LIFE SCIENCES RESEARCH OFFICE STAFF
Pippa Antonio, B.S., graduated from the University of Stirling, Scotland with distinction. She was previously
employed as a research writer for The Myelin Project in Fairfax, VA. Pippa was an Associate Staff Scientist for the
Life Sciences Research Office in 2003.
Michael Falk, Ph.D., is the Director of the Life Sciences Research Office. He received his Ph.D. in biochemistry
from Cornell University and completed postdoctoral training at Harvard Medical School. He was employed in
various capacities at the Naval Medical Research Institute (Bethesda, MD). Serving as a principal investigator, he
was a key member of the Scientific Advisory Board and Acting Director for the institute. Dr. Falk also served as
Director of the Wound Repair Program and Director of Biochemistry and Cell Biology. He has managed peer
review and subject review panels in infectious diseases, environmental sciences, military medicine, and other healthrelated fields for the National Science Foundation, Medical Research Council of Canada, and Office of Naval
Research. As the Director of the Life Sciences Research Office, Dr. Falk evaluates biomedical information and
scientific opinion for regulatory and policy makers in both the public and private sectors. He has produced seminal
white papers on infant nutrition, food labeling, food safety, and military dental research, and has organized two
international conferences.
Robin S. Feldman, B.S., M.B.A., is the Literature Specialist in the Life Sciences Research Office. She is a
seasoned information specialist with experience in the electronic acquisition, analysis, and management of scientific,
business, and regulatory information. Ms. Feldman obtained her B.S. from the George Washington University in
Washington, D.C., with a major in zoology and an M.B.A. from the University of Maryland at College Park with a
concentration in science and technology. Previously, she worked as a Biomedical Research Assistant at Consultants
in Toxicology, Risk Assessment and Product Safety, where she obtained and researched scientific literature for
private and governmental clients. At the National Alliance for the Mentally Ill, she designed and implemented a
document management and retrieval system for the Biological Psychiatry Branch of the National Institute of Mental
Health and served as Managing Editor of Bipolar Network News, a newsletter for the Stanley Foundation Bipolar
Network. At Howard Hughes Medical Institute, she oversaw the implementation of the Predoctoral Fellowship in
Biological Sciences program. While serving as Science Information Specialist at the Distilled Spirits Council of the
United States, she managed the installation of a local area network and participated in the development and
maintenance of an electronic research database. As a Report Coordinator at Microbiological Associates, Inc. she
conducted statistical analyses and prepared technical reports about toxicology studies using animal models. She also
served as Data Management Administrator for the National Toxicology Program’s sponsored studies.
Karin French, B.S. received B.S. degrees in animal science and in cell and molecular biology and genetics from the
University of Maryland at College Park. Karin worked in the Dairy Nutrition Laboratory at the university, helping
Maryland dairy farmers use milk urea nitrogen (MUN) to evaluate herd protein nutrition. She helped design and
complete studies to compare and evaluate the MUN analysis techniques used in the National Dairy Herd
Improvement Association laboratories. She joined the Life Sciences Research Office in November 2003 as an
Associate Staff Scientist.
Catherine J. Klein, Ph.D., R.D., C.N.S.D., is a Staff Scientist at the Life Sciences Research Office. She graduated
magna cum laude from the Department of Human Nutrition and Food Science at the University of Maryland at
College Park, where she also obtained her M.S. and Ph.D. in nutrition. She has completed internships in the PreProfessional Practice Program in Dietetics at the University of Maryland Medical System (UMMS) in Baltimore,
MD, the V.A. Kleinfeld Summer Internship Program at the Food and Drug Law Institute in Washington, D.C., and
most recently, Dannon’s Nutrition Leadership Institute in Wye River, MD. At UMMS, she developed system-wide
guidelines for nutrition assessment, documentation, and continuity of care. As Clinical Coordinator of Research in
the Division of Critical Care Medicine at the University of Maryland R Adams Cowley Shock Trauma Center in
Baltimore, MD, she developed, initiated, and administered research projects focused on nutrition issues in critical
care. She established a multidisciplinary nutrition task force, which resulted in improvements in clinical practice
standards. She is the primary author on nine peer-reviewed publications, including a book chapter, and has lectured
or presented at over 25 professional meetings. Dr. Klein received the Pelczar Award for Excellence in Graduate
Study from the University of Maryland Graduate School and Sigma Xi, and the Dr. E.V. McCollum Award from the
Maryland Dietetic Association. Her professional contributions include serving on the Advisory Board of the
University of Maryland Dietetics Program and as editor for the Maryland Dietetic Association.
A.4 INVITED PRESENTERS
Pamela Breimann
Senior Account Executive
Galt Associates, Inc.
Sterling, VA
Tim Dring
Associate Director
Pharmaceutical Research & Development
Johnson & Johnson
Titusville, NJ
Steven J. Dentali, Ph.D.
Vice President
Scientific and Technical Affairs
American Herbal Products Association
Silver Spring, MD
Kenneth J. Falci, Ph.D.
Director
Office of Scientific Analysis and Support
Center for Food Safety and Applied Nutrition,
U.S. Food and Drug Administration
College Park, MD
Richard L. Kingston, Pharm.D.
Vice President, PROSAR International Poison Control Center;
Associate Professor, College of Pharmacy, University of Minnesota
Minneapolis, MN
Daniel A. Kracov
Partner, Director of Administrative and Regulatory Law Practice
Patton Boggs LLP
Washington, DC
Renee Lewis
Chief Operating Officer
QED Solutions Inc.
McLean, VA
APPENDIX B
PHASE I LETTER
APPENDIX C
DIETARY SUPPLEMENTS
This appendix characterizes the types and sales of dietary supplements and provides a general overview of the
population that uses supplements.
C.1 INGREDIENTS, CLASSIFICATIONS AND FORMS OF SUPPLEMENTS
Some dietary supplements contain a single ingredient, such as vitamin E, whereas others, such as sports drink
powders, may contain complex mixtures of multiple ingredients. Based on a sample of nearly 3,000 dietary
supplement labels, it was determined that vitamin and mineral supplement products list approximately 13
ingredients per supplement, and other types of dietary supplements (i.e., botanical products) list approximately four
ingredients per supplement (U.S. Food and Drug Administration, 2003).
C.1.1 Nutritional supplements
Dietary supplementation has a long history of use in medicine for providing protein, calories, vitamins, and minerals
to the ill and malnourished to prevent or correct nutritional deficiencies. For example, nutritional supplements have
been fed to infants who were failing to thrive and to those with gastrointestinal disease who absorbed insufficient
amounts of nutrients. Nutritional supplements include vitamins, minerals, protein, carbohydrates, and fats. The
Institute of Medicine of the National Academies (1998) proposed values for tolerable upper intake levels of vitamins
and minerals because excessive intakes of these nutrients can produce adverse effects.
Vitamins are organic substances that are necessary for metabolic processes and are needed in the diet. A trace
amount of vitamins must be consumed on a regular basis. Otherwise, vitamin deficiency will cause disease with
particular defining characteristics. Sufficient amounts of water-soluble vitamins to meet the body’s requirements
must be consumed relatively frequently as part of the overall diet. In contrast, lipid-soluble vitamins can be
consumed less frequently if the body has accumulated reservoirs of these vitamins in fatty tissue and the liver.
Vitamin supplements tend to be susceptible to destruction by heat and light.
Numerous types of minerals are essential to support life. Some minerals, such as calcium, are needed in larger
quantities in the diet than other minerals, such as copper, which are only needed in very small amounts and are
referred to as “trace minerals” (Mackenzie, 2003). The minerals in supplements are resistant to destruction from
heat and light.
Protein, carbohydrate, and fat are termed macronutrients because the food in our diet consists mainly of these three
types of nutrients and they are needed in relatively large quantities to support life. Various forms of protein are used
in dietary supplements, such as protein hydrolysate and amino acids. Amino acids are the basic structural units of
protein. Essential amino acids must be obtained from the diet. In contrast, nonessential amino acids can be
synthesized in the body from intermediates of major metabolic pathways, e.g., other amino acids (Texas A&M
University, 2003). The body uses the amino acids absorbed from the diet to construct the types of functioning
protein that it needs, e.g., enzymes and structural protein (e.g., muscle).
Most diets are rich in carbohydrates, which provide the body with simple sugars, such as glucose that is needed for
cellular fuel. Carbohydrates are typically added to supplements as a sweetener and as a source of calories. Some
carbohydrate is in the form of insoluble dietary fiber, which is resistant to digestion in the upper gastrointestinal tract
and can help to maintain bowel regularity.
Fats include many different types of fatty acids. Some fatty acids are essential and must be consumed in the diet,
such as α-linoleic acid. Fatty acids in supplements come from both plant and animal sources (e.g., fish oil) and are
particularly susceptible to destruction from heat and light.
Many multivitamin and mineral preparations (e.g., Centrum® Performance™) also contain botanicals such as
ginseng root and gingko biloba leaf.
C.1.2 Botanical supplements
Botanical dietary supplements are derived from plants. Although they may contain some inherent nutrients, they are
not intended as a source of nutrients unless they also contain added nutrients. Herbal products are a subclass of
botanicals that are made from only the leaves and stems of plants (U.S. Food and Drug Administration, 1999).
Historically, botanical products were used in their raw or dried form, often consumed as tea (Glisson et al., 2003).
More potent forms are now available as extracts, powders, liquids, capsules, or tablets.
The composition of botanical products can vary due to differences in where the plant was grown, the climate, and
the method of harvesting. Many plants have not yet been fully classified and evaluated with regard to their chemical
constituents. Similar to other food products, supplements of plant origin are subject to contamination and
deterioration. Crude botanical ingredients are inherently unstable and may lose their quality unless storage and
processing controls for temperature, humidity, and light are implemented (U.S. Food and Drug Administration,
2003).
Some ethnic communities have a longstanding tradition of herbal medicine. Community members seek and support
the continued use of products such as use of Chinese or Indian herbal remedies for self-treatment of disease
(Tomassoni & Simone, 2001). As is true with other types of supplements, some botanical products are used for “off
label” purposes. Such is the case of individuals who procure and use psychoactive plant substances as a means to a
“natural high” (Tomassoni & Simone, 2001).
C.1.3 Other types of supplements and miscellaneous ingredients
Additional classifications of supplements include probiotics, prebiotics, animal derivatives, and other types of
supplements. Probiotic dietary supplements purposefully contain bacterial cultures (e.g., Lactobacillus and
Bifidobacterium). These microorganisms are consumed with the intent that they will populate the intestinal tract and
lead to beneficial effects. Prebiotic dietary supplements contain substrates that will ferment in the colon, such as
indigestible carbohydrates, and are used to promote the colonization and growth of probiotic organisms. Animal
tissue extracts and endogenous derivatives are another type of supplement. They include substances such as
chondroitin sulfate, shark cartilage, glandular (e.g., pituitary, adrenal) extracts, creatine, carnitine, glutathione, and
various enzymes. Other types of supplements sold include algae (spirulina, blue-green manna), bee pollen, royal
bee jelly, bioflavinoids, and Brewer’s yeast (U.S. Food and Drug Administration, 1999).
Dietary supplements also contain a number of substances that are not printed on the ingredient label. For instance,
processing and preparation of ingredients can create chemical changes and introduce chemical compounds into the
product that are not found in the raw source of the ingredients (Rodricks et al., 1995). Moreover, many supplements
are manufactured using solvents, binders, and lubricants that may be introduced into the final product. The number
of unlisted components is estimated to range from none to six, with four unlisted components being typical (U.S.
Food and Drug Administration, 2003). Furthermore, contamination of the product can cause unwanted and
undisclosed additional substances in the dietary supplement. Contamination may come from insects on the plant,
contact with surfaces during transport, processing, and packaging, and by potentially harmful bacterial and fungal
growth. Contamination from chemicals, microbes, aflatoxin or other natural toxins, pests, dirt, filth or other
extraneous material may be a source of product adulteration. A food is adulterated and therefore illegal, if it bears
or contains an added substance (intentional or unintentional) that renders the food harmful to health [Section 402 (a)
(1) FFDCA] (U.S. Congress, 1994). For example, the dietary supplement PC-SPES was considered adulterated after
tested lots were found to contain synthetic estrogen (diethylstilbestrol) and warfarin (White, 2002).
C.1.4 Forms of supplements
The forms of dietary supplements in the market vary widely, ranging from dried whole pieces of plant to highly
processed pills and drinks. Popular forms include tablets with various coatings (i.e., delayed-release) and capsules
of hard or soft gelatin. Supplements may also be in the form of lozenges, liquids, powders, teas, or chewable bars.
C.2 MANUFACTURING AND SALES
The U.S. Food and Drug Administration (FDA) estimated that there may be 1500 or more manufacturing and
repacking facilities for dietary supplements in the United States. Of these facilities, FDA inspects less than 5%
annually (U.S. General Accounting Office, 2001).
Most producers of dietary supplements in the United States have less than $20 million in sales per year, and are
therefore considered small manufacturers by FDA. In contrast, by 1997 there were 11 large companies that were
each selling in excess of $100 million in total annual revenues, accounting for 53% of all supplement sales (U.S.
A single lot of raw ingredient can be distributed into one to 12 batches of product, with an average of 6.5 batches
produced per lot. Very small manufacturers produce an average of 223 total batches of dietary supplements per
year. It is estimated that large companies average 300 to 550 batches of product per year (U.S. Food and Drug
Administration, 2003), but production may be as high as 3000 batches per year (Wallace, 2003).
FDA estimates that there are approximately 29,000 unique formulations packed into more than 75,000 distinctly
labeled dietary supplements in the U.S. market (U.S. Food and Drug Administration, 2001).
As of 1995, some 1500 to 1800 botanical products were sold in the U.S. market for use as dietary supplements
and/or traditional medicines (Commission on Dietary Supplement Labels, 1997). The five top selling single
botanicals from 1997 to 2000 in the United States were gingko biloba, echinacea, garlic, ginseng, and St. John’s
wort (IARC Working Group, 2002). In the 2000 mass market, sales of soy supplements and green tea grew by 90
percent and 42 percent, respectively (Dietary Supplement Information Bureau, 2003b). Products containing a
mixture of botanicals far exceeded the sales of single-item botanical products (IARC Working Group, 2002). In
2000, U.S. sales for herbs and botanicals exceeded $4.1 billion (Anonymous, 2003; Dietary Supplement Information
Bureau, 2003b).
Within the vitamin and mineral category, the six products having the greatest retail sales are multivitamins with or
without minerals, vitamin E, vitamin C, iron, calcium, and B vitamins (Commission on Dietary Supplement Labels,
1997). In 2000, U.S. sales for vitamins and minerals were approximately $6 billion and $1.39 billion, respectively
(Anonymous, 2003; Dietary Supplement Information Bureau, 2003b).
Dietary supplements are readily available via the Internet, television infomercials, mail order, multilevel marketing
firms, supermarkets, telephone orders and other retail outlets. As of 1995, the worldwide sales of dietary
supplements were in excess of $28 billion per year (Commission on Dietary Supplement Labels, 1997). U.S. sales
of all dietary supplements were estimated to be approximately $17 billion each year for 2000 and 2001
(Anonymous, 2003; Dietary Supplement Information Bureau, 2003b; U.S. Food and Drug Administration, 2001)
and reached $18 billion in 2002 (National Nutritional Foods Association, 2003). In particular, sales of “sports”
supplements more than doubled in 2000 compared with 1995 (Anonymous, 2003; Commission on Dietary
Supplement Labels, 1997). FDA expects the dietary supplement industry to grow at a rate of 12% to 14% per year
(U.S. Food and Drug Administration, 2001).
C.3 CHARACTERISTICS OF DIETARY SUPPLEMENT USE
Dietary supplements are now used by a broad spectrum of the U.S. population. Approximately 40% of 33,905
participants in a nationwide survey reported taking dietary supplements (Ervin et al., 1999). Supplements are used
for myriad reasons, although many reasons have not been substantiated by scientific studies (Commission on
Dietary Supplement Labels, 1997; Conner et al., 2003; Dietary Supplement Information Bureau, 2003a; Lanski et
al., 2003; Office of Dietary Supplements, 2003; Palmer et al., 2003):
• Enhance athletic, mental, and/or sexual performance
• Alter body weight, shape, and/or composition
• Improve or maintain nutritional and health status
• Increase energy and alertness
• Facilitate the digestive process or promote bowel regularity
• Prevent chronic disease or to self-medicate to treat disease
• “Feel better,” for “well-being,” and to “improve overall quality of life”
• Have more control over personal health matters
• Delay aging
• Aid sleep
Uncommonly reported intents have included abortifacient, treatment of menstrual symptoms, foiling of drug tests,
smoking cessation, recreational purposes to generate a “natural high,” and as a means of suicide (Palmer et al.,
2003).
C.3.1 Pediatric use
C.3.1.1 Prevalence of supplement use by children
In the Third National Health and Nutrition Examination Survey (NHANES III) conducted from 1988 to 1994, the
prevalence of dietary supplement use among children varied by age group (Ervin et al., 1999):
• ~25% for children younger than one year of age (n=2094)
• ~40% for children 1-2 years of age (n=2683)
For children two months to 11 years of age who consumed supplements (n=3857), 90% were only using one type of
supplement, which was a multivitamin with or without minerals (89.4%), a single vitamin or mineral (7.5%), or
something not specified (0.5%). Similar findings were reported for children (n=8285) approximately 3 years of age
who were included in the 1991 longitudinal follow up to the 1988 Maternal and Infant Health Survey (Yu et al.,
1997). For these children, 54.4% had consumed dietary supplements within one month of the interview. The most
popular supplement type was a multivitamin-mineral combination used by 85.4% of those who had consumed
supplements.
Of 423 adolescents, 13 to 18 years of age, interviewed during the 1994 Continuing Survey of Food Intakes of
Individuals, 34% used dietary supplements (Stang et al., 2000). Sixteen percent of adolescents used supplements on
a daily basis. Multivitamins without minerals were used by 48% of those reporting supplement use. Adolescents
who consumed supplements on a daily basis had diets that were significantly denser in several micronutrients
(supplements not included), indicating they had healthier diets than adolescents who did not use supplements (Stang
et al., 2000).
C.3.1.2 Botanical supplement use by children
Lanski et al. (2003) interviewed caregivers of 142 children aged 18 years or younger who entered an emergency
department in Atlanta, GA over a three-month period in the Fall/Winter of 2001. Forty-five percent of caregivers
reported giving their child an herbal or home remedy during the past year (route of administration, i.e. oral or
topical, not specified). Approximately one-half (53%) of children receiving herbal products received only one type
of product. Seventeen children received three or more different products. The greatest number of different herbal
products given to a single child in the last year was 11.
The most frequently reported products included aloe plant or juice (44%), and echinacea (33%) (Lanski et al., 2003).
The majority (61%) of children given herbal products were also administered a prescription medication at the same
time. The authors thought the most concerning herbal-drug combination reported was that of ephedra and albuterol
given to an adolescent with asthma.
Caregivers who administered herbal products to children were often recommended to do so by family or friends
(45%). They believed the products were safer than traditional Western medicine (30%), and in a few cases, hoped it
would negate the need for them to take the child to see a doctor (8%)( Lanski et al., 2003). Of 142 caregivers
interviewed in 2001, 77% were uncertain whether herbal products had any side effects in children (Lanski et al.,
2003). Furthermore, 94 (66%) were unsure or thought that herbal products would not interact with over-the-counter
or prescription medication.
C.3.1.3 Supplement use by sick children
A substantial portion of chronically ill children use dietary supplements. Neuhouser et al. (2001) conducted
telephone interviews with parents of 75 pediatric oncology patients identified in the Cancer Surveillance System of
western Washington State. Parents were asked if their child had used any vitamin, mineral, herbal, or other
supplements (e.g., shark cartilage) or preparations (e.g., homeopathic remedies) in the preceding 12 months. The
occurrence and severity of adverse events were also obtained. Because standard medical care for these patients
routinely includes multivitamin use, where serving sizes are consistent with recommended daily intakes,
multivitamin use was excluded from analysis.
Dietary supplements were used as adjuncts, not substitutes, for conventional medical therapy, and each parent had
several reasons for giving their child a supplement:
• 34 (45%) parents were motivated to treat side effects or symptoms of cancer or cancer treatment
• 24 (32%) parents were motivated to prevent recurrence or spread of cancer
• 37 (49%) parents were motivated to treat cold, flu, or other non-cancer illness
• 52 (69%) parents were motivated to support the general good health of their child
Single-vitamin supplements were used within 12 months of the interview by 22 of the 75 children, 59% of whom
began product use after the cancer diagnosis (Neuhouser et al., 2001). Similarly, 26 of the pediatric patients had
used an herbal supplement, 54% of whom began after the cancer diagnosis. The average estimated expenditure per
year was $79 for single-vitamins and $214 for herbal products, with an overall range of $2 to $2,737 per child
(Neuhouser et al., 2001).
Neuhouser et al. (2001) expressed concern that some children receiving chemotherapy may have received
antioxidant supplements (e.g., vitamin C), which might have reduced the effectiveness of their treatment. They were
also concerned that children were given herbal products containing yew needle and Essiac® (Essiac tea is made
from burdock root, Indian rhubarb root, sheep sorrel, and slippery elm bark). According to Neuhouser et al. (2001),
these ingredients have been associated with adverse events similar to those observed after use of cisplatin and
anthracyclines, drugs routinely used for cancer patients. Two children experienced “very severe” adverse events
(e.g., vomiting, skin irritation or sleep disturbance) following use of an herbal preparation.
C.3.2 Adult use
C.3.2.1 Prevalence of supplement use by adults
By some estimates, one-third of adults take supplements every day or nearly every day (Commission on Dietary
Supplement Labels, 1997). Use of dietary supplements by adults surveyed in NHANES III (n=18,792) ranged from
approximately 30% for men 20-29 years of age to 55% for women 80 or more years of age (Ervin et al., 1999).
Adults were more likely to use a multivitamin with or without minerals rather than other types of supplements
(Ervin et al., 1999). Foote et al. (2003) reported that nearly 50% of healthy adults using multivitamins had used the
multivitamin regularly for five or more years. Ervin et al. (1999) reported that one-third or more of adults used two
or more different supplements. Among women, 76% of those who were pregnant or lactating reported taking a
dietary supplement, typically a prenatal vitamin and/or iron supplement (Ervin et al., 1999). Fewer than 2% of
adults surveyed in NHANES III reported taking an herbal product (Radimer et al., 2000).
From 2000 through 2002, the Vitamin and Lifestyle Study was conducted on a cohort of 75,000 men and women,
aged 50 to 75 years, in western Washington State (Satia-Abouta et al., 2003). The 24-page survey questionnaire
included questions on health complaints, medical history, and type, dose, and frequency of use of supplements over
the past ten years. Satia-Abouta et al. (2003) reported the results of 45,748 participants (49% male, 92% Caucasian,
23% obese) who completed surveys by December 2001. Among these, more than 75% reported regular use of
dietary supplements. Multivitamins were used by 57% of those surveyed. The most commonly used single
supplements were vitamin E, vitamin C, calcium, folate, and selenium.
C.3.2.2 Supplement use by chronically ill adults
Patterson et al. (2002) surveyed 356 patients, aged 20-79 years, diagnosed with colorectal, breast, or prostate cancer
that were identified from the Cancer Surveillance System of western Washington State. Of these patients, 38.2%
reported using an herbal product in the past year and 64.6% reported taking a dietary supplement that was not a onea-day multivitamin type preparation (one-a-day multivitamin type preparations were excluded from analysis).
Patients with breast cancer were significantly more likely to use dietary supplements than other patients surveyed.
More than 45% of patients were motivated to use supplements to treat cancer. Dietary supplements were also used
to treat other illness and to promote general health and well-being. The average estimated expenditure per year was
$36 (n=176) for vitamin and/or mineral supplements that were not a one-a-day type of supplement and $46 (n=105)
for herbal products, with an overall range of $4 to $2,861 spent per patient.
Newman et al. (1998) reported that 352 of 435 women (81%) who had early-stage invasive breast cancer, and were
participating in the Women's Healthy Eating and Living study, were using dietary supplements at the time of the
baseline interview in 1996. The most commonly reported supplements, in order of prevalence, were vitamin E,
vitamin C compounds, multivitamin with mineral, and single mineral supplements. Participants who were within 12
months of diagnosis were more likely to use herbal products than those who had received their diagnosis 24 to 48
months previously. Participants with a stage IIIA cancer diagnosis were four times more likely than those with a
stage I diagnosis to use miscellaneous products, when adjusted for time since diagnosis. Miscellaneous products
were those not included in nutrient or herbal categories, such as Lactobacillus acidophilus, coenzyme Q-10, lecithin,
melatonin, and shark cartilage.
Schaffer et al. (2003) analyzed general health surveys returned by 15,985 adult members of a health maintenance
organization in northern California. Of 1,845 adults reporting arthritis, 18.8% were using glucosamine dietary
supplements. Of 2,117 adults reporting depression for two or more weeks in the past year, 33.4% were using St.
John’s wort dietary supplements and 15.6% of these supplement users were also using a prescription antidepressant.
C.3.3 Factors associated with use
Use of dietary supplements tends to increase with age, years of education, and income (Ervin et al., 1999; Foote et
al., 2003). Use of dietary supplements also varies by gender; women are more likely to consume supplements than
men (Ervin et al., 1999). The gender difference is particularly distinct for calcium supplements, which are used by
three times as many women as men regardless of ethnicity (Foote et al., 2003).
Overall, use of dietary supplements has tended to vary with ethnicity (Ervin et al., 1999; Foote et al., 2003). NonHispanic whites are more likely to consume supplements than non-Hispanic blacks, Mexican-Americans/Latinos or
Hawaiian-Americans (Ervin et al., 1999; Foote et al., 2003; Jasti et al., 2003). However, use of vitamin A or iron
supplements by Latinos and African-Americans was approximately twice that of other ethnic groups, regardless of
gender (Foote et al., 2003). More Japanese-Americans and whites report long-term (five or more years) use of
supplements than do other ethnic groups (Foote et al., 2003).
A subgroup of adult women (n=2868) in the 1994-1996 Continuing Survey of Food Intake by Individuals completed
the Diet and Health Knowledge Survey questionnaire. The majority (79.8%) of supplement users were not on a diet
of any kind, but the portion reporting they were on a diet (20.2%) was significantly higher than the portion of
nonusers (13.8%) who reported they were on a diet (Jasti et al., 2003). A significantly greater portion of users
(30.0%) than nonusers (20.7%) disagreed with the statement “eating a variety of foods gives all the vitamins and
minerals you need” (Jasti et al., 2003).
Healthy lifestyle characteristics are known to be associated with the use of dietary supplements. Supplement users
are more likely to have a lower body mass index, exercise regularly, eat more fruits, vegetables, and dietary fiber,
adhere to cancer screening recommendations, and to be nonsmokers than non-users (Foote et al., 2003; Greger,
2001; Hoggatt et al., 2002; Lyle et al., 1998; Neuhouser, 2003).
Use of supplements tends to be higher for those who assess themselves as being in very good or excellent health
compared with those rating their health as fair or poor (Ervin et al., 1999). In contrast, use of certain supplements is
associated with concurrent disease. Satia-Abouta et al. (2003) determined that supplement users with self-reported
coronary artery disease were 1.5 times as likely to use vitamin E supplements as those without such disease. Both
coronary artery disease and depression were associated with significantly higher supplement use among men. In the
study, calcium use was strongly associated with both acid reflux disease and frequent indigestion.
Maternal factors that positively associated with supplement use in children include non-Hispanic white ethnicity,
college education, use of dietary supplements prior to pregnancy, and higher income (Yu et al., 1997). Yu et al.
(1997) reported that children were more likely to receive single-iron supplements if their mothers were black, did
not graduate high school and had low income; the iron status of the child, i.e., anemia, was not reported.
C.4 LITERATURE CITATIONS FOR DIETARY SUPPLEMENTS
Anonymous (2003) State of the supplement, 2001. The Dietary Supplement Issue No. 12: 3.
Available at http://www.health.gov/dietsupp/final.pdf
Conner, M., Kirk, S. F., Cade, J. E. & Barrett, J. H. (2003) Environmental influences: factors influencing a woman's
decision to use dietary supplements. J. Nutr. 133: 1978S-1982S.
Dietary Supplement Information Bureau. (2003a) Exploring consumer attitudes about dietary supplement barometer
survey. Available at http://www.supplementinfo.org/latest_news/survey_results.htm. Accessed 7-17-2003.
Dietary Supplement Information Bureau. (2003b) The market for dietary supplements. Available at
http://www.supplementinfo.org/industry/marketplace.htm. Accessed 6-19-2003.
J. Epidemiol. 157: 888-897.
Glisson, J. K., Rogers, H. E. & Chambliss, W. G. (2003) Dietary supplements: important concerns for the clinician.
J. Miss. State Med. Assoc. 44: 35-38.
IARC Working Group (2002) Some traditional herbal ingredients. Introduction. In: IARC Monographs on the
Evaluation of Carcinogenic Risks to Humans. Some Traditional Herbal Medicines, Some Mycotoxins, Naphthalene
and Styrene. Volume 82. International Agency for Research on Cancer, Lyon, France. pp. 43-68.
Institute of Medicine (1998) Dietary Reference Intakes: A Risk Assessment Model for Establishing Upper Intake
Levels for Nutrients. National Academy Press, Washington, DC.
Jasti, S., Siega-Riz, A. M. & Bentley, M. E. (2003) Dietary supplement use in the context of health disparities:
cultural, ethnic and demographic determinants of use. J. Nutr. 133: 2010S-2013S.
Lanski, S. L., Greenwald, M., Perkins, A. & Simon, H. K. (2003) Herbal therapy use in a pediatric emergency
department population: expect the unexpected. Pediatrics 111: 981-985.
Lyle, B. J., Mares-Perlman, J. A., Klein, B. E., Klein, R. & Greger, J. L. (1998) Supplement users differ from
nonusers in demographic, lifestyle, dietary and health characteristics. J. Nutr. 128: 2355-2362.
Mackenzie, L. (2003) Choosing the right dietary supplement. Available at
http://www.healthylife.net/body/choosesupp.html. Accessed 6-20-2003.
National Nutritional Foods Association. (2003) Facts about the nutrition industry. Available at
http://www.nnfa.org/facts/index.htm. Accessed 8-25-2003.
Neuhouser, M. L., Patterson, R. E., Schwartz, S. M., Hedderson, M. M., Bowen, D. J. & Standish, L. J. (2001) Use
of alternative medicine by children with cancer in Washington state. Prev. Med. 33: 347-354.
Newman, V., Rock, C. L., Faerber, S., Flatt, S. W., Wright, F. A. & Pierce, J. P. (1998) Dietary supplement use by
women at risk for breast cancer recurrence. The Women's Healthy Eating and Living Study Group. J. Am. Diet.
Assoc. 98: 285-292.
Office of Dietary Supplements, National Institutes of Health. (2003) Dietary supplement use in the elderly. January
14-15, 2003. Overview. Available at http://dietary-supplements.info.nih.gov/pubs/elderly.overview.pdf. Accessed
4-25-2003.
Palmer, M. E., Haller, C., McKinney, P. E., Klein-Schwartz, W., Tschirgi, A., Smolinske, S. C., Woolf, A., Sprague,
B. M., Ko, R., Everson, G., Nelson, L. S., Dodd-Butera, T., Bartlett, W. D. & Landzberg, B. R. (2003) Adverse
events associated with dietary supplements: an observational study. Lancet 361: 101-106.
Patterson, R. E., Neuhouser, M. L., Hedderson, M. M., Schwartz, S. M., Standish, L. J., Bowen, D. J. & Marshall, L.
M. (2002) Types of alternative medicine used by patients with breast, colon, or prostate cancer: predictors, motives,
and costs. J. Altern. Complement. Med. 8: 477-485.
Radimer, K. L., Subar, A. F. & Thompson, F. E. (2000) Nonvitamin, nonmineral dietary supplements: issues and
findings from NHANES III. J. Am. Diet. Assoc. 100: 447-454.
Rodricks, J. V., Frankos, V. H. & Plunkett, L. M. (1995) Food Additives. In: Regulatory Toxicology. (Chengelis, C.
P., Holson, J. F. & Gad, S. C., eds.). Raven Press, Ltd., New York. pp. 51-81.
Satia-Abouta, J., Kristal, A. R., Patterson, R. E., Littman, A. J., Stratton, K. L. & White, E. (2003) Dietary
supplement use and medical conditions: the VITAL study. Am. J. Prev. Med. 24: 43-51.
Schaffer, D. M., Gordon, N. P., Jensen, C. D. & Avins, A. L. (2003) Nonvitamin, nonmineral supplement use over a
12-month period by adult members of a large health maintenance organization. J. Am. Diet. Assoc. 103: 1500-1505.
Stang, J., Story, M. T., Harnack, L. & Neumark-Sztainer, D. (2000) Relationships between vitamin and mineral
supplement use, dietary intake, and dietary adequacy among adolescents. J. Am. Diet. Assoc. 100: 905-910.
Texas A&M University. (2003) Amino acids. Available at http://ntri.tamuk.edu/cell/chapter3/amino-acids.html.
Accessed 6-20-2003.
Tomassoni, A. J. & Simone, K. (2001) Herbal medicines for children: an illusion of safety? Curr. Opin. Pediatr. 13:
162-169.
U.S. Congress (1994) Dietary Supplement Health and Education Act of 1994. 103-417.
U.S. Food and Drug Administration (1997) Food labeling regulation, amendments; Food regulation uniform
compliance date; and new dietary ingredient premarket notification; Final rules. Fed. Reg. 62: 49826-49858.
U.S. Food and Drug Administration. (1999) Economic characterization of the dietary supplement industry. Final
Report. Available at http://vm.cfsan.fda.gov/~comm/ds-econ4.html. Accessed 6-19-2003.
U.S. Food and Drug Administration. (2001) Dietary Supplements. In: U.S. FDA Performance Plan and Summary.
Available at http://www.fda.gov/ope/fy02plan/dietsupp.html. Accessed 1-31-2003.
U.S. Food and Drug Administration (2003) Current good manufacturing practice in manufacturing, packing, or
holding dietary ingredients and dietary supplements. [Proposed Rule]. Fed Reg. 68: 12157-12263.
U.S. General Accounting Office (2001) Health products for seniors. "Anti-aging" products pose potential for
physical and economic harm. Report No. GAO-01-1129. U.S. General Accounting Office, Washington, DC. pp. 143.
Wallace, P. (2003) Groups seeking unified stance to influence final GMPs rule. Food Chem. News: 13-15.
White, J. (2002) PC-SPES—a lesson for future dietary supplement research. J. Natl. Cancer Inst. 94: 1261-1263.
Yu, S. M., Kogan, M. D. & Gergen, P. (1997) Vitamin-mineral supplement use among preschool children in the
United States. Pediatrics 100: E4.
APPENDIX D
KEY REGULATIONS FOR DIETARY SUPPLEMENTS
AND RELATED ACTIVITIES
The relatively new expansion of federal regulations concerning dietary supplements is attributable to keen
congressional interest, due in part to robust growth of sales and product use across all demographic groups of the
U.S. population. This appendix briefly describes current key federal regulations related to dietary supplements.
D.1 DIETARY SUPPLEMENT HEALTH AND EDUCATION ACT
The U.S. Food and Drug Administration (FDA) regulates dietary supplements by enforcing the Federal Food, Drug,
and Cosmetic Act as amended by the Dietary Supplement Health and Education Act (DSHEA) on October 25, 1994,
in that dietary supplements are regulated as a subcategory of food (U.S. Congress, 1994). However, dietary
supplements are not food additives, conventional food, or intended as the sole item of a meal or diet (U.S. Congress,
1994).
According to DSHEA (U.S. Congress, 1994), a dietary supplement is a product taken by mouth that contains a
“dietary ingredient” intended to supplement the diet. The “dietary ingredient” in these products may include
vitamins, minerals, herbs, or other botanicals, amino acids, and substances such as enzymes, organ tissues, and
metabolites. Homeopathic drug preparations are not dietary supplements.
Congress was motivated to enact DSHEA as a means to improve the health status of U.S. citizens by providing laws
that characterized an adulterated (unsafe) dietary supplement yet did not unjustifiably hinder access to dietary
supplements overall. Specifically, DSHEA does not require that FDA approve a dietary supplement before the
product is sold in the market. However, the manufacturer is required to notify FDA at least 75 days in advance of
introducing a product in the market that contains a dietary ingredient that has never been present in the food supply.
FDA is responsible for taking action against any dietary supplement product in the market if the product presents an
unreasonable or significant risk of illness or injury or has false or misleading labeling (Crawford, 2002). However,
FDA has the burden of proof to claim that a dietary supplement is unsafe before the product can be removed from
the market.
DSHEA prompted the creation of the Office of Dietary Supplements (ODS), formally established on November 27,
1995, within the Office of Disease Prevention, Office of the Director, at the National Institutes of Health (NIH).
ODS was directed to conduct and coordinate scientific research and funding related to dietary supplements within
NIH, specifically to explore whether dietary supplements can maintain and improve health. Moreover, ODS serves
as the principal advisor on issues relating to dietary supplements for the Secretary of Health and Human Services,
the Director of NIH, the Director of the Centers for Disease Control and Prevention, and the Commissioner of the
Food and Drug Administration.
Also as a result of DSHEA, the President appointed a Commission on Dietary Supplement Labels to provide
recommendations on evaluating and regulating product label claims (Commission on Dietary Supplement Labels,
1997).
D.2 PRODUCT LABELING, CLAIMS, AND ADVERTISEMENT
The intent of recommendations by the Commission on Dietary Supplement Labels was to provide product users with
truthful and scientifically valid information so they would not be misled by product labeling, but could make
appropriate health care choices for themselves and their families (Commission on Dietary Supplement Labels,
1997). Recommendations by the Commission on November 24, 1997 called for consideration of changes in existing
regulations, development of new regulations, and legislative action. In addition to numerous guidance statements,
the Commission specifically recommended that FDA establish a review panel for those botanical products that
might cross-over into the nonprescription drug category, i.e., claims proposed by manufacturers that certain
botanical products had preventative or therapeutic effects. Furthermore, they recommended that federal agencies
support research on the health benefits of dietary supplements, particularly to conduct research on botanical
products. Moreover, the Commission recommended that the industry consider establishing an expert advisory
committee to provide scientific review of label statements and claims and for guidance regarding the safety, benefit,
and appropriate labeling of specific products. On September 23, 1997, FDA issued a final rule, amending food-
labeling regulations concerning statements of identity and nutrition labeling of dietary supplements that went into
effect on March 23, 1999.
D.2.1 Supplement facts
On January 6, 2000, FDA published a final rule (U.S. Food and Drug Administration, 2000) defining the types of
claims that may be used on the label and in the labeling of dietary supplements without prior review by the agency.
FDA required that certain information be present on a dietary supplement label, including a descriptive name of the
product, a statement indicating that the product is a dietary supplement, and the name and place of business of the
manufacturer, packer, or distributor. In addition, each dietary supplement must have a "Supplement Facts" panel
(Figure D-1)(U.S. Food and Drug Administration, 1999). This information panel must identify each dietary
ingredient contained in the product. For botanicals, the part of the plant used must be identified (US Food and Drug
Administration, 1997a). A total weight must be provided for proprietary blends and its individual ingredients must
be listed in descending order by weight. The label must also contain information on the suggested serving size (U.S.
The product is considered misbranded if the label bears false or misleading information or does not bear the required
information (U.S. Congress, 1994). Certain dietary supplements containing protein, iron, or psyllium husk are also
required to bear a warning label as indicated in Figure D-2 (U.S. 21CFR101.17, 2002).
Individual state laws add some variation to the regulation of dietary supplements. For example, pertaining to
product labeling, a state may restrict the serving size (e.g., up to 25 mg/serving) and may choose to prohibit the use
of a product by minors.
Figure D-1. Three examples of Supplement Facts panels for labeling dietary supplements. Reprinted from U.S.
Food and Drug Administration (1999).
[Code of Federal Regulations]
[Title 21, Volume 2]
[Revised as of April 1, 2002]
From the U.S. Government Printing Office via GPO Access
CITE: 21CFR101.17]
PART 101--FOOD LABELING--Table of Contents
[Sec. 101.17 Food labeling warning, notice, and safe handling statements; excerpts for (d), (e), and (f) only.]
(d) Protein products. (1) The label and labeling of any food product in liquid, powdered, tablet, capsule, or similar forms that derives more
than 50 percent of its total caloric value from either whole protein, protein hydrolysates, amino acid mixtures, or a combination of these, and
that is represented for use in reducing weight shall bear the following warning:
WARNING: Very low calorie protein diets (below 400 Calories per day) may cause serious illness or death. Do Not Use for Weight Reduction
in Such Diets without Medical Supervision. Not for use by infants, children, or pregnant or nursing women.
(2) Products described in paragraph (d)(1) of this section are exempt from the labeling requirements of that paragraph if the protein products
are represented as part of a nutritionally balanced diet plan providing 400 or more Calories (kilocalories) per day and the label or labeling of
the product specifies the diet plan in detail or provides a brief description of that diet plan and adequate information describing where the
detailed diet plan may be obtained and the label and labeling bear the following statement:
Notice: For weight reduction, use only as directed in the accompanying diet plan (the name and specific location in labeling of the diet plan
may be included in this statement in place of “accompanying diet plan”). Do not use in diets supplying less than 400 Calories per day without
medical supervision.
(3) The label and labeling of food products represented or intended for dietary (food) supplementation that derive more than 50 percent of
their total caloric value from either whole protein, protein hydrolysates, amino acid mixtures, or a combination of these, that are represented
specifically for purposes other than weight reduction; and that are not covered by the requirements of paragraph (d) (1) and (2) of this section;
shall bear the following statement:
Notice: Use this product as a food supplement only. Do not use for weight reduction.
(4) The provisions of this paragraph are separate from and in addition to any labeling requirements promulgated by the Federal Trade
Commission for protein supplements.
(5) Protein products shipped in bulk form for use solely in the manufacture of other foods and not for distribution to consumers in such
container are exempt from the labeling requirements of this paragraph.
(6) The warning and notice statements required by paragraphs (d) (1), (2), and (3) of this section shall appear prominently and conspicuously
on the principal display panel of the package label and any other labeling.
(e) Dietary supplements containing iron or iron salts. (1) The labeling of any dietary supplement in solid oral dosage form (e.g., tablets or
capsules) that contains iron or iron salts for use as an iron source shall bear the following statement:
WARNING: Accidental overdose of iron-containing products is a leading cause of fatal poisoning in children under 6. Keep this product out
of reach of children. In case of accidental overdose, call a doctor or poison control center immediately.
(2)(i) The warning statement required by paragraph (e)(1) of this section shall appear prominently and conspicuously on the information panel
of the immediate container label. (ii) If a product is packaged in unit-dose packaging, and if the immediate container bears labeling but not a
label, the warning statement required by paragraph (e)(1) of this section shall appear prominently and conspicuously on the immediate
container labeling in a way that maximizes the likelihood that the warning is intact until all of the dosage units to which it applies are used.
(3) Where the immediate container is not the retail package, the warning statement required by paragraph (e)(1) of this section shall also
appear prominently and conspicuously on the information panel of the retail package label.
(4) The warning statement shall appear on any labeling that contains warnings.
(5) The warning statement required by paragraph (e) (1) of this section shall be set off in a box by use of hairlines.
(f) Foods containing psyllium husk. (1) Foods containing dry or incompletely hydrated psyllium husk, also known as psyllium seed husk,
and bearing a health claim on the association between soluble fiber from psyllium husk and reduced risk of coronary heart disease, shall bear a
label statement informing consumers that the appropriate use of such foods requires consumption with adequate amounts of fluids, alerting
them of potential consequences of failing to follow usage recommendations, and informing persons with swallowing difficulties to avoid
consumption of the product (e.g., “NOTICE: This food should be eaten with at least a full glass of liquid. Eating this product without enough
liquid may cause choking. Do not eat this product if you have difficulty in swallowing.”). However, a product in conventional food form may
be exempt from this requirement if a viscous adhesive mass is not formed when the food is exposed to fluids.
(2) The statement shall appear prominently and conspicuously on the information panel or principal display panel of the package label and any
other labeling to render it likely to be read and understood by the ordinary individual under customary conditions of purchase and use. The
statement shall be preceded by the word “NOTICE” in capital letters.
Figure D-2. Federal warning label requirements for dietary supplements. Excerpted from U.S. 21CFR101.17
(2002).
D.2.2 Label claims
Health claims in food labeling are explicit or implied statements characterizing the relationship of the substance to a
disease or health-related condition, excluding nutrient deficiencies (Pape et al., 2001). Dietary supplements may not
bear health claims unless the claim has undergone premarket review by FDA and has been authorized or approved
under the rules for health claims or drugs (U.S. Food and Drug Administration, 2002). Health claims that are
approved by FDA can be included on the product label, such as the claim that intake of folic acid is associated with
reduced risk of neural tube defects, and that intake of calcium may reduce the risk of osteoporosis (U.S. Congress,
1994).
Other claims, referred to as “structure/function” claims, may also be made on the product label. Structure/function
claims describe the role of a nutrient or dietary ingredient intended to affect the structure/function in humans or
characterizes the mechanism by which a nutrient or dietary ingredient acts to maintain such structure/function.
These claims may be added to the label without FDA preapproval if the manufacturer has substantiation that the
claims are truthful and not misleading and if the label also bears the disclaimer “This statement has not been
evaluated by FDA. This product is not intended to diagnose, treat, cure, or prevent any disease” (U.S. Food and
Drug Administration, 2000). The manufacturer must notify FDA of the structure/function claim within 30 days of
marketing the product. FDA discourages structure/function claims related to pregnancy.
On July 11, 2000, the U.S. General Accounting Office (2000) provided a report to Congress in which they discussed
the extent to which federal laws and federal agency efforts were able to ensure the safety of dietary supplements and
functional foods and the accuracy of health-related claims on product labels and in advertising. Based on several
weaknesses they identified in federal controls that could potentially allow unsafe products to reach the public, they
recommended that Congress establish an expert panel to reexamine the current approach to labeling to determine
whether the distinctions in health claims and structure/function claims can be made more clear and meaningful for
the product user (U.S. General Accounting Office, 2000). They also recommended several actions that FDA should
take:
• Provide regulations and/or guidance for industry on the evidence needed to document the safety of a new
ingredient
• Clarify the distinction between dietary supplements and conventional foods, especially between dietary
supplements and functional foods
• Provide regulations and/or guidance for industry on the safety-related information required on dietary
supplement labels
• Improve the system to record and analyze adverse events associated with use of dietary supplements
FDA has since prepared a guidance document to assist manufacturers in determining whether a claim is a
structure/function claim or a disease claim (U.S. Food and Drug Administration, 2002).
D.2.3 Product advertisement
The Federal Trade Commission (FTC) is responsible for enforcing laws prohibiting false and misleading health
claims in the advertising of dietary supplements. Advertising includes print and broadcast advertisements,
infomercials, catalogs, and marketing on the Internet (Pape et al., 2001). FTC prepared a guidance document to
provide a detailed discussion of what advertisers must do to comply with requirements (Federal Trade Commission,
2001a).
FTC requires that dietary supplement manufacturers substantiate all claims involving health, safety, or efficacy in
advertising with competent and reliable scientific evidence before the claims are made. The failure to do so can
result in injunctions prohibiting the future use of false or unsubstantiated claims and financial penalty and
compensation.
For example, in the marketing of St. John's wort, FTC challenged ads claiming that product users could safely use
the product to treat a variety of diseases, including hepatitis (Federal Trade Commission, 2001c, 2001e). The ads
also falsely claimed that ingestion of the product had no known contraindications. In this case, FTC's settlement
with the manufacturer required the manufacturer to issue a warning on all labeling and advertising alerting product
users to the dangerous interactions that St. John's wort has with certain prescription drugs. Product information was
also required to contain the recommendation that product users consult a physician before taking the product with
certain kinds of drugs. Moreover, the settlement required the manufacturer to send a notice to all purchasers of the
product informing them of the settlement and offer a full refund (Federal Trade Commission, 2001b, 2001d).
D.3 CURRENT GOOD MANUFACTURING PRACTICE
DSHEA indicated that the Secretary of Health and Human Services may regulate good manufacturing practice for
dietary supplements (U.S. Congress, 1994). Subsequently, FDA (2003) proposed to establish minimum standards,
e.g. current good manufacturing practice (CGMP), for dietary supplements. The proposed CGMP rule is intended to
introduce widely accepted standards into the dietary supplement industry. Specifically, CGMP helps prevent
unsanitary manufacturing that predisposes to product contamination, and inconsistent product batches that cause
variation from labeled specifications. It is generally believed that the use of CGMP leads to consistent identification,
purity, quality, strength, and composition of finished products for sale. Hence, application of good manufacturing
practice by industry would serve to help prevent problems in production that lead to adulterated or misbranded
products, which could cause adverse effects in product users (U.S. Food and Drug Administration, 2003).
D.3.1 Proposed good manufacturing practice for the review of complaints
The proposed CGMP for dietary supplements suggests regulations for how a manufacturer should handle product
user complaints involving illness or injury that are related to manufacturing production and process (U.S. Food and
Drug Administration, 2003). Specifically, FDA proposed procedures for recording, reviewing, and investigating
consumer product complaints associated with dietary supplements (U.S. Food and Drug Administration, 2003).
This type of consumer feedback can help identify previously unknown manufacturing deviations that cause
adulteration. Note, however, that complaints specifically related to the pharmacological activity of an intended
dietary ingredient are beyond the scope of the proposed CGMP, and that the manufacturer's response to this type of
complaint is also not regulated by the proposed CGMP (U.S. Food and Drug Administration, 2003).
FDA (2003) proposed that written records of complaints be maintained by manufacturers of dietary supplements and
that at least basic information should be recorded:
1. Identification of the dietary ingredient or supplement (name, description)
2. The batch and lot number when available
3. The reporter’s name, if available
4. The nature of the complaint, including how the consumer used the product
5. The response of the manufacturer to the reporter, if any
6. Findings of the investigation and any follow-up action that was taken
FDA proposed that a manufacturer should investigate a complaint when there was a reasonable possibility of a
relationship between the quality of a dietary supplement and the adverse event. The investigation would include
batch records associated with the dietary supplement involved in the complaint, as well as records of other batches
of products that may not have met specifications and that may have been associated with an adverse event.
Laboratory testing was recommended when a possible product defect, or a failure to determine whether the product
specifications or requirements were actually met, had occurred (U.S. Food and Drug Administration, 2003).
FDA proposed that complaints of serious illness or injury that are attributable to a product defect should be
evaluated by a competent medical authority to determine if follow-up action is necessary to protect the public health.
While a “serious” problem was not defined in the proposed rule, several conditions were described as being serious,
including renal failure and significant infections, such as tuberculosis or transmissible spongiform encephalopathy
(U.S. Food and Drug Administration, 2003). FDA encouraged industry to report serious adverse events.
Specifically, FDA (2003) strongly suggested, but will not likely require, that manufacturers, distributors, and
retailers notify FDA of possible associations between a dietary supplement and a serious adverse event within 15
days of receipt of the complaint by filing a report to MedWatch.
Federal and state legislative actions and judicial opinions are expected to undergo continuing development,
propelled by product user experience, research, and litigation in this field.
D.4 LITERATURE CITATIONS FOR KEY REGULATIONS
21CFR101.17 (2002) Food labeling warning, notice, and safe handling statements. Code of Federal Regulations 21,
Volume 2: 66-72.
Available at http://www.health.gov/dietsupp/final.pdf
Crawford, L. M. (5-29-2002) Dietary supplement strategic plan cost out. (Report to Congress). [House Committee
(Report No. 106-619)]. Available at http://www.cfsan.fda.gov/~dms/ds-stra2.html. Accessed 8-1-2002.
Federal Trade Commission. (2001a) Dietary supplements: An advertising guide for industry. Available at
http://www.ftc.gov/bcp/conline/pubs/buspubs/dietsupp.pdf. Accessed 10-8-2003a.
Federal Trade Commission. (2001b) ForMor, Inc. Agreement containing consent order. Available at
http://www.ftc.gov/os/2001/06/formorconsent.pdf. Accessed 10-8-2003b.
Federal Trade Commission. (2001c) ForMor, Inc. FTC Complaint 4021. Available at
http://www.ftc.gov/os/2001/06/formorcmp.pdf. Accessed 10-8-2003c.
Federal Trade Commission. (2001d) Panda Herbal International, Inc. Agreement containing consent order.
Available at http://www.ftc.gov/os/2001/06/pandaconsent.pdf. Accessed 10-8-2003d.
Federal Trade Commission. (2001e) Panda Herbal International. FTC Complaint 4018. Available at
http://www.ftc.gov/os/2001/06/pandacmp.pdf. Accessed 10-8-2003e.
Pape, S. M., Kracov, D. A. & Rubin, P. D. (2001) Dietary Supplements and Functional Foods. A Practical Guide to
FDA Regulation. (Parisi, G. P., ed.). Thompson Publishing Group, Inc., Tampa, FL.
U.S. Food and Drug Administration (1997a) Food labeling regulation, amendments; Food regulation uniform
compliance date; and new dietary ingredient premarket notification; Final rule. Fed. Reg. 62: 49826-49858.
U.S. Food and Drug Administration. (1999) Summary of nutrition labeling rules for dietary supplements:
Supplement facts panel. Available at http://www.cfsan.fda.gov/~acrobat/hhssupp2.pdf. Accessed 8-8-2003.
U.S. Food and Drug Administration (2000) Regulations on statements made for dietary supplements concerning the
effect of the product on the structure or function of the body. Fed. Reg. 65: 999-1050.
U.S. Food and Drug Administration. (1-9-2002) Guidance for industry. Structure/function claims. Small entity
compliance guide. Available at http://vm.cfsan.fda.gov/~dms/sclmguid.html. Accessed 10-8-2003.
U.S. General Accounting Office (7-1-2000) Food Safety. Improvements Needed in Overseeing the Safety of Dietary
Supplements and "Functional Foods." Report No. GAO/RCED-00-156. Government Printing Office, Washington,
D.C. pp. 1-43.
APPENDIX E
EXAMPLE SURVEILLANCE PROGRAMS
FOR FOOD ADDITIVES
In this appendix, unique characteristics of surveillance programs for the food additives aspartame and olestra are
highlighted. As part of the approval process for these new food additives, the U.S. Food and Drug Administration
(FDA) requested that postmarketing surveillance be conducted by the manufacturers. The postmarketing
surveillance activities described in this appendix were considered, at the time, extraordinary for a food product.
Because the general principles of surveillance programs are similar, these experiences may be useful in designing a
surveillance program for dietary supplements.
E.1 SURVEILLANCE PROGRAM FOR ASPARTAME
Aspartame (NutraSweet Company, Chicago, IL) was approved as a food additive in dry form, in 1981, and as a
sweetener for carbonated beverages in 1983 (Tollefson, 1988).
The NutraSweet Company initiated a toll-free number for product users to contact the company for information
related to aspartame. The company’s Consumer Center collected individual data records (IDRs) related to ingestion
of aspartame for 12 years, from 1982 through 1993 (Butchko & Stargel, 2001). According to Butchko and Stargel
(2001), this type of program for postmarketing surveillance had never before been conducted for a food additive.
Bias introduced by negative media coverage had a direct impact on reporting (Butchko et al., 2002). For example,
a disproportionate number of IDRs originated in Arizona, coinciding with substantial negative media coverage in
that state. Figure E-1 demonstrates the effect of the media on IDRs in 1984 (Butchko et al., 1994; 2002; Butchko &
Kotsonis, 1994).
120
100
Number of reports
80
60
40
20
em
be
r
O
ct
ob
er
N
ov
em
be
D
r
ec
em
be
r
us
t
pt
Se
Au
g
Ju
ly
ne
Ju
ay
M
il
Ap
r
ch
ar
y
ar
M
y
br
u
Fe
Ja
nu
ar
0
Figure E-1. Monthly anecdotal medical complaints for aspartame in 1984 following adverse media reports in mid
January. Copyright 2002 from Regulatory Toxicology and Pharmacology, by Butchko et al. Reprinted with
permission of Elsevier Publications and H.H. Butchko.
The number of IDRs peaked at approximately 750 IDRs per year in 1985 (two years after aspartame was added to
carbonated beverages) and declined to approximately 300 per year by 1988 despite increasing numbers of products
containing aspartame and increasing product sales (Figure E-2)(Butchko et al., 1994). Product users initiated the
majority of health-related complaints. Callers were more likely to be Caucasian females between 21 and 60 years of
age (Butchko et al., 1996).
Physicians in the company’s Clinical Research Group provided medical expertise for the evaluation of the healthrelated IDRs and follow-up investigations were conducted.
Clinical information was collected by the Consumer Center for health-related calls to the aspartame information line
(Butchko et al., 1994; Butchko & Kotsonis, 1994):
• Description of symptoms for the health-related complaint
• Products associated with symptoms
• Time to onset of symptoms after product consumption
• Whether symptoms disappeared upon cessation of consumption of the product and reoccurred upon
rechallenge
• Concurrent medical conditions and medications
• Whether medical care for the symptoms was received
• Whether the product user was on a weight-reduction regimen
If medical care was provided, an attempt was made to collect and document the medical opinion of the health care
professional (Butchko et al., 1994). A standardized dictionary was used to code reported symptoms based on a
simplified version of the World Health Organization’s coding system (Butchko et al., 1994).
Figure E-2.
Trends from 1982 to 1993 in the numbers of reports of anecdotal medical complaints about
aspartame products and the numbers of marketed products with aspartame. Copyright 1994 from Regulatory
Toxicology and Pharmacology, by Butchko et al. Reprinted with permission of Elsevier Publications and H.H.
Butchko.
Anecdotal IDRs were tabulated by the company and shared monthly with FDA’s Center for Food Safety and
Applied Nutrition for use in the federal Adverse Reaction Monitoring System (Butchko et al., 1994).
The company evaluated anecdotal IDRs on a monthly, quarterly, and annual basis (Butchko et al., 1994). At the
request of FDA, the Centers for Disease Control and Prevention (CDC) evaluated IDRs associated with aspartame
collected by the company through 1984. CDC concluded that adverse events in the IDRs were generally mild in
nature and were consistent with non-specific complaints common in the general population. No specific symptoms
were identified that could be associated with aspartame.
CDC recommended that focused clinical studies would be the best way to address some of the issues raised by the
anecdotal IDRs (Butchko et al., 1994; Butchko et al., 1996; Butchko & Stargel, 2001). Several studies were
subsequently initiated in humans to investigate whether ingestion of aspartame was associated with headaches,
allergic reactions, cognitive performance, behavior, and seizure activity (Butchko & Stargel, 2001).
In addition to requesting experimental studies, FDA requested that the manufacturer collect measures of actual
intake of aspartame-containing products released in the market. These measures were expected to provide a more
reliable estimate of aspartame exposure than premarket projections (Butchko et al., 1994). Hence, a collection of
data of actual aspartame intake was obtained (MRCA Information Services, Northbrook, IL) from 1984 through
1992 using 14-day diet records in more than 2000 U.S. households, representing intake records for more than 5000
people per year for eight years (Butchko et al., 1994; Butchko & Stargel, 2001). The intakes of special
subpopulations were also monitored, such as people in weight-reduction programs who potentially could have
higher than average intakes of aspartame. Moreover, data were reviewed on aspartame intake obtained from 1-day
diaries as part of the U.S. Department of Agriculture’s Continuing Survey of Food Intakes by Individuals (Butchko
& Stargel, 2001).
By 1995, FDA had received a total of 7232 IDRs associated with aspartame and concluded at that time that data
from IDRs and the clinical and epidemiological studies were sufficient to determine that aspartame was not
associated with adverse events. Hence, FDA decided that there was no need for any further controlled clinical trials
(Butchko & Stargel, 2001).
E.2 SURVEILLANCE PROGRAM FOR OLESTRA
Olestra (Procter & Gamble, Cincinnati, OH) was the first fat substitute to be evaluated by FDA. It was approved by
FDA in 1996 for use in packaged snacks and subsequently underwent a formal scheduled FDA postmarketing
review at 30 months (U.S. Food and Drug Administration, 1998a).
In order to collect postmarketing surveillance data for FDA and provide product users with pertinent information
about olestra products, Proctor & Gamble established a consumer hotline (Slough et al., 2001). Calls alleging
adverse events were managed according to the diagram in Figure E-3. During regular business hours, trained
personnel in consumer relations departments within Procter & Gamble and Frito-Lay received health-related
complaints. Medically urgent calls received after regular business hours were routed to a university drug and poison
information center.
The majority of complaints were from adult females (1.9 female: 1 male), most under the age of 65 years (Allgood
et al., 2001). The rate of incoming complaints declined over time. An initial surge of complaints occurred when the
product was introduced into the market and transient spikes in complaints occurred following increased media
attention to the product.
Figure E-3. Procter & Gamble’s (Cincinnati, OH) creation of individual data records and management of other
consumer contacts for olestra-containing snacks. Copyright 2001 from Regulatory Toxicology and Pharmacology,
by Slough et al. Reprinted with permission of Elsevier Publications.
Data describing demographics, product consumption, and the symptoms experienced were collected using a
standardized data collection tool and were recorded electronically. Standardized terms were use to describe disease
history (World Health Organization, 2003a), concomitant medications ( World Health Organization, 2003b), and
clinical effects experienced by the product user (FDA Coding Symbols for a Thesaurus of Adverse Reaction
Terms).15 Each IDR was reviewed within 24-hours of receipt by a registered nurse in the Medical Affairs
department, with physician oversight.
E.2.2.1 Follow-up information
Procter & Gamble attempted to obtain consent for the release of medical records to allow for appropriate follow up.
The reporter and/or the product user’s health care professional were subsequently contacted for certain IDRs
requiring follow up:
• To determine if symptoms had resolved for those product users who were experiencing symptoms at the
time of the call (the duration of symptoms could often be characterized at follow up)
• To further investigate unusual symptoms
• To obtain missing/additional information that was necessary for evaluating the complaint
• To obtain medical records and/or speak with the health care professional if the product user had received
medical care
Consent to obtain medical records or physician input was not available in all cases. Because follow up was not
always possible, it was important to obtain detailed and complete information at the time of the initial contact.
IDRs were downloaded to the Medical Affairs database every business day. A commercial software package
(ALERT, DLB Systems, NJ)16 was used to manage, track, and analyze the IDRs.
The numbers of incoming IDRs were tracked as a population-based rate (census data) and as a percentage of total
product user calls. Reporting rates were based on sales volume in some test markets where this could be calculated
(Slough et al., 2001). The number of complaints by age, gender, and quantity of snack consumed were examined
and the most frequently reported symptoms were identified (Allgood et al., 2001). Summaries of the adverse event
data were submitted quarterly to FDA and used by FDA in the 30-month post-approval review.
Additional safety information was obtained through several studies, one collecting food and snack intake patterns
and measuring biochemical parameters of 6,000 people (U.S. Food and Drug Administration, 1998a). Four placebocontrolled clinical trials totaling more than 4,400 subjects were also conducted to examine the effect of (1) a single
serving of Olestra, and (2) long-term daily use (Allgood et al., 2001; Slough et al., 2001; U.S. Food and Drug
Administration, 1998a).
Accurate prevalence estimates in the general population of the types of health-related events submitted by product
users were needed to understand the background “noise” level for comparison (U.S. Food and Drug Administration,
1998a). Hence, a survey of digestive complaints was conducted in the general population and compared with the
prevalence of these conditions in the IDRs (U.S. Food and Drug Administration, 1998a).
15
Coding Symbols for a Thesaurus of Adverse Reaction Terms (COSTART) is the terminology developed and
previously used by FDA for the coding, filing and retrieving of postmarketing adverse events. In November 1997,
FDA replaced its Spontaneous Reporting System and COSTART with the Adverse Events Reporting System and the
Medical Dictionary for Regulatory Activities (MedDRA®) Terminology.
16
DLB Systems was purchased by Premier Research, which is a division of SCP Communications Inc. in New York,
NY.
An independent medical advisory panel composed of five physicians specializing in gastroenterology,
epidemiology, and/or postmarketing surveillance of pharmaceutical products periodically reviewed the IDRs and
data obtained from clinical trials (Allgood et al., 2001; Slough et al., 2001; U.S. Food and Drug Administration,
1998a).
FDA’s 30-Month Foods Advisory Committee Review was provided with summary information, obtained from
postmarketing surveillance activities:
• The most frequently reported health-related events among the IDRs
• The most frequent digestive problems from surveys of the general public
• The amount of olestra snacks that were consumed among individuals who reported symptoms
• The number of days that olestra snacks were consumed in individuals who reported symptoms
• The time to onset of symptoms, grouped by symptom
• The duration of the reported symptoms
• Demographic data of those experiencing the reported symptoms and of the subgroup seeking medical
treatment
FDA continues to permit the addition of olestra as a fat substitute in savory snack food. In 2003, FDA reviewed the
accumulated evidence to ensure that these products were not misbranded (U.S. Food and Drug Administration,
2003). FDA considered whether the potential health effects of olestra-containing foods rose to the level that
warranted special labeling and whether product users were aware of the potential gastrointestinal effects associated
with the consumption of olestra-containing foods.
As a result, FDA amended the food additive regulations (U.S. Food and Drug Administration, 2003). Olestracontaining foods will no longer be required to bear a label statement informing product users of possible
gastrointestinal symptoms from consumption of olestra. Also, a label statement informing product users of possible
effects of olestra on the absorption of some vitamins and other nutrients will no longer be required. Finally, the
listing of the vitamins in the ingredient statement of olestra-containing foods will now be followed by an asterisk
that is linked to the statement “dietarily insignificant” instead of the previous requirement for a statement informing
product users that vitamins A, D, E and K had been added.
E.3 LITERATURE CITATIONS FOR EXAMPLE SURVEILLANCE PROGRAMS FOR FOOD
ADDITIVES
Butchko, H. H. & Kotsonis, F. N. (1994) Postmarketing surveillance in the food industry: The aspartame case
study. In: Nutritional Toxicology. (Kotsonis, F. N., Mackey, M. & Hjelle, J., eds.). Raven Press, Ltd., New York,
pp. 235-250.
Butchko, H. H. & Stargel, W. W. (2001) Aspartame: scientific evaluation in the postmarketing period. Regul.
Toxicol. Pharmacol. 34: 221-233.
Butchko, H. H., Stargel, W. W., Comer, C. P., Mayhew, D. A., Benninger, C., Blackburn, G. L., de Sonneville, L.
M., Geha, R. S., Hertelendy, Z., Koestner, A., Leon, A. S., Liepa, G. U., McMartin, K. E., Mendenhall, C. L.,
Munro, I. C., Novotny, E. J., Renwick, A. G., Schiffman, S. S., Schomer, D. L., Shaywitz, B. A., Spiers, P. A.,
Tephly, T. R., Thomas, J. A. & Trefz, F. K. (2002) Aspartame: review of safety. Regul. Toxicol. Pharmacol. 35: S193.
Butchko, H. H., Tschanz, C. & Kotsonis, F. N. (1994) Postmarketing surveillance of food additives. Regul. Toxicol.
Pharmacol. 20: 105-118.
Raton, FL, pp. 183-193.
Slough, C. L., Miday, R. K., Zorich, N. L. & Jones, J. K. (2001) Postmarketing surveillance of new food ingredients:
design and implementation of the program for the fat replacer olestra. Regul. Toxicol. Pharmacol. 33: 218-223.
Tollefson, L. (1988) Monitoring adverse reactions to food additives in the U.S. Food and Drug Administration.
Regul. Toxicol. Pharmacol. 8: 438-446.
U.S. Food and Drug Administration. (6-15-1998a) Food Advisory Committee on Olestra. Volume 1 [Transcript].
Available at http://www.fda.gov/ohrms/dockets/98/transcpt/3485t1.pdf. Accessed 7-8-2003a.
U.S. Food and Drug Administration. (6-17-1998b) Food Advisory Committee on Olestra. Volume 3 [Transcript].
Available at http://www.fda.gov/ohrms/dockets/98/transcpt/3485t3.pdf. Accessed 7-8-2003b.
U.S. Food and Drug Administration (2003) Food additives permitted for direct addition to food for human
consumption; Olestra; Final rules. Fed. Reg. 68: 46363-46402.
World Health Organization. (2003a) Frequently asked questions (FAQs) about ICD (International Classification of
Diseases). Available at http://www.who-umc.org/faqs/faqicd.html. Accessed 5-20-2003.
World Health Organization. (2003b) WHO Drug Dictionary. Available at http://www.whoumc.org/pdfs/WHO_Drug_Dictionary_Guide.pdf. Accessed 9-30-2003.
APPENDIX F
EXISTING SURVEILLANCE PROGRAMS
There are several existing surveillance programs collecting adverse event information for dietary supplements.
These existing programs were not designed specifically for dietary supplements but include dietary supplements as a
component of a larger surveillance program. All three surveillance programs presented in this appendix, those of the
U.S. Food and Drug Administration (FDA), the American Association of Poison Control Centers (AAPCC), and the
World Health Organization (WHO), utilize inter-product databases, which year-by-year grow larger in total
accumulated records.
F.1 SURVEILLANCE PROGRAMS AT THE FOOD AND DRUG ADMINISTRATION
The Medical Products Reporting Program, referred to as “MedWatch,” is the FDA’s safety information and adverse
event monitoring program. It includes mandatory reporting by drug and medical device manufacturers as well as a
spontaneous reporting system. The spontaneous portion of the program receives reports about serious reactions and
problems associated with medications (prescription and nonprescription), medical devices, special nutritional
products (dietary supplements, medical foods, infant formulas), and cosmetics. Product problems involve quality,
performance, and safety concerns, which include suspected contamination, questionable stability, defective
components, poor packaging or labeling, and therapeutic failures (U.S. Food and Drug Administration, 2003f).
Serious adverse events are defined as patient outcomes that required intervention to prevent permanent impairment,
a congenital anomaly, significant disability, or hospitalization, was life threatening, or resulted in death (U.S. Food
and Drug Administration, 2003g).
F.1.1 Development of the federal surveillance program
According to Kennedy et al. (2000), the first requirement of manufacturers to report adverse events to FDA came in
response to the worldwide thalidomide tragedy of 10,000 birth defects. Specifically, the U.S. Congress included a
requirement for reporting for all products having a New Drug Application as part of the 1962 Kefauver-Harris
Amendments to the Food, Drug and Cosmetic Act of 1938 (Kennedy et al., 2000). A few years later, the first
computerized storage system for adverse events, the Spontaneous Reporting System, was initiated. The earliest
retrievable individual data records (IDRs) were submitted in 1969 (Kennedy et al., 2000). In June 1993, FDA
launched MedWatch to consolidate various reporting programs. A new standardized format (reporting form) was
then used for both voluntary and mandatory information for all FDA regulated products. In November 1997, the
Spontaneous Reporting System was replaced with the computerized Adverse Event Reporting System
(AERS)(Kennedy et al., 2000). The MedWatch form continues to be updated and used for data collection.
Traditionally, health care professionals submitted the majority of adverse event reports with pharmacists
contributing more reports than physicians (Kennedy et al., 2000). While the rules and regulations governing adverse
event reporting had historically been aimed at pharmaceutical manufacturers, new systems were aimed at health care
professionals, product users, and others (Kennedy et al., 2000). MedWatch was designed to increase the proportion
of serious reports to FDA over the previous system, not necessarily the volume of reports (Goldman, 2003). The
introduction of the MedWatch form had a mostly positive impact on completeness of data collection, except for data
on concomitant exposure to medication. After MedWatch was implemented, pharmacists improved their reporting
of laboratory and clinical information and increased the number of serious reports they submitted voluntarily
(Piazza-Hepp & Kennedy, 1995). The reporting of concomitant exposure to medications by both pharmacists and
physicians decreased significantly after MedWatch began (Piazza-Hepp & Kennedy, 1995). Piazza-Hepp and
Kennedy (1995) attributed this problematic decline to a reorganization of the reporting form whereby the data field
for concomitant exposure to medications on the MedWatch form was less prominent than on previous forms.
The FDA Center for Food Safety and Applied Nutrition (CFSAN) established the Special Nutritional Adverse Event
Monitoring System in 1993 to monitor and evaluate anecdotal reports of adverse health events associated with
foods, food additives, color additives, and dietary supplements (Tollefson, 1988). The web-accessible database was
designed to assist in searching for adverse events associated with CFSAN regulated products and to search for
specific products or ingredients that were associated with particular health-related events. The CFSAN website was
criticized for containing misleading and incorrect information and for rarely being updated (U.S. Department of
Health and Human Services. Office of Inspector General, 2001). As part of an effort to undergo systematic
improvement, the website was removed in August 1999 (U.S. Food and Drug Administration, 2002). CFSAN
received approximately 7,000 complaints annually concerning the products it regulates, approximately one-half of
which were adverse events involving illness or injury. Currently, CFSAN directs reporters of adverse events related
to the use of dietary supplements to FDA’s MedWatch system.
The numbers of incoming reports (for all FDA regulated products) increased over time, approximately 14,000 in
1980, 83,000 in 1990 (Kennedy et al., 2000), 130,000 in 1994 (Piazza-Hepp & Kennedy, 1995), 230,000 in 1998,
and 250,000 annually in recent years from domestic and international sources (Bright & Nelson, 2002; Faich, 2003).
As of FY 1999 there were more than two million IDRs in the AERS database (U.S. Food and Drug Administration,
1999).
FDA encourages reporting even if the reporter is not certain whether the product caused the event and even when all
details are not available.
The data collection form was designed for use by a health care professional (Goldman, 2003). Also, FDA systems
were designed to accept information, but not to provide clinical advice. Hence, product users are encouraged to
have their health care professional complete and submit reports, which arrive via postal service, facsimile,
telephone, and electronic mail (U.S. Food and Drug Administration, 1999).
Adult female product users submit more adverse event reports directly to FDA (rather than via their health care
provider) than do male product users (Figure F-1)(Clark, 2003). This gender difference is also evident in reporting
rates in surveillance programs for food additives (Allgood et al., 2001; Butchko et al., 1996).
50
Reporting Rate
40
30
20
10
<5
5-9
10
-14
15
-19
20
-24
25
-29
30
-34
35
-39
40
-44
45
-49
50
-54
55
-59
60
-64
65
-69
70
-74
75
-79
80
-84
85
-89
90
-94
95
-99
>1
00
0
Age Range
FEMALES
MALES
Figure F-1. Adverse event reporting rates/100,000 U.S. population for age and gender by consumers reporting
directly to the U.S. Food and Drug Administration (2003). Source of data: Food and Drug Administration's
Freedom of Information database for 1998 - 2001 averaged age-gender adjusted rates of adverse event reports; U.S.
Census Bureau population counts by age-gender stratum. Reprinted with permission from Clark J.A. Galt
Associates, Inc., Sterling, VA.
A study conducted in 1994 reported an average lag time of 58 days between the day of an event and FDA’s receipt
of a MedWatch form, with a range of two days to 2190 days (Piazza-Hepp & Kennedy, 1995). The following
factors have traditionally affected whether health care professionals report adverse events (Figueiras et al., 1999;
Goldman, 2003; Hasford et al., 2002; Juergens et al., 1992; Koch-Weser et al., 1969; Scott et al., 1988):
• The level of certainty as to whether or not the product caused the event, even though assessing causality is
not a requirement for submission
• Whether the event was expected (e.g., was a recognized dose dependent event) or unexpected (e.g., an
allergic reaction)
• Whether the event was considered serious versus trivial
• The duration of time from exposure to onset of the event
• Whether hospitalization was prolonged by the event
• The morbidity associated with the event
• Understanding the benefit of reporting
• Understanding the clinical value of submitting a single report of the adverse event
• The time required to complete and submit the report
Goldman (1998) and Hostelley (2003) underscored the need for ongoing efforts to educate health care professionals
about reporting systems. Education efforts can demystify the process of adverse event reporting and clarify how
such information is used for monitoring. These education efforts may increase participation of health care
professionals and improve IDR quality.
The standardized data collection tool used by MedWatch includes patient information, event information, and
product information (Figure F-2)(U.S. Department of Health and Human Services, 2002). This data collection form
was primarily designed for industry professionals and health care professionals to submit adverse event reports of
prescription medication, but is used for all FDA covered products, including dietary supplements.
Follow-up information is systematically sought from the reporter or manufacturer as needed. The types of follow-up
information that may be collected by FDA and included in the IDR are listed in section III of the main body of the
report, “Retrospective Review of Sample IDRs.”
All spontaneous IDRs are reviewed on a case-by-case basis by an FDA safety evaluator, usually a physician or
pharmacist (Kennedy et al., 2000).
F.1.4 Signaling from the MedWatch database
Computerized analyses of IDRs are conducted using Bayesian statistical methods (DuMouchel, 1999) to identify
potential signals (Kennedy et al., 2000).
New (unlabeled) product-event associations are used to initiate a case series. Criteria defining the case series are
set, and then IDRs that meet these criteria are retrieved from the FDA database to build the case series.
Important new product-event associations, called Monitored Adverse Reactions, may be tracked prospectively and
undergo medical review for regulatory action (Kennedy et al., 2000). Other aspects of monitored associations that
may be taken into consideration are the magnitude of the signal, the vulnerability of the population at risk, changes
in risk over time, biologic plausibility, and the preventability of the event (U.S. Food and Drug Administration,
1999, 2003d). Additional analyses of the existing data may be conducted.
Figure F-2. The MedWatch form 3500 that the U.S. Food and Drug Administration uses for spontaneous reporting
of adverse events and product problems related to the use of medications, medical devices and dietary supplements
(U.S. Department of Health and Human Services, 2002).
Based on the review of evidence in support of a product-event association, FDA may disseminate risk management
information and initiate regulatory action (U.S. Food and Drug Administration, 1999). Responses to these initiatives
have resulted in changes in labeling, product names, and packaging; informational letters to health care
professionals; restriction of the use of a product; and, on rare occasions, a withdrawal of a product from the market
(Kennedy et al., 2000). Between 1980 and 1998 there were 13 prescription medications withdrawn from the U.S.
market as a result of adverse events discovered through postmarketing surveillance (Kennedy et al., 2000).
Before broadcasting public warnings, notifying physicians, and changing package labeling, industry (and often
regulatory authorities and the scientific and medical communities) must make a judgment call as to the weight of the
evidence of a product-event association and the subsequent risk to the public.
One case demanding prompt action was that of the contamination of a nonprescription pain reliever in 1982.
McNeil Consumer Products, a subsidiary of Johnson & Johnson, was confronted with a crisis when seven people on
Chicago's west side died mysteriously. Authorities determined that each of the deceased had ingested an ExtraStrength Tylenol® capsule, which had been laced with cyanide, shortly before their death. The news of this incident
traveled quickly and was the cause of a massive, nationwide panic (Kaplan, 2003). The company quickly carried out
a mass withdrawal of 264,000 bottles of Extra-Strength Tylenol® from the market (Wee, 2000).
In contrast, the industry was criticized for moving too slowly in their response to reports of the association of Reye’s
syndrome and nonprescription medications containing salicylates (Mortimer, 1987).
According to Mortimer
(1987), preliminary evidence of the association of Reye’s syndrome in children and Bayer® aspirin was published in
1980. During the years 1981 through 1985, 1003 cases of Reye’s syndrome with 291 deaths were reported to the
Centers for Disease Control. However, a warning label was not implemented until 1986.
The FDA MedWatch system has also identified idiosyncratic and serious adverse events associated with the use of
dietary supplements (U.S. Food and Drug Administration, 2003a). For example, after receiving notice of a young
woman who experienced heart block, a potentially life-threatening condition, FDA investigated the raw materials
used to make the type of dietary supplement she was using. They determined that some materials labeled “plantain”
were contaminated with lanatosides, a constituent of digitalis lanata (U.S. Food and Drug Administration, 1997b).
The effects of digitalis can include nausea, vomiting, headache, confusion, and an abnormal heart rate and rhythm.
The contaminated materials were sold by at least six bulk distributors to 175 or more retail outlets for use as a
dietary supplement or tea (U.S. Food and Drug Administration, 1997a). The use of products that mistakenly contained
digitalis was associated with serious adverse events that were consistent with the effects of digitalis, and biomarkers
of digitalis were detected in blood samples (Letkeman, 2003; U.S. Food and Drug Administration, 1997a). As a result,
FDA issued a warning to product users against using those products that may be contaminated with lanatosides.
Although FDA was commended for this and other similar successes, its overall management of the surveillance
program used to monitor dietary supplements has been criticized by regulating bodies (U.S. Department of Health
and Human Services. Office of Inspector General, 2001; U.S. Congress, 1999) and trade associations (Wallace,
2002). In sum, FDA’s mechanisms for tracking adverse events associated with the use of dietary supplements were
considered inadequate given the popularity of the products and the potential for harm (e.g., evidence of productspecific reports related to adulteration)(Noah, 2002). The Office of the Inspector General concluded that FDA
should make improvements to their surveillance program in order to better protect product users (U.S. Department
of Health and Human Services. Office of Inspector General, 2001). For example, data collection often lacked
information considered important, such as obtaining the name and contact information of the product manufacturer.
Inadequacies such as this often delayed FDA from notifying manufacturers that a serious IDR had been filed about
their product. Moreover, the FDA computer equipment used to monitor dietary supplements did not analyze trends
in adverse events in a way that allowed potential signals of product problems to be identified (U.S. Department of
Health and Human Services. Office of Inspector General, 2001).
F.2
AMERICAN ASSOCIATION
SURVEILLANCE SYSTEM
OF
POISON
CONTROL
CENTERS
TOXIC
EXPOSURE
The Toxic Exposure Surveillance System (TESS) is a comprehensive national poisoning surveillance database
initiated in 1983 and maintained by AAPCC (Litovitz et al., 2002). It includes data on more than 33 million human
poison exposures reported by more than 60 regional poison control centers serving the United States and its
territories.
Poison Control Centers are a 24-hour telephone resource for poison information, clinical toxicology consultation,
and poison prevention education. Reports to TESS originate voluntarily from the general public and from health
professionals. In addition, TESS includes reports on patients who are managed at home or at the site of exposure
and those managed in emergency departments, hospitals, and other health care facilities.
The role of the Poison Control Center is to assess the poisoning situation and to make recommendations for
treatment and site of care (home versus hospital). The purpose of TESS is to identify hazards early, focus
preventative education, guide clinical research, and direct training. TESS data are used for postmarketing
surveillance, for routine review to prevent poisonings, to limit the morbidity and mortality in poisonings that do
occur, to evaluate product safety, and to limit animal testing.
In 2001, U.S. Poison Control Centers transmitted reports of 2,267,979 human poison exposures to TESS (Litovitz et
al., 2002). There is a cyclical nature to the rate of reporting (i.e. there are fewer calls on weekend days and a higher
number of calls are received in the summer months). For example, Figure F-3 (Watson et al., 2003) illustrates the
reporting pattern of total exposures by weekday and Figure F-4 illustrates the reporting pattern for exposure to a
particular product. The volume of product sold and media reports also influence the rate of reporting.
Figure F-3. Human Exposures to poisonous substances in terms of exposures per day as reported to the Poison
Control Centers. Copyright 2003 from American Journal of Emergency Medicine, by Watson et al. Reprinted with
permission of Elsevier.
Figure F-4. Pesticide exposures (excluding rodenticides) by day for all ages and by all routes. Reprinted with
permission from the American Association of Poison Control Centers (2003).
The Poison Control Centers rely on telephone-based reporting of adverse events. Pharmacists, nurses, and others
who have specialized training in clinical toxicology receive incoming calls. Services such as language translation
and transmission to the hearing impaired are offered.
The types of calls received by Poison Control Centers include pediatric ingestions, unintentional exposures to
medications and household products, and environmental hazards. In addition, calls are received about occupational
exposures, bites and stings, suspected food poisoning cases, and intentional overdoses and abuse of drugs and
chemicals. Reports can involve exposure to more than one substance at a time. In 2002, roughly one-half (52%) of
all potential poisonings involved environmental exposures and products not intended for ingestion (Watson et al.,
2003).
Poison Control Centers document each call involving human exposure in an electronic medical record [i.e.,
individual data record (IDR)]. Each IDR includes patient information and exposure information (Table F-1). Each
IDR also contains a narrative description of the adverse event. Follow-up information is obtained for nearly 50% of
IDRs.
AAPCC uses several methods to control the quality of the data collected and processed (Litovitz, 1998). The
computer data entry systems have automated editing, which rejects reports with errors before the data collection
form can be closed (Litovitz, 1998). In addition, thresholds are set for the different data fields to ensure that the data
sent to the national database is specific and complete.
Table F-1. Information used by Poison Control Centers to assist in managing potential poisonings
Patient information:
• Demographic data, i.e., age, gender, weight, pregnancy status
• Patient status (clinical effects)
• Past medical history
• Current location of patient
• Therapy provided and management site
• Duration of symptoms and medical outcome
Exposure information:
• Reason for the exposure, e.g., intentional overdose, therapeutic error
• Substance(s) information, i.e., identification, amount and level of certainty of amount, form (solid/liquid)
• Route of exposure
• Time since exposure
• Geographic location of exposure
Other information:
• Relationship of reporter to patient
An exact product name is obtained, when possible. Otherwise, the product is identified by a generic category.
Poison Control Centers use an existing database of products to facilitate product coding (e.g., POISINDEX® System
database)(Micromedex, 2003b). Compared with data for drugs and household products, there is limited specific
dietary supplement product information. TESS has expanded its generic coding for dietary supplements to facilitate
better characterization of dietary supplements when the specific product is not available in the database.
F.2.3 Dietary supplement exposures reported to the Toxic Exposure Surveillance System
The U.S. Poison Control Centers logged 19,468 IDRs in 2001 (Litovitz et al., 2002) and 22,928 IDRs in 2002
(Watson et al., 2003) of potential poisoning from human exposure to products classified as dietary supplements or
herbal or homeopathic products. For both years:
• ~ 47% were children younger than 6 years
• ~ 17% were 6-19 years
• ~ 35% were older than 19 years
• ~ 0.5% were of unknown age
Of these IDRs, products containing ma huang (single ingredient or mixture) represented 40.7% of exposures in 2001
(Litovitz et al., 2002) and 45% in 2002 (Watson et al., 2003). There were 12 fatalities associated with the use of
dietary supplements or herbal or homeopathic products in 2001 and four fatalities in 2002 (with six of 12 and three
of four in association with ma huang use, respectively)(Litovitz et al., 2002; Watson et al., 2003).
AAPCC has a second, separate classification of products for electrolytes and minerals (route of administration not
specified), which was associated with another 18,631 IDRs of human exposure in 2001 (Litovitz et al., 2002) and
19,287 IDRs in 2002 (Watson et al., 2003). Of these IDRs, seven exposures in 2001 and five exposures in 2002
were known to be fatal. Multimineral dietary supplements represented less than 1.5% of the exposures in this
electrolyte and mineral category (160 IDRs in 2001, 235 IDRs in 2002), with no known fatalities.
A third classification of related products are vitamin preparations, including those with minerals, which accounted
for 53,172 IDRs of non-lethal human exposures in 2001 (Litovitz et al., 2002) and 57,313 IDRs in 2002, which
included two fatalities (Watson et al., 2003).
F.2.4 Signaling from the poison control database
AAPCC, in conjunction with Centers for Disease Control, provides ongoing epidemiologic surveillance using the
TESS data. The primary focus of this effort is to identify outliers in trends of regional and national poisoning
exposure data for deaths, severe outcomes, hospitalization rates, and product categories (Litovitz et al., 2002).
Some hypothetical examples of patterns that may indicate a problem in need of investigation are listed (Litovitz,
1998):
• A clustering of a certain type of exposure in a geographic region may indicate a problem such as product
tampering.
• A significantly different outcome distribution or extent of health care intervention for one product,
compared with other products with similar uses may indicate a problem with a product formulation.
• Excessive pediatric poisonings associated with a product’s new fruit flavored syrup but not with its original
flavored syrup may indicate a problem of product-behavior interaction.
TESS data have been used for general public health surveillance, for support of regulatory actions, to develop
education and research programs, and for disaster preparedness. One key benefit of TESS has been the
identification of product risks, particularly for new products in the market. This information has led to product
reformulations, repackaging, recalls, and bans, thereby minimizing morbidity, mortality, and product liability
(Litovitz et al., 2002). Two such examples were the switch to use of child resistant closures on ethanol-containing
mouth-rinse and the significant changes made to packaging for iron supplements to prevent pediatric morbidity.
F.2.5.1 Example: ethanol-containing mouth-rinse
Using data from AAPCC, Shulman & Wells (1997) reported that there were 2,937 calls related to exposures of
ethanol-containing mouth-rinse in children younger than six years of age to Poison Control Centers in 1994.
Because ingesting 7.2 oz of a mouth-rinse containing 26.9% ethanol by volume (53.8 proof) is potentially lethal to a
child weighing 15-kg or less, Shulman & Wells (1997) advocated that FDA should require visible warning labels
and child-resistant caps for these products. One year later, during a meeting of the Nonprescription Drugs Advisory
Committee Dental Plaque Subcommittee, the concern was again raised that high levels of alcohol in some mouthrinse products necessitated label warnings that these products should not be used by very young children (U.S. Food
and Drug Administration, 1998). In 2001, there were 3,866 exposures of ethanol-containing mouth wash reported
for children under the age of 6 years, but no fatalities for this age group were reported (Litovitz et al., 2002).
Proposed rules for antigingivitis/antiplaque oral rinse products, some of which contain ethanol, include directions
for use explicitly stating that children under six years of age should not use the product (Anonymous, 2003).
F.2.5.2 Example: iron supplements
Children can obtain iron supplements from child-resistant containers that they open themselves, are opened by
another child, or are left open or improperly closed by an adult. Iron supplements were responsible for an average
of two pediatric fatalities per year over an 8-year period from 1983 to 1990, but the yearly rate was increasing. The
deaths due to iron supplements in young children increased from two in 1989 to five in 1990 to eleven in 1991
(American Association of Poison Control Centers, 1992). By that year, iron poisoning had become the single most
frequent cause of pediatric poisoning deaths. The AAPCC petitioned FDA to take action to reduce the health
hazards to young children caused by products containing iron (American Association of Poison Control Centers,
1992).
In 1997, FDA issued final regulations to require warning statements on dietary supplements containing iron in solid
dosage form (U.S. Food and Drug Administration, 1997c). Specifically, the warning label must state: “Accidental
overdose of iron-containing products is a leading cause of fatal poisoning in children under 6. Keep this product
out of reach of children. In case of accidental overdose, call a doctor or poison control center immediately.” FDA
also set requirements of unit-dose packaging for supplements containing 30 mg of iron or more per dosage unit
(U.S. Food and Drug Administration, 1997c). However, on October 17, 2003, FDA withdrew the requirement for
unit-dose packaging in response to a court ruling that FDA does not have the authority under the Dietary
Supplement Health and Education Act to require manufacturers to use unit-dose packaging for poison prevention
purposes (U.S. Food and Drug Administration, 2003b; U.S. Food and Drug Administration, 2003c). The newer
ruling did not affect the requirement for labeling.
The introduction of warnings and unit-dose packaging of potent iron supplements reduced the frequency of severe
pediatric iron overdose. In 2001, among children under the age of 6 years, there were 2,094 IDRs of exposure to
iron supplements (unspecified route), 6,232 IDRs of exposure to adult multivitamin products containing iron, and
16,135 IDRs of exposure to pediatric multivitamin products containing iron, but none of these resulted in a reported
fatality (Litovitz et al., 2002). In 2002, among children under the age of 6 years, there were 2,157 IDRs of exposures
to iron supplements (unspecified route), 6,135 IDRs of exposure to adult multivitamin products containing iron and
18,368 IDRS of exposure to pediatric multivitamin products containing iron, with none of these resulting in a
reported fatality (Watson et al., 2003).
F.3 POSTMARKETING SURVEILLANCE BY THE WORLD HEALTH ORGANIZATION
WHO developed the Collaborating Center for International Drug Monitoring (CCIDM) to identify rare adverse
events to medicinal products not previously detected in clinical trials or documented in the scientific literature
(Olsson, 1998). Since 1978, the day-to-day maintenance and operations of CCIDM has been centered at the
Uppsala Monitoring Center (UMC) in Sweden (Edwards & Olsson, 2002).
In 2003, there were 71 member-nations in the WHO Drug Monitoring Program. Each participating country
establishes a National Pharmacovigilance Center (NPC), which acts as a channel through which IDRs collected in
the respective country are sent to CCIDM (Edwards & Olsson, 2002). Reports from the United States are obtained
from FDA. Altogether, NPCs provide between 150,000 and 200,000 IDRs annually to WHO (Lindquist & Edwards,
2001). The CCIDM cumulative database now comprises over two million IDRs of adverse events for medicinal
products (Lindquist & Edwards, 2001).
F.3.1 Definitions
The definitions of the terms “adverse reaction” and “signal” used by WHO were agreed upon by consensus of more
than 30 countries participating in the program (Bate et al., 2002):
• WHO Technical Report No. 498, produced in 1972, described an adverse reaction as: “A response to a
drug which is noxious and unintended, and which occurs at doses normally used in man for the
prophylaxis, diagnosis, or therapy of disease, or for the modification of physiological function” (World
Health Organization, 2003c).
• A signal was described as: “Reported information on a possible causal relationship between an adverse
event and a drug, the relationship being unknown or incompletely documented previously. Usually more
than one report is required to generate a signal, depending on the seriousness of the event and the quality
of the information” (Edwards et al., 2002; Farah et al., 2000).
In general, WHO accepts only IDRs derived from reports submitted by health care professionals (rather than by
product users or other reporters). Therefore, the stimulus for an IDR would be a patient-physician interaction during
which there is a complaint of illness or an admission into a health care facility. If the health care professional
suspects an association with a product, and is motivated to submit it, this information is sent to the NPC in the
country of origin.
WHO expects that the reporting health care professional will obtain pertinent medical history in order to account for
all potential sources of the problem, including documentation of concomitant exposure to medications and dietary
supplements.
A secure online software management tool, called Vigibase, is now available for reporting physicians and NPCs
who are in need of a modern system for the management of IDRs. This software was produced in collaboration with
Swissmedic (Bern, Switzerland) and is accessible via the Internet so that there is no need for special servers or
licenses (World Health Organization, 2003a).
Figure F-5. The flow of adverse event reports from the source to the World Health Organization for use in signal
generation.
A flow chart of IDRs from their source to WHO is presented in Figure F-5. An audit trail process was established to
register any modifications made to an IDR. The details of changes are stored in a history table. These procedures
enable previous versions of the IDR to be retrieved with no loss of information (Lindquist & Edwards, 2001).
F.3.3.1 Data entry and coding of individual data records
Medicinal products are coded using the WHO Drug Dictionary, and adverse events are coded using WHO Adverse
Reaction Terminology (WHO-ART). An attempt has been made to create a browsable and searchable system of
adverse event terminologies with a unified format through the development of the International Classification of
Disease-WHO-ART Project. WHO-ART has been translated into French, Italian, German, and Portuguese (World
Health Organization, 2003b). Because one of the major providers of IDRs is the United States, which uses the
Medical Dictionary for Regulatory Authorities (MedDRA) for coding (International Conference on Harmonisation,
1998)17, UMC is introducing certain MedDRA terms into WHO-ART. New WHO-ART terms are added every three
months. Since July 2001, 260 new terms have been added (Kennedy et al., 2000).
In 1985, the CCIDM developed a new Critical Terms List. This is a subset of adverse event terms indicative of
serious diseases that were regarded as significantly important enough to monitor. The frequency that adverse event
terms were reported to WHO was not taken into consideration in selecting terms for inclusion on the list.
17
Farah et al. (2000) raised a number of issues related to the accuracy of international IDRs for herbal products. In
particular, data regarding herbal products are difficult to classify and to interpret, and the identification of the herb
used may be imprecise or inaccurate because there are numerous lay names for such products. Accurate
identification of constituents is further complicated by the use of different names for the same product in different
countries. An example of this problem is Artemisia absinthium, an herb that has at least 11 different common names
(wormwood, absinthium, absinth, absinthe, madderwort, wermuth, mugwort, mingwort, warmot, magenkraut and
herba absinthii), seven of which bear no resemblance to its botanical name (World Health Organization, 2003a).
Additional names may also be used that identify the parts of the plant that are used for medicinal purposes.
Moreover, genus names typically appear in the IDR, yet there may be hundreds of species in the corresponding
genus. An example of a lack of specificity due to species variability is aloe, the genus name, for which there are 330
different species (Farah et al., 2000).
Because of the potential for misclassification and confusion, Farah et al. (2000) acknowledged the need for an
internationally standardized and accepted classification source that listed all the accepted botanical names of
medicinal herbs. WHO is now standardizing their classification of all herbal products that have been mentioned in
the IDRs collected by CCIDM. For this purpose, they developed the Draft Herbal Anatomical Therapeutic
Chemical (ATC) Index for coding herbal substances (WHO Collaborating Centre for International Drug Monitoring,
2002a). This index provides information on standardized plant names for genus, species, plant part, and extracts
used. The Guidelines for Herbal ATC Classification, which was modeled on the ATC classification system for
drugs, were also developed to assist in coding herbal remedies (WHO Collaborating Centre for International Drug
Monitoring, 2002b). This experience indicates that an internationally standardized and accepted classification
source that lists the botanical names of medicinal herbs and their active components, along with derivative
substances used as ingredients in dietary supplements, will be necessary if herbal products are to be reliably coded in
reporter-submitted IDRs.
Another problem identified by Farah et al. (2000) was that of classifying mixtures. Although some single source
herbal products are used, many herbal products are complex mixtures. In addition, among herbal products with the
same formulation, the actual composition from batch to batch or between brands can vary widely.
F.3.4 The Collaborating Center for International Drug Monitoring database of individual data records
The IDR is sent by NPCs to UMC for inclusion in the international CCIDM database. Because NPCs only release
IDRs periodically to UMC, the transmission of an IDR can be delayed for several months. Estimates of the time
delay between the receipt of an adverse event and its transmission to UMC through a Freedom of Information
pathway (after confidential information has been redacted) range from 6-18 months or more (Clark, 2003; U.S. Food
and Drug Administration, 2003e). Then, once an IDR is received by UMC, an additional 18 months may pass for
WHO to process it for inclusion in the CCIDM Database (U.S. Food and Drug Administration, 2003e). Hence, the
CCIDM is not an early warning system.
Since 1983, IDRs in CCIDM have been organized by the date the adverse event occurred and not by the date the
event was reported. The CCIDM system runs on a server using the Uniplexed Information and Computing Service
operating system (UNIX) and uses “structured query language” as the coding to retrieve information and request
various actions by the relational database management system.
NPCs have unrestricted access to all information in the CCIDM database. Non-member parties, such as
pharmaceutical companies, only have access to the IDRs that were contributed by the approximately 30 countries
that have consented to the release of such data. UMC developed a procedure for release of data to third-parties with
most participating countries agreeing to the general release of information for proposed searches by any inquirer
provided it is accompanied by a caveat specifying how the data can and cannot be used. The remaining member
countries prefer to decide on a case-by-case basis whether or not they will release data.
The CCIDM database can be accessed to retrieve several categories of information, if the information was provided
in the IDR (Lindquist & Edwards, 2001):
• Administrative information (i.e., IDR number, dates, classification)
• Patient (i.e., identifier for the patient, age, gender, patient outcome)
• Medical background (i.e., previous illness, predisposing conditions)
• Death (i.e., cause of death, postmortem information)
• Related IDRs (i.e., information on a related IDR, e.g., a mother’s IDR is related to her new-born infant’s
IDR, and a link is established between these IDRs)
• Notifier (i.e., identifier for the reporter)
• Drug (i.e., medication information)
• Event (i.e., information on the adverse event)
• Link (i.e., causality assessment and dechallenge and rechallenge information)
F.3.5 Signaling from the Collaborating Center for International Drug Monitoring database
Once every three months, data accumulated in the CCIDM database are statistically analyzed using the Bayesian
Confidence Propagation Neural Network procedure, to generate a list of possible signals. The aim of the signaling
process is to identify previously undetected adverse events associated with drugs and other medicinal products
(Lindquist et al., 2000).
The list of potential signals is sent to an international panel of about 30 expert consultants for signal evaluation
(Figure F-5). The experts fill out a worksheet (Figure F-6)(World Health Organization, 2003d) and generate a
“Signal Report” of each adverse event-drug relationship, which they send to UMC and all of the NPCs. Signals
identified by this method are copied to a permanent log (history) file and are followed for 24 months. Outcomes
(i.e., whether the patient was reported to have recovered or died) are documented.
F.3.5.1 Bayesian Confidence Propagation Neural Network
WHO based their signaling analysis on Bayes Law (or Bayes Theorem), which describes a procedure for performing
inference on an unknown quantity or parameter (i.e., the probability of a product-adverse event pairing) using prior
knowledge of that parameter and information derived from data. In mathematical terms, it is represented as:
Posterior Density = (Sampling Density x Prior Density) ÷ (Marginal Density of data)
And, more simply, as
Posterior Density α (Sampling Density x Prior Density)
Specifically, WHO uses the Bayesian Confidence Propagation Neural Network procedure. According to Lindquist
et al. (2000), the strength of the association of products and events, and how these product-event combinations
change with the addition of new data (new IDRs) is analyzed by a measure of disproportionality, termed the
Information Component (IC):
IC = log
P
PP
xy
2
x
y
Where,
The probability of a product being listed in an IDR
Px =
The probability of an event being listed in an IDR
Py =
Pxy =
The probability that a specific product-event combination is listed in an IDR
Figure F-6. The Uppsala Monitoring Center signal detection worksheet used by the World Health Organization to
evaluate potential signals generated from their database (World Health Organization, 2003).
Thus, the IC value is based on four types of data:
• The total number of IDRs
• The number of IDRs with product X
• The number of IDRs with event Y
• The number of IDRs with the specific product-event combination
As the expected probability of product event-combinations (Pxy) increases over the probability predicted by an
assumption of independence between the product and event (i.e., PxPy), the IC value increases. Lindquist et al.
(2000) explained that high positive IC values indicate that a particular product-event combination is currently being
reported more frequently than statistically expected from the history of IDRs in the database. If the IC value
increases over time and the calculated confidence interval narrows, this suggests to Lindquist et al. (2000) that there
is the likelihood of a positive quantitative association between the product and the event of interest, and this could
be signaling an emerging problem.
A major issue in the consideration of the IC is the multiplicity problem, which involves setting a Type 1 error
(concluding there is an association when there is none) when hundreds (or thousands) of tests are conducted. The
multiplicity problem relates also to a concept in the statistical literature that has received considerable attention as of
late. Rather than trying to maintain an overall Type 1 error, researchers in many fields, particularly bioinformatics,
are instead attempting to determine ways of controlling the “false discovery rate” or proportion of signals that are
falsely positive. This means that emphasis is shifted away from preventing Type 1 errors to controlling the number
of Type 1 errors that are made in proportion to the number of true associations that are found.
A false positive signal may have an alternative explanation for the event and not signal a causal association between
the product and event (U.S. Food and Drug Administration, 2003e). One example of the generation of a false
positive signal from the CCIDM was Claritin®, a non-arrhythmogenic antihistamine containing loratadine
(Lindquist & Edwards, 1997b). The signal of loratadine and cardiac rhythm disorders was generated primarily from
U.S. IDRs (Lindquist & Edwards, 1997a; U.S. Food and Drug Administration, 2003e). According to comments
made by Faich to FDA (U.S. Food and Drug Administration, 2003e), key details were missing from the IDRs, which
contributed to the generation of a false positive signal. The most important missing detail was evidence of preexistent arrhythmias in many of these cases. The presence of an arrhythmic condition could have influenced the use
of the product (i.e., patients with the highest risk of having a cardiac rhythm disorder may have been assigned the
safest known drug)(Himmel et al., 1997). This kind of differential use by indication produces a disproportion as a
result of confounding bias, rather than as a result of a causal association. Hence, it is important to remember that
signals represent hypotheses that may require additional testing prior to acceptance, and that manufacturers,
clinicians, epidemiologists, and other interested parties may need to gather additional data through other
postmarketing surveillance techniques.
One potential advantage of using the large pooled data in the CCIDM database is having sufficient statistical power
to analyze signals by important subsets, such as gender, age, and concomitant exposure to drugs or other conditions
(Edwards et al., 2002). Another advantage is the ability to conduct analyses of inter-product signaling. However,
because of the long delay from the doctor’s initial observation of the patient to the inclusion and analyses by
CCIDM, the WHO system may not be as useful for detecting some kinds of warning signals as are databases that
receive more timely information.
F.3.5.2 Dietary supplement exposures reported to the World Health Organization
CCIDM contains a number of IDRs relating to use of dietary supplements. Farah et al. (2000) reported that herbal
products were associated with approximately 8,985 (0.5%) of more than two million IDRs in the CCIDM database
from reports collected in the period 1968-1997. These herbal-related IDRs contained 16,079 adverse event terms, an
average of 1.8 terms per IDR. Table F-2 presents the reporting countries contributing the greatest number of IDRs
associated with herbal products. For adults aged 20 years and older, IDRs of adverse events by women for herbal
products were more frequent than for men (Farah et al., 2000).
Table F-2. Twelve countries contributing the most number of individual data records for herbal products to the
World Health Organization Collaborating Center for International Drug Monitoring
Individual data records for herbal products
Country
1968 -1997 (%)b
(year of entry)a
Germany (1968)
20.0
France (1986)
16.5
United States (1968)
11.9
United Kingdom (1968)
11.1
Australia (1968)
9.0
Spain (1984)
6.6
Canada (1968)
6.2
Sweden (1968)
4.5
Netherlands (1968)
1.6
Denmark (1968)
1.4
Japan (1972)
1.4
Ireland (1968)
1.1
a
Data from Edwards, I. R. & Olsson, S. (2002) WHO Programme—Global monitoring. In: Pharmacovigilance.
(Mann, R. D. & Andrews, E. B., eds.). John Wiley & Sons, Ltd, Chichester, West Sussex, England. pp. 169-182;
year of entry to the World Health Organization Collaborating Center for International Drug Monitoring.
b
Data from Farah, H. H., Edwards, R., Lindquist, M., Leon, C. & Shaw, D. (2000) International monitoring of
adverse health effects associated with herbal medicines. Pharmacoepidemiology and Drug Safety 9:105-112;
these countries contributed a total of 91.3% of the total 8985 reports.
More recent data have calculated at least 11,716 IDRs associated with herbal products in the CCIDM database that
have accumulated since 1968. Among these, the ten most commonly reported events were ranked (World Health
Organization, 2003a):
• Pruritus (n=912)
• Urticaria (n=688)
• Rash (n=670)
• Rash erythematous (n=668)
• Nausea (n=515)
• Vomiting (n=441)
• Diarrhea (n=414)
• Fever (n=397)
• Abdominal pain (n=350)
• Dyspnea (n=314)
The frequencies of these events were similar to the most frequently reported adverse events in the entire CCIDM
database, in rank order from the most frequent: rash, pruritus, urticaria, fever, nausea, headache, and vomiting
(Edwards, 2001). These data indicate that allergic-type responses (pruritus, urticaria, rash) were the most common
adverse events noted by health care professionals following the use of herbal products, while the second most
common adverse event grouping was gastrointestinal distress (nausea, vomiting, diarrhea).
Among all IDRs associated with herbal products (n=8985) between 1968-1997, the most frequently reported event
that was associated with a single herbal source that was categorized as critical by CCIDM was anaphylactic
reaction/shock (n=67)(Farah et al., 2000). More recently, the top ten terms from all IDRs that were associated with
a single source herbal product or herbal mixture (n=11,716) that were categorized as critical by CCIDM were ranked
(World Health Organization, 2003a):
• Face edema (n=240)
• Hepatitis (n=216)
• Angioedema (n=186)
• Thrombocytopenia (n=162)
• Hypertension (n=140)
• Chest pain (n=129)
• Convulsions (n=129)
•
•
•
Purpura (n=127)
Death (n=124)
Dermatitis (n=124)
Through the second quarter of 2002, the most frequently reported herbal products that contained eight or less
ingredients, and for which the herbal substance was specified in the CCIDM Database, were ranked by medicinal
plant (World Health Organization, 2002):
• Senna alexandra (n=616)
• Ginkgo biloba (n=591)
• Atropa belladonna (n=541)
• Glycine max (n=448)
• Hypericum perforatum (n=440)
• Chrysanthemum cinerariaefolium (n=339)
• Valeriana officinalis (n=329)
• Mentha x piperita (n=307)
• Carthamus tinctorius (n=304)
F.3.6 Signaling from other information sources
Some signals of new product-event combinations identified by CCIDM have been confirmed by prospective
randomized controlled clinical trials. An example is renal tubular disorders and/or necrosis (acute renal failure)
following the use of ketorolac (an injectable non-steroidal anti-inflammatory drug), which was identified by UMC in
the first quarter of 1993. According to Lindquist et al. (2000), these events were not listed as potential adverse
reactions to the drug in the 2000 online versions of Martindale or the Physicians’ Desk Reference (Lindquist et al.,
2000). Subsequently, Forrest et al. (2002) reported 3 of 5634 patients who received ketorolac in a prospective
randomized trial developed acute renal failure. Another 7 of 5611 patients receiving the alternate drugs ketoprofen
or diclofenac also developed acute renal failure. There were no significant differences in the occurrence of renal
failure between treatment groups. Miwa et al. (1997) cautioned that the adverse events observed in clinical trials do
not necessarily reflect the risk of a larger population. Martindale now lists acute renal failure as a potential major
adverse reaction to ketorolac if this drug is used for more than five days duration (Micromedex, 2003a). Similarly,
the online Physicians’ Desk Reference (PDR Health, 2003) now lists kidney failure among the rare side effects that
may occur following the use of ketorolac.
Beginning in 1982, WHO published the Adverse Reaction Newsletter, a yearly review of signals generated by
CCIDM. In 2000, this publication was incorporated in its entirety into the WHO Pharmaceuticals Newsletter. It is
up to the discretion of the NPCs as to whether or not they act on the signals generated by the CCIDM.
Another WHO publication, the Uppsala Reports, communicates news and information about CCIDM to the NPCs
and is accessible on the Internet. An email server (Vigimed) provides for rapid distribution of drug safety alerts for
all NPCs and is also used by these NPCs to submit requests for information to UMC.
International Training Courses are held by UMC to train interested professionals in adverse events and adverse
event monitoring. Additionally, UMC staff members will visit countries to assist with the development and
implementation of adverse event monitoring programs (Olsson, 1998).
F.4 LITERATURE CITATIONS FOR EXISITING SURVEILLANCE PROGRAMS FOR DIETARY
SUPPLEMENTS
Anonymous (2003) 21 CFR Part 356. Oral health care drug products for over-the-counter human use;
Antigingivitis/antiplaque drug products; Establishment of a monograph. Fed. Reg. 68: 32232-32287.
American Association of Poison Control Centers, I. (1992) American Association of Poison Control Centers, Inc.
submission to FDA Docket No. 91P-0186/SUP1. Re: Accidental pediatric ingestion of iron-containing drug
products. FDA Docket No. 91P-0186
Bate, A., Lindquist, M., Edwards, I. R. & Orre, R. (2002) A data mining approach for signal detection and analysis.
Drug Saf. 25: 393-397.
Bright, R. A. & Nelson, R. C. (2002) Automated support for pharmacovigilance: a proposed system.
Raton, FL. pp. 183-193.
Clark, J. A. (2003) Signaling methods and signaling program design. (UnPub)
DuMouchel, W. (1999) Bayesian data mining in large frequency tables, with an application to the FDA spontaneous
reporting system. Am. Stat. 53: 177-190.
Edwards, I. R. (2001) The management of adverse drug reactions: from diagnosis to signal. Therapie. 56: 727-733.
Edwards, I. R., Lindquist, M., Bate, A. & Orre, R. (2002) Data mining. In: Pharmacovigilance. (Mann, R. D. &
Andrews, E. B., eds.). John Wiley & Sons, Ltd, Chichester, West Sussex, England, pp. 291-300.
Edwards, I. R. & Olsson, S. (2002) WHO Programme—Global monitoring. In: Pharmacovigilance. (Mann, R. D. &
Andrews, E. B., eds.). John Wiley & Sons, Ltd, Chichester, West Sussex, England, pp. 169-182.
Faich, G. (2003) Comments on FDA's pharmacovigilance and pharmacoepidemiologic assessment concepts.
Available at http://www.fda.gov/cder/meeting/RMtranscript3.doc. Accessed 7-18-2003.
Farah, H. H., Edwards, R., Lindquist, M., Leon, C. & Shaw, D. (2000) International monitoring of adverse health
effects associated with herbal medicines. Pharmacoepidemiol. Drug Saf. 9: 105-112.
Figueiras, A., Tato, F., Fontainas, J. & Gestal-Otero, J. J. (1999) Influence of physicians' attitudes on reporting
adverse drug events: a case-control study. Med. Care 37: 809-814.
Forrest, J. B., Camu, F., Greer, I. A., Kehlet, H., Abdalla, M., Bonnet, F., Ebrahim, S., Escolar, G., Jage, J., Pocock,
S., Velo, G., Langman, M. J., Bianchi, P. G., Samama, M. M. & Heitlinger, E. (2002) Ketorolac, diclofenac, and
ketoprofen are equally safe for pain relief after major surgery. Br. J. Anaesth. 88: 227-233.
Goldman, S. A. (1998) Limitations and strengths of spontaneous reports data. Clin. Ther. 20 Suppl C: C40-C44.
Goldman, S. A. (2003). How can the quality of spontaneously reported case reports be improved? [Presented at FDA
Risk Management Public Workshop]. Available at http://www.fda.gov/cder/meeting/RMtranscript3.doc. Accessed
7-17-2003.
Hasford, J., Goettler, M., Munter, K. H. & Muller-Oerlinghausen, B. (2002) Physicians' knowledge and attitudes
regarding the spontaneous reporting system for adverse drug reactions. J. Clin. Epidemiol. 55: 945-950.
Himmel, M. H., Honig, P. K. & Worobec, A. S. (1997) Dangers of non-sedating antihistamines. Lancet 350: 69-70.
Accessed 9-29-2003.
Juergens, J. P., Szeinbach, S. L., Janssen, M. J., Brown, T. R. & Garner, D. D. (1992) An evaluation of interventions
designed to stimulate physician reporting of adverse drug events. Top. Hosp. Pharm. Manage. 12: 12-18.
Kaplan, T. (2003) The Tylenol crisis: how effective public relations saved Johnson & Johnson. Available at
http://www.personal.psu.edu/users/w/x/wxk116/tylenol/crisis.html. Accessed 7-9-2003.
Kennedy, D. L., Goldman, S. A. & Lillie, R. B. (2000) Spontaneous reporting in the United States. In:
Pharmacoepidemiology. 3rd ed. John Wiley & Sons, Ltd., United Kingdom, pp. 151-174.
Koch-Weser, J., Sidel, V. W., Sweet, R. H., Kanarek, P. & Eaton, A. E. (1969) Factors determining physician
reporting of adverse drug reactions. Comparison of 2000 spontaneous reports with surveillance studies at the
Massachusetts General Hospital. N. Engl. J. Med. 280: 20-26.
Letkeman, C. (2003) (Third) Declaration of Caroline Letkeman. Available at http://xenu.ca/ex/cl-declaration3.html. Accessed 5-20-2003.
Lindquist, M. & Edwards, I. R. (1997a) Dangers of non-sedating antihistamines. [Author's reply]. Lancet 350: 6967.
Lindquist, M. & Edwards, I. R. (1997b) Risks of non-sedating antihistamines. Lancet 349: 1322.
Lindquist, M. & Edwards, I. R. (2001) The WHO Programme for International Drug Monitoring, its database, and
the technical support of the Uppsala Monitoring Center. J. Rheumatol. 28: 1180-1187.
Lindquist, M., Stahl, M., Bate, A., Edwards, I. R. & Meyboom, R. H. (2000) A retrospective evaluation of a data
mining approach to aid finding new adverse drug reaction signals in the WHO international database. Drug Saf. 23:
533-542.
Litovitz, T. (1998) The TESS database. Use in product safety assessment. Drug Saf. 18: 9-19.
Litovitz, T. L., Klein-Schwartz, W., Rodgers, G. C., Jr., Cobaugh, D. J., Youniss, J., Omslaer, J. C., May, M. E.,
Woolf, A. D. & Benson, B. E. (2002) 2001 Annual report of the American Association of Poison Control Centers
Toxic Exposure Surveillance System. Am. J. Emerg. Med. 20: 391-452.
Micromedex. (2003a) DISEASEDEX Emergency Medicine. Available at
http://www.micromedex.com/products/diseasedexemergency/samples/dsdxemergmigraine.pdf. Accessed 9-292003.
Micromedex. (2003b) Poisindex® System. Available at http://www.micromedex.com/products/poisindex/.
Accessed 7-8-2003.
Miwa, L. J., Jones, J. K., Pathiyal, A. & Hatoum, H. (1997) Value of epidemiologic studies in determining the true
incidence of adverse events. The nonsteroidal anti-inflammatory drug story. Arch. Intern. Med. 157: 2129-2136.
Mortimer, Jr. E. A. (1987) Reye's syndrome, salicylates, epidemiology and public health policy [Editorial]. JAMA
257: 1941.
Noah, B. A. (2002) Adverse drug reactions: Harnessing experiential data to promote patient welfare. Catholic
University Law Review 49: 449-504.
Olsson, S. (1998) The role of the WHO programme on International Drug Monitoring in coordinating worldwide
drug safety efforts. Drug Saf. 19: 1-10.
PDR Health. (2003) Ketorolac tromethamine (Torodal). Available at
http://www.pdrhealth.com/drug_info/rxdrugprofiles/drugs/tor1453.shtml. Accessed 10-10-2003.
Piazza-Hepp, T. D. & Kennedy, D. L. (1995) Reporting of adverse events to MedWatch. Am. J. Health Syst. Pharm.
52: 1436-1439.
Scott, H. D., Thacher, A., Rosenbaum, S. E., Waters, W. J., Jr. & Green, M. (1988) Adverse drug reaction reporting
systems: the United Kingdom and the United States. R. I. Med. J. 71: 179-184.
Shulman, J. D. & Wells, L. M. (1997) Acute ethanol toxicity from ingesting mouthwash in children younger than 6years of age. Pediatr. Dent. 19: 404-408.
Tollefson, L. (1988) Monitoring adverse reactions to food additives in the U.S. Food and Drug Administration.
Regul. Toxicol. Pharmacol. 8: 438-446.
U.S. Congress (1999) How accurate is the FDA's monitoring of supplements like ephedra? Serial No. 106-60:
Committee on Government Reform.
U.S. Department of Health and Human Services. Office of Inspector General (2001) Adverse event reporting for
dietary supplements. Report No. OEI-01-00-00180. Office of Evaluations and Inspections, Boston, pp. 1-110.
U.S. Food and Drug Administration. (6-12-1997a) FDA warns consumers against dietary supplement products that
may contain digitalis mislabeled as "plantain." Available at
U.S. Food and Drug Administration. (6-12-1997b) FDA warns consumers against dietary supplement products that
may contain digitalis mislabeled as "plantain" [Press Release]. Available at
U.S. Food and Drug Administration (1997c) Iron-containing supplements and drugs: Label warning statements and
unit-dose packaging requirements; Final Rule. Fed. Reg. 62: 2237-2250.
U.S. Food and Drug Administration. (12-2-1998) NonPrescription Drugs Advisory Committee. Dental Plaque
Subcommittee. Meeting transcript from December 2, 1998 Meeting. Available at
http://www.fda.gov/ohrms/dockets/ac/98/transcpt/3481t1.rtf. Accessed 7-23-2003.
U.S. Food and Drug Administration. (1999) Part 3: How does FDA conduct postmarketing surveillance and risk
assessment? Available at http://www.fda.gov/oc/tfrm/Part3.html. Accessed 4-8-2003.
U.S. Food and Drug Administration. (2002) Attention. Available at http://www.cfsan.fda.gov/~dms/aems.html.
Accessed 9-29-2003.
U.S. Food and Drug Administration (2003a) Current good manufacturing practice in manufacturing, packing, or
U.S. Food and Drug Administration. (2003b) Guidance for Industry. Iron-containing supplements and drugs: Label
warning statements. Small entity compliance guide. Available at
http://www.fda.gov/OHRMS/DOCKETS/98fr/97D-0443-GDL0001.PDF. Accessed 11-3-2003b.
U.S. Food and Drug Administration (2003c) Iron-containing supplements and drugs; Label warning statements and
unit-dose packaging requirements; Removal of regulations for unit-dose packaging requirements for dietary
supplements and drugs. Fed. Reg. 68: 59714-59715.
U.S. Food and Drug Administration. (2003d) Risk assessment of observational data: Good pharmacovigilance
practices and pharmacoepidemiologic assessment [Concept paper for comment purposes only]. Available at
http://www.fda.gov/cder/meeting/groupIIIfinal.pdf. Accessed 4-8-2003.
U.S. Food and Drug Administration. (2003e) Risk Management Public Workshop. Pharmacovigilance Practices and
U.S. Food and Drug Administration. (2003f) What is a product problem? Available at
http://www.fda.gov/medwatch/report/DESK/prodprob.htm. Accessed 9-29-2003.
U.S. Food and Drug Administration. (2003g) What is a serious adverse event? Available at
http://www.fda.gov/medwatch/report/DESK/advevnt.htm. Accessed 9-29-2003.
Wallace, P. (2002) Supplement groups question adverse events reporting plan. Food Chem. News 44: 12-13.
Watson, W. A., Litovitz, T. L., Rodgers, G. C., Jr., Klein-Schwartz, W., Youniss, J., Rose, S. R., Borys, D. & May,
M. E. (2003) 2002 annual report of the American association of poison control centers toxic exposure surveillance
system. Am. J. Emerg. Med. 21: 353-421.
Wee, H. (2000) Calculating the risks of a recall. Protecting consumers' safety and companies' brands. Available at
http://abcnews.go.com/sections/business/DailyNews/productrecalls_000831.html. Accessed 7-7-2003.
WHO Collaborating Centre for International Drug Monitoring. (2002a) Draft herbal ATC index. (Farah, M.H., ed.).
the Uppsala Monitoring Center, Uppsala, Sweden.
WHO Collaborating Centre for International Drug Monitoring. (2002b) Draft guidelines for herbal classification.
(Farah, M.H., ed.). the Uppsala Monitoring Center, Uppsala, Sweden.
World Health Organization. (2002) Most reported herbals in the WHO database - both single and multiple ingredient
drugs. Available at http://www.who-umc.org. Accessed 7-29-2003.
World Health Organization. (2003a) New software solution for pharmacovigilance centres. Available at
http://www.who-umc.org/projects.html. Accessed 4-24-2003.
World Health Organization. (2003b) Products and Services. Available at http://www.who-umc.org/prodserv.html.
Accessed 5-20-2003.
World Health Organization. (2003c) Side effect-adverse reaction. Available at http://www.whoumc.org/defs.html#SeriousAR. Accessed 9-29-2003.
World Health Organization. (2003d) UMC signal detection worksheet. Available at http://www.whoumc.org/pdfs/signalsheet.pdf. Accessed 7-29-2003.
APPENDIX G
ANCILLARY ISSUES
Postmarketing surveillance is a scientifically and technologically complex undertaking. In putting forward
recommendations for a dietary supplement surveillance program, the Life Sciences Research Office (LSRO) has, by
necessity, kept this publication relatively brief in depth and scope. Hence, LSRO refers the reader to additional
resources and identifies further research that may be of benefit to propel advances and applicability of such systems.
LSRO cites several sources of information and services as examples, however, this list is not comprehensive. LSRO
makes no recommendation or endorsement of any commercial product.
G.1 RESOURCES
G.1.1 Resources for recognition and reporting of adverse events in individual product users
G.1.1.1 Electronic filing
Computer-assisted telephone interviewing and facsimile-based data transmission may be useful for data collection.
In addition, the next generation of surveillance systems includes web-based applications and Internet-accessible data
sources. These systems are pioneering web-based communication and connectivity for public health (National
Center for Infectious Diseases, 2003). There are several benefits of electronic submissions (U.S. Food and Drug
Administration, 2001a):
• An electronic acknowledgement of receipt of the report can be automatically transmitted to the reporter.
• Electronic data can be collected via the Internet and data can be rapidly entered into the database for
analysis.
• Data transcription errors are eliminated.
• The cost of data processing is reduced.
• Data transmission between industry, regulatory authorities and others can be standardized.
Guidance documents were developed by the U.S. Food and Drug Administration (FDA) for use in electronic
submissions (Kennedy et al., 2000; U.S. Food and Drug Administration, 2001a, 2001b). These documents provide
some information about technological capabilities and formats that are relevant for electronic submission of
individual data records (IDRs) in a system for dietary supplements.
Interactive Internet data collection and analysis is now being used as a supplemental resource for health care because
it permits secure messaging for support, advice, and education. These tools may serve as models for spontaneous
reporting in surveillance programs as well as for registries and other forms of active postmarket data collection.
Some examples of Internet data collection include the “Daily Diary” pages of the websites: MyAsthma.com (2003)
and Mycysticfibrosis.com (2003).
G.1.2 Resources for creation of individual data records
G.1.2.1 Standardized data collection instruments
The data collection instrument should be designed with input from professionals with expertise in clinically
managing adverse events and those with expertise in postmarketing surveillance programs. Several questionnaires
are in use that could serve as a starting point to create a tailored form for use with dietary supplements. A copy of
the FDA MedWatch data collection form that is used for adverse events related to dietary supplements is provided in
Appendix F (see Figure F-2)(U.S. Department of Health and Human Services, 2002). In addition, the National Food
Processors Association developed a series of forms to collect complaints of illness that include questions designed to
elicit information about possible product defects, such as “Was there anything unusual about the taste or odor of the
product when eaten?” (Claims Division, 1997). For reference, Christian Benichou included examples of forms for
drug pharmacovigilance in the Appendix of his text, Adverse Drug Reactions: A Practical Guide to Diagnosis and
Management (John Wiley & sons, 1994). Recently, Health Canada developed a template form as a model to use in
reporting adverse events associated with the use of natural health products (Health Canada, 2004).
G.1.2.2 Databases of dietary supplement products
To be of value, databases of dietary supplement products need to have transparent documentation, quality, and
source information (Dwyer et al., 2003).
In 1999, the National Center for Health Statistics of the Centers for Disease Control and Prevention developed a
database of dietary supplements based on label information. They gained additional support from the National
Institutes of Health (NIH) Office of Dietary Supplements (ODS) in 2001. The database contains information
transcribed from labels of products reportedly ingested by participants of the National Health and Nutrition
Examination Surveys (NHANES) between 1988 and1994. Participants were asked specifically about their use of
vitamin and mineral supplements; information collected on other supplements was obtained sporadically (Ervin et
al., 1999). More than 2,400 prescription and nonprescription dietary supplement products were included from the
third NHANES survey (McDowell & Ervin, 2003). Further updated, the database currently contains the following
information on nearly 4000 products (Dwyer et al., 2003):
• Product name, labeled ingredients from the Supplement Facts box (but not those ingredients that are listed
beneath the Supplement Facts box), amounts of ingredients, serving size and form, manufacturer or
distributor, and suggested dosage.
• The date the product entered the database, source of information, type of formulation (e.g., pediatric,
prenatal), classification of each ingredient (e.g., vitamin, mineral, amino acid), and dietary supplement
category (e.g., multivitamin, single botanical).
When labels were not available, the information may have been obtained from other sources such as company
catalogs, the Internet, directly from the company, or the Physicians’ Desk Reference. Health claims, certifications,
and descriptions that were not part of the name were not recorded. To date, no verification of ingredients has been
undertaken. At present, this database is not available to the public.
A database of analytical composition of dietary supplements is in the development stages of construction under an
interagency agreement between the U.S. Department of Agriculture’s (USDA) Agricultural Research Service and
the ODS of NIH. Initially, the most commonly consumed supplements will be included, if methods of analysis exist
for the components of interest (U.S. Department of Agriculture, 2003). Specifically, the first data will be obtained
for vitamin and mineral preparations. In parallel with this process is the development of standard reference
materials for use in analyses of botanicals and other ingredients in dietary supplements (Dwyer et al., 2003).
The University of California, San Diego (UCSD) created a dietary supplement product database as part of the
Women's Healthy Eating and Living Study. Dietary supplement use was collected by 24-hr recall interview, and
copies were made of dietary supplement labels. The database was created using the DataEase software program
(DataEase International Inc, Trumbaull, CT) and is currently being upgraded to NetPlus, a web-based system
(NetPlus, Oklahoma City, OK). Information on more than 5,400 unique products has been entered into the database
and approximately 50 products per month continue to be added. Specific information entered includes the
supplement name, manufacturer or brand name, ingredients, and nutrient content per pill in standardized units. A
secondary file contains information on nonnutrient supplements, such as botanical products and phytochemicals
(e.g., bioflavinoids, chlorophyll). Consultative and analytic services for external researchers, including use of
database information, are available through the Nutrition Shared Resource, a core unit for the UCSD Comprehensive
Cancer Center (Newman et al., 1998).
A database of 850 dietary supplements, identified in dietary data collected from studies conducted in southeast
Queensland and New South Wales, Australia, was compiled in 1997 (Ashton et al., 1997). All compound names
and amounts for total and individual ingredients were entered, as were identification of the active ingredients, if the
information was available. The database was constructed using Microsoft® Access, which allows exportation of
data into other file formats. The database is available in ASCII file format upon request, for a fee. The database has
not been expanded beyond the original set. In addition, the active ingredients of all dietary supplements sold in
Australia are included in the Australian Register of Therapeutic Goods, a computer database maintained by the
Australian Therapeutic Goods Administration (2003).
Pennsylvania State University developed and maintains a dietary supplement database in their Diet Assessment
Center (Copeland et al., 2000; Dwyer et al., 2003). It was created as part of a large nutrition intervention clinical
trial, the Women's Intervention Nutrition Study. This is a longitudinal study of approximately 2300 women with
previously diagnosed breast cancer. Dietary supplement product information has been collected as a component of
the study since 1996. The database contains compositional information on over 5000 different supplement
formulations. To date the database is not available to other investigators. Providing the data in a format useable by
others would require additional funding.
The POISINDEX® System is a database of hundreds of commercial, biological, and pharmaceutical products and
their ingredients (Micromedex, 2003). It provides data on clinical effects, range of toxicity, and treatment protocols
for exposure. Palmer et al. (2003) noted the fairly limited information for dietary supplement products in this
database. Only 36% of 1232 dietary supplement products and 45% of 1195 dietary supplement ingredients
associated with IDRs collected by 11 U.S. poison control centers in 1998 could be identified in the 2000 edition of
the POISINDEX® System database (Palmer et al., 2003).
G.1.2.3 Coding for health events
Standardized dictionaries for uniform coding of health-related terms assist in mapping event terms. One such
dictionary is the Medical Dictionary for Regulatory Activities terminology (MedDRA®), an international medical
terminology developed under the auspices of the International Conference on Harmonisation of Technical
Requirements for Registration of Pharmaceuticals for Human Use (International Conference on Harmonisation,
1998). MedDRA® is a registered trademark of the International Federation of Pharmaceutical Manufacturers
Associations. MedDRA® is used by the FDA and others world wide for coding standard medical terminology in
surveillance programs. MedDRA® recognizes more than 46,000 terms that potentially could be used by reporters to
describe signs and symptoms of events. It translates these terms into preferred terms, of which there are more than
11,000. Terms are also linked to higher levels of group and system organ class terms for aid in data searches and
retrieval. Coding is not automatic. Coders must have medical knowledge and be familiar with the structure and
content of MedDRA®. Verbatim terms are coded by matching to the corresponding MedDRA® term without
interpreting a diagnosis (U.S. Food and Drug Administration, 1999). Standard medical nomenclature is
continuously being updated and clarified. Note that in the pharmaceutical industry, standard terms use British
spellings.
Software programs are available to help with coding of medical terminology, such as dsNavigator (Galt Associates,
2002). Programs such as this apply a comprehensive dictionary of symptoms in the process of coding the adverse
events. However, as in the other steps of signal generation, coding requires human oversight for checks and
balances to ensure that data have been properly interpreted and managed.
G.1.2.4 Coding for product ingredients
The compendia of dietary supplement ingredient specifications that have been recognized by FDA for purposes of
label identification include the U.S. Pharmacopeia National Formulary (USP-NF, 2003b; U.S. Congress, 1994). The
USP-NF (2003b) contains monographs for 28 botanical ingredients, 111 single non-botanical ingredients (i.e.,
vitamins), and 16 vitamin and mineral combination products. In the 1997 final rule for nutrition labeling and
ingredient labeling of dietary supplements, FDA specified that the common or usual name of botanical ingredients
(including fungi and algae) shall be consistent with the names standardized in the 1992 edition of Herbs of
Commerce (McGuffin et al., 2000). If the Latin binomial name of the plant is not included in the aforementioned
reference, then it is to be included on the label (in parentheses) in accordance with the form in the 1994 edition of
the International Code of Botanical Nomenclature (“Tokyo Code”)(International Association for Plant Taxonomy,
1994). Recently, FDA (2003a) amended its regulation to incorporate by reference the 2000 editions of Herbs of
Commerce (American Herbal Products Association, Silver Spring, MD; McGuffin et al., 2000) and the International
Code of Botanical Nomenclature (“Saint Louis Code”)( Grueter et al., 2000).
The World Health Organization (WHO) considers the preferred herbal substance name to be the Latin binomial
name as expressed by the herbarium of the Royal Botanic Gardens in Kew (Royal Botanic Gardens in Kew, 2003).
Fucik et al. (2002) ranked 14 references that can be used for identifying synonym herbal substance names, and their
reference of choice was the WHO Monograph on Selected Medicinal Plants (2003). From these 14 references,
WHO built a computerized herbal substance register for use in coding of adverse events associated with herbal
products (Fucik et al., 2002).
G.1.3 Resources for building a relational database of individual data records and support for signaling
Expertise will be required to specify system requirements and select an electronic database, control data migration,
address validation issues, analyze data for detection of product-event signals, and interpret electronic signals
generated by the single and aggregate IDRs.
Several companies offer technical expertise and database management services for postmarketing surveillance
programs; some also offer proprietary software to help process product user health-related complaints, code terms
using standardized dictionaries, and analyze trends in spontaneously reported data:
• Galt Associates, Inc. (Sterling, VA; 2002)18
• PAREXEL International Corporation (Waltham, MA; 2003)
• Pharmaceutical Safety Assessments, Inc. (Narberth, PA, Faich, 2003)
• PROSAR Corporation (St. Paul, MN, 2003)
• RCN Associates, Inc. (Annapolis, MD, 2003)
• Relsys International, Inc. (Irvine, CA, 2003)
• Sentrx, formerly Global Safety Surveillance, Inc., (Little Falls, NJ, 2003)
• The Degge Group, Ltd. (Arlington, VA; 2003)
Several organizations host meetings for the purposes of providing training and networking in surveillance programs,
such as the Institute for International Research (2003) and the Drug Information Association (Horsham, PA).
G.1.4 Resources for signaling from other information sources
G.1.4.1 Resources for literature review
PubMed is a free search tool, available over the Internet, providing access to citations from biomedical literature
(National Library of Medicine, 2003). It was developed by the National Center for Biotechnology Information at
the National Library of Medicine, which is located on the campus of NIH, Bethesda, MD.
Through an interagency partnership with the USDA Food and Nutrition Information Center, ODS at NIH developed
and maintains the International Bibliographic Information on Dietary Supplements database. This database provides
access to bibliographic citations and abstracts from published, international, scientific literature on dietary
supplements (National Institutes of Health, 2003). ODS also compiled a Computer Access to Research on Dietary
Supplements (CARDS) database. CARDS lists research projects pertaining to dietary supplements that were funded
by the Institutes and Centers of NIH beginning with fiscal year 1999. Projects funded by other federal agencies are
added to CARDS as they become available. ODS posts notices on its website and listserv when CARDS updates are
completed (Office of Dietary Supplements, 2003).
The College of Pharmacy at the University of Illinois at Chicago produced and maintains the NAtural PRoducts
ALERT (NAPRALERT) database (College of Pharmacy at the University of Illinois at Chicago, 2003). This
resource, which is updated monthly, contains bibliographic and factual data on “natural” products. Information
includes pharmacology, biological activity, taxonomic distribution, ethno-medicine, and chemistry of plant,
microbial, and animal (including marine) extracts. In addition, the file contains data on the chemistry and
pharmacology of secondary metabolites that are derived from natural sources and that have known structure.
The Food and Nutrition Board of the Institute of Medicine of the National Academies is developing a system of
scientific reviews for evaluating the safety of dietary supplement ingredients (Institute of Medicine, 2004). They
conducted scientific reviews on chaparral, shark cartilage, glucosamine, melatonin, saw palmetto, and chromium
picolinate as prototypes for the system (Institute of Medicine, 2004). FDA will use these and future monographs to
determine whether or not to take regulatory action on an ingredient, such as requiring warning labels on products
that contain it or acting to ban its use.
A listing of additional Internet accessible resource databases for herbal citations is available from the Herb Research
Foundation (2003).
18
John Clark, MD, MSPH, a member of the LSRO Expert Committee, is Vice President, Surveillance Services
Group, Galt Associates, Inc., in Sterling, VA.
G.1.4.2 Resources for analyses of product composition
Dwyer et al. (2003) emphasized the need for unbiased and representative data on the active ingredients in botanical
supplements and the need for rapid and widely applicable methods of analytical analyses of these and other types of
dietary supplements. In support of this effort, reference standard materials for analysis of botanical ingredients will
be needed. Dwyer et al. (2003) suggested that priority should be given to bioactive compounds that, based on
emerging science, have assumed public health significance including but not limited to carotenoids, flavinoids,
omega-3 fatty acids, and trans fatty acids.
Of particular interest is whether label claims of content can be verified with analytical data. The composition of the
product associated with an adverse event is needed for interpretation of the ingredient-event and dose-event
relationship. More refined methods for analyses and interpretation may be required as the surveillance programs for
dietary supplements evolve.
Among products covered by the USP-NF, reference standards were developed for 36 botanical ingredients and 52
single non-botanical ingredients (i.e., vitamins)(U.S. Pharmacopeia, 2003b). The U.S. Pharmacopeia (2003a) will
conduct tests of identity, composition, purity, and strength for products enrolled in its Dietary Supplement
Verification program.
For a fee, laboratories such as ConsumerLabs.com (White Plains, NY) can test a limited number of dietary products
for composition, including tests for contamination. These analyses are designed to verify whether or not the product
label accurately reflects the composition of the product. Companies such as these can also investigate the
digestibility of the product (e.g., by measuring the degradation of the product in acid solutions).
G.1.4.3 Established databases of supplement users and nonusers that may be useful for active surveillance
From a practical perspective, the use of existing databases to assess risks from dietary supplements will be
challenging, especially because health status may affect the potential of dietary supplements to produce adverse
events. Yet, some databases may be available that can further characterize dietary supplement usage patterns and
the prevalence of specific health conditions. An important consideration for the selection of a database is whether or
not medical records can be accessed through the data source. Medical record data may be needed to validate the
diagnostic data in the database or gather additional information (U.S. Food and Drug Administration, 2003b).
Furthermore, information on the brand, dose, frequency, duration of supplement use, and use of non-nutrient
supplements is preferred but was often lacking in earlier observational and intervention studies (Neuhouser, 2003).
Databases of adults. Data on both dietary supplement use and medical outcome are collected in the Cancer
Prevention Study II Nutrition Cohort, which was established to investigate the relationship of diet to the risk of
cancer, mortality, and survival (Calle et al., 2002). The cohort includes over 84,000 men and 97,000 women, aged
50-74, who were surveyed in 1992 or 1993 for information on diet, vitamin supplement use, medical history, and
current medications (Calle et al., 2002). In 1997, 1999, and 2001, additional surveys were sent to surviving
participants. The 2001 questionnaire was expanded to include questions specifically on use of herbal supplements.
Example questionnaires, which were designed for self-administration and data processing by optical scanning, are
available from the American Cancer Society (Calle et al., 2002). Blood samples were also collected from a
subgroup of 40,000 subjects and are in storage. Of 162,408 participants at baseline, 45% did not use vitamin
supplements, 45.2% reported using vitamin supplements, and data were missing for 9.8%. Self-reported cancer
diagnoses that occurred through 1997 were verified by medical record review; medical records were obtained by
2002 for 64% of 9456 incident cancers. In addition, ongoing verification continues for cancers reported during the
interval from 1997 to 1999. As of 2002, 2,485 deaths have been attributed to cancer as recorded on death
certificates.
In 1993, the Cancer Research Center of Hawaii established a multiethnic cohort for long-term follow up of diet and
disease (Kolonel et al., 2000). The cohort includes more than 215,000 participants of both genders and of
multiethnic groups from Hawaii (Japanese-Americans, Caucasians, and native Hawaiians) and California (AfricanAmericans and Latinos). Questionnaire data include dietary history and supplement use. In addition, biological
specimens are being collected to establish a biorepository of blood and urine samples on nearly 100,000 members of
the cohort. Factors associated with supplement use were examined among 100,196 participants in 1993 through
1996 who reported an absence of chronic disease (Foote et al., 2003).
A cohort of 133,479 current and retired female California teachers and administrators was established in 1995 by the
Northern California Cancer Center (Hoggatt et al., 2002). Self-administered questionnaires were completed on diet,
lifestyle factors, general health, and medical history. Use of multivitamins and vitamins A, C, E, and beta-carotene
was assessed specifically for dose, frequency, and duration. Of the total, 34,996 (26%) did not use any dietary
supplements.
The Vitamin and Lifestyle Study was conducted on a cohort of 75,000 men and women, aged 50 to 75 years in
western Washington State in 2000 through 2002 (Satia-Abouta et al., 2003). The 24-page survey questionnaire
included questions on health complaints, medical history, and dose and frequency of supplement use (multivitamins,
ten single vitamins, six minerals, and herbal supplements) over the past ten years. Self-reported medical conditions
were not verified by medical record. Satia-Abouta et al. (2003) reported the results of 45,748 participants who
completed surveys by December 2001.
Gender specific databases: women. NIH established the Women's Health Initiative (WHI) in 1991 to study
cardiovascular disease, cancer, and osteoporosis in women for 15 years (National Heart, 2003). The WHI consists of
three major studies:
• A randomized controlled clinical trial of promising but unproven approaches to prevention, including
calcium/vitamin D supplementation
• An observational study to identify predictors of disease
• A study of community approaches to developing healthful behaviors
During clinical visits from 1993 through 1998, researchers in WHI conducted in-person interviews and used a
computerized inventory procedure to collect data on dietary supplement use from 68,133 women aged 50-79 years
in the clinical trial, and 93,676 women aged 50-79 years in the observational trial. Although all supplements used
were categorized (e.g., multivitamin) by viewing the product label, only unit amounts for antioxidant nutrients were
transcribed from the label or estimated based on the amount in leading brands. Sociodemographic data, health
history, and health behavior data were also collected (Shikany et al., 2003).
The Women's Healthy Eating and Living Study is a multisite randomized controlled trial of the effectiveness of a
high carotenoid diet in reducing breast cancer events and early death in women with early-stage invasive breast
cancer (Newman et al., 1998). Between 1995 and 2000, 3088 women were assigned to the diet intervention or to a
comparison group and will be followed through 2006. Information on diet and type, dose, and frequency of dietary
supplement use was collected by 24-hr recall interview. A supporting dietary supplement database of product
information was created as described earlier in this appendix. Consultative and analytic services for external
researchers, including use of database information are available through the Nutrition Shared Resource, a core unit
for the UCSD Comprehensive Cancer Center (Newman et al., 1998). Because women in the study are given regular
follow-up medical examinations, medical records may be available for purposes of additional research.
Gender specific databases: men. Neuhouser et al. (2001) examined dietary supplement use among participants in
the Prostate Cancer Prevention Trial, a multicenter double-blinded, placebo-controlled trial of the drug finasteride.
The 15,387 male participants, aged 55 years or older, completed questionnaires between 1993 and 1997 on medical
history, food frequency, and use of dietary supplements. Of these, 44.3% used a multivitamin, 35% used single
supplements of vitamin C or E, and 10-15% used antioxidant mixtures or single supplements of vitamins A and D,
zinc, or beta-carotene at least three times per week. The database includes information for multivitamins (one-perday, B-complex, stress formulations, and antioxidant types) and for a limited number of single supplements
(vitamins A, C, D, and E; beta-carotene, niacin, calcium, iron, selenium, zinc, garlic pills, and fish oil)(Neuhouser et
al., 2001). Detailed data on herbal use was not collected.
Muntwyler et al. (2002) studied the use of supplements and subsequent mortality from cardiovascular disease and
coronary heart disease for 83,639 U.S. male physicians, who had no history of cardiovascular disease or cancer at
baseline. History of disease, medications, risk factors (including self-reported blood pressure and blood
cholesterol), and current use of multivitamins and vitamins E, C, and A were obtained by questionnaire. Data on
duration of supplement use, product brands, and dose per week was also obtained to estimate the daily supplement
dose of vitamins E and C. Of the participants, 24,270 (29%) were current users of one or more supplements, mostly
multivitamin products (20,549, 24.6%). Approximately one-half of supplement users reported taking supplements
for four years or longer. Mortality was assessed by death certificate review at a mean follow-up period of 5.5 years.
G.1.4.4 Other information
The U.S. National Survey of Digestive Complaints was conducted by Innovative Medical Research (Towson, MD)
to obtain prevalence estimates for common gastrointestinal conditions in the general population (U.S. Food and
Drug Administration, 1998).
G.2 FUTURE RESEARCH
G.2.1 Research to improve recognition and reporting of adverse events in individual product users
One area that may deserve further study is the development of methods that obtain as much relevant information
from the first encounter with the reporter as possible in as structured a manner as possible.
G.2.1.1 Research to understand the barriers to reporting
Further research may be needed to determine why male product users are under-represented in IDRs collected
through spontaneous reporting systems and whether the system is missing signals that would arise if more IDRs
were secured from male product users.
Further research may be needed to understand the barriers to reporting that may be unique for dietary supplements
and how these obstacles might be overcome. For example, product users who perceived that dietary supplements
were “natural” and could not cause any physical harm may be less likely to link a physical ailment with the use of a
supplement. Additional barriers could include the reluctance by product users to reveal their use of dietary
supplements to their health care professional (Greger, 2001) and communicate their suspicion of a link between the
product and their ailment to the health care professional, especially if the health care professional recommended use
of the product in question (Walker, 2000).
G.2.2 Research for creation of individual data records
G.2.2.1 Research to gain knowledge of ingredients and products
The development of a dietary supplement database containing analytical values for ingredients would require
extensive resources, but would be useful (Dwyer et al., 2003). According to Dwyer et al. (2003), gaps in dietary
supplement composition data exist because of limited resources and rapid turnover of product formulations in the
market. Furthermore, the pharmacological effects of many active compounds in botanical products are largely
unknown, especially in individuals with pathological conditions. More data are needed on bioactive constituents in
dietary supplements but research is hindered by the lack of analytical methods for many of the ingredients and active
constituents in dietary supplements.
Because of the practice of ingredient overages and the multitude of products consisting of complex mixtures, there
may be some value in developing computational modeling to provide information to determine when variations in
content amounts may reach potentially pharmacologically active concentrations (Rosenkranz, 2003).
The Office of Inspector General, U.S. Department of Health and Human Services (2001) recommended that
development of a monograph system for dietary supplement ingredients would be especially helpful to adequately
assess signals generated by surveillance programs. Monograph papers of particular ingredients contain safety
information, including information about conditions under which the ingredient may be pharmacologically active.
G.2.3 Research for building a relational database and to improve signaling from databases of individual data
records
As the statistical methods used in the scientific discipline of pharmacovigilance evolve, they will pave the way for
the application of these methods in other such surveillance programs. Research efforts should be targeted at
reducing and controlling false discovery rates.
G.2.4 Research to improve signaling from other information sources
Further research is needed to provide guidance for decisions of whether or not efforts above and beyond adverse
event monitoring programs are warranted (e.g. the need for active surveillance efforts)(Hostelley, 2003).
Research to extend general knowledge might include studies of market sales, dietary supplement products, dietary
supplements users, and individuals who do not use dietary supplements (Yetley, 2003). Specifically, the following
studies would be beneficial for signal generation and interpretation:
• The variability in product composition due to various factors such as season, location of plant, and plant
strain
• Supplement-drug interactions and ingredient-ingredient interactions in mixtures
• Determination of product sales numbers and establishment of accurate estimates of exposure (i.e., 90th and
95th percentiles across various consumer/risk groups)
• The prevalence estimates for medical conditions in the general population (e.g., how common is dizziness
for individuals in each of the relevant demographic categories or product user groups)
• Idiosyncratic reactions in the general product user population and the identification of sensitive subgroups
• Biological plausibility for events
• How use conditions (e.g., physical exertion, weight loss, concurrent illness, dehydration) alter the potential
for toxicity and distribution or prevalence of these conditions in the population
• The risk of cumulative, long-term effects
• Reliable biomarkers of supplement toxicity
G.3 SUMMARY OF RESOURCES AND FUTURE RESEARCH
Surveillance programs need not be technically sophisticated to achieve their goals, so long as they are organized
correctly and systematic in their application. Yet, the field of adverse event surveillance is becoming highly
technological and sophisticated. The most advanced applications are introduced in pharmacovigilance, whose data
collection techniques and data management infrastructure can serve as a resource for surveillance programs for
G.4 LITERATURE CITATIONS FOR ANCILLARY ISSUES
Ashton, B. A., Ambrosini, G. L., Marks, G. C., Harvey, P. W. & Bain, C. (1997) Development of a dietary
supplement database. Aust. N. Z. J. Public Health 21: 699-702.
Australian Therapeutic Goods Administration. (2003) Office of Complementary Medicines. Advice & Consultation.
OCM, CMEC, CHCF. Available at http://www.health.gov.au/tga/docs/pdf/facts.pdf. Accessed 3-25-2003.
Calle, E. E., Rodriguez, C., Jacobs, E. J., Almon, M. L., Chao, A., McCullough, M. L., Feigelson, H. S. & Thun, M.
J. (2002) The American Cancer Society Cancer Prevention Study II Nutrition Cohort: rationale, study design, and
baseline characteristics. Cancer 94: 2490-2501.
Claims Division, N. F. P. A. Olivetti, P. L., ed. (1997) A manual for successful resolution of consumer complaints in
the food industry. Second Edition. The Food Processors Institute, Washington, DC, pp. 1-36.
College of Pharmacy at the University of Illinois at Chicago. (2003) NAtural PRoducts ALERT Database.
Available at http://www.cas.org/ONLINE/DBSS/napralertss.html. Accessed 9-25-2003.
Copeland, T., Grosvenor, M., Mitchell, D. C., Smiciklas-Wright, H., Marsoobian, V., Blackburn, G. & Winters, B.
(2000) Designing a quality assurance system for dietary data in a multicenter clinical trial: Women's Intervention
Nutrition Study. J. Am. Diet. Assoc. 100: 1186-1190.
Dwyer, J., Picciano, M. F. & Raiten, D. J. (2003) Food and dietary supplement databases for What We Eat in
America-NHANES. J. Nutr. 133: 624S-634S.
Faich, G. A. (2003) Pharmaceutical Safety Assessments, Inc. Available at http://www.gfaich.com. Accessed 9-242003.
J. Epidemiol. 157: 888-897.
Fucik, H., Backlund, A. & Farah, M. (2002) Building a computerized herbal substance register for implementation
and use in the World Health Organization International Drug Monitoring Programme. Drug Inf. J. 36: 839-854.
Galt Associates. (2002) Galt Associates and Uppsala Monitoring Centre Offer Seamlessly Integrated Web-enabled
Solution for Drug Dictionary Browsing and Encoding. [Press Release]. Available at http://www.whoumc.org/pdfs/galt.pdf. Accessed 6-3-2003.
Grueter, W., McNeill, J., Barrie, F. R., Burdet, H. M., Demoulin, V., Figueiras, T. S., Nicolson, D. H., Silva, P. C.,
Skog, J. E., Trehane, P. & Turland, N. J. (2000) International code of botanical nomenclature (St. Louis Code).
(Grueter, W. & McNeill, J., eds.). International Association for Plant Taxonomy Koeltz Scientific Books,
Köningstein.
Health Canada (2004) Adverse reaction reporting. Available at http://www.hc-sc.gc.ca/hpfb-dgpsa/nhpddpsn/arr_hcpc_e.html. Accessed 1-2-2004.
Herb Research Foundation. (2003) Recommended links to the world of herbs. Available at
http://www.herbs.org/links/linksdata.htm. Accessed 9-25-2003.
Institute for International Research. (2003) Institute for International Research. Available at
http://www.iir.se/indexmiddle.htm. Accessed 9-25-2003.
Institute of Medicine. (2004) Dietary Supplements: A Framework for Evaluating Safety. Washington, DC, National
Academy Press.
International Association for Plant Taxonomy. (1994) International Code of Botanical Nomenclature (Tokyo Code).
Available at http://www.bgbm.fu-berlin.de/iapt/nomenclature/code/tokyo-e/. Accessed 9-24-2003.
Accessed 9-29-2003.
Kennedy, D. L., Goldman, S. A. & Lillie, R. B. (2000) Spontaneous reporting in the United States. In:
Pharmacoepidemiology. 3 ed. John Wiley & Sons, Ltd., United Kingdom, pp. 151-174.
Kolonel, L. N., Henderson, B. E., Hankin, J. H., Nomura, A. M., Wilkens, L. R., Pike, M. C., Stram, D. O., Monroe,
K. R., Earle, M. E. & Nagamine, F. S. (2000) A multiethnic cohort in Hawaii and Los Angeles: baseline
characteristics. Am. J. Epidemiol. 151: 346-357.
McDowell, M. & Ervin, B. (2003) Dietary supplement databases—Information from NHANES III and plans for
future NHANES database. Available at http://www.nal.usda.gov/fnic/foodcomp/conf/NDBC22/2-2paper.pdf.
Accessed 9-24-2003.
McGuffin, M., Kartesz, J., Leung, A. & Tucker, A. (2000) Herbs of Commerce. 2nd ed.: American Herbal Products
Association, Silver Spring, MD pp. 1-442.
Micromedex. (2003) Poisindex® System. Available at http://www.micromedex.com/products/poisindex/. Accessed
7-8-2003.
Muntwyler, J., Hennekens, C. H., Manson, J. E., Buring, J. E. & Gaziano, J. M. (2002) Vitamin supplement use in a
low-risk population of US male physicians and subsequent cardiovascular mortality. Arch. Intern. Med. 162: 14721476.
MyAsthma.com. (2003) Daily Diary pages. Available at http://www.myasthma.com/tour/tourData.asp. Accessed 924-2003.
Mycysticfibrosis.com. (2003) Daily Diary pages. Available at http://www.mycysticfibrosis.com/tour/tour04.asp.
Accessed 9-24-2003.
National Center for Infectious Diseases. (2003) Infectious Disease Surveillance. Available at
http://www.cdc.gov/ncidod/osr/inernetsurveyss.htm. Accessed 9-25-2003.
National Heart, Lung and Blood Institute. (2003) Women's Health Initiative. Available at
http://www.nhlbi.nih.gov/whi/background.htm. Accessed 9-25-2003.
National Institutes of Health. (2003) Office of Dietary Supplements. Available at http://ods.od.nih.gov. Accessed
9-25-2003.
National Library of Medicine. (2003) PubMed. Available at http://www.ncbi.nlm.nih.gov/entrez/query.fcgi.
Accessed 9-25-2003.
Neuhouser, M. L., Kristal, A. R., Patterson, R. E., Goodman, P. J. & Thompson, I. M. (2001) Dietary supplement
use in the Prostate Cancer Prevention Trial: implications for prevention trials. Nutr. Cancer 39: 12-18.
Newman, V., Rock, C. L., Faerber, S., Flatt, S. W., Wright, F. A. & Pierce, J. P. (1998) Dietary supplement use by
women at risk for breast cancer recurrence. The Women's Healthy Eating and Living Study Group. J. Am. Diet.
Assoc. 98: 285-292.
Office of Dietary Supplements, National Institutes of Health. (2003) Fact sheets on dietary supplements. Available
at http://dietary-supplements.info.nih.gov/showpage.aspx?pageid=90. Accessed 8-14-2003.
Palmer, M. E., Haller, C., McKinney, P. E., Klein-Schwartz, W., Tschirgi, A., Smolinske, S. C., Woolf, A., Sprague,
B. M., Ko, R., Everson, G., Nelson, L. S., Dodd-Butera, T., Bartlett, W. D. & Landzberg, B. R. (2003) Adverse
events associated with dietary supplements: an observational study. Lancet 361: 101-106.
Parexel.com. (2003) Parexel Strategy and Services. Available at http://www.parexel.com. Accessed 9-24-2003.
Prosar. (2003) About Prosar. Available at http://www.prosarcorp.com/abou/. Accessed 7-25-2003.
RCN Associates. (2003) Drug Safety First and Only! Available at http://www.rcnrx.com. Accessed 9-24-2003.
Relsys International, Inc. (2003) Relsys Innovative Safety Systems. Available at http://www.relsysinc.com/products/overview.html. Accessed 8-6-2003.
Rosenkranz, H. S. (2003) Application of computers, bioinformatics and modeling to predict effectiveness and safety
of dietary supplements. [Presented at the NIH Office of Dietary Supplements Dietary Supplement Use in Elderly
Workshop, January 14-15, 2003, Bethesda Maryland]. Available at http://dietarysupplements.info.nih.gov/pubs/elderly.14jan03.abst.rosenkranz.pdf. Accessed 3-21-2003.
Royal Botanic Gardens in Kew. (2003) Data & publications. Available at http://www.rbgkew.org.uk/data/.
Accessed 9-24-2003.
Satia-Abouta, J., Kristal, A. R., Patterson, R. E., Littman, A. J., Stratton, K. L. & White, E. (2003) Dietary
supplement use and medical conditions: the VITAL study. Am. J. Prev. Med. 24: 43-51.
Sentrx. (2003) Sentrx Safety Trak®. Available at http://www.sentrx.com/body_our_solutions.htm?. Accessed 7-152003.
Shikany, J. M., Patterson, R. E., Agurs-Collins, T. & Anderson, G. (2003) Antioxidant supplement use in Women's
Health Initiative participants. Prev. Med. 36: 379-387.
The Degge Group, Ltd. (2003) Drug Safety Check List. Available at http://www.deggegroup.com
U.S. Department of Agriculture. (2003) How to get information from NDL (Nutrient Data Library). Available at
http://www.nal.usda.gov/fnic/foodcomp/contact.html. Accessed 9-24-2003.
U.S. Department of Health and Human Services. Office of Inspector General (2001) Adverse event reporting for
dietary supplements. Report No. OEI-01-00-00180. Office of Evaluations and Inspections, Boston. pp. 1-110.
U.S. Food and Drug Administration. (6-15-1998) Food Advisory Committee on Olestra. Volume 1 [Transcript].
Available at http://www.fda.gov/ohrms/dockets/98/transcpt/3485t1.pdf. Accessed 7-8-2003.
U.S. Food and Drug Administration. (1999) MedDRA and pharmacovigilance - The way forward. Available at
http://www.fda.gov/cder/present/dia-699/dia628/. Accessed 9-24-2003.
U.S. Food and Drug Administration. (2001a) Electronic submission through the FDA ESTRI Gateway. Frequently
asked questions (FAQ). Available at http://www.fda.gov/oc/electronicsubmissions/interfaq.htm. Accessed 9-242003a.
U.S. Food and Drug Administration. (2001b) Guidance for Industry. Providing regulatory submissions to Office of
Food Additive Safety in electronic format — General consideration. Available at
http://www.cfsan.fda.gov/~dms/opaeguid.html. Accessed 8-25-2003b.
U.S. Food and Drug Administration (2003a) Food labeling: Ingredient labeling of dietary supplements that contain
botanicals. [Direct final rule]. Fed. Reg. 68: 51693-51698.
U.S. Food and Drug Administration. (2003b) Risk Management Public Workshop. Pharmacovigilance Practices and
http://www.fda.gov/cder/meeting/RMtranscript3.doc. Accessed 6-2-2003b.
U.S. Pharmacopeia. (2003a) USP's dietary supplement verification program overview. Available at http://www.uspdsvp.org/background/overview.html. Accessed 3-24-2003a.
U.S. Pharmacopeia. (2003b) USP dietary supplement monographs & reference standards. Available at
http://www.usp-dsvp.org/standinfo/monographs.html. Accessed 3-24-2003b.
Walker, A. M. (3-9-2000) The relationship between voluntary notification and material risk in dietary supplement
safety. Harvard School of Public Health, Cambridge, MA.
World Health Organization. (2003) WHO Monographs on Selected Medicinal Plants. Available at
http://www.who.int/medicines/library/trm/medicinalplants/monographs.shtml. Accessed 9-24-2003.
Yetley, E. (2003) Plenary Session I: Overview of dietary supplements. [Presented at the NIH Office of Dietary
Supplements Dietary Supplement Use in elderly Workshop, January 14-15, 2003, Bethesda Maryland]. Available at
http://dietary-supplements.info.nih.gov/pubs/elderly.14jan03.yetley.pdf. Accessed 3-21-2003.
APPENDIX H
ACRONYMS AND GLOSSARY
H.1 TABLE OF ACRONYMS
ACRONYM
AAPCC
CARDS
CCIDM
CDC
CFSAN
CGMP
DSHEA
FDA
FTC
IC
ICH
IDR
LSRO
MedDRA®
MET
NAPRALERT
NHANES
NIH
NPC
ODS
SOP
TESS
UMC
U.S.
USCD
USDA
USP-NF
WHI
WHO
WHO-ART
DEFINITION
American Association of Poison Control Centers
Computer Access to Research on Dietary Supplements
Collaborating Center for International Drug Monitoring
Centers for Disease Control and Prevention (U.S. Department of Health and Human Services)
Center for Food Safety and Applied Nutrition (U.S. Food and Drug Administration)
Current Good Manufacturing Practice
Dietary Supplement Health and Education Act
U.S. Food and Drug Administration (U.S. Department of Health and Human Services)
Federal Trade Commission
Information component
International Conference on Harmonisation19
Individual data record
Life Sciences Research Office, Inc.
Medical Dictionary for Regulatory Activities
Metabolife International, Inc.
NAtural PRoducts ALERT database
National Health and Nutrition Examination Survey
National Institutes of Health (U.S. Department of Health and Human Services)
National Pharmacovigilance Center
Office of Dietary Supplements (National Institutes of Health)
Standard operating procedures
Toxic Exposure Surveillance System
Uppsala Monitoring Center
United States
University of California, San Diego
U.S. Department of Agriculture
U.S. Pharmacopeia National Formulary
Women's Health Initiative
World Health Organization
World Health Organization Adverse Reaction Terminology
H.2 GLOSSARY TABLE
TERM
Adulterated
Adverse event
Adverse event
monitoring program
(or system)
Alopecia
DEFINITION
A food is adulterated (and therefore illegal) if it was prepared, packed, or held under
unsanitary conditions that may have introduced contaminants, or is otherwise
potentially harmful (e.g., contains unapproved food additives) (U.S. Congress, 1994).
An adverse event is any undesirable health-related sign or symptom that is detected in
an exposed individual after use of a product. The adverse event may or may not be an
adverse reaction to the product. Thus, the term is used whether or not the event can be
attributed to the product (Clark et al., 2001).
See “Surveillance program.”
Absence of hair from skin areas where it normally is present.
19
TERM
Asthenia
Batch/lot
Batch/lot number
Cerebrovascular
accident
Complaint
(i.e., consumer
complaint)
Contaminant
Dietary supplement
Dyspnea
Ephedra
Ephedrine alkaloid
False negative signal
False positive signal
Food additive
Guarana
Health care
professional
DEFINITION
Weakness; lack of energy or strength.
A specific quantity of a dietary ingredient or dietary supplement that was produced to
meet specifications for identity, purity, quality, strength, and composition, during a
specified time period according to a single manufacturing record during the same cycle
of manufacture (U.S. Food and Drug Administration, 2003).
Any distinctive group of letters, numbers, or symbols, or any combination of them,
from which the complete history of the manufacturing, packaging, or holding of a batch
or lot of dietary ingredients or dietary supplements can be determined (United States
Pharmacopeial Convention, 2002; U.S. Food and Drug Administration, 2003).
This general term encompasses such problems as stroke and cerebral hemorrhage.
Any communication provided to the retailer or distributor or manufacturer that contains
any allegation, written or oral, expressing consumer dissatisfaction with the quality of a
dietary ingredient or a dietary supplement related to good manufacturing practices (U.S.
Food and Drug Administration, 2003). For the purposes of this Life Sciences Research
Office (LSRO) publication, we are using the term “complaint” in a broader context to
also include health-related complaints that may be due to product quality and/or
pharmacological activity.
A foreign compound in a dietary supplement that may be dangerous to health. Noxious
contaminants generally include heavy metals, pesticide residues or pathogenic
microorganisms (U.S. Pharmacopeia, 2003).
This is a product taken by mouth that contains an ingredient intended to supplement the
diet. Such ingredients may include vitamins, minerals, herbs, botanicals, amino acids,
enzymes, organ tissues, and metabolites. They may be in the form of extracts or
concentrates, tablets, capsules, softgels, gelcaps, liquids, bars, or powders (U.S.
Congress, 1994).
Shortness of breath. It is a feeling of difficult or labored breathing that is out of
proportion to the level of physical activity.
The term “ephedra” describes any plant of the genus Ephedra and extracts from plants
of this genus, especially those extracts containing ephedrine alkaloids. Ephedra is also
known by the Chinese name “ma huang.” The pharmacological activity of ephedra
depends upon the composition of its ephedrine alkaloids, which can be affected by
species, geographic location and growing conditions (Shekelle et al., 2003).
Ephedrine alkaloids occur naturally in some plants (e.g., Ephedra sinica, Sida
cordifolia) and are biologically active. The molecular structure of ephedrine is similar
to the structure of amphetamine. Effects include stimulation of the heart and central
nervous system (Wooltorton & Sibbald, 2002).
A true signal of a product-event combination that is overlooked or dismissed.
An adverse event that occurs independently from a suspect product (no causal
relationship), but that has been identified by procedure as a product-event combination
of interest.
In its broadest sense, a food additive is any substance added to food. Legally, the term
refers to "any substance the intended use of which results or may reasonably be
expected to result-directly or indirectly-in its becoming a component or otherwise
affecting the characteristics of any food." This definition includes any substance used in
the production, processing, treatment, packaging, transportation or storage of food (U.S.
Also known as “Brazilian cocoa.” It contains chemical compounds called xanthenes,
which include caffeine, theobromine, and theophylline. Guarana is used for its
stimulating properties (an antifatigue agent), and for its diuretic and antidiarrheic
actions.
A health care professional would include those licensed to provide health care, such as a
physician, nurse, pharmacist, dentist, registered dietitian/nutritionist, naturopath, midwife, chiropractor, physical therapist, speech pathologist, psychiatrist, and psychologist.
TERM
Heart block
Individual data record
Inter-product database
Intra-product database
Ma huang
Misbranded
Off label
Palpitations
Recall
Rechallenge
Relational database
Reporter
Reporting
DEFINITION
A cardiac disorder resulting from the inability of an electrical impulse in the atria to
reach the ventricles.
Standard displays of administrative, demographic, product exposure, and adverse event
information that arise from individual user adverse experiences. Such individual user
experiences can be derived from spontaneously forwarded material, literature
publications, studies, or clinical trials, or quasi-study solicited environments (e.g.,
registries). Individual data records of adverse events are to be distinguished from
complaints that pertain to the appearance or function of products (e.g., cap not sealed,
feather in bottle). For the purposes of this publication, the term “individual data record”
includes mandatory and voluntary reports to federal surveillance systems and
documented, spontaneous, health-related complaints to industry.
Database containing information (e.g., individual data records) related to products of
different manufacturers.
Database containing information (e.g., individual data records) related to a single
product or the product lines of a single manufacturer.
The Chinese name for “ephedra.”
The product is considered misbranded if the label bears false or misleading information
or does not bear the required information (U.S. Congress, 1994).
Use of a product in ways other than those intended by the manufacturer, based on
product information.
A sensation in which a person is aware of an irregular, hard, or rapid heartbeat.
Palpitations can be felt in the chest, throat, or neck.
The manufacturer recalls from market a product that violates legislation and the level of
retrieval depends upon its potential hazard to public health (U.S. Food and Drug
Administration, 2002). FDA can seize product if it is misbranded, mislabeled, or
adulterated.
Class I (immediate hazard) recalls are for dangerous or defective products that
predictably could cause serious health problems or death. A Class I recall seeks to
retrieve product from wholesalers, retailers and consumers.
Class II (no imminent hazard) recalls are for products that might cause temporary health
problems or pose only a slight threat of a serious nature. A Class II recall reaches the
wholesale or retail level to retrieve product, depending on the nature of the problem.
Class III (no imminent hazard) recalls are for products that are unlikely to cause any
adverse health reaction, but do violate FDA labeling or manufacturing regulations.
Examples might be a container defect (e.g., plastic material delaminating, a lid that does
not seal, leak from a bottled drink), off-taste, off-color, or lack of English labeling in a
retail food. A Class III recall reaches the wholesaler level to retrieve product.
A rechallenge can be conducted to determine whether or not the adverse event reoccurs
after a second exposure to the product. To conduct a rechallenge, the use of a product is
halted. Once symptoms subside, the product is reintroduced to the individual, who is
monitored for health-related signs and symptoms. A rechallenge is sometimes selfimposed by the product user or conducted under medical supervision. Intentional
rechallenge could be carried out if the treating health care professional judges it would
likely be of clinical benefit to the patient.
A database that is capable of importing and exporting information with other databases.
The term “reporter” refers to the individual who is submitting a health-related
complaint/report. This generally includes product users, health care professionals,
relatives, and lay advocates.
For the purposes of this LSRO publication, the term “reporting” is used for the action of
submitting a voluntary consumer health-related complaint to industry and for the
submission/filing of mandatory and voluntary reports of adverse events to the federal
surveillance system.
TERM
Surveillance program
(or Surveillance
system)
Signal
Signaling
Supplement Facts
Tachycardia
Trade dress
Withdrawal of product
DEFINITION
The personnel, processes, and computer resources that are needed to carry out
postmarketing surveillance. Surveillance programs establish mechanisms by which the
collection, receipt, analysis, interpretation, and documentation of individual data
records can be consistently achieved. The primary purpose of a surveillance program is
to identify potential product-related health problems that are appropriate for rapid
dissemination of information or other actions. Synonym: “adverse event monitoring
program (or system)”
A potential product-event relationship that deserves further investigation.
The planned, periodic application of analytical methods to databases of individual data
records to detect and evaluate signals.
Information that must be on a dietary supplement label. Required information includes a
descriptive name of the product stating that it is a supplement; the name and place of
business of the manufacturer, packer, or distributor; a complete list of ingredients; and
the net contents of the product. In addition, each dietary supplement (except for some
small volume products or those produced by eligible small businesses) must have
nutrition labeling in the format of a "Supplement Facts" panel (U.S. Food and Drug
Administration, 2001). This panel must identify each dietary ingredient contained in
the product including those substances such as botanicals for which no daily value has
been established. If botanicals are listed then the part of the plant used must be
identified by its common name as well as its Latin binomial name (U.S. Food and Drug
Administration, 1997). Proprietary blends may be listed by weight but the individual
ingredients must also be listed in descending order with greatest amount first. The label
must also contain information on the manufacturers suggested serving size (U.S. Food
and Drug Administration, 1999).
Abnormally fast heartbeat, particularly applied when the rate is over 100 beats a minute.
Characteristic appearance of a product to help distinguish it from other products (e.g.,
blue capsule stamped with logo).
A firm removes product from further sale or use. This action is not considered to be a
recall because the product does not violate legislation.
H.3 LITERATURE CITATIONS FOR ACRONYMS AND GLOSSARY
Clark, J. A., Klincewicz, S. L. & Stang, P. E. (2001) Spontaneous adverse event signaling methods: Classification
and use with health care treatment products. Epidemiol. Rev. 23: 191-210.
Shekelle, P., Hardy, M. L., Maglione, M. & Morton, S. C. (2003) Ephedra and Ephedrine for Weight Loss and
Athletic Performance Enhancement: Clinical Efficacy and Side Effects. Evidence Report/Technology Assessment
Number 76 (Prepared by Southern California Evidence-based Practice Center, RAND, under Contract No 290-970001, Task Order No. 9). [Final Report]. Report No. AHRQ Publication No. 03-E022. Agency for Healthcare
Research and Quality, Rockville, MD.
United States Pharmacopeial Convention (2002) Manufacturing practices for nutritional supplements. In: The
United States Pharmacopeia. The National Formulary. 20th ed. Webcom Limited, Toronto, Ontario, pp. 2667-2673.
U.S. Food and Drug Administration. (1992) What is a food additive? Available at
http://www.cfsan.fda.gov/~dms/qa-adf15.html. Accessed 7-8-2003.
U.S. Food and Drug Administration (1997) Food labeling regulation, amendments; Food regulation uniform
compliance date; and new dietary ingredient premarket notification; Final rules. Fed. Reg. 62: 49826-49858.
U.S. Food and Drug Administration. (1999) Summary of nutrition labeling rules for dietary supplements:
Supplement Facts panel. Available at http://www.cfsan.fda.gov/~acrobat/hhssupp2.pdf. Accessed 8-8-2003.
U.S. Food and Drug Administration. (2001) Overview of dietary supplements. Available at
http://www.cfsan.fda.gov/~dms/ds-oview.html. Accessed 7-7-2003.
U.S. Food and Drug Administration. (2002) FDA recall policies. Available at
http://www.cfsan.fda.gov/~lrd/recall2.html. Accessed 7-8-2003.
U.S. Pharmacopeia. (2003) Dietary supplement lexicon. Available at http://www.uspdsvp.org/background/lexicon.html. Accessed 3-24-2003.
Wooltorton, E. & Sibbald, B. (2002) Ephedra/ephedrine: cardiovascular and CNS effects. CMAJ. 166: 633.

recommendations for adverse event monitoring programs for

Transcription

Similar documents

Reduce Your Risk

Dietary Supplements 101

Ambrotose®-The Missing Link in Human Nutrition

dietary supplements as a source of illegal/designer drugs

click - PHILSPEN Main Page

Michael Specter, author of "Denialism"

Natural Living - Lazy Acres Market