Performance of a Rapid-Test Algorithm for Confirming HIV

Transcription

Performance of a Rapid-Test Algorithm for Confirming HIV
Performance of a Rapid-Test Algorithm
for Confirming HIV Serostatus in Beira and
Chókwè, Mozambique
Arlinda Zango,1 Paulo Arnaldo,2 Karine Dubé,3 Ricardo Thompson,2 Josefo Ferro,1 Sónia
Enosse,2 Ivete Meque,1 Shelly Fischer,3 Janneke van de Wijgert,4 and Paul Feldblum3
BACKGROUND
The national HIV-testing algorithm in
Mozambique includes two HIV rapid tests,
which are followed by testing with
Vironostika® HIV-1 ELISA when it is
necessary to confirm discrepant or
indeterminate results. Possible reasons for
a discrepant or indeterminate result from
an HIV rapid test could be a recent HIV
infection; differing test sensitivity; or a
problem with the test procedure itself,
such as contamination of the blood
specimens. Discrepant or indeterminate
results may be difficult to resolve in
Mozambique, however, because ELISA
technology is not widely available. In
addition, the reagents needed for ELISA
assays do not have a long shelf life and they
are difficult to procure.
Determine HIV-1/2
METHODS
In two cross-sectional surveys to determine
HIV seroprevalence among women between
the ages of 18 and 35 in Beira (Sofala
Province) and Chókwè (Gaza Province),
Mozambique, specimens were screened
for HIV using DetermineTM HIV 1/2 and
Uni-GoldTM rapid tests serially. Per the
national HIV-testing algorithm, specimens
with discrepant results were sent to the
capital, Maputo, for confirmatory testing
with Vironostika HIV-1 ELISA. For research
purposes only, the discrepant specimens
were also tested with the SD Bioline
HIV 1/2 rapid test.
Vironostika HIV-1 ELISA
RESULTS
From the 1,182 cross-sectional specimens
that were tested (706 in Beira and 476 in
Chókwè), the results from seven specimens
(0.6%)—three in Beira and four in Chókwè—
were discrepant and required confirmatory
testing. Six of these specimens were then
tested with both SD Bioline HIV-1/2 and
Vironostika HIV-1 ELISA. Three of these six
had concordant negative results, two had
concordant positive results, and one was
indeterminate on SD Bioline and positive on
Vironostika (and tested “recent” on the
BEDTM EIA assay). This one participant was
retested for HIV six weeks later, at which
time the new specimen tested positive on
Determine, Uni-Gold, SD Bioline, and
Vironostika ELISA. The seventh specimen
with discrepant results on the rapid tests
(Determine and Uni-Gold) tested negative on
SD Bioline, but has not yet been tested by
Vironostika HIV-1 ELISA.
Table 1. Discrepant HIV rapid-test results in two cross-sectional HIV surveys in Beira and Chókwè, Mozambique
Uni-Gold HIV
Site
Participant
number
Date
Determine
Uni-Gold
SD Bioline
Vironostika
HIV-1 ELISA
1
Beira
50095
23 Feb 2010
Positive
Negative
Positive
Positive
2
Beira
50224
9 July 2010
18 Aug 2010
Positive
Positive
Negative
Positive
Indeterminate
Positive
Positive
Positive
3
Beira
50567
14 Dec 2010
Positive
Negative
Positive
Positive
4
Chókwè
60091
12 July 2010
Positive
Negative
Negative
Negative
5
Chókwè
60131
21 July 2010
Positive
Negative
Negative
Negative
6
Chókwè
60221
23 Aug 2010
Positive
Negative
Negative
Negative
7
Chókwè
60309
8 Oct 2010
Positive
Negative
Negative
Pending*
*Result pending because of unavailability of Vironostika HIV-1 ELISA test kits
Conclusions
We observed a low rate of discrepant or
indeterminate results for HIV rapid tests.
When results are discrepant or
indeterminate, however, a third rapid test,
such as SD Bioline HIV-1/2, may be a suitable
tiebreaker in Mozambique, especially in
peripheral settings where ELISA technology
SD Bioline HIV-1/2 3.0
is not readily available. Although supplies of
the three rapid tests would be a concern, this
may be less of a problem than the delays
inherent in ELISA testing and patients’ failure
to return for testing four to six weeks after
discrepant or indeterminate HIV results.
Author affiliations
1
2
enter for Infectious Disease Research/Centro de Investigação de
C
Doenças Infecciosas (CIDI), Universidade Católica de Moçambique
(UCM), Beira, Mozambique
hókwè Health Research and Training Center/Centro de
C
Investigação e Treino em Saúde de Chókwè (CITSC), National
Institute of Health, Chókwè, Mozambique
3
FHI, Clinical Sciences, Research Triangle Park, NC, USA
4
cademic Medical Center (AMC) of the University of AmsterdamA
Center for Poverty-related Communicable Diseases (AMC CPCD),
Amsterdam, The Netherlands
www.fhi.org