Incidence and predictors of luteal phase deficiency by different

Joseph B Stanford, MD
Karen C Schliep, PhD
Ahmad O Hammoud, MD
Enrique Schisterman, PhD
Sunni Mumford, PhD
Christy Porucznik, PhD


Menstrual cycle ovulatory disorders range
from luteal phase deficiency (LPD) to more
advanced cases of anovulation.
Alterations in life style (exercise, diet, and
stress) and abnormal endocrine dynamics are
known to cause anovulation; however, their
role in milder menstrual cycle phenotypes
such as LPD is less known.

Clinically, LPD may be associated with:
1. abnormal luteal phase progesterone and estradiol
(E2) production
2. shortening of the menstrual cycle (by shortening
of the luteal phase)
3. premenstrual spotting or bleeding
4. pregnancy related disorders such as infertility (via
impairment of endometrial development) and
early pregnancy loss.


While LPD is thought to occur in 3-20% of
women who are infertile and in 25-60% of
women with recurrent spontaneous abortion,
data also suggest that 6-10% of women with
normal fertility demonstrate an inadequate
luteal phase.
There is a need, thus, to better understand
normal variations in the menstrual cycle
phases and in variations that could be
pathologic.

Assess luteal phase function in the BioCycle
Study by determining cycles with:
1. Short luteal phases (< 10 days in length)
2. Incidence of premature luteinization (evidenced
by a late follicular rise in progesterone before
ovulation)
3. Premenstrual spotting
4. Abnormal luteal phase progesterone
1.
2.
Some normal cycling women express
evidence of subtle menstrual cycle disorders
manifested as LPD.
Various demographic, lifestyle,
environmental, and nutritional factors
increase the incidence of LPD.
BioCycle Study (2005-2007): Followed 259 women
from NY region for two menstrual cycles
 Inclusion:

 Ages 18-44
 Regularly menstruating

Exclusion:
 Conditions known to affect menstrual cycle function
 Self-reported over- or underweight or unusual diet

Strong compliance: 250 completed two cycles; 9
completed one cycle


BioCycle women with cycles of known length
(as determined by daily records of menstrual
bleeding) who reached peak fertility via
Clearplan® fertility monitor (based on E3G
and LH in urine) OR who reached an LH peak
serum surge > 19 ng/mL (for women not
peaking via the monitor or with cycle lengths
≥30 days)
n=241 women, 406 cycles

Observed day of ovulation (i.e., urine or serum
LH surge + 1 day) was used to classify visits into
correct cycle phase categories.
 Cycle length was defined as number of days between
menstrual bleeding.
 Day 1 of the cycle was defined as menstruating by
4pm on that day (after confirming 2 consecutive days
of bleeding).
 Follicular length: Day 1 of bleeding through day of
ovulation.
 Luteal length: Day after ovulation through last day of
cycle.

Based on previous research, we defined LPD in 5
unique ways:
1. Luteal phase duration < 10 days in length.
2. Deficient luteal phase progesterone production via
AUC analysis (i.e., lowest quintile).
3. Deficient luteal phase progesterone production via
cutpoint (i.e., ≤ 5 ng/mL for all 3 luteal visits).
4. Progesterone ≥ 2ng/mL before ovulation.
5. Premenstrual spotting (defined as ≤ 2 consecutive
days of menstrual bleeding).
Data on potential
confounders were
collected by
anthropometric
measurement (height
and weight),
questionnaires (lifestyle,
physical activity, and
reproductive history), or
interview (demographic
and dietary information).

Descriptive statistics by LPD for all 3
definitions were computed for potential
confounding variables and menstrual cycle
characteristics (i.e., overall cycle length,
premenstrual spotting, and geometric mean
hormone levels) taking into account multiple
cycles from the same women.



Nonlinear harmonic models were used to
assess the association between LPD and
hormonal patterns.
All models were adjusted for age, BMI, race.
We used centered models for mean and
amplitude and non-centered models for
phase shift assessment.
Luteal phase
< 10 days
Luteal
progesterone
<5ng/mL
Lowest
quintile of
luteal AUC
progesterone
(<54.5 ng/mL)
n=39 cycles
n=35 cycles
n=82 cycles
n=16 cycles
Luteal
phase P
<5 ng/mL
n=19
Luteal
phase
length <10
days
n=6 cycles
n=328 cycles
Luteal
phase
AUC P
<54.5
ng/mL
n=30
n=17 cycles

Luteal phase < 10 days
 Significantly (P<0.05) associated with white race;
trend (P<0.10) towards younger age, increased fiber
intake, and high physical activity.

Luteal phase progesterone <5ng/mL
 Significantly (P<0.05) associated with younger age
and decreased sleep; trend (P<0.10) towards never
having been pregnant and increased fiber intake.

Luteal phase AUC progesterone <54.5 ng/mL
 Significantly (P<0.05) associated with younger age
and never having been pregnant; trend toward
decreased alcohol intake.
Estradiol
Luteal phase < 10 days
Luteal AUC progesterone < 54.5 ng/mL
Luteal progesterone < 5 ng/mL
Mean
P=0.008
LH
Luteal phase < 10 days
Luteal AUC progesterone < 54.5
Luteal progesterone < 5 ng/mL
Amplitude
P=0.12
FSH
Luteal phase < 10 days
Luteal AUC progesterone < 54.5
Luteal progesterone < 5 ng/mL
Mean
P<0.001
Progesterone
Luteal phase < 10 days
Mean
P=0.02
Amplitude
P=0.03



Findings suggest that normal cycling women
do express evidence of subtle menstrual cycle
disorders manifested as LPD.
Age, race, sleep, and dietary factors (e.g.,
fiber intake) may be associated with LPD in
normal cycling women.
LPD may express itself in different ways,
potentially through different mechanisms.



Decreased E2 across the cycle significantly
associated with deficient luteal progesterone
production.
Increased FSH significantly associated with
short luteal phase duration and deficient
luteal progesterone production.
Decreased progesterone significantly
associated with short luteal phase duration.



Need for a better diagnostic test to assess
luteal phase insufficiency.
LPD, an independent entity causing
infertility, has not been proven.
While we show here that LPD is present in
normal cycling women, we cannot infer if it is
pathological in regards to fertility.

Further research is needed to assess what
factors influence LPD, so that preventive
strategies via lifestyle modifications or
minimally invasive therapeutic treatments
might be implemented for women suffering
from reproductive disorders due to LPD.