| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Center for Fertility and Reproductive Endocrinology (M.J.D.S., B.E.M., A.A.L.), New Britain General Hospital, New Britain, Connecticut 06050; the Department of Biological Sciences, Ohio University (A.B.L., C.G.C.), Athens, Ohio 45701; the Institute for Toxicology and Environmental Health, University of California (B.L.S.), Davis, California 95616; and the Division of Health Professions, University of Hartford (L.S.P.), West Hartford, Connecticut 06117
Address all correspondence and requests for reprints to: Mary Jane De Souza, Ph.D., FACSM, Center for Fertility and Reproductive Endocrinology, New Britain General Hospital, New Britain, Connecticut 06050.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The prevalence of amenorrhea has been variously reported from 5–46% in runners (6, 7, 8, 9, 10) to 37–44% in ballet dancers (11, 12, 13). No data are available on the prevalence or incidence of ovulatory disturbances and LPD in either competitive or recreational athletes, but as the majority of physically active women participate in either recreational or moderate levels of physical activity, it is especially important to determine the nature and frequency of menstrual phase abnormalities in that population. As recently acknowledged, scientists cannot define the scope of the potential problem (the number of women affected) or the depth of concern warranted (the health risks) for these active women without this information (14).
Attempts to characterize the actual types of LPD that occur in athletes have been few and limited to single menstrual cycle evaluations (1, 3, 15). To quantify the prevalence and incidence of ovulatory status and LPD, daily hormonal measurements of LH and progesterone in the blood or urine are necessary (16, 17). As these techniques are very time consuming and costly, it is not surprising that the documentation and characterization of LPD in athletes has been incomplete (1, 3, 4, 15). In the only attempt at a prospective observational evaluation, 20–27% of moderate and long distance runners experienced one or more episodes of anovulation during two menstrual cycles monitored at the beginning and end of a 12-month period (15), but these data are questionable because the menstrual cycles were minimally monitored. A few cross-sectional evaluations of a single menstrual cycle have revealed evidence of ovulatory disturbances and LPD in both high mileage (18) and recreational runners (3, 4, 19). As menstrual cycle status may not be consistent from cycle to cycle, detailed prospective hormonal evaluations of several menstrual cycles are necessary to estimate the incidence as well as prevalence of ovulatory disturbances. Furthermore, it is important to monitor factors that have been hypothesized to disrupt reproductive function in active women (i.e. exercise stress and energy availability).
The transduction of adverse effects at the level of the hypothalamic GnRH pulse generator to the ovary has been attributed to changes in LH pulsatility (18, 20, 21). To date, there have been no prospective observational daily evaluations of reproductive hormones, particularly FSH, across several consecutive menstrual cycles in exercising women. The importance of FSH secretion during the luteal-follicular transition on the recruitment and development of a cohort of follicles, including the dominant follicle, has been emphasized (22, 23). Alterations in daily FSH secretion during the luteal-follicular transition may impact significantly on the subsequent recruitment and maturation of tertiary follicles (22, 23, 24).
The purposes of this study were to evaluate the daily ovarian steroid and urinary FSH excretion characteristics in a small sample of regularly menstruating women participating in recreational to moderate levels of exercise and to determine the frequency (i.e. the prevalence and incidence) of LPD and anovulatory cycles in the sample during three consecutive menstrual cycles. We also evaluated the relationship of the observed menstrual phase alterations with specific exercise stress and energy availability characteristics.
Of course, frequencies in a small sample may not precisely estimate the true prevalence and incidence in the general population of regularly menstruating women runners, but they can indicate whether there is reason to be concerned. Furthermore, because we restricted our sample to regularly menstruating runners, and because the frequency of irregular menses is high in women runners, the frequencies observed in this study probably underestimate the actual prevalence and incidence of LPD and anovulation in the general population of all women runners.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Forty-six women met the following general eligibility criteria. They 1) were between the ages of 18–36 yr; 2) were in good health, as determined by a medical examination, including a normal Papanicolaou smear within the past year; 3) were free of any chronic disease including hyperprolactinemia and thyroid disease; and 4) had menstrual cycle lengths of 24–36 days. They also had 5) not experienced any recent change in menstrual status within 12 months, 6) an appropriate activity history, 7) not taken any form of hormonal therapy for at least 12 months, 8) no history of an eating disorder or depressive illness within the past 3 yr, 9) no contraindications that might preclude participation in the study, and 10) not taken any medication that would interfere in calcium metabolism within the past 3 yr. Women with oligomenorrhea (defined as irregular menstrual cycles occurring every 39–90 days) and primary or secondary amenorrhea were excluded from this study. Specific admission criteria outlined below were met by 35 of the 46 subjects screened, and these women were included in this study.
Subject categorization
The sample population consisted of a cohort of women who were initially classified as either 1) sedentary eumenorrheic women, performing no more than 1 h of aerobic activity per week for the past 12 months, with a peak oxygen uptake (VO2 peak) less than 35 mL/kg·min; or 2) recreational runners, running at least 2 h or 16 km/wk for the past 12 months, with a VO2 peak greater than 35 mL/kg·min. Each subject completed a maximal exercise test to exhaustion to document peak aerobic power (VO2peak) and a body fat analysis.
All subjects were subsequently classified after a three-cycle prospective evaluation of training status (determined by analysis of training diaries) and menstrual status (determined by hormonal evaluations) during the 3-month study period. For training status, two general categories were used: sedentary and exercising. For menstrual status, three general categories were used: ovulatory, LPD, and anovulatory, as defined below.
Training status
Training activities were recorded daily. Factors recorded on the training cards included distance and duration of each run. Heart rate by radial palpation (15 s x 4) was also recorded after each run. Any other physical activity performed for 3 or more sustained min was also recorded on the training record with an associated heart rate (15 s x 4). Training cards were monitored weekly and collected monthly. Exercise training volume was defined as the actual number of kilometers run per week, as recorded on the training logs. Exercise training hours were defined as the number of hours run per week plus the hours per week of other weight-bearing physical activity. The average exercise heart rate during each menstrual cycle served as an index of exercise intensity.
Dietary records and activity monitoring
Seven-day prospective nutritional diaries were completed every month during the study period during the early follicular phase (days 2–9). Subjects were educated on techniques to estimate food portion sizes with the use of food models and were instructed to record all food and liquid consumed throughout the week days and both weekend days. All diet records were analyzed by the same nutritionist using the Nutritionist IV Program (San Bruno, CA). During these 7-day periods, a Caltrac accelerometer (Caltronics, Vienna, VA) was worn on the nondominant hip to measure activity. The Caltrac accelerometer is a motion sensor that weighs 400 g, records vertical acceleration, and estimates energy expenditure. The 7 days of energy expenditure assessments were performed during the early follicular phase (week 1) and the luteal phase (week 3), and a cycle mean was calculated. Because there were no significant differences noted between estimated energy expenditure during weeks 1 and 3, the cycle mean was used for all analyses. The dietary intake of the menstrual cycle categories was assessed for the same 7 days corresponding to the days of energy expenditure assessments during week 1. Heart rate (15 s x 4) was also recorded upon awakening while in bed, at midday, and at midevening during these 7-day monitoring periods.
Weight, menstrual patterns, nutritional and training habits, and any unusual stressors that might affect ovulatory function were monitored throughout this study. Subjects were weighed at each weekly visit.
Menstrual categorization
Menstrual calendars were used 1 month before study and for the duration of the study to record the first and the last day of menses for each cycle. Women who had cycles less than 20 days or greater than 38 days during the 3-month monitoring period were excluded from the data analysis. Three women in the exercising group were excluded from the analyses, one for pregnancy, one for a 19-day cycle followed by a 39-day cycle, and one for a 42-day cycle. For six women in the exercising group, data were collected for only two cycles; three of the women had consistent cycles that were ovulatory, one had consistent cycles that were anovulatory, one woman had consistently LPD cycles, and two women had inconsistent cycles with one ovulatory cycle and one LPD cycle [for group comparisons, these two women were included in the exercising LPD (ExLPD) group]. For two women in the sedentary group, data were collected for only two cycles, both were consistent and ovulatory.
Experimental protocol
Volunteers were initially interviewed by phone. If eligible, they reported to the office of the principal investigator to read and sign the informed consent form, which was approved by the institutional review boards at the University of Connecticut Health Center and New Britain General Hospital.
Subjects were then required to maintain a menstrual record and to collect timed 8-h urine samples beginning on day 2, 3, or 4 of the menstrual cycle (day 1 was defined as the first day of menstrual bleeding) until the onset of the next menses every day for two or three consecutive study cycles. Timed 8-h urine samples were initiated each night upon retiring, continued throughout the night, and terminated with the first morning void. The date and time of sample collection were recorded. Subjects were provided with a toilet-type urine catch kit with prelabeled urine containers for each study cycle. All subjects were required to place an aliquot (10 mL) of each urine sample in a prelabeled tube and to store the tubes in a refrigerator. Samples were delivered to the laboratory on a weekly basis. At this visit, urine collection tubes were given to the subject for the next week. Daily urine samples were analyzed for creatinine, total FSH, LH, pregnanediol-3-glucuronide (PdG), and estrone conjugates (E1C).
Determination of menstrual phase characteristics
Ovulatory status was determined for all subjects using the following criteria. The day of the LH surge was identified by the LH peak and by the concurrence of the day of or the day after the midcycle E1C peak. There is a delay between the plasma and urinary peaks of LH, and the urinary peak was used as the day of the LH surge, because the urinary peak is temporally closer than the plasma peak to the actual release of the oocyte (1). The menstrual cycle length was defined as the number of days from day 1 of menses to the day before day 1 of the next menses. The follicular phase length was defined as the number of days from day 1 of menses up to and including the day of the LH surge. The luteal phase length was defined as the difference between the cycle length and the follicular phase length. LPD was defined as short when a luteal phase length was less than 10 days or inadequate when peak PdG excretion was less than 3 µg/mg creatinine (Cr) for 3 or more midluteal phase days (25, 26). An anovulatory cycle was defined as a cycle in which no increase in E1C was observed in concurrence with a failure of LH to rise at midcycle.
Some data were analyzed by subject groups and others by individual menstrual cycle categories. Groups and individual menstrual cycles were classified as 1) sedentary ovulatory (SedOvul), 2) exercising ovulatory (ExOvul), 3) exercising LPD (ExLPD), and 4) exercising anovulatory (ExAnov). For group comparisons, women were classified according to the menstrual cycle category they displayed in at least two of the three cycles observed. All hormonal data were analyzed by cycle category.
The urinary metabolites, E1C and PdG, were compared among the cycles by the integrated area under the curve (AUC) and by the method of mean steroid levels recently described by Winters et al. (25). E1C excretion was assessed during three different periods; 1) during follicular recruitment, defined as days 2–5; 2) during growth and development of steroidogenically active follicles, defined as days 6–12; and 3) during corpus luteum activity, defined as the day after the LH surge to the day before the next menses.
PdG excretion was assessed during two different periods: 1) during adrenal excretion, defined as days 6–10; and 2) during corpus luteum activity, defined as the day after the LH surge to the day before the onset of the next menses. Luteal phase adequacy was further examined using several methods, as modified from the method of Jordon et al. (26), including the sum of the 3-day midluteal peak PdG (sum of midluteal peak PdG ± 1 day) and the peak PdG.
Urinary analyses
The validity of the urinary techniques as representative of the 24-h pattern of excretion of LH, E1C, and PdG has been presented by other investigators (1, 25, 27). Refrigerated 10-mL aliquots of urine were delivered to the laboratory, aliquoted into polyethylene tubes, and frozen at -80 C. Samples for urinary LH were analyzed by RIA prepared in a 1:2 dilution (Diagnostic Products Corp., Los Angeles, CA). The sensitivity of the assay was 1.2 mIU/L. The inter- and intraassay variances were less than 3.4% and less than 4.7%, respectively. RIA determination of urinary LH was performed in duplicate at the Reproductive Endocrinology Laboratory at New Britain General Hospital (New Britain, CT).
Urine samples were analyzed for E1C and PdG by enzyme immunoassay as described by Munro et al. (27). E1C and PdG were both indexed to Cr excretion in the same sample to control for variations in urine volume. E1C and PdG are expressed as nanograms and micrograms per mg Cr, respectively. Urine samples in which the Cr level was less than 0.2 mg/mL were considered too dilute to yield accurate measurements; levels for these measurements were treated as missing values. The sensitivity of the E1C and PdG assays were 7.8 ng/mL and 0.15 µg/mL, respectively. Values below the sensitivity were reported at the minimum detection limits. The interassay coefficients of variation for high and low internal controls in 111 individual assays were 14.7% and 13.1% for E1C and 15.6% and 12.9% for PdG, respectively. All E1C and PdG assays were performed in duplicate at the Institute for Toxicology and Environmental Health, University of California-Davis.
Urine samples were analyzed for FSH by enzyme immunoassay as described by Qiu et al. (28). FSH was also indexed to Cr excretion in the same sample to control for variations in urine volume. FSH is expressed as nanograms per mg Cr. Urine samples in which the Cr level was less than 0.2 mg/mL were also considered too dilute to yield accurate measurements, and these measurements, too, were treated as missing values. The sensitivity of the assay, calculated at 3 SD above the zero standard, was 0.32 ng/mL. Values below the sensitivity were reported at the minimum detection limits. The intraassay coefficients of variation for FSH were 3.0% and 2.3% for the 1.3 and 4.5 ng/mL internal controls, respectively (28). All assays were performed in duplicate at the Institute for Toxicology and Environmental Health, University of California-Davis.
Peak exercise testing
The VO2 peak was determined by measurement of expired metabolic gases during a progressive treadmill test to volitional exhaustion. The treadmill test was a continuous graded test that was modified according to each subjects’ training history, and the VO2 peak was determined by previously published methods and criteria (5). During the test, the subjects breathed continuously through a Hans Rudolph valve and corrugated plastic tubing connected to a pneumotach. Expired air samples were measured using an on-line Medical Graphics Exercise System 2000 (Medical Graphics, St. Paul, MN).
Body composition testing
Body fat was determined via skinfold measurements (29) at various sites, including the tricep, subscapula, illiac crest, abdomen and thigh with a constant pressure skinfold caliper (Holtain, Ltd., UK). All measurements were made in triplicate on the right side of the body by one investigator (test-retest, r = 0.96).
Energy balance and energy availability
Energy balance was estimated by subtracting energy expenditure (estimated by 24-h Caltrac monitoring) from dietary energy intake. Energy availability was estimated by subtracting exercise energy expenditure from dietary energy intake (30). Exercise energy expenditure was estimated by evaluating activity logs. The energy cost for each specific physical activity was calculated by multiplying the estimated caloric value per minute by the number of minutes engaged in a specific activity (31). Energy expenditure due to exercise was accumulated over the 24-h period for each cycle on each of the 7 days observed. The average energy expenditure was then used as the daily exercise energy expenditure. All energy parameters were analyzed both in absolute quantities and normalized by body weight.
Statistical methods
All demographic data among the four subject groups were analyzed
via a one-way ANOVA. All training logs, menstrual cycle
characteristics, and hormonal parameters were also analyzed via a
one-way ANOVA. In an effort not to violate the assumption of
independence on the grouping fixed factor, subjects with three varying
menstrual cycles were placed in a single menstrual category where the
most frequent category was observed. The third differing cycle was
deleted from the analysis. This resulted in the deletion of 12
menstrual cycles. For the urinary hormonal data, the AUC was calculated
by the trapezoidal method. ANOVA was then performed on the AUC. A
significance level of 0.05 was used to detect the presence of
significant differences. Post-hoc analyses (least
significant squares) were performed where significant F ratios were
found. Contingency tests (
2) were performed on
categorical menstrual cycle data. Tests of monotone alternatives (32)
were used when the direction of the monotonic progression of values was
justified by previous information (i.e. peak LH).
As the direction of interest in group differences was known in advance from previous research, all post-hoc statistical comparisons were one sided. In planned post-hoc comparisons of sedentary and exercising women, the number of observations in the study provided more than a 99.95% probability of detecting differences larger than 2 SD (i.e. outside the so-called normal range) and more than an 85% probability of detecting differences larger than 1 SD (i.e. beyond the central 68% of the normal range), with a tolerable type I error rate of 5%.
In planned post-hoc comparisons of ovulatory cycles in sedentary and exercising women and of ovulatory and LPD cycles in exercising women, the number of observations in this study provided more than a 99.95% probability of detecting differences larger than 2 SD (i.e. outside the so-called normal range) and an 80% probability of detecting differences larger than 0.8 SD with a tolerable type I error rate of 5%. In planned post-hoc comparisons of ovulatory cycles or LPD cycles to anovulatory cycles, the number of observations in this study provided more than a 99% probability of detecting differences larger than 2 SD and an 80% probability of detecting differences larger than 1.1 SD, with a tolerable type I error rate of 5%.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Demographic characteristics. The demographic characteristics
of the study participants are presented in Table 1
. Thirty-five women participated in this
prospective evaluation of menstrual status. Eleven women were
sedentary, and 24 women were runners performing exercise volumes of
approximately 32.4 km/wk. The sedentary (SedOvul) and exercising
subject groups were similar with respect to age, height, and weight.
Within the exercising group, there were no differences among the
predominately ovulatory (ExOvul), luteal deficient (ExLPD), and
anovulatory (ExAnov) subjects with respect to demographic,
reproductive, and training characteristics. Therefore, these data on
exercising women were collapsed across menstrual categorization. The
sedentary subjects had significantly more (P < 0.001)
fat mass and a higher percentage of body fat than the exercising group.
The body weight of the subjects did not differ (P >
0.05) from week to week across the menstrual cycles monitored (data not
shown). The subject groups were similar with respect to reproductive
characteristics, including age of menarche and reproductive maturity
(Table 1).
|
), including distance run and time spent in other
weight-bearing physical activity per week (stairmaster,
walking/running, resistance training, hiking, tennis, and racquetball).
The volume of exercise performed by these subjects is characteristic of
moderate or recreational running. As expected, these training
parameters as well as aerobic capacity and resting heart rate were
significantly different (P < 0.05) from those of the
sedentary subjects and were representative of a moderately
exercise-trained population.
Frequency and sample prevalence and incidence. Subjects were
categorized according to the predominant type of menstrual cycle as
presented in Table 2A. Among the
sedentary women, 91% (10 of 11) were ovulatory and 9% (1 of 11)
demonstrated LPD (short luteal phase). Among the exercising women, 42%
(10 of 24) were ovulatory, 42% (10 of 24) demonstrated LPD (short
and/or inadequate luteal phase), and 16% (4 of 24) were anovulatory. A
2 analysis revealed that the frequency of LPD and
anovulation was much greater (P < 0.006) in the
exercising women than in the sedentary women. Menstrual cycle status
from cycle to cycle was very consistent in the sedentary women and very
inconsistent in the exercising women. Among the sedentary women, 91%
(10 of 11) were ovulatory during all 3 of the cycles observed, whereas
only 54% (13 of 24) of the exercising women were either consistently
ovulatory (21%, 5 of 24) or consistently demonstrating LPD or
anovulation (33%, 8 of 24). In addition to the 33% of runners whose
menstrual cycles were consistently abnormal, another 46% (11 of 24)
displayed inconsistent menstrual cycle status from cycle to cycle. A
2 analysis again revealed that the frequency of
inconsistent LPD and anovulation was much greater (P <
0.008) in the exercising women than in the sedentary women.
|
Cycle comparisons
Overall comparison by menstrual cycle status. In the sedentary
women, 28 of 31 (90%) menstrual cycles monitored were ovulatory,
whereas only 30 of 66 (45%) menstrual cycles monitored were ovulatory
in the exercising women. One sedentary woman had short luteal phases
during all 3 of the menstrual cycles monitored. Among the exercising
women, 55% of the 66 menstrual cycles monitored demonstrated either
LPD (43%; n = 28) or anovulation (12%; n = 8). Among the
LPD cycles, 47% (n = 17) displayed short luteal phases, 14%
(n = 5) were inadequate (low progesterone
concentrations), and 25% (n = 9) were short and inadequate. These
data are shown in Table 2b. Characteristic menstrual cycles in
consistently ovulatory and LPD exercising women are shown in Fig. 1. Characteristic inconsistent
consecutive menstrual cycles of an exercising woman are also shown in
Fig. 1.
|
Menstrual cycle parameters are presented in Table 3. The ExOvul and SedOvul cycles were
similar for all menstrual parameters presented in Table 3. The lengths
of the SedOvul, ExOvul, ExLPD, and ExAnov cycles were similar
(P = 0.12). The ExLPD cycles had significantly longer
follicular phases (17.9 ± 0.7 days) compared to the ExOvul
(14.8 ± 0.9 days) and SedOvul (15.9 ± 0.6 days) cycles. As
anticipated, luteal phase length was significantly shorter in the ExLPD
cycles (8.2 ± 0.5 days). The ExAnov cycles were excluded from the
statistical analyses on luteal and follicular phase lengths.
|
). During days
6–12, E1C excretion was significantly less in the ExLPD and ExAnov
cycles than in the ExOvul and SedOvul cycles. Over the entire
follicular and luteal phases, E1C excretion did not differ among the
SedOvul, ExOvul, or ExLPD cycles, but less was excreted during the
ExAnov cycles (P < 0.05). For the ExAnov cycles, the
follicular phase was assumed to be the first half of the menstrual
cycle, as a true follicular phase could not be discerned. The AUC for
E1C (P < 0.001) during the follicular phase was lower
in the ExAnov cycles than in the other cycle categories. During the
luteal phase, however, the E1C AUC was lower (P <
0.004) in both the ExLPD and ExAnov cycles than in the ovulatory
cycles. Figure 2 displays the E1C
excretion of the groups across the menstrual cycle.
|
|
.
For the ExAnov cycles, the luteal phase was assumed to be the second
half of the menstrual cycle, as a true luteal phase could not be
discerned. In the ExLPD and ExAnov cycles, significantly less
progesterone was excreted than in the SedOvul and ExOvul
cycles, as assessed by PdG AUC during the luteal phase and other
measures of PdG excretion, including mean luteal phase PdG, sum of the
3-day midluteal PdG peak, and peak PdG. Moreover, for all parameters of
PdG excretion, except for follicular phase PdG during days 6–10, there
was significantly less PdG excreted during the ExAnov cycles than
during the ExLPD cycles. Even the ExOvul cycles demonstrated some
suppression of corpus luteal progesterone excretion
compared to the SedOvul cycles (P < 0.05). The ExOvul
cycles had significantly lower progesterone excretion
during the luteal phase as assessed by all parameters of
progesterone excretion. Figure 3 displays PdG excretion of the groups
across the menstrual cycle.
|
|
. Mean FSH during the
last 5 days of the menstrual cycle differed (P < 0.05)
by cycle categories. The ExLPD cycles excreted significantly less FSH
than the SedOvul and ExOvul cycles (P < 0.007).
Paradoxically, FSH excretion during the last 5 days of the menstrual
cycle was significantly greater in the ExAnov cycles than in the
SedOvul, ExOvul, and ExLPD cycles. Urinary FSH quantified during
days 2–5 and days 6–12 of the menstrual cycle was similar
(P > 0.05) among the cycle categories.
|
Energy expenditure. In the cycles categorized as ExAnov,
subjects expended significantly more energy per 24-h day than in the
cycles categorized as ExOvul and ExLPD. However, when normalized by
body weight, the ExAnov cycles were similar to the ExOvul and ExLPD
cycles. As anticipated, in cycles categorized as SedOvul, subjects
expended significantly less energy in exercise compared to cycles
categorized as ExOvul, ExLPD, and ExAnov (Table 7).
|
. In cycles
categorized as ExAnov, dietary energy intake and dietary energy
normalized for body weight were less than those in other cycle
categories. There were no significant differences in protein energy
intake among the menstrual cycle categories, but differences
(P < 0.05) in fat and carbohydrate energy intake were
noted (see Table 7).
Energy availability and energy balance. There were no
differences (P > 0.05) observed in energy balance when
comparing the SedOvul and ExLPD cycles (Table 7), but the ExAnov cycles
did display a substantial deficit (P < 0.001) in
energy balance (-16.1 ± 2.2 Cal/kg BW·day) compared to the
SedOvul (-0.8 ± 0.8 Cal/kg BW·day), ExOvul (-3.9 ±
1.7 Cal/kg BW·day), and ExLPD (-1.4 ± 1.4 Cal/kg BW·day)
cycles. The ExOvul cycles displayed a greater (P <
0.05) deficit in energy balance than the SedOvul cycles. Energy
availability was lower (P < 0.05) in the cycles
categorized as ExOvul, ExLPD, and ExAnov than in the cycles categorized
as SedOvul, both when availability was expressed as total energy per
day and when it was normalized by body weight. In addition, cycles
categorized as ExAnov were significantly lower in energy availability
normalized by body weight than in the other menstrual cycle
categories.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This is the first detailed prospective observational study to document the frequency of luteal and ovulatory abnormalities in consecutive, asymptomatic, regular menstrual cycles of normal length in moderately active women. The exercising women in this study averaged 32 km/wk of running, a volume representative of a moderate or recreational level exercise training. None of the runners in our study trained excessively or was amenorrheic. This study clearly showed that menstrual cycle length is not an accurate marker of ovarian function in this population. Even though all the women in this study presented with repeatable menstrual cycle lengths in the normal range, their ovarian function was highly variable and frequently abnormal. We found the sample prevalence and sample incidence of abnormal ovarian function in our subjects to be 48% and 79%, respectively. Such high frequencies in asymptomatic women are remarkable. Investigators need to carefully consider this issue when designing studies in which ovarian function is important.
Another compelling finding of this study is the high frequency of inconsistent menstrual status; almost half (46%) of the exercising women presented with inconsistent menstrual status from cycle to cycle. By contrast, none of the sedentary women presented with inconsistent menstrual status from cycle to cycle; that is, they were either consistently ovulatory (91%) or consistently LPD (9%). The monitoring of only one menstrual cycle in exercising women underestimates the frequency of ovarian disturbances. On the average, studies examining a single menstrual cycle will observe only one third of the inconsistently abnormal cycles. As abnormal ovarian function was displayed by 33% of our subjects consistently and by another 46% inconsistently, earlier studies of only a single menstrual cycle (3, 25) probably underestimated the 3-month sample incidence of LPD among regularly menstruating athletes by about 100[(33% + 46%) - (33%+[46%/3])]/(33% + 46%) = 38%. Therefore, investigators of ovarian function or parameters, such as infertility, that are affected by ovarian function may miss or confound important findings if they classify exercising women by the menstrual status observed during a single cycle.
The only other attempt at a prospective observational evaluation of menstrual cycles in athletes reported that 20–27% of moderate and long distance runners experienced one or more episodes of an anovulatory cycle among two menstrual cycles monitored at the beginning and end of a 12-month period (15). As has been discussed in detail previously (33), that study did not differentiate between LPD and anovulatory cycles, due to questionable methodologies used for menstrual categorization. The accuracy in the estimation of anovulatory frequency is suspect because the menstrual cycles were assessed by pooling only a single hormonal measurement during the follicular and another during the luteal phase. This method was probably not representative of actual ovarian function, and it probably grossly underestimated the incidence of LPD. By contrast, our investigation carefully monitored daily urinary hormone levels, including the direct documentation of a LH surge, to assess ovarian function during each menstrual cycle.
Other investigators (3, 18, 19, 25) have reported the occurrence of LPD in cross-sectional evaluations of one menstrual cycle in exercising women. Loucks et al. (18) found shorter luteal phases with reduced progesterone concentrations in regularly menstruating competitive runners and triathletes compared to sedentary women. Ellison (19) found similar results in recreational runners who ran only 12 miles/week. Broocks et al. (3) observed luteal phase inadequacies in 7 of 17 recreational joggers who ran 12–18 miles/wk. These women showed evidence of disturbed folliculogenesis (determined via transvaginal follicular monitoring) and luteal phase inadequacy (determined via daily progesterone levels). Another cross-sectional evaluation of daily E1C and PdG (25) reported that 4 of 10 high mileage (42.4 miles/wk) runners and 5 of 10 active (12.3 miles/wk) women displayed anovulatory cycles or luteal phase abnormalities. Thus, our multicycle data confirm and strengthen the earlier single cycle findings of widespread luteal and ovulatory abnormalities among entirely asymptomatic, regularly menstruating women who engage in only moderate levels of physical activity.
Hormonal characteristics
Before this study, the transduction of the adverse effects of exercise training at the level of the hypothalamic GnRH pulse generator has been attributed to changes in LH pulsatility (18, 20, 21). In this study, we discovered a second gonadotropin abnormality that may act in concert with the disruption of LH pulsatility to impair ovarian function in exercising women. Like LH, alterations in FSH secretion may also significantly impact ovarian function by altering folliculogenesis (22, 23, 24, 34). Typically, elevations in FSH are observed during the luteal-follicular transition and at midcycle, and declines are usually observed during the late follicular and early luteal phases (22, 23, 35). We found that the characteristic rise in FSH release during the luteal-follicular transition appears to be blunted in exercising women with LPD. This is the first report of an abnormality in the monthly pattern of FSH excretion in such women.
FSH characteristics during the luteal-follicular transition in clinical presentations of LPD have not yet been described, but a blunted elevation in FSH during the luteal-follicular transition may in part explain the decreased estradiol excretion that we observed during the early to midfollicular phase of cycles with ovarian abnormalities. Recently, a delay in the rise in urinary FSH during the midluteal phase of conceptive cycles compared to nonconceptive cycles has been described and has permitted the detection of pregnancy 10 days after the midcycle LH surge (36).
The characteristic rise in FSH during the luteal-follicular transition is strongly related to the preceding LH surge, indicating a high degree of neuroendocrine hypothalamic regulation during this critical transition period of the menstrual cycle (23). GnRH pulse frequency may play a key role in the etiology of the rise of FSH during the luteal-follicular transition and may be critical to normal folliculogenesis (23).
A progressive decline in peak LH along the continuum from ovulatory cycles to anovulatory cycles in this study resembled similar findings reported in infertile women with LPD (37). Luteal phase defects in 63 normally cycling infertile women were associated with smaller dominant follicles and a lower midcycle ovulatory LH surge compared to women without evidence of LPD (37).
The exercising women in this study also exhibited a progressive suppression of estradiol excretion during the follicular phase, in that E1C on days 2–5 was lower in all exercise groups compared to that in the sedentary group and remained low in the ExLPD and ExAnov women on days 6–12 compared to that in the SedOvul and ExOvul groups. Days 2–5 are reflective of follicular recruitment and increased estradiol excretion (25). A delay in follicle recruitment probably explains the delay in estradiol excretion and the long follicular phases (>18 days) and short luteal phases (<10 days) in our exercising LPD women. The delay in estradiol excretion during the window of days 6–12 probably reflects a delay in the growth and maturation of steroidogenically active follicles and a subsequent delay in follicular dominance and subsequent ovulation (25). Winters et al. (25) also reported lower E1C excretion during the early follicular phase in runners. Thus, the suppressed follicular phase levels of estradiol observed in this study suggest a progressive suppression in follicular maturation as menstrual dysfunction progresses along the continuum from ovulatory to anovulatory cycles.
The characteristics most significantly differentiating the disturbed ovarian function (ExLPD and ExAnov) cycles from the ovulatory (SedOvul and ExOvul) cycles were those associated with the luteal phase. The length of the luteal phase was significantly shorter in association with the delayed follicular maturation and long follicular phases in the ExLPD cycles. Similarly, a monotonic suppression in corpus luteal progesterone production was observed as menstrual dysfunction became more extreme along the continuum from ovulatory to anovulatory cycles. The ExLPD and ExAnov cycles excreted significantly less progesterone than both the sedentary (SedOvul) and exercising (ExOvul) ovulatory cycles. Even the ExOvul cycles demonstrated suppressed corpus luteal progesterone excretion compared to the SedOvul cycles. These findings support previous work by other investigators (1, 2, 3, 18, 19) who found that regularly menstruating athletes displayed shorter luteal phases with decreased progesterone compared to sedentary women. Our data and those of previous investigators (1, 2, 3, 18, 19) support the concept that in exercising women, both luteal phase progesterone excretion and early follicular phase estrogen excretion decrease concomitantly and proportionally. Of course, a cross-sectional comparison such as this cannot distinguish whether ovulatory disturbances progress to an extreme end point (amenorrhea), whether they represent successful acclimations that could be achieved by all women, or whether they are end points in robust women.
Energy intake, expenditure, availability, and balance
The hypothalamic gonadal axis has been proposed to be very sensitive to both stress (38) and energy availability (39), although the mechanisms of these factors have not yet been determined. The role(s) of these two factors in the menstrual disturbances observed in exercising women remains controversial. In this context, energy availability may be usefully defined as dietary energy intake minus exercise energy expenditure (30, 40). This permits the stress of exercise to be independently defined as everything associated with exercise except its energy cost. So defined, exercise stress and energy availability become susceptible to carefully designed experimental investigations of their independent effects on reproductive function (40, 41). By contrast, in observational studies such as this one, the effects observed in subjects (i.e. on progesterone secretion) have already occurred some time in the past, whereas the hypothetical causal factors (e.g. exercise stress and energy availability) are confounded in the habitual behavior of the subjects.
We found no association between menstrual status and the distance our exercising subjects ran or the amount of time they spent in other weight-bearing physical activity. By contrast, we found several associations with energy-related parameters.
Low energy availability has been reported to induce low T3 syndrome (30), to slow LH pulse frequency, and to increase LH pulse amplitude in sedentary women during acute short term experiments in which exercise stress and energy availability were independently controlled (40, 41). In those 4-day experiments in habitually sedentary women, exercise had no suppressive effect on thyroid metabolism or LH pulsatility beyond the impact of its energy cost on energy availability. Energy balance under the low energy availability conditions in those experiments was approximately -22 Cal/kg BW·day.
In our study, we estimated dietary intake from 7-day prospective diet records, total energy expenditure from 24-h Caltrac monitoring, and exercise energy expenditure from an analysis of daily activity logs. By these methods, we were unable to associate differences in exercise energy expenditure with differences in ovarian function, but significant differences in energy intake, total energy expenditure, energy availability, and energy balance were observed. In the ExAnov cycles, subjects expended more energy and also consumed less than in the SedOvul, ExOvul, and ExLPD cycles. This resulted in a lower energy availability normalized by body weight and a more negative energy balance compared to those in the SedOvul, ExOvul, and ExLPD cycles. The energy availability in the ExOvul and ExLPD cycles was intermediate between those in the SedOvul and ExAnov cycles.
In short term studies examining the acute independent effects of exercise stress and energy availability, Loucks et al. were able to prevent the disruption of thyroid metabolism and LH pulsatility in exercising women by supplementing their dietary energy intake (30, 40). We do not know whether dietary supplementation would have prevented disturbances of reproductive function in our exercising women, but our observations are consistent with the hypothesis that energy deficiency is the factor that disrupts reproductive function in exercising women. It should be noted that the sedentary subjects in the studies of Loucks et al. (30, 40, 41) were placed in an acute state of low energy availability for only 4 days, and the independent effects of chronic low energy availability and exercise stress have yet to be determined in prospective experiments.
We can speculate that individuals with an even more severe reduction in energy availability may experience the pathological extreme end point of the menstrual cycle continuum, that of amenorrhea. Despite high levels of physical activity, athletes report energy intakes similar to those of sedentary women (18, 21). Most women involved in exercise programs for health, fitness, and weight control practice moderate dietary restriction and increased energy expenditure as the means to attain their goals. The exercising women in our study are probably typical of this population. The effects of dietary restriction and exercise energy expenditure are probably additive in contributing to the abnormalities we observed. Therefore, more research is needed to determine whether particular degrees of menstrual dysfunction are causally related to particular degrees of energy deficiency, and whether chronic dietary restriction and exercise energy expenditure are, indeed, additive in their effect.
Additional studies are also needed to investigate the potential clinical consequences of ovarian disturbances less extreme than amenorrhea. It has been proposed that women with ovulatory disturbances or LPD may have a lower bone mineral density (15), but recently we reported that LPD associated with decreased progesterone does not appear to adversely affect bone mass, provided normal estradiol status is maintained (33). Other potential health risks of LPD appear to be related to infertility and an increased incidence of tubal ectopic pregnancy, because women with LPD and anovulation are at a greater risk for infertility and recurrent habitual abortion, and these women also have a higher incidence of ectopic pregnancy (42, 43, 44). During the luteal phase, the endometrium is under direct stimulation by progesterone. A rapid decline in progesterone or an inadequate progesterone concentration during this period results in a degenerative endometrium, which is not receptive for implantation of a fertilized ovum or maintenance of early pregnancy (42). And finally, some epidemiological evidence has led to a preliminary hypothesis that there may be an association between a short luteal phase, low progesterone levels, and increased breast cancer risk (45, 46, 47). Given the large numbers affected, more research is necessary to further elucidate the clinical consequences of LPD and anovulation in regularly menstruating exercising women.
|
Footnotes |
|---|
Received April 15, 1998.
Revised July 30, 1998.
Accepted August 20, 1998.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. D. Vescovi, T. E. Hewett, G. D. Myer, and B. T. Zazulak Letter to the Editor * Authors' Response Am. J. Sports Med., November 1, 2008; 36(11): e4 - e5. [Full Text] [PDF]
|
|
|
|
 |
|
 J. W. Hill, J. K. Elmquist, and C. F. Elias Hypothalamic pathways linking energy balance and reproduction Am J Physiol Endocrinol Metab, May 1, 2008; 294(5): E827 - E832. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. J Waugh, J. Polivy, R. Ridout, and G. A Hawker A prospective investigation of the relations among cognitive dietary restraint, subclinical ovulatory disturbances, physical activity, and bone mass in healthy young women Am. J. Clinical Nutrition, December 1, 2007; 86(6): 1791 - 1801. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. M. Olsen, C. J. Bain, S. J. Jordan, C. M. Nagle, A. C. Green, D. C. Whiteman, P. M. Webb, and Australian Cancer Study (Ovarian Cancer) and Austr Recreational Physical Activity and Epithelial Ovarian Cancer: A Case-Control Study, Systematic Review, and Meta-analysis Cancer Epidemiol. Biomarkers Prev., November 1, 2007; 16(11): 2321 - 2330. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. O'Donnell, P. J. Harvey, J. M. Goodman, and M. J. De Souza Long-term estrogen deficiency lowers regional blood flow, resting systolic blood pressure, and heart rate in exercising premenopausal women Am J Physiol Endocrinol Metab, May 1, 2007; 292(5): E1401 - E1409. [Abstract] [Full Text] [PDF]
|
|
|
|
