Abstract
Objective:
Hirsutism, the presence of excess terminal hair in a male pattern, is a clinical marker of androgen excess in women. We used cross-sectional data from a North American preconception cohort study to evaluate the association between menstrual cycle characteristics and hirsutism.
Study Design:
Women aged 21-45 years were recruited to a North American cohort of pregnancy planners. On the baseline questionnaire, participants self-reported menstrual characteristics, which included menstrual regularity, cycle length, bleed length, and bleed heaviness. Participants provided a self-rating of hirsutism in nine distinct body areas using pictograms representing the modified Ferriman-Gallwey (mFG) score. Using their ratings, we calculated total mFG scores and defined hirsutism as mFG scores ≥8. We used log-binomial regression models to estimate prevalence ratios (PRs) for the association between menstrual characteristics and hirsutism assessed at baseline.
Results:
We included 5,542 women in the analytic cohort. Mean mFG score was 4.7, with 21.7% reporting mFG scores ≥8. Compared with women with regular menstrual cycles, irregular cycles were positively associated with mFG ≥8 (PR 1.73, 95% CI 1.56-1.91). Bleed lengths of ≥7 days compared with <3 days also showed a positive association with mFG score ≥8 (PR 1.59, 95% CI 1.16-2.19), as did heavy bleeds (PR 1.42, 95% CI 1.21-1.67) compared with moderate bleeds. Findings remained consistent when restricted to women without a prior diagnosis of polycystic ovary syndrome.
Conclusions:
In a population-based cohort of North American women, menstrual irregularity, increased cycle and bleeds lengths, and heavier menstrual bleeds were associated with self-reported hirsutism.
Keywords: androgen excess, hirsutism, menstrual cycle, menstrual irregularity, preconception cohort
1.1. INTRODUCTION
Hirsutism is defined as the presence of excess coarse, terminal hairs in an androgenic or male pattern distribution, distinct from lanugo hair or hypertrichosis [1]. Clinical classification and quantification of hirsutism is typically conducted via the modified Ferriman-Gallwey (mFG) scoring system. First developed in 1961, the original system examined 11 distinct body areas for quantification of hirsutism [2]. Subsequently, evaluation of the forearms and lower legs was removed to create the mFG scoring system [3]. This version enumerates the remaining nine body areas from zero (no coarse hairs) to four (frankly virile) and are then summed to create a “total mFG” score. While alternate scoring systems have proposed quantifying a reduced number of body areas [4, 5], the total mFG score remains the most commonly utilized tool by medical professionals to clinically evaluate hirsutism[6].
In 1981, Hatch et al. proposed a threshold of total mFG score ≥8 to define clinically-significant hirsutism with scores of 8-15 representing mild, 16-25 for moderate, and >25 for severe hirsutism [7]. In North American and predominantly non-Hispanic White cohorts, this remains a commonly-used threshold for hirsutism [8–10].
Distinct from familial hirsutism, pathologic hirsutism can be a marker of androgen excess. Other features of androgen excess include alopecia, acne, and voice deepening. The predominant disorder of androgen excess that is accompanied by menstrual irregularity is polycystic ovary syndrome (PCOS). This ovulatory disorder can present as a constellation of oligo- or anovulation, clinical or biochemical androgen excess, and/or polycystic ovaries on ultrasound. Studies have used varying diagnostic criteria to define PCOS, and prevalence estimates range from 6-10% in adult women [8, 9, 11–16].
There is scant literature examining the relation of individual menstrual cycle characteristics such as cycle or bleed length to hirsutism itself (or other surrogates of androgen excess). Examining this relation may help identify androgen excess, and capture early markers of thecal and granulosa cell disruption, which is implicated in the etiology of disorders such as PCOS. We used cross-sectional data to assess the association of various menstrual cycle characteristics such as irregularity, cycle and bleed lengths, bleed heaviness and age at menarche with hirsutism, identified by self-assessment of the mFG score.
1.2. MATERIALS AND METHODS
1.2.1. Study population and procedures
Pregnancy Study Online (PRESTO) is an ongoing web-based preconception cohort study of pregnancy planners [17, 18]. Women aged 21 to 45 years, residing in the United States or Canada, attempting pregnancy, and not currently using fertility treatments at the time of enrollment were eligible to participate. Participants were recruited primarily via online advertising. Participants completed an online baseline questionnaire on demographic, lifestyle, reproductive and medical history and were followed for reported conception via bi-monthly follow-up questionnaires for up to one year.
1.2.2. Exclusions
From June 19, 2013 to March 1, 2019, a total of 10,524 women completed the baseline questionnaire. We excluded 3,286 women who enrolled before June 2016 (when mFG scoring was added to the questionnaire). Furthermore, we excluded 1,696 women who could not report their menstrual regularity in the past year because they were using hormonal contraception (see baseline characteristics of those excluded in Supplemental Table 1), yielding a final analytic sample of 5,542 women. The study was approved by the Institutional Review Board of Boston University Medical Campus and all participants provided informed consent online.
1.2.3. Assessment of menstrual cycle characteristics
As part of the baseline questionnaire, women reported specific menstrual cycle characteristics. Participants were asked, “Within the past couple of years, has your menstrual period been regular (regular in a way so you can usually predict about when the next period will start)? Please think about those times you were not using hormonal contraceptives.” Participants reported menstrual cycles as regular or irregular, the number of cycles per year (periods/year) and, in regularly cycling women, cycle length. Cycle and bleed lengths were reported in days. Bleed heaviness was reported as “light” (10 or fewer pads or tampons per period), “moderate” (11-20 pads or tampons per period), “moderate/heavy” (21-30 pads or tampons per period) or “heavy” (≥30 pads or tampons per period). Age at menarche and time to cycle regularity after menarche were reported in whole years [19].
1.2.4. Assessment of modified Ferriman-Gallwey score
Starting in June 2016, the baseline questionnaire included the 9-item mFG score [3]. Patients provided a self-assessment of hirsutism for nine body areas: upper lip, chin, arms, thighs, buttocks, upper back, chest, upper and lower abdomen, based on their “natural state” of hair growth, before use of any hair removal techniques or interventions. Severity of hair growth in each area was scored from 0 to 4 based on reference images accompanying the questionnaire. The nine areas were summed to determine total mFG score, with a threshold of ≥8 used to define clinical hirsutism [2, 20]. Although the instrument has not been validated for self-assessment, prior studies have compared self-assessed FG scores with provider assessments [28 29]. However, because women may engage in depilation, self-assessment of hirsutism is considered clinically useful [1].
1.2.5. Assessment of covariates
Covariate information was collected from the baseline questionnaire on time trying to conceive at study entry, age, race/ethnicity, body mass index (BMI), history of smoking, physical activity, caffeine intake, alcohol intake, last method of birth control, and history of PCOS/thyroid disorders. BMI was calculated as weight (kg) divided by height squared (m2). Total metabolic equivalents (METs) per week were calculated by multiplying the average number of hours per week participating in various activities by metabolic equivalents estimated from the Compendium of Physical Activities [21].
1.2.6. Data analysis methods
We used log-binomial regression models to estimate prevalence ratios (PR) and 95% confidence intervals (CI) to model the association between menstrual cycle characteristics and mFG scores of ≥8 compared with <8. We adjusted for the number of cycles trying to conceive at study entry (months), age (<25, 25-29, 30-34, ≥35 years), BMI (<18, 18-24.9, 25-29.9, 30-34.9, ≥35 kg/m2), physical activity per week (<10, 10-19, 20-39, ≥40 MET hours/week), current smoker (yes vs no), caffeine intake (<100, 100-199, 200-299, ≥300 mg/day), alcohol intake per week (0, 1-6, 7-13, ≥14 drinks/week), and hormonal last method of contraception (yes vs no). In secondary analyses, we modeled total mFG score as a continuous outcome variable and used linear regression models to estimate the difference in mean total mFG scores by menstrual cycle characteristics.
To assess potential effect measure modification, we stratified models by BMI (<30 vs ≥30 kg/m2) and age (<30 vs ≥30 years). These variables have been shown to be important correlates of menstrual characteristics [22, 23]. In additional analyses, we restricted to the largest racial/ethnic subgroup, non-Hispanic White women, to control for racial/ethnic differences in hirsutism. To determine if the association persisted among women without known androgen excess disorders, we also restricted to women without a prior diagnosis of PCOS or thyroid disorders [12].
We used multiple imputation to impute missing data for exposures and covariates [24]. Five datasets were imputed, analyzed individually, and combined across the imputed datasets [25]. The percent of missing values for menstrual cycle characteristics was 0.4% for bleed heaviness, 0.3% for bleed length, 4.8% for cycle length, 0.3% for cycle regularity, and 0.5% for age at menarche and time to cycle regularity. There were no missing values for age. We examined the association between mFG and cycle length (days), as continuous variables, by fitting restricted cubic splines to allow for non-linear associations.
1.3. RESULTS
Baseline characteristics of participants by menstrual characteristics are described in Table 1. The cohort comprised 5,542 women, with a mean age of 30.0 years and median BMI of 29.8 kg/m2, who were predominantly non-Hispanic White (78.0%), and non-smokers (87.3%). The majority reported regular periods (68.0%) with an average cycle length of 28.7 days. Overall, 13.6% women reported a diagnosis of PCOS, 2.7% of diabetes mellitus, and 7.2% of thyroid disorders. Mean age at menarche was 12.3 years, with age of menstrual regularity of 13.7 years on average. Among women with regular cycles, 17% had an mFG score ≥8, compared with 34% of women with irregular cycles. The distribution of the mFG score stratified by cycle regularity is displayed in Figure 1. Baseline characteristics of participants excluded due to recent hormonal contraception use are described in Supplemental Table 1. While participants excluded had a similar age as included participants, they were more likely to be white, non-Hispanic, had a lower BMI, and were less likely to have a PCOS diagnosis and be a current smoker.
Table 1.
Cohort characteristicsa
| Cohort | Menstrual Regularity | Bleed length | Bleed Heaviness | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Irregular | Regular | <3 | 3-4 | 5-6 | ≥7 | Light | Moderate | Moderate/Heavy | Heavy | ||
| Number of women | 5,542 | 1,565 | 3,977 | 212 | 2,560 | 2,265 | 505 | 978 | 2,889 | 1,363 | 312 |
| Age, mean years (SD) | 30.0±4.3 | 29.1±4.2 | 30.3±4.3 | 30.4±4.2 | 30.2±4.3 | 29.8±4.3 | 29.4±4.4 | 30.2±4.2 | 30.0±4.3 | 29.7±4.4 | 29.9±4.6 |
| Attempt time at study entry, months | 3.0±13.0 | 5.0±16.0 | 3.0±11.0 | 4.0±15.0 | 3.0±12.0 | 3.0±12.0 | 5.0±17.0 | 3.0±11.7 | 3.0±12.0 | 3.0±13.8 | 6.0±17.3 |
| White non-Hispanic, % | 78.0 | 74.7 | 79.2 | 80.7 | 78.9 | 78.3 | 69.9 | 80.1 | 78.5 | 76.2 | 72.8 |
| BMI, median kg/m2 (SD) | 29.8± 8.3 | 32.2±9.0 | 28.8±7.8 | 28.2±7.3 | 29.5±8.0 | 29.6±8.4 | 32.8±9.4 | 28.5±8.1 | 29.2±8.2 | 31.2±8.5 | 33.5±9.2 |
| PCOS diagnosis, % | 13.6 | 32.1 | 5.8 | 11.2 | 10.4 | 13.5 | 28.2 | 10.1 | 11.2 | 16.8 | 28.4 |
| Mean waist circumference, inches | 34.6±7.4 | 36.4±8.0 | 33.9±7.0 | 33.4±6.6 | 34.2±7.2 | 34.7±7.3 | 37.0±8.3 | 33.6±6.8 | 34.2±7.4 | 35.7±7.7 | 37.3±8.7 |
| Diabetes diagnosis, % | 2.7 | 4.5 | 2.1 | 1.4 | 2.3 | 2.9 | 4.3 | 2.6 | 2.6 | 3.5 | 4.9 |
| Thyroid diagnosis, % | 7.2 | 7.8 | 7.1 | 8.1 | 7.7 | 6.7 | 6.5 | 9.0 | 6.9 | 6.4 | 8.2 |
| Current cigarette smokers, % | 12.7 | 16.0 | 11.4 | 8.6 | 12.1 | 12.3 | 20.4 | 11.1 | 11.0 | 14.4 | 26.3 |
| Alcohol intake ≥7 drinks/week, % | 11.8 | 9.6 | 12.6 | 12.4 | 12.7 | 11.0 | 9.8 | 11.9 | 12.2 | 11.5 | 9.7 |
| MET horns of physical activity per week, % | |||||||||||
| <10 | 16.4 | 19.0 | 15.4 | 13.1 | 15.1 | 16.9 | 22.4 | 16.3 | 15.7 | 17.5 | 18.0 |
| 10-19 | 21.2 | 24.3 | 20.7 | 18.8 | 20.5 | 23.0 | 22.6 | 18.9 | 21.7 | 23.8 | 20.8 |
| 20-39 | 32.2 | 27.9 | 33.7 | 31.6 | 33.8 | 31.2 | 28.2 | 32.7 | 33.1 | 30.3 | 30.0 |
| ≥40 | 29.7 | 28.8 | 30.1 | 36.5 | 30.7 | 28.8 | 26.9 | 32.1 | 29.5 | 28.4 | 31.3 |
| Caffeine consumption, mean mg/day (SD) | 126.6±118.7 | 115.5±113.6 | 124.5±120.5 | 108.3±102.2 | 123.9±121.3 | 122.3±117.1 | 116.6±118.7 | 121.3±166.2 | 120.9±119.9 | 124.9±143.6 | 120.0±121.2 |
| Menstrual characteristics | |||||||||||
| Years on hormonal contraception, median (SD) | 5.0±4.9 | 4.0±4.7 | 5.0±5.0 | 6.0±5.4 | 5.8±5.0 | 5.0±4.8 | 4.0±4.9 | 6.5±5.1 | 5.0±4.9 | 4.0±4.8 | 4.0±5.1 |
| Age at menarche, mean years (SD) | 12.3±1.7 | 12.4±1.8 | 12.3±1.6 | 12.7±1.7 | 12.3±1.6 | 12.3±1.7 | 12.2±1.9 | 12.4±1.6 | 12.4±1.6 | 12.2±1.7 | 12.0±2.0 |
| Age at menstrual regularity, years (SD) | 13.7±2.5 | 13.4±2.5 | 13.7±2.5 | 14.0±2.6 | 13.7±2.5 | 13.6±2.5 | 13.5±2.8 | 13.8±2.6 | 13.7±2.4 | 13.6±2.7 | 13.4±2.5 |
| Regular periods, % | 68.0 | NA | NA | 65.3 | 70.6 | 69.5 | 48.9 | 65.5 | 71.6 | 65.4 | 52.7 |
| Bleed length, % | |||||||||||
| <3 | 3.8 | 3.7 | 3.7 | 15.1 | 1.9 | 0.6 | 0.0 | ||||
| 3-4 | 46.2 | 36.8 | 50.1 | NA | NA | NA | NA | 65.3 | 51.1 | 28.6 | 17.1 |
| 5-6 | 40.9 | 41.3 | 40.7 | 17.9 | 41.7 | 55.7 | 41.2 | ||||
| ≥7 | 9.1 | 18.2 | 5.4 | 2.0 | 5.3 | 15.0 | 41.7 | ||||
| Average cycle length, mean days | 28.7±2.9 | NA | 28.7±2.9 | 29.1±3.0 | 28.5±2.8 | 28.8±2.7 | 29.3±4.1 | 28.7±3.0 | 28.6±2.9 | 28.8±2.8 | 29.2±3.5 |
| Bleed heaviness, % | |||||||||||
| Light | 17.7 | 18.0 | 17.6 | 69.8 | 24.9 | 7.6 | 3.5 | ||||
| Moderate | 52.1 | 43.1 | 55.9 | 24.8 | 57.8 | 53.3 | 30.0 | NA | NA | NA | NA |
| Moderate/Heavy | 24.6 | 28.7 | 22.8 | 5.4 | 15.3 | 33.4 | 40.8 | ||||
| Heavy | 5.6 | 10.2 | 3.8 | 0.0 | 2.1 | 5.7 | 25.8 | ||||
Means & percentages are age standardized to cohort at baseline. SD=standard deviation.
Fig. 1.
Distribution of mFG score among all women and stratified by cycle regularity
The mean mFG score for the cohort was 4.7, with 21.7% of participants having total mFG scores ≥8 (Table 2). The mFG score distribution varied across ethnic groups (Table 2), with Hispanic women reporting the highest mean mFG score (6.1) and prevalence of mFG scores ≥8 (33.7%), and Asian women (including Asian Indian, Chinese, Japanese, Korean, Vietnamese) reporting the lowest scores (3.9, 19.1% respectively), although within Asian women, Indian women reported the highest mean mFG score (8.0) of all participants.
Table 2.
Race/Ethnicity and modified Ferriman-Gallwey score (mFG)
| No. of women | mFG score, mean | mFG score ≥8, % | |
|---|---|---|---|
| Entire Cohort | 5,542 | 4.7 | 21.7 |
| White/non-Hispanic | 4,320 | 4.5 | 19.7 |
| Hispanic/Latina | 448 | 6.1 | 33.7 |
| Black/non-Hispanic | 364 | 6.0 | 29.7 |
| Asian/non-Hispanica | 84 | 3.9 | 19.1 |
| Asian Indian | 29 | 8.0 | 44.8 |
| Chinese | 31 | 1.5 | 0.0 |
| Other Asian | 24 | 3.0 | 12.5 |
| Mixed race/other race | 326 | 5.0 | 24.5 |
=includes race with less than 10 individuals including Japanese, Korean, Vietnamese, Filipino, Pakistani, Thai, Taiwanese, Mien
Unadjusted and adjusted PRs for associations between menstrual cycle characteristics and mFG scores ≥8 (i.e., hirsutism) are reported in Table 3. Compared with age at menarche of 12-13 years, age <12 years at menarche was slightly associated with mFG ≥8 (PR 1.14, 95% CI: 1.02-1.27). Older ages at menarche were not appreciably associated with mFG ≥8 in the overall analysis (PR 0.98, 95% CI: 0.82-1.17 and PR 1.06, 95% CI: 0.89-1.27 for ages 14 and ≥15 years, respectively). Irregular menstrual cycles were positively associated with hirsutism, compared with regular menstrual cycles (PR 1.73, 95% CI: 1.56-1.91). Compared with cycle lengths <28 days, cycle lengths of 28-35 days (PR 1.18, 95% CI: 1.00-1.39) and ≥36 days (PR 1.60, 95% CI: 1.06-2.42) were positively associated with mFG scores ≥8. PRs for bleed lengths of 3-4 days, 5-6 days, ≥7 days, compared with <3 days, were 1.09 (95% CI: 0.81-1.48), 1.18 (95% CI: 0.87-1.60), and 1.59 (95% CI: 1.16-2.19), respectively. Compared with moderate bleed heaviness, moderate-heavy bleeds showed a positive association with hirsutism (PR 1.18, 95% CI: 1.05-1.32), as did heavy bleeds (PR 1.42, 95% CI: 1.21-1.67). Results were consistent when mFG was modeled as a continuous outcome variable (Supplemental Table 2). We observed similar patterns in analyses restricted to non-Hispanic White women, although results were less precise (Supplemental Table 3).
Table 3.
Unadjusted and Adjusted Prevalence Ratios (PR)a for Selected Menstrual Cycle Factors and modified Ferriman-Gallwey score ≥8 vs <8
| No. of women | Unadjusted | Adjusted | |||
|---|---|---|---|---|---|
| PR | 95% CI | PR | 95% CI | ||
| Age at menarche, years | |||||
| <12 | 1,585 | 1.29 | 1.15, 1.44 | 1.14 | 1.02, 1.27 |
| 12-13 | 2,826 | 1.00 | Ref | 1.00 | Ref |
| 14 | 630 | 0.89 | 0.74, 1.07 | 0.98 | 0.82, 1.17 |
| ≥15 | 501 | 1.03 | 0.86, 1.24 | 1.06 | 0.89, 1.27 |
| Time to menstrual regularity, years | 3,766 | 1.05 | 1.02, 1.08 | 1.04 | 1.01, 1.07 |
| Cycle regularity | |||||
| Regular | 3,977 | 1.00 | Ref | 1.00 | Ref |
| Irregular | 1,565 | 2.01 | 1.82, 2.21 | 1.73 | 1.56, 1.91 |
| Number of cycles (periods/year) | 5,542 | 0.92 | 0.90, 0.93 | 0.94 | 0.93, 0.95 |
| Cycle length, daysb | 3,977 | 1.02 | 1.00, 1.05 | 1.02 | 1.00, 1.04 |
| Cycle length, daysb | |||||
| <28 | 1,087 | 1.00 | Ref | 1.00 | Ref |
| 28-35 | 2,816 | 1.17 | 0.99, 1.38 | 1.18 | 1.00, 1.39 |
| ≥36 | 74 | 1.73 | 1.14, 2.62 | 1.60 | 1.06, 2.42 |
| Bleed length, days | |||||
| <3 | 212 | 1.00 | Ref | 1.00 | Ref |
| 3-4 | 2,560 | 1.18 | 0.87, 1.61 | 1.09 | 0.81, 1.48 |
| 5-6 | 2,265 | 1.28 | 0.94, 1.75 | 1.18 | 0.87, 1.60 |
| ≥7 | 505 | 2.05 | 1.48, 2.83 | 1.59 | 1.16, 2.19 |
| Bleed heaviness | |||||
| Light | 978 | 0.89 | 0.77, 1.04 | 0.94 | 0.81, 1.09 |
| Moderate | 2,889 | 1.00 | Ref | 1.00 | Ref |
| Moderate/heavy | 1,363 | 1.28 | 1.14, 1.44 | 1.18 | 1.05, 1.32 |
| Heavy | 312 | 1.74 | 1.47, 2.06 | 1.42 | 1.21, 1.67 |
=adjusted for cycle at study entry, smoking status (current vs not current), MET hours of physical activity per week (<10, 10-19, 20-39, ≥40), age (<25, 25-29, 30-34, ≥35 years), BMI (<18, 18-24, 25-29, 30-34, ≥35 kg/m2), caffeine (<100, 100-199, 200-299, >300 mg/day), last form of birth control being hormonal (yes vs no), alcohol intake per week (0, 1-6, 7-13, ≥14 drinks/week)
=among those with regular cycles (3,977)
Among women without clinically-diagnosed PCOS (Table 4), unadjusted and adjusted associations with age at menarche, irregular cycles, bleed lengths and bleed heaviness with mFG scores ≥8 persisted. Similar patterns were observed when analyses were restricted to women without prior PCOS or thyroid disorders (Supplemental Table 4). Associations were relatively uniform across strata of age and BMI (Supplemental Table 5 and 6).
Table 4.
Unadjusted and Adjusted PRa for modified Ferriman-Gallwey score ≥8 vs <8 restricted to women without a PCOS diagnosis, n=4,800
| No. of women | Unadjusted | Adjusted | |||
|---|---|---|---|---|---|
| PR | 95% CI | PR | 95% CI | ||
| Age at menarche, years | |||||
| <12 | 1,337 | 1.34 | 1.17, 1.54 | 1.22 | 1.07, 1.40 |
| 12-13 | 2,492 | 1.00 | Ref | 1.00 | Ref |
| 14 | 562 | 0.94 | 0.76, 1.17 | 1.00 | 0.81, 1.24 |
| ≥15 | 409 | 0.95 | 0.75, 1.21 | 1.00 | 0.78, 1.27 |
| Time to menstrual regularity, years | 3,505 | 1.06 | 1.02, 1.09 | 1.05 | 1.02, 1.08 |
| Cycle regularity | |||||
| Regular | 3,741 | 1.00 | Ref | 1.00 | Ref |
| Irregular | 1,059 | 1.61 | 1.42, 1.84 | 1.47 | 1.29, 1.68 |
| Number of cycles (periods/year) | 4,800 | 0.94 | 0.92, 0.97 | 0.96 | 0.93, 0.98 |
| Cycle length, daysb | 3,741 | 1.02 | 0.99, 1.05 | 1.02 | 1.00, 1.04 |
| Cycle length, daysb | |||||
| <28 | 1,036 | 1.00 | Ref | 1.00 | Ref |
| 28-35 | 2,643 | 1.20 | 1.00, 1.43 | 1.21 | 1.01, 1.45 |
| ≥36 | 62 | 1.45 | 0.85, 2.47 | 1.35 | 0.79, 2.31 |
| Bleed length, days | |||||
| <3 | 189 | 1.00 | Ref | 1.00 | Ref |
| 3-4 | 2,295 | 1.37 | 0.92, 2.03 | 1.31 | 0.89, 1.93 |
| 5-6 | 1,959 | 1.43 | 0.96, 2.12 | 1.36 | 0.92,2.01 |
| ≥7 | 357 | 2.03 | 1.33, 3.11 | 1.76 | 1.16, 2.67 |
| Bleed heaviness | |||||
| Light | 879 | 0.91 | 0.76, 1.09 | 0.94 | 0.78, 1.12 |
| Moderate | 2,567 | 1.00 | Ref | 1.00 | Ref |
| Moderate/heavy | 1,129 | 1.20 | 1.04,1.39 | 1.13 | 0.98, 1.30 |
| Heavy | 225 | 1.58 | 1.25, 2.00 | 1.37 | 1.08, 1.73 |
=adjusted for cycle at study entry, smoking status (current vs not current), MET hours of physical activity per week (<10, 10-19, 20-39,≥40), age (<25, 25-29, 30-34, ≥35 years), BMI (<18, 18-24, 25-29, 30-34, ≥35 kg/m2), caffeine (<100, 100-199, 200-299, >300 mg/day), last form of birth control being hormonal (yes vs no), alcohol intake per week (0, 1-6, 7-13, ≥14 drinks/week)
=among those with regular cycles (3,741)
The restricted cubic spline for the association between mFG scores from cycle length was consistent with the categorical analysis, showing a slight monotonic relationship between higher mFG score and longer cycle lengths (Figure 2).
Fig. 2.
Restricted cubic spline curve of predicted modified Ferriman-Gallwey score by cycle length, days. Spline curve was adjusted for number of months attempting pregnancy (centered at mean: 3), smoking status (reference: current smoker), MET hours of physical activity per week (reference: <10 hours/week), age (reference: 25-29 years), BMI (reference: 18.5-24 kg/m2), caffeine (reference: <150 mg/day), last form of birth control being hormonal (reference: last form of birth control not hormonal), alcohol intake per week (reference: 0 drinks/week) with 4 knots placed at 26, 28, 30, and 32 days. The spline was truncated at the 99th percentile.
1.4. DISCUSSION
In tins preconception cohort study, we found that women with irregular menstrual cycles, longer menstrual cycle lengths, younger age at menarche, longer bleed lengths, and greater bleed heaviness were more likely to have hirsutism, defined as a self-reported mFG score ≥8. Age and BMI stratifications and race/ethnicity restrictions showed similar patterns to those seen in the overall cohort.
To our knowledge, there have been no studies examining the relation of specific menstrual cycle characteristics with clinical hirsutism. Studies on PCOS prevalence typically report menstrual cycle features of irregularity or oligomenorrhea [9, 10, 20], but do not address other features, such as cycle length, bleed length, and bleed heaviness.
In our cohort, the prevalence of self-reported menstrual irregularity was 32.0%, and prevalence of mFG ≥8 was 21.7%, with the latter estimate being slightly higher than values reported in similar cohorts (4.6-21.2%) [9, 10, 20]. Hispanic women constituted 8.1% of the cohort and had the highest mean mFG score of 6.1 and the highest proportion of mFG scores ≥8 (33.7%). This is consistent with existing literature, which suggests higher mean mFG scores for women of Hispanic race/ethnicity [7, 15, 26].
The observed associations of menstrual irregularity, longer cycle lengths, longer bleed lengths and bleed heaviness with hirsutism would be expected in the context of androgen excess disorders, in particular PCOS. In analyses excluding PCOS and thyroid disorders (also known to disrupt menstrual cycles), the associations between most menstrual characteristics and hirsutism were similar, suggesting that there may be undiagnosed androgen excess disorders within the population.
Given the complex relationship between the endocrine system and reproductive health, there are multiple potential confounders and effect modifiers for menstrual cycle characteristics and hirsutism. We were able to control for many of these, such as BMI, smoking, alcohol, and caffeine intake, in our analytical models [27]. Given that multiple attempts at pregnancy may be a sign of androgen excess disorders or oligo-ovulation, we also included cycle of pregnancy attempt and age at study entry.
Potential limitations of our study stem from reliance on participant self-reporting of both menstrual cycle characteristics and hirsutism scoring. As participants were asked to describe their usual menstrual cycle characteristics, the menstrual characteristics for any specific cycle could differ from the usual, and therefore be misclassified. We attempted to limit any misclassification of menstrual cycle length by asking and reporting the findings only among regular cyclers, who are more likely to be able to report this measure accurately [28]. In a prior analysis conducted in PRESTO, Wesselink et al. compared retrospective self-reported menstrual cycle characteristics with prospectively reported menstrual cycle characteristics using the fertility tracking app, FertilityFriend [27]. Results suggested that prospective cycle length was, on average, 2 days longer than self-reported cycle length and that bleed length, reported via the app, was 0.4 days longer than self-reported. In this study, we had a single measure of menstrual cycle characteristics, which did not account for variation in characteristics such as bleed length and heaviness by cycle. Other relevant data, such as age at menarche, are more readily measured retrospectively [29]. If reporting of menstrual characteristics relied on hirsutism status, differential or dependent misclassification could partly explain our results [30–32]. Misclassification from self-scoring is likely non-differential as participants were not aware of the study hypothesis. We sought to reduce misclassification of hirsutism by providing corresponding diagrams of specific body areas for each individual question. We attempted to capture participant perception and reporting of hirsutism. Prior studies have shown that self-reports may differ from those of medical providers, with self-report yielding higher mFG scores than provider assessment [30,31].However, providers are less able to report or classify based on the natural hair state, without influence of hair removal strategies. Our questionnaire specifically addressed “natural body state (without the use of hair removal procedures or treatments)”. Furthermore, while patients often report higher mFG scores than medical practitioners, these are independently important, as they both have been associated with poorer quality of life measures [30]. Additionally, participant reporting reflects self-perception of hairiness, which has received increased attention both in diagnostic evaluation and management [33]. Ethnic classification and lack of ethnic specific cutoffs also introduce bias from misclassification of mFG scores. With regards to ethnic specific cutoffs, certain populations have different predispositions to hair growth patterns. Average mFG scores are found to be lower in Asian populations [13, 34, 35] and higher in populations such as Middle Eastern/Mediterranean women [33, 36, 37]. However, based on practice patterns in clinical settings and prior studies in similar populations [9, 10, 20], the general threshold of ≥8 was appropriate for this cohort. Ultimately, we chose to use the binary cut point for clinical hirsutism to allow for higher specificity, optimal interpretation of results for clinical settings, and greater comparability with previous studies. Finally, women who are planning a pregnancy may be especially attuned to their menstrual cycles and concerned about potential health effects of hormonal disorders, which could have introduced some selection bias.
1.5. CONCLUSION
To summarize, a cross-sectional examination of the association of menstrual characteristics and hirsutism in this preconception cohort indicated that menstrual irregularity, longer cycle and bleed lengths, and heavier menstrual bleeds were associated with clinically significant hirsutism as assessed via the mFG scale. This association likely stems from known and potentially unknown androgen excess disorders among affected women.
Supplementary Material
Acknowledgments:
We would like to acknowledge Meghan Hewlett, MD for her inspiration and interest in the study of hirsutism during her time in the Mahalingaiah Lab. We acknowledge the contributions of PRESTO participants and staff. We acknowledge the in-kind donation of premium app subscriptions from FertilityFriend.com. We thank Mr. Michael Bairos for technical support in developing the study’s web-based infrastructure. All authors made significant contributions to the manuscript in accordance with the Vancouver group guidelines. the manuscript.
Statement of Financial Support: Funding was provided by Eunice Kennedy Shriver National Institute of Child Health and Human Development (R01 HD086742 and R21 HD072326) and Reproductive Scientist Development Program (RSDP K12 HD000849).
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Disclosure Statement: In the last three years, Dr. Wise has served as a consultant to AbbVie and has received in-kind donations from FertilityFriend.com, Kindara.com, Sandstone Diagnostics, and Swiss Precision Diagnostics for PRESTO. All other authors report no conflict of interest.
Declaration of interests
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
REFERENCES
- 1.Martin KA, et al. , Evaluation and treatment of hirsutism in premenopausal women: an endocrine society clinical practice guideline. J Clin Endocrinol Metab, 2008. 93(4): p. 1105–20. [DOI] [PubMed] [Google Scholar]
- 2.Ferriman D and Gallwey JD, Clinical assessment of body hair growth in women. J Clin Endocrinol Metab, 1961. 21: p. 1440–7. [DOI] [PubMed] [Google Scholar]
- 3.Hatch R, et al. , Hirsutism: implications, etiology, and management. Am J Obstet Gynecol, 1981. 140(7): p. 815–30. [DOI] [PubMed] [Google Scholar]
- 4.Knochenhauer ES, et al. , Examination of the chin or lower abdomen only for the prediction of hirsutism. Fertility and Sterility. 74(5): p. 980–983. [DOI] [PubMed] [Google Scholar]
- 5.Cook H, Brennan K, and Azziz R, Reanalyzing the modified Ferriman-Gallwey score: is there a simpler method for assessing the extent of hirsutism? Fertil Steril, 2011. 96(5): p. 1266–70 el. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Yildiz BO, et al. , Visually scoring hirsutism. Hum Reprod Update, 2010. 16(1): p. 51–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Mimoto MS, Oyler JL, and Davis AM, Evaluation and Treatment of Hirsutism in Premenopausal Women. JAMA, 2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Asuncion M, et al. , A prospective study of the prevalence of the polycystic ovary syndrome in unselected Caucasian women from Spain. J Clin Endocrinol Metab, 2000. 85(7): p. 2434–8. [DOI] [PubMed] [Google Scholar]
- 9.Knochenhauer ES, et al. , Prevalence of the polycystic ovary syndrome in unselected black and white women of the southeastern United States: a prospective study. J Clin Endocrinol Metab, 1998. 83(9): p. 3078–82. [DOI] [PubMed] [Google Scholar]
- 10.March WA, et al. , The prevalence of polycystic ovary syndrome in a community sample assessed under contrasting diagnostic criteria. Hum Reprod, 2010. 25(2): p. 544–51. [DOI] [PubMed] [Google Scholar]
- 11.The Rotterdam ESHRE/ASRM-Sponsored PCOS Consensus Workshop Group., Revised2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome (PCOS). Human Reproduction, 2004. 19(1): p. 41–47. [DOI] [PubMed] [Google Scholar]
- 12.Goodman NF, et al. , American Association of Clinical Endocrinologists, American College of Endocrinology, and Androgen Excess and Pcos Society Disease State Clinical Review: Guide to the Best Practices in the Evaluation and Treatment of Polycystic Ovary Syndrome-Part 1. Endocr Pract, 2015. 21(11): p. 1291–300. [DOI] [PubMed] [Google Scholar]
- 13.Chen X, et al. , Prevalence of polycystic ovary syndrome in unselected women from southern China. Eur J Obstet Gynecol Reprod Biol, 2008. 139(1): p. 59–64. [DOI] [PubMed] [Google Scholar]
- 14.Diamanti-Kandarakis E, et al. , A survey of the polycystic ovary syndrome in the Greek island of Lesbos: hormonal and metabolic profile. J Clin Endocrinol Metab, 1999. 84(11): p. 4006–11. [DOI] [PubMed] [Google Scholar]
- 15.Gabrielli L and Aquino EM, Polycystic ovary syndrome in Salvador, Brazil: a prevalence study in primary healthcare. Reprod Biol Endocrinol, 2012. 10: p. 96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Bozdag G, et al. , The prevalence and phenotypic features of polycystic ovary syndrome: a systematic review and meta-analysis. Hum Reprod, 2016. 31(12): p. 2841–2855. [DOI] [PubMed] [Google Scholar]
- 17.Wise LA, et al. , Design and Conduct of an Internet-Based Preconception Cohort Study in North America: Pregnancy Study Online. Paediatr Perinat Epidemiol, 2015. 29(4): p. 360–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Wesselink AK, et al. , Caffeine and caffeinated beverage consumption and fecundability in a preconception cohort. Reprod Toxicol, 2016. 62: p. 39–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Wesselink AK, et al. , Age and fecundability in a North American preconception cohort study. Am J Obstet Gynecol, 2017. 217(6): p. 667 e1–667 e8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.DeUgarte CM, et al. , Degree of facial and body terminal hair growth in unselected black and white women: toward a populational definition of hirsutism. J Clin Endocrinol Metab, 2006. 91(4): p. 1345–50. [DOI] [PubMed] [Google Scholar]
- 21.Ainsworth BE, et al. , 2011 Compendium of Physical Activities: a second update of codes and MET values. Med Sci Sports Exerc, 2011. 43(8): p. 1575–81. [DOI] [PubMed] [Google Scholar]
- 22.Hahn KA, et al. , Correlates of menstrual cycle characteristics among nulliparous Danish women. Clin Epidemiol, 2013. 5: p. 311–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Rowland AS, et al. , Influence of medical conditions and lifestyle factors on the menstrual cycle. Epidemiology, 2002. 13(6): p. 668–74. [DOI] [PubMed] [Google Scholar]
- 24.Zhou XH, Eckert GJ, and Tierney WM, Multiple imputation in public health research. Stat Med, 2001. 20(9–10): p. 1541–9. [DOI] [PubMed] [Google Scholar]
- 25.Sterne JA, et al. , Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ, 2009. 338: p. b2393. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Escobar-Morreale HF, et al. , Epidemiology, diagnosis and management of hirsutism: a consensus statement by the Androgen Excess and Polycystic Ovary Syndrome Society. Hum Reprod Update, 2012. 18(2): p. 146–70. [DOI] [PubMed] [Google Scholar]
- 27.Wesselink AK, et al. , Menstrual cycle characteristics and fecundability in a North American preconception cohort. Ann Epidemiol, 2016. 26(7): p. 482–487 e1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Small CM, Manatunga AK, and Marcus M, Validity of self-reported menstrual cycle length. Ann Epidemiol, 2007. 17(3): p. 163–70. [DOI] [PubMed] [Google Scholar]
- 29.Lundblad MW and Jacobsen BK, The reproducibility of self-reported age at menarche: The Tromsø Study. BMC Womens Health, 2017. 17(1): p. 62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Pasch L, et al. , Clinician vs Self-ratings of Hirsutism in Patients With Polycystic Ovarian Syndrome: Associations With Quality of Life and Depression. JAMA Dermatol, 2016. 152(7): p. 783–8. [DOI] [PubMed] [Google Scholar]
- 31.Wild RA, et al. , Ferriman Gallwey self-scoring I: performance assessment in women with polycystic ovary syndrome. J Clin Endocrinol Metab, 2005. 90(7): p. 4112–4. [DOI] [PubMed] [Google Scholar]
- 32.Kristensen P, Bias from Nondifferential but Dependent Misclassification of Exposure and Outcome. Epidemiology, 1992. 3(3): p. 210–215. [DOI] [PubMed] [Google Scholar]
- 33.Martin KA, et al. , Evaluation and Treatment of Hirsutism in Premenopausal Women: An Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab, 2018. [DOI] [PubMed] [Google Scholar]
- 34.Zhang HY, et al. , Clinical characteristics, metabolic features, and phenotype of Chinese women with polycystic ovary syndrome: a large-scale case-control study. Arch Gynecol Obstet, 2013. 287(3): p. 525–31. [DOI] [PubMed] [Google Scholar]
- 35.Kim MJ, et al. , Prevalence of metabolic syndrome is higher among non-obese PCOS women with hyperandrogenism and menstrual irregularity in Korea. PLoS One, 2014. 9(6): p. e99252. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Coskun A, et al. , Modified Ferriman-Gallwey hirsutism score and androgen levels in Turkish women. Eur J Obstet Gynecol Reprod Biol, 2011. 154(2): p. 167–71. [DOI] [PubMed] [Google Scholar]
- 37.Api M, et al. , Interobserver variability of modified Ferriman-Gallwey hirsutism score in a Turkish population. Arch Gynecol Obstet, 2009. 279(4): p. 473–9. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.