Abstract
Context
Sleep plays important roles in metabolic and reproductive function, and polycystic ovary syndrome (PCOS) is associated with sleep disturbances, including increased prevalence of obstructive sleep apnea.
Objective
We sought to evaluate sleep parameters in infertile women with PCOS compared with women with unexplained infertility (UI) and identify risk factors for disturbed sleep.
Methods
At private and academic ambulatory gynecology and infertility practices, we evaluated a prospective cohort of women diagnosed with PCOS or UI from 2 randomized clinical trials. We included 1603 infertile women enrolled in 2 concurrent randomized clinical trials. The main outcome measures were self-reported sleep measures.
Results
Sleep duration <6 hours (6.1% vs 2.7%; P < .001), habitual snoring (37.8% vs 19.0%; P < .001), and clinical sleepiness (12.0% vs 8.6%; P < .026) were more common in women with PCOS than those with UI. After adjusting for covariates, PCOS and elevated fasting insulin were associated (P = .010) with clinical symptoms of obstructive sleep apnea (OSA) diagnosis, whereas PCOS, elevated insulin (P = .003), WC >88 cm (P = .003), and current smoking (P = .012) were associated with habitual snoring. Clinical depression score (P < .001) and PCOS diagnosis (P = .002) were associated with perceived daytime sleepiness. Short sleep duration and clinical symptoms of OSA were not associated with conception and live birth rates.
Conclusion
Infertile women with PCOS more commonly report sleep disturbances than those with UI. Markers of insulin resistance are associated with previous diagnosis of OSA, habitual snoring, and short sleep duration. The presence of clinical symptoms of OSA or short sleep duration does not affect fertility treatment response.
Keywords: PCOS, unexplained infertility, sleep-disordered breathing, snoring, sleep apnea
Sleep is an important component of normal physiology, and sleep disorders are common in contemporary society. Abnormal sleep has been associated with health conditions and comorbidities such as obesity, hypertension, diabetes, depression, and poor quality of life. In the past decade, a wealth of data has shown that sleep disruption in pregnancy has adverse consequences on maternal health (1), yet we know little about the relationship between sleep and infertility in women (2, 3). Several of the limited studies investigating sleep in reproductive age women have demonstrated associations between disordered sleep and altered reproductive function, including menstrual irregularities, increased miscarriages, and adverse obstetrical and perinatal outcomes (4-7).
Polycystic ovary syndrome (PCOS) is associated with reproductive (8) and sleep abnormalities (9). Women with PCOS also have an increased incidence of depression (10) and metabolic dysfunction (11), both of which have been associated with sleep abnormalities. Indeed, women with PCOS have been reported to be up to 30 times more likely to have obstructive sleep apnea (OSA) than controls (12-15). Although obesity is common in women with PCOS and/or OSA, the high frequency of OSA in PCOS women is not simply explained by elevated body mass index (BMI). Insulin resistance and metabolic dysfunction have been posited to be a precursor of OSA (16), and OSA has been shown to contribute to metabolic dysregulation (17) independent of BMI. Increased prevalence of sleep-disordered breathing and greater daytime sleepiness have been reported in women with PCOS (12-14, 18) than in controls, even after adjusting for BMI. Women with unexplained infertility (UI) in whom the etiology of infertility is enigmatic, may also exhibit depression, stress, and anxiety (19), factors which have been associated with sleep disturbances, yet sleep has not been well studied in this population of infertile women. Assessment of sleep habits in women with PCOS and UI may expand the understanding of underlying pathophysiology of sleep disorders in women with infertility.
This study was undertaken as an ancillary study of 2 randomized controlled clinical treatment trials of reproductive aged women with infertility that were designed to be complementary, facilitating comparisons between 2 different populations of infertile women (20, 21). The goal for this study was to gain a better understanding of sleep parameters associated with 2 common infertility diagnoses: PCOS and UI. We hypothesized that women with PCOS would have an increased prevalence of previously diagnosed OSA and increased habitual snoring compared with women with UI; moreover, age, obesity, waist circumference (WC), insulin resistance, and the presence of metabolic syndrome would be associated with self-reported OSA risk in all women. We speculated that obesity would be associated with daytime sleepiness and fatigue regardless of OSA and sleep disruption, and that infertile women with OSA would have reduced live birth rates despite infertility treatment.
Subjects and Methods
Study Population
The study population comprised 739 women with PCOS and 864 infertile women with UI who completed a baseline sleep questionnaire (see below) and participated in concurrent randomized controlled trials conducted by the NICHD’s Cooperative Reproductive Medicine Network: Pregnancy in Polycystic Ovary Syndrome II (PPCOS II) (http://clinicaltrials.gov: NCT00719186) and the Assessment of Multiple Gestation in Ovarian Stimulation (AMIGOS) (NCT01044862).
The rationale and methods, some of the baseline data, and main outcomes from these clinical trials have been published previously (20-25). Briefly, PPCOSII was a prospective, multicenter, double-blind, randomized clinical trial comparing live birth rate in response to treatment with clomiphene citrate or letrozole for up to a total of 5 cycles in oligoovulatory women with PCOS. AMIGOS was a prospective, multicenter randomized clinical trial in couples with UI that evaluated the outcomes of live birth and multiple gestations associated with ovarian stimulation with either gonadotropins, clomiphene citrate, or letrozole in conjunction with intrauterine insemination for up to a total of 4 cycles.
For both trials, women were between the ages of 18 and 40 years. Women with PCOS were diagnosed by the modified Rotterdam Criteria, had chronic anovulation and either hyperandrogenism (clinical or biochemical) or polycystic ovaries on ultrasound, with exclusion of secondary causes of PCOS; tubal patency demonstrated in at least 1 fallopian tube, and a male partner with sperm count of at least 14 million sperm per milliliter. Women with UI had regular ovulatory menstrual cycles and a negative evaluation for underlying fertility factors, tubal patency confirmed in at least 1 fallopian tube, and a male partner with at least 5 million sperm in the ejaculate. In both trials, couples had attempted to conceive without success for at least 1 year.
For both trials, couples were recruited from a variety of sources including local practices at each academic center’s clinical site, such as radio and newspaper advertising and the internet (26). Respondents were screened for eligibility with a brief telephone interview that was standardized across all sites. Eligible couples had a formal screening visit, at which time written informed consent was obtained and study questionnaires were collected. The same clinical sites recruited for each of the 2 studies concurrently using similar methods, thus both cohorts of participants were sampled from the same clinical population base. This strategy resulted in contemporaneous, well-characterized cohorts of infertile women with PCOS and UI that are representative of their communities across the United States.
Data Collection
Demographic characteristics, and medical and social history were obtained from participants at the screening, baseline, and end of study visits. Procedures at all visits were standardized across all study sites and each site used identical case report forms prepared by the Reproductive Medicine Network Data Coordination Center at Yale University for data collection for both PPCOSII and AMIGOS studies. Anthropometric measurements were obtained by a research nurse at each site. BMI was calculated from height and weight measurements performed at the screening visit. Participants were weighed while dressed in light clothing and height was measured without shoes. Participants were assessed with standardized Ferriman–Gallwey scoring for hirsutism (27), an acne assessment, and sebum measurements. Fasting blood collected at baseline or screening prior to start of fertility treatment was used for all hormonal assays. Samples were batched and analyzed at the Ligand Assay and Analysis Core Laboratory at the University of Virginia. The immunoassay method was used for the measurement of following items (with RRID, research reference identifier): total testosterone (AB_ 2756391), sex hormone–binding globulin (SHBG, AB_2819251), androstenedione (AB_2756382), follicle-stimulating hormone (FSH, AB_2756389), luteinizing hormone (LH, AB_2756388), insulin (AB_2756390), and high-sensitivity C-reactive protein (hsCRP, RRID not available). We have previously published the quality control standards of these assays and all had inter- and intra-assays coefficients of variation <10% (23, 25). The lab studies were performed as part of the conduct of the main clinical trials which have been published before (20, 21).
All participants were asked to complete the Sleep Habits Questionnaire (28) and Patient Health Questionnaire (PHQ-9) (29) in the presence of a research nurse at baseline and end of study. The Sleep Habits Questionnaire was the standard measure used to collect data for the 10-year-long National Heart Lung and Blood Institute–sponsored multicenter Sleep Heart Health Study conducted in 10 US communities to determine the consequences of sleep-disordered breathing. The Sleep Habits Questionnaire contains 56 questions about sleep habits, sleepiness, and daytime performance, and is a valid means to identify increased sleep apnea activity. The domains include duration and quality of sleep, functional impact of sleepiness, self-reported breathing disturbances, snoring and breathing during sleep, and insomnia. A construct variable “clinical symptoms of obstructive sleep apnea” (cOSA) was defined as having a sleep duration of <6 hours and snoring. A prior diagnosis of obstructive sleep apnea (pOSA) was defined as a patient having ever been told by her doctor that she had sleep apnea. Clinical daytime sleepiness was defined as having an Epworth sleepiness total score of 10 or more. The PHQ-9 is a brief self-administered tool for depression that incorporates 9 diagnostic criteria and major depressive symptoms with scores of “0” (not at all) to “3” (nearly every day). This tool has been validated to screen for depression with major depression defined as PHQ-9 score ≥10 (30).
Ethics Approval
Both PPCOSII and AMIGOS studies were approved by the Institutional Review Boards of all participating institutions and clinical practices, and all participants provided written informed consent for their participation in the study.
Data and Statistical Analyses
Patients with polycystic ovary syndrome from the PPCOSII trial (PCOS, n = 739) and those with unexplained infertility from the AMIGOS trial (UI, n = 864) with complete sleep habit data were included for the analysis. Categorical data were reported as frequencies and percentages; differences in these measures between groups were assessed by means of a chi-squared analysis, with Fisher’s exact test being used for expected frequencies of less than 5. Continuous data were expressed as means ± SD, with a Student’s test or Wilcoxon rank-sum test being used for testing differences between 2 groups. Logistic regression was used to test the association between sleep habit profiles and related measures, which included BMI, Ferriman–Gallwey score, testosterone, measures of insulin resistance, calculated free androgen index, smoking, alcohol consumption, history of heart disease, hypertension, metabolic syndrome, WC, waist to hip ratio, triglycerides, and ethnicity, and SHBG, LH to FSH ratio, androstenedione, and hsCRP levels. Backward selection at a significance level of 0.05 was then performed for model selection for sleep profiles, including ever snored, snore frequency of 3 to 7 nights per week, pOSA, sleep <6 hours, and clinical daytime sleepiness. When the final model was obtained, the adjusted odds ratios (OR) and 95% CIs were computed with respect to the corresponding reference groups. For the analysis of sleep habit profiles and pregnancy outcome, the ORs were adjusted by the variables in the above final models, as well as infertility diagnosis (PCOS or UI), treatment (letrozole, clomiphene citrate or gonadotropins) and age. Analyses were performed with SAS software, version 9.4 (SAS Institute). Statistical significance was defined as a 2-sided P value of less than .05.
Results
Baseline Demographic Characteristics of Study Population
Women with PCOS differed from those with UI in several ways, as presented in Table 1. Women with PCOS were younger (mean age of 28.9 ± 4.2 years) than those with UI (32.2 ± 4.3 years). BMI was significantly higher in women with PCOS than in those with UI (BMI 35.2 ± 9.3 vs 26.8 ± 6.4). Both groups were racially diverse, with greater representation of Hispanic (17.2% vs 10.8%) and Black (13.1% vs 8.1%) women in the PCOS cohort. Most of the women had completed high school and had some college education (PCOS 72%, UI 88.5%), with a lower percentage of graduate degrees in the women with PCOS (11.5%) than in those with UI (26.9%). Household income differed as well, with 40.2% of PCOS women and 16.7% of women with UI reporting household income of less than $50 000. All these differences were statistically significant with P < .001 (Table 1).
Table 1.
Baseline characteristics of subjects by infertility diagnosis
| Variables | UI | PCOS | P valuea |
|---|---|---|---|
| Age (N = 1603) | 32.2 ± 4.3 | 28.9 ± 4.2 | <.001 |
| BMI (N = 1603) | 26.8 ± 6.4 | 35.2 ± 9.3 | <.001 |
| Ethnic group (N = 1603) | <.001 | ||
| Not Hispanic or Latino | 89.2 (771/864) | 82.8 (612/739) | |
| Hispanic or Latino | 10.8 (93/864) | 17.2 (127/739) | |
| Race | <.001 | ||
| White | 81.5 (704/864) | 79.0 (584/739) | |
| Black | 8.1 (70/864) | 13.1 (97/739) | |
| Asian | 6.5 (56/864) | 3.0 (22/739) | |
| Others | 3.9 (34/864) | 4.9 (36/739) | |
| FG Score (N = 1597) | 7.6 ± 5.8 | 17.0 ± 8.6 | <.001 |
| Level of education | <.001 | ||
| High school graduate or less | 7.9 (68/864) | 22.9 (169/739) | |
| College graduate or some college | 65.3 (564/864) | 65.6 (485/739) | |
| Graduate degree | 26.9 (232/864) | 11.5 (85/739) | |
| Annual household income | <.001 | ||
| <$50 000 | 16.7 (144/864) | 40.2 (297/739) | |
| ≥$50 000 | 66.4 (574/864) | 46.1 (341/739) | |
| Wish to not answer | 16.9 (146/864) | 13.7 (101/739) | |
| Glucose (N = 1578) | .259 | ||
| ≤100 | 93.2 (784/841) | 91.7 (676/737) | |
| >100 | 6.8 (57/841) | 8.3 (61/737) | |
| Insulin (N = 1577) | <.001 | ||
| ≤10 | 73.8 (620/840) | 39.2 (289/737) | |
| >10 | 26.2 (220/840) | 60.8 (448/737) | |
| HOMA Index (N = 1577) | <0.001 | ||
| <4 | 89.3 (750/840) | 61.2 (451/737) | |
| ≥4 | 10.7 (90/840) | 38.8 (286/737) | |
| Testosterone (N = 1576) | <0.001 | ||
| ≤50 | 96.0 (805/839) | 50.3 (371/737) | |
| >50 | 4.1 (34/839) | 49.7 (366/737) | |
| SHBG (N = 1574) | <.001 | ||
| ≤30 | 13.3 (111/837) | 57.1 (421/737) | |
| >30 | 86.7 (726/837) | 42.9 (316/737) | |
| Free androgen indexb (N = 1571) | 1.9 ± 2.2 | 7.8 ± 6.0 | <.001 |
| LH/FSH ratio (N = 1574) | 0.8 ± 2.5 | 2.2 ± 6.7 | <.001 |
| Androstenedione (ng/mL) (N = 1569) | 2.4 ± 1.0 | 4.2 ± 1.7 | <.001 |
| Triglycerides (mg/dL) (N = 1548) | 94.1 ± 49.7 | 116.3 ± 57.8 | <.001 |
| Waist/Hip ratio (N = 1598) | 0.8 ± 0.1 | 0.9 ± 0.1 | <.001 |
| Metabolic Syndrome (N = 1545) | <0.001 | ||
| Yes | 13.0 (105/810) | 34.0 (250/735) | |
| No | 87.0 (705/810) | 66.0 (485/735) | |
| PHQ-9 depression scorecc (N = 1603) | 1.8 ± 2.5 | 3.9 ± 4.1 | <.001 |
| hsCRP (mg/L) (N = 1575) | 3.2 ± 4.9 | 6.4 ± 7.0 | <.001 |
Data are presented as % (no./total no.) or mean ± SD.
Abbreviations: UI, unexplained infertility; PCOS, polycystic ovary syndrome; HOMA, homeostatic model assessment; SHBG, sex hormone–binding globulin; hsCRP, high-sensitivity C-reactive protein.
aChi-squared or Fisher’s exact test was used for categorical variables; Wilcoxon’s rank-sum test was used for continuous variable.
bThe free androgen index was calculated according to the following formula: (total testosterone [nanomoles per liter] ÷ sex hormone–binding globulin [nanomoles per liter]) × 100.
cTotal score of 9 items, each of which ranges from 0 (not at all) to 3 (nearly every day). Higher score indicates higher level of depression.
Metabolic and Hormonal Measures and Clinical Findings
As noted in Table 1, women with PCOS had statistically significant greater degrees of hirsutism as measured by Ferriman–Gallwey score (17.0 ± 8.6) than those with UI (7.6 ± 5.8; P = .001). As anticipated given that the definition of PCOS includes hyperandrogenism, approximately half of the women with PCOS had testosterone levels >50 ng/mL, whereas only 4.1% of the women with UI had elevated testosterone >50 ng/mL. Additionally, SHBG levels differed between the PCOS and UI cohorts with a significantly greater proportion of women with a SHBG level of ≤30 nmol/L among PCOS than UI (57.1% vs13.3%, respectively, P < .001). Both cohorts were normoglycemic, but high fasting insulin levels >10 µIU/mL were more common in PCOS compared with UI (60.8% vs 26.2% respectively; P < .001). A significantly greater preponderance of metabolic syndrome was observed in the women with PCOS compared with UI (34.0% vs13.0% respectively; P < .001). The levels of waist to hip ratio, calculated free androgen index, LH to FSH ratio, baseline androstenedione, triglycerides, and hsCRP were significantly higher in women with PCOS than with UI (P < .001 Table 1). Although the incidence of depression was low in both cohorts, the mean PHQ-9 depression score was significantly higher among women with PCOS than those with UI (3.9 ± 4.1 vs 1.8 ± 2.5; P < .001) (Table 1).
Sleep Characteristics Of Women With PCOS and UI
A total of 1598 woman completed the Sleep Habits Questionnaire, though some may have omitted a few of the questions. None reported narcolepsy or restless leg syndrome. Results of the comparison of sleep measures between women with PCOS and UI are shown in Table 2. Women with PCOS more frequently reported <6 hours sleep duration (6.1% vs 2.7%; P < .001), habitual snoring (with a snore frequency of 3-7 nights per week) (37.8% vs19.0%; P < .001), and more subjective daytime sleepiness (12.0% vs 8.6%; P < .026) than those with UI. In addition, a history of pOSA via sleep studies (7.1 % vs 1.6%, P < .002) and having cOSA, defined as reported sleep duration of <6 hours plus snoring, was more prevalent in women with PCOS than UI (4.1% vs 0.9%, P < .001) (Table 2).
Table 2.
Sleep habits in UI and PCOS subjects
| Variables | UI | PCOS | P valuea |
|---|---|---|---|
| Ever snored (N = 1598) | <.001 | ||
| Yes | 44.7 (384/860) | 63.1 (466/738) | |
| No | 47.1 (405/860) | 25.6 (189/738) | |
| Unknown | 8.3 (71/860) | 11.3 (83/738) | |
| Snore frequency (nights per week, N = 781) | <.001 | ||
| 0-2 | 81.0 (294/363) | 62.2 (260/418) | |
| 3-7 | 19.0 (69/363) | 37.8 (158/418) | |
| Sleeping an average of <6 hours (N = 1597) | <.001 | ||
| Yes | 2.7 (23/863) | 6.1 (45/734) | |
| No | 97.3 (840/863) | 93.9 (689/734) | |
| Sleeping an average of less than 5 hours (N = 1597) | .238 | ||
| Yes | 0.5 (4/863) | 1.0 (7/734) | |
| No | 99.5 (859/863) | 99.0 (727/734) | |
| cOSAb (N = 1592) | <.001 | ||
| Yes | 0.9 (8/859) | 4.1 (30/733) | |
| No | 99.1 (851/859) | 95.9 (703/733) | |
| pOSAc (N = 390) | .002 | ||
| Yes | 1.6 (2/123) | 7.1 (19/267) | |
| No | 96.8 (119/123) | 89.9 (240/267) | |
| Unknown | 1.6 (2/123) | 3.0 (8/267) | |
| Clinical sleepinessd (N = 1598) | .026 | ||
| Yes | 8.6 (74/862) | 12.0 (88/736) | |
| No | 91.4 (788/862) | 88.0 (648/736) |
Data are presented as % (no./total no.).
aChi-squared or Fisher’s exact test was used for categorical variables; Wilcoxon’s rank-sum test was used for continuous variable.
bcOSA, clinical symptoms of obstructive sleep apnea, defined as reported sleep duration of <6 hours and snore.
cpOSA, prior diagnosis of obstructive sleep apnea with sleep studies.
dClinical daytime sleepiness was defined as having an Epworth sleepiness total score of 10 or more. The total score is the sum of 8 item scores and can range between 0 and 24. The higher the score, the higher the person’s level of daytime sleepiness.
Associations of Sleep Parameters With Clinical Findings
Sleep parameters were analyzed in relation to clinical findings. As shown elsewhere (Table 1 (31)), in women with PCOS, habitual snoring of 3 to 7 nights per week was associated with elevated BMI (P < .001) and measures related to insulin resistance (HOMA index ≥4, P = 0.001; elevated insulin level, P = .001; and lower SHBG, P = .012), WC >88 (P < .001), elevated levels of waist to hip ratio, triglycerides, hsCRP, and metabolic syndrome (P = .001); whereas having short sleep duration plus snoring (cOSA) was associated with elevated HOMA index (P = .015) and insulin level (P = .027). Similarly, in women with UI, habitual snoring was associated with elevated BMI (P < .001), HOMA index ≥4 (P = .047), elevated insulin (P < .001), free androgen index (P = .008), hsCRP levels (P < .001), WC >88 (P = .002) and additionally with screening positive for depression (P = .046), but not with having metabolic syndrome (P = .206) (Table 2 (31)).
The associations between sleep habits and clinical measures in all the participants as presented in Table 3 and elsewhere (Table 3 (31)). Habitual snoring was associated with elevated BMI (P < .001), WC (P < .001), waist to hip ratio (P < .001), triglycerides (P < .001), hsCRP (P < .001), measures of insulin resistance (P < .001), and hyperandrogenism (P = .003). Short sleep duration <6 hours was associated with elevated BMI (P = .016) and WC (P = .018) as well as measures of insulin resistance (P < .001) but not with elevated testosterone level (P = .424), and clinical sleepiness was associated with positive screen for depression (P < .001) and a decreased triglyceride level (P = .029) (Table 3). Similar results were obtained when WC was used as a continuous variable. Interestingly, smoking was associated with habitual snoring (P = .002); and alcohol intake and ethnicity were not meaningfully associated with reported sleep measures.
Table 3.
Association between sleep habit profiles and related measures for both PCOS and UI subjects
| Variable | Ever snored | Snore frequency 3-7 nights per week | cOSA | Sleep <6 hours | pOSA | Clinical sleepiness | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Yes (%) | P valuea | Yes (%) | P value | Yes (%) | P value | Yes (%) | P value | Yes (%) | P value | Yes (%) | P value | |
| BMI (N = 1603) | <.001 | <.001 | .002 | .016 | .001 | .755 | ||||||
| ≤25 | 32.0 (176/550) | 13.2 (22/167) | 0.7 (4/548) | 2.7 (15/549) | 0.0 (0/92) | 9.5 (52/550) | ||||||
| 25-35 | 55.7 (330/593) | 23.4 (71/304) | 2.5 (15/591) | 4.0 (24/594) | 2.9 (4/137) | 10.8 (64/593) | ||||||
| >35 | 75.6 (344/455) | 43.2 (134/310) | 4.2 (19/453) | 6.4 (29/454) | 10.6 (17/161) | 10.1 (46/455) | ||||||
| FG score (N = 1597) | <.001 | .001 | .058 | .013 | .093 | .047 | ||||||
| <8 | 42.3 (229/541) | 25.2 (55/218) | 1.3 (7/539) | 2.4 (13/541) | 4.8 (5/104) | 8.0 (43/541) | ||||||
| 8-15 | 53.6 (309/577) | 24.2 (67/277) | 2.4 (14/575) | 4.3 (25/576) | 3.5 (5/145) | 10.2 (59/577) | ||||||
| ≥16 | 65.0 (308/474) | 36.9 (104/282) | 3.6 (17/472) | 6.1 (29/474) | 7.9 (11/139) | 12.7 (60/474) | ||||||
| HOMA index (N = 1577) | <.001 | <.001 | <.001 | <.001 | .138 | .784 | ||||||
| <4 | 47.1 (563/1196) | 22.7 (120/528) | 1.5 (18/1192) | 3.3 (40/1197) | 3.7 (9/242) | 10.4 (124/1197) | ||||||
| ≥4 | 71.8 (270/376) | 43.0 (102/237) | 5.4 (20/374) | 7.5 (28/374) | 8.5 (12/142) | 9.9 (37/375) | ||||||
| Glucose (N = 1578) | <.001 | .950 | .202 | .961 | .875 | .198 | ||||||
| ≤100 | 51.4 (748/1455) | 29.0 (200/690) | 2.3 (33/1449) | 4.3 (63/1454) | 5.4 (19/352) | 10.5 (153/1455) | ||||||
| >100 | 72.0 (85/118) | 29.3 (22/75) | 4.2 (5/118) | 4.2 (5/118) | 6.3 (2/32) | 6.8 (8/118) | ||||||
| Insulin (N = 1577) | <.001 | <.001 | <.001 | .005 | .067 | .763 | ||||||
| ≤10 | 42.0 (380/905) | 17.9 (65/364) | 1.1 (10/903) | 3.1 (28/907) | 2.5 (4/163) | 10.0 (91/906) | ||||||
| >10 | 67.9 (453/667) | 39.2 (157/401) | 4.2 (28/663) | 6.0 (40/664) | 7.7 (17/221) | 10.5 (70/666) | ||||||
| Testosterone (N = 1576) | <.001 | .003 | .100 | .424 | .015 | .093 | ||||||
| ≤50 | 49.8 (583/1171) | 25.8 (138/535) | 2.1 (24/1168) | 4.1 (48/1173) | 4.0 (10/250) | 11.0 (129/1173) | ||||||
| >50 | 62.0 (248/400) | 36.4 (83/228) | 3.5 (14/397) | 5.0 (20/397) | 8.2 (11/134) | 8.0 (32/398) | ||||||
| SHBG (N = 1574) | <.001 | <.001 | <.001 | .004 | .350 | .790 | ||||||
| ≤30 | 69.6 (368/529) | 39.2 (127/324) | 4.6 (24/525) | 6.4 (34/528) | 7.3 (12/164) | 10.6 (56/531) | ||||||
| >30 | 44.6 (464/1040) | 21.6 (95/440) | 1.4 (14/1038) | 3.3 (34/1040) | 4.1 (9/220) | 10.1 (105/1038) | ||||||
| Free androgen index (N = 1571) | 5.7 (2.4)b | <.001 | 7.2 (2.2)b | <.001 | 7.2 (2.6)b | .003 | 5.9 (1.3)b | .044 | 8.5 (2.7)b | .052 | 4.7 (0.1)b | .880 |
| LH/FSH ratio (N = 1574) | 1.4 (–0.2) | .464 | 1.7 (0.2) | .694 | 1.3 –0.2) | .368 | 1.3 (–0.2) | .175 | 1.4 (–0.1) | .797 | 1.3 (–0.2) | .265 |
| Androstenedione (ng/mL) (N = 1569) | 3.4 (0.3) | .002 | 3.6 (0.3) | .019 | 3.9 (0.6) | .028 | 3.6 (0.3) | .096 | 4.2 (0.6) | .175 | 3.1 (–0.2) | .225 |
| Smoking (N = 1603) | .023 | .002 | .464 | .946 | .660 | .206 | ||||||
| Not at all | 50.6 (500/989) | 25.6 (117/457) | 2.6 (26/987) | 4.1 (41/991) | 4.9 (11/223) | 10.5 (104/991) | ||||||
| Currently | 64.6 (113/175) | 43.4 (43/99) | 2.9 (5/173) | 4.6 (8/173) | 8.1 (5/62) | 6.3 (11/174) | ||||||
| Quit | 54.6 (237/434) | 29.8 (67/225) | 1.6 (7/432) | 4.4 (19/433) | 4.8 (5/105) | 10.9 (47/433) | ||||||
| Consume alcohol (N = 1603) | .056 | .584 | .771 | .456 | .758 | .337 | ||||||
| Not at all | 45.0 (86/191) | 24.6 (17/69) | 2.1 (4/190) | 3.7 (7/191) | 6.0 (3/50) | 9.4 (18/191) | ||||||
| Currently | 54.9 (603/1098) | 30.0 (169/563) | 2.3 (25/1096) | 4.0 (44/1099) | 5.1 (13/257) | 10.9 (119/1097) | ||||||
| Quit | 52.1 (161/309) | 27.5 (41/149) | 2.9 (9/306) | 5.5 (17/307) | 6.0 (5/83) | 8.1 (25/310) | ||||||
| Ethnic group (N = 1603) | .043 | .907 | .932 | .808 | .617 | .083 | ||||||
| Not Hispanic or Latino | 52.0 (718/1380) | 29.0 (191/659) | 2.4 (33/1375) | 4.2 (58/1378) | 5.8 (20/347) | 10.7 (147/1379) | ||||||
| Hispanic or Latino | 60.6 (132/218) | 29.5 (36/122) | 2.3 (5/217) | 4.6 (10/219) | 2.3 (1/43) | 6.9 (15/219) | ||||||
| History of heart disease (N = 1603) | .257 | 1.000 | 1.000 | .443 | .276 | 1.000 | ||||||
| Yes | 40.0 (18/45) | 26.7 (4/15) | 2.2 (1/45) | 6.5 (3/46) | 0.0 (0/9) | 8.7 (4/46) | ||||||
| No | 53.6 (832/1553) | 29.1 (223/766) | 2.4 (37/1547) | 4.2 (65/1551) | 5.5 (21/381) | 10.2 (158/1552) | ||||||
| Hypertension (N = 1603) | .003 | .200 | .800 | .665 | .922 | .483 | ||||||
| Yes | 62.3 (119/191) | 34.3 (37/108) | 2.6 (5/191) | 3.7 (7/191) | 6.0 (3/50) | 11.6 (22/190) | ||||||
| No | 52.0 (731/1407) | 28.2 (190/673) | 2.4 (33/1401) | 4.3 (61/1406) | 5.3 (18/340) | 9.9 (140/1408) | ||||||
| Waist circumference (N = 1600) | <.001 | <.001 | <.001 | .018 | .012 | .395 | ||||||
| ≤88 | 36.3 (250/688) | 13.2 (31/235) | 0.9 (6/687) | 2.9 (20/691) | 7.1 (19/268) | 10.7 (97/905) | ||||||
| >88 | 65.9 (598/907) | 35.7 (194/544) | 3.6 (32/902) | 5.3 (48/903) | 1.7 (2/120) | 9.4 (65/690) | ||||||
| Waist/Hip ratio (N = 1548) | 0.88 (0.05) | <.001 | 0.91 (0.03) | <.001 | 0.91 (0.05) | .008 | 0.89 (0.03) | .055 | 0.89 (0.00) | .965 | 0.87 (0.00) | .596 |
| Triglycerides (mg/dL) (N = 1598) | 115.0 (24.3) | <.001 | 128.6 (21.0) | <.001 | 117.5 (13.2) | .142 | 113.0 (8.8) | .198 | 149.4 (36.2) | .097 | 95.7 (–10.0) | .029 |
| Metabolic syndrome (N = 1545) | <.001 | <.001 | .015 | .442 | .049 | .480 | ||||||
| Yes | 70.7 (251/355) | 41.1 (90/219) | 4.2 (15/354) | 5.1 (18/354) | 10.2 (12/118) | 9.3 (33/354) | ||||||
| No | 48.1 (570/1185) | 23.8 (127/534) | 2.0 (23/1180) | 4.1 (49/1148) | 3.5 (9/256) | 10.6 (126/1186) | ||||||
| PHQ-9 depression score (N = 1603) | <.001 | .120 | <.001 | <.001 | .252 | <.001 | ||||||
| <10 | 52.4 (786/1500) | 28.3 (206/727) | 1.9 (29/1495) | 3.7 (56/1499) | 4.7 (16/344) | 9.3 (140/1499) | ||||||
| ≥10 | 68.1 (64/94) | 38.9 (21/54) | 9.7 (9/93) | 12.8 (12/94) | 11.1 (5/45) | 23.2 (22/95) | ||||||
| hsCRP (mg/L) (N = 1575) | 5.8 (2.7) | <.001 | 7.7 (2.7) | <.001 | 6.5 (1.8) | .071 | 5.8 (1.1) | .145 | 9.1 (3.1) | .040 | 5.3 (0.6) | .258 |
Data are presented as % (no./total no.).
Abbreviations: BMI, body mass index; cOSA, clinical symptoms of obstructive sleep apnea; FG, Ferriman–Gallwey; FSH, follicle-stimulating hormone; HOMA, Homeostatic Model Assessment; hsCRP, high-sensitivity C-reactive protein; LH, luteinizing hormone; PCOS, polycystic ovary syndrome; PHQ, Patient Health Questionnaire; SHBG, sex hormone–binding globulin; UI, unexplained infertility.
aChi-squared or Fisher’s exact test was used.
bMean value for those who had ever snored, had a snore frequency 3-7 nights per week, had a cOSA, had a sleep <6 hours, has a pOSA, and had a clinical sleepiness; the value in the parentheses is the difference in the values between those who had ever snored, had a snore frequency 3-7 nights per week, had a cOSA, had a sleep <6 hours, has a pOSA, and had a clinical sleepiness and those who did not.
Table 4 presents the findings of the analysis of reported sleep measures adjusted for other clinical factors in all the PCOS and UI women. Adjusted ORs for study and risk factors for ever-snored, snore frequency of 3 or more nights per week, cOSA, sleep of <6 hours, and clinical sleepiness are shown in the final model obtained using backwards selection for all subjects. Variables left blank were not selected in the final model. Some significant interaction effects of PCOS or UI diagnosis and risk factors were found. For the ease of the interpretation, we have reported the results from the models without interaction. As shown is Table 4, self-reported habitual snoring was significantly associated with PCOS diagnosis (OR 1.69, 95% CI 1.17-2.44, P = .005), insulin >10 mIU/mL (OR 1.82, 95% CI 1.23-2.68, P = .003), WC >88 cm (OR 2.09, 95% CI 1.29-3.38, P = .005), and current smoking (OR 1.84, 95% CI 1.14-2.97, P = .012). Reported duration of sleep of <6 hours per night was associated with both insulin resistance (HOMA >4) (OR, 2.11, 95% CI 1.27-3.50, P = .004), and depression (PHQ ≥ 10) (OR 3.16, 95%CI 1.61-6.21, P < .001). Of note, after adjustment for diagnosis and other clinical factors, perceived clinical sleepiness was associated with PCOS (OR 1.81; 95% CI 1.23-2.66; P = .002), positive PHQ depression screening (OR 2.71; 95% CI 1.59-4.61; P < .001), decreased level of testosterone (OR 0.50; 95% CI 0.32-0.79; P = .003) and decreased level of triglycerides (OR 0.995; 95% CI 0.992-0.999; P < .009) Table 4.
Table 4.
Adjusted odds ratios for study and risk factors for ever snored, snore frequency of 3 or more nights per week, cOSA, sleep of <6 hours, and clinical sleepiness in the final model obtained using backwards selection for both PCOS and UI subjects
| Variable | Ever Snored | Snore frequency 3-7 nights per week | cOSA | Sleep <6 hours | Clinical sleepiness | |||||
|---|---|---|---|---|---|---|---|---|---|---|
| OR (95% CI) | P value | OR (95% CI) | P value | OR (95% CI) | P value | OR (95% CI) | P value | OR (95% CI) | P value | |
| Diagnosis | ||||||||||
| UI | 1 | 1 | 1 | 1 | ||||||
| PCOS | 1.46 (1.06, 2.02) | .022 | 1.69 (1.17, 2.44) | .005 | 2.68 (1.16, 6.22) | .006 | 1.81 (1.23, 2.66) | .002 | ||
| BMI | ||||||||||
| ≤25 | 1 | |||||||||
| 25-35 | 2.66 (2.01, 3.52) | <.001 | ||||||||
| >35 | 6.74 (4.54, 10.02) | <.001 | ||||||||
| HOMA index | ||||||||||
| <4 | 1 | |||||||||
| ≥4 | 2.11 (1.27, 3.50) | .004 | ||||||||
| Glucose | ||||||||||
| ≤100 | 1 | |||||||||
| >100 | 2.43 (1.40, 4.22) | .002 | ||||||||
| Insulin | ||||||||||
| ≤10 | 1 | 1 | ||||||||
| >10 | 1.82 (1.23, 2.68) | .003 | 2.54 (1.18, 5.50) | .010 | ||||||
| Free androgen indexa | 1.05 (1.01, 1.09) | .013 | ||||||||
| Testosterone | ||||||||||
| ≤50 | 1 | |||||||||
| >50 | 0.50 (0.32, 0.79) | .003 | ||||||||
| Androstenedionea | 0.83 (0.75, 0.92) | <.001 | ||||||||
| Smoking | ||||||||||
| Not at all | 1 | |||||||||
| Currently | 1.84 (1.14, 2.97) | .012 | ||||||||
| Quit | 1.20 (0.83, 1.75) | .337 | ||||||||
| Consume alcohol | ||||||||||
| Not at all | 1 | |||||||||
| Currently | 1.59 (1.10, 2.30) | .014 | ||||||||
| Quit | 1.06 (0.69, 1.64) | .789 | ||||||||
| Waist circumference | ||||||||||
| ≤88 | 1 | |||||||||
| >88 | 2.09 (1.29, 3.38) | .003 | ||||||||
| Triglyceridesa | 1.004 (1.001, 1.006) | .005 | 0.995 (0.992, 0.999) | .009 | ||||||
| PHQ-9 depression score | ||||||||||
| <10 | 1 | 1 | 1 | |||||||
| ≥10 | 3.22 (1.44, 7.22) | .005 | 3.16 (1.61, 6.21) | <.001 | 2.71 (1.59, 4.61) | <.001 | ||||
Variables left blank were not selected in the final model. No significant interaction effect of study and risk factors was found.
Abbreviations: cOSA, clinical symptoms of obstructive sleep apnea; PCOS, polycystic ovary syndrome; PHQ, Patient Health Questionnaire; UI, unexplained infertility.
aOdds ratio and 95% CI are for 1-unit change.
Pregnancy Outcomes
The participants in the AMIGOS and PPCOS II clinical trials were followed from baseline through up to 4 ovarian stimulation cycles (AMIGOS) or 5 ovulation induction cycles (PPCOSII) or until pregnancy, and through 6 weeks after delivery in those who conceived. A multivariate analysis of significant factors associated with poor sleep and pregnancy outcomes was performed. The results of reported baseline sleep duration <6 hours nightly and infertility treatment outcomes of conception and live birth are shown in Table 5. Notably, short sleep duration was not associated with conception and live birth rates following treatment in women with PCOS or UI. However, insulin resistance, as indicated by an elevated HOMA-IR, was significantly associated with lower conception (OR 0.69, 95% CI 0.53-0.91, P = .008) and live birth (OR 0.71, 95% CI 0.52-0.96, P = .026) rates in the overall sample. No significant associations were found between ever snored and pregnancy outcome (conception, P = .981 or live birth, P = .662) when adjusted by treatment, age, BMI, baseline glucose, calculated free androgen index, androstenedione, triglyceride levels, and alcohol consumption status. No significant associations were found between cOSA and pregnancy outcome (conception, P = .932; or live birth, P = .129) when adjusted by treatment, age, baseline insulin, and PHQ-9 depression score, nor between clinical daytime sleepiness and conception (P = 0.215) when adjusted by treatment, age, baseline testosterone and triglyceride levels, and PHQ-9 depression score.
Table 5.
Adjusted odds ratios for sleep <6 hours and pregnancy outcomes for both PCOS and UI subjects
| Variable | Conception | Live birth | ||
|---|---|---|---|---|
| OR (95% CI) | P value | OR (95% CI) | P value | |
| Sleep <6 hours | ||||
| No | 1 | 1 | ||
| Yes | 1.08 (0.65, 1.82) | .763 | 0.89 (0.49, 1.62) | .706 |
| Treatment | ||||
| PCOS clomiphene | 1 | 1 | ||
| PCOS letrozole | 1.85 (1.35, 2.52) | <.001 | 1.58 (1.12, 2.23) | .010 |
| UI clomiphene | 1.48 (1.05, 2.13) | .025 | 1.41 (0.95, 2.09) | .092 |
| UI letrozole | 1.05 (0.73, 1.52) | .783 | 1.08 (0.71, 1.62) | .730 |
| UI gonadotropin | 2.43 (1.71, 3.44) | <.001 | 2.20 (1.50, 3.23) | .001 |
| Agea | 0.97 (0.95, 0.998) | .037 | 0.95 (0.92, 0.98) | <.001 |
| HOMA index | ||||
| <4 | 1 | 1 | ||
| ≥4 | 0.69 (0.53, 0.91) | .008 | 0.71 (0.52, 0.96) | .026 |
| PHQ-9 depression scorea | 1.20 (0.77, 1.88) | .418 | 1.06 (0.64, 1.74) | .830 |
Adjusted variables included treatment, age, HOMA index and PHQ-9 Depression Score
Abbreviations: cOSA, clinical symptoms of obstructive sleep apnea; HOMA, Homeostatic Model Assessment; PCOS, polycystic ovary syndrome; PHQ, Patient Health Questionnaire; UI, unexplained infertility.
aOdds ratio and 95% CI are for 1-unit change.
Discussion
This study was performed to characterize and compare sleep habits in women diagnosed with PCOS or UI who completed a sleep habits questionnaire prior to starting treatment for anovulation or unexplained infertility, respectively, in concurrent randomized trials of the Reproductive Medicine Network and were followed through pregnancy and delivery. We found that self-reported sleep disorders, including habitual snoring, short sleep duration <6 hours, increased daytime sleepiness, as well as OSA, were more common and occurred predominantly in women with PCOS than with UI. Notably, cOSA or pOSA was approximately 4 times as prevalent in reproductive aged women with PCOS than in those with UI, though the number with pOSA was low in both groups. After adjustment for clinical factors, PCOS and elevated fasting insulin were associated with cOSA, whereas PCOS, elevated insulin, WC >88 cm, and current smoking were associated with habitual snoring. An elevated HOMA index was associated with short sleep duration <6 hours, and elevated depression risk score and PCOS diagnosis were associated with perceived daytime sleepiness.
Our findings in this very large cohort of reproductive age women are consistent with previous studies showing that insulin resistance and obesity are critical factors affecting disordered sleep. Previous investigators have shown an association of sleep-disordered breathing with obesity and increased WC (13, 32, 33). In addition, shorter sleep duration and higher prevalence of OSA has been associated with central obesity in women (34). Suboptimal sleep (including sleep apnea) has been associated with insulin resistance, systemic inflammation, and increased cardiovascular risk in women with PCOS (8, 13). A recent prospective study of clinical, biochemical, and hormonal status and overnight polysomnography of 50 patients with untreated PCOS and 100 control subjects that included snorers found that sleep-disordered breathing occurred in 66% of the PCOS patients and in 4% of control group (OR 46.5, 95% CI 14.6-148.4; P < .001). After adjustment for BMI and WC, the observed differences in sleep-disordered breathing were not significant, suggesting that obesity is the factor most closely related to clinical sleep disorders (32).
We found that measures of insulin resistance were associated with clinical symptoms of OSA, habitual snoring and shortened sleep duration <6 hours per night. Likewise, obesity and increased WC were associated with snoring and habitual snoring. In contrast, elevated testosterone level was associated with self-reported sleep apnea in all the women (PCOS and UI) in the unadjusted analysis (P = .015) (Table 3).
From this analysis, it is not possible to determine the direction of the association of sleep disturbance and insulin resistance, whether insulin resistance affects sleep OSA or vice versa. Data from previous studies suggest that the relationship between sleep disturbances and metabolic abnormalities may be bidirectional. Sleep disturbances and OSA may augment the severity of insulin resistance in PCOS. Reduced sleep duration and quality have been shown to result in decreased insulin sensitivity in laboratory studies conducted in healthy young nonobese adults. Several small studies suggest that treatment of insulin resistance with metformin improves sleep disorders (35) and treatment of OSA with CPAP may reduce insulin resistance (36), suggesting bidirectionality of effects of insulin resistance and sleep disorders. However, in contrast to previous reports of an increased prevalence of daytime sleepiness in women with PCOS and insulin resistance (14), the current analysis shows that clinical depression rather than insulin resistance as measured by HOMA index is the factor that is significantly associated with daytime sleepiness.
It is challenging to sort out the role of hyperandrogenism in sleep disorders due to the pathophysiologic relationship of hyperandrogenism with insulin resistance and obesity in PCOS. The present study assessed the associations of reported sleep habit measures with clinical and biologic factors in women with PCOS and UI (Tables 1 and 2 (31)), showing that sleep disturbances were more strongly associated with measures of insulin resistance than elevated testosterone. However, hyperandrogenism, evidenced by elevated testosterone levels, may be a contributing factor to clinical sleep disorders by altering the apneic threshold and triggering breathing instability. Of note, testosterone administration in healthy young women resulted in an increase in ventilation, the apneic ventilatory threshold, and sensitivity to carbon dioxide (37). This suggests that the increased endogenous levels of testosterone in women with PCOS may cause changes in the ventilatory sensitivity to carbon dioxide leading to hyperventilation in the face of elevated carbon dioxide levels, thereby lowering carbon dioxide levels below the apneic threshold during sleep, with resultant central apnea (38).
In the analysis of preconception sleep parameters and associated reproductive and metabolic factors with pregnancy outcomes, we found that insulin resistance, as indicated by high HOMA-IR levels, was associated with lower conception and live birth rates. This is a novel finding of this study of women undergoing fertility treatments tracked longitudinally through conception and live birth. Importantly, the presence of reported sleep-disordered breathing prior to pregnancy establishment was not associated with conception and live birth outcomes of fertility treatment. Similarly, short sleep duration did not affect conception and live birth rates related to ovarian stimulation treatment in women with PCOS or UI. To our knowledge, this is the first longitudinal study of pregnancy outcomes after fertility treatment in women who reported sleep parameters prior to conception.
Strengths and Weaknesses
The main strength of this study is the large number of well-characterized populations of women with PCOS and UI in which variables and questionnaires were collected in a systematic fashion. Assays were batched and run in a core lab to eliminate both interindividual and interlab variation. We utilized validated instruments and questionnaires to screen for health abnormalities. For example, the Sleep Habits Questionnaire utilized in this study is a robust instrument previously validated and utilized in the large longitudinal NHLBI Sleep Heart Health Study (28). In addition, we were able to determine pregnancy outcomes in relation to the baseline sleep variables on pregnancy, the most important outcome to infertile women.
The main limitation of this study is that the measures are self-reported and at a single time point. We did not have the resources to conduct polysomnography, the gold standard for diagnosing OSA, nor is it likely that such a large population of women with infertility would have volunteered for these diagnostic studies given their location away from the home and the time requirements. However, self-reported sleep questionnaires have been estimated to have a reasonable sensitivity for the detection of sleep disorders, albeit less consistent specificity (39).
In this study of sleep habits in well-characterized women with either PCOS or UI, sleep disorders were more commonly reported by women with PCOS than UI. As hypothesized, women with PCOS reported a higher prevalence of diagnosed obstructive sleep apnea, as well as shorter sleep duration and habitual snoring. Measures of insulin resistance were associated with clinical symptoms of OSA, habitual snoring, and shortened sleep duration <6 hours per night, whereas obesity and increased WC were associated with snoring and habitual snoring. Furthermore, daytime sleepiness was most strongly associated with screening positive for clinical depression and a PCOS diagnosis, but not with measures of obesity or insulin resistance. Notably, in these populations, pregnancy and live birth outcomes after treatment for infertility were not associated with OSA diagnosis nor other prepregnancy sleep disturbances. Additional prospective longitudinal studies of sleep habits in women before conception, during pregnancy, and measurement of the pregnancy complications and birth outcomes would facilitate a better understanding of the role of sleep in reproductive health.
Acknowledgments
The authors thank the members and principal investigators of the Reproductive Medicine Network for their contributions to the original trials upon which this research is based. In addition to the authors, other members of the National Institute of Child Health and Human Development Reproductive Medicine Network were the following: University of Pennsylvania: C. Coutifaris; University of Florida: G. Christman; University of Texas Health Science Center at San Antonio: R. Robinson, R. Brzyski; University of Colorado: W. Schlaff; University of Vermont: P. Casson; SUNY Upstate Medical University: J.C. Trussell; and Eunice Kennedy Shriver National Institute of Child Health and Human Development: C. Lamar, L. DePaolo. The authors thank Drs. Rebecca Usadi, R. Mitchell Rosen, Valerie Baker for recruitment efforts on behalf of the Reproductive Medicine Network.
Financial Support: This work was supported by National Institutes of Health (NIH)/Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) Grants U10 HD39005 (to M.P.D.), U10 HD38992 (to R.S.L.), U10 HD27049 (to C.C.), U10 HD38998 (to R.A.), HD055944 (to P.R.C.), U10 HD055936 (to Y.R.S.), U10HD055925 (to H.Z.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NICHD or NIH.
Clinical Trial Information: Pregnancy in Polycystic Ovary Syndrome II (PPCOS II), NCT00719186 (registered July 17, 2008). Assessment of Multiple Intrauterine Gestations in Ovulation Stimulation (AMIGOS) NCT01044862 (registered January 7, 2010). clinicaltrials.gov.
Glossary
Abbreviations
- BMI
body mass index
- cOSA
clinical symptoms of obstructive sleep apnea
- FG
Ferriman–Gallwey
- FSH
follicle-stimulating hormone
- hsCRP
high-sensitivity C-reactive protein
- LH
luteinizing hormone
- OSA
obstructive sleep apnea
- PCOS
polycystic ovary syndrome
- PHQ
Patient Health Questionnaire
- pOSA
prior diagnosis of obstructive sleep apnea
- SHBG
sex hormone–binding globulin
- UI
unexplained infertility
- WC
waist circumference
Additional Information
Disclosures: Dr. Eisenberg- Federal Employee; Dr. Richard Legro—Consultant for Fractyl, Abbvie, Bayer; Ferring and Guerbet funding; Dr. Michael P. Diamond—NIH funding, AbbVie, Bayer and ObsEva funding, Board of Directors and Stockholder for Advanced Reproductive Care; Dr. Hao Huang—No conflicts; Dr. Christos Coutifaris—NIH funding; Dr. Louise O’Brien—NIH funding, Scientific Advisory Board for Star Legacy Foundation and Smart Human Dynamics; Dr. Yolanda Smith—No conflicts; Dr. Karl R. Hansen—NIH funding, Roche Diagnostics, Ferring International Pharmascience Center US, and AblaCare funding; Dr. Nanette Santoro—consultant for Ansh Labs, Scientific Advisory Board Menogenix, Inc. and Astellas; Dr. Heping Zhang—NIH funding
Data Availability
The datasets generated and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request. The PPCOSII and AMIGOS clinical trial datasets include parts of the data used for this analysis. These datasets are deposited in the NICHD Data and Specimen Hub (DASH), https://dash.nichd.nih.gov/.
References
- 1.Pamidi S, Pinto LM, Marc I, Benedetti A, Schwartzman K, Kimoff RJ. Maternal sleep-disordered breathing and adverse pregnancy outcomes: a systematic review and metaanalysis. Am J Obstet Gynecol. 2014;210(1):52.e1-52.e14. [DOI] [PubMed] [Google Scholar]
- 2.Lateef OM, Akintubosun MO. Sleep and reproductive health. J Circadian Rhythms. 2020;18:1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Kloss JD, Perlis ML, Zamzow JA, Culnan EJ, Gracia CR. Sleep, sleep disturbance, and fertility in women. Sleep Med Rev. 2015;22:78-87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Willis SK, Hatch EE, Wise LA. Sleep and female reproduction. Curr Opin Obstet Gynecol. 2019;31(4):222-227. [DOI] [PubMed] [Google Scholar]
- 5.Warland J, Dorrian J, Morrison JL, O’Brien LM. Maternal sleep during pregnancy and poor fetal outcomes: A scoping review of the literature with meta-analysis. Sleep Med Rev. 2018;41:197-219. [DOI] [PubMed] [Google Scholar]
- 6.O’Brien LM. Sleep disordered breathing in pregnancy. In: Current Sleep Medicine Reports; Women and Sleep, vol. 2. Amy Wolfson and Kathryn Sharkey: 2016:183-190. [Google Scholar]
- 7.Michels KA, Mendola P, Schliep KC, et al. The influences of sleep duration, chronotype, and nightwork on the ovarian cycle. Chronobiol Int. 2020;37(2):260-271. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Usadi RS, Legro RS. Reproductive impact of polycystic ovary syndrome. Curr Opin Endocrinol Diabetes Obes. 2012;19(6):505-511. [DOI] [PubMed] [Google Scholar]
- 9.Ehrmann DA. Polycystic ovary syndrome. N Engl J Med. 2005;352(12):1223-1236. [DOI] [PubMed] [Google Scholar]
- 10.Cooney LG, Dokras A. Beyond fertility: polycystic ovary syndrome and long-term health. Fertil Steril. 2018;110(5):794-809. [DOI] [PubMed] [Google Scholar]
- 11.Dunaif A, Segal KR, Futterweit W, Dobrjansky A. Profound peripheral insulin resistance, independent of obesity, in polycystic ovary syndrome. Diabetes. 1989;38(9):1165-1174. [DOI] [PubMed] [Google Scholar]
- 12.Tasali E, Van Cauter E, Ehrmann DA. Relationships between sleep disordered breathing and glucose metabolism in polycystic ovary syndrome. J Clin Endocrinol Metab. 2006;91(1):36-42. [DOI] [PubMed] [Google Scholar]
- 13.Fogel RB, Malhotra A, Pillar G, Pittman SD, Dunaif A, White DP. Increased prevalence of obstructive sleep apnea syndrome in obese women with polycystic ovary syndrome. J Clin Endocrinol Metab. 2001;86(3):1175-1180. [DOI] [PubMed] [Google Scholar]
- 14.Vgontzas AN, Legro RS, Bixler EO, Grayev A, Kales A, Chrousos GP. Polycystic ovary syndrome is associated with obstructive sleep apnea and daytime sleepiness: role of insulin resistance. J Clin Endocrinol Metab. 2001;86(2):517-520. [DOI] [PubMed] [Google Scholar]
- 15.Gopal M, Duntley S, Uhles M, Attarian H. The role of obesity in the increased prevalence of obstructive sleep apnea syndrome in patients with polycystic ovarian syndrome. Sleep Med. 2002;3(5):401-404. [DOI] [PubMed] [Google Scholar]
- 16.Vgontzas AN, Bixler EO, Chrousos GP. Sleep apnea is a manifestation of the metabolic syndrome. Sleep Med Rev. 2005;9(3):211-224. [DOI] [PubMed] [Google Scholar]
- 17.Gaines J, Vgontzas AN, Fernandez-Mendoza J, Bixler EO. Obstructive sleep apnea and the metabolic syndrome: The road to clinically-meaningful phenotyping, improved prognosis, and personalized treatment. Sleep Med Rev. 2018;42:211-219. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Chatterjee B, Suri J, Suri JC, Mittal P, Adhikari T. Impact of sleep-disordered breathing on metabolic dysfunctions in patients with polycystic ovary syndrome. Sleep Med. 2014;15(12):1547-1553. [DOI] [PubMed] [Google Scholar]
- 19.Santoro N, Eisenberg E, Trussell JC, et al. ; Reproductive Medicine Network Investigators . Fertility-related quality of life from two RCT cohorts with infertility: unexplained infertility and polycystic ovary syndrome. Hum Reprod. 2016;31(10):2268-2279. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Legro RS, Zhang H; Eunice Kennedy Shriver NICHD Reproductive Medicine Network . Letrozole or clomiphene for infertility in the polycystic ovary syndrome. N Engl J Med. 2014;371(15):1463-1464. [DOI] [PubMed] [Google Scholar]
- 21.Diamond MP, Legro RS, Coutifaris C, et al. ; NICHD Reproductive Medicine Network . Letrozole, Gonadotropin, or Clomiphene for Unexplained Infertility. N Engl J Med. 2015;373(13):1230-1240. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Legro RS, Kunselman AR, Brzyski RG, et al. ; NICHD Reproductive Medicine Network . The Pregnancy in Polycystic Ovary Syndrome II (PPCOS II) trial: rationale and design of a double-blind randomized trial of clomiphene citrate and letrozole for the treatment of infertility in women with polycystic ovary syndrome. Contemp Clin Trials. 2012;33(3):470-481. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Legro RS, Brzyski RG, Diamond MP, et al. ; National Institute of Child Health and Human Development Reproductive Medicine Network . The Pregnancy in Polycystic Ovary Syndrome II study: baseline characteristics and effects of obesity from a multicenter randomized clinical trial. Fertil Steril. 2014;101(1):258-269.e8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Diamond MP, Mitwally M, Casper R, et al. ; NICHD Cooperative Reproductive Medicine Network . Estimating rates of multiple gestation pregnancies: sample size calculation from the assessment of multiple intrauterine gestations from ovarian stimulation (AMIGOS) trial. Contemp Clin Trials. 2011;32(6):902-908. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Diamond MP, Legro RS, Coutifaris C, et al. ; National Institute of Child Health and Human Development (NICHD) Reproductive Medicine Network . Assessment of multiple intrauterine gestations from ovarian stimulation (AMIGOS) trial: baseline characteristics. Fertil Steril. 2015;103(4):962-973.e4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Usadi RS, Diamond MP, Legro RS, et al. ; Reproductive Medicine Network . Recruitment strategies in two reproductive medicine network infertility trials. Contemp Clin Trials. 2015;45(Pt B):196-200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Ferriman D, Gallwey JD. Clinical assessment of body hair growth in women. J Clin Endocrinol Metab. 1961;21(11):1440-1447. [DOI] [PubMed] [Google Scholar]
- 28.Kump K, Whalen C, Tishler PV, et al. Assessment of the validity and utility of a sleep-symptom questionnaire. Am J Respir Crit Care Med. 1994;150(3):735-741. [DOI] [PubMed] [Google Scholar]
- 29.Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med. 2001;16(9):606-613. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Moriarty AS, Gilbody S, McMillan D, Manea L. Screening and case finding for major depressive disorder using the Patient Health Questionnaire (PHQ-9): a meta-analysis. Gen Hosp Psychiatry. 2015;37(6):567-576. [DOI] [PubMed] [Google Scholar]
- 31.Eisenberg E, Legro RS, Diamond MP, et al. Data from: Sleep habits of women with Infertility. figshare. Deposited May 28, 2021. doi: 10.6084/m9.figshare.14699868.v1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Suri J, Suri JC, Chatterjee B, Mittal P, Adhikari T. Obesity may be the common pathway for sleep-disordered breathing in women with polycystic ovary syndrome. Sleep Med. 2016;24:32-39. [DOI] [PubMed] [Google Scholar]
- 33.Young T, Skatrud J, Peppard PE. Risk factors for obstructive sleep apnea in adults. Jama. 2004;291(16):2013-2016. [DOI] [PubMed] [Google Scholar]
- 34.Theorell-Haglöw J, Berglund L, Janson C, Lindberg E. Sleep duration and central obesity in women – differences between short sleepers and long sleepers. Sleep Med. 2012;13(8):1079-1085. [DOI] [PubMed] [Google Scholar]
- 35.El-Sharkawy AA, Abdelmotaleb GS, Aly MK, Kabel AM. Effect of metformin on sleep disorders in adolescent girls with polycystic ovarian syndrome. J Pediatr Adolesc Gynecol. 2014;27(6):347-352. [DOI] [PubMed] [Google Scholar]
- 36.Tasali E, Chapotot F, Leproult R, Whitmore H, Ehrmann DA. Treatment of obstructive sleep apnea improves cardiometabolic function in young obese women with polycystic ovary syndrome. J Clin Endocrinol Metab. 2011;96(2):365-374. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Ahuja D, Mateika JH, Diamond MP, Badr MS. Ventilatory sensitivity to carbon dioxide before and after episodic hypoxia in women treated with testosterone. J Appl Physiol (1985). 2007;102(5):1832-1838. [DOI] [PubMed] [Google Scholar]
- 38.Zhou XS, Rowley JA, Demirovic F, Diamond MP, Badr MS. Effect of testosterone on the apneic threshold in women during NREM sleep. J Appl Physiol (1985). 2003;94(1):101-107. [DOI] [PubMed] [Google Scholar]
- 39.Ibáñez V, Silva J, Cauli O. A survey on sleep assessment methods. PeerJ. 2018;6:e4849. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The datasets generated and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request. The PPCOSII and AMIGOS clinical trial datasets include parts of the data used for this analysis. These datasets are deposited in the NICHD Data and Specimen Hub (DASH), https://dash.nichd.nih.gov/.