Causal evidence of the relationship between polycystic ovary syndrome and obstructive sleep apnea in European and East Asian populations: a two-sample Mendelian randomization study

Introduction

Polycystic ovary syndrome (PCOS) is classified as an endocrine disorder characterized by irregular menstrual cycles, hyperandrogenism (HA), and changes in ovarian morphology (1). PCOS may have long-term effects on various health issues, including infertility, cardiovascular disease (CAD), insulin resistance, psychiatric disorders, and low-grade inflammation (2). The underlying pathogenic mechanisms involve elevated androgen levels, abnormal neuronal-reproductive-metabolic circuits (3), gut microbiota disturbances (4), and mitochondrial dysfunction (5), although these mechanisms are not yet fully understood.

In modern society, obstructive sleep apnea (OSA), defined by recurrent episodes of upper airway obstruction and intermittent hypoxemia (6), is a highly prevalent condition. Comprehensive data indicate that the overall prevalence of OSA in the general population ranges from 9% to 38% (7), influenced by factors such as body mass index (BMI), sex, age, and diagnostic criteria. OSA has emerged as a significant health issue and serves as an independent risk factor for cardiovascular, metabolic, and psychiatric disorders, including hypertension, stroke, diabetes, and depression (8,9). A recent meta-analysis involving 17 studies and 648 participants highlighted the high prevalence of OSA among women with PCOS (10). However, the considerable genetic and clinical heterogeneity of PCOS, along with common confounders like high BMI, HA, and insulin resistance, have not been fully accounted for. Additionally, due to the cross-sectional nature of existing studies, it remains unclear whether PCOS and OSA are merely associated or causally related.

Mendelian randomization (MR) has emerged as a robust method for assessing causal effects in epidemiology. The core principle of MR involves using genetic variants as instrumental variables (IVs) to estimate causal relationships between exposures and outcomes in an unbiased manner. Genetic variants, such as single nucleotide polymorphisms (SNPs), remain stable during meiosis and conception, rendering them independent of confounding factors (11). MR analyses have successfully established causal relationships between PCOS and ovarian cancer (11), BMI (12), and psychiatric disorders (13). However, the applicability of MR results derived from genetic instruments based on European ancestry may be limited for other populations. PCOS is a complex endocrine disorder with intricate genetic patterns, and geographic and ethnic variations can influence its clinical presentation (14). Notable racial and ethnic differences in PCOS phenotypes and associated metabolic dysfunctions have been identified (15). Yet, most studies exploring the relationship between PCOS and OSA have been conducted in the United States, leading to inconsistent conclusions. While Ibrahim et al. (16) posit that there is no relationship between the two, Eisenberg et al. (17) and Christ et al. (18) support a correlation. Additionally, researchers in Asia, particularly China, have reported findings that corroborate this correlation, as evidenced by studies from Yang et al. (19) and Zhang et al. (20). Given this context, further evaluation of the impact of PCOS on OSA among women of diverse ethnic backgrounds using the MR approach is warranted.

In this study, we utilized SNPs and summary datasets from European and East Asian participants to explore the causality between PCOS and OSA. We present this article in accordance with the STROBE-MR reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-24-885/rc).

Methods

We conducted a two-sample MR study using IVs predicting PCOS and OSA from the latest genome-wide association studies (GWAS). Figure 1 shows an overview of the study description. MR relies on the following three assumptions to ensure the accuracy and robustness of the causal link: i.e., the genetic predictors of exposure selected as IVs are strongly associated with the exposure; the genetic variants are independent of confounders; and the genetic variants are only linked to the outcome through affecting the exposure of interest, not through alternative pathways (21). The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013).

Figure 1 An overview of MR study design and assumption. (A) Flowchart illustrating the process for the MR analyses, including data collection, IVs selection, and statistical analysis. (B) Directed acyclic graph illustrating the assumptions of the MR methodology. MR relies on three assumptions: the genetic variants selected as instruments must (I) be associated with the exposure robustly, (II) be independent of confounders, and (III) be significantly related to the risk of the outcome only through their effect on the exposure. PCOS, polycystic ovary syndrome; GWAS, genome-wide association studies; OSA, obstructive sleep apnea; IVs, instrumental variables; SNPs, single nucleotide polymorphisms; MR, Mendelian randomization; IVW, inverse variance weighted.

Selection of IVs for MR analyses

To ensure the authenticity and accuracy of the conclusions regarding the causal link between PCOS and OSA, we implemented several quality control steps to select optimal IVs.

First, all IVs for the MR analysis reached genome-wide statistical significance (P<5×10⁻⁸), indicating a strong association with PCOS. Second, to maintain the independence of each IV, we performed a clumping procedure (R²=0.001, Kb =10,000) to remove linkage disequilibrium (LD) between SNPs, retaining only the SNP with the smallest P value. Third, we calculated the F-statistic for each SNP to estimate the strength of the instrument associated with the exposure. An F-statistic value of less than 10 in MR analyses suggests no significant association between the IV and PCOS.

Data sources and IVs selection for PCOS

In European female populations, we utilized meta-analysis data comprising seven cohorts of European descent for PCOS, which included 10,074 cases and 103,164 controls (22). Among the seven articles included in this meta-analysis, one study diagnosed PCOS cases through self-report. In contrast, the remaining studies used the National Institutes of Health (NIH) or Rotterdam criteria for diagnosis. The NIH criteria require both ovulatory dysfunction (OD) and clinical and/or biochemical HA for a diagnosis of PCOS (23). The Rotterdam criteria necessitate the presence of 2 out of 3 features: (I) OD defined by oligo- or amenorrhea; (II) clinical and/or biochemical HA; and/or (III) polycystic ovarian morphology (PCOM) for a diagnosis of PCOS (24). In East Asian female populations, we analyzed meta-analysis data from two cohorts of Asian descent for PCOS, consisting of 2,254 cases and 3,001 controls (25). All Han Chinese samples were obtained from multiple collaborating hospitals in China, and cases in this meta-analysis were diagnosed based on the Rotterdam criteria.

Data sources and IVs selection for OSA

The diagnosis of OSA was established according to the International Classification of Diseases, Tenth Revision (ICD-10) and Ninth Revision (ICD-9) codes (ICD-10: G47.3, ICD-9: 3472 A). This determination was based on subjective symptoms, clinical assessments, and polysomnographic evaluations, utilizing an Apnea-Hypopnea Index (AHI) of ≥5 events per hour or a respiratory event index of ≥5 events per hour.

The FinnGen project (www.finngen.fi/en) is a comprehensive biobank study that includes both prospective and retrospective epidemiological cohorts, as well as disease-based cohorts and hospital biobank samples from individuals of European descent. In European populations, GWAS summary statistics for OSA were obtained from the FinnGen consortium’s R4 release data, which included 11,937 cases and 164,295 controls. In East Asian populations, GWAS data were sourced from The BioBank Japan Project (BBJ), comprising 473 cases and 177,864 controls. The BBJ collects DNA and serum samples from 12 Japanese medical institutions and encompasses a large cohort of approximately 200,000 participants (26,27). Detailed information regarding the GWAS datasets used in our study is presented in Table 1.

Statistical analysis

The inverse variance weighted (IVW) method is used to obtain accurate causal estimates when the pleiotropic effects are balanced and all IVs satisfy the MR assumptions (28). In this study, IVW was the primary statistical method employed. Due to the challenge of ruling out the influence of IVs on outcomes through alternative pathways, a series of sensitivity analyses with different assumptions were conducted to assess the robustness of the associations and to examine for horizontal pleiotropy. The sensitivity analyses included the weighted median, MR-Egger, and weighted mode methods. The weighted median of SNP-specific estimates produces valid results when over 50% of the information is derived from the IVs (29). While MR-Egger regression was not used to generate causal estimates due to its low statistical power, it can detect and adjust for unbalanced horizontal pleiotropy. A significant MR-Egger intercept (P<0.05) indicates the presence of directional pleiotropy and a potentially biased IVW estimate.

To further assess the robustness of our results, we employed Cochran’s Q test to assess heterogeneity among SNPs included in each analysis. If the Q statistic is significant at P<0.05, it suggests heterogeneity among individual genetic variants and the presence of invalid instruments (30). Additionally, a leave-one-out analysis was performed to determine whether the results were influenced by a single outlying SNP (31), indicating heterogeneous SNPs. Furthermore, if the genetic variants do not account for a significant amount of variation, there may be considerable bias in the estimate due to weak IVs (32). To address this concern, SNP-specific F-statistics, approximated by the square of the beta divided by the variance for the SNP-exposure association, were calculated. An F-statistic exceeding the standard threshold 10 indicates strong genetic instruments (28).

All tests were two-sided and performed using R Version 4.0.3 with the R packages “Two sample MR” and “Mendelian Randomization”. A P value <0.05 was considered statistically significant for the MR effect estimate.

Results

GWAS of PCOS

In European female populations, 12 SNPs associated with PCOS were selected for the MR analyses according to the IV selection criteria (Table S1). In East Asian female populations, 13 SNPs associated with PCOS were selected using the same criteria (Table S2). The details of these SNPs, including the strength and magnitude of their associations with PCOS, are presented in Tables S1,S2. Overall, the selected SNPs exhibited a high F-statistic (>10), indicating strong genetic instruments.

MR analysis

Figure 2 illustrates the forest plots of the causal effect estimates of PCOS on OSA using four different MR methods. In the IVW analysis, the odds ratio (OR) for PCOS affecting OSA was 1.133 [95% confidence interval (CI): 1.037–1.239, P=0.006], indicating that PCOS significantly increases the risk of OSA in the European population. Conversely, the IVW analysis found no causal effect of PCOS on OSA in the East Asian population (OR =1.061, 95% CI: 0.888–1.268, P=0.51).

Figure 2 Forest plots for the causal association between PCOS and OSA. IVW, inverse variance weighted; MR, Mendelian randomization; OR, odds ratio; CI, confidence interval; PCOS, polycystic ovary syndrome; OSA, obstructive sleep apnea.

Sensitivity analysis

Figure 3 presents scatter plots depicting the associations between SNPs and OSA against those with PCOS, allowing for the visualization of causal effect estimates for each individual SNP on OSA. Figure 4 displays the results of the leave-one-out analysis, illustrating the observed associations after the removal of each SNP one at a time. Figure 5 features funnel plots to identify potential outliers.

Figure 3 Scatter plot for the causal association between PCOS and OSA. MR, Mendelian randomization; SNP, single nucleotide polymorphism; PCOS, polycystic ovary syndrome; OSA, obstructive sleep apnea.

Figure 4 Leave-one-out regression analysis for the causal association between PCOS and OSA. MR, Mendelian randomization; PCOS, polycystic ovary syndrome; OSA, obstructive sleep apnea.

Figure 5 Funnel plot for the causal association between PCOS and OSA. MR, Mendelian randomization; SE, standard error; PCOS, polycystic ovary syndrome; OSA, obstructive sleep apnea.

The significance of the results was dismissed for the associations between PCOS and OSA in both European (OR =1.097, P=0.09) and East Asian populations (OR =1.085, P=0.51). Additionally, there were no indications of heterogeneity or directional pleiotropy, as assessed by the Cochran’s Q test (P_IVW =0.20; P_MR-Egger =0.17) and the MR-Egger test (P=0.52) (Table 2). The leave-one-out analysis estimates suggested that the observed associations remained consistent even after excluding each individual SNP, indicating that the results were not driven by any single SNP. No outliers were visually identified in the scatter plot (Figure 3), and the funnel plot appeared roughly symmetrical, lacking obvious heterogeneity (Figure 5).

Discussion

Numerous studies have observed that women with PCOS may have an increased risk of OSA. However, important confounding factors affecting OSA risk have not been adequately addressed in these studies. The purpose of our study was to conduct a two-sample MR analysis to evaluate the causal relationship between PCOS and OSA. To differentiate the effects of PCOS on OSA across diverse ethnicities, we utilized large-scale GWAS data from European and East Asian participants. Our findings indicate an association between PCOS and OSA in the European population, suggesting that European women with PCOS are at an increased risk for OSA. Conversely, no association was found between PCOS and OSA in the East Asian population in this MR study. To our knowledge, this is the first MR research exploring the causal association between PCOS and OSA among women of diverse ethnic backgrounds.

PCOS is the most prevalent endocrine disorder among women of reproductive age, affecting approximately 7% of reproductive-aged women (33). Obesity, a well-recognized risk factor for OSA, is commonly observed in women with PCOS (7,34). Thus, it is plausible that OSA and PCOS may coexist, with PCOS potentially contributing to the comorbidities associated with OSA (35). Several small-scale cross-sectional studies have investigated the prevalence of OSA in women with PCOS. A recent systematic review further supports this, revealing a markedly higher prevalence of OSA (35%, 95% CI: 22.2–48.9%) in women with PCOS compared to controls (OR =3.83, 95% CI: 1.43–10.24) (35). The mechanisms linking obesity to OSA involve various factors, including a smaller upper airway due to increased parapharyngeal fat deposition, altered neural compensatory responses, reduced lung volume, and increased breathing workload (36). Beyond weight gain, other factors such as insulin resistance, HA, low progesterone, and elevated oxidative stress may also elevate the risk of OSA in women with PCOS (35). A large longitudinal study assessing the relationship between PCOS and OSA in a European population found that women with PCOS exhibited a higher risk of developing OSA across all three BMI categories [BMI <25 kg/m²: OR =1.91 (95% CI: 0.92–3.97), P=0.08; BMI =25–29.99 kg/m²: OR =2.25 (95% CI: 1.33–3.81), P<0.001; BMI ≥30 kg/m²: OR =2.10 (95% CI: 1.72–2.56), P<0.001], with stronger associations noted in overweight or obese women (37). In addition to BMI, factors such as anovulation, hirsutism, and treatment with metformin were identified as independent predictors of OSA development in women with PCOS (37). Hirsutism, a clinical feature of hyperandrogenemia and a hallmark of PCOS was noted in a study by Chatterjee et al. (38), which found that women with PCOS and sleep-disordered breathing (SDB) exhibited more severe HA compared to those without sleep issues. Furthermore, Dexter et al. (39) reported a case of a non-obese woman with a testosterone-secreting ovarian tumor who presented with clinically significant OSA, which resolved post-surgery. This case supports the theory that the association between PCOS and OSA may be partly due to elevated testosterone levels, which significantly influence pharyngeal patency and ventilatory drive (35). On the other hand, a study involving 53 women with PCOS and 452 control premenopausal women found that PCOS patients were more likely to suffer from SDB, with insulin resistance identified as a stronger risk factor than BMI or testosterone. This underscores the independent association of insulin resistance with an increased risk of OSA in women with PCOS, as more than 50% of women with PCOS experience insulin resistance (40). A prospective study indicated that healthy lean men with OSA had 27% lower insulin sensitivity and 37% higher total insulin secretion compared to control subjects, suggesting an association between OSA and insulin resistance (41). Additionally, insulin may enhance androgen production by acting directly on the ovaries (42) or increase bioavailable testosterone by reducing sex hormone-binding globulin (SHBG) production in the liver (43), indicating an indirect regulatory role of insulin in modulating androgen levels.

Our study provides evidence of a causal relationship between PCOS and OSA in the European population, while no association was observed in the East Asian population. This discrepancy may be attributed to significant variations in both PCOS phenotypes and OSA incidence between these populations, underscoring the need for population-specific investigations. Several studies have highlighted global differences in PCOS phenotypes among women from various racial and ethnic groups, particularly concerning the prevalence of obesity, insulin resistance, and diabetes mellitus. For instance, Lo et al. conducted a large community-based retrospective study in Northern California, which included 12,734 women with PCOS. They found that the incidence of obesity was lower in Asian women with PCOS compared to White women, even after adjusting for confounding factors such as age and BMI (44). Additionally, hirsutism in East Asian women with PCOS has been observed to be less pronounced (45). Conversely, ethnic differences in OSA prevalence have also been reported, with certain groups showing increased risk. A community-based survey indicated a lower prevalence of snoring and sleep apnea among ethnic Chinese compared to Malays and Indians (46). Furthermore, studies have demonstrated a higher frequency and greater severity of OSA in African Americans compared to European Americans (47-49). These ethnic variations may result from differences in upper airway anatomy and possibly physiological factors, such as craniofacial morphology (50).

Our study incorporated several strengths. A major advantage was the two-sample MR design, which utilized large GWAS summary data, helping to mitigate the influence of potential confounding factors. Additionally, we considered cross-ancestry factors that could affect genetic and lifestyle variables, allowing us to capture population-specific associations between PCOS and OSA. However, there are some limitations in this study. First, while the two-sample MR analysis used GWAS summary datasets, it was not possible to assess non-linear correlations between exposures and outcomes. Second, the OSA datasets included both male and female participants, which may lead to collider bias (51). Finally, although this study primarily focused on the causal relationship between PCOS and OSA, the underlying mechanisms remain to be elucidated.

Conclusions

In conclusion, our study found an association between PCOS and OSA in the European population, indicating that European women with PCOS are at an increased risk for OSA. However, no association was observed between PCOS and OSA in the East Asian population within this Mendelian randomization study. Clinicians should maintain a high index of suspicion for OSA in women with PCOS.

Acknowledgments

Funding: This research was funded by the National Key Research and Development Program of China (No. 2022YFC2703803), National Natural Science Foundation of China (No. 82088102), CAMS Innovation Fund for Medical Sciences (No. 2019-I2M-5-064), Collaborative Innovation Program of Shanghai Municipal Health Commission (No. 2020CXJQ01), Key Discipline Construction Project (2023-2025) of Three-Year Initiative Plan for Strengthening Public Health System Construction in Shanghai (No. GWVI-11.1-35), Shanghai Clinical Research Center for Gynecological Diseases (No. 22MC1940200), Shanghai Urogenital System Diseases Research Center (No. 2022ZZ01012), and Shanghai Frontiers Science Research Center of Reproduction and Development.

Footnote

Reporting Checklist: The authors have completed the STROBE-MR reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-24-885/rc

Peer Review File: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-24-885/prf

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-24-885/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). No patients were involved in the design, recruitment, or conduct of this study, so ethical approval was not needed.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Zierau L, Cortes R, Thomsen SF, et al. Asthma severity and fertility outcome in women with polycystic ovary syndrome: a registry-based study. ERJ Open Res 2018;4:00138-2017. [Crossref] [PubMed]
2. Yu JN, Nam GE, Han K, et al. Association between menstrual cycle irregularity and tinnitus: a nationwide population-based study. Sci Rep 2019;9:14038. [Crossref] [PubMed]
3. Liao B, Qiao J, Pang Y. Central Regulation of PCOS: Abnormal Neuronal-Reproductive-Metabolic Circuits in PCOS Pathophysiology. Front Endocrinol (Lausanne) 2021;12:667422. [Crossref] [PubMed]
4. He FF, Li YM. Role of gut microbiota in the development of insulin resistance and the mechanism underlying polycystic ovary syndrome: a review. J Ovarian Res 2020;13:73. [Crossref] [PubMed]
5. Zeng X, Huang Q, Long SL, et al. Mitochondrial Dysfunction in Polycystic Ovary Syndrome. DNA Cell Biol 2020;39:1401-9. [Crossref] [PubMed]
6. Sforza E, Thomas T, Barthélémy JC, et al. Obstructive sleep apnea is associated with preserved bone mineral density in healthy elderly subjects. Sleep 2013;36:1509-15. [Crossref] [PubMed]
7. Redline S, Azarbarzin A, Peker Y. Obstructive sleep apnoea heterogeneity and cardiovascular disease. Nat Rev Cardiol 2023;20:560-73. [Crossref] [PubMed]
8. Eastwood PR, Malhotra A, Palmer LJ, et al. Obstructive Sleep Apnoea: From pathogenesis to treatment: Current controversies and future directions. Respirology 2010;15:587-95. [Crossref] [PubMed]
9. Park JG, Ramar K, Olson EJ. Updates on definition, consequences, and management of obstructive sleep apnea. Mayo Clin Proc 2011;86:549-54; quiz 554-5. [Crossref] [PubMed]
10. Kahal H, Kyrou I, Uthman OA, et al. The prevalence of obstructive sleep apnoea in women with polycystic ovary syndrome: a systematic review and meta-analysis. Sleep Breath 2020;24:339-50. [Crossref] [PubMed]
11. Soliman GA, Schooling CM. Causal association between mTOR-dependent EIF-4E and EIF-4A circulating protein levels and type 2 diabetes: a Mendelian randomization study. Sci Rep 2020;10:15737. [Crossref] [PubMed]
12. Brower MA, Hai Y, Jones MR, et al. Bidirectional Mendelian randomization to explore the causal relationships between body mass index and polycystic ovary syndrome. Hum Reprod 2019;34:127-36. [Crossref] [PubMed]
13. Jin L, Yu J, Chen Y, et al. Polycystic Ovary Syndrome and Risk of Five Common Psychiatric Disorders Among European Women: A Two-Sample Mendelian Randomization Study. Front Genet 2021;12:689897. [Crossref] [PubMed]
14. Lizneva D, Suturina L, Walker W, et al. Criteria, prevalence, and phenotypes of polycystic ovary syndrome. Fertil Steril 2016;106:6-15. [Crossref] [PubMed]
15. VanHise K, Wang ET, Norris K, et al. Racial and ethnic disparities in polycystic ovary syndrome. Fertil Steril 2023;119:348-54. [Crossref] [PubMed]
16. Ibrahim S, Mehra R, Tantibhedhyangkul J, et al. Sleep and obstructive sleep apnea in women with infertility. Sleep Breath 2023;27:1733-42. [Crossref] [PubMed]
17. Eisenberg E, Legro RS, Diamond MP, et al. Sleep Habits of Women With Infertility. J Clin Endocrinol Metab 2021;106:e4414-26. [Crossref] [PubMed]
18. Christ JP, Shinkai K, Corley J, et al. Metabolic and endocrine status associate with obstructive sleep apnea risk among patients with polycystic ovary syndrome. J Clin Sleep Med 2024;20:871-7. [Crossref] [PubMed]
19. Yang R, Gao C, Yan Y, et al. Analysis of the proportion and clinical characteristics of obstructive sleep apnea in women with polycystic ovary syndrome. Sleep Breath 2022;26:497-503. [Crossref] [PubMed]
20. Zhang Q, Wang Z, Ding J, et al. Effect of obstructive sleep apnea on in vitro fertilization outcomes in women with polycystic ovary syndrome. J Clin Sleep Med 2024;20:31-8. [Crossref] [PubMed]
21. Sekula P, Del Greco M F, Pattaro C, et al. Mendelian Randomization as an Approach to Assess Causality Using Observational Data. J Am Soc Nephrol 2016;27:3253-65. [Crossref] [PubMed]
22. Day F, Karaderi T, Jones MR, et al. Large-scale genome-wide meta-analysis of polycystic ovary syndrome suggests shared genetic architecture for different diagnosis criteria. PLoS Genet 2018;14:e1007813. Erratum in: PLoS Genet 2019;15:e1008517. [Crossref] [PubMed]
23. Zawadzki JK, Dunaif A. Diagnostic criteria for polycystic ovary syndrome: toward a rational approach. In: Dunaif A, Givens JR, Haseltine F, et al. editors. Polycystic ovary syndrome. Boston: Blackwell Scientific Publications; 1992:377-84.
24. Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome. Fertil Steril 2004;81:19-25. [Crossref] [PubMed]
25. Shi Y, Zhao H, Shi Y, et al. Genome-wide association study identifies eight new risk loci for polycystic ovary syndrome. Nat Genet 2012;44:1020-5. [Crossref] [PubMed]
26. Nagai A, Hirata M, Kamatani Y, et al. Overview of the BioBank Japan Project: Study design and profile. J Epidemiol 2017;27:S2-8. [Crossref] [PubMed]
27. Hirata M, Kamatani Y, Nagai A, et al. Cross-sectional analysis of BioBank Japan clinical data: A large cohort of 200,000 patients with 47 common diseases. J Epidemiol 2017;27:S9-S21. [Crossref] [PubMed]
28. Burgess S, Scott RA, Timpson NJ, et al. Using published data in Mendelian randomization: a blueprint for efficient identification of causal risk factors. Eur J Epidemiol 2015;30:543-52. [Crossref] [PubMed]
29. Bowden J, Davey Smith G, Haycock PC, et al. Consistent Estimation in Mendelian Randomization with Some Invalid Instruments Using a Weighted Median Estimator. Genet Epidemiol 2016;40:304-14. [Crossref] [PubMed]
30. Greco M FD, Minelli C, Sheehan NA, et al. Detecting pleiotropy in Mendelian randomisation studies with summary data and a continuous outcome. Stat Med 2015;34:2926-40. [Crossref] [PubMed]
31. Burgess S, Thompson SG. Interpreting findings from Mendelian randomization using the MR-Egger method. Eur J Epidemiol 2017;32:377-89. [Crossref] [PubMed]
32. Davies NM, Holmes MV, Davey Smith G. Reading Mendelian randomisation studies: a guide, glossary, and checklist for clinicians. BMJ 2018;362:k601. [Crossref] [PubMed]
33. Dunaif A. Polycystic ovary syndrome in 2011: Genes, aging and sleep apnea in polycystic ovary syndrome. Nat Rev Endocrinol 2011;8:72-4. [Crossref] [PubMed]
34. Dumesic DA, Oberfield SE, Stener-Victorin E, et al. Scientific Statement on the Diagnostic Criteria, Epidemiology, Pathophysiology, and Molecular Genetics of Polycystic Ovary Syndrome. Endocr Rev 2015;36:487-525. [Crossref] [PubMed]
35. Kahal H, Kyrou I, Tahrani AA, et al. Obstructive sleep apnoea and polycystic ovary syndrome: A comprehensive review of clinical interactions and underlying pathophysiology. Clin Endocrinol (Oxf) 2017;87:313-9. [Crossref] [PubMed]
36. Fogel RB, Malhotra A, White DP. Sleep. 2: pathophysiology of obstructive sleep apnoea/hypopnoea syndrome. Thorax 2004;59:159-63. [Crossref] [PubMed]
37. Kumarendran B, Sumilo D, O'Reilly MW, et al. Increased risk of obstructive sleep apnoea in women with polycystic ovary syndrome: a population-based cohort study. Eur J Endocrinol 2019;180:265-72. [Crossref] [PubMed]
38. Chatterjee B, Suri J, Suri JC, et al. Impact of sleep-disordered breathing on metabolic dysfunctions in patients with polycystic ovary syndrome. Sleep Med 2014;15:1547-53. [Crossref] [PubMed]
39. Dexter DD, Dovre EJ. Obstructive sleep apnea due to endogenous testosterone production in a woman. Mayo Clin Proc 1998;73:246-8. [Crossref] [PubMed]
40. Dunaif A, Segal KR, Futterweit W, et al. Profound peripheral insulin resistance, independent of obesity, in polycystic ovary syndrome. Diabetes 1989;38:1165-74. [Crossref] [PubMed]
41. Pamidi S, Wroblewski K, Broussard J, et al. Obstructive sleep apnea in young lean men: impact on insulin sensitivity and secretion. Diabetes Care 2012;35:2384-9. [Crossref] [PubMed]
42. Diamanti-Kandarakis E, Dunaif A. Insulin resistance and the polycystic ovary syndrome revisited: an update on mechanisms and implications. Endocr Rev 2012;33:981-1030. [Crossref] [PubMed]
43. Franks S, Mason H, Willis D. Follicular dynamics in the polycystic ovary syndrome. Mol Cell Endocrinol 2000;163:49-52. [Crossref] [PubMed]
44. Lo JC, Feigenbaum SL, Yang J, et al. Epidemiology and adverse cardiovascular risk profile of diagnosed polycystic ovary syndrome. J Clin Endocrinol Metab 2006;91:1357-63. [Crossref] [PubMed]
45. Chen X, Yang D, Mo Y, et al. Prevalence of polycystic ovary syndrome in unselected women from southern China. Eur J Obstet Gynecol Reprod Biol 2008;139:59-64. [Crossref] [PubMed]
46. Ng TP, Seow A, Tan WC. Prevalence of snoring and sleep breathing-related disorders in Chinese, Malay and Indian adults in Singapore. Eur Respir J 1998;12:198-203. [Crossref] [PubMed]
47. Redline S, Leitner J, Arnold J, et al. Ventilatory-control abnormalities in familial sleep apnea. Am J Respir Crit Care Med 1997;156:155-60. [Crossref] [PubMed]
48. Redline S, Tishler PV, Schluchter M, et al. Risk factors for sleep-disordered breathing in children. Associations with obesity, race, and respiratory problems. Am J Respir Crit Care Med 1999;159:1527-32. [Crossref] [PubMed]
49. Ancoli-Israel S, Klauber MR, Stepnowsky C, et al. Sleep-disordered breathing in African-American elderly. Am J Respir Crit Care Med 1995;152:1946-9. [Crossref] [PubMed]
50. Li KK, Kushida C, Powell NB, et al. Obstructive sleep apnea syndrome: a comparison between Far-East Asian and white men. Laryngoscope 2000;110:1689-93. [Crossref] [PubMed]
51. Fry A, Littlejohns TJ, Sudlow C, et al. Comparison of sociodemographic and health-related characteristics of UK Biobank Participants With Those of the General Population. Am J Epidemiol 2017;186:1026-34. [Crossref] [PubMed]