Impact of obstructive sleep apnea on insulin resistance and glucose tolerance in women with polycystic ovary syndrome

Abstract

Context: Insulin resistance, impaired glucose tolerance, and type 2 diabetes are common in women with polycystic ovary syndrome (PCOS). Obstructive sleep apnea (OSA) has been linked to metabolic dysfunction. We studied women with and without PCOS to determine the extent to which OSA is responsible for insulin resistance and glucose intolerance in PCOS.

Methods: In a prospective design, 52 women with PCOS and 21 women without PCOS of similar age and body mass index had an overnight polysomnogram and a 75-g oral glucose tolerance test.

Results: Twenty-nine women (56%) with PCOS had OSA compared with four controls (19%) (adjusted odds ratio 7.1; 95% confidence interval, 1.7-45.7; P = 0.01). PCOS women with OSA were more insulin resistant than those without OSA [homeostasis model assessment (HOMA) index 5.7 +/- 0.4 vs. 3.5 +/- 0.4; P = 0.006] after controlling for age, body mass index, and ethnicity. Impaired glucose tolerance was found in 16 of 29 (55%) PCOS women with OSA and only six of 23 (26%) of those without OSA (unadjusted P = 0.049). Insulin resistance and glucose intolerance were highly correlated with the presence and severity of OSA. Among PCOS women with normal glucose tolerance, the presence of OSA was associated with a nearly 2-fold higher fasting insulin level and HOMA index. The severity of OSA was a highly significant predictor of the fasting concentrations of glucose and insulin as well as the 2-h glucose concentration and HOMA index.

Conclusions: OSA is a highly prevalent and important determinant of insulin resistance, glucose intolerance, and type 2 diabetes in PCOS.

Trial registration: ClinicalTrials.gov NCT00203996.

Figures

Figure 1

Metabolic and hormonal variables in study subjects. A, All PCOS women (n = 52) vs. controls (n = 21); B, PCOS women without OSA (n = 23) vs. controls without OSA (n = 17); C, normal glucose tolerant (NGT) PCOS women without OSA (n = 17) vs. normal glucose tolerant controls without OSA (n = 16). *, P < 0.05; **, P < 0.01.

Figure 2

Mean (± sem) concentrations of glucose (top) and insulin (bottom) in response to a 75-g oral glucose challenge. All subjects have normal glucose tolerance. Solid heavy lines depict PCOS women with OSA, solid thin lines depict PCOS women without OSA, and dashed lines depict control women without OSA. Plasma glucose concentrations at time 0, 60, 90, and 120 min after glucose ingestion did not differ between PCOS women and controls without OSA; at 30 min, the difference approached (P = 0.06) but did not reach significance. Insulin values during the OGTT are virtually identical in PCOS and control women without OSA.

Figure 3

Prevalence of impaired glucose tolerance (IGT) among control women without OSA, PCOS women without OSA, and PCOS women with mild, moderate, and severe OSA. The prevalence of impaired glucose tolerance increased in direct proportion to the severity of OSA.

Figure 4

Relationships between the severity of OSA (as assessed by the AHI) in women with PCOS and the fasting levels of glucose and insulin, HOMA index, 2-h glucose concentration after glucose challenge, and free testosterone concentrations. After controlling for age, BMI, and ethnicity, the severity of OSA (as assessed by the AHI value) was a highly significant predictor of the fasting concentrations of glucose and insulin as well as the 2-h glucose concentration and HOMA index. In contrast, no significant association between AHI and free testosterone levels was detected.