Glucose and lipid metabolism in non-diabetic, non-obese patients with obstructive sleep apnea: sex differences

Obstructive sleep apnea (OSA) is associated with glucose and lipid disturbances and insulin resistance. However, glucose and lipid disturbances and insulin resistance in OSA are often attributed to confounding obesity and/or diabetes. Studies on nondiabetic, nonobese OSA patients are very limited.

This cross-sectional study retrospectively analyzed non-diabetic, non-obese adults who underwent a home sleep apnea testing and collected fasting blood samples before or after the sleep study to measure glucose and lipids. This study was designed as a cross-sectional study and therefore can only demonstrate associations between variables, but not causality.

Among the 191 participants (mean age 48.94 years, 68.06% male) included in the study, 83.77% had OSA. The high-density lipoprotein cholesterol (HDL-C) level in OSA participants was significantly lower (0.99 vs. 1.12 mmol/L, p = 0.036), and the triglyceride-glucose (TyG) index was significantly higher (8.74 vs. 8.45, p = 0.016), while there was no significant difference in the levels of total cholesterol, low-density lipoprotein cholesterol (LDL-C), and non-HDL-C. Correlation analysis by sex showed that AHI was significantly positively correlated with fasting plasma glucose (r = 0.373), non-HDL-C (r = 0.280), and TyG index (r = 0.337) in female participants, while AHI was only significantly negatively correlated with HDL-C (r = -0.194) in male participants. Multivariable analysis revealed that compared with non-OSA individuals, OSA severity in women was independently associated with fasting plasma glucose (AHI ≥ 5: β = 0.55, 95 % CI 0.13 to 0.98; AHI ≥ 15: β = 0.60, 95% CI 0.13 to 1.07) and TyG index (AHI ≥ 5: β = 0.37, 95% CI 0.08 to 0.66; AHI ≥ 15: β = 0.39, 95% CI 0.07 to 0.71; AHI ≥ 30: β = 0.53, 95% CI 0.08 to 0.98). In contrast, among men, OSA severity showed independent associations with triglycerides (15 ≤ AHI < 30: β = 1.00, 95% CI 0.05 to 1.95) and HDL-C (AHI ≥ 15: β = -0.17, 95% CI -0.33 to -0.01; AHI ≥ 30: β = -0.22, 95% CI -0.38 to -0.06).

Our study supports the claim that there are sex differences in glucose and lipid metabolic disorders in non-diabetic, non-obese OSA participants: women mainly showed elevated fasting plasma glucose and TyG index, while men showed dyslipidemia with elevated triglycerides and decreased HDL-C. These findings highlight the need to consider sex differences when assessing OSA-related metabolic risks.