The association between obesity indices and obstructive sleep apnea is modified by age in a sex-specific manner

Background The beneficial effects of weight reduction on obstructive sleep apnea (OSA) are highly variable. Whether or not the variability is associated with the effects of age and sex remains unclear. This study examined this issue with large cross-sectional data. Method Anthropometric measurements, polysomnographic variables, biochemical indicators, and medical history were collected for each participant. Multivariable linear regression with interaction terms was used to estimate the modification effect of age on the associations between OSA severity (assessed by apnea-hypopnea index, AHI) with obesity indices (body mass index, BMI; neck circumference, NC; waist circumference, WC; waist-to-hip ratio, WHR) in a sex-specific manner, and vice versa. To facilitate interpretation of the results, participants were further classified into six subpopulations according to both sex and age, and population-specific beta-coefficients were calculated and compared. Results A total of 5756 adults (4600 men) with suspected OSA were included in the study. BMI, NC, WC, and WHR were all positively correlated with AHI after adjusting for potential confounders in all populations. In men, these associations were much stronger and more significant in younger than older individuals (P for interaction < 0.001). For example, a 10% increase in BMI was independently associated with a 32% increase in AHI for men < 40 years old, whereas the corresponding increases were 21% and 17% for men 40-60 and > 60 years old, respectively. By contrast, no modification effect of age was observed in women (P for interaction > 0.05). A 10% increase in BMI was associated with 26%, 27%, and 24% increases in AHI for women < 40, 40-60, and > 60 years old, respectively. Conclusions Age modifies the associations between obesity indices and OSA severity in a sex-specific manner. These findings may broaden the understanding of age- and sex-related heterogeneities in the pathogenic role of obesity in OSA, and may be beneficial for individualized risk evaluation and treatment management for patients with OSA.