Sex- and age-dependent associations between waist-to-hip ratio and cognitive impairment: a cross-sectional study of rural adults aged 40 years and older in northwestern China

Background – Obesity is a modifiable risk factor for cognitive impairment; however, body mass index fails to capture fat distribution. Waist-to-hip ratio (WHR), reflecting central adiposity, may better characterize obesity-related differences in cognitive impairment. Nevertheless, whether the association between WHR and cognitive impairment varies by sex and age remains unclear. Methods – This cross-sectional study included 1, 792 adults aged ≥40 years from rural China. Cognitive impairment was defined using education-adjusted cutoffs on the Chinese Mini-Mental Status. WHR and clinical characteristics were collected by trained investigators. Multivariable logistic regression, restricted cubic splines, stratified analyses, and interaction analyses were performed. Results – In the total population, WHR was positively associated with the odds of cognitive impairment (OR = 1.245 per SD, 95% CI: 1.022–1.517, P = 0.029), with borderline evidence of nonlinearity (P_overall = 0.019, P_nonlinear = 0.090). In sex-stratified analyses, the association in females showed evidence of nonlinearity, with the estimated odds remaining relatively stable below a WHR of 0.88 and increasing at higher levels (P_overall = 0.014, P_nonlinear = 0.049), whereas no significant association was observed in males. Age-stratified analyses demonstrated a significant association in the middle-aged group (40–59 years, P_overall = 0.016, P_nonlinear = 0.143) but not in those ≥60 years. Further interaction analyses showed that age appeared to modify the association between WHR and cognitive impairment in males (OR = 0.462 per SD, 95% CI: 0.258–0.828, P_interaction = 0.010), but not in females. Specifically, the positive association between WHR and cognitive impairment was observed only in middle-aged males (OR = 1.652 per SD, 95% CI: 1.052–2.595, P = 0.029), but not in older males (OR = 0.804 per SD, 95% CI: 0.492–1.314, P = 0.384). Conclusion – Our findings show that WHR is associated with screening-defined cognitive impairment, with patterns that may differ by sex and age. Females exhibited a possible nonlinear association, whereas in males the association was mainly observed in midlife. These findings highlight central adiposity as an important correlate of cognitive impairment. Given the rural single-region sample, larger longitudinal studies with clinically adjudicated outcomes are warranted to assess generalizability and clarify temporal associations.