Sex differences in dementia risk: considering underlying medical conditions

Background: Underlying medical conditions may explain inconsistent reports of the association between sex and dementia risk. The current study aimed to explore the association between sex and the risk of incident dementia, considering a broad range of medical conditions. Methods: This prospective national birth cohort study consisted of 53,224 members of a nonprofit health maintenance organization. Participants were born between 1922 and 1946 and entered the cohort on January 1, 2002, aged 55 to 80, without a dementia diagnosis. The cohort was followed up for 18 years to January 1, 2020. Dementia was ascertained based on medical diagnoses. Cox regression models were fitted to quantify the association between sex and the risk of incident dementia with hazard ratios (HR) and their 95% confidence intervals (CI), unadjusted and, in the primary model, adjusted for background demographic factors and 33 medical conditions, classified as ten medical domains. Complementary analyses examined adjustment for each medical domain, and sensitivity analyses provided sex-specific estimates for each demographic or medical domain. Results: The analytic cohort of 53,224 participants had a mean (SD) age at cohort entry of 64.3 (7.08) years (Males: N = 24,489; 46.01%; M=63.47 (6.39); Females: N = 28,735; 53.99%; M=64.35 (6.78)). During follow-up, 8,373 participants (15.73%) were diagnosed with dementia (Females: 17.36%; N = 4,987; Males: 13.83%; N = 3,386). Sex differences in the risk of dementia were null after adjustment for demographic factors and medical conditions (unadjusted HR = 1.08 [95% CI = 1.03–1.13, P = 0.001]; adjusted HR = 1.02 [95% CI, 0.97–1.08; p = 0.38]). Complementary analyses showed that when accounting for some conditions (i.e., circulatory, respiratory, metabolic, digestive, or nervous system diseases; cancer; and injuries), females were at an elevated dementia risk compared to males. However, after accounting for rheumatic and genitourinary diseases, the association between sex and dementia was attenuated to null, and when accounting for psychiatric disorders, males were at greater risk. Conclusions: In this prospective birth cohort, the association between sex and the risk of incident dementia changed when background medical conditions were considered, possibly explaining previous inconsistent reports. Future dementia risk and prevention studies may wish to adequately explore sex differences in medical history.