Associations of domain-specific cognitive function with all-cause mortality over 11 years: the Australian Diabetes, Obesity and Lifestyle (AusDiab) study.
BACKGROUND Poor cognitive function is associated with increased mortality; however, whether these associations are consistent across different cognitive domains remains incompletely understood. METHODS We examined associations between cognitive function and all-cause mortality in a population-based cohort of Australian adults (N = 4561, median age = 60.0 years). In addition to global cognitive function (Mini-Mental State Examination [MMSE]) and premorbid verbal ability, participants underwent clinical assessments targeting the specific cognitive domains of processing speed and memory. Cox proportional hazards models with age as the time scale estimated hazard ratios (HR) for all-cause mortality per 1-standard deviation (SD) increase in cognitive scores. Analyses were adjusted for sociodemographic (including education), lifestyle, and clinical factors. RESULTS In this cohort with predominantly normal MMSE scores (98%), followed for a median of 11.3 years, 566 deaths occurred. Cognitive function was associated with lower mortality across most domains, as evidenced by 22% reduced mortality risk per 1-SD increase in global cognitive function (HR 0.78, 95% CI [0.67, 0.90]), 32% per 1-SD increase in processing speed (HR 0.68, [0.59, 0.78]) and 20% per 1-SD increase in memory (HR 0.80, [0.71, 0.91]), in fully adjusted models. Premorbid verbal ability was not associated with mortality (HR 0.92, [0.83, 1.02]). Associations were similar in men and women, and robust across sensitivity analyses. CONCLUSIONS Processing speed and memory were independently associated with mortality after covariate adjustment. The presence of these associations in a cohort with mostly preserved cognitive function underscores the potential of domain-specific cognitive measures to serve as indirect prognostic markers of mortality and to inform risk stratification.