Cognitive decline and risk of stroke in older women: a cohort study

Impaired cognitive function has been linked to an elevated risk of stroke. However, the effect of cognitive change, conceptualized as a dynamic process, has received limited attention. We therefore aimed to estimate the effect of cognitive decline on the risk of stroke in older women.

Our analyses were based on data from the cognitive cohort of the Women's Health Study composed of 6,377 US female health professionals aged 65 years and older at baseline. We included women without a history of stroke who had complete information on global cognitive performance change. Women were categorized into quintiles based on their 4-year global cognitive performance change and followed for self-reported incident stroke confirmed by a physician Endpoints Committee through December 31, 2022. We used multivariable-adjusted Cox proportional hazards regression models to estimate hazard ratios (HRs) and corresponding 95% confidence intervals (CIs) of the association between cognitive change and stroke. Secondary analyses assessed the risk for stroke subtypes, lowest 10% and 20% of the cognitive change distribution, and changes in verbal memory.

5,093 women with a mean age of 66 years (SD = 3.93) were analyzed. The mean follow-up time was 11.6 years per person during which 302 incident stroke cases were reported. Compared to the 5th quintile (greatest improvement), the adjusted HR for stroke was 0.76 (95%CI [0.53-1.10]) for the 4th, 0.91 (95%CI [0.64-1.29]) for the 3rd, 0.95 (95%CI [0.67-1.34]) for the 2nd, and 1.02 (95%CI [0.72-1.45]) for the 1st quintile characterized by the greatest decline. Similar patterns were observed in secondary analyses. Women in the lowest 10% of global cognitive change exhibited a non-significant increase in the risk of stroke compared to the rest of the distribution (HR_adj = 1.25 (95%CI [0.86-1.82]).

Large declines in cognitive performance were not strongly associated with stroke risk in this cohort of older female health professionals. However, future studies should further explore whether cognitive trajectories may influence risk of stroke and assess the generalizability of our findings, particularly to men.