Living in the United States Stroke Belt and cognitive change in the Women's Health Study: a cohort study

Residence in the United States Stroke Belt has been linked to a higher stroke mortality rate, raising the question of whether this increased risk is also associated with other brain health outcomes, such as cognitive change, defined as within-person changes in cognitive performance over time. This study aimed to investigate if residing in the Stroke Belt is associated with cognitive decline compared to residing outside this region in a large cohort of women.

We used data from the cognitive cohort of the Women's Health Study, a U.S.-based cohort of predominantly White women currently or formerly employed in health professions. Our analyses included 6,362 women aged ≥65 years at baseline who underwent cognitive assessments. The exposure was residency in the Stroke Belt. Cognitive change was assessed as the primary outcome using repeated measures of a global cognitive composite score. Linear mixed-effects models were applied, adjusted for age, education, and lifestyle factors, measured by Life's Simple 7 score with random effects for each of the participating women (fully adjusted model). Secondary analyses used logistic regression to examine the association between Stroke Belt residency and the odds of belonging to the first quintile or first decile of the cognitive change distribution (secondary outcome), representing women with the greatest cognitive decline.

Of the participating women, 1,131 (17.8%) resided within the Stroke Belt. Compared to women residing outside the Stroke Belt, women residing in the Stroke Belt did not have significantly different rates of cognitive decline. The fully adjusted difference in change in the global cognitive composite score from the first to the last assessment between women living inside and outside the Stroke Belt was β = -0.03, SE = 0.02, p = 0.233. Women residing in the Stroke Belt had higher odds of being in the first quintile (adjusted odds ratio (OR) 1.20, 95% confidence interval (CI) 1.00-1.42) and in the first decile (adjusted OR 1.43, 95% CI 1.14-1.78) of the cognitive change distribution.

Residence in the U.S. Stroke Belt was not associated with overall differences in cognitive decline trajectories, although secondary analyses suggested a possible increase in more pronounced decline among women living in the region.

ClinicalTrials.gov Identifier: NCT00000479; retrospectively submitted: October 27, 1999.