Impact of Stroke on Cognition in Old Age: Comparison of Two Population-Based Cohorts, Born Up to 30 Years Apart and Followed from Age 70 to 85

Abstract

Little is known about birth cohort differences in the impact of stroke on cognitive aging. Given improved poststroke rehabilitation and better treatments for vascular health risk, we may expect a reduction in the stroke impact in later-born cohorts. We tested this prediction using data from two cohorts, born in 1901–1907 (n = 1,155) and 1930 (n = 919), identified from the same city population at the same age of 70 and subsequently measured on the same cognitive outcomes (i.e., spatial ability, perceptual–motor speed, and reasoning) at ages 70, 75, 79, and 85. We fitted multiple-group second-order latent growth curve models to the data, regressing the first-order cognitive factor on the time-varying stroke variable and controlling for relevant covariates. Findings revealed moderate to large average cognitive decline (d = −.45) following stroke, and the impact was relatively similar across cohorts (1901–1907: d = −.52; 1930: d = −.39). However, there was a stroke by age by cohort interaction, implying that the stroke impact increased with age in the 1901–1907 cohort (dage ≤ 75 = −.42; dage ≥ 79 = −.70) but decreased in the 1930 cohort (dage ≤ 75 = −.53; dage ≥ 79 = −.17). We found no evidence for lagged effect of stroke beyond the impact on measures following the incidence. Our hypothesis was only partially supported, as the impact of stroke was reduced in the later-born cohort but solely at higher ages.