Early life blood pressure and cognitive function in mid/late life: A synthetic longitudinal cohort analysis

ABSTRACT BACKGROUND Hypertension is a well-studied dementia risk factor. The effects of hypertension on cognitive aging may vary by life stage, yet most prior studies have focused on mid- or late-life blood pressure. This gap reflects the lack of longitudinal cohorts spanning both childhood blood pressure measurements and later-life cognitive assessments. To address this gap, we examined the association between early-life blood pressure and mid/late-life cognition using a synthetic cohort approach. METHODS Using data fusion as a guiding framework, we pooled the Bogalusa Heart Study (BHS), which provides early-life blood pressure data (ages 4-16), with the CARDIA study, which provides mid/late-life cognitive data (ages 58-70), to create a synthetic longitudinal cohort. The cohorts overlapped between ages 17-57. Cognition comprised raw and z-scores measures of executive function, memory, and processing speed. Early life systolic blood pressure (SBP) was defined as a time-weighted average of SBP measurements between ages 4 and 16. We pooled data by exact- and distance-matching 10 BHS participants to each CARDIA participant on hypothesized mediators & confounders of the early life SBP and cognition relationship that overlapped in age of measurement between the two cohorts, including sociodemographics and ∼midlife measures of blood pressure, cognition, and vascular risk factors. Confounder-adjusted random intercept models were used to estimate the association between early-life SBP and cognitive outcomes. RESULTS There was no significant association between early life SBP and mid/late-life cognitive raw scores: executive function [β (95% CI) = 0.285 (−0.867, 1.437)], memory [−0.019 (−0.308, 0.269)], or processing speed [−0.188 (−1.611, 1.236]. Likewise, no associations were observed when using Z-scored cognitive outcomes. CONCLUSIONS We observed no evidence that early life SBP was associated with mid/late-life cognition, although interpretation of these findings depends on strong assumptions about cohort comparability and linkage validity that underpin the synthetic cohort approach. While our results lacked precision, we demonstrate a possible signal in the expected direction (i.e., greater systolic blood pressure results in worse cognitive functioning across all domains) that could be the focus of future work.