Vascular risk factors and cognitive function in midlife women: The study of women’s health across the nation (SWAN)

Abstract

Modifiable cardiovascular risk factors are promising targets for dementia prevention. The preclinical course of dementia spans decades, and the effects of risk factors are likely cumulative over the life-span. Thus, better understanding the relation of cardiovascular risk factors to cognition in midlife, prior to development of impairment is critical. We examined longitudinal associations of modifiable risk factors with cognition in the SWAN cohort of middle aged women. SWAN is a community based multi-site, multi-ethnic longitudinal study of the menopause transition. Pre- or early peri-menopausal women aged 42-52, were recruited at 7 sites between 1996-1997. Analyses include 1701 women who completed cognitive evaluations at the 4th annual study visit (Mean age 46 years, 78% had > high school education). Cognitive assessments included processing speed (Symbol Digit Modalities Test (SDMT)), Verbal episodic memory immediate and delayed recall (East Boston Memory Test (EBMT)), and working memory (Digit Span Backward (DSB)). Tests were repeated at the 6th-10thSWAN visit s(median 4 assessments). Glucose, insulin, and lipids were measured using standardized assays, and waist, BMI, blood pressure, and medical history were assessed using standardized protocols. Risk factors were dichotomized based on established criteria. Linear mixed effects models with random intercept were used to examine the relation of baseline cardiovascular risk factors to cognitive change, adjusting for age, education, race/ethnicity, depressive symptoms, alcohol use, smoking, and menopause status. Models included a spline at the third assessment, to reflect that cognitive performance showed improvement consistent with learning effects over the initial 3 assessments, and showed no significant change thereafter. Women with hypertension, high triglycerides or glucose, diabetes or high HOMA-insulin resistance index showed a significantly reduced rate of improvement in SDMT over the initial 3 cognitive assessments. Elevated glucose and diabetes were associated with less improvement in DSB. Improvement in EBMT delayed recall was reduced among those with obesity, or high HOMA-IR (All p< 0.05). In this healthy middle-aged cohort without evidence of cognitive decline, factors related to pre-diabetes were associated with diminished learning effects over repeated cognitive assessments, particularly tests of executive function. Further research is needed to determine if early control of cardiovascular risk factors can curtail cognitive decline.