Associations of Sex, Age, and Cardiometabolic Risk Profiles With Brain Structure and Cognition: A UK Biobank Latent Class Analysis

Research output: Contribution to journalArticleResearchpeer-review

3 Citations (Scopus)


BACKGROUND AND OBJECTIVES: It is unknown whether there are sex-related profiles of cardiometabolic health that contribute differently to age-related changes in brain health during midlife. We studied how latent classes of middle-aged individuals clustering by age, sex, menopause, and cardiometabolic health were associated with brain structure and cognitive performance. METHODS: Health, brain, and abdominal MRI data from the UK Biobank cohort (men and women aged >40 years in the United Kingdom) were used. We applied latent class analysis to identify groups of individuals based on age, sex, menopausal status, and cardiometabolic health. We examined associations of class membership with brain volumes (total brain volume [TBV], gray matter volume [GMV], white matter volume [WMV], hippocampal volume, and white matter hyperintensity volume) and cognitive performance. RESULTS: Data were available for 36,420 individuals (mean age 64.9 years, 48.5% women). Eight latent classes differing in age, sex, and cardiometabolic risk were identified. Class 1 (reference class) included individuals with the lowest probability of older age and cardiometabolic risk, and the healthiest levels of brain volumes and cognition. In those aged >60 years, but not in those aged 50-60 years, the negative associations of age with TBV, GMV, and WMV were greater in the class comprising healthier older women than classes comprising older men of varying cardiometabolic and vascular health. There were no age-class interactions for cognitive test performance. DISCUSSION: Latent class analysis detected groups of middle-aged individuals clustering by cardiometabolic health. The relationship of age with brain volumes varies by sex, menopausal status, and cardiometabolic health profile.

Original languageEnglish
Pages (from-to)e1853-e1865
Number of pages14
Issue number17
Publication statusPublished - 25 Oct 2022

Cite this