Life course, genetic, and neuropathological associations with brain age in the 1946 British Birth Cohort: a population-based study

Aaron Z Wagen,William Coath,Ashvini Keshavan,Sarah-Naomi James,Thomas D Parker,Christopher A Lane,Sarah M Buchanan,Sarah E Keuss,Mathew Storey,Kirsty Lu,Amy Macdougall,Heidi Murray-Smith,Tamar Freiberger,David M Cash,Ian B Malone,Josephine Barnes,Carole H Sudre,Andrew Wong,Ivanna M Pavisic,Rebecca Street,Sebastian J Crutch,Valentina Escott-Price,Ganna Leonenko,Henrik Zetterberg,Henrietta Wellington,Amanda Heslegrave,Frederik Barkhof,Marcus Richards,Nick C Fox,James H Cole,Jonathan M Schott

DOI: https://doi.org/10.1016/S2666-7568(22)00167-2

Abstract:Background: A neuroimaging-based biomarker termed the brain age is thought to reflect variability in the brain's ageing process and predict longevity. Using Insight 46, a unique narrow-age birth cohort, we aimed to examine potential drivers and correlates of brain age. Methods: Participants, born in a single week in 1946 in mainland Britain, have had 24 prospective waves of data collection to date, including MRI and amyloid PET imaging at approximately 70 years old. Using MRI data from a previously defined selection of this cohort, we derived brain-predicted age from an established machine-learning model (trained on 2001 healthy adults aged 18-90 years); subtracting this from chronological age (at time of assessment) gave the brain-predicted age difference (brain-PAD). We tested associations with data from early life, midlife, and late life, as well as rates of MRI-derived brain atrophy. Findings: Between May 28, 2015, and Jan 10, 2018, 502 individuals were assessed as part of Insight 46. We included 456 participants (225 female), with a mean chronological age of 70·7 years (SD 0·7; range 69·2 to 71·9). The mean brain-predicted age was 67·9 years (8·2, 46·3 to 94·3). Female sex was associated with a 5·4-year (95% CI 4·1 to 6·8) younger brain-PAD than male sex. An increase in brain-PAD was associated with increased cardiovascular risk at age 36 years (β=2·3 [95% CI 1·5 to 3·0]) and 69 years (β=2·6 [1·9 to 3·3]); increased cerebrovascular disease burden (1·9 [1·3 to 2·6]); lower cognitive performance (-1·3 [-2·4 to -0·2]); and increased serum neurofilament light concentration (1·2 [0·6 to 1·9]). Higher brain-PAD was associated with future hippocampal atrophy over the subsequent 2 years (0·003 mL/year [0·000 to 0·006] per 5-year increment in brain-PAD). Early-life factors did not relate to brain-PAD. Combining 12 metrics in a hierarchical partitioning model explained 33% of the variance in brain-PAD. Interpretation: Brain-PAD was associated with cardiovascular risk, and imaging and biochemical markers of neurodegeneration. These findings support brain-PAD as an integrative summary metric of brain health, reflecting multiple contributions to pathological brain ageing, and which might have prognostic utility. Funding: Alzheimer's Research UK, Medical Research Council Dementia Platforms UK, Selfridges Group Foundation, Wolfson Foundation, Wellcome Trust, Brain Research UK, Alzheimer's Association.