Additional selection bias may arise because women cognitive decline before age 65, statistically more likely to be those with early surgical menopause, were excluded.

Early and surgical menopause are linked to increased AD risk and faster cognitive decline, but mechanisms remain unclear. Shorter reproductive lifespan (RLS: menarche to menopause) is at midlife associated with gray matter loss, but its association with late-life brain volume trajectories is uncertain.

Female participants in the Religious Orders Study/Memory and Aging Project cohorts were analyzed. Associations between reproductive factors, brain volumes (total, gray and white matter volume), and global cognition were evaluated using mixed-effects models adjusted for age, education, smoking, and cohort. Global AD neuropathology was examined using linear models.

In the entire cohort (n=2,398; mean age=78.1±7.8; global cognition=0.12±0.58 at baseline), older menopause age (Est=0.0008, P=0.03), and MHT exposure (Est=0.28, P=0.00008), were both associated with slower cognitive decline. Later surgical menopause was associated with lower AD neuropathology (Est=–0.005, P=0.05) and AD risk (OR=0.97, 95%CI 0.94–0.99, P=0.03). Unexpectedly, in the subset with longitudinal MRIs (n=430), older menopause age was associated with faster total brain volume loss (Est=–0.005, P=0.02). Compared to the larger cohort, at study entry this group were younger (76.8±7.5) with higher cognition (0.37±0.46).

Extended estrogen exposure–via later menopause or timely MHT–appears to confer cognitive and pathological resilience to AD. The paradoxical association observed here between later menopause and faster brain atrophy, highlights possible cohort biases. Indeed, the younger age and cognitive preservation in the MRI subsample at study entry potentially reflected survivor effects; those with early decline (perhaps likeliest to have early surgical menopause) were excluded, introducing selection bias. Altogether, these findings point to the importance of careful interpretation and replication when analyzing the contribution of mid-life contributors, such as reproductive lifespan, to later-life cognitive decline.