Hormone replacement therapy and dementia risk among postmenopausal women: Identifying responsive subgroups in the UK Biobank.

INTRODUCTION: Modifiers of the association between hormone replacement therapy (HRT) use and dementia risk and responsive subgroups, are unknown. METHODS: In 183,450 postmenopausal women from the UK Biobank (mean 13.3 years follow-up), multivariable Cox regression models evaluated associations between HRT use (≥1 year) and all-cause dementia and dementia subtypes. Stratified analyzes assessed modification by menopause type, lifetime estrogen exposure, apolipoprotein E (APOE) ɛ4 genotype, and age at HRT initiation. RESULTS: Over 2.43 million person-years, 3,948 dementia cases were identified. HRT use was associated with lower all-cause dementia (hazard ratio [HR] 0.90; 95% CI, 0.84-0.96), with stronger associations in surgical menopause (0.74, 0.65-0.84), APOE ɛ4 carriers (0.87, 0.80-0.95) and women with lower endogenous estrogen exposure (0.84, 0.77-0.93). Initiation between ages 46 and 56 conferred the greatest benefit. DISCUSSION: HRT use was associated with lower dementia risk, and these findings may help inform more tailored approaches to HRT prescribing in relation to women's lifetime dementia risk.