Abstract
INTRODUCTION: Age-related elevation of gonadotropins may contribute to cognitive decline, while apolipoprotein E epsilon 4 (APOE ε4) is an established risk factor for Alzheimer's disease (AD). This study investigated the associations of follicle-stimulating hormone (FSH) and luteinizing hormone (LH) levels with global cognition and dementia in adults aged over 80. METHODS: A total of 509 adults (440 males and 69 females) were included in this cross-sectional analysis, comprising 337 with normal cognition (NC), 97 with Alzheimer's disease-related dementia (AD-D), and 75 with vascular dementia (VD). Cognitive status was assessed using Mini-Mental State Examination (MMSE). Plasma gonadotropins and sex hormones were measured by chemiluminescence. Multivariable linear and logistic regression models, adjusted for potential confounders, were employed. RESULTS: Follicle-stimulating hormone concentrations were significantly elevated in males with dementia and females with VD compared to NC. LH concentrations were significantly elevated in VD across sexes compared with NC. Neither estradiol nor total testosterone differed across groups. Continuous LH, rather than FSH, was significantly associated with MMSE scores in the total cohort and males after adjusting covariates (both p < 0.05). When dichotomized by median (19.9 IU/L for males; 67.1 IU/L for females), FSH was significantly associated with MMSE after further adjusting for LH (both p < 0.05). A significant interaction between high FSH and APOE ε4 carrier status on cognitive impairment was observed (p < 0.05). After multivariate adjustment including LH, elevated FSH was independently associated with higher AD-D risk both when defined by median split (total sample: OR = 3.109; males: OR = 3.597) and as a continuous variable (total: OR = 1.033; males: OR = 1.048). In contrast, higher continuous LH was linked to lower AD-D risk in the total cohort and males, regardless of FSH adjustment. Neither FSH nor LH concentrations were associated with VD risk after adjusting for covariates. The area under the receiver operating characteristic curve for FSH in predicting AD-D in males was 0.600, with an optimal cutoff value of 28.4 IU/L. CONCLUSION: Elevated FSH and reduced LH may be associated with poorer cognition and an increased risk of AD-D in very old Chinese adults, particularly in males.