Longitudinal Trajectories of Cognitive Decline and Temporal Lobe Atrophy Based on Baseline Gonadotropins and Testosterone

Min Zhao^1,2Jiwei Jiang^2,3Linlin Wang⁴Shiyi Yang⁵Wenyi Li^1,2Qiwei Ren^1,2Huiying Zhang^1,2Tianlin Jiang^1,2Shirui Jiang^1,2Junya Zhou^1,2Jun Xu^1,2*

• ¹Department of Neurology, Beijing Tiantan Hospital Affiliated to Capital Medical University, Beijing, China
• ²China National Clinical Research Center for Neurological Diseases, Beijing, China
• ³Health Management Center, Beijing Tiantan Hospital Affiliated to Capital Medical University, Beijing, China
• ⁴Department of Neurology, Capital Medical University Affiliated Beijing Friendship Hospital, Beijing, China
• ⁵Department of Critical Care Medicine, Sichuan Academy of Medical Sciences and Sichuan People's Hospital, Chengdu, China

Introduction: Although previous studies have reported associations between gonadotropins, testosterone, and Alzheimer's disease (AD), their longitudinal relationships with cognitive decline and temporal lobe atrophy remain insufficiently characterized. This study examined the association between baseline hormone levels and cognitive decline and temporal lobe volume loss trajectories, and whether these associations vary by sex or APOE ε4 status. Methods: This study included 490 participants (378 MCI/112 AD; 311 men/179 women; mean age = 75.01±7.52) from the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort. Baseline plasma levels of gonadotropins (FSH, LH) and total testosterone (TT) were measured using Luminex xMAP multiplex immunoassay. Cognitive decline was assessed longitudinally through MMSE and ADAS-Cog 13 scores. Temporal lobe atrophy was quantified using tensor-based morphometry of 1.5T MRI scans, with bilateral temporal lobe volumes scaled to a normalized reference (1,000=baseline). Linear mixed effects models were employed to relate baseline plasma hormones to longitudinal cognitive performance and temporal lobe volume. Results: Longitudinal analyses showed that higher baseline FSH levels were associated with faster cognitive decline (MMSE: β = −0.025, P = 0.012; ADAS-cog: β = 0.066, P = 0.020) and accelerated temporal lobe atrophy (β = −0.115, P = 0.005) in fully adjusted models. LH was also associated with faster temporal lobe atrophy (β = −0.077, P = 0.034), while TT showed no significant association. Sex-stratified analyses showed that higher TT was associated with slower MMSE decline in women (β = 0.022, P = 0.032) but not in men (P = 0.762), with a significant sex interaction (P-interaction = 0.020). Modification effects of APOE ε4 status on cognition and temporal lobe volume changes were not observed for FSH, LH, or TT. Discussion: The results indicate that in individuals across the AD spectrum, elevated gonadotropin levels may exert deleterious, domain-specific effects on cognitive decline or temporal lobe atrophy. Women with lower TT levels may experience faster cognitive progression. Although future studies incorporating additional longitudinal hormone measurements and cognitive trajectories are warranted, our results underscore the importance of gonadotropins and testosterone in AD progression.