Comorbidity patterns and sex differences in late-onset Alzheimer's disease using electronic records

In this retrospective study, we applied unsupervised learning techniques to electronic medical records from UCSF to characterize Alzheimers disease comorbidity patterns. Given the well-known female predominance in Alzheimers disease, we performed sex-stratified analyses to explore differences in comorbidity patterns based on sex. Findings were evaluated using an independent UC-Wide dataset. Among 8,363 patients in the UCSF dataset, we observed five data-driven comorbidity clusters related to cardiovascular conditions, gastrointestinal disorders, and frailty-related conditions such as pneumonia and pressure ulcers, as well as groups with higher and lower overall comorbidity burden. Sex-stratified analyses within clusters identified variation in comorbidity patterns, including circulatory diseases in males in Cluster 2 and bladder stones in females in Cluster 3. Key results demonstrate partial consistency in the UC-Wide dataset. Our study provides a structured characterization of comorbidity heterogeneity in AD and highlights sex-related differences, though findings should be interpreted as descriptive and hypothesis-generating. HighlightsO_LIAmong an EHR cluster analysis, Alzheimers disease patients stratify into five data-set derived comorbidity clusters. C_LIO_LIClinical patterns and comorbidity burdens vary across patient subgroups. C_LIO_LISex-specific differences influence comorbidity profiles in selected patient clusters. C_LIO_LIHighly granular cluster configurations lack robust cross-dataset reproducibility. C_LI