Comorbidity sequence, sex, and APOE-genotype forecast Alzheimer's disease diagnosis

Alzheimer's disease (AD) is a highly heterogeneous neurodegenerative disorder and the leading cause of dementia characterized by the progressive accumulation of non-modifiable (age, female sex, APOE-ε4 genotype) and modifiable factors [hypertension (HTN), diabetes, obesity (OB), hyperlipidemia (HLP), depression (DEP)]. However, the temporal sequencing and interaction patterns between comorbidity burden and biological subgroups defined by sex and APOE genotype remain not fully understood.

We applied the Cumulative Event Method (CEM), a novel process mining framework, to longitudinal UK Biobank (UKB) data. Event logs tracked five modifiable risk factors across sex- and APOE-ε4-stratified analyses to identify distinct longitudinal comorbidity patterns associated with AD. Sex-specific findings were validated in an independent CureMD cohort.

Among 1,916 UK Biobank participants, CEM identified 203 distinct comorbidity sequences across 7,316 clinical events. Females more frequently exhibited a hypertension-preceding-AD sequences than males (7.0% vs. 3.8%; p = 0.005), while males exhibited earlier metabolic-vascular patterns involving hyperlipidemia and hypertension (7.7% vs. 4.5%; p = 0.0085). APOE-ε4 carriers exhibited accelerated multi-comorbidity patterns, whereas non-carriers more frequently transitioned from hypertension to non-AD (p = 1 × 10^-4). External validation in CureMD confirmed sex-specific patterns across 191 sequences and 5,176 events.

Longitudinal comorbidities patterns preceding AD differ by sex and APOE genotype, supporting Alzheimer's as a multisystem failure disease with subgroup-specific comorbidity sequences and clinically relevant windows for precision prevention.