To examine whether a joint cardiometabolic health–frailty phenotype is associated with incident atrial fibrillation (AF), incident heart failure (HF) and progression to AF–HF comorbidity, and to identify candidate circulating protein signals related to this phenotype.

Among 319,048 UK Biobank participants without AF or HF, cardiometabolic status (healthy [CMH] or unhealthy [CMU]) and Frailty status (non-frail [NF] or pre-frail/frail [F]) defined four phenotypes. Cox and clock-forward multistate models estimated disease-specific and transition-specific hazards. Proteomic feature selection and exploratory mediation analyses included 28,206 participants with Olink plasma proteomic data.

During follow-up, 17,844 participants developed AF, 5,410 developed HF, and 2,403 and 791 progressed from AF and HF, respectively, to comorbidity. Compared with CMH-NF, CMU-F was associated with the greatest risks of incident HF (hazard ratio (HR), 2.76; 95% confidence interval (CI), 2.59–2.95) and AF (HR, 1.49; 95% CI, 1.43–1.55). Corresponding multistate HRs were 2.98 (95% CI, 2.76–3.22) for baseline-to-HF, 1.45 (1.39–1.51) for baseline-to-AF, 1.76 (1.57–1.98) for AF-to-comorbidity, and 1.27 (1.02–1.57) for HF-to-comorbidity. Positive additive interactions indicated excess risk from coexisting cardiometabolic unhealthiness and frailty. Thirteen proteins showed statistical mediation signals for both outcomes; FABP4 yielded the largest positive mediated proportions (AF, 49.8%; HF, 32.9%) and transition-specific signals for baseline-to-AF and AF-to-comorbidity.

Coexisting cardiometabolic unhealthiness and frailty identified individuals at greatest risk of AF, HF, and subsequent comorbidity. Exploratory proteomic analyses identified FABP4 as the most consistent preliminary signal, which warrants further validation in independent cohorts.