Multidimensional Cardiovascular-Geriatric Vulnerability Phenotypes in Older Adults With Coronary Heart Disease: An Exploratory Latent Class Analysis

Older adults with coronary heart disease (CHD) frequently exhibit concurrent physical, cognitive, functional, social, and cardiovascular vulnerabilities. We aimed to identify clinically interpretable multidimensional vulnerability phenotypes and characterize their baseline profiles. This single-centre observational study included 178 adults aged ≥ 60 years with chronic CHD. Following exploratory indicator-set development, latent class analysis was performed using nine binary indicators spanning physical, functional, social, cognitive, and cardiovascular domains. Two- to six-class solutions were compared using information criteria, entropy, posterior classification, class size, response profiles, and clinical interpretability; 60-day outcomes were examined separately. AIC and sample-size-adjusted BIC favoured three classes, whereas BIC favoured two. A four-class solution was retained as an exploratory compromise, with entropy of 0.761, all classes comprising ≥ 9.6% of participants, and class-specific average posterior probabilities of 0.808–0.884. The phenotypes were cardiac dysfunction-dominant vulnerability (9.6%), physical-social-cognitive vulnerability (28.1%), lower-burden functionally preserved (52.2%), and multidomain high-burden vulnerability (10.1%). No clear between-class differences were detected in exploratory 60-day outcomes, although event counts were limited. Vulnerability in older adults with CHD was heterogeneous in composition rather than simply ordered along a frailty-severity continuum. Multidimensional phenotyping may complement conventional cardiovascular assessment, although the retained class structure requires external and longitudinal validation.