Socio-clinical Phenotyping in Patients with Atherosclerotic Cardiovascular Disease: A Latent Class Analysis

In a common disease population such as atherosclerotic cardiovascular disease (ASCVD), latent classes may uncover subgroups of patients that can be distinguished by combinations of several factors instead of a single factor. In this study, we sought to identify clinical, demographic, and social sub-phenotypes of ASCVD, using latent class analysis (LCA), and assess the risk of all-cause and cardiovascular mortality across the identified socio-clinical classes. LCA is a statistical technique employed to uncover hidden class divisions within a set of individuals by utilizing a mix of categorical and/or continuous observed variables. Using the National Health Interview Survey (NHIS) between 2013 and 2018, a nationwide self-reported survey, linked to the National Mortality Index (NDI), we included participants 18 years and older who reported a history of ASCVD in the US. The main outcome of this study is all-cause and cardiovascular mortality. There were 17,807 patients with a mean (standard deviation, SD) age of 66.9 (13.5). In summary, the three classes derived from LCA can be described as follows: Class 1 is characterized by non-Hispanic White individuals with a low comorbidity burden, Class 2 consists of older individuals with a higher proportion of men, and Class 3 includes younger individuals, predominantly non-Hispanic Black and Hispanic, with a greater burden of comorbidities. In multivariable models, the adjusted Hazard ratio (aHR) with 95% confidence intervals (95% CIs) were 1.678 (1.458-1.930) in class 2 and 2.255 (1.931-2.633) in class 3 (p<0.001) for the all-cause long-term mortality. ASCVD sub-phenotype (latent class) of younger, female, non-Hispanic Black or Hispanic individuals with a high burden of comorbidities and unfavorable SDoH was associated with the highest risk of mortality compared with other classes. Our approach may inform future work to understand the heterogeneity among demographic, clinical and social risk factors in the ASCVD population, and classify mortality risk based on these key population characteristics.