Key points are not available for this paper at this time.
Background: Hispanics are a highly admixed population. Pooled Cohort Equation (PCE) excluded them in its derivation. We analyzed Hispanic background and genetic ancestry in relation to atherosclerotic cardiovascular disease (ASCVD) risk. Methods: We included 966 Hispanics from Multi-Ethnic Study of Atherosclerosis (MESA), aged 44-75 years and free of diabetes/prior CVD. Statin users and those with LDL-C>190 mg/dL were excluded. Continental genetic ancestry (West African AA, European EA) was estimated by ADMIXTURE software. Predicted ASCVD events (myocardial infarction, stroke, coronary heart disease death) based on non-Hispanic White (NHW) and non-Hispanic Black (NHB) PCE versions, were compared with observed events based on 10-years Kaplan-Meier Estimates and presented as predicted-to-observed (P/O) ratios across Hispanic backgrounds and >median genetic ancestry. Risk discrimination was assessed by Harrell’s C-statistic based on Cox models. Results: MESA Hispanics included 504 Mexicans ME, 284 Caribbean CA and 178 Other Hispanics OH. At 10-years, there were 179 ASCVD events: 107 ME; 49 CA; 23 OH. PCE overestimates ASCVD risk (reflected by P/O ratio >1.0). Using the NHW PCE, all Hispanics had a similar P/O ratio compared to NHW in the low-risk category, and lower P/O ratio compared to NHW in the high-risk category. The extent of overestimation varied across disaggregated Hispanic background (least in ME intermediate risk; most in CA high risk category). Higher P/O ratios were observed among Hispanics of >median AA ancestry and lowest in Hispanics of >median EA ancestry (Figure 1). Harrell’s C-statistic showed no statistically significant difference within groups. Conclusion: Disaggregating Hispanics by background and ancestry led to variability in PCE performance with greater overestimation of ASCVD risk for those of CA background or with >median AA compared to those of ME background or >median EA. The relationship between ethnicity, ancestry, and ASCVD risk in Hispanics should be further explored.
Nriagu et al. (Tue,) studied this question.