Comparison of Nonblood-Based and Blood-Based Total CV Risk Scores in Global Populations

THOMAS A. GAZIANO; VILMA IRAZOLA

Global Heart

Elsevier

Año: 2016

2211-8160

Background: Cost-effective primary prevention of cardiovascular disease (CVD) in low- and middle-incomecountries requires accurate risk assessment. Laboratory-based risk tools currently used in high-incomecountries are relatively expensive and impractical in many settings due to lack of facilities.Objectives: This study sought to assess the correlation between a non-laboratory-based risk tool and 4 commonlyused, laboratory-based risk scores in 7 countries representing nearly one-half of the world?s population.Methods: We calculated 10-year CVD risk scores for 47,466 persons with cross-sectional data collected from16 different cohorts in 9 countries. The performance of the non-laboratory-based risk score was comparedwith 4 laboratory-based risk scores: Pooled Cohort Risk Equations (ASCVD [Atherosclerotic CardiovascularDisease]), Framingham, and SCORE (Systematic Coronary Risk Evaluation) for high- and low-risk countries.Rankings of each score were compared using Spearman rank correlations. Based on these correlations, wemeasured concordance between individual absolute CVD risk as measured by the Harvard NHANES (NationalHealth and Nutrition Examination Survey) risk score, and the 4 laboratory-based risk scores, using both theconventional Framingham risk thresholds of >20% and the recent ASCVD guideline threshold of >7.5%.Results: The aggregate Spearman rank correlations between the non-laboratory-based risk score and thelaboratory-based scores ranged from 0.915 to 0.979 for women and from 0.923 to 0.970 for men. Whenapplying the conventional Framingham risk threshold of >20% over 10 years, 92.7% to 96.0% of women and88.3% to 92.8% of men were equivalently characterized as ?high? or ?low? risk. Applying the recent ASCVDguidelines risk threshold of >7.5% resulted in risk characterization agreement for women ranging from 88.1%to 94.4% and from 89.0% to 93.7% for men.Conclusions: The correlation between non-laboratory-based and laboratory-based risk scores is very high forboth men and women. Potentially large numbers of high-risk individuals could be detected with relativelysimple tools.