Abstract Body

Background: Hepatitis C (HCV) infection may confer increased risk for cardiovascular disease (CVD). The Framingham Risk Score (FRS) is commonly used to predict 10-year coronary heart disease (CHD) risk, but its performance in HCV-infected persons is not well known. Methodology: We utilized the Electronically Retrieved Cohort of HCV Infected Veterans (ERCHIVES), a national cohort of HCV-infected and age, sex, and race-matched uninfected subjects from 2001-2009. Male subjects ages 20-79 with calculable FRS for Hard CHD (calculated using age, sex, systolic blood pressure, antihypertensive treatment, smoking status, high-density lipoprotein and total cholesterol level) and without baseline CVD, diabetes, HIV and chronic hepatitis B were included. Incident myocardial infarction (MI) rates were calculated and Cox regression used to determine the hazard ratio (HR) for MI by HCV status, adjusting for age and race. C-statistics were used to determine risk discrimination by the FRS, by HCV status and race/ethnicity. Results: Of 316,514 eligible subjects, 232,816 had calculable FRS: 111,436 HCV-infected and 121,380 uninfected. The table summarizes baseline characteristics and FRS. Age, race/ethnicity, BMI, and baseline FRS (in 10-year CHD risk percent) were similar between HCV-infected and uninfected groups. HCV-infected subjects had higher rates of drug and alcohol abuse and lower LDL-cholesterol. Median (IQR) follow-up was 1437 (758-2059) days in the HCV group and 1547 (824-2159) days in the uninfected. Incident MI rate in the HCV and uninfected groups was 7.11 and 7.70 per 1,000 person years, respectively. Adjusted for age and race, the HR for incident MI in HCV vs uninfected persons was 0.92 (95% CI 0.87-0.96, p<0.001). C-statistics were 0.62 (HCV) and 0.61 (uninfected), and by racial/ethnic group, 0.61 (White), 0.61 (Black), and 0.62 (Hispanic). Conclusions: There was no difference in FRS performance by HCV status. The FRS performed similarly poorly across HCV status and racial/ethnic groups in this Veteran Affairs cohort. Because we used incident MI as the CHD outcome measurement and hard CHD includes not only acute MI but also non-MI coronary insufficiency and coronary death, the event rate measured likely underestimates the true CHD event rate and the degree to which FRS underpredicts incident CHD may be even greater. FRS performance in this cohort may be confounded by risk modifying variables and competing risk of non-MI death over the follow-up period.