Abstract 15955: Widely Varying Prevalence of Diastolic Dysfunction by Different Classification Criteria: The Cardia Study

Abstract

Background: Diastolic dysfunction is an important target for investigations attempting to understand heart failure with preserved ejection fraction. The many existing classification criteria for diastolic dysfunction could identify widely different case mixes in large cohort studies with research echocardiographic data. We compared prevalence of diastolic dysfunction at the Y25 exam in CARDIA using definitions of diastolic dysfunction proposed by the American Society of Echocardiography (ASE), Rochester County Epidemiology Project (Redfield), and the VALIDD trial. Methods: 3474 CARDIA participants had echocardiographic data at Y25. Echocardiographic exams were performed at the CARDIA Y25 exam. At each center, trained echo techs obtained images by a standardized protocol and scans were read centrally. Standard 2D and M-mode assessments were obtained. Published binary definitions of diastolic dysfunction were applied; pulmonary venous flow and mitral inflow at peak Valsalva maneuver measures were not available in CARDIA, so the (modified) Redfield definition of diastolic dysfunction was based on two of the four variables included in the definition. Logistic regression was used to determine demographic predictors of diastolic dysfunction. Results: The mean age in the sample was 50.2 years, mean BMI was 30.2 kg/m2, 46.7% of the sample was African American and 53.3% was white. The prevalence of diastolic dysfunction according to different definitions was as follows: 1.6 % according to the ASE definition, 24.3% according to the VALIDD definition, and 32.0% according to the (modified) Redfield definition. Using the ASE definition, 38.9% of individuals were unclassifiable for diastolic dysfunction. After adjustment for age, sex, and BMI, African Americans had 7.0 (95% C.I.: 3.4-14.4) times the odds of having diastolic dysfunction compared to whites according to the ASE definition. The comparable OR was 1.9 (95% C.I.: 1.6-2.2) for the Redfield Definition and 1.9 (95% C.I.: 1.6-2.2) for the VALIDD definition. Discussion: Existing definitions for diastolic dysfunction are difficult to implement in cohort studies and provide widely varying estimates for the prevalence of diastolic dysfunction as well as for predictors of diastolic dysfunction.