Abstract 11622: Temporal Trends and Predictors of Hospital Mortality among Patients with Concomitant Coronary and Valvular Heart Diseases: A 25-year Longitudinal Study in New York State

Abstract

Background: Valvular heart disease (VHD) and coronary artery disease (CAD) often coexist in the same patient. However, data are limited on temporal trends and predictors of hospital mortality among these high risk patients who increasingly undergo combined valve (VS) and coronary artery bypass surgery (CABG).

Methods: We conducted a longitudinal analysis of records from the New York Statewide Planning and Research Cooperative System (SPARCS) inpatient database (1983-2007); each bore a principal or secondary ICD-9 code for VHD±CAD±CABG. Linear regression was used to model temporal trends in hospital deaths; χ²tests and multivariable logistic regression were used for in-hospital mortality prediction.

Results: We identified 1,881,356 patients with VHD±CAD; among these patients, coexisting CAD (n=645,240) was associated with increased mortality during the study period (CAD=6.4% vs No CAD=4.8%, p<.001), rising by 136 deaths/yr over 25 yrs. VS was performed in 58,347 patients with VHD+CAD; 16,980 (29%) of these patients underwent isolated VS while 41,367 (71%) underwent VS+CABG. Though death rates among the surgical patients declined during the study period, they were consistently highest among patients with coexisting CAD (Figure). By multivariable analysis, CABG was the strongest independent predictor of mortality among VHD+CAD patients who underwent VS (p<.001); other independent predictors were age >65 yrs, female gender, non-White race, heart failure and mitral VHD (p<.001 all).

Conclusions: CAD presence adds substantially to risk of hospital death among patients with VHD. CABG does not appear to mitigate this risk and may even be associated with increased intra- or peri-operative mortality during hospitalization. However, lack of data on extent and severity of disease among these patients precludes firm inferences about this association. Randomized clinical trials are needed to better define optimal treatment options in this patient population.