Predicting 15-Year Mortality in Adults With Congenital Heart Disease Using Disease Severity and Functional Indices

Abstract

Background: Disease severity and functional indices are widely used for risk stratification of patients with congenital heart disease (CHD). The predictive value of these classification systems for assessing long-term mortality is unknown. We aimed to determine and compare the predictive value of disease severity and functional indices for 15-year mortality in adults with CHD.

Methods: Between 2000 and 2002, we categorized 629 patients with CHD (median age, 24 years; 60% were men) on 5 indices: disease complexity scores based on criteria of Task Force 1 of the 32nd Bethesda Conference; Disease Severity Index; New York Heart Association functional class; Ability Index; and Congenital Heart Disease Functional Index (CHDFI). Harrell's concordance statistics index (C-index) was calculated for each classification system through Cox hazard regression analysis to evaluate their performance on predicting all-cause and cardiac mortality over the subsequent 15 years.

Results: Over the 15-year follow-up period, 40 patients died, resulting in a mortality rate of 4.56 per 1000 person-years. The CHDFI showed the highest discrimination ability for all-cause mortality (C-index = 0.74; P < 0.001) and cardiac mortality (C-index = 0.76; P < 0.001). The C-index for the other classifications ranged from 0.58 to 0.71 for all-cause mortality and 0.55 to 0.67 for cardiac mortality. The CHDFI showed statistical superiority toward the Disease Severity Index (P < 0.01).

Conclusions: These results suggest that the Task Force 1 of the 32nd Bethesda Conference, New York Heart Association functional class, Ability Index, and CHDFI could aid in predicting long-term mortality. The CHDFI demonstrated the highest discrimination ability and emphasizes the importance to integrate both anatomic and physiological variables to predict long-term mortality.