Validation of the Danish Comorbidity Index for Acute Myocardial Infarction for predicting one-year mortality in patients with venous thromboembolism.

INTRODUCTION:The Danish Comorbidity Index for Acute Myocardial Infarction (DANCAMI) was developed to predict one-year mortality after myocardial infarction. We validated DANCAMI in predicting one-year mortality after venous thromboembolism (VTE). MATERIALS AND METHODS:We identified all first-time VTE patients in Denmark during 2000-2015. Using Cox regression, we assessed the performance of DANCAMI to predict one-year all-cause mortality using Nagelkerke's R2, Harrell's C-Statistic, the net reclassification index (NRI), and the integrated discrimination improvement (IDI). We compared the performance of DANCAMI with the Charlson Comorbidity Index (CCI) and the Elixhauser Comorbidity Index (ECI) and evaluated whether DANCAMI comorbidities not included in the CCI predicted one-year mortality. We stratified the analyses by type (deep vein thrombosis [DVT] and pulmonary embolism [PE]) and presence of provoking risk factors. RESULTS:We identified 108,824 VTE patients of whom 20,649 (19%) died within one year. The R2, C-Statistic, NRI, and IDI for DANCAMI were 0.35, 0.76, 0.63, and 0.098 for VTE overall; 0.43, 0.80, 0.70, and 0.105 for DVT; and 0.24, 0.71, 0.54, and 0.083 for PE. The R2 and C-Statistic for VTE overall were 0.35 and 0.76 for CCI and 0.33 and 0.75 for ECI. After adjusting for age, sex, and all CCI comorbidities, seven DANCAMI comorbidities, not included in the CCI, predicted increased mortality. DANCAMI performed better than the CCI and ECI in predicting mortality after provoked VTE, including provoked DVT and PE. CONCLUSION:DANCAMI performed comparable to existing comorbidity indices in predicting one-year mortality after first-time VTE overall, but better after provoked VTE.