Applicability and performance of heart failure prognostic scores in dilated cardiomyopathy: the real-world experience of an Italian referral center for cardiomyopathies

Background: The performance of heart failure (HF) risk models is validated in the general population with HF but in specific aetiological settings, and specifically in dilated cardiomyopathy (DCM), has scarcely been explored. We tested eight of the main prognostic scores used in HF in a large real-world population of patients with DCM. Methods: We included 784 consecutive DCM patients enrolled, both inpatients and outpatients, enrolled between January 2000 and December 2017. The risk of 1 and/or 3-year all-cause mortality/heart transplantation/durable left ventricular assist device (LVAD) implantation (D/HTx/LVAD) was estimated in our cohort according to the following risk scores SHFM, 3-CHF, CHARM, MAGGIC, GISSI-HF, MECKI, Barcelona Bio-HF, Krakow score and their accuracy calculated through the receiver operator characteristic (ROC) curve analysis.Results: During a median follow-up of 5.8 years (Interquartile Range 3.2-7.6 years), 191 patients (20%) died or underwent HTx/LVAD (158 deaths, 30 heart transplantations, and 3 LVAD implantations). The high missing rate allowed to calculated only four prognostic models (MAGGIC, CHARM, 3-CHF and SHFM). All the scores overestimated the rate of D/HTx/LVAD. The prognostic accuracy was suboptimal for MAGGIC (AUC 0.754) and CHARM (AUC 0.720) scores and only modest for 3-CHF (AUC 0.677) and SHFM (AUC 0.667).Conclusions: Main prognostic scores for the risk stratification of HF are only partially applicable to real-world patients with DCM. MAGGIC and CHARM scores showed the best accuracy, despite the overestimation of risk. Our findings corroborate the need of specific risk scores for the prognostic stratification of DCM. Clinical perspective:: What is new? The present study is the largest analysis in literature which investigate how the main existing heart failure prognostic risk scores performed in a real-world of dilated cardiomyopathy population, both in- and outpatients. What are the clinical implications?DCM is a stand-alone model of heart failure, where the performance of multiple heart failure prognostic scores for the risk stratification is quite limited. The need for contemporary, dedicated prognostic scores in this disease is increasingly evident.