A multicenter evaluation of the HeartMate 3 risk score

The Heartmate 3 (HM3) risk score (HM3RS) was derived and validated internally from within the Multicenter Study of MagLev Technology in Patients Undergoing Mechanical Circulatory Support Therapy with HeartMate 3 (MOMENTUM 3) trial population and provides 1- and 2-year mortality risk prediction for patients in those before HM3 left ventricular assist device (LVAD) implantation. We aimed to evaluate the HM3RS in nontrial unselected patients, including those not meeting inclusion criteria for MOMENTUM 3 trial enrollment.Patients who underwent HM3 LVAD implant at 1 of 7 US centers between 2017 and 2021, with at least 1-year follow-up, were included in this analysis. Patients were retrospectively assessed for their eligibility for the MOMENTUM 3 trial based on study inclusion and exclusion criteria. HM3RS risk discrimination was evaluated using time-dependent receiver operating characteristic curve analysis for 1-year mortality for all patients and further stratified by MOMENTUM 3 trial eligibility. Kaplan-Meier curves were constructed using the HM3RS-based risk categories.Of 521 patients included in the analysis, 266 (51.1%) would have met enrollment criteria for MOMENTUM 3. The 1- and 2-year survival for the total cohort was 85% and 81%, respectively. There was no statistically significant difference in survival between those who met and did not meet enrollment criteria at 1 (87% vs 83%; p = 0.21) and 2 years postimplant (80% vs 78%; p = 0.39). For the total cohort, HM3RS predicted 1-year survival with an area under the curve (AUC) of 0.63 (95% confidence interval [CI]: 0.57-0.69, p < 0.001). HM3RS performed better in the subset of patients meeting enrollment criteria: AUC 0.69 (95% CI:0.61-0.77, p < 0.001) compared to the subset that did not: AUC 0.58 (95% CI: 0.49-0.66, p = 0.078).In this real-world evidence, multicenter cohort, 1- and 2-year survival after commercial HM3 LVAD implant was excellent, regardless of trial eligibility. The HM3RS provided adequate risk discrimination in "trial-like" patients, but predictive value was reduced in patients who did not meet trial criteria.