External Validation of the Minnesota Pectoralis Muscle Risk Score to Predict Mortality after Ventricular Assist Device Implantation

Purpose We previously developed a post LVAD mortality prediction model using multicenter data incorporating pectoralis muscle mass and tissue attenuation obtained on preoperative CT scans. The purpose of this analysis was to evaluate the performance of this model in a prospective cohort including the most contemporary LVADs. Methods Since 2016, all patients who underwent first time LVAD implantation at the University of Minnesota had chest CTs performed prior to surgery (n=82). None of the patients included in the Minnesota Pectoralis Muscle Risk Score (MPRS) development cohort were included in this analysis. Unilateral pectoralis muscle mass indexed to body surface area (PMI) and attenuation (approximated by mean Hounsfield units; PHUm) were measured from pre-operative CT scans. MPRSs were calculated on all patients using the following variables: PHUm, PMI, African American race, creatinine, total bilirubin, body mass index, bridge to transplant status and presence or absence of contrast. Cox regression analysis was performed to test the association between the MPRS and post LVAD mortality. Receiver-operating characteristic (ROC) curves were then generated to test model discrimination at 30 and 90 days post LVAD implantation. Results The prospective cohort included 62% HeartMate 3, 17% HVAD, and 20% HeartMate 2 devices. Each two point increase in MPRS was associated with a 2.6 fold increase in the mortality after LVAD (HR 2.6, 95% CI 1.3-5.4, p=0.01). The area under the curve for the MPRS was 0.73 at 30 days and 0.72 at 90 days. Conclusion The Pectoralis Muscle Risk Score had favorable discrimination in this contemporary prospective cohort, which was largely comprised of patients receiving HeartMate 3 devices. These skeletal muscle measures, which are associated with heart failure stage and frailty, continue to add important information to pre-operative risk assessment.