Results:

A total of 15,138 patients were included in the INTERMACS cohort. Patients who had an LVAD with a prior BTR strategy were more likely to be young, female with lower body mass index, and more often had nonischemic cardiomyopathies, and had a relatively new diagnosis of heart failure. They were less likely to have an implantable-cardioverter defibrillator (ICD), chronic kidney disease, lower INTERMACS profiles (higher acuity of illness), and more likely to require dialysis or require a biventricular assist device, and less likely to require a temporary mechanical percutaneous device. At a median of 323.3 days, 192 (1.3%) of the INTERMACS population had cardiac recovery, which required LVAD explantation or device deactivation. Recovery occurred in 2.1% of the patients during the first month, 14.6% by 3 months, and 80% of recoveries by 2 years post-LVAD implantation. 11.2% of patients who were implanted as a BTR experienced cardiac recovery versus 1.2% of patients who were non-BTR.

A competing regression analysis in the INTERMACS cohort identified multiple univariable predictors of cardiac recovery. Patients with complex congenital heart disease and restrictive heart disease were less likely to experience cardiac recovery. Six independent predictors of cardiac recovery were identified: age <50 years, nonischemic etiology, time from cardiac diagnosis <2 years, no ICD, serum creatinine ≥1.2 mg/dl, and LV end-diastolic dimension <6.5 cm (c-statistic of 0.85; 95% confidence interval [CI], 0.82-0.88; p < 0.0001). A prognostic score was created called the INTERMACS Cardiac Recovery Score (I-CARS). The score was used to identify three groups of cardiac recovery: low probability (0-3 points), intermediate probability (4-6 points), and high probability (7-9 points). I-CARS was applied to the BTR and non-BTR groups with an area under the receiver operating characteristics curve of 0.84 (95% CI, 0.81-0.87). Pharmacotherapy showed the use of beta-blockers, angiotensin-converting enzyme inhibitors or angiotensin-receptor blockers, and mineralocorticoid receptor antagonists to be associated with cardiac recovery in a univariable analysis.

The validation cohort consisted of 190 patients who were mostly male, median age 60 years, and New York Heart Association class IV patients. I-CARS stratified patients on the basis of their probability of cardiac response and recovery. Patients with a high probability score did have improvement in myocardial function.