Abstract 4362331: Prevalence and Predictors of Left Ventricular Systolic Dysfunction during Treatment with Mavacamten for Obstructive Hypertrophic Cardiomyopathy

Background: Mavacamten is effective at improving symptomatic left ventricular outflow obstruction in patients with hypertrophic cardiomyopathy (HCM). Its main adverse effect is left ventricular systolic dysfunction (LVSD). However, predictors of LVSD are currently unknown. Research Question/Hypothesis: We sought to determine the prevalence and predictors of LVSD (LV ejection fraction <50%) in patients treated with mavacamten at a single center for at least 8 weeks. Methods: The clinical course of consecutive patients monitored according to the Risk Evaluation and Mitigation Strategies (REMS) program by serial echocardiographic and clinical examinations was investigated. Results: A total of 180 patients, median age 70 (interquartile range 58, 77) years, 58% women, 78% white, were followed for 58 (24, 87) weeks, of whom 21 (12%) patients developed LVSD, 4 (19%) at 4 weeks, 4 (19%) at 8 weeks, 4 (19%) at 12 weeks, 4 (19%) at 24 weeks and 5 (24%) >24 weeks after drug initiation or dose increase; 19 (90%) of LVSD events were detected during scheduled exams and 2 (10%) during hospital admissions for acute illness. The event was associated with acute decompensated heart failure in 5 (24%) patients; persistent atrial fibrillation (AF) was the precipitating factor in 2 (10%). History of AF (48% vs 24%, p=0.04), and left bundle branch block (LBBB) or ventricular pacing (57% vs 22%, p=0.001) were more common in those who developed LVSD compared with those who did not. In a multivariable model adjusted for age and sex, both factors were statistically significant, OR=4.3 (p=0.004) for LBBB/pacing and OR=2.6 (p=0.05) for history of AF. The addition of late gadolinium enhancement (LGE) extent to the model (in 67% patients with available cardiac MRI) was statistically significant (OR=1.2 per 1%, p=0.02). Permanent (29% vs 8%) and temporary (81% vs 28%, p<0.001) drug discontinuations were more common in patients with LVSD. The peak provoked outflow gradient was 100 (76, 121) mmHg before and 13 (8, 22) mmHg during mavacamten treatment (p=0.76 between groups). Two deaths occurred (1 in a patient with prior LVSD though 7 months after recovery of LV function). Conclusion: In this single-center study, LBBB/ventricular pacing, history of AF and LGE burden were associated with LVSD during mavacamten treatment for symptomatic obstructive HCM. Dose adjustment for patients with these risk markers may be considered and continued surveillance of LV function is warranted.