Identification of risk factors for incident left ventricular systolic dysfunction and predictors of cardiac recovery in patients with systemic sclerosis.

BACKGROUND Cardiac involvement in systemic sclerosis (SSc) is a leading cause of death. We sought to investigate predictors of incident left ventricular systolic dysfunction (LVSD) and cardiac recovery in SSc. METHODS 2,303 patients in the Johns Hopkins Scleroderma Center Research Registry and 13,209 echocardiograms were analyzed. We identified predictors associated with incident LVSD defined by transitions in left ventricular (LV) ejection fraction (EF) states (EF≥50% declining to <50% and EF>35% dropping to ≤35% [severe LVSD]) by fitting multivariate logistic regression models with time-varying and invariant variables. Variables associated with cardiac recovery were identified by fitting multivariate logistic regression models using important variables identified from random forest analysis. RESULTS Male sex, Black race, diffuse skin disease, higher mRSS, echocardiographic evidence of pulmonary hypertension (PH), kidney disease, and atrial fibrillation (AFib) were associated with increased odds of incident LVSD (EF<50%), while anti-centromere and anti-topoisomerase-1 were protective. Male sex, higher mRSS, PH, skeletal myopathy, kidney disease, AFib, and anti-Ku antibodies were associated with higher odds of incident severe LVSD (EF≤35%). For previous EF<50%, tendon friction rubs were associated with lower odds of cardiac recovery, and anti-RNA polymerase III (POLR3) with higher odds. For previous EF≤35%, diabetes was associated with lower odds of recovering to EF>35%. CONCLUSIONS Distinct demographic, SSc-specific and cardiac characteristics associate with increased risk of incident LVSD in SSc, with skeletal myopathy and anti-Ku antibodies being important risk factors for severe disease. Some patients improve, which is more likely in anti-POLR3-positive patients.