Characteristics and longer-term outcomes of paravalvular leak after aortic and mitral valve surgery.

Abstract

Paravalvular leak (PVL) is often seen after aortic (AV) and mitral valve (MV) surgery, either due to infection or valve dehiscence. We sought to describe predictors of longer-term outcomes in patients who developed PVL after AV and MV surgery and were considered eligible for reoperative cardiac surgery (RCS). We studied 495 such patients (65 ± 14 years, 65% men, 47% with MV PVL) who presented at our center between January 2003 and December 2011. Patients with severe mitral/aortic stenosis, patients with less than mild PVL, and those with prohibitive risk precluding RCS were excluded. Society of Thoracic Surgeons (STS) score was calculated. Primary endpoint was mortality. At baseline, mean STS score and left ventricular ejection fraction were 5.8 ± 4% and 52 ± 12%, respectively. In total, 105 (21%) had infective PVL and 72% had moderate or greater PVL. At a median of 8 days, 351 (71%) patients underwent RCS to repair PVL (3% in-hospital postoperative mortality), and at 6.6 ± 4 years, 230 (47%) patients died. On multivariable Cox survival analysis, greater STS score (hazard ratio or HR 1.35), mitral versus aortic PVL (HR 1.66), infectious etiology (HR 2.05), and greater right ventricular systolic pressure (HR 1.09) were associated with greater longer-term mortality, whereas surgery (HR 0.58) was associated with improved longer-term survival (all P < .05). Patients who develop mild or greater PVL after AV/MV surgery have a high rate of longer-term mortality, despite excellent perioperative outcomes. Greater STS score, right ventricular systolic pressure, infectious etiology, and MV (vs AV) involvement were all independently associated with long-term mortality, whereas RCS for PVL closure was associated with improved longer-term survival.