Risk of reoperative valve surgery for endocarditis associated with drug use.

Abstract

We aimed to quantify incidence and operative risks associated with reoperative valve surgeries (RVS) in patients with drug-associated infective endocarditis in a multi-center setting. We formed a registry of patients with drug-associated infective endocarditis who underwent valve surgeries at 8 US centers between 2011 and 2017. Outcomes of first-time valve surgery (FVS) and RVS were compared. Multivariable logistic regression models related RVS to 30-day mortality. Poisson regression models were fitted to evaluate temporal trends in overall case volume and proportions of patients undergoing RVS. The cohort consisted of 925 patients with drug-associated infective endocarditis who underwent a valve surgery, of which 652 were FVS and 273 were RVS. Patients undergoing FVS had fewer comorbidities than those undergoing RVS. Overall case volume increased from 108 in 2012 to 229 cases in 2017 (P < .001). The proportion of redo valve cases increased from 19% in 2012 to 28% in 2017 (P < .001). The 30-day mortality in RVS was higher compared with FVS (8.1% vs 4.8%; P = .049). An increase in unadjusted mortality rates were observed as the number of prior cardiac surgeries increased, from 4.8% in FVS to 11.8% in ≥3 RVS. Multivariable model demonstrated that RVS was associated with an increased risk of 30-day mortality (odds ratio, 2.22; 95% confidence interval, 1.22-4.06; P = .010). An increasing proportion of valve surgery for drug-associated infective endocarditis is for RVS. Despite being young and harboring few comorbidities, the RVS cohort is still susceptible to increased risk of 30-day mortality compared with those undergoing FVS.