Can prosthesis type influence the recurrence of infective endocarditis after surgery for native valve endocarditis? A propensity weighted comparison.

OBJECTIVES

Our goal was to investigate whether the incidence of valve-related adverse events might be different depending on the valve substitute after valve replacement for left-sided native valve endocarditis.

METHODS

We assessed the long-term freedom from recurrence, reoperation and survival of 395 patients who had valve replacements for native valve endocarditis (314 mechanical vs 81 biological). Age <18 years, reoperation, prosthetic endocarditis, right valve involvement, valve repair and homograft implants were the main exclusion criteria. The balance between the 2 groups was addressed by weighting the results on the inverse of the propensity score.

RESULTS

After inverse probability of treatment weighting (IPTW), freedom from recurrence of infective endocarditis was not significantly different (mechanical 84.1 ± 3.2% vs 50.6 ± 21.7%; P = 0.29) nor was freedom from reoperation different (mechanical 85.7 ± 3.1% vs biological 50.9 ± 21.9%; P = 0.29). Excluding competing deaths, patients receiving a bioprosthesis had a similar subdistribution hazard of the above end points compared to recipients of a mechanical valve [recurrence IPTW: hazard ratio (HR) 1.631, 95% confidence interval (CI) 0.756-3.516; P = 0.21; reoperation IPTW-HR 1.737, 95% CI 0.780-3.870; P = 0.18]. Mechanical valves were associated with improved long-term survival (34.9 ± 5.8% vs 10.5 ± 7.4% at 30 years; P = 0.0009; in particular: aortic valve subgroup 41.6 ± 9.3% vs 10.1 ± 8.2%; P < 0.0001), although the hazard of cardiovascular mortality did not favour either valve type (IPTW: HR 1.361, 95% CI 0.771-2.404; P = 0.29).

CONCLUSIONS

Our analysis showed a clinical trend in favour of mechanical valves as valve substitutes for native valve endocarditis, especially in the aortic position. In view of long-term freedom from adverse events, the choice of the valve type should be tailored according to patient characteristics and specific clinical conditions.