Editorial commentary: surgical therapy for Staphylococcus aureus prosthetic valve endocarditis: proceed with caution (Caveat Emptor).

Embraced by treatment guidelines from various organizations (eg, the American Heart Association, the European Society of Cardiology), the role of surgery in the treatment of left-sided infective endocarditis (IE) has been expanding in recent decades. The generally accepted indications for surgical intervention are based almost entirely on observational studies, and in some instances on expert opinion. Recent studies utilizing large datasets of patients with native valve IE have used sophisticated analytic techniques to assess the benefits of surgical treatment of IE. These studies adjust for treatment selection bias (propensity for surgery), prognostic variables, and comorbidities, as well as survival bias expressed as the time of surgery. At least 2 of these investigations have demonstrated survival benefits of early valve surgery (EVS) in terms of mortality in hospital and/or at 6 months after surgery [1, 2]. Prosthetic valve infective endocarditis (PVIE) represents a special management category, given its distinct pathogenesis and severe complications, especially in early PVIE (ie, within 12 months after valve implantation). Current guidelines [3, 4] and some studies advocate aggressive surgical intervention for patients with early PVIE, particularly when caused by staphylococci or complicated by severe heart failure (New York Heart Association [NYHA] class III–IV), valve dehiscence, or paravalvular abscess. At large tertiary hospitals with cardiac surgical programs, surgical intervention occurs in 42%–49% of patients with PVIE [5, 6]. This is not unexpected given the abovementioned severe complications, which are unlikely to respond to antimicrobial therapy alone [5, 7]. Nevertheless, Lalani et al [7] were unable to demonstrate reduced in-hospital or 1-year mortality with valve replacement compared with medical therapy for PVIE. Using data collected prospectively between 1 January 2000 and 31 December 2006 by the multicenter International Collaborative on Endocarditis–Prospective Cohort Study (ICE-PCS), these investigators examined mortality rates among 1025 patients with definite PVIE, 490 (47.8%) of whom underwent surgery during their index hospitalization. After controlling for therapy selection bias using the inverse probability of treatment weighting, a propensity variable, in regression models, EVS was associated with lower mortality during hospitalization and at 1 year. However, when controlling for both treatment selection, as well as for survivor bias by including surgery as a time-dependent variable in a Cox proportional hazards model, the survival benefit of EVS was lost. The issue of whether all patients with Staphylococcus aureus (SA) PVIE should undergo early prosthetic valve replacement, irrespective of the presence or absence of heart failure and/or intracardiac complications, remains controversial. For example, both John et al [8] and Attaran et al [9] concluded that SA PVIE was an indication for prompt valve surgery even in the absence of intracardiac complications. Habib et al [10] similarly identified SA PVIE as a marker for poor early and late outcomes, and strongly recommended EVS for such patients. In contrast, Sohail et al [11] recommended EVS for SA PVIE only in the presence of the abovementioned hemodynamic or periannular complications. In the current issue of Clinical Infectious Diseases, Chirouze and colleagues studied 168 well-characterized cases of Received 21 October 2014; accepted 24 October 2014; electronically published 10 November 2014. Correspondence: Arnold S. Bayer, MD, Los Angeles Biomedical Research Institute, at Harbor-UCLA, 1124 W Carson St, Bldg RB2, Rm 225, Torrance, CA 90502 (abayer@labiomed.org). Clinical Infectious Diseases 2015;60(5):750–2 © The Author 2014. Published by Oxford University Press on behalf of the Infectious Diseases Society of America. All rights reserved. For Permissions, please e-mail: journals. permissions@oup.com. DOI: 10.1093/cid/ciu877