Early surgery fails to edge out medical therapy for endocarditis

One-year mortality rates remained similar for patients with prosthetic valve endocarditis who received early surgery or medical therapy in an observational study that adjusted for biases. The study was published online July 15 in JAMA: Internal Medicine.

Guidelines from the American College of Cardiology, the American Heart Association and the European Society of Cardiology recommend early surgery to treat infective endocarditis patients, especially in those with complications such as heart failure. But the recommendations largely are based on expert opinion and observational studies because of a paucity of clinical trial data, wrote Tahaniyat Lalani, MD, MPH, of the Naval Medical Center Portsmouth in Portsmouth, Va., and colleagues.

Lalani et al used a prospective observational study, the International Collaboration on Endocarditis-Prospective Cohort Study (ICE-PCS), in an effort to overcome some of the limitations of previous small and single-center studies. ICE-PCS includes 64 sites from 28 countries.

Their primary outcomes were all-cause in-hospital and one-year mortality. Of 1,025 patients diagnosed with definite prosthetic valve endocarditis in ICE-PCS between 2000 and 2006, 47.8 percent underwent early surgery and 52.2 percent received medical therapy. The researchers defined early surgery as replacing or repairing the infected prosthetic valve during the initial endocarditis hospitalization.

The early surgery group had a larger proportion of complications compared to the medical therapy group. In unadjusted analyses, in-hospital mortality in the surgery group was 22 percent vs. 26.7 percent in the medical therapy group and one-year mortality was 27.1 percent vs. 36.6 percent, respectively.

To address methodological concerns, they applied propensity score-based methods to adjust for treatment effects and kept surgical patients in the medical treatment group until the date of surgery to reduce survival bias. The one-year survival advantage persisted in analyses adjusting for treatment selection bias but disappeared after controlling for survival bias.

A subgroup analysis found lower incidence in-hospital mortality in the highest surgical propensity quintile of early surgery patients and a lower one-year mortality in the fourth and fifth quintiles.

 “[O]ur results emphasize that survival bias and timing of surgery should be considered when evaluating the treatment effect on mortality,” Lalani et al wrote. “Although patients underwent surgery at a median of eight days after admission, the potential benefit of earlier intervention was not evaluated and may influence outcome.”