CardioExchange Archive

In the first five years after valve replacement approximately 3-6% of patients will develop prosthetic valve endocarditis (PVE). These patients are much more likely to die. Now a large observational study — the first of its type — has found that surgery is no better than medical therapy in reducing mortality in these patients. The report from the ICE-PCD (International Collaboration on Endocarditis  — Prospective Cohort Study) investigators appears in JAMA Internal Medicine.

The trial followed 1,025 PVE patients, about half of whom had surgery (47.8% versus 52.2% who received medical therapy). Mortality, both in-hospital and at 1 year, was significantly lower in the surgery group. This was true using both unadjusted data and a multivariable model using a propensity score. However, after adjusting for survivor bias there were no significant differences between the two groups:

In-hospital mortality:

• Unadjusted hazard ratio: 0.68, CI 0.53-0.87
• Multivariable model: 0.44, CI 0.38-0.52
• Multivariable model controlling for survivor bias: 0.90, CI 0.76-1.07

One-year mortality:

• Unadjusted hazard ratio: 0.68, CI 0.55-0.87
• Multivariable model: 0.57, CI 0.49-0.67
• Multivariable model controlling for survivor bias: 1.04, CI 0.89-1.23

The investigators observed a significant improvement in mortality in the subgroup of surgical patients who had strong indications for surgery.

The findings, write the investigators, suggest that “survival bias and timing of surgery should be considered when evaluating the treatment effect on mortality.” Since patients in the study had surgery at a median of 8 days after hospital admission, the investigators pointed out  that their study did not rule out possible benefits of very early surgery after admission.

In an accompanying editorial, Ann Bolger writes that the anatomic features of PVE  “are dramatic and intuitively seem important to treat given that they are unlikely to improve with antibiotic therapy alone.” However, she writes,

“Impressive as these visible anatomic and functional features of PVE may be, perhaps we overvalue them in our overall assessment of patient risk. Guidelines have consistently emphasized these features as indicators of a need for surgery. However, their presence does not guarantee poor immediate outcomes. It is important, as we discuss treatment options with the patient and care team, that we be nuanced in our assessment of the patient’s comorbidities, as well as the individual patient’s ability to tolerate some of the functional sequelae of prosthetic infection.”