By Michael H. Crawford, MD, Editor

SYNOPSIS: In a study of the relationship between the duration of antibiotic therapy and a positive surgically excised heart valve culture in patients with infective endocarditis, researchers observed positive valve culture incidence decreases exponentially on antibiotic therapy for 14 days then plateaus with no effect after 21 days of therapy.

SOURCE: Gisler V, Dürr S, Irincheeva I, et al. Duration of pre-operative antibiotic treatment and culture results in patients with infective endocarditis. J Am Coll Cardiol 2020;76:31-40.

In infective endocarditis (IE), the timing of valve replacement is challenging. Operate too early, and the new valve goes into an infected area. Wait too long, and the patient may succumb to the acute hemodynamic burden or other complications. Investigators from Switzerland sought to assess the temporal association of preoperative antibiotic (AB) therapy and excised valve culture (VC) results in patients with IE who were undergoing valve surgery.

This was a registry study of 231 IE patients at one center from 2005 to 2016 who met inclusion criteria for this analysis. Patients with other indications for surgery, not on AB at the time of surgery, or insufficient data were excluded (n = 121). Among the patients included in the study, 58 returned a positive VC of the execised valve and 173 were negative. The primary endpoint was the relationship between the duration of AB therapy and a positive VC result. The secondary endpoint was in hospital mortality. A McCabe-Jackson score was determined from clinical variables and dichotomized as non-fatal vs. rapid or ultimately fatal. Four categories of bacteria were considered: Staphylococcus (Staph), viridans group Streptococci (Strep), Enterococcus (Entero), and other species.

Staph and Entero IE exhibited a higher rate of positive VC compared to the others. Those with positive VC spent fewer days on AB therapy (6 days vs. 13 days; P < 0.001). In-hospital mortality was low in this study (5%), but those with positive VC compared to negative VC died more often (42% vs. 24%; P < 0.001). Although Strep and Staph were the most common causes of IE, Staph and Entero were more likely to return positive VC (odds ratio [OR] for Entero, 6.35; 95% confidence interval [CI] 1.94-20.78; P < 0.002; OR for Staph, 3.93; 95% CI, 1.57-9.84; P < 0.003). The analysis of the response to AB therapy on positive VC was non-linear. There was a rapid decrease in the first seven days of therapy, a slower decrease from seven to 13 days, a plateau from 14 to 21 days, and no further effect after 21 days. The adjusted probability of a positive VC compared to days of AB therapy showed the nadir of positive VC was at about 15 days for all organisms (except Entero, in which it was 20 days). The authors concluded that after seven days of appropriate AB therapy for IE, the effect on VC results after surgical excision were minor. AB therapy after 21 days did not affect VC positivity.


Frequently, the management of patients with IE results in a war of wills over if and when to operate. In general, the surgeons want as much AB treatment as possible. Cardiologists push for early surgery. The results of recent observational studies have supported earlier surgery in that those operated on early die less often. However, these studies suffer from survivor bias. If one could adjust for such bias, early surgery actually could reduce survival. Surgeons worry performing a procedure before the full course of AB is complete would involve placing a prosthetic valve in an infected field, which could result in an increase in valve-related complications. Gisler et al addressed this issue. At their institution, all explanted valves are cultured, allowing them to analyze the relationship between the duration of AB therapy and a positive VC. They showed the incidence of a positive VC declined rapidly during the first seven days of therapy, decreased slower over the next seven days, and then plateaued for the next seven days. After 21 days, there was no effect of AB therapy on VC positivity. Current guidelines list several Class I level of evidence B indications for early surgery: heart failure, fungal or resistant organisms, heart block, annular abscess, penetrating lesions, and persistent evidence of infection after five to seven days of AB therapy. Class IIa B recommendations include recurrent emboli, increasing vegetation size, severe valve regurgitation, and vegetation > 10 mm in diameter or mobile. If these clinical indications for surgery exist, the Gisler et al data suggest not waiting to put more AB into the patient, especially if past seven days of therapy. The caveat here is Gisler et al confirmed what other studies have shown, that a positive VC after surgery does increase the mortality rate. Predictors of finding a positive VC were Entero and Staph infection. Perhaps contracting one of these infections should be included in the guidelines for earlier surgery. Still, Gisler et al also showed the duration of AB therapy before the positive VC curve plateaued was 20 days for Entero and 15 for all other organisms. This would tend to favor waiting for two or three weeks of therapy if there were no other clinical indications to operate. Besides the study’s retrospective, observational design, there were other weaknesses. It was relatively small, and there were scant data reported about the clinical decision-making process for surgery. Also, the duration of IE before therapy was not considered. In addition, the patient’s care team made the decision for surgery. Over the 12-year study period, there was a trend toward earlier surgery.

There are many factors to weigh in the decision for surgery before a full course of AB is completed. Generally, clinical factors are more important than the duration of AB therapy for this decision. Gisler et al have added to our knowledge of the relationship between the duration of therapy and a valve with persistent infection, which is a factor that increases post-surgical mortality.