June 27th, 2012
Panel: Early Surgery for Infective Endocarditis
CardioExchange has asked three experts to weigh in on findings from a randomized trial, published in the New England Journal of Medicine, comparing early surgery with conventional treatment for infective endocarditis.
Findings from the Trial
Seventy-six patients with left-sided infective endocarditis, severe valve disease, and large vegetations were randomized to undergo early surgery or conventional treatment. All participants received beta-lactam–based antibiotic therapy. The primary endpoint — a composite of in-hospital death and embolic events within 6 weeks after randomization — occurred in 3% of the early-surgery group and 23% of the conventional treatment group (1 vs. 9 patients; hazard ratio, 0.10; P=0.03). The difference was due entirely to a greater incidence of embolism in the conventional-treatment group. The two groups did not differ in all-cause mortality at 6 months. No additional embolic events occurred between 6 weeks and 6 months.
Opinions from the Experts
Question: Will the findings from this trial change your practice? If so, how? If not, why not?
Although this is an excellent study, given its limitations, it will not prompt a change in practice for me. These limitations include the relatively small cohort size and the unusual distribution of pathogens that were identified (a predominance of streptococcal species). The work does focus us on the timing of surgical intervention, which deserves much more investigation, as there was no controversy as to need of surgery in the study cohort.
Surgeons can be reluctant to operate early on patients with infective endocarditis for two reasons. First, if the organisms are not sensitive to the antibiotic therapy, the operation may be associated with fulminating sepsis and vasoplegia, which can be fatal. Second, the presence of small foci of vegetations disseminated on the mitral valve leaflets may preclude the ability to perform a valve repair. With conventional treatment, the major risk in delaying the operation is the possibility of systemic emboli, particularly to the central nervous system.
In this trial, the antibiotic treatment with beta-lactam agents and, in about a third of the participants, additional amino-glycoside successfully controlled the bacteremia in all but one patient. The authors acknowledge the relatively lower incidence of Staphylococcus aureus (11%) in their trial than in other studies. Nevertheless, I wonder about the near-total absence of resistant organisms in the whole randomized group of patients. Did the selection process exclude such patients, given that in most practices, resistant gram-negative and S. aureus organisms are found? Endocartitis involving such organisms often requires changing the antibiotic regimen. The favorable outcome of the patients with early surgery is plausible for patients who have suffered from an infection with sensitive streptococcus organism. To generalize the trial findings to staph-resistant infections, one would need to use antimicrobial agents with a very low incidence of in vitro resistance, such as daptomycin, perioperatively (to provide greater assurance of controlling bacteremia).
Surgeons must concern themselves with whether to perform mitral valve repair or replacement in the early infectious process. This study validates such concern, but the authors do not discuss it. Of the patients with involvement of the mitral valve, only 8 of 22 (37%) in the early-surgery group underwent mitral valve repair, compared with 11 of 17 (65%) in the conventional-treatment group. The difference — 28% — equals the incidence of emboli in the conventional group. Some would consider a valve replacement to be a complication, compared with a valve repair. Therefore, there is a trade-off between emboli and valve replacement. Of course, the disability caused by emboli to the central nervous system is greater than the morbidity associated with a mitral valve replacement, and 5 conventionally treated patients in this trial developed cerebral emboli. On the other hand, the aortic valve patients all underwent replacement, and there is no concern about a lower incidence of reparability in those patients.
From this study, one can conclude that it is reasonable to perform early surgery in the patients with aortic valve disease who have streptococcal infection and large vegetations, but one cannot reach a similar conclusion for patients with nonstreptococcal infections or infections of the mitral valve.
These findings favor early aggressive treatment for infective endocarditis. The numbers suggest that operating on 5 patients within 48 hours of making the diagnosis would prevent one occurrence of death or peripheral embolization over 6 months.
The study was well performed, but the participants were highly selected, which may limit applicability. Of 134 screened patients with infective endocarditis, 44 were initially excluded for unspecified reasons, and 26 of those underwent urgent surgery. Fourteen additional patients were excluded because of clinical reasons or refusal. The remaining 76 patients, recruited over 5 years, had infective endocarditis with a large left-sided vegetation and severe valve disease — yet no more than mild heart failure, no prior major embolic stroke with risk of hemorrhagic transformation, and no serious coexisting condition.
Unfortunately, patients with infective endocarditis are usually more complicated. Just 3 weeks ago, I cared for a patient with infective endocarditis and a large vegetation. My patient had an infection of a tissue aortic valve prosthesis. He also had several coexisting conditions, including chronic systolic heart failure and severe peripheral arterial disease. While awaiting a final decision regarding re-do cardiac surgery, he deteriorated and ultimately died.
In practice, infective endocarditis patients commonly have complicating conditions, probably because those conditions predispose them to infective endocarditis. For patients with infective endocarditis without complicating conditions, however, this informative study suggests that surgery within 48 hours of making the diagnosis is the best strategy.
What’s your take on the implications of this study for clinical practice? Feel free to pose questions to our experts.