Bacteremia after oral surgical procedures and antibiotic prophylaxis

Abstract: The use of antibiotic prophylaxis in patients predisposed to infective endocarditis who are undergoing oral surgical procedures is widely accepted. Because no controlled clinical trials of antibiotic regimens for the prevention of endocarditis in humans have been done, recommendations are based on results of, e.g., prophylactic studies in animal models of endocarditis, in vitro susceptibility data of pathogens causing endocarditis, procedure-related studies of bacteremia and studies on the efficacy of antimicrobial prophylaxis for prevention of postsurgical bacteremia. The antibiotic regimens usually include agents that are highly active against viridans streptococci, since these are the most common causative bacteria in endocarditis and are frequently recovered from blood during various invasive oral procedures. Minimizing the occurrence of postoperative bacteremia has been considered to be the most important factor in the prevention of infective endocarditis and several clinical studies have shown a marked reduction in bacteremia, following dental extraction by the use of antibiotic prophylaxis. In the present studies, a lysis-filtration blood culture technique was used to investigate the incidence, type and magnitude of bacteremia following common invasive procedures performed in the oral cavity and the effects of prophylaxis with current recommended antibiotics on bacteremia in patients undergoing dental extraction. Blood samples for culture were obtained before, during and 10 minutes after the procedure. Bacteremia occurred in 100% of patients after dental extraction, in 70% after dental scaling, in 55% after third-molar surgery and after tonsillectomy, and in 20% after endodontic treatment. The incidences are generally higher than those previously reported, indicating a high sensitivity of the blood culture technique used. Anaerobic bacteria, such as actinomyces and lactobacilli, were usually predominant, while aerobic bacteria, such as staphylococci, enterococci, corynebacteria and haemophilus were isolated in low numbers. Viridans streptococci were isolated in high amounts from all five treatment groups, most frequently following dental extraction (85%) and dental scaling (55%). The elimination of bacteria from blood was rapid. In 90% of the patients, with viridans streptococci during treatment, the microorganisms were eliminated within 10 minutes. Prophylactic administration of penicillin V, amoxicillin or cefaclor did not reduce the incidence or the magnitude of bacteremia after dental extraction, as compared to placebo. No difference in bacteremia after dental extraction was noted when prophylaxis with erythromycin and clindamycin were compared. The absence of reduction in bacteremia in the prophylaxis groups was not due to high bacterial resistance to the drugs, since most of the bacterial strains isolated had MICs and MBCs well below the individual antibiotic serum concentration during the extraction. Since bacteremia after dental extraction is rapidly eliminated, the ineffectiveness of the prophylaxis is probably due to the brief period when microorganisms are exposed to the antibiotics. On the basis of the present observations, a protective effect against endocarditis probably results from clearance of microorganisms at a later stage of the infection process, rather than during the initial period of transient bacteremia. By use of MIC breakpoints recommended by the NCCLS, virtually all bacteria isolated from blood after dental extraction were classified as susceptible to penicillin V, ampicillin, cefaclor, erythromycin and clindamycin. Cefaclor was the least active drug and MBCs ±128 mg/L were observed in some viridans streptococci. These results indicate that cefaclor is not suitable as an alternative to the drugs recommended for prophylaxis of endocarditis. In a quality control study comparing the E test method with the reference agar dilution method for NUC determination of bacteria isolated from blood after dental extraction, the agreement was high. The E test MICs agreed within ±1 dilution step with the agar dilution MICs in more than 90% of the aerobic bacteria (140 strains) for penicillin V, cefaclor, clindamycin and erythromycin, respectively. For anaerobic bacteria (311 strains), the agreement was 93%, 52%, 90% and 94% for the antibiotics tested. The discrepancy in agreement between the two methods concerning determination of the cefaclor MIC for anaerobic bacteria might be related to instability of cefaclor in the agar medium during anaerobic conditions or to the release and diffusion process of cefaclor from the E test strip.