Elimination of bacteraemia after dental extraction: comparison of erythromycin and clindamycin for prophylaxis of infective endocarditis

Transcription

1 Journal of Antimicrobial Chemotherapy (99) 7, Elimination of bacteraemia after dental extraction: comparison of erythromycin and clindamycin for prophylaxis of infective endocarditis G. Hall', C. E. Nord'* and A. Heimdahl* "Department of Oral Surgery; b Department of Immunology, Microbiology, Pathology and Infectious Diseases, Huddinge University Hospital, Karolinska Institute, S-I 8, Sweden Erythromycin and clindamycin are currently recommended for antibiotic prophylaxis of infective endocarditis in predisposed patients allergic to penicillin undergoing oral invasive procedures. Thirty-eight healthy patients were randomized to receive either erythromycin ( g) or clindamycin (0. g) orally.5 h prior to dental extraction. Blood samples for microbiological investigation were collected before, during and 0 min after surgery and were processed by lysis filtration under anaerobic conditions. The incidence of bacteraemia with viridans streptococci was 79% in the erythromycin group and 7% in the clindamycin group. No statistically significant difference was noted in incidence or magnitude of bacteraemia with viridans streptococci or anaerobic bacteria between the two groups, at any sampling time. Ninety-six aerobic and anaerobic strains recovered from the blood samples were tested for their susceptibility to erythromycin and clindamycin as well as to penicillin V and ampicillin. The antimicrobials were found to be highly active against the majority of bacteria except for some enterococci, staphylococci and veillonella. Protection from endocarditis by prophylaxis with erythromycin or clindamycin must be due to elimination of bacteria at a later stage in the development of the disease, rather than by elimination of bacteria from blood during the short period of postoperative bacteraemia. Introduction Infective endocarditis is in general a bacterial infection of the endocardial surface of the heart. A bacteraemic period followed by adherence of microorganisms to fibrin-platelet vegetations formed at the site of heart endocardial or valvular lesions and subsequent bacterial multiplication are fundamental events for the development of the disease (Freedman, 987). Viridans streptococci are the most common causative bacteria in endocarditis, being found in 7-8% of cases (Young, 987), indicating that the oral cavity is the main source of transient asymptomatic bacteraemia. Invasive procedures in the oral cavity such as dental scaling and extraction induce bacteraemia in 70-00% of patients (Heimdahl et al., 990). Downloaded from by guest on 0 October 08 Corresponding author /9/ $.00/0 99 The British Society for Antimicrobial Chemotherapy

2 78 G. Hall et al. Though no controlled clinical trials of antimicrobial prophylaxis of infective endocarditis in predisposed patients have been carried out, guidelines have been published by several national committees. The recommendations are based largely on indirect data from studies of bacteraemia, pharmacokinetics, clinical experience, compliance with recommended antimicrobial drugs and animal experiments. Reduction of postoperative bacteraemia has been considered crucial to the prophylaxis of endocarditis. Several clinical studies have shown a drastic reduction of bacteraemia following dental extraction when prophylactic penicillins are used (Shanson, Cannon & Wilks, 978; Baltch et al., 98; Head et al., 98; Roberts, Radford & Holt, 987). Other reports claim that mechanisms other than killing of bacteria in blood by penicillins may be responsible for prevention from endocarditis (Berney & Francioli, 990; Hall et al., 99). The American Heart Association recommends erythromycin, g orally h before the procedure followed by 0.5 g h later, for endocarditis prophylaxis in penicillin-allergic patients undergoing invasive oral procedures. Clindamycin, 0. g orally h before the procedure and 0.5 g h later may be used as an alternative for individuals who cannot tolerate erythromycin (Dajani et al., 990). The British Society for Antimicrobial Chemotherapy proposes clindamycin as the drug of choice for patients allergic to penicillin and recommends 0. g as a single oral dose h before treatment (Simmons et al., 99). Since few reports have been published concerning the efficacy of antimicrobials for patients allergic to penicillin (Phillips et al., 97; Shanson et al., 985; Sefton et al., 990) the present study was undertaken to compare the effects of prophylaxis with erythromycin and clindamycin on the incidence, type and magnitude of bacteraemia in patients undergoing dental extraction. The in-vitro activity of erythromycin, clindamycin, penicillin V and ampicillin was also determined against the bacteria isolated. Patients Materials and methods The study included 8 patients ( males and females; mean age: 8.5 years, range 5 to 7 years) who were referred to the Department of Oral Surgery, Huddinge University Hospital, Sweden for dental extraction. Patients with a history of allergic reactions or systemic symptoms following clindamycin or erythromycin therapy were excluded as were pregnant women and patients with cardiovascular, renal, hepatic or gastrointestinal diseases. None of the patients had received any antimicrobial agents for the weeks preceding treatment. Except for analgesics and oral contraceptives, none of the patients was on any medication. Informed consent was obtained from all patients, and the guidelines for human trials were followed as set by the Ethics Committee of Huddinge University Hospital. Downloaded from by guest on 0 October 08 Dental extraction Thirty-eight patients had one single tooth extracted by the same surgeon (G. H.) because of dental caries or chronic periradicular osteitis. Local analgesia was achieved by injection of prilocaine (0 g/l) with felypressine ( g/l) (Citanest, Octapressin, Astra, Sodertalje, Sweden).

3 Bacteraemia after dental extraction 785 Antibiotic administration The study was designed as a randomised double-blind trial. Nineteen patients ( males and 5 females; mean age 7 years, range 5-7 years) were randomized to receive two 0.5 g erythromycin stearate tablets (Abboticin, Abbott, England).5 h before dental extraction and 9 patients (0 males and 9 females; mean age 50 years, range 5-7 years) were randomized to receive two 0. g clindamycin capsules (Dalacin, Upjohn, England).5 h before dental extraction. Blood sampling procedure An indwelling catheter was placed in the cubital vein of the arm under strict aseptic conditions. Blood samples (8. ml) for isolation of microorganisms were taken before, during and again 0 min after dental extraction by means of the catheter into Vacutainer tubes containing sodium polyanetholsulphonate (SPS,.7 ml of a 0.5% solution; Becton Dickinson & Co., Rutherford, NJ). The catheter was rinsed by accurate flushing with 5 ml NaCl, followed by aspiration of 5 ml blood before each blood sample was drawn. The samples were processed immediately. Blood samples for antibiotic assay were also obtained at the time of dental extraction. Blood culture technique The blood samples were injected into bottles with 0.9 L of a lysing solution, ph 0, containing 0.08% Na CO, 0.005% Triton X-00 (Rohm and Haas, Darmstadt, Germany) and 0.00 L of streptokinase-streptodornase (Varidase, Lederle Laboratories, Spain) to avoid clogging of the filters. Vacuum filtration was performed in a 0.5 nm pore filter system (Millipore AB, Solna, Sweden) under a continuous flow of nitrogen (Heimdahl et al., 985). After filtration, the niters were placed on brain heart infusion agar plates (Difco Laboratories, Detroit, Michigan) for anaerobic incubation at 7 C for 0 days. Identification of microorganisms Aerobic and anaerobic microorganisms were identified using the methods described by Lennette et al. (985). Quantitative counts were estimated from the numbers of colonies visible on the filters. Assay of antibiotics in serum The serum concentrations of erythromycin and clindamycin were determined by an agar diffusion method using paper discs and Micrococcus luteus ATCC 9 as indicator strain (Jailing et al., 97). The medium used was Antibiotic Sensitivity Medium (PDM-AB BioDisk Sweden), ph 7.. Downloaded from by guest on 0 October 08 Antibiotic susceptibility tests The minimal inhibitory concentrations of erythromycin, clindamycin, penicillin V and ampicillin were determined as recommended by the National Committee for Clinical

4 78 G. Hall et at. Laboratory Standards, for the different bacterial strains (NCCLS, 990a,*). The bacterial isolates were inoculated onto agar plates (Antibiotic Sensitivity Medium, supplemented with 5% horse blood) containing two-fold dilution steps ranging from 8 to 0.0 mg/l of the antimicrobials. An agar plate without antimicrobials was included as well as reference strains with known MICs. Aerobically-incubated plates were examined after h incubation at 7 C and anaerobic plates were incubated at 7 C in anaerobic jars (GasPak, BBL, Cockeysville, Maryland, USA) for 8 h. The reference strains used were Staphylococcus aureus ATCC 9, Bacteroides fragilis ATCC 585, Clostridium perfringens ATCC and Streptococcus pyogenes RAF-M. The MIC was defined as the lowest concentration of the antibiotic inhibiting growth; a single colony or faint haze was disregarded. MIC breakpoints for susceptible and resistant categories used were according to the guidelines of NCCLS (99), whenever possible. For viridans streptococci, MICs and MBCs of erythromycin, clindamycin, penicillin V and ampicillin were determined by a standard two-fold broth macrodilution method (NCCLS, 99a) from 8 to 0.0 mg/l. Antibiotic Sensitivity Medium was used as substrate and the MIC was read as the concentration of penicillin in the first tube that showed no visible growth after h of incubation at 7 C in 5% CO. The reference strain used was 5. aureus ATCC 9. The MBC was determined by plating 50 ^L from the last tube with visible growth and all subsequent tubes without visible growth on agar plates supplemented with 5% horse blood. The agar plates were incubated at 7 C in 5% CO for h. The MBC was defined as the lowest concentration of antibiotic that killed 99.9% or more of the initial inoculum (5 x 0 cfu/ml) and tolerance was defined as MBC/MIC ratio ^. fi-lactamase assay 0-Lactamase production was determined by the nitrocefin assay (Olsson, Dornbusch & Nord, 977). Statistical analysis Difference in incidences of bacteraemia between the two groups were analyzed by use of a two sided chi-square test. The Wilcoxin rank sum test was used to compare the groups concerning magnitude (cfu) of microorganisms isolated. Results Incidence, type and magnitude of bacteraemia All 8 blood samples obtained before dental extraction showed no growth. Anaerobic bacteria dominated the findings of postextraction bacteraemia and Gram-positive strains greatly outnumbered Gram-negative bacteria. Aerobic bacteria other than viridans streptococci were recovered infrequently (Table I). The incidence of bacteraemia found during dental extraction was 79% in the erythromycin group and 8% in the clindamycin group. Ten minutes after extraction the incidence had decreased to 58% and 5% respectively. Table II shows the incidence, type and magnitude of bacteraemia in relation to prophylaxis and blood sampling time. Downloaded from by guest on 0 October 08

5 Bacteraemia after dental extraction 787 Table I. Different species of microorganisms isolated from blood after a single dental extraction in relation to prophylaxis, number of patients, total number of strains" and total number of cfu' in each patient group Species isolated Aerobes Corynebacterium xerosis E. faecalis Micrococcus spp. S. epidermidis Total number Viridans streptococci S. intermedius S. mitis Streptococcus mulans S. sanguis Streptococcus salivarius Total number Anaerobes Peptostreptococcus spp. V. parvula Actinomyces israelii Actinomyces meyeri Actinomyces naeshwdii Actinomyces odontolyticus Actinomyces viscosus Actinomyces spp. Bifidobacterium dentium Bifidobacterium spp. Eubacterium aerofaciens Eubacterium contortion Eubacterium lentum Eubacterium tenue Eubacterium spp. Lactobacillus acidophilus Lactobacillus casei Lactobacillus delbrueckii Lactobacillus fermentum Lactobacillus plant arum Lactobacillus rogosae Propionibacterium acnes Bacteroides distasonis Bacteroides ureolyticus Bacteroides sp. Total number Erythromycin group no. patient strains cfu ' Clindamycin group no. patient strains cfu Downloaded from by guest on 0 October 08 "Strain; means all isolates within the same species, characterized by the same morphology, biochemical reactions and antimicrobial susceptibility. *cfu, estimated from number of colonies on the filters in each patient group. 'Empty entries indicate no isolate recovered.

6 Table II. Results of blood cultures after a single dental extraction in relation to antibiotic prophylaxis and sampling time. Per cent of patients with bacteraemia, its type and magnitude Prophylactic agent (no. of patients) Erythromycin, g (9) Clindamycin, 0. g (9) Bacteraemia during surgery % of patients 79 (.05') 8 (0.7) 'Median cfu per millilitre in positive samples. Bacteraemia lomin after surgery % of patients 58 (0.0) 5 (0.0) Viridans streptococci isolated during surgery % of patients 79 (0.) (0.) Viridans Anaerobic Anaerobic miauwu streptococci bacteria bacteria isolated 0 min isolated during isolated 0 min after surgery surgery after surgery % of patients % of patients % of patients (0.) 7 (0.) 58 (.5) 8 (0.) 5 (0.) (0.) s a. Downloaded from by guest on 0 October 08

7 Bacteraemia after dental extraction 789 No statistically significant difference was noted in incidence or magnitude of total bacteraemia, bacteraemia with viridans streptococci or bacteraemia with anaerobic bacteria between the two groups, at any sampling time. Bacteraemia with viridans streptococci The overall incidence of viridans streptococcal bacteraemia was 79% in the erythromycin group and 7% in the clindamycin group (two patients in the clindamycin group developed bacteraemia lomin after extraction only). Strains of Streptococcus intermedius were most frequently isolated followed by strains of Streptococcus mitis and Streptococcus sanguis in both prophylaxis groups. Bacteraemia with anaerobic bacteria The incidence of anaerobic bacteraemia during dental extraction was 58% in the erythromycin group and 7% in the clindamycin group. Species of Actinomyces, Eubacterium and Lactobacillus were most commonly recovered while Propionibacterium acnes and Bacteroides species were isolated from single patients only. The total number of cfu of anaerobic species recovered in the erythromycin group was about four times that isolated in the clindamycin group. However, two patients showed extremely high counts which represented 85% of the total of anaerobic bacteria isolated from the erythromycin group. This fact was also reflected in high recovery rate of specific species such as Veillonella parvula, Bifidobacterium spp. and Lactobacillus rogosae (Table I). Serum concentrations of erythromycin and clindamycin The drugs were administered 90 min prior to the planned extraction, but due to clinical circumstances, time of blood sampling varied from 85 to 0 min and the mean time of sampling was 95 min in both groups. At dental extraction the mean serum concentration (±S.D.) of erythromycin was. ±.7 mg/l (range mg/l) and of clindamycin 5.5 ±. mg/l (range.7-9. mg/l). Susceptibility of isolates in relation to the individual antibiotic serum concentration during dental extraction In the erythromycin group all strains, except isolates of Veillonella parvula, had MICs below the individual serum concentrations. One isolate of Staphylococcus epidermidis was the only strain with an MBC exceeding the individual serum concentration of erythromycin. In the clindamycin group, two Enterococcus faecalis strains and one S. sanguis were recovered with MICs or MBCs above the respective individual serum concentrations. Downloaded from by guest on 0 October 08 Susceptibility of isolates to erythromycin, clindamycin, penicillin V, and ampicillin The MICs determined by the agar dilution method for the 79 viridans streptococci were within one two-fold dilution step of the MICs obtained by the macrobroth dilution method. MBCs were determined for 8 strains under aerobic conditions.

8 790 G. Hall et al. All viridans streptococci (except one S. mitis strain) were classified as susceptible (MIC ^0.5 mg/l) to erythromycin and clindamycin (Figure). MBCs > 0.5 mg/l of erythromycin were observed in nine strains and of clindamycin in four strains (mostly S. sanguis and S. mitis, MBCs - mg/l). Penicillin V inhibited 78 of 79 strains of viridans streptococci at the breakpoint of ^0.5 mg/l while 5 isolates were classified as susceptible to ampicillin. No strains were resistant (^ mg/l) to penicillin V or ampicillin (Figure). MBCs of ampicillin were > mg/l for strains of S. sanguis and S. intermedius of which strains had MBCs of mg/l and the rest <mg/l. Ten of these also showed MBCs > mg/l of penicillin V (three strains required mg/l, and the rest ^ mg/l). The MBC/MIC ratio was ^ for all these isolates which were by definition tolerant. Aerobic species other than viridans streptococci were highly susceptible to the antibiotics tested with exception of the /?-lactamase-producing staphylococci (MICs ^ mg/l, MBCs ^8 mg/l) and the enterococci resistant to clindamycin (MICs : mg/l) with MBCs of 8 mg/l for most antibiotics tested. The susceptibility pattern of anaerobic strains to the antibiotics tested are shown in the Figure. Except for two strains of V. parvula none was resistant (>mg/l) to clindamycin, penicillin V or ampicillin. No breakpoint is recommended by NCCLS for anaerobic bacteria to erythromycin. Ten V. parvula strains had MICs exceeding mg/l of erythromycin. 00 Downloaded from by guest on 0 October 08 - n o N S 0 <n : d d o o o o o o o o mg/l Figure. Cumulative percentages of 79 viridans streptococci ( ) and anaerobic isolates ( ) inhibited by erythromycin (a), clindamycin (b), penicillin V (c) and ampicillin (d). Slashed lines indicate breakpoint antibiotic concentrations (NCCLS) for the category "susceptible" for viridans streptococci.

9 Bacteraemia after dental extraction 79 Discussion During the decades before the 990s, a high oral loading dose of erythromycin was recommended for penicillin-allergic patients by different national committees for prevention of infective endocarditis in patients at risk undergoing oral procedures. The basic aim was to produce serum concentrations of antimicrobials high enough to kill viridans streptococci released into the blood stream following an invasive procedure. The dosage has been a matter of discussion because erythromycin stearate causes side effects such as nausea and abdominal discomfort in more than 0% of the patients when a loading dose of.5 g is used (Shanson et al., 985). Considerable individual variation in the serum erythromycin stearate concentrations has also been reported (Meier, Liithy & Siegenthaler, 98) and for this reason an additional dose h postoperatively has been recommended. During the 990s, the international trend has changed from the recommendation of erythromycin to clindamycin for endocarditis prophylaxis in patients allergic to penicillins. The use of clindamycin is supported by pharmacokinetic studies (Meier et al., 98) and animal data (Glauser & Francioli, 98), and the drug has a low frequency of side effects when used as a single dose (Endocarditis Working Party of BSAC, 990). Shanson et al. (985) found that a.5 g oral loading dose of erythromycin stearate given h before dental extraction reduced the incidence of bacteraemia with viridans streptococci from % in control patients to 5%. Malinverni et al. (988) reported that erythromycin gluceptate suppressed postextraction bacteraemia in rats when samples were cultured on blood agar plates. However, when care was taken to eliminate erythromycin from the blood sample by a lysis-centrifugation process, the incidence and magnitude of bacteraemia was similar to that in control rats. In a clinical study by Sefton et al. (990) on the efficacy of erythromycin stearate and josamycin base in the prevention of bacteraemia following dental extraction, no reduction was noted compared with a placebo group. These results indicated that a substitute for erythromycin was required. In the present study we have standardized the surgical procedure to one single dental extraction. The lysis-filtration technique used eliminates some of the disadvantages occasionally associated with ordinary blood-culture systems. These may give a low recovery rate of some bacteria due to inhibition of one microorganism by other more rapidly growing isolates, or inhibition of growth due to phagocytic cells, antimicrobial substances and residues of antimicrobial drugs in blood (Washington, 987). Gram-positive bacteria dominated the postoperative bacteraemia and reflected the expected microflora at the site of surgery (Evaldson et al., 98). It should be noted that only three strains of Bacteroides spp. and no Porphyromonas, Prevotella or Fusobacterium species were isolated. Differences in magnitude of specific species was observed between the two prophylaxis groups which was due mainly to the high recovery rate from individual patients. No preoperative bacteraemia was noted in the 8 patients despite evidence of chronic dentoalveolar pathology in most cases. Endocarditis caused by anaerobic bacteria remains uncommon although cases due to Bacteroides spp., Fusobacterium spp., Propionibacterium acnes (Felner & Dowell, 970) and Lactobacillus spp. (Sussman et al., 98) have been reported. The viridans group Downloaded from by guest on 0 October 08

10 79 G. HaU et al. of streptococci accounts for almost half the cases of endocarditis with S. sanguis and S. mitis predominating (Young, 987). In the present study, these species were recovered from more than 70% of all patients with viridans streptococcal bacteraemia. Staphylococci are second to streptococci as a cause of endocarditis, have a predilection for the tricuspid valve in drug addicts and are often involved in prosthetic valve endocarditis (Braimbridge & Eykyn, 987). Despite the oral cavity not being the normal habitat for these organisms in immunocompetent patients, some studies indicate high frequencies in post extraction bacteraemia (Roberts et al., 987). In the present study S. epidermidis was recovered only in low numbers possibly due to precautions taken to minimize contamination during blood sampling and laboratory processing. Other aerobic bacteria are rare findings in endocarditis but a few cases due to Actinobacillus spp., Haemophilus spp., Enterococcus spp., and Corynebacterium spp. have been reported (Bayliss et al., 98). The two last-mentioned species were recovered only occasionally in the present study. The overall incidence of bacteraemia with viridans streptococci due to extraction was 79% in the erythromycin group and 7% in the clindamycin group. This corresponds closely to the results of our previous report where penicillin V and amoxycillin were used for prophylaxis of postextraction bacteraemia (Hall et al., 99). The high incidences of bacteraemia observed in the present study cannot be explained by bacterial resistance to the drugs since more than 90% of all strains in each group had MICs and MBCs below the individual antibiotic serum concentration during dental extraction. Since bacteraemia is rapidly cleared by the reticulo-endothelial system (Silver, Martin & McBride, 975), time of exposure of the microorganisms to the antimicrobial agent in blood must be too short to be effective. Results from in-vitro time-killing studies of viridans streptococci support this statement (Glauser et al., 98; Lowy et al., 98). Experimental animal studies indicate that inhibition of bacterial growth after attachment to damaged heart valves may be the most important mechanism for successful prophylaxis against endocarditis using penicillin (Moreillon et al., 98; Berney & Francioli, 990). In-vitro antibiotic breakpoint concentrations are set to guide the clinician to chose the most appropriate therapeutic or prophylactic antibiotic (Report of Working Party of BSAC, 988). Applying NCCLS breakpoints, 99% of tested viridans strains were susceptible to erythromycin, clindamycin and penicillin V, while 8% of the isolates were susceptible to ampicillin and remaining strains were classified as intermediate. Varying frequencies of resistance of oral streptococci to antimicrobial agents have been reported. However, it is difficult to compare the results obtained in this investigation with other reports because of differences in methodology and criteria for defining antibiotic susceptibility. Harrison et al. (985) found no erythromycin-resistant oral streptococci at a breakpoint of 0.5 mg/l which is in agreement with our results. Longman et al. (99) isolated streptococci from the oral cavity with MICs >.5 mg/l of erythromycin in about 5% of patients. Our study support the results of Williams et al. (99) who observed no resistance at 0.5 mg/l of clindamycin in 78 viridans streptococci tested, while other authors have reported occasional isolates with high MICs (Phillips et al., 97; Wilcox et al., 99). Bacterial killing has been assumed to be the mechanism of action of prophylactic antimicrobial drugs (Durack & Petersdorf, 97). Nevertheless, experimental animal studies have shown that successful prophylaxis against endocarditis can be achieved by Downloaded from by guest on 0 October 08

11 Bacteraemia after dental extraction 79 the use of bacteriostatic antibiotics (Glauser & Francioli, 98; Malinverni et al., 988) and by use of sublethal concentrations of penicillin (Francioli & Glauser, 985). This study showed that the bacteraemia after dental extraction did not differ significantly when prophylaxis with erythromycin and clindamycin were compared. The great majority of bacterial species isolated from blood samples after dental extraction from patients on prophylaxis with erythromycin or clindamycin were highly susceptible to these agents as well as to penicillin V and ampicillin. If a protective effect against endocarditis is exerted by administration of erythromycin or clindamycin, the antimicrobials must interfere with bacteria during crucial steps in the development of endocarditis other than during the initial phase of transient bacteraemia. Acknowledgements This work was supported by the Swedish Medical Research Council and the Swedish National Association against Heart and Chest Diseases. References Baltch, A. L., Pressman, H. L., Hammer, M. C, Sutphen, N. C, Smith, R. P. & Shayegani, M. (98). Bacteremia following dental extractions in patients with and without penicillin prophylaxis. American Journal of the Medical Sciences 8, 9 0. Bayliss, R., Clarke, C, Oakley, C. M., Sommerville, W., Whitfield, A. G. W. & Young, S. E. (98). The microbiology and pathogenesis of infective endocarditis. British Heart Journal 50, 5-9. Berney, P. & Francioli, P. (990). Successful prophylaxis of experimental streptococcal endocarditis with single-dose amoxicillin administered after bacterial challenge. Journal of Infectious Diseases, 8-5. Braimbridge, M. V. & Eykyn, S. J. (987). Prosthetic valve endocarditis. Journal of Antimicrobial Chemotherapy 0, Suppl. A, Dajani, A. S., Bisno, A. L., Chung, K. J., Durack, D. T., Freed, M., Gerber, M. A. et al. (990). Prevention of bacterial endocarditis. Journal of the American Medical Association, 99-. Durack, D. T. & Petersdorf, R. G. (97). Chemotherapy of experimental streptococcal endocarditis: I. Comparison of commonly recommended prophylactic regimens. Journal of Clinical Investigation 5, Endocarditis Working Party of the British Society for Antimicrobial Chemotherapy. (990). Antibiotic prophylaxis of infective endocarditis. Lancet i Evaldson, G., Heimdahl, A., Kager, L. & Nord, C. E. (98). The normal human anaerobic microflora. Scandinavian Journal of Infectious Diseases, Suppl. 5, 9-5. Felner, J. M. & Dowell, V. R. (970). Anaerobic bacterial endocarditis. New England Journal of Medicine 8, Francioli, P. & Glauser, M. P. (985). Successful prophylaxis of experimental streptococcal endocarditis with single doses of sublethal concentrations of penicillin. Journal of Antimicrobial Chemotherapy 5, Suppl. A, Freedman, L. R. (987). The pathogenesis of infective endocarditis. Journal of Antimicrobial Chemotherapy 0, Suppl. ^,. Glauser, M. P., Bernard, J. P., Moreillon, P. & Francioli, P. (98). Successful single-dose amoxicillin prophylaxis against experimental streptococcal endocarditis: evidence for two mechanisms of protection. Journal of Infectious Diseases 7, Glauser, M. P. & Francioli, P. (98). Successful prophylaxis against experimental streptococcal endocarditis with bacteriostatic antibiotics. Journal of Infectious Diseases, Hall, G., Hedstrom, S. A., Heimdahl, A. & Nord, C. E. (99). Prophylactic administration of penicillins for endocarditis does not reduce the incidence of postextraction bacteremia. Clinical Infectious Diseases 7, Downloaded from by guest on 0 October 08

12 79 G. Hall et al. Harrison, G. A., Stross, W. P., Rubin, M. P., Davies, R. M. & Speller, D. C. (985). Resistance in oral streptococci after repeated three-dose erythromycin prophylaxis. Journal of Antimicrobial Chemotherapy 5, 7-9. Head, T. W, Bentley, K. C, Millar, E. P. & devries, J. A. (98). A comparative study of the effectiveness of metronidazole and penicillin V in eliminating anaerobes from postextraction bacteremias. Oral Surgery, Oral Medicine, Oral Pathology 58, 5-5. Heimdahl, A., Hall, G., Hedberg, M., Sandberg, H., Soder, P.-O., Tuner, K. et al. (990). Detection and quantitation by lysis-filtration of bacteremia after different oral surgical procedures. Journal of Clinical Microbiology 8, Heimdahl, A., Josefsson, K., von Konow, L. & Nord, C. E. (985). Detection of anaerobic bacteria in blood cultures by lysis filtration. European Journal of Clinical Microbiology, 0-7. Jailing, B., Malmborg, A. S., Lindman, A. & Boreus, L. O. (97). Evaluation of a micromethod for determination of antibiotic concentrations in plasma. European Journal of Clinical Pharmacology, Lennette, E. H., Balows, A., Hausler, W. J. & Shadomy, H. J., Eds. (985). Manual of Clinical Microbiology, th edn. American Society for Microbiology. Washington, DC. Longman, L. P., Pearce, P. K., McGowan, P., Hardy, P. & Martin, M. V. (99). Antibiotic-resistant oral streptococci in dental patients susceptible to infective endocarditis. Journal of Medical Microbiology, -7. Lowy, F. D., Neuhaus, E. G., Chang, D. S. & Steigbigel, N. H. (98). Penicillin therapy of experimental endocarditis induced by tolerant Streptococcus sanguis and nontolerant Streptococcus mitis. Antimicrobial Agents and Chemotherapy, 7-7. Malinverni, R., Overholser, C. D., Bille, J. & Glauser, M. P. (988). Antibiotic prophylaxis of experimental endocarditis after dental extractions. Circulation 77, 8-7. Meier, B., Luthy, R. & Siegenthaler, W. (98). Endokarditis-Prophylaxe mit Amoxycillin, Clindamycin oder Erythromycin Eine pharmakokinetische Betrachung. Schweizerische Medizinische Wochenschrift, 5-. Moreillon, P., Francioli, P., Overholser, D., Meylan, P. & Glauser, M. P. (98). Mechanisms of successful amoxicillin prophylaxis of experimental endocarditis due to Streptococcus intermedius. Journal of Infectious Diseases 5, National Committee for Clinical Laboratory Standards. (990a). Methods for Antimicrobial Susceptibility Testing of Anaerobic Bacteria-Second Edition: Approved Standard M-A. NCCLS, Villanova, PA. National Committee for Clinical Laboratory Standards. (990ft). Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria that Grow Aerobically Second Edition: Approved Standard M7-A. NCCLS, Villanova, PA. National Committee for Clinical Laboratory Standards. (99a). Methods for Determining Bactericidal Activity of Antimicrobial Agents: Tentative Guideline M-T. NCCLS, Villanova, PA. National Committee for Clinical Laboratory Standards. (99A). Performance Standards for Antimicrobial Susceptibility Testing; Fourth Informational Supplement: MI0O-S. NCCLS, Villanova, PA. Olsson, B., Dornbusch, K. & Nord, C. E. (977). Susceptibility to beta-lactam antibiotics and production of beta-lactamases in Bacteroides fragilis. Medical Microbiology and Immunology, 8-9. Phillips, I., Warren, C, Harrison, J. M., Sharpies, P., Ball, L. C. & Parker, M. T. (97). Antibiotic susceptibilities of streptococci from the mouth and blood of patients treated with penicillin or lincomycin and clindamycin. Journal of Medical Microbiology 9, 9-0. Report by a Working Party of the British Society for Antimicrobial Chemotherapy. (988). Breakpoints in in-vitro antibiotic sensitivity testing. Journal of Antimicrobial Chemotherapy, Roberts, G. J., Radford, P. & Holt, R. (987). Prophylaxis of dental bacteraemia with oral amoxycillin in children. British Dental Journal, Sefton, A. M., Maskell, J. P., Kerawala, C, Cannell, H., Seymour, A., Sun, Z.-M. et al. (990). Comparative efficacy and tolerance of erythromycin and josamycin in the prevention of Downloaded from by guest on 0 October 08

13 Bacteraemia after dental extraction 795 bacteraemia following dental extraction. Journal of Antimicrobial Chemotherapy 5, Shanson, D. C, Akash, S., Harris, M. & Tadayon, M. (985). Erythromycin stearate,.5 g, for the oral prophylaxis of streptococcal bacteraemia in patients undergoing dental extraction: efficacy and tolerance. Journal of Antimicrobial Chemotherapy 5, Shanson, D. C, Cannon, P. & Wilks, M. (978). Amoxycillin compared with penicillin V for prophylaxis of dental bacteraemia. Journal of Antimicrobial Chemotherapy, -. Silver, J. G., Martin, L. & McBride, B. C. (975). Recovery and clearance rates of oral microorganisms following experimental bacteraemias in dogs. Archives of Oral Biology 0, Simmons, N. A. and the Endocarditis Working Party of the British Society for Antimicrobial Chemotherapy. (99). Recommendations for endocarditis prophylaxis. Journal of Antimicrobial Chemotherapy, 7-8. Sussman, J. I., Baron, E. J., Goldberg, S. M., Kaplan, M. H. & Pizzarello, R. A. (98). Clinical manifestations and therapy of lactobacillus endocarditis: report of a case and review of the literature. Reviews of Infectious Diseases 8, 77-. Washington, J. A. (987). The microbiological diagnosis of infective endocarditis. Journal of Antimicrobial Chemotherapy 0, Suppl. A, 9-. Wilcox, M. H., Winstanley, T. G., Spencer, R. C. & Douglas, C. W. (99). Susceptibility of endocarditis streptococci to clindamycin. Lancet 0, 0. Williams, J. D., Maskell, J. P., Shain, H., Chrysos, G., Sefton, A. M., Fraser, H. Y. et al. (99). Comparative in-vitro activity of azithromycin, macrolides (erythromycin, clarithromycin and spiramycin) and streptogramin RP against oral organisms. Journal of Antimicrobial Chemotherapy 0, 7-7. Young, S. E. J. (987). Aetiology and epidemiology of infective endocarditis in England and Wales. Journal of Antimicrobial Chemotherapy 0, Suppl. A, 7-5. {Received 7 June 995; accepted November 995) Downloaded from by guest on 0 October 08