| Sign In to gain access to subscriptions and/or personal tools. |
Efficacy of Antibiotic Prophylactic Regimens for the Prevention of Bacterial Endocarditis of Oral Origin
1 Special Needs Unit, School of Medicine and Dentistry, Santiago de Compostela University (Spain); and Correspondence: * corresponding author, c.scully{at}eastman.ucl.ac.uk
Despite the controversy about the risk of individuals developing bacterial endocarditis of oral origin, numerous Expert Committees in different countries continue to publish prophylactic regimens for the prevention of bacterial endocarditis secondary to dental procedures. In this paper, we analyze the efficacy of antibiotic prophylaxis in the prevention of bacteremia following dental manipulations and in the prevention of bacterial endocarditis (in both animal models and human studies). Antibiotic prophylaxis guidelines remain consensus-based, and there is scientific evidence of the efficacy of amoxicillin in the prevention of bacteremia following dental procedures, although the results reported do not confirm the efficacy of other recommended antibiotics. The majority of studies on experimental models of bacterial endocarditis have verified the efficacy of antibiotics administered after the induction of bacteremia, confirming the efficacy of antibiotic prophylaxis in later stages in the development of bacterial endocarditis. There is no scientific evidence that prophylaxis with penicillin is effective in reducing bacterial endocarditis secondary to dental procedures in patients considered to be "at risk". It has been suggested that there is a high risk of severe allergic reactions secondary to prophylactically administered penicillins, but, in reality, very few cases have been reported in the literature. It has been demonstrated that antibiotic prophylaxis could contribute to the development of bacterial resistance, but only after the administration of several consecutive doses. Future research on bacterial endocarditis prophylactic protocols should involve the re-evaluation of the time and route of administration of antibiotic prophylaxis, and a search for alternative antimicrobials.
Key Words: endocarditis • bacteremia • dentistry • antibiotic prophylaxis
Bacterial endocarditis is an uncommon infection of the cardiac endothelium, with an incidence ranging between one and five cases/100,000 population/yr (Bouza et al., 2001; Cecchi et al., 2004). Despite advances in diagnosis, antimicrobial therapy, and treatment of complications, bacterial endocarditis continues to be responsible for substantial morbidity and mortality (Heiro et al., 2006; Tran and Kjeldsen, 2006; Smith et al., 2007). The main complications deriving from bacterial endocarditis are valvular destruction, heart failure, and emboli, requiring cardiac surgical treatment within 3 mos for almost 30% of patients (Heiro et al., 2006). In recent cohort series, some authors found a prevalence of bacterial endocarditis mortality of 12–16% (Heiro et al., 2006; Tran and Kjeldsen, 2006; Smith et al., 2007), and suggested, as predictive factors: age greater than 65 yrs, longer ’pump time’, unstable hemodynamic status (Smith et al., 2007), prosthetic valve endocarditis, and Staphylococcus aureus infection (Tran and Kjeldsen, 2006). In the majority of retrospective studies published during the 1990s, the oral cavity was identified as the portal of entry of the causative microbial agent in 14–20% of patients with bacterial endocarditis (Manford et al., 1992; Sandre and Shafran, 1996; Hricak et al., 1998; Sekido et al., 1999), and the prevalence of bacterial endocarditis of possible oral origin in recently published series was similar to that reported in older series (Krcmery et al., 2003; Loupa et al., 2004; Knirsch et al., 2005). In 1995, for the first time, the oral origin of two cases of bacterial endocarditis was demonstrated by molecular biological techniques, since a complete concurrence of identity was observed between the blood and oral cavity isolates (Fiehn et al., 1995). Recently, Streptococcus mutans was isolated and characterized by the application of broad-range polymerase chain-reaction with DNA sequencing in heart valve and dental plaque specimens from a patient with bacterial endocarditis (Nomura et al., 2006).
• Causal Organisms Broad-range polymerase chain-reaction is used to target commonly shared bacterial 16S rRNA genes (via pan-bacterial primers). Subsequently, direct sequencing is used to detect and differentiate bacteria. This molecular method improves the microbiological diagnosis of bacterial endocarditis, not only for fastidious, slow-growing, and/or non-culturable bacteria, but also for easy-to-culture pathogens (Breitkopf et al., 2005). The results of using the polymerase chain-reaction technique to identify bacterial DNA in excised valvular tissue of patients with bacterial endocarditis confirmed that streptococci were the predominant bacteria, followed by staphylococci (Lang et al., 2004). It has even been suggested that this molecular approach should be included as an additional major criterion in the Duke’s classification scheme (Millar et al., 2001). However, the costs and quality of databases are major obstacles to implementation of 16S rRNA gene sequencing in the routine clinical microbiology laboratory (Boudewijns et al., 2006). Other disadvantages of DNA-based methods are that they do not distinguish between living and dead organisms, results are sometimes difficult to interpret, and DNA may persist without any evidence of infection (Rovery et al., 2005). In consequence, currently, the polymerase chain-reaction method is especially useful when the causative agent of bacterial endocarditis is fastidious, or when the specimen is taken during antimicrobial treatment (Kotilainen et al., 2006). Broad-range polymerase chain-reaction has allowed for the detection of oral pathogens such as Streptococcus mutans and Actinobacillus actinomycetemcomitans in about 70% and 31% of cardiovascular tissue samples, respectively (Nakano et al., 2006, 2007). So far, in contrast, the results of polymerase chain-reaction-based screening to detect bacteremia following dental procedures did not show a good correlation with those obtained by means of standard blood cultures (Kinane et al., 2005; Savarrio et al., 2005).
• Pathogenesis
The normal cardiac valve endothelium is resistant to colonization by bacteria and does not activate platelet aggregation (Rodgers et al., 1983). Endothelium injury mainly results from turbulent blood flow caused by pre-existing cardiac lesions, such as a defective valve or a congenital defect (high-velocity jet stream from a high- to low-pressure chamber, or created by a pressure gradient across a narrowed orifice between 2 chambers, or a high-velocity jet), but in some cases damage follows direct endocardium trauma (cardiac surgery or catheterism), or may be associated with malignancy and connective tissue disorders (Rodbard, 1963; Lopez et al., 1987; Maté del Tio et al., 1997). Damage to the endothelium triggers platelet and fibrinogen-fibrin deposition at the site. This sterile platelet-fibrin aggregate is called ’non-bacterial thrombotic endocarditis’ (Rodgers et al., 1983), and is more receptive to colonization by bacteria than is the intact endothelium. Transient bacteremia is a common event and occurs as a consequence of skin or mucosal injuries, which are normally laden with an endogenous flora. During transient bacteremias, microorganisms can adhere to this non-bacterial thrombotic endocarditis, which provides them with both nutrition and protection from host defenses, especially leukocytes. Bacteria that are typically found in patients with bacterial endocarditis, such as Streptococcus viridans, Staphylococcus aureus, or Enterococcus spp., adhere more avidly to endothelium than do species that seldom cause bacterial endocarditis. The mechanism of this adherence is poorly understood. Bacterial exopolysaccharides, such as dextrans produced by Streptococcus sanguis and mutans, act as adhesins that bind to injured valves (Scheld et al., 1978). Although the majority of organisms present in the bloodstream after dental manipulations do not produce dextran, over 50% of the endocarditis cases are caused by dextran-producing oral streptococci (Hehre and Neill, 1946). Microbial surface components recognizing adhesive matrix molecules, such as clumping factors A and B, fibronectin-binding proteins, and FimA factor, have also been implicated in Gram-positive bacteria adherence (Burnette-Curley et al., 1995; Schwarz-Linek et al., 2004; Rindi et al., 2006). Other compounds that may mediate adherence include fibrinogen, laminin, and type IV collagen. Bacterial-platelet interactions appear to be important in both the induction of a vegetation and its enlargement after colonization. Interactions between streptococci and platelets may involve 3 sites (Herzberg, 1996):
The host defenses include leukocytes, antibodies and complement, and platelet-derived bactericidal proteins (thrombocidins). Paradoxically, the host also responds to invasion by activation of the clotting system by monocyte tissue thromboplastin and release of pro-inflammatory cytokines (IL-1, IL-6, IL-8), which increases fibrin deposition and promotes growth of the vegetation (Thompson et al., 1982; Dankert et al., 2006). If the bacteria trapped within the thrombus is resistant to thrombocidins, large numbers of tightly packed microcolonies grow (Marrie et al., 1987). The resultant vegetation is composed of successive layers of fibrin and clusters of bacteria, with rare red cells and leukocytes, almost always covered by a layer of fibrin on the luminal surface. The pathophysiology of bacterial endocarditis includes: (1) local destructive effects of infection, leading to incompetent valvular function, congestive heart failure, or atrioventricular conduction abnormalities; (2) embolization of fragments of the vegetation, resulting in infection or infarction of the surrounding tissue (spleen, kidney, meninges, brain, bone, pericardium, synovium, or vitreous humor); (3) the hematogenous seeding of remote sites during continuous bacteremia (septic arthritis, osteomyelitis, splenic or kidney abscesses, meningeal or brain abscesses); and (4) host response, including release of pro-inflammatory cytokines that provoke "constitutional symptoms" (fever, fatigue, anorexia, weight loss, night sweats) and deposition of pre-formed immune complexes (glomerulonephritis, Osler’s nodes, rheumatological manifestations). As we have described, in the classic model of the development of bacterial endocarditis of oral origin, a sterile vegetation is invaded by oral microorganisms as a consequence of bacteremia. However, it has been suggested that a bacteremia of oral origin, instead of directly inducing the onset of bacterial endocarditis, could favor the initial thickening of the cardiac valves from atherosclerosis, making them more susceptible to bacterial adherence and subsequent colonization (Drangsholt, 1998). In this new pathogenic model, several initial episodes of bacteremia would affect the cardiac valve endothelium over a long period, until a later bacteremia of days’ or weeks’ duration caused bacterial adherence and colonization and established cardiac infection. Consequently, this model suggests that bacterial endocarditis of oral origin is a chronic disease with a long latency period and a series of well-defined stages, but there is little evidence for this model, and there are few studies on experimental animals of the long-term effects of such bacteremias of oral origin on cardiac endothelium (Brodala et al., 2005).
In 1955, the American Heart Association published the first protocol for the prevention of bacterial endocarditis associated with dental procedures (American Heart Association, 1955). Since then, many Expert Committees in different countries have drawn up different prophylactic regimens. In 1995, the European Society of Cardiology, together with a group of experts from the International Society of Chemotherapy, published a European Consensus on bacterial endocarditis prophylaxis (Leport et al., 1995). In 2004, both the European Society of Cardiology and the British Cardiac Society, together with the Royal College of Physicians of London, drew up guidelines on the prevention of bacterial endocarditis associated with dental procedures (Horstkotte et al., 2004; Dental aspects of endocarditis prophylaxis, 2004). The British Society of Antimicrobial Chemotherapy published its first antibiotic prophylaxis regimens for bacterial endocarditis in 1982 (Antibiotic prophylaxis of infective endocarditis, 1982); these were revised and modified in 1986 (Simmons et al., 1986), 1990 (Antibiotic prophylaxis of infective endocarditis, 1990), 1992 (Simmons et al., 1992), and 2006 (Gould et al., 2006). The American Heart Association has published nine bacterial endocarditis prophylaxis protocols (American Heart Association, 1955, 1960, 1972; Hussar, 1965; Kaplan, 1977; Shulman et al., 1984; Dajani et al., 1990, 1997; Wilson et al., 2007), the latest revision being in 2007 (Wilson et al., 2007). The most recent bacterial endocarditis prophylaxis protocols published by the European Society of Cardiology, British Cardiac Society-Royal College of Physicians, British Society of Antimicrobial Chemotherapy, and American Heart Association are summarized in Tables 1  and 2.
Despite the frequent updating of these antibiotic prophylaxis guidelines, the incidence of bacterial endocarditis of possible oral origin has evidently not decreased. Various reasons have been suggested to explain this situation: an increase in the number of patients susceptible to bacterial endocarditis and in the practice of "at risk" procedures; the lack of compliance with the prophylactic regimens for the prevention of bacterial endocarditis by doctors, dentists, and patients; and the lack of efficacy of prophylactic protocols (Delahaye and De Gevigney, 2001).
• American Heart Association Guidelines, 2007 — In the protocols on the prevention of bacterial endocarditis associated with dental procedures, published by the American Heart Association in the 1960s (American Heart Association, 1960; Hussar, 1965), the profile of the individual considered to be "at risk" of bacterial endocarditis was a patient with rheumatic heart disease or a congenital heart disease. In the first guidelines published by the British Society of Antimicrobial Chemotherapy in 1982 (Antibiotic prophylaxis of infective endocarditis, 1982), patients considered to be "at risk" of bacterial endocarditis included those with disturbances of the endocardium due to congenital or acquired disease, those with heart valve disease, and those with prosthetic heart valves. Since then, several Expert Committees have defined the heart conditions catalogued as "at risk" of bacterial endocarditis, also discussing the controversy concerning the administration of antibiotic prophylaxis in cases of mitral stenosis without valve incompetence (Leport et al., 1995). In recent years, most updated guidelines have restricted prophylaxis to high-risk patients (Gould et al., 2006). Following the last American Heart Association guidelines, bacterial endocarditis prophylaxis for dental procedures should be recommended only for patients with underlying cardiac conditions associated with the highest risk of adverse outcome from bacterial endocarditis. These cardiac conditions are: prosthetic cardiac valve, previous bacterial endocarditis, congenital heart disease (unrepaired defect, completely repaired defect during the first 6 mos after the procedure, and repaired defect with residual alterations), and cardiac transplantation recipients who develop cardiac valvulopathy (Wilson et al., 2007). — In 1960, the American Heart Association stated that the dental procedures in which prophylaxis was indicated were tooth extractions and gingival treatments, specifying that these procedures frequently caused "transitory bacteremias", and that the bacteremias were of a higher intensity in patients with oral infections (American Heart Association, 1960). Some years later, the American Heart Association recognized the impossibility of predicting which dental procedures could be responsible for causing bacterial endocarditis. The treatments associated with gingival bleeding—in which antibiotic prophylaxis was recommended—included scaling, and among the procedures in which prophylaxis was not indicated were included the adjustment of orthodontic appliances and the exfoliation of primary teeth (Kaplan, 1977). The British Society of Antimicrobial Chemotherapy has recently summarized the indications for antibiotic prophylaxis for high-risk patients, stating that it should be given for "all dental procedures involving dento-gingival manipulation or endodontics" (Gould et al., 2006). Following the most recent American Heart Association guidelines, prophylaxis is recommended for all dental procedures that involve manipulation of gingival tissue or the periapical region of teeth, or perforation of the oral mucosa. This includes procedures such as biopsies, suture removal, and placement of orthodontic bands, but it does not include routine anesthetic injections through non-infected tissue, the taking of dental radiographs, placement of removable prosthodontic or orthodontic appliances, placement of orthodontic brackets, or adjustment of orthodontic appliances. There are other events for which prophylaxis is not recommended, such as shedding of deciduous teeth and trauma to the lips or oral mucosa (Wilson et al., 2007). — In 1960, the American Heart Association pronounced itself in favor of administering an antibiotic prophylactic regimen that consisted of several injections of penicillin from 2 days before up to 2 days after the session of dental treatment (American Heart Association, 1960). The British Society of Antimicrobial Chemotherapy, in its first guidelines, suggested a single prophylactic regimen, for all patients considered to be "at risk" of bacterial endocarditis, of a single dose of amoxicillin before the dental procedure (Antibiotic prophylaxis of infective endocarditis, 1982). The British Society of Antimicrobial Chemotherapy substituted penicillin V—previously recommended by the American Heart Association (Kaplan, 1977)—for amoxicillin, due to its more favorable pharmacokinetic and pharmacodynamic characteristics (Antibiotic prophylaxis of infective endocarditis, 1982). The American Heart Association was the first to recommend erythromycin in patients with a history of allergy to penicillin (Kaplan, 1977), but in 1992, the British Society of Antimicrobial Chemotherapy definitively replaced erythromycin with clindamycin in patients allergic to penicillin (Simmons et al., 1992). In 1995, the European Society of Cardiology performed a critical review of the prophylaxis protocols drawn up by the different national committees, noting clear differences between countries, although all included a simple or standard regimen and another more complex regimen for use in special circumstances. In general, the standard guidelines consisted of the oral administration of a single dose of antibiotic which, in the majority of countries, was amoxicillin, and clindamycin was the antibiotic of choice in patients allergic to the betalactams (Leport et al., 1995). The prophylactic protocol currently updated by the American Heart Association continues recommending amoxicillin as the antibiotic of choice for oral prophylaxis. For individuals who are allergic to penicillins or amoxicillin, the use of cefalexin or another first-generation oral cephalosporin, clindamycin, azithromycin, or clarithromycin is recommended. Because of possible cross-reactions, a cephalosporin should not be administered to patients with a history of anaphylaxis, angioedema, or urticaria after treatment with any form of penicillin, including ampicillin or amoxicillin. Patients who are unable to tolerate an oral antibiotic may be treated with ampicillin, ceftriaxone, or cefazolin, administered intramuscularly or intravenously. For ampicillin-allergic patients who are unable to tolerate an oral agent, prophylaxis is recommended with parenteral cefazolin, ceftriaxone, or clindamycin (Wilson et al., 2007). It has been estimated that the number of cases of bacterial endocarditis that result from a dental procedure is very small. In consequence, the American Heart Association concluded that only an extremely small number of cases of bacterial endocarditis might be prevented by antibiotic prophylaxis for dental procedures, even if such prophylactic regimens were 100% effective. Finally, this Expert Committee stated the need for prospective placebo-controlled studies of antibiotic prophylaxis of bacterial endocarditis, to evaluate the efficacy of bacterial endocarditis prophylaxis (Wilson et al., 2007).
• Prevalence and Duration There are many published studies on the efficacy of antibiotic prophylaxis administered by oral route in the prevention of bacteremia secondary to dental procedures, but they have important differences with respect to the type and dose of antibiotic used, and the time of administration (Shanson et al., 1978, 1985; Josefsson et al., 1985; Roberts et al., 1987; Sefton et al., 1990; Göker and Güvener, 1992; Hall et al., 1993, 1996b; Aitken et al., 1995; Vergis et al., 2001; Lockhart et al., 2004; Diz Dios et al., 2006) (Table 3 ). Probably influenced by the idea that bacteremia secondary to dental procedures is of a transitory nature (Dajani et al., 1997), few studies have been published on the effect of antimicrobial prophylaxis on the duration of post-dental manipulation bacteremia.
Recent studies have confirmed the efficacy of amoxicillin in the prevention of bacteremia following dental manipulation. In a paper published in 2001, a reduction of almost 80% in the prevalence of post-extraction bacteremia after prophylaxis with 3 g of amoxicillin was reported (Vergis et al., 2001). In children, 50 mg/kg bodyweight of amoxicillin significantly reduced bacteremias secondary to nasal intubation (from 18% to 4%), restorative dental treatment and professional oral hygiene (from 20% to 6%), and tooth extractions (from 76% to 15%) (Lockhart et al., 2004). These authors also tested the effect of the prophylactic regimen on the duration of post-dental manipulation bacteremia: Forty-five minutes after the completion of treatment, the percentages of positive blood cultures were 14% in the "placebo group" vs. 0% in the "amoxicillin group" (Lockhart et al., 2004). In a trial published in 2006, the efficacies of different antibiotics in the prevention of bacteremia following dental extractions were evaluated. The prevalences of bacteremia in the "control group" and the "amoxicillin group" were 96% and 46%, respectively, at 30 sec after dental extractions, and 20% and 4%, respectively, 1 hr later (Diz Dios et al., 2006). In contrast, other authors did not find that prophylaxis with penicillin V or amoxicillin significantly reduced the prevalence or magnitude of post-extraction bacteremia (Hall et al., 1993). The percentages of positive blood cultures obtained in this series during surgery were 95% in the "placebo group", 90% in the "penicillin V group", and 85% in the "amoxicillin group". Furthermore, antibiotic prophylaxis did not affect the type of bacteremia: The predominant microorganisms in all three groups were Streptococcus intermedius, and the most frequent obligate anaerobic bacteria were Actinomyces spp., Peptostreptococcus spp., and Veillonella spp. These authors attributed their results to the use of the lysis-filtration technique, among other factors, since this provides a higher percentage of bacterial isolates with respect to other culture methods. However, it has been recently shown that the Bactec system is more sensitive than the lysis-filtration technique for detecting post-extraction bacteremia of very low intensity (Lucas et al., 2002a). In this trial (Hall et al., 1993), the percentage of positive post-extraction blood cultures at 10 min after completion of the surgery was 80% in the "placebo group" and 60% in the "amoxicillin group", although the authors concluded that amoxicillin did not significantly influence the duration of the post-extraction bacteremia. Calculating the ß-error and the statistical power of the results obtained in this study, we found that the probability of obtaining statistical significance, if it existed, was only 29% (when it should be greater than 80%), due to the sample size (20 persons in each group), which was so small that the investigators could not exclude the possibility that the lack of significance might be due to chance (Calatayud and Martín, 2003). With respect to orally administered antibiotics with bacteriostatic activity, 1 study published in 1985 compared the effect of a 500 mg dose of erythromycin with that of 2 g of penicillin V on the prevalence of bacteremia secondary to the removal of impacted or partially erupted mandibular third molars (Josefsson et al., 1985). There were no differences in the prevalence of bacteremia between the "prophylaxis groups", although the percentage of positive blood cultures 10 min after completion of the surgical manipulation was lower in the patients receiving antibiotics than in controls. One study supported that erythromycin stearate (1.5 g) significantly reduced the frequency of positive post-extraction blood cultures (from 43% to 15%) (Shanson et al., 1985). Other authors compared the effect of oral administration of 1.5 g of erythromycin with that of 1.5 g of josamycin (Sefton et al., 1990). In contrast to the findings in previous series, and based on the percentages of positive blood cultures detected (60% and 70% in the "erythromycin group" and "josamycin group", respectively, vs. 65% in the "placebo group"), these authors stated that none of these macrolides significantly affected the frequency of blood cultures due to Streptococcus spp. (Sefton et al., 1990). One study evaluated the prophylactic effect of 600 mg oral clindamycin vs. 1.5 g of erythromycin, and found clindamycin to be more active than erythromycin (Aitken et al., 1995). A similar study, however, did not find significant differences in the percentage of positive blood cultures or in the concentrations of the bacterial isolates during the extraction or 10 min after completion (Hall et al., 1996b). The prevalence of bacteremia during the extraction was 79% in the "erythromycin group" and 84% in the "clindamycin group" and, at 10 min, 58% and 53%, respectively. Although the number of bacteremias of streptococcal etiology was not affected by the type of antibiotic administered, the proportion of those produced by obligate anaerobic bacteria was reduced to half in the patients receiving clindamycin, compared with those on erythromycin (Hall et al., 1996b). To date, we have found only 2 studies where the results of a "clindamycin group" were compared with those obtained in a "control group" (Göker and Güvener, 1992; Diz Dios et al., 2006). One of these studies reported a 44% rate of bacteremia secondary to the removal of impacted mandibular third molars in a "control group" vs. 40% in the "clindamycin group" (Göker and Güvener, 1992). These authors also found no effect of clindamycin on the prevalence of post-extraction bacteremia 1 hr after completion of the manipulation, since this was 28% in the "placebo group" and 24% in the "clindamycin group" (Göker and Güvener, 1992). These findings could be attributed to the use of low doses of clindamycin (150 mg instead of the 600 mg recommended in the latest bacterial endocarditis prophylaxis protocols) (Horstkotte et al., 2004; Dental aspects of endocarditis prophylaxis, 2004; Gould et al., 2006; Wilson et al., 2007). All the blood cultures obtained at 24 hrs in the patients receiving antibiotic prophylaxis were sterile, although these findings are not a reflection of the efficacy of a single dose, since these authors administered the clindamycin every 6 hrs for 4 days (Göker and Güvener, 1992). In the other study, the prevalences of bacteremia in the "control group" and the "clindamycin group" were 96% and 85%, respectively, at 30 sec after dental extractions, and 20% and 22%, respectively, 1 hr later (Diz Dios et al., 2006).
Some authors have evaluated the efficacy of parenteral antibiotic prophylaxis (intramuscular or intravenous) on the reduction of the prevalence of bacteremia following dental manipulations (Elliott and Dunbar, 1968; Baltch et al., 1982a,b; Hess et al., 1983; Kaneko et al., 1995; Roberts and Holzel, 2002) (Table 4
One trial detected a 21% rate of positive post-extraction blood cultures in children with cardiopathies who had received prophylaxis with intramuscular penicillin (Hess et al., 1983). Other authors analyzed various intravenous prophylactic regimens in children with congenital heart defects undergoing dental treatment under general anesthetic (Roberts and Holzel, 2002). The most frequently applied protocols were ampicillin (mean dose, 627 mg), and teicoplanin in combination with amikacin (6 mg/kg bodyweight and 50 mg/kg bodyweight, respectively). Similar percentages of positive blood cultures were found with both regimens (17% and 22%, respectively); these rates were significantly lower than those reported in children who did not receive antibiotic prophylaxis (Roberts and Holzel, 2002). In contrast, another trial found intravenous vancomycin to be fairly ineffective in the prevention of bacteremia secondary to dental extractions, since 38% of the patients who received this glycopeptide had positive post-manipulation blood cultures (Kaneko et al., 1995). To provide new prophylactic antimicrobial alternatives, some authors have investigated the prevalence of bacteremia following dental manipulations after the administration of antibiotics not included in the bacterial endocarditis prophylaxis protocols recommended by the Expert Committees, and in force at the time of study. One study evaluated the efficacy of 2 g of oral metronidazole, compared with 2 g oral penicillin V or placebo (Head et al., 1984). Although penicillin V prophylaxis was associated with a lower prevalence of post-extraction bacteremia (20% vs. 52% in the "metronidazole group" and vs. 84% in the "placebo group"), it is interesting to note that Gram-negative obligate anaerobes were isolated from the blood cultures of four (16%) patients receiving penicillin G, whereas they were not identified in any of the cultures from patients receiving metronidazole (Head et al., 1984). One trial reported that an intravenous bolus of 400 mg of teicoplanin significantly reduced the post-extraction streptococcal bacteremia (from 32% in the "control group" to 2% in the "teicoplanin group"); its efficacy was also superior to that of intramuscular 1 g of amoxicillin given 20 to 30 min before anesthetic induction (Shanson et al., 1987). Another study demonstrated that patients receiving 1.5 g of cefuroxime intravenously had a significantly lower rate of bacteremia after multiple dental extractions than did controls. This finding was also seen after 10 min (79% and 23%, respectively) and 30 min (69% and 20%, respectively) of commencement of the surgical manipulation (Wahlmann et al., 1999). In contrast, other authors did not achieve a reduction in the prevalence of post-extraction bacteremia after the oral administration of 200 mg of oxafloxacin and 375 mg of sultamicillin 1 hr before the intervention, findings which were probably affected by the low doses used (Göker and Güvener, 1992). One study found that oral 1 g of cefaclor did not influence the prevalence or magnitude of the post-extraction bacteremia caused by viridans group streptococci or obligate anaerobic bacteria (either during the manipulation or 10 min later) (Hall et al., 1996a). In a recent paper, the effect of a single oral dose of 400 mg of moxifloxacin showed this to be effective, significantly reducing the prevalence (at 30 sec) and duration (at 15 min and 1 hr) of positive blood cultures secondary to dental extractions (54%, 21%, and 4%, respectively, in the "moxifloxacin group" vs. 96%, 64%, and 20%, respectively, in the "control group") (Diz Dios et al., 2006). From the first data published in 1956, another appealing line of research arose, that of topically administered antibiotic prophylaxis (Bender and Pressman, 1956). In one trial, the use of pre-operative topical vancomycin (starting 4 days before the dental treatment) resulted in a lower percentage of positive post-curettage or post-extraction blood cultures (25% and 69%, respectively), than after the application of a placebo (47% and 94%, respectively), but the differences between the groups were not significant (Bartlett and Howell, 1973). Other authors have evaluated the effect of topically applied amoxicillin on the prevalence of post-extraction bacteremia (Vergis et al., 2001). The study group was comprised of 10 controls and 15 patients who performed a double mouthwash with amoxicillin for 1–2 min. Although the amoxicillin reduced the percentage of post-extraction bacteremia compared with that in controls (53% vs. 90%), the sample size was too small to reach statistical significance (Vergis et al., 2001).
• Intensity The small number of colony-forming units (CFU)/mL usually detected in positive blood cultures associated with dental manipulations probably makes it difficult to interpret the possible effects of antibiotic prophylaxis on the intensity of post-manipulation bacteremia (Roberts et al., 2000; Lucas et al., 2002b).
• Animal Models In 1970, a study demonstrated that the insertion of a polyethylene catheter into a rabbit heart led to the appearance of a small sterile vegetation, which favored bacterial colonization after the injection of an infectious bolus (Garrison and Freedman, 1970). Since then, animal models have been key to the study of the efficacy of antibiotics in the prevention of bacterial endocarditis. The common characteristic of all these studies is that the animals, at the time of inoculation of the microorganisms into the bloodstream, had serum levels of antibiotics similar to those detected in humans after receiving standard prophylactic doses. These investigations have enabled workers to define accurately the time-course of cardiac infection and the influence of the size of the bacterial inoculum in its etiopathogenesis. However, results obtained on the efficacy of antibiotic prophylaxis in experimental models cannot necessarily be extrapolated to humans (Mizen and Woodnutt, 1988; Vogelman et al., 1988). Other problems have also been attributed to this type of study, such as the need for placement of an intra-cardiac catheter to induce the initial lesion, and the size of the bacterial inoculum necessary to infect the animals (Glauser and Francioli, 1987). The first studies revealed that antibiotic prophylaxis was effective, preventing the onset of experimental bacterial endocarditis in rabbits, although prophylactic regimens varied in efficacy (Durack and Petersdorf, 1973). "Bacterial death" was said to be the mechanism responsible for the success of prophylaxis (Southwick and Durack, 1974). Modifying the experimental model previously described (Garrison and Freedman, 1970), investigators produced sterile cardiac vegetations by inserting a polyethylene catheter into the hearts of rabbits and then infecting them by the intravenous injection of 108 CFU of a Streptococcus sanguis strain originally isolated from a patient with bacterial endocarditis (Durack and Petersdorf, 1973). The success of antibiotic prophylaxis was determined based on the results of the conventional culture of the endocardial vegetations, 24 hrs after the administration of the antibiotics. The results of this study revealed that intramuscular penicillin G (150 mg/kg bodyweight), combined with streptomycin (15 mg/kg bodyweight) or procaine penicillin (250 mg/kg bodyweight) and penicillin G (150 mg/kg bodyweight) combined with benzathine penicillin (7.5 mg/kg bodyweight) were effective in the prevention of bacterial endocarditis. In contrast, lower doses of penicillin G did not significantly reduce the number of rabbits that developed cardiac infection. In consequence, this trial demonstrated the need to achieve high serum concentrations of penicillin and to maintain the bactericidal activity in the serum for 6–8 hrs as essential in preventing bacterial endocarditis. In this study, the efficacy of other bactericidal agents, such as vancomycin, was also demonstrated, while antibiotics such as cefalexin, cefaloridine, and rifampicin did not prevent cardiac infection. To explain these findings, it was suggested that the bactericidal agents used in the prophylaxis protocols had to cause "total bacterial death", since the survival of only 0.1% of the bacteria could still be sufficient to cause bacterial endocarditis (Durack and Petersdorf, 1973). In a trial published in 1986, different oral regimens of amoxicillin (associated with probenecid), and an intramuscular combination of penicillin G, streptomycin, and probenecid were administered 1 hr after the inoculation of 104, 106, and 108 CFU of penicillin- and amoxicillin-sensitive Streptococcus sanguis to rabbits with experimentally induced cardiac vegetations (Pujadas et al., 1986). The combination of penicillin G, streptomycin, and probenecid provided protection against all the bacterial concentrations used. The administration of a single dose of amoxicillin and probenecid prevented cardiac infection in rabbits infected with 104 CFU, but its efficacy decreased with larger bacterial inoculations. However, the prophylaxis based on 2 oral doses of amoxicillin and probenecid (the second dose administered 10 hrs after the first) was completely effective in preventing bacterial endocarditis, independently of inoculum size, and these authors therefore favored the administration of a second dose of amoxicillin and the incorporation of probenecid in patients considered to be at "high risk" of bacterial endocarditis and about to undergo a dental procedure that involved the potential passage of "high magnitude bacterial inocula" into the bloodstream (Pujadas et al., 1986). Soon the dilemma arose as to whether the "bacterial death" induced by the antibiotic prophylaxis could fail in patients with bacterial endocarditis caused by Streptococcus spp. showing tolerance to amoxicillin (minimum bactericidal concentration/minimum inhibitory concentration [MBC/MIC] = 32 mg/L) (Glauser et al., 1983). A study of the efficacy of a single dose of amoxicillin on the prevention of bacterial endocarditis in rats, after the inoculation of Streptococcus spp. with different sensitivity profiles to this beta-lactam, found that amoxicillin was bactericidal in an inoculum-independent manner against bacteria not tolerant of this antibiotic, signifying that it was active against inocula up to 1000 times higher than the Infective Dose90 (ID90) (Glauser et al., 1983). Furthermore, protection against bacteria tolerant to amoxicillin was principally due to the inhibition of bacterial adherence to the cardiac vegetations, through an inoculum-dependent action, with a reduction in efficacy when the size of the bacterial inoculum was increased (Glauser et al., 1983). Confirming this theory, other authors, in in vitro studies, observed that the adhesion of Streptococcus sanguis to fibrin and platelet coagula decreased after the administration of penicillin or amoxicillin, probably as a consequence of the effects of these antibiotics on bacterial structures such as teichoic and lipoteichoic acids, or on certain specific proteins of the cell surface (Christensen et al., 1985; Kusser et al., 1985; Nealon et al., 1986). A study published in 1987 on the prevention of experimentally induced streptococcal bacterial endocarditis in rabbits showed that amoxicillin was much more effective than cephradine (Longman et al., 1987). In the first 4 hrs, the serum levels of amoxicillin were higher than the MICs and MBCs of the Streptococcus sanguis inoculated, whereas the concentrations of cephradine were lower than the MBC of this microorganism, underlining the importance of achieving high serum levels of the antibiotics for several hrs for the prevention of bacterial endocarditis (Longman et al., 1987). A trial was designed with a computer-controlled system for the continuous infusion of antibiotic, aiming to simulate in rats the pharmacokinetic characteristics of a single dose of 3 g of amoxicillin in humans (Fluckiger et al., 1994). In this experiment, detectable blood levels of amoxicillin were achieved for more hours than after the injection of a single intravenous bolus (9 hrs vs. 4.5 hrs), enabling these rats to be protected against much larger bacterial inocula (up to 100 times higher). Consequently, these authors stated that the prolonged presence of amoxicillin in the bloodstream led to the activation of other defense mechanisms. On this basis, it was speculated that amoxicillin, under "non-bactericidal" conditions, could inhibit the growth of bacteria adherent to cardiac vegetations, favoring their elimination by host defense mechanisms such as microbicidal platelet proteins (Fluckiger et al., 1994). Based on this hypothesis, and taking into account that bacterial growth on the vegetations is significant by 4 hrs after the onset of the bacteremic episode (Moreillon et al., 1986), the efficacy of the post-bacteremic phase administration of amoxicillin in the prevention of experimental endocarditis was evaluated in rabbits and rats (James et al., 1987; Berney and Francioli, 1990); these studies demonstrated that prophylaxis was effective if it was administered up to 2 hrs after an ID90, but that this effect was not achieved when the beta-lactam was administered 4 or 6 hrs after injection of the bacterial inoculum (James et al., 1987; Berney and Francioli, 1990). Another mechanism of action of amoxicillin proposed in the prevention of bacterial endocarditis is the separation of the bacteria adhering to the vegetation due to structural modifications in their cell walls (Lowy et al., 1983a). However, in vitro experiments have not confirmed this after exposure to bacteriostatic concentrations of amoxicillin for 4 hrs, a period that simulates the time of exposure to amoxicillin in vivo (Moreillon et al., 1986). Nevertheless, this mechanism of action could not be definitively excluded, due to the methodological difficulties involved in demonstrating a phenomenon in which probably only a small number of bacteria are involved (Berney and Francioli, 1990). A study compared the efficacy of single doses of clarithromycin and clindamycin in the prophylaxis of experimentally induced streptococcal bacterial endocarditis in rats, demonstrating that both prophylactic regimens were successful in preventing the onset of bacterial endocarditis after the inoculation of an ID90 or up to 100 times higher (Vermot et al., 1996). Since it is rare to find such large inocula from the bloodstream after dental procedures, it has been suggested that these antibiotics provided wide safety margins for bacterial endocarditis prophylaxis in humans (Vermot et al., 1996). It has been demonstrated that clindamycin inhibits the adherence of the viridans group streptococci to sterile vegetations by interfering with glycocalyx synthesis (Dall et al., 1990). In 1997, a trial was published on the efficacy of azithromycin and clarithromycin in the prevention of bacterial endocarditis in rabbits after the intravenous infusion of 5 x 105 CFU of Streptococcus milleri, comparing this with the effects of other antimicrobial agents frequently used for prophylaxis (amoxicillin, erythromycin, and clindamycin) (Rouse et al., 1997). Cardiac infection developed in 88% of animals not receiving prophylaxis, in 9% of those receiving erythromycin, and in only 0–2.5% of those receiving the other antibiotics (Rouse et al., 1997). It has been demonstrated that azithromycin is effective in preventing experimental streptococcal endocarditis (protecting 94% of animals challenged with Streptococcus oralis), but, against methicillin-resistant Staphylococcus aureus-challenged animals, it was less effective than vancomycin (59% and 94%, respectively) (Tsitsika et al., 2000). Recently, other authors have investigated the prophylactic efficacy of fluoroquinolones (antibiotics not included in the bacterial endocarditis prophylaxis protocols) in preventing streptococcal aortic valve endocarditis (Katsarolis et aI., 2000; Sakka et al., 2005). One of these trials reported that a single dose of moxifloxacin (15 mg/kg intravenously administered) prevented endocarditis in 80% of rabbits challenged with 107 CFU of Streptococcus oralis. In consequence, these authors concluded that moxifloxacin was at least as effective as ampicillin in preventing streptococcal endocarditis (Sakka et al., 2005). To achieve a better simulation of the specific conditions of bacterial endocarditis secondary to dental manipulations in humans, investigators designed a study on the efficacy of single doses of amoxicillin and erythromycin on the prevention of bacterial endocarditis after extracting teeth with periodontal disease from rats with sterile cardiac vegetations induced by a catheter (Malinverni et al., 1988). In the animals receiving antibiotic prophylaxis, the percentage of positive post-extraction blood cultures fell by only between 20% and 40% with respect to the rates observed in controls. However, the prophylactic administration of amoxicillin or erythromycin was effective in preventing bacterial endocarditis (only 10% and 7%, respectively, of the rats receiving prophylaxis with these antibiotics developed bacterial endocarditis vs. 89% of the control rats), allowing these authors (Malinverni et al., 1988), in agreement with other researchers (Glauser et al., 1983), to suggest the potential participation of other protective mechanisms, distinct from "bacterial death" in the bloodstream, in the success of antibiotic prophylaxis.
• Human Studies It has been stated that antibiotic prophylaxis does not prevent the onset of bacterial endocarditis in 100% of cases (Durack, 1995). In 1983, a series was published including 52 cases of bacterial endocarditis in patients with predisposing cardiac lesions, with a history of previous dental procedures, and who had received a prophylactic regimen with penicillin or erythromycin (Durack et al., 1983). The prophylactic protocol recommended by the American Heart Association in 1977 (Kaplan, 1977) had been administered in only six of the 52 cases; consequently, the majority of the patients had received "inadequate" prophylactic cover, allowing it to be suggested that, in many patients developing bacterial endocarditis, this could be due to the use of regimens differing from the guidelines published by the various Expert Committees. Many surveys performed among physicians and general dental practitioners have demonstrated that the type and dose of antibiotic administered to patients susceptible to bacterial endocarditis undergoing "at risk" dental procedures varied considerably, with the percentage of responders adhering to the American Heart Association and British Society of Antimicrobial Chemotherapy protocols varying between 6% and 96% (Hashway and Stone, 1982; Nelson and van Blaricum, 1989; Vuille and Bloch, 1992; Forbat and Skehan, 1993). Recently, in a survey performed in Spain among "high risk" patients, only 34% confirmed that they had been informed by their cardiologists about the need for antibiotic prophylaxis in some specific situations (Caballero-Borrego et al., 2006). Furthermore, it has been suggested that the efficacy of the preventive strategy could be affected by the low compliance index of the patients, since, in a series of 318 Dutch patients in whom antibiotic prophylaxis for bacterial endocarditis was indicated, only 22% complied adequately with the prescription (van der Meer et al., 1992a). A retrospective analysis of 52 cases of bacterial endocarditis, attributed to a possible failure of the antibiotic prophylaxis, could not confirm that the profiles of antimicrobial sensitivity of the isolates responsible for the cardiac infection represented a risk factor in those persons in whom the prophylactic regimen was administered correctly (Durack et al., 1983). However, several cases are to be found in the literature, such as one that referred to a patient receiving intravenous erythromycin during a surgical procedure for maxillary sinusitis (and also on the following days), who later developed bacterial endocarditis due to erythromycin-resistant Streptococcus sanguis (the isolate presented a MIC value to erythromycin of 40 mg/L) (Eng et al., 1982). On this subject, it has been indicated that the principal reason for re-evaluation of the current prophylactic protocols for bacterial endocarditis is the growing prevalence of bacterial resistance, particularly in Streptococcus spp. (Durack, 1998). Furthermore, it has been demonstrated that these resistances are associated with the expression of certain genes that may be transferred between bacteria in the oral ecosystem (Durack, 1998). However, the efficacy of prophylaxis with amoxicillin in the prevention of bacterial endocarditis in rabbits that had previously been inoculated with isolates of Streptococcus sanguis resistant to this beta-lactam has been demonstrated, leading to the suggestion that the profiles of antimicrobial sensitivity tested in vitro do not represent a predictive factor for the efficacy of antibiotic prophylaxis (Longman et al., 1992). Some authors have suggested that, even assuming that antibiotic prophylaxis was effective in 100% of cases, its administration would prevent only a small number of cases of bacterial endocarditis (van der Meer et al., 1992b; Strom et al., 1998; Duval et al., 2006). In a retrospective analysis of 427 cases of bacterial endocarditis, of which 275 were patients classified as "at risk", 64 patients (23%) had undergone a dental procedure requiring prophylaxis in the 180 days prior to the onset of symptoms of bacterial endocarditis, but only 17 had received a bacterial endocarditis prophylaxis regimen. These findings enabled the authors to conclude that: "Assuming an incubation period of 180 days, the administration of prophylaxis would have avoided cardiac infection in 47 patients, representing only 17% of the bacterial endocarditis diagnosed in patients with heart disease undergoing procedures requiring prophylaxis" (van der Meer et al., 1992b). It was estimated that, even with antibiotic prophylaxis with a 100% efficacy, the incidence of bacterial endocarditis would be reduced by only two cases/1,000,000 population/yr, considering a frequency of five cases of bacterial endocarditis/100,000 population/yr in the general population (Strom et al., 1998). In a recent study, the estimated risk of developing bacterial endocarditis was 1 in 10,700 and 1 in 54,300 for persons with prosthetic- and native-valve-predisposing cardiac conditions, respectively, after undergoing an "at risk" dental procedure and not having received prophylaxis (Duval et al., 2006). The Cochrane group published a meta-analysis that evaluated whether the prophylactic administration of penicillin before the performance of invasive dental treatments in patients "at risk" of bacterial endocarditis affected the prevalence of the cardiac infection (Oliver et al., 2004). Of the 108 studies analyzed, only one (van der Meer et al., 1992b) satisfied the inclusion criteria. In consequence, these authors concluded that there is currently no scientific evidence that prophylaxis with penicillin is effective in reducing bacterial endocarditis secondary to dental procedures in patients considered to be "at risk" (Oliver et al., 2004). For some experts, the effectiveness of the prophylactic administration of antibiotics in the prevention of bacterial endocarditis could be limited even if administered correctly, since it is difficult to identify the precise moment at which the bacteremia responsible for the cardiac infection occurs, and because there are cases of bacterial endocarditis in patients with unknown pre-existing cardiac valve abnormalities but eligible for prophylaxis (Oakley, 1987; van der Meer, 2002). It has been recently suggested that the use of antibiotic prophylaxis to prevent bacterial endocarditis caused by dental procedures must be reassessed, based on the above recent evidence (Wahl and Pallasch, 2005). These authors consider that at least 99% of those given antibiotic prophylaxis for bacterial endocarditis are unlikely to receive any benefit (Wahl and Pallasch, 2005).
Although it is impossible to calculate accurately the risk of onset of severe allergic reactions secondary to the administration of penicillins, it has been estimated that this could occur in 0.04–0.2% of patients (Finch, 1990; International Rheumatic Fever Study Group, 1991), with an associated mortality of one case per 60,000 exposures to penicillin (16 per million prescriptions) (Idsoe et al., 1968). Other authors have suggested that the mortality indices could be considerably higher, one case per 2000–2500 exposures to penicillin (Atkinson and Kaliner, 1992). In the mid-1980s, the first studies were published that evaluated the benefit of antibiotic prophylaxis, taking into account the number of bacterial-endocarditis-related deaths avoided by the prophylaxis, and the number of deaths caused by severe allergic reactions to penicillins (Bor and Himmelstein, 1984; Clemens and Ransohoff, 1984; Tzukert et al., 1986). One of these studies estimated that 47 cases of bacterial endocarditis could occur for every 10 million dental procedures performed in patients with mitral valve prolapse, if antibiotic prophylaxis was not administered, and that two of these would lead to the death of the patient. If prophylaxis with penicillin was prescribed at each visit, the number of cases of bacterial endocarditis would fall to five, and there would be no deaths due to this cause, but up to 175 individuals could die as a consequence of allergic reactions to the penicillin (Bor and Himmelstein, 1984). Other authors calculated the annual mortality rate due to bacterial endocarditis of dental origin, and that attributable to antibiotic prophylaxis in patients with rheumatic heart disease and undergoing dental treatment, estimating that, in a population of 100 million, there would be approximately 26 deaths per yr due to bacterial endocarditis secondary to dental procedures. Considering a population of 3.4 million patients with rheumatic heart disease who would presumably attend a dental consultation once a yr, with the routine administration of prophylactic cover with penicillin, there would be around 136 deaths annually due to anaphylactic shock (Tzukert et al., 1986). In the United States, it was estimated that allergic reactions to penicillin could be responsible for 400 to 800 deaths per yr, whereas the administration of antibiotic prophylaxis would prevent only between 240 and 480 cases of bacterial endocarditis, implying a higher mortality index associated with allergic reactions to penicillins (Tzukert et al., 1986; Pallasch, 1989), particularly if we consider that the bacterial endocarditis of oral etiology caused by Streptococcus viridans is lethal in less than 10% of cases. Assuming an incidence of bacterial endocarditis of 11–50 cases/1,000,000 population/yr, and an associated mortality rate of 25–40%, vs. a mortality due to anaphylactic reactions secondary to penicillin of 16 cases/1,000,000 population/yr, investigators concluded that the mortality due to bacterial endocarditis exceeded that due to allergic reactions to penicillin only when the maximum values of the incidence of bacterial endocarditis and associated mortality were assumed (50 cases/1 million population/yr and 40%, respectively) (Pallasch, 1989). It has been suggested that the development of allergic reactions secondary to the use of penicillins depends on the dose and duration of treatment (Dukes and Aronson, 2000). A review published in 1999 stated that there have been no cases reported in the literature of serious allergic reactions in patients with no history of allergy to penicillin after receiving a single oral dose of 2 g of amoxicillin (Wynn et al., 1999). However, a case of an allergic reaction to oral prophylactic amoxicillin administered prior to dental treatment has been recently reported (Sadaghiani et al., 2005). To our knowledge, only one case of pseudomembranous colitis secondary to a single dose of 600 mg of clindamycin has been reported (Bombassaro et al., 2001). In contrast, it has been reported that almost half the patients receiving a prophylactic regimen of oral erythromycin (1.5 g 1 hr before the procedure and 0.5 g 6 hrs later) had adverse gastrointestinal effects (Sefton et al., 1990). Although some authors stated that the use of antibiotics for prophylactic purposes in dental practice was not a significant abuse of antibiotics and, therefore, of the development of bacterial resistance, others contradict this statement in relation to the administration of several prophylactic doses (Tong and Rothwell, 2000). It has been shown that a single dose of 3 g of amoxicillin did not lead to a significant increase in the number of salivary-resistant streptococci, but that the administration of a second or third dose caused a marked rise in the percentage of resistant isolates, which persisted for 4 to 7 wks (Woodman et al., 1985; MacGregor and Hart, 1986). In a trial published in 1987, 4 g of phenoxymethyl penicillin were orally administered over a period of 10 hrs (divided into 3 doses) to 29 healthy volunteers who were not carriers of penicillin-resistant Streptococcus viridans, and by 6 hrs, streptococci resistant to this beta-lactam were isolated in nine individuals (31%), and these persisted for 9 days (Leviner et al., 1987). The prevalence of amoxicillin-resistant Streptococcus spp. was also investigated in a group of 12 volunteers taking 2 oral doses of 3 g of amoxicillin (at an interval of 8 hrs), repeating the process weekly on five occasions. After the administration of the final dose, all the individuals had amoxicillin-resistant oral Streptococcus spp., still present in all the individuals 13 wks after the final dose of the antibiotic (Southall et al., 1983). With respect to the prophylactic protocols involving macrolides, in a study published in the 1908s, 3 doses of erythromycin stearate (total dose, 2 g) were administered on two occasions to 10 healthy persons, at an interval of 1 wk. After the second dose, erythromycin-resistant Streptococcus spp. were isolated from the oral cavities of all the participants, and isolates with very high MIC values (ranging between 16 and > 256 mg/L) were detected in four persons. These resistant isolates remained detectable for up to 23 wks after the administration of the drug in eight of the 10 volunteers, and in five people, the resistance persisted for up to 43 wks (Harrison et al., 1985). In 1990, a trial was designed to evaluate whether the administration of 2 doses of erythromycin and josamycin affected the selection of resistant streptococci in the oral cavity (Maskell et al., 1990). These authors observed that the low proportion of resistant microorganisms present before the administration of the antibiotic prophylaxis increased significantly 48 hrs after administration of the macrolides. The percentages of streptococci resistant to the antibiotics administered, with MICs of 1, 4, and 64 mg/L, were 23%, 17%, and 6%, respectively, for erythromycin, and 13%, 6%, and 4%, respectively, for josamycin (Maskell et al., 1990). In contrast, other authors demonstrated resistant oral streptococci in patients who had received a prophylactic dose of penicillin V (2 g initially and 1 g 6 hrs later) on 3 consecutive Mondays, but that these represented only 0.0003–0.41% of the total streptococcal population cultured (Fleming et al., 1990). In consequence, these authors concluded that prophylaxis with penicillin V administered on a number of consecutive occasions (with an interval of 1 wk between administrations) did not lead to significant levels of resistance in the Streptococcus spp. in the oral cavity (Fleming et al., 1990).
• Antibiotic of Choice The onset of bacterial endocarditis of oral origin involves the presence of a previous bacteremia (Drangsholt, 1998). More than half of the studies published on antibiotic prophylaxis and post-dental manipulation bacteremia have investigated the efficacy of the prophylactic administration of penicillins (Elliott and Dunbar, 1968; Shanson et al., 1978; Baltch et al., 1982a,b; Hess et al., 1983; Roberts et al., 1987; Hall et al., 1993; Vergis et al., 2001; Lockhart et al., 2004), and their results confirm the efficacy thereof. However, there are fewer studies on the effect of the prophylactic administration of other recommended antibiotics (clindamycin, azithromycin, and cephalosporins) on the prevention of post-dental manipulation bacteremia (Josefsson et al., 1985; Sefton et al., 1990; Göker and Güvener, 1992; Aitken et al., 1995; Hall et al., 1996a; Diz Dios et al., 2006), and their results do not confirm the efficacy of these antibiotics. To date, most of the papers on the prevalence of post-dental manipulation bacteremia have confirmed that most of the bacteria isolated in the blood cultures are sensitive to the antibiotics recommended in the prophylaxis protocols by the Expert Committees (Shanson et al., 1978; Josefsson et al., 1985; Sefton et al., 1990; Hall et al., 1993, 1996b; Aitken et al., 1995). However, increasing resistance to the beta-lactams, macrolides, and lincosamides has recently been found in oral bacteria (Groppo et al., 2005), and this could restrict their use for bacterial endocarditis prophylaxis. Recently, new antibiotics, such as the fluoroquinolones, have been shown to be successful in the prophylaxis of bacteremia following tooth extractions in humans (Diz Dios et al., 2006), and to prevent endocarditis in animal models (Sakka et al., 2005). It has been shown that the inefficacy of some antibiotic prophylactic regimens for the prevention of post-dental manipulation bacteremia does not necessarily imply that these cannot prevent the development of bacterial endocarditis (Glauser et al., 1983; Berney and Francioli, 1990; Dall et al., 1990). However, more scientific evidence of the effect of antibiotic prophylaxis on the prevalence and duration of bacteremia following dental procedures is needed, with analysis of the influence of the increasing prevalence of bacterial resistance in the oral ecosystem (Durack, 1998). Prophylactic alternatives such as oral antiseptics (i.e., chlorhexidine) (Tomás et al., 2007) and peptides that can interfere with bacterial adhesion should also be explored (Ito, 2006).
• Time of Administration of the Antibiotic Prophylaxis In the majority of experimental studies of bacterial endocarditis in animals, the responsible microorganisms are inoculated directly into the bloodstream. What these studies are actually investigating, therefore, is the effect of antibiotics administered after the induction of bacteremia (Pelletier et al., 1975; Glauser and Francioli, 1982; Glauser et al., 1983; Pujadas et al., 1986). In contrast, some authors have suggested that the success of antibiotic prophylaxis in preventing post-dental manipulation bacteremia in humans could be due to the action of the antibiotic on the bacteria in the oral cavity, before these invade the bloodstream (Bender et al., 1984; Aitken et al., 1995). In addition to the concentration that the antibiotic reaches in the oral cavity and the antimicrobial sensitivity of oral bacteria, another factor which must be taken into account is the time of exposure of the bacteria in the oral cavity to the antibiotic. Some authors have suggested that "the contact time of the antibiotic with the bacteria in the gingival sulcus could be too short to ensure success of the prophylaxis in the prevention of post-dental manipulation bacteremia" (Hess et al., 1983). After a detailed review of the literature, we were interested to note that, in several of studies that reported the efficacy of antibiotic prophylaxis in the prevention of post-dental manipulation bacteremia, the antibiotic was administered at least 2 hrs before the dental procedure (Shanson et al., 1978; Lockhart et al., 2004; Diz Dios et al., 2006). This situation could favor the activity of the antibiotic at a "local level" and, therefore, its efficacy in the prevention of bacteremia following dental manipulations. If one assumes that this is a possible mechanism of local action (Bender et al., 1984; Aitken et al., 1995), the time recommended for administration of antibiotic prophylaxis could be questioned. An increase in the time of contact between the bacteria and the antibiotic in the oral cavity might decrease the prevalence and size of a post-dental manipulation bacteremia.
• Route of Administration of the Antibiotic Prophylaxis The efficacy of intravenous antibiotics for the prevention of post-dental manipulation bacteremia is similar to that found in series in which the antibiotic prophylaxis was administered orally (Baltch et al., 1982a,b; Roberts et al., 1987; Lockhart et al., 2004). The mechanism of action of antibiotic prophylaxis at a dentoalveolar level proposed by some authors could represent a possible explanation (Bender et al., 1984; Aitken et al., 1995). The oral administration of therapeutic doses of amoxicillin and clindamycin gives rise to high antibiotic concentrations in the gingival fluid (3–4 mg/L and 1–2 mg/L, respectively) (García et al., 1996). Despite the fact that the intravenous route provides higher serum antibiotic concentrations at the time of the manipulation than does oral administration (Roberts et al., 1987), several authors have demonstrated that the oral amoxicillin and clindamycin provide high serum concentrations in the first and second hrs after ingestion (15 and 25 mg/L, and 4.5 and 4.8 mg/L, respectively), with high levels at 4–6 hrs (5 mg/L and 2 mg/L, respectively), and detectable levels still present after 9–10 hrs (0.7 mg/L and 0.2 mg/L, respectively) (Shanson et al., 1978; Dan et al., 1997). Taking into account that bacterial growth in the vegetations starts significantly at around 4 hrs after the onset of bacteremia (Moreillon et al.,1986), it has been stated that the prolonged presence of amoxicillin and clindamycin in the bloodstream probably led to the activation of other defense mechanisms (Dall et al., 1990; Fluckiger et al., 1994). The low compliance with intravenous guidelines by both patients and practitioners is clear and not surprising. In consequence, it may be important to carry out more studies on the efficacy of oral prophylaxis in the prevention of post-dental manipulation bacteremia and experimental bacterial endocarditis in comparison with intravenous prophylaxis.
Received for publication December 17, 2006. Revision received May 21, 2007. Accepted for publication August 30, 2007.
Journal of Dental Research, Vol. 86, No. 12,
1142-1159 (2007) This article has been cited by other articles:
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
TOP
ABSTRACT
(1) BACTERIAL ENDOCARDITIS OF...
