Antibiotics for Prevention of Endocarditis during Dentistry: Time To Scale Back?

Antibiotics should be given to prevent endocarditis when a predisposed patient undergoes dental treatment. This 50-year-old doctrine is based on a logical premise: Dental procedures induce bacteremias with bacterial species that often cause endocarditis, and endocarditis carries high morbidity and mortality; therefore, antibiotics should be administered to susceptible patients before dental procedures to prevent endocarditis [1]. Furthermore, antibiotics definitely can prevent endocarditis in experimental animals [1-3]. On these grounds, prophylaxis against endocarditis is widely recommended [1, 4]. Although compliance has been less than perfect [5] and apparent failures of prophylaxis are not rare [6], the practice is now so entrenched that failure to give prophylaxis has generated many malpractice claims [7]. However, its effectiveness in humans, let alone its cost-effectiveness, has not been proven and probably never will be [1].

The foundation for the practice of endocarditis prevention has been questioned repeatedly [8-14] but not seriously shaken. As a colleague wryly commented, “The less the evidence there is, the more antibiotic we give.” Now, Strom and colleagues [15] have stirred the pot with a major case–control study of 273 adults in which no link was found between endocarditis and dental procedures. The incubation period before onset of symptoms when endocarditis follows a dental procedure is usually short [16], but in this study, the calculated risk for endocarditis was no higher in the first month after dental treatment than after 2 or 3 months. This finding is consistent with a lack of a link between the two events. Of course, it does not prove that dental procedures never cause endocarditis-occasionally, they do [17, 18]. The authors simply argue that this happens too rarely to justify routine use of antibiotic prophylaxis. This supports the conclusions of an earlier study by van der Meer and colleagues, which was based on a large survey in the Netherlands [10, 14]. In contrast to the apparent irrelevance of dental treatment, Strom and colleagues' findings emphasized the importance of underlying cardiac valvular abnormalities and quantitated the major risks posed by prosthetic valves and by previous endocarditis, which carry odds ratios of 75 and 37, respectively.

In this study, no particular type of dental work was significantly linked to infective endocarditis, not even tooth extraction [15]. Tooth extractions, however, are conspicuous in case reports of endocarditis after dental work [16, 17]. Furthermore, extractions are more likely than most other dental procedures to produce bacteremias [18]. Strom and colleagues report that 6 of 273 patients with endocarditis had had tooth extractions within 2 months compared with 0 of 273 controls [15]. These numbers are too small to be statistically significant, but they indicate the real possibility of a type II error with regard to extractions. This caveat is critical, but it should not detract from the credibility of the main conclusions.

In summary, three major studies [10, 14, 15] and several outcomes analyses and commentaries [8, 9, 12, 13] have presented substantial grounds on which to challenge the value of the unvalidated practice of administering antibiotics before dental procedures to prevent endocarditis.

The time has come to scale back on prophylaxis against endocarditis before dental treatment. In the matrix of procedures related to predisposing conditions, prophylaxis should be downgraded to “not recommended” for most dental procedures except extractions and gingival surgery (including implant placement) and for most underlying cardiac conditions except prosthetic valves and previous endocarditis. When any one or more of these four high-risk factors is present, prophylaxis should follow the present American Heart Association (AHA) guidelines [4]. Mitral valve prolapse is not included among the underlying conditions requiring routine prophylaxis because although a significant number of cases of endocarditis do occur in patients with mitral valve prolapse, the denominator of susceptible persons is large and the risk encountered by an individual patient with mitral valve prolapse is lower than that for patients with prosthetic valves, previous endocarditis, or both [15]. Furthermore, the prognosis for cure of viridans streptococcal endocarditis in a patient with mitral valve prolapse is good [9, 12]

These proposed changes could eliminate most of the prophylactic doses currently given to dental patients. What benefits could be reaped? Less antibiotic use, fewer side effects [9, 12], more convenience for patients, less work for health care providers, and lower costs for third-party payers. Perhaps more important, some of the effort and expense currently expended on routine prophylaxis could be refocused on high-risk procedures and patients. The quantitative data now available [10, 14, 15] suggest, as a first approximation, that it should be possible to retain at least 80% of any putative benefits from currently recommended prophylaxis for less than 20% of the costs (the “80/20 rule”).

Selection for antibiotic resistance may provide another argument for reassessing the costs and benefits of prevention. Viridans streptococci, traditionally sensitive to antibiotics, have recently become more resistant [19]. Because they are feeble pathogens in normal hosts, this sea-change has not achieved the notoriety attracted by emerging resistance among pneumococci, enterococci, and staphylococci; however, it may have its own dark significance. At the entrance to the body, the oropharynx shelters a standing army of viridans streptococci, which have the potential not only to accept resistance genes from passing bacteria but also to donate them. This has already occurred in the case of resistant pneumococci [20]. Although the magnitude of selection pressure due to overuse of antibiotics for endocarditis prophylaxis is small compared with overuse of antibiotics for treatment, it may not be negligible. In any event, the case against ill-considered use of any unnecessary antibiotics is today stronger than ever.

What problems might arise if we scale back? Some additional cases of endocarditis might occur. The mathematics indicate that if any real increase did occur, it would be undetectably small [10, 14, 15]. Furthermore, when promptly diagnosed and treated, endocarditis after dental treatment has a reasonably good prognosis. Certainly, best practice indicates the need for vigilance and follow-up to achieve optimal results, and meticulous education of both health care providers and susceptible patients might lead to better outcomes than would unselective use of routine antibiotic prophylaxis. Another potential problem is that a major change in the current recommendations, which are widely known, might cause confusion among health care providers. In addition, evaluation of the merits of malpractice claims, already tendentious, might become even more convoluted. These problems could be minimized if the AHA, the Infectious Diseases Society of America, and other veterans of the guidelines wars successfully rise to the challenge of crafting appropriate new recommendations. Brief, simple instructions should help to avoid confusion and may result in better compliance. Inappropriate malpractice claims could be minimized by emphasizing that antibiotics are not generally required except for the subgroup of high-risk patients. Of course, a health care provider could choose the option of using prophylaxis in lower-risk patients, according to his or her clinical judgment. It should be re-emphasized that such guidelines are “not intended as the standard of care or as a substitute for clinical judgment” [4].

Of note, children were excluded from Strom and colleagues' study [15]. The issues of prophylaxis for infants and children at risk because of major congenital heart conditions or cardiac surgery and for patients undergoing nondental procedures require further discussion and are not addressed here.

Recommendations for the prevention of endocarditis have evolved more or less continuously since the first AHA pronouncement in 1955. In all, there have been nine major revisions of the AHA recommendations, plus many variations published by other authorities in the United States and other countries. At first, the thinking of the AHA committee was steered primarily by pediatric cardiologists who were experts on rheumatic fever. Later, the influence of infectious diseases specialists was felt, resulting in more extensive use (and probably overuse) of antibiotics. Enthusiasm for parenteral regimens, fueled by news of many studies showing that antibiotics could prevent experimental endocarditis [1-3], was later moderated in favor of simpler and more convenient oral regimens [4]. The emphasis may now shift again; the study by Strom and colleagues [15] suggests that we will increasingly feel the hand of epidemiologists and outcomes analysts on the wheel.

Are the current recommendations devolving toward eventual extinction? No. The concept of prophylaxis is valid and should be retained, but in a restricted and better-focused form. Committees will meet, controversy will erupt, and guidelines will continue to evolve. Eventually, consensus is needed. “The dogs may bark, but the caravan moves on.” At the end of the road, we are simply seeking the best possible outcomes for our patients at risk.

• Copyright ©2004 by the American College of Physicians