The first major change is the addition of a regimen suitable for outpatient treatment of infective endocarditis in patients with strains of streptococci that are highly susceptible to penicillin. This regimen consists of ceftriaxone, 2 g administered intravenously or intramuscularly once daily for 4 weeks. In the past, outpatient treatment with oral agents was not advocated because of the possibility of variable absorption of agents and lack of patient compliance. Furthermore, the previously used parenteral regimens for home therapy required two or three daily injections. The long half-life of ceftriaxone, which permits once-daily injection, is a major advantage in terms of cost and convenience. The daily injection can be given by a visiting nurse or family member in the home or can be given at a physician's office or an emergency department. An intravenous central line catheter may be required. Home therapy should be reserved for hemodynamically stable patients who have no complications of endocarditis.

Two studies have been published on the treatment of streptococcal endocarditis with once-daily ceftriaxone [3, 4]. In these studies, one possible relapse occurred among 70 patients treated for 4 weeks, yielding a bacteriologic cure rate of more than 98%. These results are similar to those seen with all other previously recommended regimens. Preliminary results from studies currently in progress suggest that once-daily injection of ceftriaxone plus an aminoglycoside for 14 days may be equally effective.

A second major change in the American Heart Association's recommendations is the recognition that enterococci are becoming increasingly resistant not only to aminoglycosides but also to penicillin G and vancomycin. Traditionally, enterococci are moderately resistant to penicillin G or ampicillin; a median of 2 µg of penicillin per mL is necessary for inhibition. Therefore, large daily doses have been used to treat enterococcal endocarditis. Penicillin alone, however, is not bactericidal for most strains of enterococci, and the addition of an aminoglycoside is necessary to provide the synergistic bactericidal activity necessary for reliable curative therapy.

For many years, streptomycin was the aminoglycoside successfully used in combination with penicillin. However, about 25 years ago, researchers noted that enterococci were becoming highly resistant to streptomycin (high resistance defined as not inhibited by 2000 µg of the drug per mL). With this high-level resistance, the synergistic bactericidal activity that occurs when penicillin is combined with an aminoglycoside did not occur. Gentamicin, which retained activity against streptomycin-resistant enterococci, was then increasingly substituted for streptomycin. However, enterococcal strains highly resistant to gentamicin began to appear.

Currently, at least 20% of enterococcal strains are highly resistant to gentamicin, and even more are highly resistant to streptomycin [5]. The gentamicin-resistant enterococci are also very resistant to all other aminoglycosides except streptomycin, to which less than half retain susceptibility.

A much more recent phenomenon is the appearance of enterococcal strains that are highly resistant to penicillin G and ampicillin (that is, not inhibited by 32 µg of the drugs per mL) and of strains that are highly resistant to vancomycin [6]. Vancomycin had previously been a reliable substitute for penicillin. When enterococci that are resistant to both penicillin and vancomycin cause endocarditis, no therapy is reliably effective. Strains of Enterococcus faecium are particularly likely to be resistant to penicillin G, gentamicin, or vancomycin [6]. In addition, penicillinase production has been noted in rare strains of E. faecalis, which are usually also resistant to aminoglycosides [6]. Ampicillin-sulbactam inhibits these bacteria.

The clinical significance of resistance to aminoglycosides, penicillin G, or vancomycin is sufficiently important that in vitro susceptibility testing is required for all enterococci isolated from patients with suspected endocarditis. When high-level resistance to gentamicin is found, susceptibility to streptomycin should be tested. With high-level resistance to both streptomycin and gentamicin, single-drug therapy with penicillin G, ampicillin, or vancomycin should be given for at least 8 weeks; the expected relapse rate is 50%. With relapse, cardiac surgery with valve resection must be considered.

Effective therapeutic regimens for infections with pathogens resistant to penicillin G and vancomycin are unlikely to be found. However, teicoplanin (an investigational glycopeptide) may be effective—it is active against some vancomycin-resistant strains of enterococci. Similarly, synercid (a combination of two pristinamycins), which has activity against many strains of E. faecium, is being studied in clinical trials. A less likely possibility is that an unusual combination of antibiotics that is found to have in vitro activity may be successful in treating endocarditis. For example, although combinations including a fluoroquinolone and rifampin have been successful in treating resistant enterococcal infections other than endocarditis [7], these regimens will probably not be effective in endocarditis in which high serum bactericidal activity is needed. Cardiac surgery may provide the only chance of cure.

The third major modification in the American Heart Association's new recommendations is the addition of guidelines for treating infections with organisms other than streptococci, enterococci, and staphylococci. Specifically, regimens are recommended for use in endocarditis caused by members of the HACEK group (Haemophilus, Actinobacillus, Cardiobacterium, Eikenella, and Kingella). These organisms are responsible for as many as 10% of cases of endocarditis. Formerly, these bacteria were almost uniformly susceptible to ampicillin, and standard therapy consisted of ampicillin with or without gentamicin for 4 weeks. Currently, many of these organisms have become resistant to ampicillin by acquiring the ability to produce penicillinase.

The new recommendation is to treat infections with these organisms with ceftriaxone, 2 g intramuscularly or intravenously once daily for 4 weeks; this regimen has proven successful in limited trials [8]. Ampicillin plus gentamicin can be used if the organism is susceptible to ampicillin. Not included in the recommendations is the fact that ampicillin-sulbactam plus gentamicin should be effective therapy for penicillinase-producing HACEK strains.

Two new points that are raised in the new recommendations are based on recent data and are worth emphasizing. First is the accumulating data showing that short-course therapy (2 weeks) with a penicillinase-resistant penicillin plus an aminoglycoside is adequate therapy for selected patients with methicillin-susceptible Staphylococcus aureus endocarditis restricted to the tricuspid valve [9-11]. Second is the observation that vancomycin may be inferior to penicillinase-resistant penicillins in treating S. aureus endocarditis caused by methicillin-susceptible S. aureus [9, 12-14].

The 2-week regimen is particularly appealing for use in injection drug users, in whom S. aureus endocarditis restricted to the tricuspid valve is frequently found. In general, injection drug users are difficult to keep in the hospital for prolonged periods, and problems are associated with long courses of parenteral therapy because of the lack of availability of intravenous access. Nafcillin plus tobramycin (both given intravenously for 2 weeks) and cloxacillin plus gentamicin, tobramycin, or amikacin (both given intravenously for 2 weeks) have proven effective in treating uncomplicated S. aureus endocarditis restricted to the tricuspid valve [9-11]. In these studies, about 94% of 134 patients were cured with this 2-week course. However, not all of the patients were proven to have endocarditis, and most of the treated patients were selected; this introduces the possibility of bias in the exclusion of problematic patients. Therefore, one must be cautious in selecting candidates for a 2-week course. Injection drug users with S. aureus endocarditis that is restricted to the tricuspid valve and is susceptible to methicillin and gentamicin [or tobramycin] are appropriate candidates if 1) clinical and bacteriologic response is seen within 96 hours of therapy; 2) no hemodynamic compromise, systemic emboli, renal insufficiency, or metastatic lesions develop; and 3) as suggested by DiNubile [15], vegetations are no larger in diameter than 2 cm. The regimen should be nafcillin, 2 g intravenously every 4 hours, plus gentamicin, 1 mg/kg of body weight intravenously every 8 hours, for 2 weeks. Restriction of infection to the tricuspid valve should be shown both clinically and by transesophageal echocardiography.

Of interest is the observation in one study [9] that vancomycin plus tobramycin for 2 weeks failed in two of three patients treated with this regimen. This finding agrees with other observations of an in vitro bactericidal response to vancomycin that was slower than the response to penicillinase-resistant penicillins [13]. Similarly, results have been reported in patients in which the response to vancomycin (as measured by duration of fever and blood-culture positivity) seemed to be slower than the response to penicillinase-resistant penicillins [12-14].

Explanations for these results include faster renal clearance in some patients, an inherently slower bactericidal rate of vancomycin at high inocula, and worse penetration of vancomycin into vegetations [14]. Although these interesting observations require more substantiation, they do suggest the prudence of using a penicillin rather than vancomycin for S. aureus endocarditis when possible.

In summary, the American Heart Association's new guidelines for treating infective endocarditis [1] update previous recommendations and include some major modifications: 1) the addition of a ceftriaxone regimen suitable for reliable outpatient therapy for streptococcal endocarditis; 2) the highlighting of emerging high resistance of enterococci not only to aminoglycosides but also to penicillin and vancomycin; and 3) the addition of recommended regimens for treating endocarditis caused by HACEK organisms. The new guidelines raise two other clinically important points: 1) Two-week regimens with a penicillinase-resistant penicillin plus an aminoglycoside may be adequate for methicillin-susceptible S. aureus endocarditis restricted to the tricuspid valve; and 2) vancomycin may be less effective than penicillinase-resistant penicillins for treating S. aureus endocarditis.