Short-Course Antibiotic Therapy for Right-Sided Endocarditis Caused by Staphylococcus aureus in Injection Drug Users

Background

The duration of antibiotic therapy is critical in predicting outcome in patients with endocarditis. Two weeks of therapy with penicillin alone cures a substantial proportion of patients with streptococcal endocarditis [8, 9]. The relapse rate of more than 20% with 2 weeks of penicillin monotherapy is nevertheless unacceptably high when compared with the relapse rates after 4 weeks of treatment (<2%) [10, 11]. However, 2 weeks of combination therapy with penicillin and an aminoglycoside antibiotic has essentially the same success rate as 4 weeks of penicillin therapy (with or without 2 weeks of treatment with an aminoglycoside) in patients with exquisitely penicillin-susceptible streptococcal endocarditis [10-14]. Similar data do not exist for staphylococcal endocarditis, but because of the virulence of S. aureus, standard practice has been to treat endocarditis caused by this organism, even when confined to the tricuspid valve, for a minimum of 4 weeks.

Right-sided endocarditis is in general a less aggressive infection than left-sided disease [15]. This difference can be at least partially explained by the greater hemodynamic significance of left-sided compared with right-sided heart valves. In experimental animals, tricuspid valve endocarditis responds more promptly and completely to antibiotic therapy. The number of organisms in right-sided vegetations and the extent of extra-cardiac metastatic infection is also less than in experimental left-sided disease. Resistance rarely develops among the bacteria residing in tricuspid valve vegetations. Antibiotic penetration differs between left- and right-sided vegetations; the period during which the concentration of antibiotics exceeds the minimum bactericidal concentration generally is longer for tricuspid than for aortic vegetations. Other differences between right- and left-sided endocarditis have been attributed to disparities in host immune responses and oxygen saturation. All the experimental evidence is consistent with the clinical observation that right-sided endocarditis is a more benign and responsive infection than left-sided disease.

Substantial data from in vitro and animal experiments suggest that combination therapy with β-lactam and aminoglycoside antibiotics synergistically and rapidly kill S. aureus[16, 17]. When young drug users with staphylococcal endocarditis receive nafcillin and gentamicin, they respond faster clinically and clear their bacteremia sooner, without undue toxicity, than they do with nafcillin alone [7]. However, the addition of an aminoglycoside for the initial 2 weeks of a standard course of a cell wall-active antibiotic does not confer any clinically important advantage over single-drug therapy [7].

Clinical Experience

Two recent prospective studies indicate that combination therapy with a penicillinase-resistant penicillin and aminoglycoside for 2 weeks is adequate treatment for most patients with uncomplicated, exclusively right-sided endocarditis caused by S. aureus[18, 19]. Although most patients had direct evidence of endocarditis (for example, presumed septic pulmonary emboli or abnormal echocardiograms), some patients were accepted into these trials simply on the basis of community-acquired S. aureus bacteremia that had no apparent source. Importantly, the criteria for enrollment closely paralleled usual clinical practice in diagnosing and treating right-sided endocarditis in parenteral drug users. Ninety-two percent (115 of 125) of the patients enrolled in these two trials were cured after 2 weeks of combination antibiotic therapy. The 10 failures included 5 relapses: 2 of the 3 patients treated with vancomycin instead of a penicillinase-resistant penicillin; 1 patient whose bacteremia persisted through day 4 of treatment; 1 patient whose therapy was interrupted on days 11 and 12 of treatment; and 1 patient who developed a subcutaneous abscess near the end of therapy. Cure was achieved in 4 patients for whom short-course treatment was planned but whose course of therapy was prolonged because of persistent signs of infection after 2 weeks. One patient died on day 10 of therapy; therefore, his death was unaffected by the projected duration of treatment. No deaths occurred in patients completing 2 weeks of therapy, either from the initial episode or a relapse of endocarditis.

In a third report, published in Spanish [20], the investigators evaluated 14 drug users with right-sided S. aureus endocarditis treated with cloxacillin (except for 2 patients given vancomycin) and gentamicin for 2 weeks; no deaths or relapses occurred after short-course therapy. No complications were seen in these patients compared with a concurrent but nonrandomized control group that received conventional therapy for 4 weeks.

Perspectives

Almost all injection drug users analyzed thus far with regard to short-course therapy (>95%) have been treated with a penicillinase-resistant penicillin and an aminoglycoside. Cefazolin may be less effective than a penicillinase-resistant penicillin in treating serious staphylococcal infections [21]. Staphylococci produce various β-lactamases, at least one of which can hydrolyze cefazolin at rates significantly higher than other first-generation cephalosporins and penicillinase-resistant penicillins [22]. In addition, vancomycin may not as rapidly or effectively treat patients with staphylococcal endocarditis caused by methicillin-susceptible S. aureus, which indicates that the drug, not the organism, is the critical variable [23, 24].

Gentamicin probably should be substituted for the tobramycin or amikacin that was used in two of the clinical trials just described. Of the available aminoglycoside antibiotics, it is the cheapest and most active in vitro against staphylococci. Unfortunately, little attention has been paid to aminoglycoside resistance in S. aureus and its effect on the synergistic killing that usually occurs when a penicillinase-resistant penicillin is combined with an aminoglycoside against susceptible strains [15, 16, 25-28]. Many, but not all, aminoglycoside-resistant staphylococci have also been resistant to methicillin. Whether high-level resistance to the aminoglycoside antibiotic negates the benefits of combination therapy (as in the treatment of enterococcal endocarditis) remains an unanswered and important question [29]. If the organism displays high-level resistance (minimal inhibitory concentration [MIC] > 500 µg/mL) to gentamicin [30], another aminoglycoside antibiotic to which it is susceptible might provide synergy, but the major gentamicin-modifying enzyme inactivates most other aminoglycosides [31, 32].

The clinical studies of short-course treatment implicitly advocate the close monitoring of patients during the 2 weeks of therapy and imply that evidence of complications or extravalvular infection developing during therapy should prompt an extended duration of antibiotic treatment.

The potential risks of short-course combination therapy include the theoretical possibility of a higher relapse rate than that associated with a longer duration of antibiotic treatment and an increased incidence of nephrotoxicity because of the addition of an aminoglycoside antibiotic. In the three studies described above, no relapses were identified in patients who had promptly and completely responded to therapy with a penicillinase-resistant penicillin and aminoglycoside. Because most of these patients were young and were selected because they had normal renal function and were hemodynamically stable, clinically significant aminoglycoside nephrotoxicity was not observed.

Most, but not all, patients described in the reports of abbreviated courses of antibiotic therapy were intravenous drug users with tricuspid valve endocarditis. Extrapolation of these results to patients with other right-sided sites of involvement is reasonable but less well documented.

Of the 134 patients summarized in Table 1, 4 had “bad outcomes.” These data yield a composite failure rate of 3% (95% CI, 1% to 7%.)

Conclusions

On the basis of the published reports, we can define the group of injection drug users with endocarditis who are candidates for short-course therapy with a penicillinase-resistant penicillin and gentamicin. Potential candidates for short-course therapy, who may include more than half the injection drug users admitted to some hospitals with endocarditis [19], are infected with methicillin-sensitive S. aureus that is confined to the right-sided heart valves, as indicated by clinical and perhaps echocardiographic evaluation. These patients could tolerate a penicillinase-resistant penicillin and low-dose aminoglycoside therapy. Candidates would clinically and bacteriologically respond within 96 hours of initiation of combination therapy. Most importantly, there should be no evidence of hemodynamic compromise, systemic embolic complications, or metastatic infection, either at the initiation or completion of 2 weeks of combination therapy, except that septic pulmonary emboli developing during the first few days of treatment do not constitute a contraindication to short-course therapy. Although the prognostic implications of large vegetations are still unclear, patients with echocardiographic vegetations greater than 2 cm³ might best be excluded from short-course therapy [3, 33].

At this point, neither cefazolin nor vancomycin should be substituted for a penicillinase-resistant penicillin in short-course regimens. Gentamicin is the aminoglycoside antibiotic of choice for combination therapy. On the basis of the dosing schedules used in the clinical trials [18-20] and extrapolation from the treatment of enterococcal endocarditis [34], goals for peak and trough levels of gentamicin should be approximately 3.0 and 0.5 microns/mL, respectively. If the staphylococcal isolate shows high-level gentamicin resistance (MIC > 500 µg/mL), combination therapy of any duration should probably be avoided, and the patient should receive a standard 4- to 6-week course of therapy with a β-lactam antibiotic.

The decision to discontinue treatment after only 2 weeks should not be finalized until the patient is assessed on day 14 of therapy; if any evidence suggests continuing infection or complications, therapy with an appropriate β-lactam antibiotic should be extended to at least 4 weeks.

Cautions

No comparative data exist to prove that short-course combination therapy is as safe and effective as standard regimens of longer duration, even in highly selected patients. In addition, the efficacy of a penicillinase-resistant penicillin used as single-agent therapy for 2 weeks has never been evaluated; therefore, the rationale for adding an aminoglycoside antibiotic to short-course regimens is only theoretical.

The power to exclude statistically and clinically important differences between short and standard treatment courses is limited by the relatively few patients studied and the lack of matched controls. Pending further studies (ideally randomized, controlled studies), short-course antibiotic therapy can justifiably be used for patients meeting the criteria described above. The published data, however, should not be extrapolated to patients with left-sided valvular involvement, those who are infected with organisms other than methicillin- and aminoglycoside-susceptible S. aureus, and those with complicated courses, renal insufficiency, intolerance to penicillinase-resistant penicillin or aminoglycoside antibiotics, slow responses to therapy, or any evidence of metastatic foci (outside the lungs) or persistent infection.