Clarithromycin Therapy for Mycobacterium avium Complex Disease in Patients with AIDS: Potential and Problems

The quality and duration of survival for patients with human immunodeficiency virus (HIV) infection have improved substantially during the past decade. Much of this improvement has been because of better management of opportunistic infections. However, progress with Mycobacterium avium complex, a pathogen that ultimately infects 30% to 50% of patients with the acquired immunodeficiency syndrome (AIDS) in North America, has until recently been disappointing, especially in terms of prophylaxis and therapy.

The development of new rifamycin and macrolide antibiotics has increased optimism among health care providers and patients searching for remedies for disease caused by this Mycobacterium species. Rifabutin, a newer rifamycin, has been approved by the U.S. Food and Drug Administration for prophylaxis of M. avium complex based on two large, controlled trials [1]. Clarithromycin, a newer macrolide, is the only drug that has been approved to date for the acute treatment of disseminated M. avium complex disease. This approval was based largely on unpublished data (2, 3; Minutes of the Open Public Meeting of the Food and Drug Administration Antiviral Advisory Committee, 10 May 1993). For health care professionals and patients, an important question is, “Do we truly have therapy that can produce sustained clinical benefit?”

In this issue, Chaisson and colleagues [4] present some of the data that led to the approval of clarithromycin for treatment of disseminated M. avium complex infection. The data clearly show evidence for in vivo microbiological activity and for short-term clinical benefit. Are these data sufficient to establish clear-cut therapy recommendations, or, despite their promising nature, are there still gaps in our knowledge and some ominous issues that must be resolved?

Chaisson and colleagues [4] randomly assigned patients with M. avium complex bacteremia to receive twice-daily clarithromycin at doses of 500 mg, 1000 mg, and 2000 mg for 12 weeks. For each group, mycobacteremia decreased during the 12 weeks, and this decrease was accompanied by diminution of fever and sweats. These aspects of the data were quite encouraging. It appeared that the 500-mg twice daily dose cleared mycobacteremia more slowly than did higher doses.

Unfortunately, some less sanguine developments were noted. First, patients who initially showed a response to their therapy often recrudesced symptomatically and microbiologically. The median duration for attaining culture negativity for M. avium complex was 43 days at the 500-mg dose and 59 days at the 1000-mg dose. Although symptomatic improvement was frequent, it seldom lasted longer than 6 weeks (Minutes of the Open Public Meeting of the Food and Drug Administration Antiviral Advisory Committee, 10 May 1993). Breakthrough isolates were substantially less sensitive to clarithromycin than were initial isolates. Second, treatment-limiting toxicity caused by clarithromycin occurred in 23%, 20%, and 40% of patients treated with 500-mg, 1000-mg, or 2000-mg twice daily doses, respectively. Some type of adverse reaction occurred in more than 50% of patients. Lastly, and perhaps most disturbingly, patients treated with the two higher doses of clarithromycin had higher death rates than did those treated with the 500-mg dose.

It is not surprising that clarithromycin monotherapy for mycobacterial disease leads to the development of drug resistance. Initial clinical trials of streptomycin monotherapy for tuberculosis in the era before AIDS showed the high likelihood that resistance would promptly emerge, often within 3 months. Combination therapy to prevent this development of resistance has been used since the availability of para-aminosalicylic acid and then isoniazid [5-8]. Although in vitro and animal models of M. avium complex infection have shown that overall activity can be increased with combination therapy, and perhaps that the development of resistance can be decreased, this has yet to be translated into human treatment of noticeably increased effectiveness [9]. However, most of the published clinical trials in patients with this disease were done before incorporation of macrolides into treatment regimens [10-12]. Thus, it is logical to hope that combination therapy that includes a macrolide could provide substantial benefits in patients with M. avium complex infections; appropriate trials are in progress.

The product label for clarithromycin for treating M. avium complex infection states that it should be used as combination therapy and not as monotherapy. This was based on the previously shown limitations of monotherapy, including that presented in this issue [4], and on the presumption—rather than on actual data—that combination therapy would be better. By using its accelerated approval mechanism, the Food and Drug Administration was able to require as a condition for approval that the pharmaceutical sponsor do additional clinical trials to evaluate clarithromycin as part of combination therapy (21 CFR 314.500-560).

Combination therapy for M. avium complex infection, however, is not without its problems. Given the toxic effects of clarithromycin and those of other therapies commonly used to treat M. avium complex infection, patients may have trouble tolerating combination regimens with agents currently available. A recently published trial [13] of 1 month of monotherapy with ethambutol, ciprofloxacin, or rifampin had rates of discontinuation due to toxicity that ranged from 4% to 20%. In a trial of a four-drug combination regimen (ciprofloxacin, rifampin, ethambutol, and clofazimine [Lamprene; Geigy Pharmaceuticals, Ardsley, New York]), 46% of patients had to discontinue therapy with at least one drug because of toxicity [10]. Thus, when these drugs are added to clarithromycin to create 2-, 3-, or 4-drug regimens, tolerance will likely be a major problem because clarithromycin alone is associated with toxic side effects.

Drug interactions in combination regimens also pose dangers and problems that we are only beginning to appreciate. Three examples provide an introduction to some of the problems that can arise when clarithromycin, rifabutin, fluconazole, or zidovudine are used in various combinations. First, clarithromycin in high doses (1- to 2-g doses) interferes with the absorption of zidovudine by as much as 30%, potentially decreasing the efficacy of antiretroviral therapy [14]. Second, rifabutin can induce hepatic cytochrome P450 enzymes, increasing the metabolism of drugs that are metabolized by the liver [15]. For zidovudine, rifabutin can decrease the area under the curve by 30%. For clarithromycin, rifabutin can decrease mean serum levels by 60% (Wallace R. Personal communication). Third, fluconazole can increase serum levels of rifabutin [16]; the mechanism appears to be inhibition of hepatic P450 enzymes, which would increase serum levels of rifabutin, clarithromycin, or zidovudine [17].

Are these interactions a curiosity for kineticists or are they clinically important? Narang and colleagues [18] did not measure the effect of fluconazole on rifabutin drug levels, but in an intriguing retrospective analysis, they found that patients receiving rifabutin and fluconazole in a controlled trial had fewer episodes of M. avium complex bacteremia than did patients who received only rifabutin. Fluconazole had no antimycobacterial effect on the placebo arm of that trial, suggesting that the benefit, if real, might have been due to higher rifabutin levels by the mechanism indicated above.

Two recent reports [19, 20] describe uveitis associated with high doses of rifabutin. Although the mechanism for uveitis is not clear, evidence suggests that concomitant clarithromycin may be an important factor. It may be that clarithromycin slows metabolism of rifabutin, permitting unusually high levels. Rifabutin may subsequently reverse the slow metabolism, lowering clarithromycin levels and rifabutin levels. These complex interactions need more intense study.

A totally unexpected finding in the trial by Chaisson and colleagues [4] was the inverse dose relation for patient survival. This differential mortality, also noted in an open-label clinical trial with clarithromycin, was particularly striking during the first 12 weeks when patients were receiving monotherapy at their randomly assigned dose (Minutes of the Open Public Meeting of the Food and Drug Administration Antiviral Advisory Committee, 10 May 1993). The risk for death was significantly increased for those randomly assigned to either 1000 mg or 2000 mg compared with those assigned to 500 mg. Although described by Chaisson and colleagues as a survival advantage for the 500-mg group, the results are better thought of as a survival difference because it is uncertain whether survival was enhanced in the 500-mg group or worsened at the other doses. No explanation is available yet for this result, such as baseline differences between treatment groups. A review of the causes of death failed to yield any clues. Because the 500-mg twice-daily dosage regimen had the best survival results and was shown to be microbiologically active, it was the regimen chosen for marketing approval. Clearly, however, additional trials are necessary to confirm or refute this dose effect on survival.

Clarithromycin has impressive activity against M. avium complex. This drug and the related macrolide azithromycin currently form the cornerstone for the optimism that more effective prophylaxis and therapy for M. avium complex infection can be attained. Much work needs to be done, however, to adequately show that we can devise single- or multiple-drug regimens that are truly more effective, well tolerated, and safe.

• Copyright ©2004 by the American College of Physicians