Mycobacterium avium complex (MAC) is an important cause of illness in patients with advanced HIV infection and is associated with shortened survival [1–4]. Although treatment of bacteremic MAC disease is recommended , optimal regimens for managing this infection have not been defined. Clarithromycin, one of the newer macrolide antibiotics, has been shown to have considerable activity against MAC disease in patients with and without HIV infection [6–8]. However, treatment of MAC bacteremia with clarithromycin alone results in rapid emergence of clarithromycin-resistant organisms resulting in recrudescent disease . Combining other antimycobacterial agents with clarithromycin will presumably prevent the development of drug resistance, but neither the number nor type of drugs with which clarithromycin should be combined is currently known .
A US Public Health Service Task Force has listed a number of agents that can be combined with macrolides for the treatment of MAC bacteremia . Two agents that have been adopted in clinical use include ethambutol, which has modest activity against MAC in vivo , and clofazimine, which was found to be interchangeable with amikacin in a multidrug regimen for MAC bacteremia . The relative contribution of either of these agents to a combination regimen including clarithromycin is unknown. The purpose of this study was to compare the safety and activity of a regimen of clarithromycin and ethambutol with or without clofazimine for treating MAC bacteremia in patients with HIV infection. We selected ethambutol as our second drug because it is more active than clofazimine against MAC , and we randomly assigned patients to receive or not receive clofazimine, to determine whether this improved clinical outcomes.
The study was conducted at 22 HIV specialty centers in the United States. Eligible patients were individuals with serologically confirmed HIV infection with at least one blood culture positive for MAC within the 4 weeks before entry. Treatment with antimycobacterial therapy was permitted for up to 14 days between the time of culture and start of study medication. Patients were excluded if they had other active opportunistic infections or malignancies that required acute therapy; if their Karnofsky performance status was less than 40; if they were judged unlikely to comply with therapy; or if there were serious derangements in baseline laboratory studies (haemoglobin <7 g/dl, platelet count <50×109/l, neutrophil count <600×106/l, alanine aminotransferase or aspartine aminotransferase >10 times the upper limit of normal). All patients gave written informed consent, and the trial was approved by the institutional review boards at each participating site.
Consenting patients had a detailed medical examination at enrollment and had blood drawn for laboratory studies. Patients were randomly assigned to one of two treatment groups by a computer-generated allocation code prepared in blocks of 100. Patients were treated with clarithromycin 500 mg twice daily and ethambutol (800 mg daily for patients weighing <60 kg and 1000 mg daily for patients weighing ≥ 60 kg) or with the same doses of clarithromycin and ethambutol with the addition of clofazimine 100 mg daily. The study was not blinded.
Patients were evaluated monthly for clinical response, adverse reactions and laboratory studies. A detailed history of signs and symptoms was taken at each visit. Patients who were withdrawn from the study were asked to return for clinical evaluation within 48 h. The planned duration of the study was 24 weeks, after which patients were permitted to continue on their assigned treatment at the discretion of the investigator. During this follow-up phase, patients were evaluated every 3 months. All patients were followed for survival.
Blood was obtained for mycobacterial culture twice before entry and monthly for 6 months. Specimens of blood (10 ml) were placed in vacutainer tubes containing sodium polyanetholesulfate (SPS) and shipped overnight at ambient temperature to a central laboratory (Microbiology Specialists, Inc., Houston, Texas, USA). Specimens were mixed with a 3.3% sodium deoxycholate and 0.35% SPS lysing solution. After centrifugation, the supernatant was decanted and the pellet resuspended in 0.2% bovine serum albumin to a volume of 2.0–2.5 ml. Two 0.5 ml samples were inoculated onto Middlebrook 7H10 agar plates and one sample was placed into a BACTEC 7H12 broth vial (Becton-Dickinson, Sparks, Maryland, USA). An additional 0.5 ml aliquot was diluted with 7H9 broth to yield a 1:100 dilution and was then inoculated onto a Middlebrook 7H10 plate. The agar plates were incubated in 8–10% CO2 at 35°C for 6 weeks and the BACTEC vials were incubated at 35°C for 2–4 weeks or until the growth index was >500. Speciation of mycobacteria recovered in BACTEC culture was performed with hybridized DNA-RNA probes (Gen Probe, San Diego, California, USA). Testing of isolates for susceptibility to clarithromycin, ethambutol and clofazimine was performed in BACTEC 7H12 broth by the method of Heifets et al. . Colony counts per millilitre of blood were calculated by quantifying the number of colonies present on the agar plates.
Data were analysed on an intention-to-treat basis for safety and survival. Microbiological data were analysed for patients with a positive blood culture for MAC at study enrollment. Differences between treatment groups in categorical variables at enrollment and follow-up were assessed by the Mantel-Haenzel χ2 test or Fisher's exact test. Continuous variables were compared with Student's t test or the Wilcoxon rank-sum test for non-normally distributed variables. Time to event after study entry analyses were conducted for response to treatment and survival using Kaplan-Meier survival curve methodology and the log-rank test for comparison between the treatment groups. A Cox proportional hazards analysis using death as the dependent variable was performed to assess the association of treatment and other clinical features with survival. Two-tailed P values at the 0.05 level were used for all comparisons.
One hundred and six patients with documented MAC bacteremia were enrolled in the study between February 1993 and March 1994; 89 (84%) patients had a positive blood culture on the day of enrollment. All 106 patients were studied for clinical response and survival; the 17 patients with negative enrollment blood cultures were excluded from microbiological analyses. Table 1 shows the baseline demographic and clinical characteristics of the study patients. The patients were predominantly men and 42% were racial or ethnic minority group members. All patients had advanced immunodeficiency, with median CD4 lymphocyte counts of 10 × 106/l. At entry, 93% of patients complained of fevers, 78% had night sweats, 91% had experienced weight loss and 40% had diarrhea. The duration of symptoms and bacteremia were similar in both groups. Patients assigned to receive clofazimine had a significantly higher median weight than patients assigned to two-drug therapy. There were no significant differences between groups in prior AIDS-related conditions or baseline hematologic or serum chemistry indices. Sixteen per cent of patients assigned to two drugs and 14% of patients assigned to three drugs had received rifabutin prophylaxis of MAC.
The median number of c.f.u. of MAC at entry was 152 c.f.u./ml blood (2.2 logs) for patients assigned to two drugs and 1907 c.f.u./ml blood (3.3 logs) for patients assigned to three drugs (P < 0.001). Significant reductions in mycobacterial colony counts in blood were achieved with both regimens at 4 weeks and thereafter (Table 2). The magnitude of decline was significantly greater at all visits except week 8 for patients assigned to receive clofazimine, probably reflecting higher baseline colony counts. At the last visit for which a culture was available, the median decline from baseline was 1.8 logs for patients assigned to two drugs and 2.3 logs for patients assigned to three drugs (not significant). However, the mean percentage reduction in the number of c.f.u. in blood did not differ significantly between groups.
During the course of treatment, 65% of patients receiving two drugs and 54% of those receiving three drugs had negative blood cultures (not significant). The median time to a negative culture was 58 days for patients receiving two drugs and 63 days for patients receiving three drugs (P = 0.40, log-rank). Following conversion of blood cultures to negative, five out of 31 patients (16%) assigned to the two-drug regimen and five out of 22 patients (23%) receiving the three-drug regimen had a relapse of bacteremia (not significant).
The 90-day Kaplan-Meier probability of relapse of bacteremia after blood cultures turned negative was 0.20 for patients assigned to two drugs and 0.15 for patients assigned to three drugs (not significant).
Resistance to clarithromycin (minimum inhibitory concentration >8 μg/ml) was found in the baseline MAC isolate of one patient assigned to the three-drug arm; this patient did not respond to therapy. All other baseline isolates were susceptible to clarithromycin. Resistance to clarithromycin developed during therapy in one patient receiving two drugs, although this individual had a decrease in mycobacterial load with treatment.
Symptoms of MAC disease responded promptly to combination therapy (Table 3). At any time after initiating therapy, fevers had resolved or improved in 89% of patients receiving two drugs and 86% of patients receiving three drugs, whereas night sweats had improved or resolved in 87 and 84% of patients, respectively. Karnofsky performance scores remained stable during the study, with no clinically significant differences between the two groups.
Adverse reactions to study medications were common. Overall, 36% (20 out of 55) of patients on the two-drug regimen had an adverse event attributed to study medication, whereas 33% (17 out of 51) of patients receiving the three-drug regimen had adverse events. The most common side-effects of medication were nausea, vomiting and abdominal pain. Gastrointestinal associated adverse events occurred in 27% of patients receiving two drugs and 24% of patients receiving three drugs (P = 0.84). Six patients (11%) receiving two drugs and seven patients (14%) receiving three drugs had treatment stopped because of adverse drug reactions suggestive of lack of tolerance. The mean duration of on-study therapy was 128 days for patients in the two-drug arm and 105 days in the three-drug arm (P = 0.22).
During the study, 21 out of 55 patients (38%) assigned to the two-drug regimen died, whereas 31 out of 51 patients assigned (61%) to the three-drug regimen died (P = 0.032). No deaths were attributed to study medications. Figure 1 shows the time to death, by Kaplan-Meier analysis. The difference in survival time between the two treatment groups was significant by log-rank test (P = 0.012). Median survival was 199 days for patients with baseline MAC colony counts ≥ 1000 c.f.u./ml, 303 days for patients with colony counts between 100 and 999 c.f.u./ml, and >465 days for patients with colony counts <100 c.f.u./ml (P = 0.01, lowest versus highest group, log-rank). Stratification of patients by baseline level of bacteremia showed that patients receiving clofazimine had significantly higher mortality than those who did not receive the drug when the baseline number of c.f.u./ml blood was between 100 and 1000 (Table 4). Patients who continued antiretroviral therapy after enrollment had longer survival (median, >465 days) than those who did not (median, 201 days; P = 0.07). In a Cox regression analysis that included baseline hematocrit, CD4 count, Karnofsky score, antiretroviral therapy, and baseline MAC colony counts, a higher relative hazard (RH) of death was associated with assignment to clofazimine [RH, 1.79; 95% confidence interval (CI), 1.02–3.17], baseline colony count ≥ 1000 c.f.u./ml (RH, 1.95; 95% CI, 1.02–3.73) and not receiving antiretroviral therapy after entry (RH, 1.63; 95% CI, 0.94–2.81).
This study shows that the combination of clarithromycin and ethambutol was effective in treating disseminated MAC bacteremia in patients with HIV infection, and that the addition of clofazimine to this regimen did not improve outcomes. Rates of culture conversion, symptom resolution and emergence of drug-resistant disease were all similar in the two treatment groups. In addition, patients receiving clofazimine as well as clarithromycin and ethambutol had significantly greater mortality than patients treated with only clarithromycin and ethambutol. Despite randomization, however, patients receiving three drugs had significantly higher baseline mycobacterial colony counts. This imbalance in assignment to treatment arms was due to chance. Patients were well matched with respect to other clinical and laboratory parameters. Moreover, because colony counts were not determined until several weeks after randomization, bias in assigning patients to treatment arms could not have occurred.
Several previous studies have shown that clarithromycin is very active in treating MAC bacteremia in patients with HIV infection [6, 7]. When the drug is administered alone, however, high rates of relapse occur with clarithromycin-resistant organisms . When combination therapy is begun after 6–8 weeks of clarithromycin monotherapy, drug resistance still occurs in 20–40% of patients . We found that the combination of ethambutol with clarithromycin and ethambutol was highly effective in preventing the emergence of clarithromycin resistance. Only one patient treated with either combination regimen in this trial developed drug resistance during treatment. This underscores the importance of combination therapy for MAC disease and demonstrates that ethambutol alone is effective for inhibiting clarithromycin-resistant clones. Only one patient had an initial isolate that was resistant to clarithromycin. This suggests that the use of a two-drug regimen to treat MAC bacteremia in patients with HIV infection is very unlikely to promote emergence of drug-resistant strains. Dautzenberg et al. have previously reported higher rates of clarithromycin resistance during combination therapy . In that study, however, all patients were initially treated with clarithromycin alone, and other drugs were added subsequently. Patients in our trial received combination therapy from the outset. If clarithromycin and azithromycin resistance is present at the start of therapy as a result of the use of these macrolides for preventive therapy, then initial therapy with at least two other active agents would be necessary [14,15].
The role of clofazimine in therapy of MAC has been uncertain. Early studies in patients with AIDS suggested that this agent had little activity [16,17]. When clofazimine was used in a multidrug regimen instead of amikacin, a drug with appreciable in vitro activity against MAC , results comparable to an amikacin-containing regimen were seen [11,19]. However, when clofazimine was administered alone to patients with MAC bacteremia, no change in mycobacterial load was seen over 2 weeks . In contrast, ethambutol produced a 0.5 log decline in mycobacteremia during the same period . Our data suggest that clofazimine does not contribute to antimycobacterial activity when combined with clarithromycin and ethambutol. These findings are consistent with several other recent trials that suggest little antimycobacterial effect from clofazimine [20, 21].
The higher mortality seen in patients treated with the three-drug regimen is a serious concern. The proportion of patients receiving clofazimine who were withdrawn from therapy for adverse events was similar to that for patients only receiving only clarithromycin and ethambutol. Moreover, no patient was felt by the investigators to have died of drug toxicity. It is possible, however, that unappreciated drug toxicity from clofazimine could have contributed to the higher mortality. Patients assigned to the three-drug regimen, however, had a 1 log greater median number of organisms in their blood at entry than patients assigned to two drugs, despite random assignment to treatments. Horsburgh et al.  have shown that higher blood mycobacterial load is associated with higher mortality in patients with AIDS. In addition, Kemper et al. have reported that patients with transient MAC bacteremia had better survival than patients with sustained and higher-level mycobacteremia . Our multivariate analysis demonstrated independent effects of high-grade mycobacteremia and assignment to clofazimine on survival. Given the lack of evidence that clofazimine contributes to clinical responses, the survival difference found in this study is a compelling reason for not using clofazimine in the therapy of HIV-related MAC disease.
Despite the effectiveness of several agents in preventing MAC disease in patients with HIV [14, 24, 25], a considerable proportion of patients with advanced HIV disease continue to develop bacteremia with this organism . Effective, well-tolerated treatment regimens for MAC infections are therefore still required. Our data suggest that the combination of clarithromycin and ethambutol is clinically and microbiologically active, well tolerated, and associated with lower toxicity and mortality than the same agents plus clofazimine. It is not yet known whether the addition of other agents more active than clofazimine, such as rifabutin, would improve outcomes further [27, 28]. A recent study by Shafran et al. showed efficacy of a regimen of clarithromycin, rifabutin and ethambutol for the treatment of MAC bacteremia in patients with HIV infection . Patients treated with this regimen had better microbiological responses and survival than patients treated with rifampin, ethambutol, clofazimine and ciprofloxacin. These investigators also found that a higher dose of rifabutin was associated with improved clearance of MAC from the blood, albeit with more toxicity. The proportion of patients in that study treated with clarithromycin, ethambutol and rifabutin who had sterilization of blood cultures (69%), however, was similar to that seen in patients receiving clarithromycin and ethambutol in our study (65%). Additional studies to assess the contribution of rifabutin to multidrug regimens for MAC bacteremia in patients with AIDS are under way. At present, we suggest that clarithromycin and ethambutol be considered as an effective therapy for disseminated MAC infection in patients with HIV infection.
The authors thank the patients who volunteered for this study, and the physicians and other health care workers who referred and provided care for them.
1. Horsburgh Jr CR: Mycobacterium avium-complex infection in the acquired immunodeficiency syndrome
. N Engl J Med
2. Hoover DR, Saah AJ, Bacellar H, et al.
: Clinical manifestations of AIDS in the era of Pneumocystis prophylaxis
. N Engl J Med
3. Chaisson RE, Keruly J, Richman DD, Moore RD: Incidence and natural history of Mycobacterium avium-complex disease in advanced HIV infection treated with zidovudine
. Am Rev Respir Dis
4. Chin DP, Hopewell PC, Yajko DM, et al.
: Mycobacterium avium-complex in the respiratory or gastrointestinal tract and the risk of M. avium-complex bacteremia in patients with human immunodeficiency virus infection
. J Infect Dis
5. US Public Health Service Task Force on Prophylaxis and Therapy for Mycobacterium avium
Complex: Recommendations on prophylaxis and therapy for disseminated Mycobacterium aviumcomplex for adults and adolescents infected with human immunodeficiency virus
6. Dautzenberg B, Truffot C, Legris S, et al.
: Activity of clarithromycin against Mycobacterium avium infection in patients with the acquired immune deficiency syndrome
. Am Rev Respir Dis
7. Chaisson RE, Benson CA, Dube MP, et al.
: Clarithromycin therapy for bacteremic Mycobacterium avium complex disease: a randomized, double-blind, dose-ranging study in patients with AIDS
. Ann Intern Med
8. Wallace RJ, Brown BA, Griffith DE, et al.
: Initial clarithromycin monotherapy for Mycobacterial avium-intracellulare complex lung disease
. Am Rev Respir Dis
9. Chin DP, Hopewell PC: How to treat bacteremicMycobacterium aviumcomplex disease
10. Kemper CA, Havlir D, Haghighat D, et al.
: The individual micro-biologic effect of three antimycobacterial agents, clofazimine, ethambutol, and rifampin, onMycobacterium aviumcomplex bacteremia in patients with AIDS
. J Infect Dis
11. Kemper CA, Meng T-C, Nussbaum J, et al.
: Treatment of Mycobacterium avium complex bacteremia in AIDS with a four-drug oral regimen: rifampin, ethambutol, clofazimine, and ciprofloxacin
. Ann Intern Med
12. Heifets LB, Lindholm-Levy PJ, Comstock RD: Clarithromycin minimal inhibitory and bactericidal concentrations againstMycobacterium avium
. Am Rev Respir Dis
13. Dautzenberg B, Saint Marc T, Meyohas MC, et al.
: Clarithromycin and other antimicrobial agents in the treatment of disseminatedMycobacterium aviuminfections in patients with acquire immunodeficiency syndrome
. Arch Intern Med
14. Pierce M, Lamarca A, Jablonowski H, et al.
: A randomized trial of clarithromycin as prophylaxis against disseminatedMycobacterium aviumcomplex infection in patients with advanced acquired immunodeficiency syndrome
. N Engl J Med
15. Centers for Disease Control and Prevention: USPHS/IDSA guidelines for the prevention of opportunistic infections in persons infected with human immunodeficiency virus: a summary
16. Masur H, Tuazon C, Gill V, et al.
: Effect of combined clofazimine and ansamycin therapy on Mycobacterium avium- Mycobacterium intracellulare bacteremia in patients with AIDS
. J Infect Dis
17. Agins BD, Berman DS, Spicehandler D, et al.
: Effect of combined therapy with ansamycin, clofazimine, ethambutol, and isoniazid forMycobacterium aviuminfection in AIDS
. J Infect Dis
18. Bermudez LE, Young LS: Activities of amikacin, roxithromycin, and azithromycin alone or in combination with tumor necrosis factor againstMycobacterium aviumcomplex
. Antimicrob Agents Chemother
19. Chiu J, Nussbaum J, Bozzette S, et al.
: Treatment of disseminatedMycobacterium aviumcomplex infection in AIDS with amikacin, ethambutol, rifampin, and ciprofloxacin
. Ann Intern Med
20. May T, Brel F, Beuscart C, et al.
: French randomized open trial of two clarithromycin combination therapies forMycobacterium aviumcomplex bacteremia: first results
. 35th Interscience Conference on Antimicrobial Agents and Chemotherapy
. San Francisco, September 1995 [abstract LB-19].
21. Dube M, Sattler F, Torriani F, et al.
: Prevention of relapse of MAC bacteremia in AIDS: a randomized study of clarithromycin plus clofazimine, with or without ethambutol
. Third Conference on Retroviruses and Opportunistic Infections
. Washington, DC, January-February 1996 [abstract 206].
22. Horsburgh CR, Metchock B, Gordon SM, et al.
: Predictors of survival in patients with AIDS and disseminatedMycobacterium aviumcomplex disease
. J Infect Dis
23. Kemper CA, Havlir D, Bartok AE, et al.
: Transient bacteremia due toMycobacterium aviumcomplex in patients with AIDS
. J Infect Dis 1994, 170
24. Havlir DV, McCutchan JA, Bozzette SA, et al.
: A double-blind randomized study of weekly azithromycin, daily rifabutin, and combination azithromycin and rifabutin for the prevention ofMycobacterium aviumcomplex in AIDS patients
. N Engl J Med
25. Nightingale SD, Cameron DW, Gordin FM, et al.
: Two controlled trials of rifabutin prophylaxis against Mycobacterium avium-complex infection in AIDS
. N Engl J Med
26. Moore RD, Chaisson RE: The natural history of opportunistic disease in an urban HIV clinical cohort
. Ann Intern Med
27. Hoy J, Mijch A, Sandland M, et al.
: Quadruple-drug therapy for Mycobacterium avium-intracellulare bacteremia in AIDS patients
. J Infect Dis
Members of the M93–069 Study Investigators included Ian Baird (Columbus, Ohio); Constance Benson (Rush Presbyterian/St Luke's Medical Center, Chicago, Illinois); Jack Bissett (Austin Infectious Disease Consultants, Austin, Texas); Richard E. Chaisson (Johns Hopkins University School of Medicine, Baltimore, Maryland); W. Jeffrey Fessel (Kaiser Permanente Medical Center, San Francisco, California); Jeffrey Galpin (Shared Medical Research Foundation, Tarzana, California); Hermes Garcia (Caribbean Medical Research Network, Inc., Col. Metro Guaynabo, Puerto Rico); Stephen Hauptman (Thomas Jefferson University Hospital, Philadelphia, Pennsylvania); Wilbert Jordan (Paramount, California); Philip Keiser (University of Texas Southwestern Medical Center, Dallas, Texas); Carol Kemper (Santa Clara Valley Hospital, San Jose, California); Thomas Klein (Klein and Slotten Medical Associates, Chicago, Illinois); Christopher Lahart (Veterans Affairs Medical Center, Houston, Texas); Craig Lindquist (Marin County Specialty Clinic, Greenbrae, California); Donna Mildvan (Beth Israel Medical Center, New York, New York); Mark Pierce (Vanderbilt University, Nashville, Tennessee); Joel Ruskin (Kaiser Permanente Medical Center, Los Angeles, California); William B. Smith (LCRC, New Orleans, Louisiana); Corklin Steinhart (Miami, Florida); Charles van der Horst (University of North Carolina, Chapel Hill, North Carolina); Ernest Wong (Lafayette, California); and Bienvenido Yangco (St Joseph's Infectious Disease Research Institute, Tampa, Florida).
28. Shafran S, Singer J, Phillips P, et al.
: Canadian randomized open-label trial of combination therapy forMycobacterium aviumcomplex bacteremia: final results
. N Engl J Med