Share this article on:

A phase II/III trial of antimicrobial therapy with or without amikacin in the treatment of disseminated Mycobacterium avium infection in HIV-infected individuals

Parenti, David M.1; Williams, Paige L.2; Hafner, Richard3; Jacobs, Michael R.4; Hojczyk, Peter5; Hooton, Thomas M.6; Barber, Thomas W.7; Simpson, Gail8; der Horst, Charles van9; Currier, Judith10; Powderly, William G.11; Limjoco, Marissa2; Ellner, Jerrold J.4AIDS Clinical Trials Group Protocol 135 Study Team

Evaluation of antiretroviral guidelines

Objective: To determine the clinical and microbiologic benefit of adding amikacin to a four-drug oral regimen for treatment of disseminated Mycobacterium avium infection in HIV-infected patients.

Design: A randomized, open-labeled, comparative trial.

Setting: Outpatient clinics.

Patients: Seventy-four patients with HIV and symptomatic bacteremic M. avium infection.

Interventions: Rifampin 10 mg/kg daily, ciprofloxacin 500 mg twice daily, clofazimine 100 mg every day, and ethambutol 15 mg/kg orally daily for 24 weeks, with or without amikacin 10 mg/kg intravenously or interamuscularly 5 days weekly for the first 4 weeks.

Main outcome measure: Clinical and microbiologic response at 4 weeks; quantitative level of bacteremia with M. avium.

Results: No difference in clinical response was noted with the addition of amikacin to the four-drug oral regimen, and only 25% in either group had a complete or partial response at 4 weeks. A comparable quantitative decrease in bacteremia was noted in both treatment groups, with 16% of patients being culture-negative at 4 weeks and 38% at 12 weeks. Toxicities were mainly gastrointestinal. Amikacin was well tolerated. Median survival was 30 weeks in both groups.

Conclusions: The addition of amikacin to a four-drug oral regimen of rifampin, ciprofloxacin, clofazimine, and ethambutol did not provide clinical or microbiologic benefit.

1George Washington University Medical Center, Washington, DC

2Department of Biostatistics, Harvard School of Public Health, Boston, Massachusetts

3Division of AIDS, National Institute of Allergy and Infectious Diseases, Belhesda, Maryland

4Case Western Reserve University and University Hospitals, Cleveland, Ohio

5Frontier Science and Research Technology Foundation, Amherst, New York

6the University of Washington, Seattle, Washington

7Boston City Hospital, Boston, Massachusetts

8Harbor/University of California, Torrance, California

9University of North Carolina, Chapel Hill, North Carolina

10Beth Israel Hospital, Boston, Massachusetts

11Washington University, St Louis, Missouri, USA.

See Appendix.

Note: Presented in part at the 33rd Interscience Conference on Antimicrobial Agents and Chemotherapy, New Orleans, October, 1993, and the Second National Conference on Human Retroviruses and Related Infections, Washington, DC, January-February 1995.

Sponsorship: Supported in part by contracts AI-25867, AI-25879, AI-27664, AI-27659, AI-27660, AI-25868, AI-25903 from the National Institute of Allergy and Infectious Diseases, National Institutes of Health, and by the General Clinical Research Center; units were funded by the National Center for Research Resources MOI RR00046.

Requests for reprints to: Dr David M. Parenti, Division of Infectious Diseases, The George Washington University Medical Center, 2150 Pennsylvania Avenue NW, Washington, DC 20037, USA.

Date of receipt: 29 July 1998; accepted: 15 September 1998.

Back to Top | Article Outline


Disseminated infection with Mycobacterium avium is a common opportunistic infection in individuals infected with HIV. Susceptible individuals have profound depression of CD4 lymphocyte counts, and the prevalence of infection at postmortem may be as high as 50-70% [1–4]. Treatment regimens are based on combination therapy for improved efficacy and to prevent the emergence of resistance. Antimicrobial susceptibility testing has been performed by a variety of techniques in vitro (broth dilution, agar dilution, macrophage culture) [5–10] and drug activity has also been evaluated in vivo, particularly utilizing the beige mouse model [11–14].

Antimicrobials with demonstrated activity against M. avium, either as single agents or in combination, include macrolides (clarithromycin, azithromycin), ethambutol, clofazimine, quinolones (ciprofloxacin, sparfloxacin), rifamycins (rifampin, rifabutin) and aminoglycosides (amikacin, liposomal gentamicin) [15]. The susceptibility of M. avium to amikacin varies depending on the assay system, with 100% susceptible by broth dilution [5–7], and 56-100% by agar dilution [5,8]. A post-antibiotic effect has also been demonstrated in vitro [10]. Synergy with amikacin in vitro has been noted for ciprofloxacin and imipenem [12], ethambutol and rifampin [16], and ethambutol and roxithromycin [6]. In the J774 macrophage assay, killing was noted for 63% of isolates when amikacin was used alone, and 100% when amikacin was combined with rifabutin and ethambutol [9]. In the beige mouse model a significant reduction in M. avium colony counts in spleen, liver and lung was demonstrated when amikacin was used alone [11,12] or in combination with clofazimine [11] or rifapentine [13]. Similar reductions were noted when amikacin was administered as a single daily dose, or in two or three divided doses [11]. Amikacin has also been shown in the beige mouse model to delay the development of macrolide resistance [14].

Several uncontrolled trials have suggested that amikacin may have use in the therapy of disseminated M. avium infection when employed in combination drug regimens [16–19]. In a trial of 17 patients conducted by the California Collaborative. Treatment Group, a mean reduction greater than 1 log10 colony-forming units (CFU)/ml blood was noted after 4 weeks of treatment with a regimen employing rifampin, ethambutol, ciprofloxacin and amikacin [20]. Amikacin was administered for the first month at a dose of 7.5 mg/kg daily. Compared with baseline, a statistically significant decrease in the number of patients reporting fever or night sweats was also noted after 4 weeks of treatment. A subsequent single arm trial with 31 patients receiving a similar oral regimen (rifampin, clofazimine, ethambutol, ciprofloxacin) without amikacin revealed comparable clinical and microbiologic responses [21]. The disadvantages of amikacin therapy include the need for parenteral administration, cost, and potential toxicity.

This study was designed as a phase II/III multicenter, randomized, open-label comparative trial of a four-drug oral regimen (rifampin, ethambutol, clofazimine and ciprofloxacin), with and without parenteral amikacin, in symptomatic HIV-infected individuals with disseminated M. avium infection. The study was conducted from 1991 to 1992, prior to the availability of macrolide therapy (clarithromycin, azithromycin).

Back to Top | Article Outline

Materials and methods


Twenty-six sites recruited patients for this study as participants in AIDS Clinical Trials Group (ACTG) protocol 135. HIV-infected patients were eligible for participation if they (i) were at least 13 years of age, (ii) had isolation of M. avium from blood, bone marrow, or tissue biopsy within 12 weeks of enrollment, (iii) had a history of fever (> 38°C) for at least 14 of the previous 30 days, and (iv) had a Karnofsky score of at least 50. Patients were excluded if they (i) had received antimycobacterial therapy (exclusive of isoniazid or pyrazinamide) within the previous 4 weeks, (ii) were receiving acute therapy for opportunistic infections, or (iii) had allergy to study medications or concurrent use of nephrotoxic agents. In addition, patients with the following laboratory criteria were excluded from the study: hemoglobin <7.0 g/dl, absolute neutrophil count <750 > 106/l, platelet count <50 < 109/l, calculated creatinine clearance <30 ml/min (Cockcroft-Gaulte equation), bilirubin >2.0 mg/dl, and aspartate/alanine aminotransferase levels more than five times normal. Those with negative blood cultures on enrollment were discontinued from the study.

Patients were randomly assigned in equal proportions to two treatment arms for a projected treatment period of 24 weeks, with follow-up to 60 weeks. Patients in both arms received rifampin 10 mg/kg orally daily (up to 600 mg daily; Marion Merrel Dow Pharmaceuticals, Kansas City, Missoruri, USA), ciprofloxacin 500 mg orally twice daily (Miles/Bayer, West Haven. Connecticut, USA), clofazimine 100 mg orally every day (CIBA-Geigy Corporation, Summi. New Jersey, USA), and ethambutol 15 mg/kg orally daily (Lederle Laboratories, Pearl River, New York, USA). Patients in the amikacin arm received amikacin 10 mg/kg intravenously or intramuscularly 5 days per week for the first 4 weeks (Bristol-Myers Squibb, Princeton, New York, USA). Patients were instructed to keep diaries collecting information on maximum daily temperatures, use of antipyretic agents, and episodes of diarrhea. In the event of adverse effects, patients were to discontinue the most likely drugs involved, with the stepwise re-addition of drugs in an order specified in the protocol. Adverse experiences were graded using the standardized Division of AIDS table for grading severity of adverse experiences.

Informed consent was obtained from the patients or their parents or guardians, and human experimentation guidelines of the US Department of Health and Human services and those of the authors' institutions were followed in the conduct of the clinical research.

Back to Top | Article Outline


Quantitative blood cultures (10 ml blood) for M. avium were performed after overnight transport to a central laboratory (University Hospitals of Cleveland Microbiology Laboratory, Cleveland, Ohio, USA), using isolator tubes (Dupont, Wampole, Massachusetts, USA). Preliminary studies and a literature report demonstrated that a 24 h delay in processing had no significant effect on colony counts [22]. Tubes were inverted 20 times to ensure through mixing. Quantitative cultures were performed by plating 0.5, 0.1 and 0.01 ml on Middlebrook 7H10 agar plates (Remel, Lenexa, Kansas, USA) in duplicate. Plates were sealed and incubated at 37°C for up to 8 weeks and examined twice weekly for growth. Bactec 13A vials were also inoculated with 5 ml of specimen, incubated at 37°C for 49 days and read twice weekly in a Bactec 460 instrument (Becton Dickinson Diagnostic Instrument Systems, Towson, Maryland, USA) for detection of 14CO2 production. Isolates were confirmed as M. avium by a gene probe specific for rRNA of M. avium-intracellulare complex organisms (Acuprobe, GenProbe Corp., San Diego, California, USA). The minimal level of bacteremia detectable was 0.5 CFU/ml, and all negative cultures were given this value. Susceptibility testing was performed by determining minimal inhibitory concentations (MIC) of agents utilizing a microdilution assay in 7HSF broth according to the method of Yajko et al. [23]. Wells containing 100 μl of medium containing serially diluted antimicrobial agents were inoculated with 105 CFU/ml of test isolates and incubated at 37°C for 14 days in humidified ambient atmosphere. Susceptibility was determined on all first and last isolates from each patient, and interpreted according to the following susceptible breakpoints: rifampin ≤ 4.0 mg/ml, ciprofloxacin ≤ 2.0 mg/ml, clofazimine ≤ 1.0 mg/ml, ethambutol ≤ 4.0 mg/ml, and amikacin ≤ 4.0 mg/ml. Breakpoints for M. avium have not been established, and those used were based in part on breakpoints suggested by Heifets [24].

After enrollment, patients were observed for 1-2 weeks during which baseline symptoms and quantitative microbiology were obtained. Quantitative cultures for M. avium were obtained at baseline (mean of two cultures) and at weeks 2, 4, 8, 12 and 24 of treatment. Interval histories, targeted clinical examinations, and hematologic and clinical chemistry testing were performed at weeks 2, 4, 8, 10, 12, 16, 20 and 24. Patients with negative baseline cultures were removed from the study at 6 weeks but were followed for survival until study closure.

Back to Top | Article Outline

Response criteria

Stringent criteria were utilized for evaluation of clinical and microbiologic responses. The major endpoints were (i) improvement in clinical symptoms associated with M. avium infection, as reflected by decreases in the proportion of days with fever, diarrhea, and antipyretic use, and (ii) resolution or reduction of bacteremia, as reflected by decreases in quantitative blood cultures. For the purposes of statstical design, the primary endpoint was designated to be the proportion of patients demonstrating a complete response to therapy after 4 weeks of therapy, and a secondary endpoint was evaluated at 12 weeks.

A complete clinical response was defined as resolution of fever, antipyretic use, and diarrhea for at least 4 weeks. A partial response was defined as a decrease of at least 50% in the proportion of days with fever since the last visit compared with baseline, in combination with a greater than 50% decrease in proportion of days of either diarrhea or antipyretic use. Responses were graded as inadequate if they did not qualify as a complete or partial response, or unknown if clinical symptoms required for classification of response were not collected.

A complete microbiologic response was classified as two consecutive negative cultures at least 1 week apart. A partial response was defined as at least 1 log10 decrease in bacteremia (CFU/ml) compared with baseline in two consecutive cultures more than 1 week apart. Responses were graded as inadequate if they did not qualify as a complete or partial response, or unknown if cultures required for classification of microbiologic response were not collected or were unavailable.

One of the secondary objectives of the trial was to assess the durability of clinical and microbiological responses by evaluating the number of patients who relapsed after achieving a complete or partial response. Clinical relapse was defined as either (i) a complete clinical response and then a recrudescence of any of the three symptoms of fever, diarrhea, or antipyretic use, or (ii) a partial response and then at least 50% increase in the proportion of febrile days or days of antipyretic use since the last clinic visit. Microbiological relapse was defined as the re-emergence of positive cultures following a complete microbiological response or an increase of at least 1 log10 CFU/ml compared with the previous blood culture following a partial microbiological response.

Back to Top | Article Outline

Statistical analysis

The trial was designed to detect a 35% benefit in the clinical response rates (50% for the non-amikacin arm and 85% for the amikacin arm) with 80% power based on a one-sided significance level of α = 0.05. A one-sided significance level was considered appropriate since amikacin would only be considered as an addition to the four-drug regimen if the response was beneficial, in view of its added expense and inconvenience. the trial was originally designed to enroll 60 patients with at least one positive culture at baseline; however, due to the high death rate among patients in this population, the target sample size was increased to 90 patients in March 1992. Statistical comparisons were performed on an intent-to-treat basis using χ2 tests. Mantel-Haenszel corrected χ2 tests, Wilcoxon rank-sum tests, and Fisher's exact tests. Log-rank tests are also used to compare survival distributions and times to treatment modification or discontinuation.

Back to Top | Article Outline


Seventy-nine patients were accrued from 26 ACTG sites over a 17-month period from April 1991 to September 1992, with 37 patients randomized to each arm. Patients were observed during the baseline observation period for a median 9.5 days (range, 3-40 days). All but two patients (95%) had at least 6 days of pre-entry observation, and all but four (90%) had at least 7 days. Baseline demographics, health status, and laboratory values were comparable in the two treatment groups (Table 1). The treatment groups were balanced with respect to baseline susceptibility to all agents. Eleven patients were receiving study drug at the time of randomization: eight received ciprofloxacin, three received ethambutol, and three received rifampin.

Table 1

Table 1

The mean duration of treatment was 12.1 days (median, 10 days), and there was no difference in distribution between the amikacin and non-amikacin groups. The median baseline Karnofsky score was 70 and the median absolute CD4 cell count was 10 × 106/1, reflecting the advanced nature of HIV infection in this population. The proportion of patients with selected symptoms during the baseline period was 74% for fever, 68% for antipyretic use, and 30% for diarrhea. The mean level of M. avium in peripheral blood cultures at baseline was 1.75 ± 0.18 log10 CFU/ml for the non-amikacin group and 2.11 ± 0.20 log10 CFU/ml for the amikacin group (P = 0.19).

Of the 79 patients enrolled in this trial, five were never randomized because of patient request (n = 2), initiation of clarithromycin (n = 1), severe illness precluding completion of observation period (n = 1), and inability to maintain an accurate diary during the observation period (n = 1). Thirty-seven patients were randomized to each arm, although two randomized patients withdrew prior to receiving treatment, one refusing amikacin infusion and the other refusing all further contact. Two randomized patients were withdrawn when screening cultures (pre-observation/baseline) failed to reveal M. avium. Attrition was considerable during the course of the study with only 17 patients completing 24 weeks of treatment, and 11 completing the 60-week study period. Thirty-five (47%) of the patients died during the study period. Reasons for permanent discontinuation of study drugs included death (n = 5), treatment toxicity (n = 7), clinical endpoint reached (n = 6), toxicity/patient request (n = 14), patient request (n = 14), and investigator request (n = 6). Those listed as toxicity/patient request were those that refused stepwise re-addition of medication after temporary discontinuation due to toxicity. Some patients discontinuing therapy due to patient request or investigator request did so to pursue macrolide therapy, which became available during the course of this trial.

The time to permanent drug discontinuation was similar for both arms, and there were no apparent differences for individual agents. In general, when one drug was permanently discontinued, all were discontinued. Toxicities were primarily gastrointestinal (nausea, vomiting, abdominal pain, diarrhea); 24% of these gastrointestinal toxicities were grade 3, and none were grade 4. Fourteen (40%) out of 35 patients receiving amikacin had dosage reductions primarily due to changes in weight or decreases in the calculated creatinine clearance during therapy. Only five patients required permanent discontinuation of amikacin. There were no differences in hematologic, gastrointestinal, or renal toxicities, and no grade 3 or 4 toxicities due to amikacin. There were no reported instances of ototoxicity, although this was not formally tested.

One of the primary endpoints of the trial was the proportion of complete clinical responders after 4 weeks of treatment based on an intent-to-treat analysis. Although the proportion of febrile days decreased from 74% at baseline to 44% at week 4 (a 41% decrease), only one patient had a complete response at this timepoint (Table 2). Approximately 25% had a partial clinical response at 4 weeks and there was no difference in the proportion of complete or partial responders between treatment groups. If those patients with unknown responses or those who were not evaluated at this timepoint were excluded from the analysis, 34% would be classified as having had a complete or partial response at 4 weeks (n = 59), 62% at 8 weeks (n = 42), and 59% at 12 weeks (n = 34); however, the number of evaluable patients still receiving treatment decreased at each timepoint. There were no statistically significant differences between treatment groups at any of these three timepoints. Median survival was 30 weeks in both groups (log-rank test, P = 0.83).

Table 2

Table 2

M. avium was recovered from 69 patients at baseline, with a mean 1.93 log10 CFU/ml blood (range, 0-3.7 log10 CFU/ml; Table 3, Fig. 1), with quantitative bacteremia slightly higher in the amikacin arm. The mean levels for weeks 0, 4, and 12 are shown by treatment arm in Table 3 and Fig. 1. There was no difference between treatment arms in the mean decrease in bacteremia (Table 3): at 4 weeks the mean decrease was 0.56 log1 CFU/ml in the amikacin group compared with 0.75 log10 CFU/ml for the non-amikacin group (P = 0.957), and at 12 weeks the mean decrease was 1.47 log1 CFU/ml in the amikacin group versus 1.25 log10 CFU/ml in the non-amikacin group (P = 0.604). A similar proportion of complete or partial microbiologic responders was noted in both treatment groups (intent-to-treat; P = 0.812, Fisher's exact test); however, only 16% were culture-negative at 4 weeks, and 38% at 12 weeks. At 4 weeks, 3% were classified as complete microbiologic responders and 27% as partial responders. If those patients with unknown responses or those who were not evaluated at this timepoint were excluded from the analysis, 48% would be classified as having had a complete or partial microbiologic response at 4 weeks, and 90% at 12 weeks; however, again the number of evaluable patients decreased at each timepoint.

Fig. 1

Fig. 1

Table 3

Table 3

The concordance between clinical responses and microbiologic response was only 50%. At 4 weeks, only seven (44%) out of 16 patients with a complete or partial clinical response had a complete or partial microbiologic response, and only seven (37%) out of 19 patients with a complete or partial microbiologic response had a complete or partial clinical response. At 8 weeks, clinical relapse was noted in 18% of patients with complete or partial responses at week 4 (Fisher's exact test, P = 0.665). After 12 weeks, clinical relapses were apparent in 19% of those with complete or partial responses at week 8, with no statistically significant differences between treatment arms (Fisher's exact test, P = 0.455). There were no microbiologic relapses.

Susceptibility of isolates was assessed by determination of the MIC at which 50 and 90% of isolates were inhibited (MIC50, MIC90), percentages of isolates susceptible at breakpoint concentrations of each agent, and change in susceptibility during therapy. MIC50/MIC90 for 65 initial isolates were as follows: rifampin 8/> 16, ethambutol 8/16, clofazimine 0.5/1.0, ciprofloxacin 4/16, and amikacin 8/16. Almost all isolates were susceptible to clofazimine (97%), whereas fewer isolates were susceptible to the other agents: ethambutol 22%, amikacin 34%, ciprofloxacin 37%, and rifampin 60%. MIC50 and MIC90 were the same for 29 pairs of first and last isolates, although MIC of last isolates were higher than those of the first isolates in 3-24% of cases. Interestingly, more last isolates were susceptible to amikacin, whereas susceptibility rates for ciprofloxacin, clofazimine, rifampin and ethambutol remained relatively constant. The proportion of patients with a clinical or microbiologic response after 4 weeks of treatment was comparable between patients with baseline isolates resistant to amikacin versus those with isolates susceptible to amikacin.

Back to Top | Article Outline


Treatment of disseminated M. avium infection has evolved considerably since the beginning of the HIV epidemic. Initially, treatment with multiple oral agents was felt to be only minimally effective and associated with poor patient tolerability. It was in this milieu that amikacin was advocated as being potentially beneficial [16–20]. Subsequent clinical and laboratory research has identified the macrolides clarithromycin and azithromycin as important therapeutic agents in this infection. ACTG 135 was conducted in the pre-macrolide era, and was designed and powered to determine whether amikacin, when added to a multidrug oral regimen, provided significant clinical or microbiologic benefit given its cost, difficulty in administration, and potential toxicity.

The addition of amikacin as initial therapy to a four-drug oral regimen of rifampin, ciprofloxacin, ethambutol and clofazimine did not result in significant improvement in clinical response as measured by a proportional change in fever, antipyretic use, or diarrhea. A complete or partial clinical response was noted at 4 weeks in only 27% of patients, although a 41% decrease in the proportion of febrile days was noted at 4 weeks. Previous studies have reported defervescence at 4 weeks in 71-73% of patients receiving combination oral regimens similar to that employed in this study [20,21], and in 43-81% with clarithromycin therapy [25]. Only symptomatic patients were included in this trial compared with 84-93% [20,21] in previous studies. The lower rate of symptomatic improvement in this study may be related to differences in how baseline symptoms were collected and clinical responses were measured. Patient populations in these studies were comparable in terms of CD4 cell count and baseline quantitative bacteremia. We did not test whether the addition of amikacin to a failing regimen or during relapse would be useful.

A complete or partial microbiologic response was noted at 4 weeks in only 30% of enrolled patients on an intent-to-treat basis (48% if non-evaluable/unknown patients were excluded from the analysis). A mean decrease in bacteremia of 0.65 log10 CFU/ml was noted at 4 weeks, and only 16% of cultures were negative at 4 weeks, with 38% negative at 12 weeks. There was again no benefit from the addition of amikacin to the four-drug oral regimen. The utilization of a non-macrolide-containing regimen in this study should have favored the demonstration of a favorable amikacin effect. Both arms contained clofazimine, which has shown no efficacy in recent therapeutic and prophylactic clinical trials [26,27]. Although initial isolates were only variably susceptible by broth microdilution assay to the agents employed in this trial, the role of antimicrobial susceptibility in predicting clinical or microbiologic outcome could not be determined by this study.

Previous studies of combination oral regimens such as that employed in this study have reported decreases in mean quantitative bacteremia at 4 weeks of 1.4-1.5 log10 CFU/ml [20,21] and culture negativity rates of 17% [21]. The reason for the difference in bacteremia reduction was not readily apparent. Baseline bacteremia was 2.1-2.7 log10 CFU/ml, slightly higher than the 1.93 log10 CFU/ml noted in this study. Mean values were calculated using a limit of detectability of mycobacteremia of 0.5 log10 CFU/ml (therefore 0.5 log10 CFU/ml was the value of a negative culture), whereas other studies have used 0.1 log10 CFU/ml [21]; however, this should have lowered the mean decrease only slightly because only 16% of cultures were negative. Clarithromycin-containing regimens have resulted in decreases in mean bacteremia of 1.5-2.5 log10 CFU/ml, and 15-35% sterilization of blood cultures at 4 weeks [25,28].

It is interesting that only a 50% concordance could be demonstrated between clinical and microbiologic responses. In the dose-ranging study of clarithromycin therapy, concordance was higher but still did not reach statistical significance [25]. Improvement in bacteremia may not be a true reflection of M. avium tissue burden. In recent studies, quantitative cultures of bone marrow do not appear to correlate closely with quantitative mycobacteremia [29], and bone-marrow cultures remained positive at 4 weeks despite clearance of bacteremia [28]. In contrast, recent autopsy data has shown a correlation between quantitative blood cultures and reticuloendothelial tissue burden [30]. These differences may be related to the obvious advanced stage of disease in the autopsy study.

As previously described, the four oral agents were poorly tolerated when used in this combination regimen, with primarily gastrointestinal toxicity. The drop-out rate of 20% at 4 weeks and 46% at 12 weeks is similar to that seen in previous trials [21,25]. This is a reflection of poor medication tolerability, the advanced stage of HIV infection, and other superimposed opportunistic infections. The median survival in both groups in this series was 30 weeks, which was not dissimilar from the 24-38 weeks reported for other prospective and retrospective studies [21,31,32], including those utilizing macrolides [25,33]. The poor prognosis despite therapy may give support to the concept of antimicrobial prophylaxis of M. avium. Given the potential toxicity of amikacin, especially ototoxicity, there is no evidence to support its use as a component of initial therapy; however, it may still be useful in treatment of patients failing macrolide therapy or in patients with relapse.

Back to Top | Article Outline


1. Macher AK, Kovacs JA, Gill V, et al.: Bacteremia due toMycobacterium avium-intracellularein the acquired immuno-deficiency syndrome. Ann Intern Med 1983, 99:782–785.
2. Hawkins CC, Gold JWM, Whimbey E, et al.: Mycobacterium aviumcomplex infections in patients with the acquired immunodeficiency syndrome. Ann Intern Med 1986, 105:184–188.
3. Reichert CM, O'Leary TJ, Levens Dl., Simrell CR, Macher AM: Autopsy pathology in the acquired immunodeficiency syndrome. Am J Pathol 1983, 112:357–382.
4. Wallace JK, Hannah JB: Mycobacterium aviumcomplex infection in patients with the acquired immunodeficiency syndrome. A clinicopathologic study. Chest 1988, 93:926–932.
5. Inderlied CB, Young LS, Yamada JK: Determination ofin vitrosusceptibility ofMycobacterium aviumcomplex isolates to antimicrobial agents by various methods. Antimicrob Agents Chemother 1987, 31:1697–1702.
6. Rastogi, Goh KS, Bryskier A: Activities of roxithromycin used alone and in combination with ethambutol, rifampin, amikacin, ofloxacin, and clofazimine againstMycobacterium avium complex. Antimicrob Agents Chemother 1994, 38:1433–1438.
7. Heifets L, Lindholm-Levy P: Comparison of bactericidal activities of streptomycin, amikacin, kanamycin, and capreomycin againstMycobacterium avium and M. tuberculosis. Antimicrob Agents Chemother 1989, 33:1298–1301.
8. Davis C, Carpenter JL, Trevino S, Koch J, Ognibene AJ: In vitrosusceptibility ofMycobacterium aviumcomplex to antibacterial agents. Diagn Microbiol Infect Dis 1987, 8:149–155.
9. Yajko DM, Nassos PS, Sanders CA, Hadley WK: Killing by antimicrobial agents of AIDS-derived strains ofMycobacterium aviumcomplex inside cells of the mouse macrophage line J774.Am Rev Respir Dis 1989, 140:1198–1203.
10. Bermudez LE, Wu M, Young LS, Inderlied CB: Postantibiotic effect of amikacin and rifapentine againstMycobacterium avoiumcomplex. J Infect Dis 1992, 166:923–926.
11. Gangadharam PRJ, Perumal VK, Podapati NR, Kesavalu L, Iseman MD: In vivo activity of amikacin alone or in combination with clofazimine, rifabutin or both against acute experimental Mycobacterium avium complex infections in beige mice. Antimicrob Agents Chemother 1988, 32:1400–1403.
12. Inderlied CB, Kolonoski PT, Wu M. Young IS: Amikacin, ciprofloxacin, and imipenem treatment for disseminatedMycobacterium aviumcomplex infection of beige mice. Antimicrob Agents Chemother 1989, 33:176–180.
13. Kolonoski PT, Wu M, Inderlied CB, Young I.S: Amikacin plus rifapentine for disseminatedMycobacteriumcomplex infection in beige mice. 26th Interscience Conference on Antimicrobial Agents and Chemotherapy. New Orleans 1986 [abstract 1103].
14. Ji B, Lounis N, Truffot-Pernol C, Grossel J: Effectiveness of various antimicrobial agents againstMycobacterium aviumcomplex in the beige mouse model. Antimicrob Agents Chemother 1994, 38:2521–2529.
15. Benson CA: Treatment of disseminated disease due to theMycobacterium aviumcomplex in patients with AIDS. Clin Infect Dis 1994, 18 (suppl 3:S237-S242.
16. Baron FJ, Young LS: Amikacin, ethambutol, and rifampin for treatment of disseminatedMycobacterium avium-intracellulareinfections in patients with acquired immune deficiency syndrome. Diagn Microbiol Infect Dis 1986, 5:215–220.
17. De Lalla F, Maserati R, Scarpellini P, et al.: Clarithromycin-ciprofloxacin-amikacin for therapy ofMycobacterium avium-Mycobacterium intracellularebacteremia in patients with AIDS. Antimicrob Agents Chemother 1992, 36:1567–1569.
18. Benson CA, Kessler HA, Pottage JC, Trenholme GM: Successful treatment of acquired immunodeficiency syndrome-relatedMycobacterium aviumcomplex disease with a multiple drug regimen including amikacin. Arch Intern Med 1991, 151:582–585.
19. Polis MA, Tuazon CU: Clues to the early diagnosis ofMycobacterium avium-intracellulareinfection in patients with acquired immunodeficiency syndrome. Arch Pathol Lab Med 1985, 109:465–466.
20. Chiu, Nussbaum J, Bozzette S, et al.: Treatment of disseminatedMycobacterium aviumcomplex infection in AIDS with amikacin, ethambutol, rifampin, and ciprofloxacin. Ann Intern Med 1990, 113:358–361.
21. Kemper CA, Meng TC, Nussbaum J, et al.: Treatment ofMycobacterium aviumcomplex bacteremia in AIDS with a four-drug oral regimen: rifampin, ethambutol, clofazimine, and ciprofloxacin. Ann Intern Med 1992, 116:466–472.
22. von Reyn CF, Hennigan S, Niemczyk S, Jacobs NJ: Effect of delays in processing on the survival ofMycobacterium avium-M. intracellularein the Isolator blood culture system. J Clin Microbiol 1991, 29:1211–1214.
23. Yajko DK, Nassos PS, Hadley WK: Broth microdilution testing of susceptibilities to 30 antimicrobial agents ofMycobacterium aviumstrains from patients with acquired immune deficiency syndrome. Antimicrob Agents Chemother 1987, 31:1579–1584.
24. Heifets LB (Ed): Dilemmas and realities in drug susceptibility testing ofM. aviumandM. intracellulareand other slowly growing nontuberculous mycobacteria. In Drug Susceptibility in the Chemotherapy of Mycobacteria Infections. Boca Raton (FL): CRC Press, Inc.; 1991:123–141.
25. Chaisson RE, Benson CA, Dube MP, et al.: Clarithromycin therapy for bacteremicMycobacterium aviumcomplex disease. Ann Intern Med 1994, 121:905–911.
26. Chaisson RE, Keiser P, Pierce M, et al.: Clarithromycin and ethambutol with or without clofazimine for the treatment of bacteremicMycobacterium aviumcomplex disease in patients with HIV infection. AIDS 1997, 11:311–317.
27. Abrams DI, Mitchell TF, Child CC, et al.: Clofazimine as prophylaxis for disseminatedMycobacterium aviumcomplex infection in AIDS. J Infect Dis 1993, 167:1459–1463.
28. DATRI 007 Study Group: Determination of impact of a standard therapy onMycobacterium aviumcomplex infection in blood and bone marrow in HIV-positive patients. Second National Conference on Human Retroviruses. Washington, DC, January-February 1995 [abstract 112].
29. DATRI 007 Study Group: Mycobacterium avium complex disease in HIV-infected patients is a uniform infection of bone marrow that does not correlate with the level of infection in blood. Second National Conference on Human Retroviruses. Washington, DC, January-February 1995 [abstract 7].
30. Torriani F, Behling C, McCutchan JA, Haubrich R, Keays L, Havlir D: DisseminatedMycobacterium aviumcomplex: correlation between blood cultures and tissue burden in AIDS patients. Second National Conference on Human Retroviruses. Washington, DC, January-February 1995 [abstract 115].
31. Horsburgh CR, Metchock B, Gorgon SM, Havlik JA, McGowan JE, Thompson SE: Predictors of survival in patients with AIDS and disseminatedMycobacterium aviumcomplex disease. J Infect Dis 1994, 170:573–577.
32. Kerlikowski KM, Katz MH, Chan AK, Perez-Stable EJ: Antimycobacterial therapy for disseminatedMycobacterium aviumcomplex infection in patients with acquired immuno-deficiency syndrome. Arch Intern Med 1992, 152:813–817.
Back to Top | Article Outline


In addition to the authors, the following institutions and investigators participated in this trial: M. Power, N. Briggs (Division of AIDS, NIAID, Bethesda, Maryland); P.E. Sax, K. McLaughlin (Harvard/Boston City Hospital, Boston, Massachusetts); A.C. Collier, M.A. Paradise, L. Corey (University of Washington, Scattle, Washington); J.T. Carey, M.M. Lederman, C. Johnson (Case Western Reserve University, Cleveland, Ohio); M. Meyers, M. Royal (Washington University, St Louis, Missouri); R. Soeiro, D. Stein, E. Jenny, B. Zingman (Albert Einstein College of Medicine, New York); M. Cameron, M. Packard (Duke University, Durham, North Carolina); P. Kloser, J.M. Oleske, J.M. Picardi, E.M. Connor (UMDNJ-New Jersey Medical School, Newark, New Jersey); W. Borkowsky, M. Minter, N. Deygoo (Bellevue Hospital Center, New York); C. Sanders, A. Gurtman, B. Simpson, D. Mildvan (Mount Sinai Medical Center, New York); G.L. Simon, Susan LeLacheur (George Washington University, Washington, DC); D.M. Mushat, J. Zachary (Tulane University, New Orleans, Louisiana); R. Murphy, C. Benson, J. Pulvirenti, N. French (Northwestern University, Chicago, Illinois); E. Telzak, M. Dickmeyer, S.E. Krown (Memorial Sloan-Kettering, New York); M. Sands, D. Naglieri-Prescod (Baystate Medical Center, Springfield, Massachusetts); F.T. Valentine, E. McCarthy (ACTG 135 Protocol Study Team).

33. Chin DP, Reingold AL, Stone EN, et al.: The impact ofMycobacterium aviumcomplex bacteremia and its treatment on survival of ALDS patients: a prospective study. J Infect Dis 1994, 170:578-584.

Mycobacterium avium; MAC; AIDS; HIV; bacteremia; amikacin

© 1998 Lippincott Williams & Wilkins, Inc.