Clarithromycin Therapy for Bacteremic Mycobacterium avium Complex Disease

Methods

Patients

Patients were eligible for enrollment if they had HIV infection and a blood culture that grew M. avium complex within 60 days before entry. Patients could not have received any antimycobacterial agents for at least 4 weeks before entry into the study; those with a history of severe allergic reactions to erythromycin were excluded. Patients were enrolled at six centers, three within the National Institute of Allergy and Infectious Diseases AIDS Clinical Trials Group (ACTG) and three sites funded by Abbott Laboratories. Written informed consent was obtained from all study participants. Enrollment began 1 May 1991 and was completed by 30 November 1991.

Patients were randomly assigned to treatment with one of three doses of clarithromycin (Biaxin, Abbott Laboratories, North Chicago, Illinois): 500 mg twice daily, 1000 mg twice daily, or 2000 mg twice daily. Neither patients nor study staff were aware of the dose group to which patients were assigned. Randomization was done in blocks of six and was stratified by site of enrollment. Because the safety of 2000 mg of clarithromycin twice daily in patients with HIV infection was not established when the study began, the first group of 36 patients were randomly assigned to receive 500 mg or 1000 mg twice daily. A 2:1 randomization scheme was used, with two patients assigned to the 500-mg group for each patient assigned to the 1000-mg group. After the start of the study, the safety of 2000 mg of clarithromycin twice daily was shown in a phase 1 trial [17], and the second group of 36 patients were randomly assigned in a 2:1 manner to receive clarithromycin at 2000 mg or 1000 mg twice daily. Thus, after the first 72 patients were enrolled, 24 had been randomly assigned to each dose group. The remainder of the patients were randomly assigned 1:1:1 to the three dose groups. For each dose, patients took four capsules—of which one, two, or four were clarithromycin at 500 mg—and the remainder were identical-appearing placebos.

Patients were evaluated weekly for 2 weeks, then biweekly through week 12. At each study visit, an interval history was taken, a targeted physical examination was done, and blood was collected for culture and other monitoring tests. After 12 weeks of therapy, patients having a clinical or bacteriologic response were permitted to extend their therapy for up to 1 year, with the option to add other antimycobacterial agents at the discretion of the investigator.

Mycobacteriologic Analyses

Quantitative blood cultures for mycobacteria were done twice before initiating study therapy and every 2 weeks for the 12 weeks of the study. Cultures were done in the Mycobacteriology Clinical Reference Laboratory at the National Jewish Center for Immunology and Respiratory Medicine in Denver, Colorado, by methods previously reported [18]. In brief, 5 to 8 mL of blood was obtained in previously labeled vacutainer tubes containing 1.7 mL of sodium polyanetholsulfate solution as an anticoagulant. Specimens were shipped at ambient temperature by overnight mail to Denver; shipping of mycobacterial cultures in this manner does not affect viability or quantitative results [19].

After receipt in the laboratory, cultures were mixed with 20 mL of a lysing solution that contained 0.45% sodium deoxycholate and 0.09% sodium polyanetholsulfate. After manual mixing, tubes were centrifuged at 3500 g for 30 minutes in a refrigerated centrifuge, and the supernatant was decanted. The pellet was resuspended, and 0.2% bovine albumin was added to yield a final volume of 2.5 mL. The suspension was inoculated onto two 7H11 agar plates, with 0.5 mL distributed on each plate. A third 7H11 agar plate was inoculated with 0.5 mL of the suspension diluted to 1:100. The plates were then incubated at 37 °C in the presence of 5% CO₂. The remaining 1.0 mL of suspension was inoculated into liquid 7H12 medium in a BACTEC vial (Becton Dickinson Diagnostic Instrument Systems, Sparks, Maryland), which was incubated at 37 °C.

Growth in the BACTEC vial was monitored radiometrically daily, and results were reported as positive if the growth index was more than 20. Cultures were considered negative if no growth was observed within 2 weeks in BACTEC. All culture-positive vials were incubated until the growth index was more than 500 to permit speciation and drug susceptibility testing. Speciation was done using hybridized DNA-RNA probes for M. avium and M. intracellulare (AccuProbe, Gen-Probe, San Diego, California). Quantitative colony counts were determined by counting the number of colonies of mycobacteria on the agar plates, followed by a calculation that took into account the concentration coefficient and the initial volume of each blood specimen. If colonies were too numerous to count on duplicate plates, colonies on the plate inoculated with a diluted 1:100 sample were counted. Susceptibility testing of isolates to clarithromycin was done in 7H12 broth at pH 7.4 by the dilution method of Heifets as previously described [18, 20]. Because peak serum concentrations of clarithromycin are 4 to 10 µg/mL with drug doses used in this study, we selected a minimal inhibitory concentration of 4 µg/mL as the cut-point for susceptibility and selected more than 32 µg/mL as the cut-point for resistance.

Statistical Analyses

Analyses were done to compare baseline characteristics among groups and to compare treatment outcomes within and among groups. Differences in categorical variables were assessed with the continuity-corrected chi-square test. Normally distributed continuous variables within groups were compared with the paired t-test, and skewed variables were compared with the Wilcoxon rank-sum test. Initial analyses were on an intent-to-treat basis, and all randomly assigned patients were included in safety and survival analyses.

Bacteriologic analyses were done on all patients with a positive baseline culture and at least one follow-up culture. Median values for colony-forming units (CFUs) of M. avium complex/mL of blood at baseline were compared with follow-up values within and among groups using the Kruskal-Wallis test for three groups. The time to achieve cultures that were negative for M. avium complex (sterilization) in blood was estimated using the product-limit method, with differences assessed by the log-rank test. Survival was also estimated by the product-limit method. To assess the effect of baseline characteristics and treatment on survival, Cox proportional-hazards analysis was done.

Results

Between 1 May and 30 November 1991, 154 patients were enrolled in the study, 108 at the three ACTG sites and 46 at the other sites. Baseline characteristics of the study population are shown in Table 1. Baseline demographic characteristics were similar. Mean ages of patients in the three groups ranged from 34 to 38 years, and most patients were men. Thirty-one percent of patients enrolled were members of racial or ethnic minorities. Mean weights of participants were 60 to 62 kg. The mean baseline Karnofsky performance score was lower in patients assigned to 1000 mg of clarithromycin twice daily than in patients assigned to 500 mg or 2000 mg twice daily. The median time from diagnosis of M. avium complex infection to enrollment was 84 days for patients assigned to 500 mg twice daily, 64 days for patients assigned to 1000 mg twice daily, and 59 days for patients assigned to 2000 mg twice daily.

Baseline laboratory values of patients in the study are shown in Table 2. All patients were moderately anemic. Of note, serum lactate dehydrogenase and alkaline phosphatase levels were increased across all groups, with no statistical differences among groups. Patients were profoundly immunocompromised; median CD4 cell counts were 5 to 6/mm³ in all groups.

The number of CFUs of M. avium complex/mL of blood at baseline and during follow-up are shown by dose group in Table 3. Median numbers of CFUs/mL at baseline were 456 to 610 or 2.7 to 2.8 log 10 (P > 0.2). Within 2 weeks, the number of CFUs/mL of blood markedly decreased, with a greater decrease at the two higher doses than at the dose of 500 mg twice daily (P < 0.05). At 2 weeks, blood cultures negative for M. avium complex were obtained in 11% of patients receiving 500 mg of clarithromycin twice daily, 31% of those receiving 1000 mg twice daily, and 29% of those receiving 2000 mg twice daily (P = 0.08). By week 4, a more than 2.5 log reduction in CFUs/mL of blood was seen in patients receiving doses of 1000 mg or 2000 mg twice daily, and an 1.5 log reduction was observed in patients taking 500 mg twice daily. By week 8, the median number of CFUs/mL of blood was 1 for patients receiving 500 mg of clarithromycin twice daily, 0 for patients receiving 1000 mg, and 1 for patients receiving 2000 mg of clarithromycin twice daily. The proportion of patients remaining in the study at week 8 who had negative blood cultures for M. avium complex by dose group was 41% for the 500-mg group, 53% for the 1000-mg group, and 41% for the 2000-mg group. These reductions were sustained until week 12, when statistically nonsignificant increases in median CFUs of M. avium complex/mL of blood were observed for patients in each dose group.

During 12 weeks of study, most patients continuing to receive medication became culture negative on at least one follow-up visit (Figure 1). The time until the blood was culture negative for M. avium complex was shorter in patients treated with higher doses of clarithromycin. The median time until the blood was culture negative was 29 days for patients receiving the 2000-mg dose, 43 days for patients receiving the 1000-mg dose, and 54 days for patients receiving the 500-mg dose (P = 0.04). The median time to a 10-fold (>1 log 10) decrease in M. avium complex in the blood was 15 days for patients receiving 2000 mg twice daily, 16 days for patients receiving 1000 mg twice daily, and 29 days for patients receiving 500 mg twice daily (P < 0.05). After achieving a more than 1-log reduction in mycobacteremia, the median time to a return to a less than 1-log decrease in bacteremia compared with baseline was 85 days for patients receiving 2000 mg twice daily, 111 days for patients receiving 1000 mg, and 120 days for patients receiving 500 mg.

View larger version:

• In this window
• In a new window

Figure 1. Patients received 500 mg, 1000 mg, or 2000 mg twice daily. Rows at the bottom represent the number of uncensored patients at specific time points by dose. The overall difference among the three treatment groups was = 0.04. Kaplan-Meier estimates of the probability of blood culture positivity for M. avium-complex during treatment by dose of clarithromycin.P

Symptoms of M. avium complex disease were prominent at enrollment and decreased substantially with all three doses of clarithromycin. Patients with at least a 1-log decrease in mycobacteremia were more likely to have improvement in fever and night sweats (71% improved) than those patients whose colony counts did not decrease by 1 log (44% improved; P = 0.08). More patients randomly assigned to the 2000-mg dose reported resolution of fever and night sweats at weeks 4 through 8 (P < 0.05; (Table 4).

Of 152 initial isolates of M. avium complex, 150 (99%) were susceptible to clarithromycin, with minimal inhibitory concentrations of 4.0 µg/mL or less of clarithromycin. Two patients had initial isolates with minimal inhibitory concentrations of 128 µg/mL and 512 µg/mL, respectively. During the first 12 weeks of therapy, however, increases in minimal inhibitory concentration to more than 32 µg/mL were seen in 21% of all patients. Overall, during the acute and follow-up treatment periods, 46% of all patients developed minimal inhibitory concentrations of more than 32 µg/mL. Emergence of resistance was associated with recrudescence of symptoms and increases in the number of CFUs of M. avium complex in blood in most patients from whom resistant isolates were recovered. No association was observed between clarithromycin dose and development of resistance. The earliest development of resistance was at 8 weeks, and the median time at which a resistant isolate was first identified was 16 weeks. Patients with higher baseline bacterial counts tended to develop resistance earlier than did patients with lower baseline counts. The mean time at which resistance developed was 138 days for patients with baseline counts of 10 to 99 CFUs/mL of blood compared with 95 days for patients with more than 10 000 CFUs/mL of blood at baseline (P = 0.03).

Adverse reactions to clarithromycin were reported in a large proportion of patients, with gastrointestinal symptoms (nausea, vomiting, abdominal pain, diarrhea) reported by more than 50% of patients at all doses. The frequency and severity of gastrointestinal symptoms were greater at the highest dose (Table 5). Patients receiving the 2000-mg dose were more likely to discontinue therapy because of toxicity than were those receiving the 1000-mg or 500-mg dose (40% compared with 16% or 15%; P = 0.003). The primary adverse reaction requiring discontinuation of clarithromycin at all doses was gastrointestinal toxicity, and this was most pronounced in the 2000-mg group. Two patients had a severe, dose-limiting rash that was attributed to clarithromycin. The overall proportion of patients who failed to finish 12 weeks of therapy because of adverse reactions, death, or withdrawal from the study was 58% for patients assigned to 2000 mg twice daily, 43% for patients assigned to 1000 mg twice daily, and 28% for patients assigned to 500 mg twice daily (P = 0.01).

Twenty-nine patients died during the acute therapy phase. The proportion of patients who died during the first 12 weeks of study was lower in the 500-mg group (3 of 53, 5.7%) than in the 1000-mg (13 of 51, 25%) and 2000-mg groups (14 of 50, 28%; P = 0.007). Median survival after randomization was 203 days for patients assigned to 2000 mg twice daily, 215 days for those assigned to 1000 mg twice daily, and 249 days for those assigned to 500 mg twice daily (P = 0.08, Figure 2. Most deaths in the acute treatment phase occurred in patients who had discontinued therapy with study drugs because of toxicity or disease progression. The median time from withdrawal from study to death was 7 days [range, 0 to 64 days] in the 2000-mg group, 12 days (range, 0 to 46 days) in the 1000-mg group, and 51 days (range, 0 to 65 days) in the 500-mg group. A Cox proportional-hazards analysis adjusting for baseline Karnofsky performance status showed that the relative hazard of death in patients assigned to 1000 mg twice daily compared with those assigned to 500 mg twice daily was 1.18 (95% CI, 0.09 to 2.01; P = 0.54) and that the relative hazard of death for patients assigned to 2000 mg twice daily was 2.01 (CI, 1.19 to 3.58; P = 0.009). Survival was not associated with level of bacteremia at baseline, time from first positive blood culture to enrollment, or CD4 cell count.

View larger version:

• In this window
• In a new window

Figure 2. Patients received 500 mg, 1000 mg, or 2000 mg twice daily. Rows at the bottom represent the number of uncensored patients at specific time points by dose. The overall difference among the three treatment groups was = 0.08. The difference between the 500-mg and 1000-mg groups was = 0.20. The difference between the 500-mg and 2000-mg groups was = 0.02. Survival estimates for patients by dose of clarithromycin.PPP

Discussion

Our study shows that clarithromycin therapy resulted in significant reductions in the quantity of mycobacteremia in patients with AIDS who have disseminated M. avium complex disease. A rapid diminution in the quantity of mycobacteremia was observed with each of the three doses of clarithromycin studied, and higher doses were associated with more rapid elimination of M. avium complex from the blood (cultures negative for M. avium complex). Ultimately, most patients continuing treatment with clarithromycin had negative blood cultures on at least one occasion. In addition to a microbiological response to therapy, patients had improvements in constitutional symptoms of M. avium complex disease. However, clinical relapse and the emergence of clarithromycin-resistant M. avium complex after several months of therapy were observed in a large proportion of patients who continued to receive clarithromycin, regardless of dose.

The response of M. avium complex disease to clarithromycin monotherapy is remarkable in that no other agent has resulted in such large decreases in the number of these organisms in the blood of patients with AIDS. After 1 to 2 months of therapy with clarithromycin, CFUs of M. avium complex/mL of blood decreased by 2.7 to 2.8 log, with most patients becoming culture negative. In contrast, 1 month of monotherapy with azithromycin, another extended spectrum macrolide, resulted in a mean decrease of 1.4 log of CFUs of M. avium complex in blood [15]. Clarithromycin and azithromycin have not been directly compared in a randomized trial. No other single agent used to treat M. avium complex bacteremia has produced more than a 1-log reduction in CFUs/mL of blood in colony counts of M. avium complex during a similar period of time [21, 22].

Several previous studies of multidrug therapy for M. avium complex disease in patients with AIDS have shown clinical responses and cultures negative for M. avium complex (sterilization) in a high proportion of patients. Hoy and coworkers [12] reported that 21 of 28 patients treated for at least 1 month with isoniazid, ethambutol, clofazimine, and rifabutin became culture negative and had resolution of constitutional symptoms. Chiu and coworkers [10] reported a 1.5-log mean decrement in CFUs for M. avium complex in blood after 1 month of therapy with rifampin, ethambutol, ciprofloxacin, and amikacin. Kemper and colleagues [11] used a four-drug oral regimen of rifampin, ethambutol, ciprofloxacin, and clofazimine to treat 41 patients with AIDS and M. avium complex bacteremia. They noted a median 1.5-log reduction in bacteremia, and blood cultures became negative for M. avium complex in 42% of patients. In contrast, clarithromycin monotherapy resulted in negative blood cultures in 50% of patients continuing therapy at 4 to 8 weeks, depending on dose, and the median decrease in bacteremia was more than 2.7 log. Our data support the conclusions of Dautzenberg and coworkers [15], whose study suggested that clarithromycin had the capacity to eliminate M. avium complex from the blood of patients with AIDS who had disseminated M. avium complex infection. Our study included more patients (154 compared with 15) with substantially higher levels of mycobacteremia.

Clarithromycin was generally well tolerated at doses of 500 mg and 1000 mg twice daily, although dose-related gastrointestinal side effects were observed. In general, macrolide antibiotics cause nausea, vomiting, taste perversion, abdominal discomfort, and diarrhea, particularly at high doses. Previous studies of clarithromycin indicated that doses greater than 1000 mg twice daily might be tolerated less well [17]. The significantly higher incidence of adverse reactions in patients receiving the 2000-mg twice daily dose suggests that this dose should not be used in patients with AIDS. Although this dose did result in the fastest decrease in quantitative mycobacteremia, the 1000-mg twice daily dose was nearly as effective and better tolerated. A larger proportion of patients survived and were able to continue therapy with clarithromycin at 500 mg twice daily through 12 weeks than with either of the higher doses.

We found that patients randomly assigned to the lowest dose of clarithromycin (500 mg twice daily) had the longest median survival. Moreover, during the first 12 weeks of study, significantly fewer patients assigned to the 500-mg dose of clarithromycin died than did patients assigned to higher doses. This early difference does not appear to be a consequence of more severely ill patients being assigned to higher doses. Poor survival of patients receiving the 2000-mg dose may have been related to the intolerance of clarithromycin experienced by 40% of patients assigned to this dose compared with 15% of patients receiving the 500-mg dose. However, only 16% of patients assigned to 1000 mg terminated treatment early because of adverse reactions. Thus, the early survival advantage seen with the 500-mg dose is not fully understood. Because 500 mg of clarithromycin twice daily was clinically and microbiologically active, although the time to response was slower than with higher doses, use of this dose in the initial treatment of AIDS-related M. avium complex bacteremia is recommended (see package insert for clarithromycin). Additional studies of the effect of various doses of clarithromycin on the clinical course of M. avium complex disease and survival are warranted.

The occurrence of clinical and microbiologic relapse of M. avium complex bacteremia with resistant organisms is of concern. Although almost all isolates were susceptible to clarithromycin at baseline, resistance began to appear at 8 to 10 weeks of therapy, and a resistant organism was eventually recovered from approximately one half of treated patients. Resistant isolates were usually associated with a recrudescence of symptoms and signs of disease. Development of clarithromycin resistance has been shown to result in cross-resistance to azithromycin but not to other antimycobacterial agents [18]. The mechanism by which resistance develops in M. avium complex is not known. The prevalence of clarithromycin-resistant organisms in cultures of M. avium complex from untreated patients is 1/10⁸[14, 18]. It is likely that under selective pressure from a single antimicrobial agent, a small proportion of innately drug-resistant organisms eventually become predominant. This is consistent with the observation that more than one half of patients with pulmonary tuberculosis treated with streptomycin alone or isoniazid alone developed drug-resistant isolates within 3 months [23, 24].

Alternatively, the mechanism by which clarithromycin inhibits mycobacterial DNA synthesis may directly result in the acquisition of drug resistance. Ji and colleagues [25] have shown in a murine model of M. avium complex bacteremia that monotherapy with clarithromycin and azithromycin results in high-level resistance after 1 month, whereas other classes of drugs (including aminoglycosides, rifamycins, and ethambutol) given for the same amount of time did not induce resistance. Although this could reflect acquired resistance, it is also possible that the greater antimycobacterial activity of macrolides permits the expression of primary drug-resistant strains over time, as has been suggested for standard antituberculous drugs [26].

Dautzenberg and coworkers [27] have recently reported that clarithromycin given in combination with one or more other antimycobacterial agents, including clofazimine, ethambutol, and fluoroquinolones, is associated with the emergence of clarithromycin-resistant isolates in 14% to 23% of patients. Prospective studies of combination therapy for M. avium complex bacteremia are necessary to determine optimal regimens. The United States Public Health Service Task Force on Mycobacterium avium complex [28] has recommended that patients treated with a macrolide receive at least one other drug, such as ethambutol, clofazimine, rifampin, rifabutin, ciprofloxacin, or amikacin. However, some experts would use at least three or four drugs. Further study is necessary to determine the optimal regimen for treating disseminated M. avium complex disease.

Appendix

Members of the AIDS Clinical Trials Group Protocol 157 Study Team included Edward W. Stool, MD (AMI Park Plaza Hospital, Houston, Texas); Rob Roy MacGregor, MD (University of Pennsylvania, Philadelphia, Pennsylvania); Thomas Buehner, MD (San Diego Community Research Group, San Diego, California); Albert W. Wu, MD, Grace Link Barnes, BSN, MPH, and Rebecca Becker, PA-C (Johns Hopkins University, Baltimore, Maryland); Pamela Urbanski, MSN (Rush Medical College, Chicago, Illinois); William Richardson, RN (University of Southern California, Los Angeles, California); Richard Hafner, MD, and Dennis Dixon, PhD (National Institute of Allergy and Infectious Diseases, Bethesda, Maryland); David W. Feigal, MD, MPH (University of California, San Diego, San Diego, California); Melinda Dellerson, BA, Sinder Gupta, RPh, David Henry, PhD, and Seymour Schlager, MD, PhD (Abbott Laboratories, Abbott Park, Illinois).

Abbreviation

CFU: colony-forming units