We did a multicenter clinical trial of the efficacy of single-drug therapy with clarithromycin for patients with cutaneous infection due to M. chelonae, especially in patients with disseminated infection.

Methods

Top Methods Results Author & Article Info References

Clinical isolates of rapidly growing mycobacteria submitted to the Nocardia/Mycobacteria Research Laboratory of the University of Texas Health Center between February 1991, and February 1992, were screened for isolates of M. chelonae. Using standard methods, most isolates had been identified as belonging to the M. chelonae group (M. chelonae and M. abscessus) [6]. They were identified as species chelonae on the basis of use of citrate as a sole carbon source [6] and antibiotic resistance patterns that included high-level resistance to cefoxitin (MIC 256 µg/mL) and susceptibility to tobramycin (MIC 8 µg/mL) [3].

Antimicrobial susceptibility to these agents as well as to clarithromycin was tested using the broth microdilution technique of Swenson and colleagues as described for rapidly growing mycobacteria [7]. Microdilution plates were prepared and dispensed with the Mini-Quick Spense II reagent dispenser (Dynatech Laboratories; Chantilly, Virginia). Susceptibility to clarithromycin was defined as an MIC of 4 µg/mL.

After a clarithromycin-susceptible isolate of M. chelonae was identified, the primary care physician for the patient was contacted and the patient was entered into the study protocol. Entry criteria included ability to give informed consent; patient reliability; absence of a history of macrolide allergy; presence of active infection that had not been totally surgically excised; and, if a woman, not pregnant and, if premenopausal, taking acceptable birth control. Patients received clarithromycin at an oral dose of 500 mg twice a day after meals for 6 months. No other antimicrobial agents were given. Safety monitoring included routine clinic visits, complete blood counts, liver and renal function tests, and routine urinalysis. Follow-up after therapy continued for at least 6 months. Follow-up wound cultures for mycobacteria were taken whenever possible during and after therapy. Signed informed consent was obtained from the first 10 patients before entry into the study. After the Food and Drug Administration approved clarithromycin in late 1991, this study was formally discontinued. However, additional patients were treated using the same protocol except for the absence of pharmaceutical company supervision and signed informed consent.

Results

Top Methods Results Author & Article Info References

Twenty patients with cutaneous infections due to M. chelonae were identified during the study period. These patients were under the care of 18 different physicians. All pretreatment isolates were susceptible to clarithromycin. Two patients had minimal disease and were not considered candidates for therapy. One patient was considered unreliable, and one physician did not have the time to do the necessary paperwork. The remaining 16 patients were entered into the study. Two patients died of unrelated diseases within 1 month of beginning therapy. The remaining 14 patients completed at least 3 months of therapy and are described in detail. Characteristics of these patients and their types of clinical disease are shown in Table 1. Patients ranged in age from 18 to 79 years. All were immune suppressed, and all were taking either corticosteroids (93%) or cyclophosphamide (7%). The most common underlying diseases were rheumatoid arthritis (5 patients), autoimmune disorders (4 patients), and organ transplantation (3 patients). Eleven of the 14 patients (79%) had disseminated cutaneous infection, manifested by multiple cutaneous lesions. Three patients had localized infections.

All patients had positive acid-fast smears and cultures of one or more skin lesions within 2 months of initiation of therapy. The MICs to clarithromycin were determined before entrance into the study. All infecting isolates met biochemical and antimicrobial susceptibility criteria as isolates of M. chelonae [3, 6].

The results of therapy for the 14 patients are shown in Table 2. Because of a high MIC to clarithromycin of 1.0 µg/mL, patient 2 was given clarithromycin, 1.0 g twice a day, but the dosage was lowered to 500 mg twice a day because of nausea. Two patients (patients 8 and 12) had the dose increased from 500 mg to 1000 mg twice a day. (One dosage was inadvertently increased by someone not familiar with the protocol and the other was increased because new skin lesions developed while taking the medication therapy.) The dose for patient 12 was again lowered to 500 mg because of nausea. Three patients received one other antimicrobial agent concurrent with the clarithromycin, although the patients' isolates were resistant in vitro to these agents. One patient (patient 2) received amikacin, for which the isolate had an MIC of more than 32 µg/mL, and another patient (patient 3) received trimethoprim/sulfamethoxazole, for which the isolate had an MIC of more than 128 µg/mL (sulfamethoxazole component). These agents were given because of uncertainty of whether clarithromycin alone would be effective or would result in acquired resistance (despite the directives of the protocol); these were two of the earliest patients to be entered in the trial. Patient 7 was taking ciprofloxacin (MIC for the M. chelonae of more than 8 µg/mL) for a previously diagnosed Pseudomonas osteomyelitis.

The clinical response of the patients was excellent, with rapid resolution of disease in patients with small numbers of lesions and decrease in the size and amount of drainage in patients with extensive disease. All acid-fast smears and cultures taken 4 weeks or longer after the start of therapy were negative. New skin lesions appeared during therapy on six of the patients, but results of the biopsies of these patients were acid-fast smear negative and culture negative. Patient 12, who was culture negative with therapy despite recognized noncompliance, discontinued therapy on her own after 3.5 months. Although no new skin lesions developed, a previous lesion became acid-fast bacilli-smear positive and culture positive 1 month later. This isolate was resistant to clarithromycin. Two patients (patients 4 and 14) had clearing of their skin lesions but died of unrelated disorders 3 months into therapy. The remaining 11 patients have completed therapy given for a mean of 6.8 months (range, 4.5 to 9 months). Nine of the 11 patients have discontinued therapy after at least 6 months (mean, 7.1 months; range, 6 to 12 months) with no evidence of relapse. Patients 1 and 2 had some continued drainage at the time their therapy was discontinued (Table 2). The drainage has gradually decreased with time, however, and multiple cultures tested after the patients finished taking the medication have remained negative. Examples of clinical responses to therapy are shown in Figure 1(patient 1) and Figure 2(patient 7).

View larger version (70K): [in this window] [in a new window]   Figure 1. Disseminated skin disease. Left. Before treatment. Multiple skin lesions of the leg in patient 1 who had underlying rheumatoid arthritis. Right. During treatment. Close-up of the same area of the leg of patient 1 after 2 months of therapy with clarithromycin.

View larger version (94K): [in this window] [in a new window]   Figure 2. Mycobacterium chelonae osteomyelitis. Left. Before treatment. Cellulitis-osteomyelitis caused by M. chelonae infection at the site of a previous transmetatarsal amputation in a 46-year-old man (patient 7) receiving cyclophosphamide for rheumatoid arthritis. Right. After treatment. Follow-up photograph of the same foot after 6 months of therapy with clarithromycin.

No abnormalities of blood cell counts or liver function tests that could be ascribed to clarithromycin occurred during the study. (One patient had baseline abnormal liver function tests.) Patient 2, the only patient taking amikacin, had abnormal renal function test results that cleared when the amikacin was withdrawn but clarithromycin was continued. Patient 3, who was taking trimethoprim/sulfamethoxazole, developed a diffuse rash that resulted in withdrawal of both drugs after 9 months of therapy.

Previous drug therapy of cutaneous disease due to M. chelonae has been generally poor. Although approximately 80% of isolates are moderately susceptible to erythromycin (as defined by MICs by the National Committee for Clinical Laboratory Standards for bacteria that grow aerobically) [2] with MICs of 1 to 4 µg/mL [3], these values are close to the peak achievable serum levels for this drug. Clinical response to high-dose erythromycin has varied, and some successfully treated cases have been reported [8, 9]. We are aware of other successfully treated cases that have not been reported, although clinical and microbiologic relapse has been common after discontinuation of therapy in patients with underlying immunosuppression. (RJ Wallace, Jr. Unpublished observations.) Approximately 20% of individual strains have susceptible MICs for doxycycline [1, 3] and ciprofloxacin [1, 4], and successful therapy with these agents has also been reported [10, 11]. However, many other patients have failed such therapy, and the disease in most patients has persisted for long periods, often years [12]. Interestingly, although the disease often results in varying degrees of morbidity due to pain, extremity edema, and low-grade fever, death resulting from untreated or uncontrolled infection due to M. chelonae has been rare.

In the current study, treatment with the newer macrolide, clarithromycin, resulted in clearing of the skin lesions and no relapses of disease with a minimum of 5 months follow-up in 10 of 11 patients and a minimum of 6 months follow-up in 9 of 11 patients. One noncompliant patient, who discontinued her therapy after only 3.5 months, relapsed with a drug-resistant isolate within 1 month of discontinuing therapy. Despite the fact that this patient was receiving essentially monotherapy, she was the only patient with a case of acquired drug resistance that we observed. Tolerance to clarithromycin at this dose was excellent and side effects, other than a bitter taste, were few. Although few patients were treated, the low MICs of all isolates of M. chelonae tested and the excellent response of these patients suggest that clarithromycin may be the current therapy of choice for cutaneous (disseminated) infection due to this species of rapidly growing mycobacteria. Additional patients with long-term follow-up after treatment was withdrawn (preferably at least 6 months) are needed. Whether lower dosages of drug (for example, 250 mg twice a day) or shorter duration of therapy (for example, 4 months) is possible have not yet been studied. Although the MICs of other newer macrolides such as azithromycin and roxithromycin are approximately 10 times higher against isolates of M. chelonae [5], these MICs are still within a reasonable therapeutic range given the high tissue concentrations of these drugs and the preliminary success of one of these drugs (azithromycin) against disseminated infection in patients with the acquired immunodeficiency syndrome caused by Mycobacterium avium complex isolates [13], whose MICs are even higher to azithromycin than are those of M. chelonae. Thus, treatment with other newer macrolides against M. chelonae may also be effective but remains to be evaluated.

Given the theoretic risk for mutational drug resistance with monotherapy of any mycobacterial species, combination therapy of M. chelonae with clarithromycin as one of the components seems preferable to single-drug therapy. Unfortunately, the generally poor activity of other oral agents usually means that the newer macrolides are the only available active oral agents. Injectable agents such as tobramycin, amikacin, and imipenem are active against most isolates of M. chelonae [3, 18] but are less desirable because of their cost, toxicity, and need for long-term intravenous access and almost certain hospitalization. Additional classes of oral agents with activity against M. chelonae are clearly needed.

All of the patients in our study were being treated with either prednisone (93%) with or without other suppressive agents or cyclophosphamide (cytoxan) alone (7%) when their disease developed. Results of a previous study of 100 patients with cutaneous M. chelonae showed that 92% of patients with disseminated disease (53 cases) and 23% of patients with localized disease (47 cases) were taking corticosteroids [1]. In three of 14 of the current patients, the corticosteroids were withdrawn (patients 6, 8, and 11; see Tables 1 and 2). It is probable that this factor played a role in resolution of disease or absence of relapse after discontinuation of therapy, or both. In 11 of 14 patients in the study, however, the immunosuppressive therapy could not be discontinued.

Until recently, most human disease-producing isolates of rapidly growing mycobacteria were simply identified as belonging to the "M. fortuitum complex." Demonstration that the M. fortuitum and the M. chelonae group were strongly associated with different disease syndromes [14-17] and that they had striking differences in antimicrobial susceptibility patterns [3, 4, 18] prompted most laboratories to separate the two groups. Recent studies of DNA homology have clearly established the species status of M. fortuitum, M. peregrinum, M. chelonae, and M. abscessus [19, 20]. Separation of these four species is readily done biochemically [6, 21] and, for the last two species, by high-performance liquid chromatography [22], although the method is available only in a few specialized laboratories. Isolates of M. abscessus have MICs of 1 µg/mL to clarithromycin, but approximately 50% show fourfold or greater increases in MICs with a 1 day-longer incubation time [5]. A major concern is whether these latter isolates will prove to be resistant clinically. Until this concern is resolved, we strongly recommend that all patients with disease due to the M. chelonae group (nitrate negative, iron uptake negative) have their isolates identified as to species and, if possible, as to both short and long incubation times for the MIC. All relapse isolates should be tested for the possibility of acquired resistance to clarithromycin, which was observed in one of the patients in our study.