MAC lung infections are a growing public health problem. The ATS / IDSA 2007 guidelines for the treatment of these non-tuberculous mycobacterial infections recommend the use of a macrolide or azalide (clarithromycin or azithromycin), rifampicin or rifabutin and ethambutol.

For MAC disseminated infections, several studies have compared combinations containing clarithromycin or azithromycin and found no significant difference in efficacy. No randomized controlled trials have been performed for pulmonary infections to compare clarithromycin and azithromycin in terms of efficacy. Clarithromycin is often used as a first-line treatment in France, but its tolerance is often poor, particularly in terms of risk of hepatitis, metallic taste in the mouth, nausea or vomiting, and it interacts with many drugs via cytochrome p450 . In particular, it increases the toxicity of rifabutin, in particular in terms of uveitis. Azithromycin has fewer side effects especially less digestive toxicity and drug interactions than clarithromycin.

The hypothesis is therefore that the efficacy of azithromycin would be non-inferior in comparison with that of clarithromycin.

1. Scientific rationale and general description:

   Nontuberculous mycobacteria (NTM) represent all mycobacteria not belonging to the Mycobacterium tuberculosis complex. This group of mycobacteria was discovered recently and has demonstrated pulmonary pathogenicity. The presence of a NTM in a respiratory specimen is therefore not sufficient to confirm the diagnosis of a lung infection. The ATS/IDSA 2007 guidelines require clinical, radiological and microbiological criteria to distinguish infection from colonisation.

   The main NTM isolated and responsible of NTM pulmonary infections in France is Mycobacterium avium complex (MAC). MAC organisms are common in many environmental sites, including water, soil and in animals. Up to now, the main hypothesis for transmission is inhalation of MAC via aerosol during the shower. Three classical clinical and radiological patterns are described for MAC: the fibrocavitary lung disease, the nodular bronchiectatic disease and the hypersensitivity disease. This last pattern classically doesn't need any antibiotics. So, this study will focus only on the two first patterns: fibrocavitary form and nodular-bronchiectatic form.

   The limitations for effective therapy were the absence of antimicrobial agents with low toxicity and good in vivo activity against the organism. The major therapeutic advance is the introduction of the newer macrolides, clarithromycin and azithromycin, which have substantial efficacy in vitro and clinical activity against MAC. Structurally, azithromycin is an azalide. However, because of the close similarity of azalides to macrolides, the term "macrolide" is often used to refer to azithromycin and clarithromycin. As macrolides are the cornerstone of MAC treatment, their use in monotherapy is responsible of selection of macrolides MAC resistant strains that are associated with microbiological and clinical relapse. This is why all international guidelines and all experts recommended the use of macrolides with at least two companion drugs, classically rifampicin and ethambutol.

   Treatment of MAC lung disease is long and not well tolerated. Despite treatment, microbiological failure or early relapse occurs in 20-30% of patients. One of the main failure explanations in NTM disease is the spontaneous treatment stop by the patient, due to adverse effects. Even if clarithromycin is often used in first line in France and is the only one macrolide with indication in its Summary of Product Characteristics, its tolerance is often poor. Adult patients generally cannot tolerate clarithromycin at more than 1,000 mg/day. The most common toxicities seen with clarithomycin are gastro-intestinal (metallic taste, nausea, and vomiting) and it interacts with numerous drugs via P450 cytochrome. Moreover, clarithromycin enhances rifabutin toxicity, especially uveitis. Azithromycin toxicity is dose and serum-level related. Most adult patients with MAC lung disease do not tolerate azithromycin doses of greater than 300 mg/day because of frequent adverse events, including gastrointestinal symptoms (primarily diarrhea).

   The main hypothesis is the non-inferiority of azithromycin containing regimen efficacy in comparison to clarithromycin. Moreover differences in intracellular concentration of azithromycin and clarithromycin may result in differences in efficacy and safety. The investigators also suggest that intracellular concentration of azithromycin or clarithromycin may be correlated with microbiological success at 6 months of treatment. If this study could demonstrate that azithromycin is not inferior to clarithromycin in term of efficacy, azithromycin could be used in first line to improve patient tolerance and adherence to the treatment.

2. Study design and randomization method:

The CLAZI study is a multicentre, controlled open-label clinical trial studying two treatment regimens containing two molecules, rifampin and ethambutol, in combination with a third randomized molecule, either clarithromycin or azithromycin.

41 centres located in France will participate.

After having confirmed the presence of all inclusion criteria and the absence of all exclusion criteria, and after having obtained the patient's free and informed consent, the patient will be included and randomized to one of the treatment regimens. An interactive web response system will be used to randomize patients using a minimization algorithm stratified on centre, radiological form (nodular bronchiectatic form and fibrocavitary form).

3. Study treatment procedure:

After inclusion, the patients will receive the treatment as follow:

• Rifampin 10 mg/kg once daily on an empty stomach (600 mg for patients weighing more than 50 kg, 450 mg for patients weighing less than 50 kg) at least 30 minutes before a meal, in combination with ethambutol 15 to 20 mg/kg once daily and clarithromycin 1,000 mg/day in two divided doses
• Rifampin 10 mg/kg once daily on an empty stomach (600 mg for patients weighing more than 50 kg, 450 mg for patients weighing less than 50 kg) at least 30 minutes before a meal, in combination with ethambutol 15 to 20 mg/kg once daily and azithromycin 250 mg once daily

Treatments will not be provided because they are prescribed in their indication. They will be issued in the dispensary. The compliance of treatment will be evaluated for each period by the investigator, from the book of compliance, given to the patient at the beginning of treatment and bring each consultation.

4. Study procedures: The planned inclusion period is 48 months. However, the number of inclusions will be evaluated at 18 months in order to request extension of the inclusion period, if necessary. Each patient will be followed for the first 12 months of treatment.

Monitoring will be performed throughout treatment, comprising:

• Monitoring of liver function tests (transaminases, gamma GT, alkaline phosphatase, bilirubin) weekly for the first two weeks, then monthly throughout treatment. Liver function tests will be repeated more frequently whenever an abnormality is detected.
• CBC, clotting parameters at M1 then only in the presence of abnormal liver function tests.
• Monthly ophthalmological examination (colour vision and, if abnormal, visual field examination) and an additional consultation if the patient reports any symptoms.
• Depending on the patient's concomitant treatment and any possible drug interactions, pharmacological assays of the patient's usual medications (e.g. digoxin) left to the discretion of the patient's clinician.
• Electrocardiogram two weeks after starting treatment
• Clinical surveillance

A nested pharmacokinetic study will be done on 100 patients with peak serum and mononuclear cells concentration of azithromycin and clarithromycin and their main metabolites (D azithromycin and 14OH clarithromycin respectively). Indeed, a correlation could exist between intracellular concentration and microbiological success. These dosages will be done at 1 and 6 months.

Dosages in hair samples will be done at M6, as it is the best reflect of macrolides use. Indeed, most drugs present in the bloodstream also reach and stay in the forming hair where their concentrations can be measured and interpreted as an overall chronic exposure.

For all patients, a baseline assessment will be performed on inclusion and the same criteria will then be assessed at 3, 6 and 12 months in order to determine the clinical and radiological outcome.

Long-term follow-up of the included patients will be proposed, with recording of any new adverse effects emerging at the end of treatment (between 12 and 18 months), any relapses, and survival at 18, 24, 36 months and, whenever possible, 5 years.

If a patient who initially clears sputum, but then turns positive again while still on therapy could be a failure of treatment or a re-infection. To distinguish between failure and re-infection, genotyping will be performed.

Relapse is defined by the presence of two positive cultures on respiratory specimens obtained at least 6 months after stopping treatment.

Cure is defined by classical bacteriological criteria, i.e. 2 negative cultures during the last 3 months of treatment and 3 years after stopping treatment. When specimens are not obtained 3 years after stopping treatment, the definition of clinical cure will be based on the absence of relapse at 3 years in survivors.

5. Statistical methods: Quantitative variables will be presented as the mean ± standard deviation or the median [range] and qualitative variables as percentage.

Analysis of the primary point:

Non-inferiority will be established if the upper limit of the two-sided 95% confidence interval of the difference in the proportion of 6-month negative sputum rate between the two groups (Control arm - Experimental arm) is lower than the non-inferiority margin (10%). Intention-to-treat analysis will be performed followed by a per protocol analysis. A multiple imputations analysis will be performed if more than 5% of the data are missing for the primary endpoint. Farrington-Manning p-value for non-inferiority will be calculated and a two-sided p-value < 0.05 will be required for statistical significance.

Analyses of secondary endpoints:

• The tolerance and safety will be compared between the two arms with a χ2 or Fisher exact test as appropriate.
• Clinical improvement will be compared with Student or Wilcoxon test and radiological improvement with χ2 or Fisher exact test.
• The 3, and 12-month sputum conversion rate will be compared with a χ2 or Fisher exact test as appropriate and asymptotic or exact 95% confidence intervals will be given.
• 12-month death will be compared with χ2 or Fisher exact test and Kaplan-Meier curves will be displayed graphically.
• Serum and mononuclear cells concentration of azithromycin and clarithromycin and their metabolites will be compared with Student or Wilcoxon test.
• For each arm, association between azithromycin/clarithromycin intracellular concentration in circulating mononuclear cells and microbiological success at 6 month will be assessed with a logistic regression model.
• For each arm, association between azithromycin/clarithromycin intracellular concentration in circulating mononuclear cells and tolerability of treatment will be assessed with a logistic regression model.
• Association between MAC species (avium, intracellulare and chimaera) and 6-month conversion will be assessed with a logistic regression model with adjustment by the treatment arm.

Given the substantial number of analyses for secondary endpoints, the step-down Bonferonni procedure will be used to control the type I error rate.

Statistical analysis will be performed using SAS® version 9.4 (SAS Institute, Cary, NC).

Inclusion Criteria:

• patients 18 years of age or older
• having a positive Mycobacterium avium complex sample showing the ATS / IDSA infection criteria and requiring treatment
• ATS / IDSA infection criteria combine clinico-radiological criteria, associated with microbiological criteria
• the exclusion of any other diagnosis on the thoracic CT, fibroscopy and bacteriological samples

Exclusion Criteria:

• Known hypersensitivity to one of the study molecules (rifampicin, ethambutol, azithromycin, clarithromycin)
• Relapse of an MAC infection,
• Strain resistant to macrolides, based on genotyping susceptibility testing (genotyping susceptibility testing must be done before inclusion)
• Treatment that interacts with cytochrome p450 that can not be replaced by another therapeutic,
• HIV serology 1 and 2,
• Renal insufficiency with creatinine clearance less than 30 ml / min,
• Pregnancy and breast feeding,
• Contra-indication to one of the antibiotics,
• Impossibility to follow the protocol due in particular to drug addiction according to the investigator,
• Limited life expectancy, less than 6 months,
• Patient already participating in a clinical trial on a medical treatment or a therapeutic strategy for non-tuberculous mycobacteria.