Azithromycin in Patients With Cystic Fibrosis Chronically Infected With Pseudomonas aeruginosa:  A Randomized Controlled Trial FREE

Cystic fibrosis (CF) is characterized by a recurrent cycle of pulmonary infection and inflammation. Pseudomonas aeruginosa is the most common pathogen in patients with CF and by age 18 years, 80% of patients are chronically infected. Neutrophils are the dominant inflammatory cells and much of the airway destruction characteristic of CF is secondary to neutrophil-derived proteases and oxidants.¹ Treatment strategies for CF lung disease have included antibiotics, mucolytics, and anti-inflammatory therapies.²

There is evidence suggesting that macrolide antibiotics may be beneficial for patients with CF. Macrolide antibiotics substantially reduced morbidity and mortality in patients in Japan with diffuse panbronchiolitis.^3- 4 Diffuse panbronchiolitis shares many clinical features with CF; patients are often infected with mucoid strains of P aeruginosa and mortality is secondary to chronic progressive lung disease. During the late 1990s, 2 small open-label studies suggested that macrolide antibiotics, particularly azithromycin, may benefit patients with CF.^5- 6 In 2002, 2 small trials also reported benefit.^7- 8 The mechanism of action of macrolide antibiotics is not known, but in vitro studies suggest that macrolide agents may decrease the production of bacterial virulence factors,^9- 10 modulate the host inflammatory response, or both.¹¹

We conducted a randomized, double-blind, placebo-controlled, multicenter trial in patients with CF chronically infected with P aeruginosa to determine if azithromycin administered for 24 weeks was safe, improved lung function, and decreased pulmonary exacerbations.

The trial was conducted at 23 CF Foundation accredited care centers in the United States between December 15, 2000, and May 2, 2002, and was coordinated by the CF Therapeutics Development Network Coordinating Center in Seattle, Wash. The institutional review board at each center approved the study and each participant or their parent/guardian voluntarily consented to participate. Eligibility criteria included a documented diagnosis of CF; age 6 years or older; weight of 25 kg or more; chronic infection with P aeruginosa, defined by a positive respiratory tract culture 1 year or more before screening and at screening; and a forced expiratory volume in 1 second (FEV₁) of 30% or more of predicted.¹² Exclusion criteria were Burkholderia cepacia complex isolated from the respiratory tract at screening or within 2 years of screening; nontuberculous mycobacteria within 2 years of screening or acid-fast bacillus smear positive at screening; history of biliary cirrhosis or portal hypertension, splenomegaly on examination, or liver function results (ie, aspartate serum transferase, alanine transferase, γ-glutamyltransferase phosphate) of 2 or more times the upper limit of normal; use of intravenous antibiotics, quinolones, or other oral antibiotics within 14 days of screening; use of systemic corticosteroids (≥20 mg of prednisone daily) within 30 days of screening; or initiation of tobramycin solution for inhalation,¹³ recombinant human dornase alfa inhalation solution,^13- 14 or high-dose ibuprofen¹⁵ within 60 days of screening. Long-term use of these medications was permissible.

Approximately equal allocation to azithromycin and placebo groups was planned to achieve reasonable balance within categories of disease severity (≥60% or <60% of the predicted value of FEV₁) and weight (≥40 kg or <40 kg) and within each study site. Because of the large number of sites and the relatively small number of participants at each site, an allocation strategy used an algorithm that recalculated the proportion of participants at each site with each set of baseline clinical characteristics, before each randomization assignment.¹⁶ Randomization assignments and a series of blinded drug kit numbers were generated by the CF Therapeutics Development Network Coordinating Center. Kits were distributed to the centers and a centralized secure randomization system at the coordinating center was used to assign drug kit numbers to patient identification numbers at each site. All study personnel and participants were blinded to treatment assignment. The randomization codes for each participant were revealed to the researchers once recruitment, data collection, and data analyses were completed.

Azithromycin was supplied as 250-mg tablets and the placebo was identically packaged. Participants who weighed less than 40 kg were instructed to take 1 tablet 3 days a week (Monday, Wednesday, and Friday) and participants who weighed 40 kg or more were instructed to take 2 tablets on the same 3 days per week. The study drug was discontinued if a participant had an allergic reaction; a life-threatening adverse event, not including hospitalization for a pulmonary exacerbation; an adverse event that was considered intolerable by the site's study team or research participant; or if nontuberculous mycobacteria grew from a sputum sample obtained at screening. For participants with toxicity thought to be related to study drug (eg, gastrointestinal adverse effects), the study protocol included provisions for a step-down in the dosing regimen. Day-to-day care of the patient was left to the discretion of the CF care teams.

Medical history, physical examination, spirometry, oximetry, and chest radiograph were obtained at the screening visit (day −14). Clinical evaluations and spirometry were performed at days 0 (randomization), 28, 84, 168 (end of active treatment phase), and 196. Adverse events and concomitant medications were recorded during each visit and by telephone calls conducted on days 56, 112, and 140. Sputum samples or throat swabs were obtained at screening and on day 168. Additional sputum was collected from participants at 12 centers on days 28 and 168 to measure azithromycin concentrations.¹⁷

Blood samples to monitor hematology, liver function, and creatinine were obtained at screening, day 28, and day 168. The Cystic Fibrosis Quality of Life Questionnaire was administered to all English-speaking participants and parents/guardians on days 0 and 168.^18- 20 Audiology testing was performed on participants at 11 centers on days 0 and 168. Hearing loss was defined as at least 10 dB increase in auditory threshold at 1 or more of the following frequencies: 500, 1000, 2000, 3000, 4000, 6000, and 8000 Hz.

The primary outcomes of the study were to determine if azithromycin was associated with a change in FEV₁ from day 0 to completion of therapy (day 168) and whether it was safe, as determined by adverse events including self-reported symptoms, physical findings, laboratory tests, audiology, and microbiology. Secondary outcomes included change in forced vital capacity (FVC), change in body weight, time until first pulmonary exacerbation, number of pulmonary exacerbations, hospitalization rates, use of nonquinolone oral antibiotics, changes in inflammatory markers, and quality of life. Exacerbations were defined as the use of intravenous antipseudomonal antibiotics or oral quinolones for 7 or more days.

Pulmonary function testing was performed in accordance with American Thoracic Society standards.²¹ The FEV₁ and FVC measurements were expressed in liters and referenced to a healthy population as a percentage of predicted by using a modification of the Knudson equations.¹²

A standard microbiologic evaluation of sputum and throat cultures for potential pathogens was performed by the Therapeutics Development Network Microbiology Core Laboratory.²² Acid fast bacillus smear and culture for nontuberculous mycobacteria were performed on sputum specimens only.²³

Levels of IL-8 and neutrophil elastase activity were determined by the Therapeutics Development Network Inflammatory Markers Core Laboratory by using an enzyme immunoassay (R&D Systems, Minneapolis, Minn) and a kinetic spectrophotometric assay (Sigma Chemicals, St Louis, Mo), respectively.²⁴

We hypothesized that the mean (SD) decrease in FEV₁ in the placebo group would be 0.05 (0.25) L and the mean (SD) increase in the azithromycin group would be 0.11 (0.28) L after 168 days of treatment. With these assumptions, a sample size of 88 participants per treatment group provided 97.5% power to detect a significant difference between treatment groups (2-sided P = .05). Participants were screened until a maximum of 185 participants were eligible for participation to account for attrition.

Analyses were conducted by the CF Therapeutics Development Network Coordinating Center according to a prespecified analysis plan and performed with a modified intent-to-treat principle that included all randomized participants who had at least 1 postbaseline measurement for the primary outcome. For the primary efficacy analysis, it was assumed a priori based on the results of other successful therapies for CF^13- 14 that after an initial response to study drug, the FEV₁ of each participant would follow a linear pattern of decline, incline, or stabilization for the remainder of the treatment period. Given this assumption and to improve efficiency in the estimation of each participant's 168-day change from baseline, a derived variable approach²⁵ was used to estimate the change from baseline to day 168 for each participant. This approach maximizes the use of repeated FEV₁ measurements obtained throughout the study from each participant. For each participant, a piecewise linear regression vs time (actual days of day 0, 28, 84, and 168 visits) was fit to his/her FEV₁ data by using day 28 as the change point at which a maximum initial response would be observed. This regression model was used to estimate the 168-day change for each participant. The mean 168-day change across participants using these derived estimates was then compared between treatment groups by using a .05 level 2-sided t test. If a participant had insufficient data to compute the change using this approach (eg, were lost to follow-up), his/her day 0 FEV₁ measurement was subtracted from the last available measurement to estimate the change from baseline included in the primary analysis. The relative change in FEV₁% predicted, calculated as

{[(FEV₁% predicted at day 168 – FEV₁% predicted at day 0) × 100]/FEV₁% predicted at day 0},

was also compared between treatment groups by using a .05 level 2-sided t test. Between-group comparisons of proportions were performed by using χ² test or the Fisher exact test, as appropriate. Mean changes between treatment groups in continuous outcome measures were compared with a 2-sample t test. Differences between groups in the linear rate of change from baseline in weight were estimated and tested using repeated measures regression with robust variance estimation.²⁶ Time to first exacerbation was assessed by using Cox proportional hazards regression model and graphically displayed using Kaplan-Meier method estimates. S-PLUS version 6.1 (Insightful Corp, Seattle, Wash) and SAS version 8.0 (SAS Institute, Cary, NC) were used for all statistical analyses. P<.05 was considered statistically significant.

Of the 251 participants screened for this study, 185 (74%) were randomized: 87 to the azithromycin group and 98 to the placebo group (Figure 1). Four participants in the azithromycin group and 6 in the placebo group were lost to follow-up. All but 1 participant in the placebo group, who was lost to follow-up before the day 28 visit, were included in the efficacy analysis. The median number of participants randomized per site was 8 (range, 3-17) at each of the 23 study sites. The baseline characteristics and randomization strata were similar between treatment groups (Table 1). Comparable proportions of participants in each group used long-term concomitant medications at baseline; recombinant human dornase alfa inhalation solution and tobramycin solution for inhalation were used by 75% and 61% of participants in the azithromycin group and 67% and 57% of participants in the placebo group, respectively. Compliance was monitored by the number of pills dispensed and returned, and was similar in both groups. On average, 93% of the weekly dosages in the azithromycin group and 89% of the weekly dosages in the placebo group were used.

Participants in the azithromycin group had an estimated 0.097-L (SD, 0.26) improvement in FEV₁ at day 168 while participants in the placebo group had an estimated 0.003-L (SD, 0.23) improvement (mean difference, 0.094 L; 95% confidence interval [CI], 0.023-0.165; P = .009). As measured by the mean (SD) relative change from baseline in FEV₁% predicted, the azithromycin treatment group had a 4.4% (13.6%) improvement and the placebo group had a 1.8% (10.7%) decline (mean difference, 6.2%; 95% CI, 2.6%-9.8%; P = .001). The improvement in FEV₁ was noted at day 28 and sustained until day 168 but by day 196 (4 weeks after the active treatment was discontinued), FEV₁ returned to near baseline (Figure 2).

An improvement in FVC paralleled that observed for FEV₁. The mean (SD) improvement in FVC at day 168 was 0.14 (0.32) L in the azithromycin group and 0.02 (0.26) L in the placebo group (mean difference, 0.12 L; 95% CI, 0.04-0.20; P = .01). This corresponded with a 5.0% treatment effect in the relative change in FVC% predicted as the azithromycin group improved by 3.7% (SD, 11.8%) and the placebo group declined by 1.3% (SD, 9.0%) (mean difference, 5.0%; 95% CI, 1.9%-8.0%; P = .002).

Among the adverse events reported, only nausea, diarrhea, and wheezing were found to statistically differ between participants receiving azithromycin and those receiving placebo (Table 2). Nausea occurred in 17% more participants in the azithromycin group (95% CI, 5%-29%; P = .01), diarrhea in 15% more (95% CI, 4%-25%; P = .009), and wheezing in 13% more (95% CI, 4%-23%; P = .007). The majority of adverse events were described as mild or moderate in intensity. There were no statistically significant differences in laboratory abnormalities between azithromycin and placebo groups.

Two participants reported hearing impairment (1 in azithromycin group and 1 in placebo group), and 2 other participants reported tinnitus (1 in azithromycin group and 1 in placebo group). As judged by audiology testing, there was no evidence to suggest that hearing loss was more frequent in the azithromycin group (8 [18%] of 44) than in the placebo group (12 [24%] of 50; risk ratio, 0.8; 95% CI, 0.3-1.7).

The original dosage was either modified, or stopped and restarted in 8 participants (4 in each group). The majority of serious adverse events were related to hospitalization for pulmonary exacerbations. Study drug was discontinued in 3 participants in the azithromycin group and 2 in the placebo group. In the azithromycin group, study drug was discontinued because of sore feet and bruising (n = 1), sinusitis (n = 1), or rash and ankle pain (n = 1).

The microbiologic profiles of the participants in each treatment group were similar at screening (Table 3). At day 168, 10% fewer participants in the azithromycin group had newly detected methicillin-susceptible Staphylococcus aureus (95% CI, –19% to –3%; P = .01). However, S aureus was not more likely to be eradicated in the azithromycin group; S aureus was eradicated from 18% of participants in the azithromycin group and 12% of participants in the placebo group (95% CI, –16% to 5%; P = .46). There was little difference between the 2 groups in the emergence or eradication of multidrug-resistant strains of P aeruginosa or other potential pathogens, including nontuberculous mycobacteria. P aeruginosa density decreased by 0.3 log colony forming units at day 168 in the azithromycin group and increased by 0.2 log colony forming units in the placebo group (mean difference, 0.5 log colony forming units; 95% CI, 0.4-0.6; P = .06).

The rate of weight change was found to be higher among participants in the azithromycin group than in the placebo group . Participants in the azithromycin group gained an average 0.7 kg more than participants in the placebo group by day 168 (95% CI, 0.1-1.4 kg; P = .02).

Participants in the azithromycin group had less risk of experiencing an exacerbation than participants in the placebo group (hazard ratio, 0.65; 95% CI, 0.44-0.95; P = .03) (Figure 3). There was a reduction in the number of participants hospitalized and the mean number of days of nonquinolone oral antibiotic use in the azithromycin group (Table 4). Although not statistically significant, the azithromycin group experienced a 47% reduction in hospital days and a 39% reduction in the days of intravenous antibiotic use.

Participants in the azithromycin group reported improvements in physical functioning on the CF Quality of Life Questionnaire in contrast with those in the placebo group (mean difference, 2.7; 95% CI, 0.1-5.3; P = .05). No significant differences were found between the groups in the scores for psychosocial functioning, body image factors, or total CF Quality of Life Questionnaire (Table 5).

The mean (SD) elastase levels in the azithromycin and placebo groups were similar at baseline (1.9 [0.4] log µg/mL in both groups). The mean (SD) increase in elastase from day 0 to day 168 was 0 (0.5) log µg/mL in the azithromycin group compared with 0.2 (0.5) log µg/mL in the placebo group (mean difference, –0.2; 95% CI, –0.4 to 0; P = .01). The IL-8 levels were similar at baseline (5.2 [0.5] log pg/mL in both groups) and did not change significantly during the study (mean difference, –0.1; 95% CI, –0.3 to 0.1; P = .14).

Treatment of patients with CF chronically infected with P aeruginosa with azithromycin for 168 days (24 weeks) led to improvement in clinical parameters associated with morbidity and mortality in CF.²⁷ Pulmonary function and nutritional status improved, pulmonary exacerbation rates decreased, and there was some evidence of improved quality of life. No unexpected adverse events were identified and the drug was generally well tolerated.

The concomitant medications recombinant human dornase alfa inhalation solution and tobramycin solution for inhalation were widely used by participants during this study. The results suggested that participants had clinical benefits when azithromycin was added to other therapies proven to be effective in patients with CF. In addition, the impact of azithromycin was comparable with that reported in previous studies of these and other CF therapies. In this trial, the relative change in FEV₁% predicted was 6.2% compared with the 5.6% and 11.0% treatment effects observed in studies of recombinant human dornase alfa and tobramycin solution for inhalation, respectively.^13- 14

This clinical trial of azithromycin expands on and confirms the trends observed in 2 other recently reported smaller trials that demonstrated a benefit from azithromycin in patients with CF.^7- 8 Equi et al⁷ performed a double-blind, placebo-controlled trial in England that included 41 children, 20 of whom did not have persistent infection with P aeruginosa, defined as 3 positive cultures during the year before the trial. The crossover design of this trial included a 2-month washout period between 2 6-month treatment phases. The median relative improvement in FEV₁% predicted between the azithromycin and placebo treatment phases was 5.4%. Oral antibiotic use was reduced but the number of pulmonary exacerbations and courses of intravenous antibiotics did not statistically differ between azithromycin and placebo treatment phases. No subjective reports of adverse effects were elicited. Wolter et al⁸ conducted a randomized, double-blind, placebo-controlled trial in Australia in 60 stable adults with CF chronically infected with P aeruginosa. Because of the relatively small number of participants, treatment assignments resulted in significant differences between the azithromycin and placebo groups; at baseline, participants randomized to the azithromycin group had better lung function and were, on average, taller and heavier. These baseline differences required statistical adjustment for baseline FEV₁ in the analysis of efficacy. Nevertheless, the azithromycin group had a 3.6% relative improvement in FEV₁% predicted and the number of courses of intravenous antibiotics and hospital days were reduced. Despite differences in patient populations, study design, and treatment regimens, all 3 trials demonstrated clinical improvement associated with azithromycin treatment. Our trial allowed us to more completely assess safety and evaluate several important secondary outcomes.

The mechanism of action of azithromycin in patients with CF has not been elucidated. In vitro data suggest that azithromycin has bactericidal activity against P aeruginosa in stationary phase growth,²⁸ but azithromycin had minimal effect on bacterial density in this trial. Although azithromycin has activity against S aureus, this pathogen was not eradicated. The relative importance of Chlamydia pneumoniae and Mycoplasma pneumoniae in patients with CF is uncertain and was not evaluated in this trial, but the beneficial effect of azithromycin is unlikely to be mediated solely by impact on these organisms. Macrolide antibiotics show a modest degree of in vitro synergy with antipseudomonal antibiotics against gram-negative organisms from patients with CF.²⁹ An anti-inflammatory mechanism of action has been suggested for macrolide antibiotics³⁰ but none of the clinical trials have shown statistically significant reductions in IL-8. The impact on neutrophil elastase noted in this trial was modest and of uncertain clinical significance. In the study by Wolter et al,⁸ C-reactive protein declined in the azithromycin group but the C-reactive protein level was highly correlated with lung function as patients with normal C-reactive protein had better lung function. Macrolide antibiotics have been shown to decrease sputum viscosity^30- 31 but this has not been systematically studied. We speculate that the increased wheezing reported by participants in the azithromycin group was because of increased mobilization of respiratory tract secretions. Future studies should assess changes in sputum rheology. Finally, a direct effect on the CF transmembrane conductance regulator has been postulated.³²

The emergence of macrolide-resistant nontuberculous mycobacteria is of concern when considering chronic azithromycin therapy for patients with CF. In a multicenter study conducted in 1000 patients with CF aged 10 years or older, the prevalence of nontuberculous mycobacteria was approximately 12%, although many patients did not fulfill American Thoracic Society criteria for active disease.^33- 34 Because of these concerns, potential participants in the azithromycin study were screened for nontuberculous mycobacteria. Physicians considering azithromycin therapy should assess patients with CF for nontuberculous mycobacteria before and every 6 months after initiating azithromycin.²³

The results of this trial suggest several avenues for future research. Only 19 participants aged 6 to12 years were enrolled and therefore further investigations are needed in younger patients with CF infected with P aeruginosa. This study did not assess the efficacy of azithromycin in patients infected with B cepacia complex or uninfected with P aeruginosa. Evaluation of the effects and safety of azithromycin beyond 24 weeks is needed as the decline in lung function after discontinuing azithromycin suggests that sustained benefit could be derived from more prolonged therapy. The emergence of resistance to macrolide antibiotics in oral flora, potentially S aureus, and nontuberculous mycobacteria should be monitored during long-term therapy.

Despite the remaining questions, azithromycin is medication that is currently available. Care providers of patients with CF should consider this therapy for patients with CF aged 6 years or older and chronically infected with P aeruginosa.