Azithromycin versus placebo for the treatment of HIV-associated chronic lung disease in children and adolescents (BREATHE trial): study protocol for a randomised controlled trial

BREATHE (Bronchopulmonary function in response to azithromycin treatment for chronic lung disease in HIV-infected children) is a two-site, double-blind, placebo-controlled, individually randomised trial in which we intend to enrol 400 perinatally HIV-infected children and adolescents aged 6–19 years with HIV-associated chronic lung disease who have been receiving ART for a minimum of 6 months. Participants will be enrolled from the outpatient HIV clinics in Harare, Zimbabwe, and Blantyre, Malawi.

Participants are randomly assigned to receive either azithromycin or placebo in a 1:1 ratio. Weight-band azithromycin (10–19.9 kg, 250 mg; 20–29.9 kg, 500 mg; 30–39.9 kg, 750 mg; > 40 kg, 1250 mg/week) or placebo will be given weekly under direct observation by a treatment monitor identified within the family for a total of 12 months.

HIV-infected children and adolescents with HIV-associated chronic lung disease attending outpatient HIV clinics in Harare and Blantyre will be enrolled in the trial. Individuals aged 6–19 years will be approached together with their guardians during their routine HIV outpatient visits and provided with information about the study. Once informed consent is obtained for screening, eligibility will be established using a multi-step screening procedure. Chronic lung disease will be established by spirometry (forced expiratory volume in 1 second [FEV₁] z-score less than −1.0) with no reversibility (< 12% improvement in FEV₁ after salbutamol 200 μg inhaled using a spacer). Spirometry will be performed using the EasyOne™ spirometer (ndd Medical Technologies Inc., Andover, MA, USA) by trained research staff certified in performing spirometry and following the American Thoracic Society guidelines. Those who meet the criteria for chronic lung disease will undergo further tests, including a urine pregnancy test (for girls who have reached menarche), serum creatinine, alanine aminotransferase (ALT), an electrocardiogram and screening for tuberculosis (TB). For TB screening, we will use the Xpert™ MTB/RIF (Cepheid, Sunnyvale, CA, USA) to test one sputum sample obtained either spontaneously or through induction.

Perinatally HIV-infected children and adolescents aged 6–19 years who have been receiving ART for at least 6 months for HIV-associated chronic lung disease, who have a firm home address and a stable guardian, and with consent from the guardian and assent from the participant (for those aged < 18 years; those aged ≥ 18 years to consent independently) will be eligible for inclusion. Exclusion criteria will include having a condition that may prove fatal during the study period (e.g., malignancy), TB or acute respiratory tract infection at the time of screening, pregnancy or breastfeeding, history of cardiac arrhythmia, a prolonged QTc interval, abnormal creatinine clearance or elevated ALT, known macrolide hypersensitivity, and concomitant use of digoxin and/or fluconazole (or other drugs known to prolong the QTc interval) (Table 1).

Eligible individuals will be randomised to either azithromycin or placebo. The allocation ratio will be 1:1 by block randomisation with variable-length blocks stratified by site. The randomisation schedule and allocation list will be generated using Stata™ version 14 software (2015 release; StataCorp, College Station, TX, USA) by an independent statistician not connected with the study. The allocation list will be sent directly to the study pharmacists at both trial sites, who will prepare the study medication. Participants and study personnel will therefore be masked to treatment allocation. The pharmacist will be provided with only randomly allocated numbers assigning participants to arm 1 or arm 2, with each number linked to a study number, and therefore the pharmacist will also remain blinded. It will not be possible to have pre-prepared medication packs, owing to changes of dose with weight.

After randomisation, spirometry will be repeated, and cardiopulmonary fitness will be evaluated using the incremental shuttle walk test. Transthoracic echocardiography will be performed by a trained echocardiographer using the SonoSite™ M-Turbo echocardiography system (FUJIFILM SonoSite, Bothell, WA, USA) at the Malawi site and the Mindray™ DC-N6 echocardiography system (Mindray, Shenzhen, China) at the Zimbabwe site to assess right heart function and pulmonary hypertension. The research nurse will take a rectal swab, nasal aspirate, and sputum and blood samples. The blood sample will be used to measure CD4 count with the Pima™ Analyser (Alere, Orlando, FL, USA) and to perform viral load testing (Xpert™ HIV-1 Viral Load; Cepheid). The study drug will be dispensed by the study pharmacist. After the enrolment visit, follow-up visits will be scheduled at 2 weeks and 3-monthly thereafter. Participants will be followed for a minimum of 12 months to ascertain the primary outcome and for a further 6 months (i.e., 6 months following completion of study drug) to investigate durability of effect (if any) of the intervention. Procedures undertaken at the follow-up visits are summarised in Fig. 1.

Potential risks include an allergic reaction or adverse reaction to the medication or placebo. Examples of potential side effects include nausea, vomiting, diarrhoea and skin reactions. Each of these risks and any other unexpected outcomes will be monitored at study visits. Criteria for stopping the study drug are allergic reaction, QTc prolongation > 500 milliseconds, hepatic toxicity with ALT > 80 IU/L, concomitant use of drugs known to produce QTc prolongation, pregnancy during the course of the study and any adverse reaction with DAIDS severity grade > 3. In case of any serious adverse events, the institutional review boards and the data and monitoring safety board (DSMB) will be notified. Unblinding of treating physicians will be carried out in case of pregnancy or other events after decision by the DSMB.

The primary outcome is the mean difference between trial arms in FEV₁ z-scores (generated using Global Lung Function Initiative reference standards) between trial arms at 12 months after initiation of the study drug, adjusted for site and baseline FEV₁ z-score. The secondary outcomes are described in Table 2.

Other outcomes will be the effect of azithromycin on (1) diversity and composition of respiratory and gut microbiome of children with HIV-associated chronic lung disease, (2) antibiotic resistance of bacteria colonising the respiratory tract, and (3) biomarkers of systemic inflammation.

Nasopharyngeal and sputum samples will be frozen and stored at −80 °C for later batch processing, which will include routine bacterial culture for respiratory pathogens, antimicrobial susceptibility testing and 16S ribosomal RNA (rRNA) amplicon sequencing (microbiome analysis). Microbiome analysis of stool samples will also be done.

Archived plasma samples (isolated and stored at −80 °C) will be used to quantify the levels of soluble markers of immune activation and microbial translocation, including high-sensitivity C-reactive protein, d-dimers, interleukin-6, β₂-microglobulin and soluble CD14 using multiplex bead assays (Luminex, Austin, TX, USA). Levels of bacterial 16S rRNA will be quantified using qRT-PCR to measure the extent of microbial translocation over time. The levels of these markers will be correlated with FEV₁ z-scores and compared between trial arms.

Research nurses and assistants collect the data at baseline and follow-up and record it on electronic clinical record forms using Google Nexus™ tablets (Google, Mountain View, CA, USA) running OpenDataKit software. In both countries, data will be uploaded, processed and saved to a Microsoft Access database (Microsoft, Redmond, WA, USA) before being exported to Stata version 14.0 software and merged into a single database backed up on a regular basis. Consistency checks and checks for missing data are performed both at data entry and fortnightly once the database has been merged. Laboratory data will be recorded on paper forms and entered into the database using optical character recognition. Spirometric data will be uploaded directly from the spirometer. Echocardiographic data will be recorded on electronic clinical record forms, and these will be processed and saved to the Microsoft Access database for merging with the main trial database.

The primary analysis will be done on a modified intention-to-treat basis [29]. Secondary analyses will include a per-protocol analysis which will be defined using all adherence data prior to unblinding the study. Continuous outcomes will be compared between treatment groups, adjusting for site and baseline measures of the outcome and using linear regression to estimate the mean difference and corresponding 95% CI. Time-to-event outcomes will be assessed using Cox proportional hazards regression and graphically displayed using Kaplan-Meier estimates. Between-group comparisons of binary outcomes will be analysed with logistic regression to estimate ORs and 95% CIs. Count data (e.g., number of hospitalisations) will be analysed using Poisson regression to estimate incidence rate ratios and 95% CIs. All analyses will be adjusted for site. Pre-specified effect modification analyses will include site and baseline severity of lung disease.

The following assumptions were made to calculate the sample size:

Under these assumptions, a sample size of 400 recruited participants and 300 participants with outcome data (25% unassessable based on previous studies of Zimbabwean children) will enable 80% power to detect a difference in trial arm means ranging between 0.17 and 0.23, an effect size (difference in means/SD) of 0.32 (Fig. 2).

The trial will be conducted in accordance with the principles of the Declaration of Helsinki, in compliance with the protocol approved by relevant ethics committees, and according to good clinical practice standards [30]. No change in the protocol will be implemented without the prior review and approval of the regulatory authorities, except where it may be necessary to eliminate an immediate hazard to the trial participants. Details concerning the enrolment process can be found in the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) checklist (Additional file 1) [31].

The trial sponsor is the London School of Hygiene and Tropical Medicine (LSHTM). The trial will be monitored by the Clinical Trials Unit of the LSHTM and externally by the University of Zimbabwe Research Support Centre in Zimbabwe and the Malawi-Liverpool Wellcome Trust Clinical Research Programme Clinical Trials Unit in Malawi.

Results of this trial will be published on completion in a peer-reviewed scientific journal. The funder will not be involved in the analysis or interpretation of the data. The full anonymised dataset will be made available no longer than 18 months after completion of the trial.