Analysis of antimicrobial resistance in the LAKANA trial of azithromycin administration in Mali: A sub-study of a cluster randomized clinical trial

Background Mass azithromycin administration has been used to reduce under-five mortality in sub-Saharan Africa, but concerns remain about its potential impact on antimicrobial resistance (AMR). Methods and findings The Large-scale Assessment of the Key health-promoting Activities of two New mass drug administration regiments with Azithromycin (LAKANA) trial was a cluster-randomized, placebo-controlled trial conducted in Mali, enrolling infants aged 1−11 months randomized to placebo, twice-yearly azithromycin, or quarterly azithromycin for 2 years. The trial did not find any mortality benefit amongst treated children. The current analysis is a sub-study of the parent trial examining AMR outcomes. Repeated cross-sectional nasopharyngeal and rectal swabs were collected from children aged 4–14-months in villages randomized to placebo, twice-yearly or quarterly azithromycin, and from untreated contacts aged 49–59-months at baseline, 12-, 24-, and 36-months post-village enrollment across 59 LAKANA trial villages. Phenotypic resistance to macrolide and nonmacrolide antibiotics was assessed in Streptococcus pneumoniae ( S. pneumoniae ) and Escherichia coli ( E. coli ) isolates using disk diffusion or E-tests. The primary noninferiority analysis evaluated macrolide resistance prevalence at 36-months among younger children and at 24-months among older contacts, with predefined margin of +10 percentage points. A total of 5,025 nasopharyngeal and 5,021 rectal samples were analyzed. At baseline, among 4–14-month-old children, macrolide resistance in S. pneumoniae was 16.6%, 15.5%, and 11.1% in the placebo, twice-yearly and quarterly azithromycin groups, respectively, and 5.0%, 5.0%, and 0.0% in E. coli . At 36-months, corresponding prevalences were 13.4%, 14.5%, and 17.2% for S. pneumoniae , and 5.1%, 6.4%, and 9.1% for E. coli . The difference between placebo and twice-yearly groups was 1.0% (95% CI [−4.7, 6.8]) for S. pneumoniae and 1.3% (95% CI [−2.4, 4.9]) for E. coli ; corresponding differences for quarterly azithromycin group was 3.8% (95% CI [−3.3, 10.9]) for S. pneumoniae and 4.0% (95% CI [−0.7, 8.7]) for E. coli. Among untreated older children, there were no marked inter-group differences in macrolide resistance either at baseline or at 24-months. The study may not have been large enough to detect a small increase in macrolide resistance in S. pneumoniae among young children in villages that received quarterly azithromycin. Conclusions Twice-yearly azithromycin administration to 1–11-month-old infants for 2 years in Mali did not increase macrolide resistance in treated children or untreated contacts and met the predefined noninferiority margin in tested isolates. Quarterly administration, however, cannot exclude a modest dose-dependent increase of macrolide resistance in S. pneumoniae and E. coli .