Effect of Mass Azithromycin Distributions on Childhood Growth in Niger: A Cluster-Randomized Trial

Abstract

Importance: Mass azithromycin distributions may decrease childhood mortality, although the causal pathway is unclear. The potential for antibiotics to function as growth promoters may explain some of the mortality benefit.

Objective: To investigate whether biannual mass azithromycin distributions are associated with increased childhood growth.

Design, setting, and participants: This cluster-randomized trial was performed from December 2014 until March 2020 among 30 rural communities in Boboye and Loga departments in Niger, Africa, with populations from 200 to 2000 individuals. Communities were randomized in a 1:1 ratio to biannual mass distributions of azithromycin or placebo for children ages 1 to 59 months. Participants, field-workers, and study personnel were masked to treatment allocation. Height and weight changes from baseline to follow-up at 4 years were compared between groups. Data were analyzed from June through November 2021.

Interventions: Participants received azithromycin at 20 mg/kg using height-based approximation or by weight for children unable to stand every 6 months at the participants' households. Placebo contained the vehicle of the azithromycin suspension.

Main outcomes and measures: Longitudinal anthropometric assessments were performed on a random sample of children before the first treatment and then annually for 5 years. Height and weight were the prespecified primary outcomes.

Results: Among 3936 children enrolled from 30 communities, baseline characteristics were similar between 1299 children in the azithromycin group and 2637 children in the placebo group (mean 48.2% [95% CI, 45.5% to 50.8%] girls vs 48.0% [95% CI, 45.7% to 50.3%] girls; mean age, 30.8 months [95% CI, 29.5 to 32.0 months] vs 30.6 months [95% CI, 29.2 to 31.6 months]). Baseline anthropometric assessments were performed among 2230 children, including 985 children in the azithromycin group and 1245 children in the placebo group, of whom follow-up measurements were available for 789 children (80.1%) and 1063 children (85.4%), respectively. At the prespecified 4-year follow-up visit, children in the azithromycin group gained a mean 6.7 cm (95% CI, 6.5 to 6.8 cm) in height and 1.7 kg (95% CI, 1.7 to 1.8 kg) in weight per year and children in the placebo group gained a mean 6.6 cm (95% CI, 6.4 to 6.7 cm) in height and 1.7 kg (95% CI, 1.7 to 1.8 kg) in weight per year. Height at 4 years was not statistically significantly different between groups when adjusted for baseline height (0.08 cm [95% CI, -0.12 to 0.28 cm] greater in the azithromycin group; P = .45), and neither was weight when adjusted for height and baseline weight (0.02 kg [95% CI, -0.10 to 0.06 kg] less in the azithromycin group; P = .64). However, among children in the shortest quartile of baseline height, azithromycin was associated with a 0.4 cm (95% CI, 0.1 to 0.7 cm) increase in height compared with placebo.

Conclusions and relevance: This study did not find evidence of an association between mass azithromycin distributions and childhood growth, although subgroup analysis suggested some benefit for the shortest children. These findings suggest that the mortality benefit of mass azithromycin distributions is unlikely to be due solely to growth promotion.

Trial registration: ClinicalTrials.gov Identifier: NCT02048007.

Figure 1.. Study Flow Diagram

There were 30 clusters drawn from the same pool of eligible communities as the main Macrolides Oraux pour Réduire les Décès avec un Oeil sur la Résistance (MORDOR) trial and randomized to biannual mass administration with azithromycin or placebo for 5 years. A random sample of children was selected from the baseline census for anthropometric monitoring and followed annually for 5 years. ^aData were missing from 1 entire community owing to technical problems; this community was known to be treated. Population and drug coverage estimates omit this community with missing data. ^bOne community refused participation after month 36 and was not included in estimates of population or drug coverage. ^cOne azithromycin cluster refused to participate after month 36, but follow-up data from earlier points was included in analyses. ^dMeasurements from a malfunctioning scale were excluded from 3 communities in the azithromycin group and 5 communities in the placebo group at the month 24 study visit.

Figure 2.. Height Over Time

A random sample of children was monitored annually over 5 years in the azithromycin and placebo group. Each thin line indicates an individual child’s growth curve, colored according to the baseline quartile of height; heavy lines, mean of the treatment group in the respective subgroup.