An intriguing study was published in last month’s New England Journal of Medicine. It was picked up by several news agencies, including Wired which published a piece by science journalist Maryn McKenna (the author of the “Big Chicken” book) that briefly made it to the top of Digg. A large cluster randomized trial of 1,533 communities (190,238 children at baseline census) in Malawi, Tanzania and Niger showed that 6-monthly azithromycin administered to children between the age of 1 month and 5 years reduced childhood mortality by an average of 13.5% compared to placebo (14.6 deaths per 1,000 person-years in communities that received azithromycin vs. 16.5 deaths per 1,000 person-years in communities that received placebo). The trial took place between 2014 and 2016.

The investigators – at least one of whom must have been a Tolkien fan – had previously studied the impact of mass administration of azithromycin for the prevention of trachoma (river blindness) in Ethiopia in 2006, and had incidentally noticed a resulting reduction in childhood deaths in communities that received the antibiotic. This finding had led to the initiation of this clinical cluster randomized trial 8 years later, sufficiently powered to detect a 10% reduction in childhood mortality.

The following image, obtained via screen capture of the supplementary material, shows the general study scheme. The investigators also wished to detect the impact of mass administration of azithromycin on subsequent antibiotics resistance (the additional 30 communities), but these results were not included in the published paper.

A closer read of the paper suggests that virtually all the results tilting towards benefit of mass azithromycin administration occurred in young children between 1-5 months of age, and in communities in Niger. Childhood mortality for communities involved in this study were by the highest in Niger as well (azithromycin arm: 22.5 per 1,000 person-years vs. 9.1 and 5.4 in Malawi and Tanzania respectively; placebo arm: 27.5 per 1,000 person-years in Niger vs. 9.6 and 5.5 in Malawi and Tanzania respectively).

The causes of death were largely unknown, although “verbal autopsies” were obtained for 250 childhood deaths in the communities in each country. Malaria figured prominently, especially in the Niger, along with diarrheal diseases and pneumonia. How azithromycin reduced childhood mortality especially in the Niger was not clear at all, given that it was administered only twice a year, although the authors suggested that this could be due to the prophylactic effect of the drug against different infectious diseases including the top 3 most common causes of death listed above. Would the impact have been even greater (or would it taper off) if the antibiotics were administered more frequently, i.e. quarterly or every 2 months?

The other obvious concern is antimicrobial resistance. Mass administration of azithromycin for trachoma treatment is well known to result in the subsequent appearance of macrolide-resistant bacteria, including Streptococcus pneumoniae and Escherichia coli. On balance, however, this is the lesser harm in the short-medium term (as opposed to increased rates of trachoma/blindness and childhood mortality) for those communities where childhood mortality from infection is very high.

Although I believe that the impact of antibiotics will probably wane as the socioeconomic status of the communities/countries rise and as access to clean water and health services improve. Giving antibiotics in a similar fashion (as prophylaxis) to children in Singapore – where the under-5 mortality is very low – would probably have many downsides and virtually no clear-cut benefits.