The Real-Life Effect of Vitamin a Supplementation Campaigns on All-Cause Mortality. Systematic Review and Meta-Analysis

Abstract

Background: WHO recommends high-dose vitamin A supplementation (VAS) in vitamin A-deficient areas. Many countries deliver VAS in national campaigns. Based on randomised trials conducted in the 1980s-1990s, such campaigns are assumed to reduce all-cause mortality by 12-24%. There has been no analysis of the real-life effect of VAS campaigns on all-cause mortality.

Methods: We conducted a systematic review and a meta-analysis of the real-life effect of “VAS” campaigns (defined as VAS +/- simultaneous deworming) on all-cause mortality. In March 2024 we searched Pubmed and Embase for studies involving children<5 years of age, that investigated VAS campaigns based on individual-level data and reported all-cause mortality. We did not include modelling studies. ROBINS-I was used to assess the methodological quality. Some campaigns provided VAS in combination with vaccines (oral polio vaccine (OPV) or measles vaccine (MV)). Meta-estimates were calculated as random (prioritized) and fixed effect estimates, overall and by biological sex and age group. 

Findings: We identified five observational studies covering the period from 1999-2019, including a total of 163,996 children and 5,449 deaths. Four studies were of median risk of bias, one of critical risk of bias. Overall, the meta-analysis showed no benefit of VAS campaigns on all-cause mortality, the random effect estimate being 1.07 (0.82-1.38)(fixed effect estimate 1.17 (1.03-1.32)). In males, the meta-estimate was 1.18 (0.91-1.52) and in females 1.06 (0.85-1.33). Among infants between 6-11 months, the effect estimate was 1.24 (1.00-1.54). In combined “VAS+OPV” campaigns the effect estimate was 1.05 (0.97-1.14); for “VAS+MV” it was 1.17 (0.88-1.57). For “Any VAS” (VAS +/- deworming and vaccines), the estimate was 1.07 (0.99-1.16). In comparison, the estimate for “OPV-only” campaigns across the same studies was 0.78 (0.71-0.85)(p for same effect of “Any VAS” vs “OPV-only” campaigns<0.001; p for same effect of “VAS+OPV” vs “OPV-only”<0.001).

Interpretation: There was no evidence that VAS campaigns contribute to lowering all-cause mortality. On the contrary, there were signals of harm in the youngest children. In the same settings, “OPV-only” campaigns had markedly beneficial effect. The effect of VAS may have changed over time. VAS campaigns should be stopped until randomised trials have documented their current effect.  

Funding: No funding was available for the work.

Declaration of Interest: The authors have no conflicts of interest that are directly relevant to the content of this article.

Keywords: Vitamin A supplementation, low- and middle income countries, children, all-cause mortality, oral polio vaccine, campaigns