Two different doses of supplemental vitamin A did not affect mortality of normal-birth-weight neonates in Guinea-Bissau in a randomized controlled trial.
What this study found
The lower and higher neonatal vitamin A doses did not differ in their effect on mortality, and neither dose reduced infant mortality versus placebo. Mortality rate ratio for 25,000 IU versus 50,000 IU was 0.96 (95% CI 0.67, 1.38); 50,000 IU versus placebo was 1.31 (0.89, 1.93); and 25,000 IU versus placebo was 1.26 (0.85, 1.86). When both vitamin A groups were combined, the mortality rate ratio versus placebo was 1.28 (0.91, 1.81). Subgroup analyses by sex also showed no clear benefit, with boys at 0.79 (0.48, 1.30) and girls at 1.21 (0.70, 2.08).
- Study & population
- Three-arm randomized controlled trial in Guinea-Bissau among healthy normal-birth-weight neonates (>2500 g) enrolled at birth and given Bacille Calmette-Guérin vaccination.
- Intervention
- Neonatal vitamin A supplementation was given as a single oral dose at birth, using retinyl palmitate.
- Key limitation
- The main analysis was limited by relatively few deaths (160) over 4125 person-years, which reduces precision for detecting modest effects.
View source
Open PDFOriginal abstract
Whether neonatal vitamin A supplementation (NVAS) should be policy in areas with vitamin A deficiency is debated. We observed that a smaller dose of vitamin A may decrease mortality more than a larger dose and conducted a randomized, double-blind, placebo-controlled trial in Guinea-Bissau with the primary aim of comparing the effect of 50,000 with 25,000 IU neonatal vitamin A on infant mortality. The secondary aim was to study the effect of NVAS vs. placebo, including a combined analysis of NVAS trials. Between 2004 and 2007, normal-birth-weight neonates were randomly assigned in a 1:1:1 ratio to be administered 2 different doses of vitamin A (50,000 or 25,000 IU) or placebo. Infant mortality rates (MRs) were compared in Cox models providing MR ratios (MRRs). Among 6048 children enrolled, there were 160 deaths in 4125 person-years (MR = 39/1000). There was no difference in mortality between the 2 dosage groups: the MRR for 25,000 vs. 50,000 IU was 0.96 (95% CI: 0.67, 1.38). Neither dose of NVAS was associated with lower mortality than placebo (MRR = 1.28; 95% CI: 0.91, 1.81). In a combined analysis of the present trial and 2 previous NVAS trials in Guinea-Bissau, the effect of receiving NVAS (any dose) vs. placebo was 1.13 (95% CI: 0.94, 1.36) and differed significantly (P = 0.01) between boys (0.80; 95% CI: 0.58, 1.09) and girls (1.35; 95% CI: 1.04, 1.75). We could not confirm that a smaller dose of neonatal vitamin A reduces mortality more than a larger dose. We confirmed 2 other trials in Guinea-Bissau that showed no beneficial effect of NVAS. This trial was registered at clinicaltrials.gov as NCT00168610.