Provitamin A–biofortified maize increases serum β-carotene, but not retinol, in marginally nourished children: a cluster-randomized trial in rural Zambia

Background: Vitamin A deficiency remains a nutritional concern in sub-Saharan Africa. Conventionally bred maize hybrids with high provitamin A carotenoid concentrations may have the potential to improve vitamin A status in maize-consuming populations. Objective: We evaluated the efficacy of regular provitamin A carotenoid–biofortified “orange” maizemeal (∼15 μg β-carotene/g) consumption in improving vitamin A status and reducing vitamin A deficiency in children. Design: This was a cluster-randomized controlled trial in the rural farming district of Mkushi, Zambia. All 4- to 8-y-old children in an ∼400-km² area were identified and grouped by proximity into clusters of ∼15–25 children. We randomly assigned clusters to 1) orange maizemeal (n = 25), 2) white maizemeal (n = 25), or 3) a parallel, nonintervention group (n = 14). Children in intervention clusters (n = 1024) received 200 g maizemeal for 6 d/wk over 6 mo; the maizemeal was prepared according to standardized recipes and served in cluster-level kitchens. Staff recorded attendance and leftovers. We collected venous blood before and after the intervention to measure serum retinol, β-carotene, C-reactive protein, and α₁-acid glycoprotein. Results: Intervention groups were comparable at baseline, and vitamin A status was better than anticipated (12.1% deficient on the basis of serum retinol <0.7 μmol/L). Although attendance at meals did not differ (85%), median daily maize intake was higher in white (154 g/d) than in orange (142 g/d) maizemeal clusters. At follow-up, mean serum β-carotene was 0.14 μmol/L (95% CI: 0.09, 0.20 μmol/L) higher in orange maizemeal clusters (P < 0.001), but mean serum retinol (1.00 ± 0.33 μmol/L overall) and deficiency prevalence (17.1% overall) did not differ between arms. Conclusion: In this marginally nourished population, regular biofortified maizemeal consumption increased serum β-carotene concentrations but did not improve serum retinol. This trial was registered at clinicaltrials.gov as NCT01695148.