Background: Serum retinol concentration determined by high-performance liquid chromatography (HPLC) is recommended by the World Health Organization to assess population vitamin A status. This assay is expensive, technically demanding and rarely available in developing countries. Our objective was a) to assess the diagnostic performance of proxy markers in detecting vitamin A deficiency and b) to derive decision rules based on these markers to estimate vitamin A deficiency prevalence.
Methods: A survey was conducted in 15 rural primary schools in Eastern Province, Kenya, with 375 children aged 6 to 12 years (25 randomly selected per school). Serum retinol concentration <0.70 μmol/L by HPLC was used to define vitamin A deficiency. Proxy markers for vitamin A deficiency were serum concentrations of retinol binding protein (RBP), transthyretin, retinol measured by fluorometry and RBP:transthyretin molar ratio.
Results: The prevalence of vitamin A deficiency (HPLC) was 18%. Transthyretin and RBP showed the best diagnostic performance individually, with area-under-the-curve (AUC) values of 0.96 and 0.93. When combined, and with C-reactive protein added, the AUC increased to 0.98. A simple decision rule {(-15.277 × [RBP, μmol/L] - 7.013 × [Transthyretin, μmol/L] + 0.367 × [C-reactive protein, mg/L] + 24.714) > 0.496} yielded prevalence estimates of vitamin A deficiency that is unbiased by diagnostic error.
Conclusions: The combination of transthyretin, RBP and C-reactive protein concentrations could eventually replace retinol concentration by HPLC in resource-poor settings as the preferred method to assess the population burden of vitamin A deficiency.