RESUMO
OBJECTIVE: To compare the prevalence of vitamin A deficiency (VAD) among Cambodian preschool children as determined by the retinol-binding protein-enzyme immunoassay (RBP-EIA) and direct measurement of serum retinol by high-performance liquid chromatography (HPLC). SETTING AND SUBJECTS: Sera from 359 children were randomly selected from archived specimens collected in a national VAD prevalence survey in Cambodia. METHODS: Sera were first analyzed for retinol content by HPLC and then subjected to analysis using RBP-EIA to determine serum RBP concentrations. National Institute of Standards and Technology and control sera were used to ensure quality and accuracy for each set of analyses. To classify VAD, the same cutoff point of <0.70 micromol/l was employed for each indicator. RESULTS: Overall, the prevalence of VAD based on serum retinol was 22.3% (95% confidence interval (CI): 18.0, 26.6), whereas the RBP-EIA indicated a VAD prevalence of 20.9% (95% CI: 16.7, 25.1). A simple linear regression model indicated an R2 of 0.79, and a receiver operating curve analysis revealed an area under the curve of 0.92. CONCLUSIONS: We found no significant difference between the results of RBP-EIA compared to retinol analyzed by HPLC in estimating the prevalence of VAD. Use of the test could enable public health authorities to assess the extent of VAD and track progress in control programs in resource-poor settings.
