ABSTRACT
Objectives: To assess retinol binding protein (RBP) as an indicator of vitamin A status in children and women, during a pilot program fortifying cooking oil with retinyl palmitate in West Java. Methods: Surveys measured serum retinol using High-Performance Liquid Chromatography (as gold standard), and RBP using Enzyme-Linked Immunosorbent Sandwich Assay, among children 6-59 months (n=753) and 5-9 years (n=175), women 15-29 years not reportedly pregnant (n=170) and lactating mothers (n=303). C-reactive protein and α-1-acid glycoprotein were measured to adjust serum retinol for sub-clinical inflammation using established correction factors. Results: At baseline, serum retinol concentrations (μg/dL, mean and standard deviation) were 33.2±14.0, 33.0±0.91, 40.9±18.7, and 44.2±18.5 for children 6-59 months, 5-9 years, women and lactating mothers. Prevalences of vitamin A deficiency (retinol <20μg/dL) were 14.9%, 10.9%, 10.0% and 5.3%. Mean RBP concentrations (mg/kg) were 1.00±0.24, 1.01±0.22, 1.24±0.35, 1.34±0.37, respectively. Serum retinol and RBP concentrations correlated positively (Pearson's correlation coefficient r=0.54, 0.50, 0.51, and 0.48 in the four groups). In the 2nd survey a year later, retinol and RBP averages were 3.2-17.4% and 3.7-13.7% increased among groups; correlation patterns were similar. Using RBP<0.80 μmol/L as cut-off (lacking a WHO recommendation), deficiency prevalence, sensitivity and specificity of RBP were 15%, 36% and 88% in children 6-59 months. Conclusions: RBP can be an indicator of vitamin A status in child and woman populations, saving cost compared to serum retinol-based monitoring. The lesser retinol-RBP concordance compared to earlier studies suggest that precision of measurements in this study may have been less than typical.