Plasma retinol-binding protein predicts plasma retinol concentration in both infected and uninfected Cameroonian women and children

Abstract

Variation in the relationship between plasma retinol-binding protein (RBP) and retinol (ROH) has implications for vitamin A (VA) status assessment using RBP. Our objectives were to identify factors affecting the RBP:ROH relationship and to derive and evaluate population-specific RBP cutoffs for VA deficiency (VAD) in Cameroon. Plasma RBP, C-reactive protein (CRP), α1-acid-glycoprotein (AGP), and ROH concentrations were compared in a subsample of women 15-49 y (n = 121) and children 12-59 mo (n = 123) included in a national survey conducted in 2009. Plasma RBP and ROH were highly correlated (r = 0.94 for women; r = 0.96 for children; P < 0.001). Pregnancy and lactation altered the RBP:ROH relationship in women, but obesity, elevated CRP and AGP, age, and VA status did not. Among children, age altered the RBP:ROH relationship, but sex, stunting, VA status, and elevated CRP and AGP did not. Cutoffs for VAD derived using regression analysis were <1.17 μmol RBP/L for women (corresponding to <1.05 μmol ROH/L) and <0.83 μmol RBP/L for children (corresponding to <0.70 μmol ROH/L). The sensitivity and specificity of derived cutoffs were 81.8 and 93.0% for women and 94.7 and 88.9% for children, respectively. The infection-adjusted prevalence of low VA status (<1.17 μmol RBP/L) was 21.9% (95% CI = 18.7-25.0%) among women. Among children, the infection-adjusted prevalence of VAD (<0.83 μmol RBP/L) was 35.0% (95% CI = 31.1-39.0%). In conclusion, VAD remains a public health problem in Cameroon. The RBP:ROH relationship should be considered in surveys using RBP to assess VA status, and use of population-specific cutoffs may be advisable.