A comparative study on indicators of vitamin A status and risk factors for sensitivity and specificity of the methods to detect vitamin A deficiency

Abstract

Background Serum retinol (SR) and retinol-binding protein (RBP) are commonly used indicators, but they are affected by infections and inflammation. This study aimed to assess the sensitivity and specificity of VA indicators to detect vitamin A deficiency (VAD) in 36–59-month-old children living in a rural area in Burkina Faso. Methods In a community-based study, two cross-sectional surveys were carried out from November 2016 to September 2017 in the health district of Dandé in Burkina Faso. The surveys included 115 children 36–59 months old. Indicators of VA and inflammation assessed in all children included SR, RBP and total liver VA reserves (TLR) estimated by retinol isotope dilution, and inflammation markers (C-reactive protein (CRP) and alpha 1-acid glycoprotein (AGP)). We calculated the sensitivity, specificity, positive and negative predictive values. In addition, the effects of inflammation, helminth infection, and season on sensitivity and specificity were assessed. Results The prevalence of VAD assessed by SR (< 0.7 µmol/L), RBP (< 0.7 µmol/L), and TLR (< 0.1 µmol/g liver) were, respectively, 30.9%, 33.3%, and 0%. Compared to TLR, the specificity, positive predictive value, and negative predictive value of SR were 71.1%, 0%, and 100%, and of RBP, were 68.9%, 0%, and 100%, respectively. The sensitivity was indeterminable for SR and RBP. The specificity of SR and RBP was lower during the dry season. Elevated CRP (> 5.0 mg/L) and AGP (> 1.0 g/L) were detected in 1.9% and 28.6% of children, respectively. The adjustment of VA indicators for inflammation improved SR’s specificity to 75.9% and decreased RBP’s specificity to 67.8%. Conclusion No cases of VAD were identified by TLR. However, (inflammation-adjusted) SR and RBP had varying accuracy in the estimation of VAD.