Vitamin A biomarkers were associated with α(1)-acid glycoprotein and C-reactive protein over the course of a human norovirus challenge infection.

Abstract

Retinol binding protein (RBP) is used as a proxy for retinol in population-based assessments of vitamin A deficiency (VAD) for cost-effectiveness and feasibility. When the cut-off of < 0·7 μmol/l for retinol is applied to RBP to define VAD, an equivalence of the two biomarkers is assumed. Evidence suggests that the relationship between retinol and RBP is not 1:1, particularly in populations with a high burden of infection or inflammation. The goal of this analysis was to longitudinally evaluate the retinol:RBP ratio over 1 month of follow-up among fifty-two individuals exposed to norovirus (n 26 infected, n 26 uninfected), test whether inflammation (measured as α-1-acid glycoprotein (AGP) and C-reactive protein (CRP)) affects retinol, RBP and the ratio between the two and assess whether adjusting vitamin A biomarkers for AGP or CRP improves the equivalence of retinol and RBP. We found that the median molar ratio between retinol and RBP was the same among infected (0·68) and uninfected (0·68) individuals. AGP was associated with the ratio and RBP individually, controlling for CRP, and CRP was associated with both retinol and RBP individually, controlling for AGP over 1 month of follow-up. Adjusting for inflammation led to a slight increase in the ratio among infected individuals (0·71) but remained significantly different from the expected value of one. These findings highlight the need for updated recommendations from the WHO on a cut-off value for RBP and an appropriate method for measuring and adjusting for inflammation when using RBP in population assessments of VAD.