Effects of subclinical infection on plasma retinol concentrations and assessment of prevalence of vitamin A deficiency: meta-analysis.

BACKGROUND Vitamin A deficiency adversely affects child morbidity and survival. This deficiency is estimated by measurement of plasma retinol concentrations, but because plasma retinol is reduced by clinical and subclinical infection, this proxy measure can lead to overestimation. Infection and trauma are accompanied by rises in concentrations of acute-phase proteins in plasma. We aimed to estimate vitamin A deficiency more accurately by measuring changes in plasma retinol and acute-phase proteins associated with subclinical infection or convalescence. METHODS We analysed data for concentrations of plasma retinol and one or more acute-phase proteins (alpha1-acid-glycoprotein, alpha1-antichymotrypsin, C-reactive protein, or serum amyloid A) from 15 studies of apparently healthy individuals. We generated summary estimates of differences in retinol concentrations for incubation, early, and late convalescent phases of infection between people with none and those with one or more raised acute-phase proteins. We compared these groups in two, three, and four group analyses. We also compared a subgroup of apparently healthy preschool (1-5 years) children with results from all other studies. FINDINGS For all four proteins, retinol values were much higher in people with normal concentrations of protein, than in individuals with raised concentrations (16% higher for alpha1-antichymotrypsin, 18% for alpha1-acid-glycoprotein, 25% for C-reactive protein, and 32% for serum amyloid A). Estimates of the reduction in plasma retinol for individuals with infection compared with healthy individuals, were 13% (incubation), 24% (early convalescent), and 11% (late convalescent). Estimates of vitamin A deficiency in individuals with no raised acute-phase proteins (healthy group) were much the same as those obtained by adjustment of plasma retinol concentrations in the whole group using acute-phase proteins. INTERPRETATION We recommend that surveys to estimate vitamin A deficiency should include measurements of serum C-reactive protein and alpha1-acid-glycoprotein concentrations. Information about acute-phase proteins will enable plasma retinol concentrations to be corrected where sub-clinical infection exists, and the healthy sub-group to be identified.