TEST-RETEST REPRODUCIBILITY OF THE RELATIVE DOSE-RESPONSE FOR VITAMIN-A STATUS IN GUATEMALAN ADULTS - ISSUES OF DIAGNOSTIC SENSITIVITY

The relative dose reponse (RDR) test has been used as a functional measure of whole-body stores of vitamin A in humans. We have examined the reproducibility of the RDR procedure in a population of Guatemalan adult subjects who would be expected to show a moderate prevalence of hypovitaminosis A. Fifty one subjects were administered a standard RDR test, and the plasma samples were analyzed for retinol, tocopherol, retinol binding protein (RBP) and prealbumin (PAL). Thirty-four of the subjects underwent repeat RDR tests 7 d later. Plasma levels in fasted subjects were as follows: retinol, 1.35 ± 0.30 μmol/L; RBP, 37.8 ± 7.7 mg/L; PAL, 187.0 ± 39.0 mg/L; and tocopherol, 16.6 ± 6.2 μmol/L. RDRs ranged from -35.2% to +63.1%, with a mean of 2.6 ± 10.4%. Overall, we observed poor within-subject reproducibility of the RDR procedure whether expressed numerically or by diagnostic classification. Moreover, in contrast to previous studies in children, we observed fewer positive RDR tests than would be expected for the population studied. Nevertheless, the mean RDR was inversely proportional to fasting retinol levels, thus confirming the validity of the biological basis of the RDR procedure in humans. Because of high intra-individual variability with this test, investigators should be cautious when using the RDR procedure in serial studies to monitor the efficacy of therapeutic interventions or subject compliance to dietary regimens.