Background: The relative dose response (RDR) test, which quantifies the increase in serum retinol after vitamin A administration, is a qualitative measure of liver vitamin A stores. Particularly in preterm infants, the feasibility of the RDR test involving blood is critically dependent on small sample volumes. Objectives: This study aimed to assess whether the RDR calculated with retinol-binding protein 4 (RBP4) might be a substitute for the classical retinol-based RDR test for assessing vitamin A status in very preterm infants. Methods: This study included preterm infants with a birth weight below 1,500 g (n = 63, median birth weight 985 g, median gestational age 27.4 weeks) who were treated with 5,000 IU retinyl palmitate intramuscularly 3 times a week for 4 weeks. On day 3 (first vitamin A injection) and day 28 of life (last vitamin A injection), the RDR was calculated and compared using serum retinol and RBP4 concentrations. Results: The concentrations of retinol (p < 0.001) and RBP4 (p < 0.01) increased significantly from day 3 to day 28. On day 3, the median (IQR) retinol-RDR was 27% (8.4-42.5) and the median RBP4-RDR was 8.4% (-3.4 to 27.9), compared to 7.5% (-10.6 to 20.8) and -0.61% (-19.7 to 15.3) on day 28. The results for retinol-RDR and RBP4-RDR revealed no significant correlation. The agreement between retinol-RDR and RBP4-RDR was poor (day 3: Cohen's κ = 0.12; day 28: Cohen's κ = 0.18). Conclusion: The RDR test based on circulating RBP4 is unlikely to reflect the hepatic vitamin A status in preterm infants.