Hepatic stores of vitamin A were estimated in 31 Bangladeshi surgical patients (15 males and 16 females) by the deuterated-retinol-dilution (DRD) technique and by analysis of the vitamin A concentration of a liver biopsy specimen obtained during previously scheduled abdominal surgery. Patients ranged in age from 21 to 65 y and had an average body mass index (BMI: in kg/m2) of 17.7 +/- 3.4. They received 0.753 mumol [2H4]retinyl acetate/kg body wt orally 9-11 d before surgery. Hepatic vitamin A reserves were estimated according to Furr et al (Am J Clin Nutr 1989;49:713-6) by using a single plasma isotopic-ratio measurement (18-25 d postdose). Estimated mean hepatic vitamin A stores were similar by both techniques, 0.110 +/- 0.072 mmol (by DRD) compared with 0.100 +/- 0.067 mmol (by biopsy). Regression analysis was used to compare results of the DRD and biopsy techniques. A significant linear relation was found between the two techniques (r = 0.75, P < 0.0001), and the least-squares regression line was not significantly different from y = x (P = 0.09). The results indicate that the DRD technique provided a very good estimate of hepatic vitamin A reserves for this population. However, a wide prediction interval was observed for estimates of hepatic vitamin A reserves for individual subjects. Thus, further refinement of the prediction model is necessary to improve estimates of hepatic vitamin A reserves for individual subjects.