To determine the efficacy of cooking food in iron pots to prevent anemia in premature infants, a longitudinal study on iron nutritional status was conducted in preterm, healthy infants from families of low socioeconomic level between mo 4 and 12 of life. The infants were divided randomly into two groups. The study group consisted of 22 infants whose food was cooked in iron pots; the control group consisted of 23 infants whose food was cooked in aluminum pots. Supplemental iron [2 mg/(kg.d)] was recommended from 15 d to 12 mo of age for both groups. At 12 mo of age, the group fed food cooked in iron pots had significantly better hematologic values than the group fed food cooked in aluminum pots. Differences included hemoglobin (116 +/- 16 vs. 103 +/- 20 g/L, P = 0.02), hematocrit (0.35 +/- 0.04 vs. 0.31 +/- 0.05, P = 0.005), mean corpuscular volume (72.1 +/- 10.4 vs. 62.7 +/- 11.1 fL, P = 0.005), free erythrocyte protoporphyrin (0.78 +/- 0.60 vs. 1.46 +/- 0.94 mol/L, P = 0.006) and serum ferritin (median 5 vs. 0 g/L, P = 0.001). No significant differences between groups were observed in serum iron concentration, total iron-binding capacity or transferrin saturation. Iron deficiency anemia (hemoglobin </= 110 g/L) was observed in 36.4% (8 of 22) of infants in the group fed food cooked in iron pots and in 73.9% (17 of 23) of the infants fed food cooked in aluminum pots (P = 0.03). These results indicate that the iron added to food cooked in iron pots is bioavailable. However, this increased iron availability was insufficient to satisfy the high iron requirements of this group of preterm infants.