Abstract
Bioavailability of iron (Fe) from a low-Fe infant formula was determined by erythrocyte incorporation of 58Fe 14 d after administration in ten healthy, non-Fe-deficient infants. Two feeding protocols were compared, with each infant acting as his/her own control. At 140 and 154 d of age, infants were fed 1000 g of 58Fe-labelled formula (1.44 mg total Fe/1000 g) as six feeds over 24 h (Protocol A) or as two feeds/day on three consecutive days (Protocol B). A water solution with 57Fe and ascorbic acid was given separately as a reference dose in both study protocols. Erythrocyte incorporation of 58Fe and 57Fe was determined by thermal ionisation mass spectrometry. Geometric mean 58Fe incorporation was 7.6% (range 3.3-13.5%) with Protocol A as compared to 10.6% (range 6.7-18.6%) with Protocol B (P = 0.05); paired t test. Inter-individual variability of 58Fe was not reduced by correcting for the incorporation of 57Fe from the reference dose, or by correcting for plasma ferritin concentration. Fractional erythrocyte incorporation of Fe from low-Fe infant formula was in the same range as our earlier published data on erythrocyte incorporation of Fe from human milk extrinsically labelled with 58Fe (Davidsson et al. 1994a). The methodological evaluations included in this study clearly indicate the importance of using standardised study protocols when evaluating Fe bioavailability in infants. Corrections of erythrocyte incorporation data based on plasma ferritin or erythrocyte incorporation of Fe from a reference dose were not found to be useful.
