The study included 10 VLBW infants who received EPO plus iron (2–4 mg/kg/d) as part of their normal clinical care. The study protocol was approved by the Memorial University Ethics Committee and informed written permission was obtained from parents/guardians. The baseline period was that time before the administration of EPO and iron isotopes. EPO treatment began when infants were tolerating 75% of feeds (non-iron fortified formula or human milk) by the enteral route. On two separate occasions, before commencement of EPO treatment, and again after 2 weeks of EPO treatment stable non-radioactive isotopes of iron (57Fe, 58Fe) were administered enterally. At baseline and again at 2 and 4 weeks of treatment, a blood sample was collected by venipuncture and a 12-hour urine sample was collected. We thereby tested the two hypotheses that firstly, early supplemental iron intakes (2–4 mg/kg/d) result in oxidant stress, and that secondly, incorporation of ingested iron is affected by type of feed (human milk or formula) that the infant receives.
Subjects and feedings
Eligibility required birth weight < 1500 g, < 34 weeks gestation receiving oral or gastric feeds (75%), stable respiratory status as defined by FIO2 ≤ 30%, and a mean airway pressure of 8 cm of water or less, if on assisted ventilation. Infants were not eligible if they received transfusions within the previous week or during the 4 week study period, had any major congenital defect, liver disease, necrotizing enterocolitis, grade III or IV intracranial hemorrhage or culture-proven sepsis.
All infants received parenteral nutrition with no added iron before study entry. During the study, 3 infants received formula (Similac Special Care, Ross Products Division, Abbott Laboratories, Columbus, OH) and the remaining 7 infants received their own mother's milk fortified with Enfamil Human Milk Fortifier (Mead Johnson Nutritionals, Evansville, IN). No iron or vitamin supplementation were given on the day of isotope administration, otherwise management of nutrition was under the direction of the medical team.
Beginning at study entry, r-HuEPO (R.W. Johnson Pharmaceutical Research Institute, Raritan, NJ) was administered S.C. in a dose of 200 U/kg 3 times a week for six weeks. Iron was given concomitantly in graded doses of 2 mg/kg/d wks 1 and 2 and 4 mg/kg/d, wks 3 and 4.
Administration of 57Fe and 58Fe test doses
At baseline (
57Fe dosing) and again on d 14 (
58Fe dosing) each infant received one dose with a regular feed (as described previously). A 24 hour equilibration time was used and is sufficient to allow isotope to mix with native iron in human milk [
10]. The average daily volume of milk at baseline was 148 ± 20 mL/kg and at week 2, 158 ± 18 mL/kg (x ± SD).
Enriched
57Fe (95 weight %
57Fe, Atomergix Inc. Farmingdale, NY), and
58Fe (93.05 weight %
58Fe, Cambridge Isotope Labs, Andover, MA) were converted to ferrous chloride as described previously [
13]. Enriched oxides were dissolved in a small amount of Aqua Regia (HCl: HN0
3, 3:1), diluted, pH adjusted to 5 with NaOH, brought up to a volume of 20 mL with saline and filtered through a 0.2 um filter and stored in glass vials until use. Before use, each vial was checked for purity from pyrogens with a test kit (Limulus Ambeocyte Lysate Test Kit: Whittaker Bioproducts, Walkersville, MD).
To prepare the enriched formulas, an amount of solution which provided approximately 1800 ug 57Fe/kg or 450 ug 58Fe/kg was added directly to the milk using a pre-weighed syringe. 57Fe was equilibrated with vitamin C overnight in milk alone. 58Fe was mixed in milk with vitamin C and a solution of natural abundance ferrous sulfate (Fer-In-Sol, Mead Johnson, Inc., Evansville, IN). The final iron content of the first feed with 57Fe was 0.2 mg/kg. The final iron content of the latter feed was set to 4 mg/kg made up from 58Fe, endogenous iron in the formula as labeled, (human milk was considered to have negligible iron) with the majority of iron from ferrous sulfate. This final dosage matched the usual treatment dose (4 mg/kg/d) provided with EPO at this time period. Milk administered by bottle was flushed with 5 mL of a 10% glucose solution.
Laboratory methods
Hemoglobin (Hgb) concentration was determined in whole blood by Coulter Counter (Coulter Electronics, Inc., Hialeah, FL). Whole blood was centrifuged and plasma removed. Red blood cells (RBCs) were washed 3 times in isotonic saline and both plasma and RBCs frozen at -80°C until analysis.
For isotopic analyses, iron was obtained from RBCs according to Fomon et al. [
14]. Two week study periods were used as 2 weeks is the time required for enrichment of RBCs (14). For each of 10 infants, 3 separate samples were analyzed for a total of 20 estimations of iron absorption, plus 10 baseline measurements. Briefly, after microwave digestion with concentrated nitric acid (HNO
3), in closed vessels, iron was selectively extracted with xylene that contained thenoyltrifluoroacetone and processed as reported previously [
15]. Samples were analysed for
58Fe/
54Fe and
57Fe/
54Fe ratios by inductively coupled plasma mass spectrometry (ICP-MS, Plasma Quad3, Thermo Elemental, Franklin, MA US).
The quantity of administered
57Fe and
58Fe label incorporated into erythrocytes 14 days after each dose was calculated as described previously [
14];
Total circulating iron was estimated as; Fe
circ = BV * Hgb * 3.47. BV is blood volume, assumed to be 0.085 L/kg [
16], Hgb is hemoglobin concentration in g/L, and 3.47 is the mg/g concentration of iron in Hgb. The following assumption was made; of the isotropic iron that is absorbed, by 14 d, 80% is incorporated into erythrocytes [
14].
All glass and plastic beakers, flasks and pipettes were acid washed and only distilled deionised water was used throughout as previously reported [
17]. Iron concentrations were determined by atomic absorption spectrometry (17). NIST standard reference material spectrophotometric grade iron was run as a reference standard and to determine a normalization factor applied to each sample. To measure oxidant stress in plasma, malondialdehyde (MDA) was measured by HPLC [
18]. Total radical antioxidant protection (TRAP) as a measure of the ability to resist oxidative stress was measured using an oxygen electrode (YSI Model 5331) according to Wayner et al. [
19]. In erythrocytes, superoxide dismutase (SOD) and catalase (CAT) were determined spectrophotometrically [
20,
21]. F
2-Iso-prostanes as a measure of oxidant stress in urine were measured using a commercially available ELISA kits (Oxford Biomedical Research Inc., Oxford, MI). Information on oxygen exposure, medications, transfusions daily nutrient intakes and other data was recorded from nursing records.
Differences in descriptive data for study subjects were assessed by Student's t tests (SPSS- version 10, SPSS Inc., Chicago, IL). Pearson Correlation Coefficients were determined between iron absorption and study weight, birthweight, gender and gestational age. Differences in iron incorporation by type of feed and oxidative stress were measured by both parametric and nonparametric tests with consistent results.
Sample size was determined from measurements taken from previous studies in similar infants (13, 17) based on iron incorporation. We expected absorption in the F group to be 10 % with an SD of 5 %. In the HM fed group we expected absorption of 30% with an SD of 5 based on our earlier work (13, 17). Thus with an expected difference of 20 %, a power of 0.8 and an alpha of 0.05 we would need 5 subjects in each group. Significance for all assays was assigned to p < 0.05 (two-tailed).