PEDIATRICS Vol. 104 No. 2 August 1999, pp. 210-215

Comparison Between Two and Five Doses a Week of Recombinant Human Erythropoietin for Anemia of Prematurity: A Randomized Trial

Mark S. Brown, MD^* and Julian F. Keith III, MD

From the ^* Department of Pediatrics, Presbyterian/St Luke's Medical Center, Denver, Colorado; and the  Department of Pediatrics, Naval Medical Center, San Diego, California.

	ABSTRACT

Top Abstract Methods Results Discussion Conclusion References

Objective.  To compare the erythropoietic response between two and five times a week dosages of recombinant human erythropoietin (r-EPO) using the same weekly dose, 500 U/kg, in very low birth weight (VLBW) infants.

Study Design.  Eighty VLBW infants were stratified into two gestational age groups and randomized to receive 500 U/kg of r-EPO either two or five times a week; 72 infants completed at least 4 weeks of study. The primary outcome variable was absolute reticulocyte counts at 4 weeks. Secondary outcome variables were hematocrits, transfusions, iatrogenic blood losses, infections, and serum ferritins. Multiple regression analysis was used to evaluate the secondary outcome variables.

Results.  By 4 weeks, absolute reticulocyte counts were higher in the infants given r-EPO five times a week [mean (SEM)]: 173 000/mm³ (15 000) vs 220 000/mm³ (18 000), two versus five doses per week, respectively. Hematocrits, 34.9% (0.9) vs 34.1% (0.8), and transfusions per infant, 2.06 (0.4) vs 2.11 (0.4), were not different between the groups. Additionally, 79% of the variance in the amount of blood transfused was accounted for by iatrogenic blood loss, the latter primarily associated with number of days ventilated. Episodes of sepsis and necrotizing enterocolitis were significantly associated with decreased absolute reticulocyte counts and increased transfusions.

Conclusions.  More frequent dosing of the same weekly amount of r-EPO produced a significant and sustained increase in stimulated erythropoiesis in VLBW infants. The importance of this finding on reducing transfusions was not able to be demonstrated because this study was not intended to differentiate transfusions. In this population of infants and at the dose level of r-EPO, iatrogenic blood loss contributed more to transfusions than a lower level of erythropoiesis, the former primarily associated with mechanical ventilation. Based on this and other studies, when VLBW infants are at risk for greater phlebotomy losses, it may be justifiable to use more vigorous r-EPO treatment, and when at lower risk to use less frequent dosing to enhance cost-effectiveness.  Key words:  anemia, prematurity, transfusions, erythropoietin.

Recombinant human erythropoietin (r-EPO) has been effective in reducing transfusions in very low birth weight (VLBW) infants,^1-8 although the most suitable dosage schedule remains unclear because of differences in weekly dose and frequency used in these trials.⁹ Weekly doses have ranged from 150 to 1400 U/kg a week and frequency from two to seven times a week. There is a dose-dependent response based on weekly dose,^3,10 but it is uncertain what additional effects different dose frequencies have on the erythropoietic response.

Maintenance doses of r-EPO in children and adults are usually effective when given once or twice a week, often at higher doses to improve cost-benefit.^11-16 However, there are enough differences between VLBW infants and older children and adults that extrapolation of dosing to VLBW infants is difficult. Unlike children and adults, blood volume in VLBW infants is increasing rapidly with growth during the period of the anemia of prematurity, placing a higher demand on erythropoiesis to maintain a stable hematocrit. Further, the red blood cell life span in premature infants is shorter than in children and adults, and iatrogenic blood loss during intensive care represents a greater proportionate loss. These factors, along with differences in pharmacokinetics of r-EPO between adults and VLBW infants,^17,18 create enough uncertainty about the lowest effective frequency of r-EPO dose that further study in VLBW infants seemed warranted.

The null hypothesis was that the same weekly dose of r-EPO, begun by the end of the first week of life and divided into two equal doses, would be as effective in stimulating erythropoiesis as when divided into five equal doses in infants 30 weeks' of gestational age. The primary outcome was absolute reticulocyte counts after 4 weeks of treatment. Secondary outcome variables included hematocrits, transfusions, iatrogenic blood losses, infections, and serum ferritins.

	METHODS

Top Abstract Methods Results Discussion Conclusion References

Study Participants

The study was conducted in the neonatal intensive care unit of Presbyterian/St Luke's Medical Center, Denver, CO, between January 15, 1993 and April 18, 1994. The study was approved by the institutional review board and parental consent was obtained before enrollment.

A total enrollment of 76 patients was projected based on: 1) absolute reticulocyte counts from treated infants in the North American multicenter r-EPO trial;⁷ 2) an  level of 0.05; 3) a  level of 0.20; and 4) the ability to detect at least a 20% difference in absolute reticulocyte counts at 4 weeks--a difference chosen empirically. Inclusion criteria were: 30 weeks' gestation by dates or early ultrasound and confirmed by examination, 7 days of age, absence of life-threatening major congenital abnormalities, absence of significant hemolytic disease, platelet count >75 000 and absence of significant coagulopathy, and anticipated hospital stay 4 weeks. Study participants were dropped from the study if the parents withdrew informed consent or if there were contraindications to subcutaneous injection.

A total of 134 infants were eligible and screened, 80 were entered. Of those infants not entered, early transfer was anticipated for 24, parents declined participation for 20, and 10 infants died in the delivery room or during the first week of life. After enrollment, but before completing 4 weeks of treatment, 6 infants died and 2 infants were transferred. For those infants who died before completing the 4 weeks of treatment, the next eligible infant was assigned to the same dosage schedule to maintain equal allocation and power. Seventy-two infants completed at least 4 weeks of treatment, 36 on each dosage schedule. All but 1 of the 31 infants that were 27 weeks of age completed 8 weeks of study. This latter group, as a strata of the original study population, was evaluated at 8 weeks to assess the persistence of the r-EPO dose frequency effect.

Study Design

Infants were stratified for gestational age 27 weeks and 28 to 30 weeks, and then randomized to two different dosage schedules of r-EPO, two times a week and five times a week, each totaling 500 U/kg a week. The reason for stratification was to allow all infants to be available at 4 weeks and the younger ones at 8 weeks. r-EPO was administered for 4 weeks and then continued to discharge or 8 weeks, whichever occurred sooner. The primary endpoint was absolute reticulocyte counts at 4 weeks in all infants. The effect of dose frequency on several secondary outcome variables was also evaluated. At 4 weeks, these variables were hematocrit, transfusion, iatrogenic blood loss, infection, serum ferritin, and, at 8 weeks, were absolute reticulocyte count, hematocrit, iatrogenic blood loss, and infection. Feedings were begun depending on the infant's clinical condition. When infants were receiving at least 60 kcal/kg/d of feedings, oral iron supplements were started at an initial dose of 3 mg/kg/d and increased to 6 to 10 mg/kg/d for the remainder of the study. Additional vitamin supplements included vitamin E (15 units/d), folate (50 µg/d), and a preparation of vitamins A, C, and D.

Erythropoietin Administration

Commercially available r-EPO (Procrit) was provided by Ortho Biotech, Raritan, NJ. Doses were prepared by the on-site pharmacist and administered subcutaneously in the lateral thigh with a 30-gauge needle; volume of the dose was maintained between 0.3 and 0.6 mL.

Laboratory Monitoring

Blood was obtained by peripheral venipuncture for complete blood count, differential, and manual reticulocyte count at the beginning of the study, weekly, and at the end of the study. Iron studies included weekly serum ferritin, serum iron, and transferrin levels. The additional blood samples totaled 1.9 mL a week. Tests were performed on automated instruments in the clinical laboratory, and results were available to those caring for the infants. To provide a comparable estimate of red blood cell production, absolute reticulocyte count/mm³ was calculated using the red blood cell count/mm³ and proportion of reticulocytes from manual counts. Manual reticulocyte counts were done using an optical device that reduces variability but has been shown to produce slightly lower counts than traditional manual counts.¹⁹ Ferritin was measured immunologically on the Ciba-Corning ACS:180 analyzer (Ciba-Corning Diagnostics Corp, Medfield, MA). Serum iron was measured by Kodak Ektachem (Johnson & Johnson Kodak Clinical Diagnostics, Rochester, NY). Transferrin was measured as turbidity using transferrin antibody (Beckman, Galway, Ireland).

Transfusions

Transfusions were given as 15 mL/kg throughout a 4-hour period. Transfusions and iatrogenic blood losses were totaled by week of occurrence. Transfusion criteria were those used in the North American multicenter trial.⁷ In brief, these included: 1) for hematocrit 30% if: a) undergoing surgery; or b) heart rate >180 beats/minute for 24 hours, or increased respiratory rate >80 beats/minute for 24 hours, or <10 g weight gain/d for 4 days while receiving 100 kcal/kg/d; or c) respiratory disease supported with 1/4 liter oxygen requirement or 35% oxygen, or mechanical ventilation, or continuous positive airway pressure <6 cm H₂O; or d) significant apnea and bradycardia in 24 hours while receiving therapeutic doses of methylxanthines; or 2) for hematocrit <35% if respiratory disease was supported with >1/4 liter per minute or >35% oxygen or mechanical ventilation with a mean airway pressure 6 cm of H₂O.

Statistical Analysis

Comparison of single-measure, continuous data between infants was made by a two-sided t test for independent samples if variances were equal; otherwise means were compared by Welch analysis of variance for unequal variances. For nonparametric data, comparisons were made with Wilcoxon rank sums. Unless otherwise specified, parametric data are given as mean ± SEM. The Fisher's exact test or the ² test was used when comparisons of nominal data were made. Univariate correlations were analyzed with correlation matrices. Variables that had a correlation coefficient >0.20 and treatment group were entered into multivariate analysis. Stepwise multiple regression was used to assess the overall effect of univariate correlations, singularly or as interactions, on secondary outcome variables. Absolute reticulocyte counts from infants with infections were compared using analysis of variance with Tukey correction for multiple comparisons. All analyses were conducted using the SAS and JMP software system (SAS Institute Inc, Cary, NC).

	RESULTS

Top Abstract Methods Results Discussion Conclusion References

Study Participants

The infants in the two study groups were comparable at entry, (Table 1) except for prestudy hematocrit and incidence of early sepsis, which were greater in the two dose a week group. The effect this may have had on subsequent measures of illness was not sustained and by 4 and 8 weeks the groups were comparable in phlebotomy losses, transfusion, days ventilated, incidence of sepsis, and days of antibiotics.

Absolute Reticulocyte Counts

Reticulocyte counts were slightly lower than those from the North American multicenter trial on the same five times a week dose. This may have been because of the change in technique of manual reticulocyte counts for this study.¹⁹ Because this was internally consistent, it did not effect the results or conclusions when comparing the responses between the two dosage schedules. After 2 weeks of r-EPO administration, absolute reticulocyte counts had increased to a plateau; those infants who received r-EPO five times a week had significantly higher absolute reticulocyte counts (Fig 1A). This difference was present at 4 weeks for all infants 30 weeks' gestation, and at 8 weeks for infants 27 weeks' gestation (Table 2). By multivariate analysis at 4 weeks, 26% of the variance in absolute reticulocyte counts was associated with amount of transfused blood, 6% with the treatment group, and 5% with the interaction between transfused blood volume and iatrogenic blood losses, overall R² = 0.37. Transfusions and increased iatrogenic blood loss are typical of less mature infants and both had a stronger negative association with absolute reticulocyte counts than either gestational age or days ventilated. Infants 27 weeks' gestation had significantly lower absolute reticulocyte counts by 4 weeks than more mature infants by analysis of variance, P = .02. By 8 weeks in the 27 weeks' gestation infants, the only correlation with absolute reticulocyte counts was treatment group, R² = 0.24. 

View larger version (14K): [in this window] [in a new window]   Fig. 1.   A, mean (±SEM) weekly absolute reticulocyte counts for the two dosage schedules of r-EPO at two doses of 250 U/kg each per week () and five doses of 100 U/kg each per week (). Up to 4 weeks, all infants 30 weeks are included, represented by the solid lines (), and from 4 to 8 weeks only infants 27 weeks' gestation are included, represented by the dashed lines (). At 4 and 8 weeks, absolute reticulocyte counts were significantly different between the treatment groups (P = .046 and 0.03, respectively). B, mean (±SEM) weekly hematocrits for the two dosage schedules of r-EPO at two doses of 250 U/kg each per week () and five doses of 100 U/kg each per week (). Up to 4 weeks, all infants 30 weeks are included, represented by the solid lines (), and from 4 to 8 weeks only infants 27 weeks' gestation are included, represented by the dashed lines (). At 4 and 8 weeks, hematocrits were not significantly different between the treatment groups.

Hematocrit

At 4 weeks in all infants and at 8 weeks in the 27 weeks' gestation infants, there was not a significant difference in hematocrits between the two treatment groups (Fig 1B). If the small difference at 8 weeks, 32.9% vs 34.%1, was a real difference it would have taken a sample size of 55 infants per group to demonstrate a significant difference ( = 0.05,  = 0.20).

Transfusions

At 4 and 8 weeks the amount of transfusions did not differ between treatment groups (Table 2). By multivariate analysis at 4 weeks, 79% of the variance in transfusions was associated with amount of iatrogenic blood loss, 3% with absolute reticulocyte count, and 1% each with gestational age and days of antibiotics, overall R² = 0.84. By multivariate analysis at 8 weeks, 83% of the variance in transfusions was associated with iatrogenic blood loss, and 3% with absolute reticulocyte count, overall R² = 0.86.

Iatrogenic Blood Loss

At 4 and 8 weeks, the volume of iatrogenic blood loss per infant did not differ between treatment groups (Table 2). At 4 weeks, the maximum amount of iatrogenic blood loss that was sustained without receiving a transfusion was 30 mL. In multivariate analysis, 65% of the variance in iatrogenic blood loss was associated with days ventilated, 3% with days of antibiotics, and 1% with gestational age, overall R² = 0.69. Transfusion was not entered as an independent variable into the multiple regression because of the problem of collinearity. At 8 weeks, the maximum amount of iatrogenic blood loss that was sustained without receiving a transfusion was 31 mL. By multivariate analysis, 56% of the variance in iatrogenic blood loss was associated with days ventilated and 9% with days of antibiotics, overall R² = 0.66.

Infections

Serious infections beyond the first week, namely sepsis and necrotizing enterocolitis, were associated with both a decrease in the absolute reticulocyte count and an increase in transfusions, despite r-EPO treatment. During r-EPO treatment, there were 17 episodes of gastrointestinal infection, 5 associated with systemic sepsis, and 27 episodes of sepsis, most associated with central line infections. Sepsis was defined as a positive blood culture that was clinically significant and treated with antibiotics for at least 5 days. The incidence of sepsis was not significantly different between the treatment groups. The incidence of both sepsis and gastrointestinal infection was consistent with our unit experience from previous years in this birth weight category without r-EPO treatment. Gastrointestinal complications included: suspected necrotizing enterocolitis,²⁰ n = 8; definite necrotizing enterocolitis, n = 5; surgical necrotizing enterocolitis, n = 3; and obstruction, n = 1. Mean absolute reticulocyte counts for the week of infection were: necrotizing enterocolitis, 112 000/mm³; necrotizing enterocolitis and sepsis, 66 000/mm³; and sepsis, 140 000/mm³; P < .05.

Iron

Iron sufficiency was estimated by serum ferritin, based on previous work.^21,22 At 2 weeks, all ferritins were >50 ng/mL (µg/L), and by 3 weeks those infants whose ferritin level was <50 ng/mL (µg/L) had been started on oral iron. At 4 weeks, 8 infants had a ferritin level <50 ng/mL (µg/L), 7 were in the five dose a week group and all were receiving at least 6 mg/kg/d of oral iron. At 4 and 8 weeks, ferritin levels did not differ between treatment groups. There was no correlation between absolute reticulocyte counts and either serum ferritin, serum iron, transferrin, or transferrin saturation. Because transfusions increase ferritin, for analysis of ferritin it was useful to separate infants based on whether they had received a transfusion. By 4 weeks in untransfused infants, n = 24, the median serum ferritin was 77.5 ng/mL (µg/L) and range was 23 to 249 ng/mL (µg/L). In multivariate analysis, 41% of the variance in ferritin was associated with initial ferritin, 19% with days of antibiotics, and 8% with total amount of oral iron administered, overall R² = 0.68. In transfused infants, n = 48, the median serum ferritin was 180 ng/mL (µg/L) and range was 33 to 639 ng/mL (µg/L). In transfused infants at 4 weeks by multivariate analysis, 55% of the variance in ferritin was associated with number of transfusions, 6% with initial ferritin, and 4% with the interaction between transfusions and days ventilated, overall R² = 0.65.

	DISCUSSION

Top Abstract Methods Results Discussion Conclusion References

More frequent administration of the same weekly r-EPO dose, five times a week, in VLBW infants resulted in a sustained increase in erythropoietic response compared with two times a week. Neither hematocrit nor transfusions differed between treatment groups. Transfusions were primarily associated with iatrogenic blood losses, with only a small contribution by differences in absolute reticulocyte counts. Although there was a difference in erythropoietic response between treatment groups, we were unable to demonstrate a decrease in transfusions because of the strong relationship between transfusions and iatrogenic blood loss. Iatrogenic blood loss primarily depended on days of mechanical ventilation and days of antibiotics, the latter a marker for concerns about or occurrence of sepsis. Sepsis, whether or not it included necrotizing enterocolitis, was associated with decreased absolute reticulocyte counts and increased transfusions.

Ferritin was used to evaluate iron sufficiency. From previous work we believed ferritin was more reliable than serum iron or transferrin saturation,^21,22 although ferritin levels need to be interpreted with caution during infection because, as acute phase reactants, they may be elevated. By 4 weeks, only 8 infants had ferritin levels <50 ng/mL (µg/L), all were receiving at least 6 mg/kg per day of oral iron. This is reassuring that iron deficiency was not likely to have been present. Of importance was that 7 of these infants were from the five dose a week group, confirming that as the rate of erythropoiesis increases in VLBW infants, closer attention needs to be paid to iron sufficiency.

The most likely explanation for the difference in dose frequency response is the longer amount of time that serum erythropoietin levels were elevated above than baseline in the five dose a week group. The peak serum erythropoietin concentration would have been higher in the two dose a week group and the area under the curve for serum erythropoietin would have been the same for both groups, making these unlikely explanations for the higher erythropoietic response in the five dose a week group. If the twice a week dose were doubled to attempt to compensate for the shorter time above baseline, it would result in only a 5% increase in the time above baseline; this is based on calculations from published pharmacokinetic data in long-term treated infants.¹⁸

The treatment assignment was blinded to those doing the absolute reticulocyte counts and these counts provided an unbiased estimate of the erythropoietic response to r-EPO dose frequency. On the other hand, the dose schedule was known to those caring for the study infants. This creates a potential for bias in a possible difference in transfusion practices that might have altered the true hematocrit or transfusion experience between the groups. Had this occurred, it would have produced a higher hematocrit in one group with either higher or equivalent transfusion amounts, or the same hematocrit in one group with higher transfusion amounts. Instead, at 4 weeks both hematocrit and transfusion amounts were the same between groups. This is reassuring that the lack of difference in hematocrit was not because of over-transfusion in one or the other of the treatment groups.

r-EPO has not been as effective in reducing transfusions during the time when infants are ventilated. In the current study, regardless of treatment group, ventilated infants had greater iatrogenic blood losses, and hematocrits were maintained by transfusions; this accounts for the association between transfusions and iatrogenic blood loss. These findings are consistent with those from the North American multicenter trial^7,23 and confirms the report by Soubasi et al.² To decrease iatrogenic blood losses during mechanical ventilation in VLBW infants requires prudent use of blood gases and laboratory tests, more reliance on transcutaneous or in-line monitoring in place of blood gases, and application of point-of-care techniques for laboratory tests that use minimal blood samples.

It is more problematic to reduce iatrogenic blood losses and transfusions during sepsis evaluations and treatment. Infections are known to depress erythropoiesis in adults.²⁴ That there was a decrease in erythropoiesis during sepsis and necrotizing enterocolitis during treatment with r-EPO in the current study suggests that there is a relative refractoriness to erythropoietin during infection in newborns. During sepsis and necrotizing enterocolitis in newborns, there is an increase in certain cytokines, namely interleukin-6 and tumor necrosis factor,²⁵ that can depress erythropoiesis in vitro.²⁶ During such an infection in VLBW infants, it is unclear whether to increase, stop, or continue r-EPO, although there is evidence to suggest that in similar situations erythropoiesis may respond to an increase in r-EPO.^26,27

	CONCLUSION

Top Abstract Methods Results Discussion Conclusion References

In summary, not only is the erythropoietic response to r-EPO in VLBW infants dose-dependent, but also dose frequency-dependent. In the current study, more frequent dosing of the same weekly amount of r-EPO produced a significant and sustained increase in stimulated erythropoiesis in VLBW infants, but the importance of this finding on reducing transfusions was not able to be demonstrated because this study was not intended to differentiate transfusions. In this population of VLBW infants treated by the end of the first week of life and at the dose level of r-EPO, iatrogenic blood loss contributed more to transfusions than a lower level of erythropoiesis, the former primarily associated with mechanical ventilation. Based on the current study and those using r-EPO during the first week of life,^6,8 when VLBW infants are at risk for greater phlebotomy losses, it may be reasonable to use more vigorous r-EPO treatment, and when at lower risk for such losses to use less frequent dosing to enhance cost-effectiveness. To what extent this strategy decreases donor exposure and compliments a reduction in iatrogenic blood loss requires further investigation. Iron supplementation is necessary by 2 to 3 weeks depending on the level of erythropoiesis, particularly in those infants who have not received a transfusion. During serious infections, especially sepsis and necrotizing enterocolitis, the erythropoietic response to r-EPO seems blunted, and whether to adjust the r-EPO dose needs further evaluation.

	ACKNOWLEDGMENTS

This study was supported by Ortho Biotech, Raritan, NJ, and P/SL Community Foundation Grant 9014.

Kim Dohren provided expertise with graphics and Eileen Delaney, RN, Steve Aschfort, RhD, and Erin Walsh, RhD, assisted with data collection and drug handling. Michael A. Jones, PharmD, analyzed the pharmacokinetic curves for the two different doses. The support in conducting this study and caring for these infants by Barbara Quissell, MD, Peter Honeyfield, MD, Delphine Eichorst, MD, Bruce Reddix, MD, Jeffrey B. Hanson, MD, and the NICU nurses is deeply appreciated.

	FOOTNOTES

Received for publication Aug 31, 1998; accepted Jan 8, 1999.

Reprint requests to (M.S.B.) 1601 East 19th Ave, Suite 5300, Denver, CO 80218. E-mail: marksbrown{at}compuserve.com

	ABBREVIATIONS

r-EPO, recombinant human erythropoietin; VLBW, very low birth weight.

	REFERENCES

Top Abstract Methods Results Discussion Conclusion References

1. Carnielli V, Montini G, Da Riol R, Dall Amico R, Cantarutti F Effect of high doses of human recombinant erythropoietin on the need for blood transfusions in preterm infants. J Pediatr. 1992; 121:98-102 [CrossRef][Medline]
2. Soubasi V, Kremenopoulos G, Diamandi E, Tsantali C, Tsakiris D In which neonates does early recombinant human erythropoietin treatment prevent anemia of prematurity? Results of a randomized, controlled study. Pediatr Res. 1993; 34:675-679 [Medline]
3. Messer J, Haddad J, Donato L, Astruc D, Matis J Early treatment of premature infants with recombinant human erythropoietin. Pediatrics. 1993; 92:519-523 [Abstract/Free Full Text]
4. Meyer MP, Meyer JH, Commerford A, Recombinant human erythropoietin in the treatment of the anemia of prematurity: results of a double-blind, placebo-controlled study. Pediatrics. 1994; 93:918-923 [Abstract/Free Full Text]
5. Maier RF, Obladen M, Scigalla P, The effect of epoetin beta (recombinant human erythropoietin) on the need for transfusion in very-low-birth-weight infants. N Engl J Med. 1994; 330:1173-1178 [Abstract/Free Full Text]
6. Ohls RK, Osborne KA, Christensen RD Efficacy and cost analysis of treating very low birth weight infants with erythropoietin during their first two weeks of life: a randomized, placebo-controlled trial. J Pediatr. 1995; 126:421-426 [CrossRef][Medline]
7. Shannon KM, Keith JF III, Mentzer WC, Recombinant human erythropoietin stimulates erythropoiesis and reduces erythrocyte transfusions in VLBW preterm infants. Pediatrics. 1995; 95:1-8 [Abstract/Free Full Text]
8. Ohls RK, Harcum J, Schibler KR, Christensen RD The effect of erythropoietin on the transfusion requirements of preterm infants weighing 750 grams or less: a randomized, double-blind, placebo-controlled study. J Pediatr. 1997; 131:661-665 [CrossRef][Medline]
9. Strauss RG Recombinant erythropoietin for the anemia of prematurity: still a promise, not a panacea. J Pediatr. 1997; 131:653-655 [Medline]
10. Halperin DS, Felix M, Wacker P, Lacourt G, Babel J-F, Wyss M Recombinant human erythropoietin in the treatment of infants with anaemia of prematurity. Eur J Pediatr. 1992; 151:661-667 [CrossRef][Medline]
11. Abraham PA, St Peter WL, Redic-Kill KA, Halstenson CE Controversies in determination of epoetin (recombinant human erythropoietin) dosages. Clin Pharmacokinet. 1992; 22:409-415 [Medline]
12. Zachée P Controversies in selection of epoetin dosages. Drugs. 1995; 49:536-547 [Medline]
13. Ongkingco JRC, Ruley EJ, Turner ME, Fragale MR Efficacy of once- versus thrice-weekly subcutaneous recombinant human erythropoietin in children receiving continuous cycling peritoneal dialysis. Am J Nephrol. 1994; 14:14-18 [Medline]
14. Nomoto Y, Kawaguchi Y, Kubota M, A multi-center study with once a week or once every two weeks high-dose subcutaneous administration of recombinant human erythropoietin in continuous ambulatory peritoneal dialysis. Perit Dial Int. 1994; 14:56-60
15. Cheng IKP, Chan CY, Yu L, Fang GX, Wei D Correction of anemia in patients on continuous ambulatory peritoneal dialysis with subcutaneous recombinant erythropoietin twice a week: a long-term study. Clin Nephrol. 1991; 35:207-212 [Medline]
16. Lui SF, Wong KC, Li PKT, Lai KN Once weekly versus twice weekly subcutaneous administration of recombinant human erythropoietin in haemodialysis patients. Am J Nephrol. 1992; 12:55-60 [Medline]
17. Brown MS, Jones MA, Ohls RK, Christensen RD Single-dose pharmacokinetics of recombinant human erythropoietin in preterm infants after intravenous and subcutaneous administration. J Pediatr. 1993; 122:655-657 [Medline]
18. Widness JA, Veng-Pedersen P, Peters C, Pereira LM, Schmidt RL, Lowe LS Erythropoietin pharmacokinetics in premature infants: developmental, nonlinearity, and treatment effects. J Appl Physiol. 1996; 80:140-148 [Abstract/Free Full Text]
19. Nazi KM The Miller disk: an improvement in the performance of manual reticulocyte counts. Lab Med 1986; 17:742-744
20. Bell MJ, Ternberg JL, Feigin RD, Neonatal necrotizing enterocolitis: therapeutic decisions based on clinical staging. Ann Surg. 1978; 187:1-7 [Medline]
21. Brown MS Effect of transfusion and phlebotomy on serum ferritin levels in low birth weight infants. J Perinatol. 1996; 16:39-42 [Medline]
22. Brown MS, Shapiro H Effect of protein intake on erythropoiesis during erythropoietin treatment of anemia of prematurity. J Pediatr. 1996; 128:512-517 [CrossRef][Medline]
23. Shannon KM, Mentzer WC, Abels RI, Recombinant human erythropoietin in the anemia of prematurity: results of a placebo-controlled pilot study. J Pediatr. 1991; 118:949-955 [CrossRef][Medline]
24. Zucker S, Friedman S, Lysik RM Bone marrow erythropoiesis in the anemia of infection, inflammation, and malignancy. J Clin Invest. 1974; 53:1132-1138
25. Harris MC, Costarino AT, Sullivan JS, Cytokine elevations in critically ill infants with sepsis and necrotizing enterocolitis. J Pediatr. 1994; 124:105-111 [CrossRef][Medline]
26. Means RT, Krantz SB Inhibition of human erythroid colony-forming units by  interferon can be corrected by recombinant human erythropoietin. Blood. 1991; 78:2564-2567 [Abstract/Free Full Text]
27. Schreiber S, Howaldt S, Schnoor M, Recombinant erythropoietin for the treatment of anemia in inflammatory bowel disease. N Engl J Med. 1996; 334:619-623 [Abstract/Free Full Text]