Serum erythropoietin levels during infancy: Associations with erythropoiesis

Serum erythropoietin levels during infancy: Associations with erythropoiesis P a m e l a J. Kling, MD, R o b e r t L. S c h m i d t , BS, Robin A. Roberts, BS, BA, a n d J o h n A. Widness, MD From the Department of Pediatrics, The University of Iowa College of Medicine, Iowa City, Iowa

Objective: To determine plasma erythropoietin levels and their association with hemoglobin and reticulocyte counts in healthy term infants. Design: We compared plasma erythropoietin levels measured in serial blood samples obtained every 4 weeks during the first 6 months of life with one another and with levels in term fetuses and healthy adults. Correlation analysis was applied to examine for associations of erythropoietin with hemoglobin and with reticulocyte count. Results: Plasma erythropoietin levels were lowest in the first and highest in the second postnatal months, a pattern reciprocal to that observed for hemoglobin during the period of physiologic anemia. The erythropoietin level was negatively correlated with hemoglobin (p <0.0001) and positively correlated with reticulocytes (p <0.000 I). The slope of the inverse relationship of hemoglobin and plasma erythropoietin in infants was similar to those previously reported for anemic fetuses and premature infants, but much less steep than for anemic children and adults. Conclusion: This study is the first to report simultaneous patterns of change observed in plasma erythropoietin, hemoglobin, and reticulocytes during normal infancy. These patterns are consistent with postnatal perturbations in tissue oxygenation and suggest a major role for erythropoietin in the regulation of erythropoiesis during normal infancy, but at a lower hemoglobin concentration than for older children and adults with pathologic anemia. (J Pediatr 1996;128:791-6) There are few data on normal plasma erythropoietin levels during the physiologic anemia in healthy growing infants. During this period major developmentally associated changes that influence tissue oxygenation occur] These changes include increased partial pressure of oxygen in arterial blood as oxygen exchange switches from in utero dependency on the placenta to ex utero dependency on the lungs, a switch from fetal hemoglobin synthesis to adult hemoglobin synthesis, an increase in intraerythrocyte 2,3-diphosphoglycerSupported by National Institutes of Health grants No. 5T32 HD 07287 (PJK) and No. PO1 I-IL46925(JAW). Submitted for publication Aug. 24, 1995; accepted Feb. 6, 1996. Reprint requests: Pamela J. Kling, MD, University of Arizona Health Sciences Center, Department of Pediatrics, Steele Memorial Children's Research Center, 1501 N. Campbell Ave., Tucson, AZ 85724. Copyright © 1996 by Mosby-Year Book, Inc. 0022-3476/96/$5.00 + 0 9/20/72556

ate concentration, and a switch in the site of and mechanisms regulating erythropoietin production. 2-4 Although the regulation of erythropoiesis during development is complex and incompletely understood, the net culmination of the aforementioned processes is the normal postnatal decline in hemoglobin concentration--the early anemia of infancy. 3 Pathologic anemia in fetuses and premature infants is associated with relatively low plasma erythropoietin levels relative to comparably anemic adults. 5-7 Although this suggests that erythropoietin functions as a regulator of erythropoiesis in such individuals, a similar response has not been reported in normal infants during the early physiologic anemia. Our primary objectives in this study were to define the pattern of plasma erythropoietin levels in healthy term infants during the first 6 months of life, and to determine whether the postnatal changes in plasma erythropoietin levels were associated with concomitant changes in hemoglobin and reticulocytes. A secondary objective was to compare

791

792

Kling et al.

the anticipated inverse relationship of plasma erythropoietin with hemoglobin in normal term infants with data reported previously for fetuses, premature infants, children, and adults with anemias.

METHODS Subjects. After approval was obtained for this study from the institutional review board, written consent was obtained from one or both parents. Formula-fed term infants >~37 weeks of gestation were eligible for study. Offspring of mothers with significant medical problems were excluded, as were infants with anemia as indicated by a positive Coombs test result, perinatal infection, congenital malformations, or intrauterine growth retardation (birth weight <10th percentile). Infants with acute illness that interrupted normal enteral intake for more than 2 days were also excluded. Two comparison groups were studied: normal term fetuses (n = 41) delivered by cesarean section before onset of labor and healthy nonsmoking adults (n = 58) meeting our blood bank's criteria for whole blood donation. Study design. Beginning at 4 weeks Of age, blood sampling in the infant group was performed every 28 days (i.e., ' 'monthly") for the first 7 months of life. The first 2 months' samples were obtained within 2 days of the specified day, with samples thereafter obtained within 4 days of the specified day. All infants were fed commercially prepared ironfortified formula(label claim ---12 mg iron/L). The use of iron and vitamin supplements was discouraged. Foods other than formula were discouraged before 20 weeks. Thereafter solid foods were allowed at the parents' discretion. Laboratory determinations. Plasma erythropoietin levels were measured in triplicate with a double-antibody radioimmunoassay. 8 The lower sensitivity of the assay is 2 mU/ml. The intraassay and interassay coefficients of variation for three plasma pools spanning the useful range of the assay varied from 1.4% to 3.5%, and 13.1% to 16.7%, respectively. Because the biologic activity of erythropoietin is correlated with the logarithm of its concentration, the logarithmically transformed plasma erythropoietin concentrations were used in the analyses. Hemoglobin concentration and erythrocyte counts were determined with an automated counter (M430 Coulter Counter, Coulter Corp., Hialeah, Fla.). After staining with brilliant cresyl blue, reticulocytes were counted and results expressed as a percent. 9 Absolute reticulocyte count was the product of erythrocyte count and reticulocyte percentage. Data analysis. Differences between the seven monthly study periods in the infant group and between the two comparison groups were determined with one-way analysis of variance. This more conservative procedure was used in preference to repeated measures analysis of variance because of missing data points. If the F ratio from this comparison demonstrated a significant difference, post hoc testing with

The Journal of Pediatrics June 1996

the Bonferroni correction for multiple unpaired t testing (two-tailed) was applied to determine which monthly comparisons were significantly different. When infant data were compared with the fetal or adult group, the total number of comparisons was 7, compared with 21 needed for all monthly combinations in the infant group. Associations between study variables were examined with general estimating equations applicable to sequential data containing missing values, unequally spaced observations, and covariates that change with time by repeated measures using generalized estimating equations. 1° With this method, simple and multiple regression coefficients are expressed as standardized z scores. Results are expressed as the mean _+ SEM. An alpha level <0.05 was considered significant.

RESULTS The mean birth weight of the infant group was 3658 gm (range 2720 to 4505 gm); this was similar to the birth weight of normal term fetuses of 3328 gm (range 2530 to 4140 gin). Significant monthly changes occurred in plasma erythropoietin concentration during the first 6 months of life in normal infants; the lowest erythropoietin levels occurred at month 1 and the highest at month 2 (F6, 131 = 8.1, p <0.001) (Fig. 1, A). The infants' plasma erythropoietin values also differed significantly compared with both term fetuses and adults (FT, 171 = 13,0, p <0.001, and F7, 188 = 6.2, p <0.001, respectively). Infant postnatal monthly hemoglobin concentration significantly changed in a pattern consistent with the normal physiologic anemia (F6, 159 = 23,0, p <0.001) (Fig. 1, B). These hemoglobin concentrations differed from those of fetuses or adults (/77, 177= 28.0, p <0.001, and FT, 216 = 37.1, p <0.001, respectively). The pattern of change for hemoglobin was reciprocal to and negatively associated with plasma erythropoietin (z = -5.60, p <0.0001). Although correlations of available oxygen 11 and oxygen-releasing capacity 12 with plasma erythropoietin were also highly significant (z = -3.86 and z=-2.59, respectively), neither of these ancillary hemoglobin-derived indicators of tissue oxygenation improved on that of hemoglobin alone. The significant postnatal changes observed in infant reticulocyte counts followed a pattern similar to those for plasma erythropoietin (F6, 104= 23.1, p <0.0001) (Fig. 1, C). These changes demonstrated significant associations with plasma erythropoietin (z = 5.49, p <0.0001) and with hemoglobin (z = -3.37, p <0.0001). Multiple linear regression performed to determine whether one of these variables was more dominant in its association with reticulocyte counts identified erythropoietin as significant (z = 3.31, p <0.0001), whereas hemoglobin was not (z = -1.87, not significant). The relationship of paired plasma erythropoietin and hemoglobin data in the healthy term infant group was exam-

The Journal of Pediatrics Volume 128, Number 6

Kling et al.

a.

793

50-

g30-

0 ~'~20-

4t

•

,~T

~10-

,,ll

i'~

m

0

bl

Fetal 'Mclllhl 'Month2 Moth3 'Mo~4 'Monfll5'Month6 'Month7 I ~k~It~ [

180-

.l 150 ,.ID O O 120 ID

90 C.

Fetal Mc~thl 'Moth2 Moalh3 'MoNa4 Mc~h5 'Irish6 Month7 Adult

140~-'~ 120 ! 100802 ~9

-

602

-

o 40:

E

"4= 202 0

Fetal 'Momhl 'Moth2 'Month3 'Mc~tl4 'Ma~5 'Month6 'Mon~7 ' Adult

Fig. 1. A, Plasma eryttn-opoietin levels, B, hemoglobin concentrations, and C, reticulocyte counts for normal term infants during the first 7 months of life. A and B also contain data for the two comparison study groups (i.e., normal term fetuses delivered by cesearean section before labor and healthy, nonsmoking normal adults). Median values are indicated by the horizontal lines inside the rectangles. The upper and lower boundaries of the rectangles define the upper and lower quartiles, and the thick vertical lines indicate the 10th and 90th percentiles. The vertical numbers to file left of the quartile boxes represent the mean _+ SD, with number of samples in parenthesis. The values to the right of the the quartile boxes indicate statistical significance detected by post hoc testing atp <0.05 with the other study periods or with the comparision groups. As an example, in Month 1 for plasma erythropoietin, the vertical notation, "Fetal, 2, 3 & Adult," indicates significant differences in plasma erythropoietin concentrations for infants at 1 month of age compared with fetal umbilical cord plasma at birth, with infants at months 2 and 3 of age, and with normal adults. ined by simple linear regression and compared with similar data published for midfetal through adult life. A highly significant inverse relationship was observed for paired infant plasma erythropoietin and hemoglobin data pooled from all

7 months (r -- -0.51, p <0.000!, n = 138) (Fig. 2). When the slope of this relationship in normal infants was compared with those previously reported in the presence of pathologic anemia at other periods Of development, the slope of the in-

794

Kling et al.

The Journal of Pediatrics June 1996

A 17-38wkFetuses(Moyaet al. 1993) i00,000-

0 VLBWInfants(Brownet al. 1983)

X Normal TermInfants(presentstudy)

10,000 Q"

\

O Children(Bray et al. 1992)

1,000-

LI.I o

10010

[ ] Adults(Gareiaetal. 1982)

N A.......................... \ "

1 ....... 20 40

, .... . ...... 60 80 100

,...,..., 120 140

160

Hemoglobin (g/L) Fig. 2. Relationship of hemoglobin concentration and the log of plasma erythropoietin concentration at different developmental periods. The regression relationship for the normal term infants from this study is indicated by the thick dashed line. Shown for comparisonare the regression slopes of 17- to 38-week gestationalfetuses at cordocentesis,13very low birth weight infants (<1500 gm) before hospital discharge (2 to 17 weeks of age),6 children with iron deficiency (age 3.5 ± 3.7 years),14and adults with anemia from a variety of causes but without renal disease.15Note that the regression lines for the normal term infants (thick dashed line) and the 17- to 38-week fetuses (thin, finely dotted line) overlap. The horizontal dimensions of these regression lines indicate the range of values included in the specific studies shown.

fants' regression was similar to those of fetuses in the second haft of pregnancy13 and to very low birth weight premature infants.6 The slopes of these three regressions in developmentally immature individuals were markedly flatter that those observed in children14 or adults. 15 DISCUSSION

This study is the first to show that postnatal plasma erythropoietin levels in normal term infants undergo a predictable postnatal pattern of change inversely associated with both the decrease and increase in hemoglobin concentrations that occur during the normal physiologic anemia of infancy. These data also document for the first time that the magnitude of the decrease and increase in postnatal monthly plasma erythropoietin levels exceeds the boundaries of plasma erythropoietin levels found in normal adults and that an increase during infancy is equivalent to that observed among normal term fetuses. 16 Finally, a highly significant direct association of plasma erythropoietin with reticulocyte count in normal infancy has not been previously shown. Together these findings suggest an important physiologic role for erythropoietin during normal infancy. Two previous, smaller cross-sectional studies of plasma erythropoietin levels in term infants17' 18observed no change in plasma erythropoietin and no significant relationship of erythropoietin and hemoglobin. A more recent study by Yamashita et al. 19 of infants at less frequent, more sporadic intervals also reported a significant inverse associ-

ation of hemoglobin and erythropoietin. 19 These authors did not present reticulocyte data. Although Yamashita et al. reported that the mean of 7 plasma erythropoietin values from postnatal days 7 to 50 was less than that in a normal adult population and the mean of 24 values obtained between days 51 to 200 was similar to that in a normal adult reference population, no rise in erythropoietin levels with physiologic anemia was reported. 19 Our observation of significantly elevated plasma erythropoietin values during the nadir of hemoglobin at month 2 was probably the result of better defined and more frequent sampling in a greater number of infants. The oxygen-hemoglobin dissociation curve at birth in preterm and term infants is shifted to the left compared with the dissociation curve of normal children and adults. 12 This is the result of higher concentrations of fetal hemoglobin, lower levels of intraerythrocyte 2,3-diphosphoglycerate, and poorer responsiveness of fetal hemoglobin to 2,3-diphosphoglycerate relative to adult hemoglobin. It has been suggested that expressions of hemoglobin-oxygen delivery incorporating adjustments made for factors shifting the oxygen dissociation curve are more physiologically important than hemoglobin alone. In contrast to findings in very low birth weight premature infants supporting this speculation,6 in the present study neither oxygen-releasing capacity 12 nor available oxygen 11,2o improved on the relationship observed between erythropoietin and hemoglobin in normal term infants. The significant inverse association found between plasma eryttu-opoietin and hemoglobin in normal infants is consis-

The Journal of Pediatrics Volume 128, Number 6

tent with observations made in individuals with pathologic anemia occurring during other periods of life.6' 13-15In comparing the slope of the infants' semilogarithmic hemoglobin-erythropoietin plots with those of anemic fetuses during the last haft of pregnancy 13 and with anemic premature infants, 6 strikingly similar regression slopes were observed (Fig. 2). Only marginally steeper slopes and slightly higher absolute plasma erythropoietin levels are observed among fetuses and term infants. In contrast, the hemoglobin-erythropoietin slopes observed for anemic young children 14 and anemic adults 15 are much steeper than any of the three developmentally less mature groups. Because the range of hemoglobin values in very low birth weight and term infants was less than the other groups and the slope of the hemoglobin-erythi'opoietin plots is greatly influenced by the range of hemoglobin values, the slope comparisons are somewhat limited. However, the observed differences were striking, and all studies except that of Bray et al.14 used the same erythropoietin radioimmunoassay, limiting the methodologic variability. The switch in site of erythropoietin production from fetal liver to the postnatal kidney 2' 21 may be responsible for the developmental difference in the hemoglobin-erythropoietin relationship. Although this study was not designed to study developmental mechanisms, it supports the in vitro findings of decreased translational efficiency in the liver compared with kidney 2 and additional negative regulatory elements that repress postnatal hepatic erythropoietin expression.4 Alternatively, this developmental difference in the hemoglobin-erythropoietin relationship is consistent with erythropoietin undergoing more rapid catabolism in early development. This latter hypothesis is supported by pharmacokinetic data in immature human subjects and sheep in which erythropoietin plasma clearances two to five times greater than adults were demonstrated. 22-24 In summary, plasma erythropoietin levels measured in healthy term infants during the first half year of life demonstrated changes that were inversely associated with hemoglobin concentration and directly associated with reticulocyte count. Increased erythropoielin levels were observed as the normal postnatal decrease in hemoglobin occurred. This increase coincided with evidence of a stimulation of erythropoiesis. Despite the flatter slope of the hemoglobinerythropoietin relationship in normal term in/ants relative to comparably anemic children and adults, our data provide evidence that erythropoietin plays a major role in the regulation of erythropoiesis during the physiologic anemia of normal infancy. We thank Leon F. Burmeister, PhD, for statistical advice, and Ronald R. Rogers, BS, Steven E. Nelson BA, and James A. Kohler, BA, for technical assistance.

Kling et al.

795

REFERENCES

1. Dallman PR. Anemia of prematarity. Ann Rev Med 1981; 32:143-60. 2. Eckardt K-U, Ratcliffe PJ, Tan CC, Bauer C, Kurtz A. Age-dependent expression of the erythropoieitin gene in rat liver and kidneys. J Clin Invest 1992;89:753-60. 3. Finne PH, Halvorsen S. Regulation of erythropoiesis in the ferns and newborn. Arch Dis Child 1972;47:683-7. 4. Semenza GL. Regulation of erythropoietin production: new insights into molecular mechanisms of oxygen homeostasis. Hematol Oncol Clin North Am 1994;8:863-84. 5. Buchanan GR, Schwartz AD. Impaired erythropoietin response in anemic premature infants. Blood 1974;44:347-52. 6. Brown MS, Garcia IF, Phibbs RH, Dallman PR. Decreased response of plasma immunoreactive erythropoiefin to "available oxygen" in anemia of prematurity. J Pediatr 1984;105: 793-8. 7. Stockman JA HI, Graebe JE, Clark DA, McClellan K, Garcia IF, Kavey REW. Anemia of prematttrity: determinants of tile erythropoietin response. J Pediatr 1984;105:786-92. 8. Georgieff MK, Landon MB, Mills MM, et al. Abnormal iron distribution in infants of diabetic mothers: spectrum and maternal antecedents. J Pediatr 1990;117:455-61. 9. Fomon SJ, Janghorbani M, Ting BTG, et al. Erythrocyte incorporation of ingested 58-iron by infants. Pediatr Res 1988; 24:20-4. 10. Davis CS. A computer program for regression analysis of repeated measures using generalized estimating equations. Computer Methods and Programs in Biomedicine 1993;40:1531. 11. Jones JG, Holland BM, Veale KEA, Wardrop CAJ. 'Available oxygen,' a realistic expression of the ability of the blood to supply oxygen to tissues. Scand J Haematol 1979;22:7782. 12. Delivoria-Papadopoulos M, Roncevic N, Oski FA. Postnatal change in oxygen transport of term, premature, and sick infants: the role of red cell 2,3-diphosphoglycerate and adult hemoglobin. Pediatr Res 1971;5:235-45. 13. Moya FR, Grannum PAT, Widness JA, Clemons GK, Copel JA, Hobbins JC. Erythropoietin in human fetuses with immune hemolytic anemia and hydrops fetalis. Obstet Gyneco11993;82: 353-8. 14. Bray GL, Taylor B, O'Donnell R. Comparison of the erythropoietin response in children with aplastic anemia, transient erythroblastopenia, and iron deficiency. J Pediatr 1992; 120:52832. 15. Garcia IF, Ebbe SN, Hollander L, Cutting HO, Miller ME, Cronkite EP. Radioimmunoassay of erythropoietin: circulating levels in normal and polycythemic human beings. J Lab Clin Med 1982;99:624-35. 16. Widness JA, Clemons GK, Garcia IF, Schwartz R. Plasma inlmunoreactive erythropoietin in sequentially studied normal women during and after pregnancy. Am J Obstet Gynecol 1984; 194:646-50. 17. Hhg~ P, Cotes PM, Till JA, Shinebourne EA, Halvorsen S. ls oxygen supply the only regulator of erythropoietin levels? Serum immunoreactive erythropoietin during the first 4 months of life in term infants with different levels of arterial oxygenation. Acta Pediatr Scand 1987;76:907-13. 18. Hellebostad M, H~gh P, Cotes PM. Serum immunoreactive erythropoietin in healthy normal children. Br J Haematol 1988; 70:247-50.

796

Kling et al.

The Journal of Pediatrics June 1996

19. Yamashita H, Kukita J, Ohga S, Nakayama H, Akazawa K, Ueda K. Serum erythropoietinlevels in term and preterm infant during the first year of life. Am J Pediatr Hematol Oucol 1994;16:213-8. 20. Wardrop CAJ, Holland BM, Veale KEA, Jones JG, Gray DP. Nonphysiologicanemia of prematurity. Arch Dis Child 1978; 53:855-60. 21. Zanjani ED, AscensaoJL, McGlave PB, BanisadreM, Ash RC. Studies in the liver to kidney switch of erythropoietinproduction. J Clin Invest 1981;67:1183-8. 22. Widness JA, Veng-Pedersen P, Peters C, Pereira L, Schmidt

R, et al. Erythropoietinpharmacokineticsin very low birthweight infants:demonstrationof developmentaldifferences,nonlinearity, and treatment effects. J Appl Physiol 1996;80:140-8. 23. Kling PJ, Widness JA, Guillery EN, Veng-PedersenP, Peters C, deAlarconPA. Pharmacokineticsand pharrnacodynamicsof erythropoietin during therapy in an infant with renal failure. J Pediatr 1992;121:822-5. 24. Brown MS, Jones MA, Ohls RK, ChristensenRD. Single-dose pharmacokineticsof recombinanthuman erythropoietinin preterm infants after intravenousand subcutaneousadministration. J Pediatr 1993;122:655-7.

BOUND VOLUMES AVAILABLE TO SUBSCRIBERS Bound volumes of the 1996 issues of The Journal of Pediatrics axe available to subscribers (only) from the Publisher, at a cost of $81.00 for domestic, $103.79 for Canadian, and $97.00 for international subscribers, for Vol. 128 (January-June) and Vol. 129 (July-December), shipping charges included. Each bound volume contains subject and author indexes, and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram, with the Journal name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Mosby-Year Book, Inc., Subscription Services, 11830 Westline Industrial Dr., St. Louis, MO 63146-3318, USA/800-453-4351, or 314-4534351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular Journal subscription.