Secondary Logo

Share this article on:

Detection of Darbepoetin Alfa Misuse in Urine and Blood: A Preliminary Investigation


Medicine & Science in Sports & Exercise: October 2007 - Volume 39 - Issue 10 - p 1742-1747
doi: 10.1249/mss.0b013e31811e9d55
BASIC SCIENCES: Original Investigations

Introduction: Darbepoetin alfa is a modified erythropoietin (EPO) molecule with a longer serum half-life than recombinant human erythropoietin (rhEPO). Because the detection period of rhEPO in urine is only 2-3 d after the last injection, blood algorithms have been developed in order to expand the detection window of rhEPO misuse. The main objectives were to establish the period of detection of darbepoetin alfa by isoelectric focusing (IEF) and examine the applicability of blood algorithms and individual variations in blood variables in an antidoping context.

Methods: Six recreationally active males and six recreationally active females had 0.78 μg·kg−1·wk−1 of darbepoetin alfa administered for 3 wk. Blood and urine samples were collected continuously during and after administration. Urine samples were analyzed by IEF and immunoblotting for darbepoetin alfa, and blood samples were analyzed for erythropoietic sensitive blood variables on a hematological analyzer.

Results: Darbepoetin alfa was detected in 8 of 12 samples at 10 d after the last injection. Ten subjects showed variations in hemoglobin concentration [Hb] > 10%, whereas only three males and one female exceeded suggested upper [Hb] limits of 17.0 and 16.0 g·dL−1, respectively. Four subjects exceeded the 1:1000 ON- as well as the OFF-model cutoff limit.

Conclusion: The large number of samples containing detectable amounts of darbepoetin alfa at 10 d into the washout period stipulate the possibility of a 7-d window of detection after administration, wherein a sample would be regarded as an adverse analytical finding. The marked variations in all examined blood parameters could be used for the targeting of urine samples. These preliminary findings open up for larger scale studies with more frequent urine sampling in the washout period on elite athletes.

1The Copenhagen Muscle Research Centre, Rigshospitalet, University of Copenhagen, Copenhagen, DENMARK; 2Hormone Laboratory, Aker University Hospital, Oslo, NORWAY, and 3Center for Idraet, University of Aarhus, Aarhus, DENMARK

Address for correspondence: Jakob Morkeberg, Department of Anaesthesiology, Bispebjerg Hospital, Bispebjerg Bakke 23, DK 2400 Copenhagen NV, Denmark; E-mail:

Submitted for publication November 2006.

Accepted for publication April 2007.

The objective of any kind of blood doping is to increase the amount of circulating hemoglobin in the blood (16). One of the methods by which an increased amount of hemoglobin can be attained is by administration of pharmacological agents such as recombinant human erythropoietin (rhEPO) (8). Modification of this molecule has led to rhEPO analogues with different qualities than the original rhEPO molecule (12). Darbepoetin alfa is a rhEPO analogue (9). The EPO molecule consists of a peptide portion with a stable amino acid sequence and a highly variable carbohydrate component (21-23). The carbohydrate chains are terminated with a negatively charged sugar molecule, a sialic acid (21). It has been demonstrated that the number of sialic acids correlates positively with its serum half-life (13). Since darbepoetin alfa has been designed with more sialic acids residues than rhEPO, the presence of darbepoetin alfa in serum after injections should last approximately three times longer than that of rhEPO (12). The difference between the number of sialic acid residues and thereby the net negative charge makes it possible to separate EPO analogues from endogenous EPO (uhEPO) by isoelectric focusing (IEF) and has shown dominant bands of darbepoetin alfa clustered in the acidic area of the electropherogram (10). In contrast, administered rHuEPO has shown dense isoforms in the least acidic area (15). Because of the longer serum half-life of darbepoetin alfa, a longer detection window after cessation of treatment compared with rHuEPO would be expected.

Although IEF is capable of detecting rhEPO misuse, the period of detection is limited to 2-3 d. Furthermore, as the hemoglobin concentration [Hb] is still increased weeks after the last injection, the athlete can avoid penalization by terminating injections before competition. Therefore, an increased interest in developing blood-based screening tools with longer detection periods than the direct EPO test has evolved. Algorithms and cutoff limits, based on changes in hemoglobin concentration ([Hb]) and serum erythropoietin (s-EPO) during (ON-he model) and [Hb] and reticulocytes after (OFF-hr model) EPO treatment, have been developed and refined to screen for potential EPO abusers (14,20). Despite the great specificity of blood algorithms, they are considered inadequate proof of doping abuse in most federations. However, some international federations have, from a health perspective, suggested that male and female athletes with [Hb] above 17.0 and 16.0 g·dL−1, respectively, should be temporarily prevented from participating in competitions. Furthermore, a reticulocyte percentage > 2.0 and/or ≤ 0.2 are in cycling used as indicators of doping abuse. Such fixed cutoffs do not take large interindividual differences in baseline levels into account. Therefore, Malcovati et al. (17) have proposed the implementation of an individual hematological passport including individual baseline levels and reference ranges. This proposal has received much interest in recent time and the World Anti-Doping Agency (WADA) has initiated an approach to use individual blood profiles in future antidoping work.

The period of darbepoetin alfa detection in urine from healthy subjects after administration is unknown. In addition, the effect of darbepoetin alfa on selected biomarkers of erythropoiesis, and, hence, blood algorithms in healthy males and females have never been examined. The main objectives of this investigation were to determine the duration of detection in urine after cessation of treatment and to examine the variations in blood variables during and after darbepoetin administration in recreationally active males and females.

Because of a longer serum half-life compared with rhEPO (12), we hypothesized that darbepoetin alfa could be detected by IEF at 10 d after the last injection, and that the administration protocol would induce marked fluctuations in blood parameters typically used in antidoping testing.

Back to Top | Article Outline



In total, 12 (6 males and 6 females) healthy, recreationally active subjects took part in the study. Subject characteristics (mean ± SD) for males were: age, 26.0 ± 3.2 yr; height, 181.3 ± 3.5 cm; body mass, 79.2 ± 8.6 kg; BMI, 24.1 ± 2.6; and V˙O2max, 58 ± 7 mL·min−1·kg−1. The corresponding values for females were: age, 24.5 ± 4.5 yr; height, 167.1 ± 7.2 cm; body mass, 57.6 ± 12.8 kg; BMI, 20.4 ± 3.0; and V˙O2max, 51 ± 9 mL·min−1·kg−1. All subjects were nonsmokers and did not take part in any kind of organized sports. No training was engaged during the study period. The study was performed in accordance with the Helsinki II declaration and was approved by the local ethics committee of Copenhagen, Denmark (KF 01-070/03). All subjects were informed orally and in writing. Written informed consent was obtained from each participant.

Back to Top | Article Outline


All subjects had subcutaneous injections of 0.78 μg·kg−1·wk−1 of darbepoetin alfa (Aranesp, Amgen Europe B.V., Netherlands) administered weekly for 3 wk, corresponding to 156 IU·kg−1·wk−1 of epoetin beta. Dosages of 40-50 IU·kg−1 of epoetin beta injected three times per week (120-150 IU·kg−1·wk−1) for 2-5 wk have shown to rapidly elevate the hematocrit in studies from which the blood algorithms were developed (14). Furthermore, corresponding dosages have increased V˙O2max by 5 mL·min−1·kg−1 (8%) in well-trained subjects (4,6). Compared with standard practices, the dosage was higher than dosages used for the treatment of patients with chronic renal failure (0.45 μg·kg−1·wk−1) (9) but lower than dosages used for the treatment of chemotherapy-induced anemia in cancer patients (2.25 μg·kg−1·wk−1) (7). One week before the study until 2 wk after, female and male subjects were orally supplemented each day with 200 and 100 mg of iron, respectively.

Urine samples were collected at baseline and 2 or 3 d after each injection. After the last injection, samples were collected every week for a period of 2 months. Urine samples were analyzed according to the method by Lasne et al. (15). The samples were up concentrated from 20 mL to 50 μL and were analyzed by isoelectric focusing, separating the different molecular glycoforms on a gel. The proteins were transferred to a membrane and recognized by a monoclonal antibody specific for EPO. The proteins were detected by a chemiluminescence substrate that is read by a charge coupled device camera (LAS 1000 Plus, Fujifilm) (Fig. 1). The placement of the glycoforms on the final picture reveals whether the sample contains human urinary EPO (uhEPO), rhuEPO, or darbepoetin alfa. The four glycoforms representing darbepoetin alfa in the IEF gels are located in the most acidic part of the gel. The intensity of the chemiluminescent signals represents the amount of protein blotted onto the membrane. Lanes containing detectable amounts of darbepoetin alfa were evaluated by two main criteria described in the WADA Technical Document-TD2004EPO ( The acceptance criteria were performed visually. The identification criteria, defining whether the sample could be regarded as an adverse analytical finding for darbepoetin alfa or not, was automatically performed by quantitative analysis using the software GASepo (ARC Seibersdorf Research GmbH, Austria) (5). The software is developed for evaluating rhEPO and darbepoetin alfa separated by isoelectric focusing in doping control. The reference substance used in these analyses was Aranesp (Amgen, Thousand Oaks, CA).



Blood sampling was standardized as nonfasting samples with subjects seated for 10 min before venous puncture at baseline and on days 3, 7, 10, 14, and 21 after initiation of administration and on days 7, 10, 14, 17, 21, 24, 31, 38, 45, and 52 after cessation of administration (washout period).

[Hb], hematocrit, and reticulocyte count were measured using an automatic hematology system (ADVIA 120, Hematology System; Bayer Diagnostics, Tarrytown, NY). The coefficient of variation of RBC, [Hb], and reticulocyte count were 1.2, 0.9, and 12.5%, respectively. EPO and soluble transferrin receptor (sTfR) were measured in serum samples. After coagulation, serum was separated from the red blood cells by centrifugation for 15 min at 3500 rpm at 4°C and thereafter stored in Eppendorf tubes at −20°C until analyzed. sTfR and s-EPO concentrations were determined in duplicate by ELISA-kits (R&D systems Inc., Quantakine IVD human sTfR or Quantakine IVD erythropoietin, ELISA). The detection limit for sTfR was 0.5 nM, whereas the intra- and interassay coefficients of variation were 7.1 and 6.4% for mean sTfR levels of 18 and 11 nM, respectively. The detection limit of s-EPO was 0.6 mIU·mL−1, whereas the intra- and interassay coefficients of variation for mean s-EPO levels at 6 mIU·mL−1 were 7.8 and 10.3%, respectively.

Back to Top | Article Outline

Statistical analyses.

A one-way repeated-measures analysis of variance (ANOVA) was performed to test for significance in the selected hematological variables. After a significant F test, pairwise differences were identified using Tukey's honestly significant difference (HSD) post hoc procedure. Phenotypic data are presented as means ± SD, whereas results are presented as means ± SEM. Statistical analyses were performed using a statistical package (SigmaStat, Version 3.11). The significance level was set at P ≤ 0.05.

Back to Top | Article Outline


Results showed that darbepoetin alfa was detectable in 11 out of 12 samples taken 2-3 d after each of the first two injections and in every sample 2-3 d after the last two injections (Fig. 2). All samples except one fulfilled the identification criteria and were considered as an adverse finding for darbepoetin alfa. One week later (10 d after the last injection), darbepoetin alfa was detectable in samples from 8 out of 12 subjects. Only two samples fulfilled both the acceptance criteria and identification criteria for an adverse analytical finding. For samples taken 17 and 24 d after the last injection, the numbers of adverse analytical findings were two and one, respectively.



Darbepoetin alfa administration gradually increased [Hb], reaching maximum levels 17 d after the last injection (Fig. 3). In total, one third of the male and female subjects exceeded 17.0 and 16.0 g·dL−1 in [Hb], respectively (Table 1). In 10 out of 12 subjects, variations greater than 10% in [Hb] were observed (Table 1). Three days after each injection, there was a marked (six- to ninefold) but short-lasting increase in s-EPO, decreasing to levels not different from baseline at 7 d after injections (Fig. 3). After the last injection, s-EPO levels decreased, reaching values significantly lower (P = 0.019) than baseline only in males at 17 d after treatment. One week after the first injection, reticulocyte levels were 168% (P < 0.001) and 116% (P < 0.001) higher than baseline levels in males and females, respectively. During the following 2 wk, reticulocyte levels were above 2% in all subjects (Fig. 3). The levels remained significantly elevated during the rest of the treatment period. Although levels were significantly suppressed inboth genders after treatment, only one subject (male) experienced a reticulocyte percentage below 0.3% (Table 1). sTfR levels doubled during the treatment period in both genders, reaching peak values at 21 d after the first injection. sTfR remained elevated until 2 wk into the washout period (Fig. 3).





ON-model (he) scores were significantly different from baseline values 10 d after the first injection and during the rest of the administration period in both genders. Scores from 33% of the subjects exceeded the 1-in-1000 false-positive cutoff values (Table 1). Scores that exceeded the cutoffs were obtained in every subject at least once and, at most, twice. All subjects had increases in ON-model scores by more than 15% (Table 1). OFF-model (hr) scores were significantly different from baseline values from day 7 until day 31 after administration in males and from day 10 until the end of the washout period in females. Scores exceeding the 1-in-1000 cutoff were present in 33% of all subjects (Table 1). Increases by more than 25% in the OFF-model scores were observed in all subjects (Table 1).

The statistical power was > 0.8 in all analysis except in the s-EPO postadministration analysis for males (power = 0.68) and females (power = 0.26).

Back to Top | Article Outline


We found that darbepoetin alfa was detected by IEF in 46 of 48 urine samples during and 2-3 d after the treatment period, giving a sensitivity (true adverse analytical finding/[true adverse analytical finding + false negative finding]) of 96%. In 8 out of 12 samples, or 67% of the samples taken 10 d after the last injection, detectable amounts of darbepoetin alfa were found. In comparison, the hematological 1-in-1000 ON- as well as OFF-model cutoff score only deemed positive scores in 4 out of 12 subjects once during the sampling period. All subjects had increases in ON- and OFF-model scores of 15 and 25%, respectively. Moreover, 10 out of 12 subjects had variations in Hb above 10%.

Our data reveal that darbepoetin alfa was detected in the majority of samples at 10 d after cessation of treatment. Data from a corresponding rhEPO trial, where subjects had 50 IU·kg−1 of epoetin alfa administered subcutaneously on nine different occasions during a 3-wk period, revealed that rhEPO was detected in 50% of the urine samples at 7 d after the end of treatment (9). Because of the overlapping charge distributions and, thereby, the overlapping bands between endogenous and recombinant EPO (Fig. 1), the number of tests with rhEPO traces are not equal to the number of EPO tests reported as adverse analytical findings. Results from the laboratory in Oslo show that at 7 d after the last rhEPO injection, none of the samples were categorized as adverse analytical findings, although many of them contained traces of rhEPO (18). It was, therefore, concluded that the period of detection is maximally 3 d after the last injection. Although a direct comparison between separate and distinct studies with different populations should be done with caution, darbepoetin alfa seems to be detectable by IEF and immunoblotting for a longer period than rhEPO. This is probably explained by darbepoetin alfa's longer serum half-life. Although only two of the test results were considered adverse analytical findings in an antidoping context, our results stipulate the possibility of adverse analytical findings in the majority of samples taken 6-7 d after injections.

Even though most sport federations are not willing to sanction/penalize athletes on the basis of changes in their blood profiles, blood-based algorithms have been developed as a screening tool for targeting potential rhEPO abusers (19). The algorithms were developed by comparing the levels of different blood variables from a population of recreational active athletes (training volume = 1.5-20 h·wk−1) or former elite cross-country skiers treated with rhEPO and a large cohort of supposedly "clean" elite athletes (14). Because a period of rhEPO treatment results in an immediate, marked, and persistent increase in s-EPO and a delayed increase in [Hb], the combination of these two parameters led to the development of the ON-he model (14). The OFF-model was based on the concomitant increase in [Hb] and decrease in the percentage of reticulocytes after administration.

Our results show a detection rate of 33% by the ON- and OFF-model. In contrast, Gore et al. (14) found rates of detection of 80% and above in ON- as well as OFF-model scores at certain time points after administering rhEPO dosages of 50 UI·kg−1 for 3 wk followed by "titration" dosages of 20 UI·kg−1 for 5 wk. Besides the limited number of subjects in our study (N = 12) compared with the study by Gore et al. (N = 57), it must be stated that these algorithms are based on data from and developed for athletes. Athletes in general (14), and endurance athletes in particular (24), seem to have higher baseline scores than sedentary people and non-endurance-trained athletes, respectively. This is attributable to higher baseline [Hb] and [s-EPO]-the latter probably explained by a higher red cell turnover caused by exercise (25). Because the relative increase in [Hb] during treatment with erythropoiesis stimulating agents such as darbepoetin alfa or epoetin alfa/beta does not depend on the initial [Hb] level (7), a higher initial baseline value makes less room for changes before reaching a standardized cutoff limit. Furthermore, this delta value is decreased by up to 3 wk after descent from altitude (20-23 d at 2650-3000 m) mainly by an increase in [Hb] (1), which should be taken into consideration when testing athletes. Effects are smaller or negligible for simulated altitude (1).

With these differences in mind, a suitable alternative to fixed cutoffs could be an individual hematological passport. With a hematological passport, it would be possible to continuously monitor athletes, thereby giving information of changes in blood variables that could be caused by any kind of illegal practice. This makes it suitable for detecting all kinds of blood manipulations irrespective of the presence of a direct testing procedure. Even autologous transfusions, which are undetectable at the moment, might be revealed by continuously measuring different blood variables.

On the basis of 2506 hematological determinations from 923 professional football players, Malcovati et al. found the 95th percentile of the coefficient of variation distribution for [Hb] to be lower than 5% during the year (17). The authors conclude that increases in [Hb] above 10% could not be explained by natural variations. All subjects in our study increased their [Hb] with more than 7%, whereas 10 out of 12 subjects had increases above 10%. On the other hand, only 4 out of 12 exceeded the upper [Hb] limits of 17.0 and 16.0 g·dL−1 for males and females, respectively. Upper [Hb] limits leave room for large individual fluctuations without reaching upper limits. This issue is further highlighted by evaluating individual changes in algorithm scores. All subjects experienced changes in ON- and OFF-model scores of more than 15 and 25%, respectively. Four of these subjects were close to, but did not exceed, the 1:1000 OFF-model cutoff limit. Individual variations in [HB] > 10% and 1:1000 OFF-model scores > 25%, therefore, seem to be more relevant than universal cutoff limits in the detection of blood-stimulating substances.

Our data show a pronounced and long-lasting increase (> 90%) in the percentage of reticulocytes during treatment in all subjects with absolute levels above 2%. Similar perturbations are seen during rhEPO administration (19) and autologous blood transfusions (11) and are much greater than during altitude exposure (2). After treatment, only one subject had a reticulocyte percentage below 0.3. Because of the continuous turnover of red blood cells, even during periods of pronounced polycythemia caused by blood doping (3), a marked decrease (< 0.3%) in the reticulocyte percentage is rare. Until the implementation of a hematological passport, a higher limit seems to be more suitable.

In conclusion, 67% of all samples collected at 10 d into the washout period contained detectable amounts of darbepoetin alfa. This finding stipulates the possibility of a 7-d window of detection after administration, wherein samples would be regarded as adverse analytical findings. Data suggest that changes in [Hb] > 10% or ON- and OFF-model > 15 and 25%, respectively, should elicit a follow-up urine test at an appropriate time and stipulate a short, health-related quarantine until [Hb] or the ON- and OFF-models are normalized within an individually based reference range. These preliminary findings on recreationally active subjects open up for larger-scale studies with more frequent urine sampling in the washout period from a larger cohort of subjects representing the target population, the elite athletes.

We give special thanks to the volunteer subjects. This study was supported by The Ministry of Culture's Sport Research Committee (KIF).

Back to Top | Article Outline


1. Ashenden, M. J., C. J. Gore, R. Parisotto, K. Sharpe, W. G. Hopkins, and A. G. Hahn. Effect of altitude on second-generation blood tests to detect erythropoletin abuse by athletes. Haematologica 88:1053-1062, 2003.
2. Ashenden, M. J., A. G. Hahn, D. T. Martin, P. Logan, R. Parisotto, and C. J. Gore. A comparison of the physiological response to simulated altitude exposure and R-HuEpo administration. J. Sports Sci. 19:831-837, 2001.
3. Ashenden, M. J., K. Sharpe, R. Damsgaard, and L. Jarvis. Standardization of reticulocyte values in an antidoping context. Am. J. Clin. Pathol. 121:816-825, 2004.
4. Audran, M., R. Gareau, S. Matecki, et al. Effects of erythropoietin administration in training athletes and possible indirect detection in doping control. Med. Sci. Sports Exerc. 31:639-645, 1999.
5. Bajla, I., I. Hollander, M. Minichmayr, G. Gmeiner, and C. Reichel. GASepo-a software solution for quantitative analysis of digital images in Epo doping control. Comput. Methods Programs Biomed. 80:246-270, 2005.
6. Birkeland, K. I., J. Stray-Gundersen, P. Hemmersbach, J. Hallen, E. Haug, and R. Bahr. Effect of RhEPO administration on serum levels of STfR and cycling performance. Med. Sci. Sports Exerc. 32:1238-1243, 2000.
7. Bokemeyer, C., M. S. Aapro, A. Courdi, et al. EORTC guidelines for the use of erythropoietic proteins in anaemic patients with cancer: 2006 update. Eur. J. Cancer 43:258-270, 2007.
8. Breidbach, A., D. H. Catlin, G. A. Green, I. Tregub, H. Truong, and J. Gorzek. Detection of recombinant human erythropoietin in urine by isoelectric focusing. Clin. Chem. 49:901-907, 2003.
9. Cases, A. Darbepoetin alfa: a novel erythropoiesis-stimulating protein. Drugs Today (Barc.) 39:477-495, 2003.
10. Catlin, D. H., A. Breidbach, S. Elliott, and J. Glaspy. Comparison of the isoelectric focusing patterns of darbepoetin alfa, recombinant human erythropoietin, and endogenous erythropoietin from human urine. Clin. Chem. 48:2057-2059, 2002.
11. Damsgaard, R., T. Munch, J. Morkeberg, S. P. Mortensen, and J.Gonzalez-Alonso. Effects of blood withdrawal and reinfusion on biomarkers of erythropoiesis in humans: implications for anti-doping strategies. Haematologica 91:1006-1008, 2006.
12. Egrie, J. C., E. Dwyer, J. K. Browne, A. Hitz, and M. A. Lykos. Darbepoetin alfa has a longer circulating half-life and greater in vivo potency than recombinant human erythropoietin. Exp. Hematol. 31:290-299, 2003.
13. Elliott, S., J. Egrie, J. Browne, et al. Control of RHuEPO biological activity: the role of carbohydrate. Exp. Hematol. 32:1146-1155, 2004.
14. Gore, C. J., R. Parisotto, M. J. Ashenden, et al. Second-generation blood tests to detect erythropoietin abuse by athletes. Haematologica 88:333-344, 2003.
15. Lasne, F., and J. de Ceaurriz. Recombinant erythropoietin in urine. Nature 405:635, 2000.
16. Lippi, G., M. Franchini, G. L. Salvagno, and G. C. Guidi. Biochemistry, physiology, and complications of blood doping: facts and speculation. Crit. Rev. Clin. Lab. Sci. 43:349-391, 2006.
17. Malcovati, L., C. Pascutto, and M. Cazzola. Hematologic passport for athletes competing in endurance sports: a feasibility study. Haematologica 88:570-581, 2003.
18. Nissen-Lie, G., K. Birkeland, P. Hemmersbach, and V. Skibeli. Serum STfR levels may indicate charge profiling of urinary R-HEPO in doping control. Med. Sci. Sports Exerc. 36:588-593, 2004.
19. Parisotto, R., C. J. Gore, A. G. Hahn, et al. Reticulocyte parameters as potential discriminators of recombinant human erythropoietin abuse in elite athletes. Int. J. Sports Med. 21:471-479, 2000.
20. Parisotto, R., M. T. Wu, M. J. Ashenden, et al. Detection of recombinant human erythropoietin abuse in athletes utilizing markers of altered erythropoiesis. Haematologica 86:128-137, 2001.
21. Rush, R. S., P. L. Derby, D. M. Smith, et al. Microheterogeneity of erythropoietin carbohydrate structure. Anal. Chem. 67:1442-1452, 1995.
22. Rush, R. S., P. L. Derby, T. W. Strickland, and M. F. Rohde. Peptide-Mapping and evaluation of glycopeptide microheterogeneity derived from endoproteinase digestion of erythropoietin by affinity high-performance capillary electrophoresis. Anal. Chem. 65:1834-1842, 1993.
23. Sasaki, H., B. Bothner, A. Dell, and M. Fukuda. Carbohydrate structure of erythropoietin expressed in Chinese-hamster ovary cells by a human erythropoietin Cdna. J. Biol. Chem. 262:12059-12076, 1987.
24. Sharpe, K., W. Hopkins, K. R. Emslie, et al. Development of reference ranges in elite athletes for markers of altered erythropoiesis. Haematologica 87:1248-1257, 2002.
25. Weight, L. M., M. J. Byrne, and P. Jacobs. Haemolytic effects of exercise. Clin. Sci. (Lond.) 81:147-152, 1991.


©2007The American College of Sports Medicine