Secondary Logo

Journal Logo

Increased body iron stores in elite road cyclists


Medicine & Science in Sports & Exercise: May 2002 - Volume 34 - Issue 5 - p 876-880
APPLIED SCIENCES: Physical Fitness and Performance

DEUGNIER, Y., O. LORÉAL, F. CARRÉ, A. DUVALLET, F. ZOULIM, J. P. VINEL, J. C. PARIS, D. BLAISON, R. MOIRAND, B. TURLIN, Y. GANDON, V. DAVID, A. MÉGRET, and M. GUINOT. Increased body iron stores in elite road cyclists. Med. Sci. Sports Exerc., Vol. 34, No. 5, pp. 876–880, 2002.

Background One third of French elite road cyclists were found to have hyperferritinemia on antidoping control tests performed during the Tour de France in 1998.

Purpose This study was undertaken to determine whether hyperferritinemia corresponded to elevated body iron stores or not and, affirmatively, what were its mechanism, its clinical consequences, and its spontaneous course.

Methods 83 elite road male cyclists presenting with hyperferritinemia, defined as serum ferritin level greater than 300 μg·L−1, were studied with respect to consumption of iron and other drugs, serum iron tests, HFE mutations, and hepatic iron concentration (HIC; N < 35 μmol·g−1 dry weight).

Results All cyclists were asymptomatic and had normal physical and cardiac examination. Their median (range) serum ferritin, serum iron, and transferrin saturation levels were 504 μg·L−1 (306–1671), 20 μmol·L−1 (8.5–36.3), and 39% (20–76), respectively. HIC was increased in 24/27 up to 187 μmol·g−1. Allelic frequency of the H63D mutation was increased in cyclists when compared to controls (P = 0.04). However, iron tests did not differ according to HFE genotypes. Most cyclists (89%) had been supplemented with iron. The median iron supplementation was 25.5 g (range: 1.4–336) and correlated well (P = 0.002) with serum ferritin. Evolution of serum ferritin levels did not differ whether cyclists had been continuing iron supplementation or not.

Conclusion Hyperferritinemia in elite road cyclists accounted for increased body iron stores caused by and persisting after cessation of excessive iron supplementation. Even when mild, iron excess may expose to long-term complications and should be removed, at least at the time when professional cyclists retire. To prevent iatrogenic iron overload, supplementation with iron must be done according to serum ferritin follow-up and not either blindly or on the basis of serum iron determination only.

Service des Maladies du Foie and INSERM-U522, Laboratoire de Physiologie, Laboratoire d’Anatomie Pathologique B, Département de Radiologie, and Laboratoire de Génétique Moléculaire and UMR 6061 CNRS, CHU Pontchaillou, Rennes, FRANCE; Service de Médecine du Sport, CHU Cochin, Paris, FRANCE; Service d’Hépatologie et de Gastro-Entérologie, CHU Hôtel Dieu and INSERM-U271, Lyon, FRANCE; Service des Maladies de L’Appareil Digestif, CHU Purpan, Toulouse, FRANCE; Service des Maladies de L’Appareil Digestif et de la Nutrition, CHU Claude Huriez, Lille, FRANCE; Service de Gastro-Entérologie, CH, Troyes, FRANCE; and Fédération Française de Cyclisme, Rosny sous Bois, FRANCE

Submitted for publication May 2001.

Accepted for publication October 2001.

In 1998, because of concerns about widespread doping in professional cycling, the French Cycling Organization launched systematic biological control tests in French elite road cyclists and showed that more than one third had hyperferritinemia (A. Mégret and M. Guinot, personal communication). This was quite unexpected because male endurance athletes usually have normal or even low body iron stores (4). The present study was undertaken to determine whether hyperferritinemia in elite road cyclists corresponded to elevated body iron stores or not and, affirmatively, what were its mechanisms, especially with respect to iron supplementation and to the C282Y and H63D mutations on the HFE gene involved in genetic hemochromatosis, its clinical consequences, and its spontaneous course.

Back to Top | Article Outline



All French elite road cyclists enrolled into French teams and presenting with serum ferritin levels greater than the upper limit of normal (300 μg·L−1) were contacted by the medical staff of the French Cycling Organization in order to enter the study and to benefit from the following records: (i) consumption of legal (iron, vitamins B and C, analgesics, nonsteroid anti-inflammatory drugs, etc.) and illegal (erythropoietin (EPO), steroids, growth hormone, insulin, etc.) drugs; (ii) physical examination with special reference to general, cutaneous, hepatic, cardiac, and osteoarticular symptoms and signs of iron overload; (iii) determination of serum iron tests (serum iron, transferrin saturation, and serum ferritin); (iv) search for conditions associated with hyperferritinemia, such as hemolysis (blood cell count), increased serum alanine (ALT) and aspartate (AST) transaminases, and inflammatory syndrome (C reactive protein; CRP); (v) testing for C282Y and H63D HFE mutations; (vi) 12-lead resting electrocardiogram and transthoracic echocardiography; and (vii) assessment of hepatic iron concentration by either magnetic resonance imaging (MRI) in those who had serum ferritin levels comprised between 300 and 800 μg·L−1, or liver biopsy in those who had either increased serum ALT levels or serum ferritin levels greater than 800 μg·L−1. In addition, serum ferritin levels determined 12 and 6 months before the study were recorded when available.

Data on drug consumption were recorded using a preestablished questionnaire listing all known drugs used by road cyclists. Daily dosage, duration, and administration mode were documented for each treatment, allowing for the calculation of cumulative dose, especially for iron. Biochemical tests were determined on fasting serum samples after, at least, 3 d without any physical training. Serum iron was assayed using the Feren colorimetric method (Kone Instruments, Norderstedt, Germany). Transferrin saturation was determined on the basis of serum iron and serum transferrin assayed by an immunoturbidimetric method (Kone Instruments). Serum ferritin was determined by an immunoenzymatic method (Kit Abbott Imx B22192, Abbott Laboratories, Abbott Park, IL). HFE testing was performed by amplification-restriction in the Laboratory of Molecular Genetics in Rennes. For each subject, three spots of whole blood were done on a filter paper card (Schleicher & Schuell ref. 10539977), stored at room temperature and mailed to Rennes. Then a disk of blood spot of 3 mm diameter was punched out and incubated for 30 min in sterile water to discard red cells. The amplification reaction was then carried out in 50 μL containing 30 pmol of each primer, 200 μM dNTP, and 1 μL of Advantaq polymerase 50 X (Clontech, Palo Alto, CA) in MgCl23 mM, tris HCl pH9 80 mM, (NH4)2SO420 mM. Amplification was performed for 94°C 1 min and 32 cycles (94°C 30 s, 55°C 45 s, 72°C 1 min). The primers were as follows: 5′ GCT GAT CTG ACT GCT CTC C 3′ and 5′ GAA AAA GCA AGT TAA AGC 3′. The conditions used for the PCR product restriction by RsaI were as previously described (9). Allelic frequencies of HFE mutations in cyclists were compared with those of a control population from Brittany, France, composed of 254 random healthy sedentary adults tested for C282Y and H63D (8). Determination of hepatic iron concentration (N < 35 μmol·g−1 dry weight) was made either by MRI as described previously (7) using 1.5-Tesla devices calibrated according to Gandon et al. (http: // or on liver biopsy according to the biochemical method of Barry and Sherlock (2). For echocardiography, M-mode recordings were taken at the tip of the mitral valve leaflets from the two-dimensional image of the long-axis left ventricle. Measurements of the left ventricular end-diastolic and end-systolic diameters and of interventricular and posterior wall thickness were performed at the peak of the R-wave according to Penn convention guidelines (17).

Descriptive and univariate analysis was performed using StatView 4.5 software (Abacus Concepts Inc, Berkeley, CA). Results are presented as median (range) for quantitative data and as percentage (number) for qualitative data. For serum ferritin, logarithm transformation was performed before statistical analysis because variable distribution was skewed. Wilcoxon test was used to compare values of a qualitative variable between two groups. Correlation between two quantitative data was assessed using Spearman test. Paired t-test was performed to compare serum ferritin levels at different times. Distribution of HFE genotypes in controls and cyclists were compared using Fisher exact test.

The study was conducted during autumn and winter in 1999 when cyclists were on vacation, usually at home. Six French referral centers—Lille, Lyon, Paris, Rennes, Toulouse, and Troyes—were selected. Cyclists were asked to attend the outpatients clinic of the closest center for clinical and biochemical evaluation. When hyperferritinemia was confirmed, a second visit was proposed for MRI (or liver biopsy), electrocardiogram, and echocardiography. All cyclists gave their written informed consent. The study was approved by the Ethical Committee of Rennes (#99/29-240).

Back to Top | Article Outline


Bioclinical data.

A total of 198 cyclists were contacted. Ninety-nine (50%) accepted to enter the study and attended the first visit. Hyperferritinemia was confirmed in 83 who composed the study group. Of these, 30 accepted cardiac evaluation and 27 accepted either MRI (N = 25) or liver biopsy (N = 2). The 83 cyclists were males aged from 23 to 37 yr (median: 28) with a body mass index ranging from 18.7 to 25.4 kg·m−2 (median: 21.9). All were asymptomatic and had normal physical examination. Their main biochemical results are summarized in Table 1. Serum ferritin was lower than 600 μg·L−1 in 58/83 (70%), comprised between 600 and 800 μg·L−1 in 15/83 (18%), and greater 800 μg·L−1 in 10/83 (12%). Serum iron was increased in 13/83 (16%). Transferrin saturation exceeded 45% in 21/57 (37%). Hepatic iron concentration was normal in 3, comprised between 35 and 80 μmol·g−1 in 15, and greater than 80 in 9. Serum ferritin and hepatic iron concentration levels were well correlated (r = 0.67, P = 0.001) as shown in Figure 1. Serum AST and ALT levels were increased in 11/83 (13%) and 6/83 (7%) cyclists, respectively, and did not correlate with neither serum ferritin levels nor hepatic iron concentration. Blood cell count and CRP were normal in all subjects. Results of HFE typing are shown in Table 2. As a whole, 57% of cyclists and 46% of controls had either the C282Y or the H63D mutation (difference not significant). Allelic frequency of H63D was significantly (P = 0.04) increased in cyclists when compared with healthy sedentary controls. This was due to homogeneous excess in H63D heterozygotes, H63D homozygotes, and compound heterozygotes (subjects heterozygous for both C282Y and H63D mutations). Iron tests did not differ significantly whether cyclists were wild homozygotes or not (Fig. 2) and carried the H63D mutation or not. In the four compound heterozygotes, transferrin saturation only was increased when compared with cyclists with other genotypes (57% vs 39%, P = 0.004).

Table 1

Table 1



Table 2

Table 2



Most cyclists (73/83, 88%) had been supplemented with iron. However, the amount of supplemented iron could be calculated precisely in 23 cyclists only. The largest part of iron supplementation had been taken orally during training periods and short races. The median cumulative oral iron intake was calculated as 23.4 g (range: 1–320) over a period of 3.9 yr (range: 1.5–15.7). Parenteral supplementation had been administered for shorter periods, especially during long stage races such as Tour de France and Giro, and accounted for an additional supplementation of 3.2 g (range: 0.4–47.5). As a whole, the median total iron supplementation was calculated as 25.5 g (range: 1.4–336) and correlated well (r = 0.66 -P = 0.002) with serum ferritin (Fig. 3). Evolution of serum ferritin levels did not differ whether cyclists had been continuing iron supplementation or not after the discovery of hyperferritinemia (Fig. 4). All cyclists had been taking various vitamin tablets for discontinuous periods over previous years. Seven admitted that they had been treated with erythropoietin without giving details. These cyclists did not differ from others with respect to iron status and blood cell count. No cyclist admitted that he had been treated by steroids, insulin, or growth hormone.





Cardiac evaluation was normal according to training level in all the 30 subjects who accepted to be investigated (16). Left ventricular end-diastolic and end-systolic diameters were 57.5 mm (range: 46–64) and 36.1 mm (range: 26–46), respectively. Diastolic left ventricular posterior wall and interventricular septal thickness were 9.9 mm (range: 7–15.2) and 11.2 mm (range: 7–16) respectively.

Back to Top | Article Outline

Histological data.

Two cyclists (TL14 and TR05) had a liver biopsy. Both were asymptomatic, had normal examination, and were free of both HFE mutations. TL14 had increased serum iron (32 μmol·L−1), transferrin saturation (77%), serum ferritin (1050 μg·L−1), and serum ALT (80 IU·L−1). He had been treated with oral and intramuscular iron up to 14 months before liver biopsy was performed but was unable to give more details. Liver biopsy was normal except for the presence of a marked parenchymal siderosis mimicking genetic hemochromatosis. Hepatic iron concentration was 187 μmol·g−1. TR05 had increased serum ferritin (1020 μg·L−1) only. He had been treated with intramuscular iron for a 1-yr period in 1998 (cumulative dose of 2.5 g) but not supplemented with iron tablets. Liver biopsy was also normal except for the presence of mixed parenchymal and mesenchymal iron overload. Hepatic iron concentration was 88 μmol·g−1.

Back to Top | Article Outline


Present data indicate that increased serum ferritin levels detected in French elite road cyclists were related to excessive enteral and parenteral iron supplementation. Indeed, no cause of isolated hyperferritinemia, such as inflammatory syndrome, chronic liver disease, or myolysis, was found whereas almost all cyclists declared that they had been given iron and presented with increased body iron stores well correlated to the amount of supplemented iron.

It is of note that the allelic frequency of the H63D mutation was significantly increased in cyclists when compared with healthy controls (8). However, this increase was slight. Thus, before suggesting that the H63D mutation may give some subtle metabolic advantage predisposing to higher performances, more data are needed on its allelic frequency in cyclists with normal serum iron tests and in controls matched according to geographical origin. Otherwise, it is unlikely that this mutation had played a significant role in the development of iron overload in cyclists because (i) serum and hepatic iron tests did not differ significantly whether cyclists had the H63D mutation or not and (ii) previous studies have shown that C282Y homozygosity and compound heterozygosity for both mutations were the only HFE genotypes associated with significantly increased body iron stores (3,13).

As shown by studies performed in long-distance runners, serum ferritin levels are usually normal in male endurance athletes despite some of them having low serum iron after heavy training (4). In addition, serum ferritin has been shown to correlate negatively with time of physical activity in leisure sportsmen (11), and there is no evidence that iron supplementation results in improving performances in athletes, except in case of iron deficiency (15). Therefore, there are no strong data suggesting that professional cyclists should be regularly and systematically supplemented with iron. The reasons why cyclists were so widely given iron remain unclear. Most cyclists claimed that they had been taking iron for two main reasons: the ability of iron to prevent or reduce leg cramps, especially during long stage races, and the finding of low serum iron on biochemical follow-up. However, such indications of iron supplementation are not supported by scientific studies (15). Moreover, it is well admitted that, unlike serum ferritin, serum iron does not reflect accurately body iron stores and can be either normal or low in patients with iron overload conditions (13,14). Actually, iron is widely considered as an aspecific and safe fortifying drug and widely taken by endurance athletes, most often by self-administration and without any biochemical follow-up. In addition, as declared by defendants involved in the recent “erythropoietin trial” held in Lille, France, and widely reported in the European popular press, most professional cyclists from every country have been using illegal drugs, especially erythropoietin, between 1995 and 1999. In this context, it is likely that many more than seven cyclists in the present study had been treated by erythropoietin and that iron supplementation was also given to enhance EPO-induced erythropoiesis (12) and to prevent iron deficiency related to drug-induced increased erythropoiesis (10).

Because there are no regulation mechanisms of iron excretion from the body (1), it is unlikely that iron excess can be spontaneously and rapidly removed. This is supported by the finding that serum ferrin levels in cyclists who had stopped iron supplementation 1 yr before the study remained stable. However, normalization of body iron stores is necessary because iron excess, even when mild, is associated with higher long-term risks of hepatic and extrahepatic carcinomas (6) and of cardiovascular diseases (5). This implies that iron overloaded cyclists should be followed and that iron excess should be removed in those keeping abnormal iron stores. However, in the present study, all cyclists rejected the proposal of venesection for fear of losing their endurance.

The present demonstration of iatrogenic iron overload in French elite road cyclists should lead (i) to advise cyclists that supplementation with iron must be done according to serum ferritin follow-up and not either blindly or on the basis of serum iron determination only, (ii) to set up a systematic follow-up of serum ferritin during professional life of cyclists, and (iii) to propose venesection therapy to those who keep increased iron stores at the time of retirement to prevent long-term complications of iron overload.

The authors are indebted to Mrs. Michèle Perrin for conducting the protocol as clinical research assistant. They thank Pr. P. Brissot, Pr. D. Guyader, and Dr. F. Lainé for allowing them to study those cyclists they examined.

This study was supported by grants from the French Ministère de la Jeunesse et des Sports, the Fondation Festina and the Association Fer et Foie.

Address for correspondence: Yves Deugnier, M.D., Clinique des Maladies du Foie and INSERM U522, CHU Pontchaillou, 35033 Rennes, France; E-mail

Back to Top | Article Outline


1. Andrews, N. C. Disorders of iron metabolism. N. Engl. J. Med. 341: 1986–1995, 1999.
2. Barry, M., and S. Sherlock. Measurement of liver iron concentration in needle-biopsy specimens. Lancet 1: 100–103, 1971.
3. Burt, M. J., P. M. George, J. D. Upton, et al. The significance of haemochromatosis gene mutations in the general population: implications for screening. Gut 43: 830–836, 1998.
4. Cook, J. The effect of endurance training on iron metabolism. Sem. Hematol. 31: 146–154, 1994.
5. de Valk, B., and J. Marx. Iron, atherosclerosis, and ischemic heart disease. Arch. Intern. Med. 159: 1542–1548, 1999.
6. Deugnier, Y., and O. Loréal. Iron as a carcinogen. In Hemochromatosis, J. Barton and C. Edwards (Eds.). Cambridge: Cambridge University Press, 2000, pp. 239–249.
7. Gandon, Y., D. Guyader, J. F. Heautot, et al. Hemochromatosis. diagnosis and quantification of liver iron with gradient-echo MR imaging. Radiology 193: 533–538, 1994.
8. Jezequel, P., M. Bargain, F. Lellouche, F. Geffroy, and I. Dorval. Allele frequencies of hereditary hemochromatosis gene mutations in a local population of west Brittany. Hum. Genet. 102: 332–333, 1998.
9. Jouanolle, A. M., P. Fergelot, G. Gandon, J. Yaouanq, J. Y. Le Gall, and V. David. A candidate gene for hemochromatosis: frequency of the C282Y and H63D mutations. Hum. Genet. 100: 544–547, 1997.
10. Karltwasser, J. P. Erythropoietin and iron. Kidney Int. 55: S49–S56, 1999.
11. Lakka, T., K. Nyssonen, and J. Salonen. Higher levels of conditioning leisure time physical activity are associated with reduced levels of stored iron in Finnish men. Am. J. Epidemiol. 15: 148–160, 1994.
12. Major, A., F. Mathy-Loic, R. Rohling, K. Gautski, and C. Brugnara. The effect of intravenous iron on the reticulocyte response to recombinant erythropoietin. Br. J. Haematol. 98: 5–9, 1997.
13. Moirand, R., A. M. Jouanolle, P. Brissot, J. Y. Le Gall, V. David, and Y. Deugnier. Phenotypic expression of Hfe mutations: a French study of 1110 unrelated iron-overloaded patients and relatives. Gastroenterology 116: 372–377, 1999.
14. Moirand, R., A. M. Mortaji, O. Loreal, F. Paillard, P. Brissot, and Y. Deugnier. A new syndrome of liver iron overload with normal transferrin saturation. Lancet 349: 95–97, 1997.
15. Nielsen, P., and D. Nachtigall. Iron supplementation in athletes. Sports Med. 26: 207–216, 1998.
16. Pellicia, A., B. Maron, A. Spataro, M. Proschan, and P. Spirito. The upper limit of physiologic cardiac hypertrophy in highly-trained elite athletes. N. Engl. J. Med. 324: 1378–1391, 1991.
17. Sahn, D. J., A. Demaria, J. Kisslo, and A. Weyman. Recommendations recording quantification in M-mode echocardiography: results of a survey of echocardiographic measurements. Circulation 58: 1072–1083, 1978.


©2002The American College of Sports Medicine