- Email: [email protected]
Screening for haematological and iron-related abnormalities in elite athletes—Analysis of 576 cases
Journal of Science and Medicine in Sport (2008) 11, 329—336
ORIGINAL PAPER
Screening for haematological and iron-related abnormalities in elite athletes—–Analysis of 576 cases Kieran E. Fallon Department of Sports Medicine, Australian Institute of Sport, Australia Received 17 November 2006 ; received in revised form 12 February 2007; accepted 18 February 2007 KEYWORDS Screening; Iron; Haematology; Elite; Athlete
Summary The aim of this prospective, cohort study was to determine the clinical and performance related utility of haematological and iron-related screening in elite athletes. Three hundred and three male and 273 female elite athletes underwent routine medical screening over a three-year period. In association with a standard medical consultation, a full blood count and iron-related variables were measured. Ten male athletes had a serum ferritin less than 30 ng/mL and satisfied AIS criteria for iron supplementation. In only one case was a disorder identified which was not expected following the clinical history and examination. Fifty-two female athletes had a serum ferritin less than 30 ng/mL and satisfied AIS criteria for iron supplementation. In the females, there were no instances in which a medical condition was identified which was not expected following the clinical history and examination. In both groups, clinically non-significant abnormalities were generally minor or isolated reductions in haemoglobin and/or haematocrit, and alterations in red cell parameters or single measures of iron status. Conclusion: Screening for haematological and iron-related abnormalities in athletes has a low yield. Due to the critical nature of the effects of anaemia and low serum ferritin on some aspects of performance it is reasonable to perform a haemoglobin and a serum ferritin on male and female athletes entering an elite training program. Isolated abnormalities which are close to the limits of their normal ranges and not accompanied by symptoms or signs of illness can almost certainly be ignored. © 2007 Sports Medicine Australia. Published by Elsevier Ltd. All rights reserved.
Introduction Only two prospective studies related to the clinical utility of screening of haematological and ironrelated variables in elite athlete populations have E-mail address: [email protected].
been published,1,2 with both of these detailing follow-up of screened athletes. One recent publication assessed serum ferritin levels as a marker for iron overload in amateur and professional cyclists and included follow-up for the presence of mutations in the HFE gene for heamochromatosis.3 In view of the large number of screening tests per-
1440-2440/$ — see front matter © 2007 Sports Medicine Australia. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.jsams.2007.02.017
330 formed worldwide in elite athletes, assessment of the results from a large number of elite athletes from various sports is required so that appropriate recommendations for such screening may be made. This large prospective study was performed to significantly add to the database from which such recommendations may be made.
Methods This study was approved by the Ethics Committee of the Australian Institute of Sport. All procedures conformed to the National Health and Medical Research Council guidelines for experimentation with human subjects and all subjects gave their informed written consent before participation. All athletes (303 males and 273 females) presenting at the Sports Science/Sports Medicine Centre at the Australian Institute of Sport for routine screening during a three-year period were entered into the study. They underwent a standard medical examination by specialist sports physicians and at its conclusion blood (12 mL) was drawn under aseptic technique from a vein in the cubital fossa. Blood was drawn immediately following the assumption of the recumbent position and was transported immediately to an adjacent haematology laboratory where analysis was performed on the same day. Due to training demands and, on some occasions, the demand for screening large numbers of athletes in short time periods, the timing of blood collection was not standardised in relation to periods of training and time of day. Therefore, parameters potentially affected by training (WCC, platelets) and diurnal variation (serum iron) may have been perturbed from true baseline levels. The haematology analysers used in this study (Advia 120, Bayer Diagnostics, Tarrytown, USA using Bayer reagents) are capable of measuring a large number of variables in a full blood count but only a clinically relevant selection was included in this study—–white blood cell count (WCC), red blood cell count (RCC), haemoglobin (Hb), haematocrit (HCT), mean cell volume (MCV), mean cell haemoglobin concentration (MCHC) and platelet count (PL). The following iron-related parameters were measured on a Hitachi 911 analyser (Roche Diagnostics, Indianapolis, USA) using Roche reagents—–serum iron, ferritin, transferrin, percent transferrin saturation. Normal ranges were those of the AIS laboratory. A level of less than 30 ng/mL is used at the AIS as an indication for iron supplementation based on sensitivity and specificity data published by Mast et al.,4 common patterns of supplementa-
K.E. Fallon tion elsewhere5 and data from a study in which the relationship between serum ferritin and soluble transferrin receptor in elite athletes indicated that if the serum ferritin was greater than or equal to 22 ng/mL, soluble transferrin receptor was not beyond the upper limit of normal.6 Athletes were divided into three groups dependent upon their results—–all results normal, results abnormal with no clinically obvious explanation and results abnormal with an explanation apparent from history and clinical examination. The results of those in the third group were excluded from the statistical analysis. The significance of abnormal results was assessed by follow-up consultations with the athletes and, in some cases repeat blood tests or further investigations. Due to differing normal ranges for some parameters, results for males and females were analysed separately. Simple descriptive statistics (mean, confidence intervals, standard deviation, median, minimum, maximum and range) were performed on Statistica, version 5, Statsoft, Tulsa.
Results For both males and females, sports represented and numbers in each sport are detailed in Table 1. Statistical analysis of group parameters is detailed in Table 2.
Table 1 sport
Numbers of male and female screenings by
Sport
Male screenings
Female screenings
Rugby league Rowing Soccer Rugby union Basketball Boxing Waterpolo Volleyball Swimming-disabled Cycling Swimming Archery Triathlon Athletics Athletics-disabled Gymnastics Netball Skeleton Ice skating
54 45 38 32 19 19 16 15 13 13 13 8 8 6 3 3 0 0 0
0 18 30 0 20 0 55 2 15 16 10 9 4 4 1 31 37 22 1
Screening for haematological and iron-related abnormalities in elite athletes Table 2
331
Haematological and iron-related screening parameters—–males (n = 244) and females (n = 142)
Parameter
Mean (males)
Confidence interval 95% (males)
Mean (females)
Confidence interval 95% (females)
White cell count (×109 L−1 ) Red cell count (×1012 L−1 ) Haemoglobin (g/dL) Haematocrit MCV (fL) MCHC (g/dL) Platelets (×109 L−1 ) Serum iron (mol/L) Serum ferritin (ng/mL) Transferrin (g/L) Transferrin saturation (%)
7.32 5.09 15.1 0.43 84.3 35.3 277 19.1 84.8 2.69 32
7.11—7.56 5.05—5.13 15.0—15.2 0.43—0.43 83.9—84.7 35.2—35.4 270—284 18.5—19.7 78.5—91.1 2.65—2.73 31—33
7.10 4.55 13.7 0.39 85.8 35.1 295 18.4 62.6 2.80 29
6.79—7.41 4.26—4.84 13.6—13.8 0.39—0.39 82.2—89.4 35.0—35.2 286—304 17.5—19.3 58.0—67.2 2.74—2.86 28—30
Male athletes All 303 male athlete screenings were performed on scholarship holders of the Australian Institute of Sport all of whom participated at the elite level of competition. They ranged in age from 13 to 41 years (mean 18 years). In 132 (43.5%) athletes no abnormality was found. One hundred and two screenings in males (33.7%) revealed minor abnormalities which were either close to the limits of normal and unaccompanied by symptoms or findings on physical examination or thought to be related directly to training. These abnormalities are summarised in Table 3. Sixty-nine screenings in males (22.8%) revealed abnormalities thought to be significant enough to require further investigation.
Red blood cell abnormalities Four athletes had haemoglobin concentrations greater than or equal to 0.5 g/dL below the normal range. Two cases were asymptomatic and had normochromic, normocytic red cells and normal iron studies. On repeat testing one case remained
unchanged and one returned to the normal range. The abnormalities in both cases were almost certainly due to dilutional anaemia related to training. In the third case reduced haemoglobin (12.8 g/dL) and haematocrit (0.35) were associated with mild microcytosis and normal iron studies. Haemoglobin electrophoresis was normal and a repeat blood count revealed persistent mild anaemia (12.5 g/dL) with normochromic, normocytic red cells. In the fourth case reduced haemoglobin (12.9 g/dL) and haematocrit (0.36) was associated with low serum iron and transferrin saturation and a proven urinary tract infection. Following resolution of the infection a mild normochromic, normocytic anaemia persisted (13.2 g/dL) with iron studies consistent with low iron stores (ferritin 11 g/L). One athlete demonstrated persistent microcytic red cells (and microcytic reticulocytes) with normal MCHC in the absence of anaemia and with normal iron studies. Haemoglobin electrophoresis was normal.
White blood cell and platelet abnormalities Twelve athletes had raised white blood cell counts. Nine cases (eight associated with neutrophilia
Table 3 Males—–minor abnormalities which were either close to the limits of normal and unaccompanied by symptoms or findings on physical examination or thought to be related directly to training Parameter
Number
Range
Increased MCHC
28
36.2—39.1 g/dL
Decreased haemoglobin and haematocrit
15
13.3—13.9 g/dL 0.36—0.40
Decreased MCV (absence of anaemia, normal MCHC, normal iron studies) Isolated decreased haematocrit Isolated decreased transferrin Increased white cell count (asymptomatic)
14 10 8 5
73.5—80 fL 0.37—0.40 1.76—2.29 g/L (11.01—13.21) × 109 L−1
332 and one with monocytosis) were associated with symptoms and signs of infection. One case was associated with an acute groin injury and two had no clear explanation. Two athletes had minor increases in the platelet count, both of which returned to the normal range on repeat testing.
Abnormalities in iron-related variables Eleven athletes had isolated increases in serum iron. Three of these athletes were using iron supplements. Three, who were not, had a normal level on repeat testing. For geographical and temporal reasons five were to be reviewed by their local general practitioners. Two athletes had isolated reductions in serum iron, consistent with their symptoms of upper respiratory tract infection. Two athletes demonstrated reductions in both serum iron and transferrin saturation. In one case this was associated with an upper respiratory tract infection but in the other asymptomatic case further investigation revealed no clear cause. One athlete had an isolated increase in serum ferritin (317 ng/mL) explained by the use of iron supplements at the time of screening. Ten athletes returned serum ferritin levels below 30 ng/mL which is our limit for recommendation for supplementation. The range was 11.4—29.5 ng/mL and in two cases both serum iron and transferrin saturation were below normal. There was no obvious clinical explanation for these cases and each athlete was referred for dietary analysis and iron supplementation. Twenty-four athletes (7.9% of the whole male sample) had increased transferrin saturation. Five of these elevations were associated with increased serum iron and were explained by the use of iron supplements. Twelve cases had normal iron studies when the test was repeated. Three cases remained elevated on repeat testing and underwent genetic testing for haemochromatosis. One athlete was homozygous for the C282Y mutation, one homozygous for the H63D mutation and one heterozygous for the C282Y mutation. Two cases had histories of having very varying iron studies over time. One had previously had negative tests for HFE mutations and one remained under review by his general practitioner. The two other cases were referred for repeat tests in their home states with a request that abnormal tests be reported to the researcher. It is presumed that these tests were normal.
Female athletes All 273 screenings in female athletes were performed on scholarship holders of the Australian
K.E. Fallon Institute of Sport who participated at the elite level of competition. They ranged in age from 9 to 34 years (mean 20 years). Ninety-four (34.4%) had no abnormality found. Forty-six female screenings (16.8%) revealed minor abnormalities which were close to the limits of normal and unaccompanied by symptoms or findings on physical examination or thought to be related directly to training. These abnormalities are summarised in Table 4. One hundred and thirty three female screenings (48.7%) revealed abnormalities thought to be significant enough to require further investigation.
Red blood cell abnormalities One athlete had a haemoglobin level greater than or equal to 0.5 g/dL below the normal range (11.4 g/dL). This was associated with a reduced haematocrit (0.32) and MCV. Iron studies were normal. This athlete had had endoscopy for possible coeliac disease which was normal. No clear cause was found. One athlete had a low haematocrit (0.32) with haemoglobin close to normal (11.7 g/dL) both of which normalised on a repeat sample. Red cell parameters and iron studies were normal.
White blood cell and platelet abnormalities Fifteen athletes had elevated white cell counts. Ten cases were related to infections. Two had no clear explanation but may have been related to recent training and three cases tested on the same day may be explained by a laboratory error. One athlete had a reduced WCC and reduced serum iron and transferrin saturation. Although associated with no symptoms the presence of atypical lymphocytes suggested the presence of a subclinical infection. Fifteen athletes had increased platelet counts but only three of these were above 450 × 109 L−1 . All returned to normal on repeat testing.
Abnormalities in iron-related variables Six athletes, all of whom had symptoms and signs of infection, demonstrated serum iron and transferrin saturation below normal limits. Two athletes had low transferrin saturation. One returned to within normal limits on repeat testing and the other also had an increased WCC and signs of rhinitis suggesting the presence of a negative acute phase reactant response. Four athletes had isolated raised serum ferritin. One returned to normal on repeat testing,
Screening for haematological and iron-related abnormalities in elite athletes
333
Table 4 Females—–minor abnormalities which were either close to the limits of normal and unaccompanied by symptoms or findings on physical examination or thought to be related directly to training Parameter
Number
Range
Increased MCHC Decreased haematocrit Decreased MCV (absence of anaemia, normal MCHC, normal iron studies) Increased white cell count (asymptomatic)
11 11 9
36—36.7 g/dL 0.33—0.36 77—80 fL
one was negative for mutation of the HFE gene, one was heterozygous for the H63D mutation and one who had multiple previous and repeat levels around the upper limit of normal, associated with normal other iron-related variables, was to be followed up by her general practitioner at 12 months. Twenty-six athletes had increased serum iron, 25 of these being isolated findings. Eight were using iron supplements. Three had repeat testing, all of these returning to normal. The others were asked to have a repeat test in 12 months. One athlete had increased serum iron, ferritin and transferrin saturation but was negative on genetic testing for haemochromatosis. Thirteen athletes had increased transferrin saturation combined with increased serum iron. Seven were using iron supplements. Two returned to normal on repeat testing. Two were negative on genetic testing for haemochromatosis. One athlete had been previously investigated and it was felt that high values were normal for her and one athlete was referred for follow-up by her general practitioner but no follow-up results were forwarded to the researcher. Fifty-two athletes (19% of all screenings) had a serum ferritin less than 30 ng/mL. In 12 athletes serum ferritin was below 20 ng/mL. There were no cases of anaemia and no athlete in the low ferritin group had microcytic, hypochromic red cells. Only one athlete had a pathological explanation—–coeliac disease.
Discussion—–male athletes Sixty-nine male subjects were investigated for significant abnormalities. A satisfactory explanation was found in 56 cases. One case of microcytosis and one of an increase in serum iron accompanied by increased transferrin satu-
4
(11.01—12.54) × 109 L−1
ration could not be explained. It is likely that two cases of increased white cell count and two cases of increased platelets were related to training. Five cases of isolated increased serum iron and two cases of isolated increased transferrin saturation were referred, due to time constraints, to the athletes’ local general practitioner for follow-up, with a request to notify the researcher of any abnormal findings. None were notified. Presuming that those followed-up by their general practitioners are normal, screening detected a clinically significant problem in only one male athlete. He was diagnosed with haemochromatosis. Ten athletes had serum ferritin levels less than 30 ng/mL which led to iron supplementation. One of these athletes had a serum ferritin level less than 20 ng/mL which may have had an impact on performance.
Discussion—–female athletes One hundred and thirty three female subjects were investigated for potentially significant abnormalities. A satisfactory explanation was found in 111 cases. One case of an isolated finding of a haemoglobin level greater than 0.5 g/dL below the normal range was presumed to represent dilutional anaemia. Two cases of isolated increased white cell count were probably related to recent training and three were probably related to a laboratory error. Due to athletes being based in other states, 14 cases of an isolated increase in serum iron, 1 of an isolated increase in serum ferritin and 1 with increases in serum iron and transferrin saturation were referred to the athletes’ local general practitioner for follow-up, with a request to notify the researcher of any abnormal findings. None were notified. Presuming that those followed-up by their general practitioners are normal, screening did not detect a clinically significant problem in any female athlete. Fifty-two athletes had serum ferritin levels
334 less than 30 ng/mL which led to iron supplementation. Twelve of these athletes had serum ferritin levels less than 20 ng/mL which may have had an impact on performance.
General comments The results obtained in both males and females in this group of athletes are remarkably similar to those obtained in an earlier study of the screening of 294 elite Australian athletes.2 In aggregate these two studies provide us with data from 870 elite Australian athletes. The only major difference between these studies and data from the only other comprehensive screening study yet published, that of the Israeli Olympic team, was that that study identified four cases of iron deficiency anaemia and two cases of anaemia related to vitamin B12 deficiency in a total of 114 athletes. No cases of such anaemia were found during the current study. The mean, combining the results of 75 males and 39 females, for serum ferritin in the Israeli study, 66.1 ng/mL, was similar to that in this study (males, 84.8 ng/mL; females, 62.6 ng/mL) and in the previous Australian study (males, 84.7 ng/mL; females, 54.4 ng/mL).2 In the present study, when athletes with low serum ferritin levels were included in the analysis the mean values were, for males, 82.3 ng/mL and females, 52.0 ng/mL. In a significant proportion of the athletes screened the abnormalities found were isolated, close to the limits of their normal ranges and not accompanied by symptoms or signs of illness. In general they appeared to be of no clinical significance suggesting that these ‘‘abnormal’’ results are consequent upon the usual method of definition of the normal range. A number of minor abnormalities which have been found in both this and the previous Australian screening study warrant discussion. These include increased MCHC, reduced haemoglobin with or without reduced haematocrit, reduced MCV, reduced transferrin and increased white blood cell count. These were generally isolated findings and none deviated greatly from the normal range. Mean corpuscular haemoglobin concentration is defined as the average concentration of haemoglobin in grams in a decilitre of erythrocytes. It is calculated by dividing the haemoglobin concentration by the haematocrit. The spherocyte is the only erythrocyte which demonstrates hyperchromia (increased MCHC). Spherocytes are present in hereditary spherocytosis, autoimmune haemolytic anaemias, Heinz-body anaemias and in severe burns. Based on the lack of other haemato-
K.E. Fallon logical changes in athletes with an increase in MCHC (reticulocytosis, increased percentage of hyperchromic red cells, reduced mean red cell volume and increased haemoglobin distribution width) it is most unlikely that even the most common of these disorders, hereditary spherocytosis, was present. Based on the findings even a recessive form of this disease is unlikely.7 It is not known if high levels of training have a specific effect on MCHC. Microcytosis (reduced mean red cell volume) in the presence of anaemia is indicative of iron deficiency or thalassaemia but may also be seen in the anaemia of chronic disease. In this group of athletes the isolated nature of this finding (including absence of anaemia and normal iron studies) and the absence of clinical features of relevant illnesses suggest that this finding is of little significance. Again, it is not known if high levels of training have a specific effect on MCV. Reductions in haemoglobin, haematocit or both, in association with normochromic, normocytic red cells and normal iron studies, in athletes, is generally thought to be related to the dilutional anaemia consequent upon training.8 High levels of serum iron can occur as the result of multiple blood transfusions, iron injections and oral supplements, lead poisoning, liver and kidney disease and haemochromatosis. Each of these diagnoses is usually associated with significant findings on clinical history and examination or associated changes in other blood parameters. These diagnoses are highly unlikely in elite athletes with normal histories and examinations and where the increase in serum iron as an isolated finding. It has, however, been shown that young women who are heterozygous for the C282Y mutation for haemochromatosis have significantly higher serum iron, transferrin saturation and haemoglobin when compared with those without such mutations. It has been suggested that this could have a protective effect against iron deficiency.9 Despite the uncertainty of the clinical significance of an isolated increased serum iron level, it may be of interest to assess athletes with minor increases in these parameters for the presence of this mutation. Serum transferrin is a negative acute phase reactant and can be reduced in acute or chronic inflammatory states and infections. It may also be reduced in nephrotic syndrome and thalassaemia. The lack of symptoms and signs of illness and the isolated nature of this finding make the presence of any of these diagnoses unlikely in the subjects of this study. Increases in white blood cell count, in the absence of symptoms and signs of infection or inflammatory disease have been described in a
Screening for haematological and iron-related abnormalities in elite athletes number of studies involving exercise. Significant increases in neutrophils have been demonstrated after relatively mild exercise such as a walk of 30 min duration10 and elevations in total white cell count and neutrophils can persist for at least 6 h following even such short duration exercise.11 Increases in white cell count in otherwise well athletes are therefore likely to be related to training and are very unlikely to represent a pathological process. The true baseline WCC could be assessed after a period of 12 h of rest but this may be difficult in athletes who often train twice a day. Haematological and iron related screening of this group of athletes may underestimate the number of abnormalities that could be found in other athletic groups. Because of the elite nature of the athletes involved, one could expect that they would have been tested previously and that abnormalities identified had been corrected. Screening athletes for haematological and ironrelated disorders is not universal. In a survey of US National Collegiate Athletic Association Division 1-A schools only 43% of schools responding to a survey reported screening female athletes for iron deficiency.12 However, a full blood count and serum ferritin has been previously recommended for athletes of both sexes entering into an elite training program and this suggestion is supported by the somewhat surprising finding of five cases of anaemia in 114 Israeli Olympic athletes when they were screened with a full blood count, iron studies and a biochemical profile. This study supports a recommendation for screening for iron deficiency for athletes involved in an elite training program. In view of the lack of significance of many of the abnormalities found in this study it would be reasonable to assess only haemoglobin and ferritin. For better or worse, however, automated blood analysers produce results in which many other parameters are reported and appreciation of minor abnormalities so demonstrated could lead to over investigation. The issue of general population screening for haemochromatosis remains unresolved13 but the detection of the suggestion of the presence of this disease by assessment of iron-related parameters may be a welcome advantage of screening athletes.
Conclusion Due to the critical nature of the effects of anaemia and low iron stores on some aspects of performance it is reasonable to perform a haemoglobin level and
335
a serum ferritin on both male and female athletes entering an elite training program. Isolated abnormalities which are close to the limits of their normal ranges and not accompanied by symptoms or signs of illness can almost certainly be ignored. This applies particularly to increased MCHC, reduced haemoglobin with or without reduced haematocrit, reduced MCV, reduced transferrin and increased white blood cell count. It may be, however, reassuring to repeat the test several weeks after the initial test as many of the initially abnormal results will be normal on repeat testing.
Practical implications • It is reasonable to perform a haemoglobin level and a serum ferritin on both male and female athletes entering an elite training program. • Isolated abnormalities in the parameters assessed in this study which are close to the limits of their normal ranges, and not accompanied by symptoms or signs of illness, can almost certainly be ignored. • This applies particularly to increased mean cell haemoglobin concentration, reduced haemoglobin with or without reduced haematocrit, reduced mean cell volume, reduced transferrin and increased white blood cell count.
Disclosure Financial support for this research was provided from the Departmental Discretionary Research Budget of the Australian Institute of Sport.
References 1. Eliakim A, Nemet D, Constantini N. Screening blood test in members of the Isreali National Olympic team. J Sports Med Phys Fitness 2002;42:250—5. 2. Fallon KE. Utility of haematological and iron-related screening in elite athletes. Clin J Sport Med 2004;14: 145—52. 3. Lippi G, Schena F, Franchini M, Salvagno GL, Guidi GC. Serum ferritin as a marker of potential biochemical iron overload in athletes. Clin J Sport Med 2005;15:356—8. 4. Mast AE, Blinder MA, Gronowski AM, Chumley C, Scott MG. Clinical utility of the soluble transferrin receptor and comparison with serum ferritin in several populations. Clin Chem 1998;44:45—51. 5. Neilsen P, Nachtigall D. Iron supplementation in athletes. Current recommendations. Sports Med 1998;26:207—16.
336
K.E. Fallon
6. Pitsis GC, Fallon KE, Fallon SK, Fazakerly R. Response of soluble transferrin receptor and iro-related parameters to iron supplementation in elite, iron-depleted, nonanaemic female athletes. Clin J Sport Med 2004;14:300—4. 7. Conway AM, Vora AJ, Hinchliffe RF. The clinical relevance of an isolated increase in the number of circulating hyperchromic red blood cells. J Clin Path 2002;55:841—4. 8. Shaskey DJ, Green GA. Sports haematology. Sports Med 2000;29:27—38. 9. Datz C, Haas T, Rinner H, Sandhofer F, Patsch W, Paulweber B. Heterozygosity for the C282y mutation in the haemochromatosis gene is associated with increased serum iron, transferrin saturation and haemoglobin in young women: a protective role against iron deficiency? Clin Chem 1998;44:2429—32.
10. Nieman DC, Henson DA, Austin MD, Brown VA. Immune response to a 30 min walk. Med Sci Sports Exerc 2005;37:57—62. 11. Mc Carthy DA, Grant M, Marbut M, Watling M, Wade AJ, Macdonald I, et al. Brief exercise indices an immediate and a delayed leucocytosis. Br J Sports Med 1991;24:191— 5. 12. Cowell BS, Rosenbloom CA, Skinner R, Summers SH. Policies on screening female athletes for iron deficiency in NCAA division 1-A institutions. Int J Sport Nut Excer Metab 2003;13:277—85. 13. Qaseem A, Aronson M, Fitterman N, Snow V, Weiss KB, Owens DK. Screening for hereditary haemochromatosis: a clinical practice guideline from the American College of Physicians. Ann Intern Med 2005;143:517—21.
Available online at www.sciencedirect.com
ORIGINAL PAPER
Screening for haematological and iron-related abnormalities in elite athletes—–Analysis of 576 cases Kieran E. Fallon Department of Sports Medicine, Australian Institute of Sport, Australia Received 17 November 2006 ; received in revised form 12 February 2007; accepted 18 February 2007 KEYWORDS Screening; Iron; Haematology; Elite; Athlete
Summary The aim of this prospective, cohort study was to determine the clinical and performance related utility of haematological and iron-related screening in elite athletes. Three hundred and three male and 273 female elite athletes underwent routine medical screening over a three-year period. In association with a standard medical consultation, a full blood count and iron-related variables were measured. Ten male athletes had a serum ferritin less than 30 ng/mL and satisfied AIS criteria for iron supplementation. In only one case was a disorder identified which was not expected following the clinical history and examination. Fifty-two female athletes had a serum ferritin less than 30 ng/mL and satisfied AIS criteria for iron supplementation. In the females, there were no instances in which a medical condition was identified which was not expected following the clinical history and examination. In both groups, clinically non-significant abnormalities were generally minor or isolated reductions in haemoglobin and/or haematocrit, and alterations in red cell parameters or single measures of iron status. Conclusion: Screening for haematological and iron-related abnormalities in athletes has a low yield. Due to the critical nature of the effects of anaemia and low serum ferritin on some aspects of performance it is reasonable to perform a haemoglobin and a serum ferritin on male and female athletes entering an elite training program. Isolated abnormalities which are close to the limits of their normal ranges and not accompanied by symptoms or signs of illness can almost certainly be ignored. © 2007 Sports Medicine Australia. Published by Elsevier Ltd. All rights reserved.
Introduction Only two prospective studies related to the clinical utility of screening of haematological and ironrelated variables in elite athlete populations have E-mail address: [email protected].
been published,1,2 with both of these detailing follow-up of screened athletes. One recent publication assessed serum ferritin levels as a marker for iron overload in amateur and professional cyclists and included follow-up for the presence of mutations in the HFE gene for heamochromatosis.3 In view of the large number of screening tests per-
1440-2440/$ — see front matter © 2007 Sports Medicine Australia. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.jsams.2007.02.017
330 formed worldwide in elite athletes, assessment of the results from a large number of elite athletes from various sports is required so that appropriate recommendations for such screening may be made. This large prospective study was performed to significantly add to the database from which such recommendations may be made.
Methods This study was approved by the Ethics Committee of the Australian Institute of Sport. All procedures conformed to the National Health and Medical Research Council guidelines for experimentation with human subjects and all subjects gave their informed written consent before participation. All athletes (303 males and 273 females) presenting at the Sports Science/Sports Medicine Centre at the Australian Institute of Sport for routine screening during a three-year period were entered into the study. They underwent a standard medical examination by specialist sports physicians and at its conclusion blood (12 mL) was drawn under aseptic technique from a vein in the cubital fossa. Blood was drawn immediately following the assumption of the recumbent position and was transported immediately to an adjacent haematology laboratory where analysis was performed on the same day. Due to training demands and, on some occasions, the demand for screening large numbers of athletes in short time periods, the timing of blood collection was not standardised in relation to periods of training and time of day. Therefore, parameters potentially affected by training (WCC, platelets) and diurnal variation (serum iron) may have been perturbed from true baseline levels. The haematology analysers used in this study (Advia 120, Bayer Diagnostics, Tarrytown, USA using Bayer reagents) are capable of measuring a large number of variables in a full blood count but only a clinically relevant selection was included in this study—–white blood cell count (WCC), red blood cell count (RCC), haemoglobin (Hb), haematocrit (HCT), mean cell volume (MCV), mean cell haemoglobin concentration (MCHC) and platelet count (PL). The following iron-related parameters were measured on a Hitachi 911 analyser (Roche Diagnostics, Indianapolis, USA) using Roche reagents—–serum iron, ferritin, transferrin, percent transferrin saturation. Normal ranges were those of the AIS laboratory. A level of less than 30 ng/mL is used at the AIS as an indication for iron supplementation based on sensitivity and specificity data published by Mast et al.,4 common patterns of supplementa-
K.E. Fallon tion elsewhere5 and data from a study in which the relationship between serum ferritin and soluble transferrin receptor in elite athletes indicated that if the serum ferritin was greater than or equal to 22 ng/mL, soluble transferrin receptor was not beyond the upper limit of normal.6 Athletes were divided into three groups dependent upon their results—–all results normal, results abnormal with no clinically obvious explanation and results abnormal with an explanation apparent from history and clinical examination. The results of those in the third group were excluded from the statistical analysis. The significance of abnormal results was assessed by follow-up consultations with the athletes and, in some cases repeat blood tests or further investigations. Due to differing normal ranges for some parameters, results for males and females were analysed separately. Simple descriptive statistics (mean, confidence intervals, standard deviation, median, minimum, maximum and range) were performed on Statistica, version 5, Statsoft, Tulsa.
Results For both males and females, sports represented and numbers in each sport are detailed in Table 1. Statistical analysis of group parameters is detailed in Table 2.
Table 1 sport
Numbers of male and female screenings by
Sport
Male screenings
Female screenings
Rugby league Rowing Soccer Rugby union Basketball Boxing Waterpolo Volleyball Swimming-disabled Cycling Swimming Archery Triathlon Athletics Athletics-disabled Gymnastics Netball Skeleton Ice skating
54 45 38 32 19 19 16 15 13 13 13 8 8 6 3 3 0 0 0
0 18 30 0 20 0 55 2 15 16 10 9 4 4 1 31 37 22 1
Screening for haematological and iron-related abnormalities in elite athletes Table 2
331
Haematological and iron-related screening parameters—–males (n = 244) and females (n = 142)
Parameter
Mean (males)
Confidence interval 95% (males)
Mean (females)
Confidence interval 95% (females)
White cell count (×109 L−1 ) Red cell count (×1012 L−1 ) Haemoglobin (g/dL) Haematocrit MCV (fL) MCHC (g/dL) Platelets (×109 L−1 ) Serum iron (mol/L) Serum ferritin (ng/mL) Transferrin (g/L) Transferrin saturation (%)
7.32 5.09 15.1 0.43 84.3 35.3 277 19.1 84.8 2.69 32
7.11—7.56 5.05—5.13 15.0—15.2 0.43—0.43 83.9—84.7 35.2—35.4 270—284 18.5—19.7 78.5—91.1 2.65—2.73 31—33
7.10 4.55 13.7 0.39 85.8 35.1 295 18.4 62.6 2.80 29
6.79—7.41 4.26—4.84 13.6—13.8 0.39—0.39 82.2—89.4 35.0—35.2 286—304 17.5—19.3 58.0—67.2 2.74—2.86 28—30
Male athletes All 303 male athlete screenings were performed on scholarship holders of the Australian Institute of Sport all of whom participated at the elite level of competition. They ranged in age from 13 to 41 years (mean 18 years). In 132 (43.5%) athletes no abnormality was found. One hundred and two screenings in males (33.7%) revealed minor abnormalities which were either close to the limits of normal and unaccompanied by symptoms or findings on physical examination or thought to be related directly to training. These abnormalities are summarised in Table 3. Sixty-nine screenings in males (22.8%) revealed abnormalities thought to be significant enough to require further investigation.
Red blood cell abnormalities Four athletes had haemoglobin concentrations greater than or equal to 0.5 g/dL below the normal range. Two cases were asymptomatic and had normochromic, normocytic red cells and normal iron studies. On repeat testing one case remained
unchanged and one returned to the normal range. The abnormalities in both cases were almost certainly due to dilutional anaemia related to training. In the third case reduced haemoglobin (12.8 g/dL) and haematocrit (0.35) were associated with mild microcytosis and normal iron studies. Haemoglobin electrophoresis was normal and a repeat blood count revealed persistent mild anaemia (12.5 g/dL) with normochromic, normocytic red cells. In the fourth case reduced haemoglobin (12.9 g/dL) and haematocrit (0.36) was associated with low serum iron and transferrin saturation and a proven urinary tract infection. Following resolution of the infection a mild normochromic, normocytic anaemia persisted (13.2 g/dL) with iron studies consistent with low iron stores (ferritin 11 g/L). One athlete demonstrated persistent microcytic red cells (and microcytic reticulocytes) with normal MCHC in the absence of anaemia and with normal iron studies. Haemoglobin electrophoresis was normal.
White blood cell and platelet abnormalities Twelve athletes had raised white blood cell counts. Nine cases (eight associated with neutrophilia
Table 3 Males—–minor abnormalities which were either close to the limits of normal and unaccompanied by symptoms or findings on physical examination or thought to be related directly to training Parameter
Number
Range
Increased MCHC
28
36.2—39.1 g/dL
Decreased haemoglobin and haematocrit
15
13.3—13.9 g/dL 0.36—0.40
Decreased MCV (absence of anaemia, normal MCHC, normal iron studies) Isolated decreased haematocrit Isolated decreased transferrin Increased white cell count (asymptomatic)
14 10 8 5
73.5—80 fL 0.37—0.40 1.76—2.29 g/L (11.01—13.21) × 109 L−1
332 and one with monocytosis) were associated with symptoms and signs of infection. One case was associated with an acute groin injury and two had no clear explanation. Two athletes had minor increases in the platelet count, both of which returned to the normal range on repeat testing.
Abnormalities in iron-related variables Eleven athletes had isolated increases in serum iron. Three of these athletes were using iron supplements. Three, who were not, had a normal level on repeat testing. For geographical and temporal reasons five were to be reviewed by their local general practitioners. Two athletes had isolated reductions in serum iron, consistent with their symptoms of upper respiratory tract infection. Two athletes demonstrated reductions in both serum iron and transferrin saturation. In one case this was associated with an upper respiratory tract infection but in the other asymptomatic case further investigation revealed no clear cause. One athlete had an isolated increase in serum ferritin (317 ng/mL) explained by the use of iron supplements at the time of screening. Ten athletes returned serum ferritin levels below 30 ng/mL which is our limit for recommendation for supplementation. The range was 11.4—29.5 ng/mL and in two cases both serum iron and transferrin saturation were below normal. There was no obvious clinical explanation for these cases and each athlete was referred for dietary analysis and iron supplementation. Twenty-four athletes (7.9% of the whole male sample) had increased transferrin saturation. Five of these elevations were associated with increased serum iron and were explained by the use of iron supplements. Twelve cases had normal iron studies when the test was repeated. Three cases remained elevated on repeat testing and underwent genetic testing for haemochromatosis. One athlete was homozygous for the C282Y mutation, one homozygous for the H63D mutation and one heterozygous for the C282Y mutation. Two cases had histories of having very varying iron studies over time. One had previously had negative tests for HFE mutations and one remained under review by his general practitioner. The two other cases were referred for repeat tests in their home states with a request that abnormal tests be reported to the researcher. It is presumed that these tests were normal.
Female athletes All 273 screenings in female athletes were performed on scholarship holders of the Australian
K.E. Fallon Institute of Sport who participated at the elite level of competition. They ranged in age from 9 to 34 years (mean 20 years). Ninety-four (34.4%) had no abnormality found. Forty-six female screenings (16.8%) revealed minor abnormalities which were close to the limits of normal and unaccompanied by symptoms or findings on physical examination or thought to be related directly to training. These abnormalities are summarised in Table 4. One hundred and thirty three female screenings (48.7%) revealed abnormalities thought to be significant enough to require further investigation.
Red blood cell abnormalities One athlete had a haemoglobin level greater than or equal to 0.5 g/dL below the normal range (11.4 g/dL). This was associated with a reduced haematocrit (0.32) and MCV. Iron studies were normal. This athlete had had endoscopy for possible coeliac disease which was normal. No clear cause was found. One athlete had a low haematocrit (0.32) with haemoglobin close to normal (11.7 g/dL) both of which normalised on a repeat sample. Red cell parameters and iron studies were normal.
White blood cell and platelet abnormalities Fifteen athletes had elevated white cell counts. Ten cases were related to infections. Two had no clear explanation but may have been related to recent training and three cases tested on the same day may be explained by a laboratory error. One athlete had a reduced WCC and reduced serum iron and transferrin saturation. Although associated with no symptoms the presence of atypical lymphocytes suggested the presence of a subclinical infection. Fifteen athletes had increased platelet counts but only three of these were above 450 × 109 L−1 . All returned to normal on repeat testing.
Abnormalities in iron-related variables Six athletes, all of whom had symptoms and signs of infection, demonstrated serum iron and transferrin saturation below normal limits. Two athletes had low transferrin saturation. One returned to within normal limits on repeat testing and the other also had an increased WCC and signs of rhinitis suggesting the presence of a negative acute phase reactant response. Four athletes had isolated raised serum ferritin. One returned to normal on repeat testing,
Screening for haematological and iron-related abnormalities in elite athletes
333
Table 4 Females—–minor abnormalities which were either close to the limits of normal and unaccompanied by symptoms or findings on physical examination or thought to be related directly to training Parameter
Number
Range
Increased MCHC Decreased haematocrit Decreased MCV (absence of anaemia, normal MCHC, normal iron studies) Increased white cell count (asymptomatic)
11 11 9
36—36.7 g/dL 0.33—0.36 77—80 fL
one was negative for mutation of the HFE gene, one was heterozygous for the H63D mutation and one who had multiple previous and repeat levels around the upper limit of normal, associated with normal other iron-related variables, was to be followed up by her general practitioner at 12 months. Twenty-six athletes had increased serum iron, 25 of these being isolated findings. Eight were using iron supplements. Three had repeat testing, all of these returning to normal. The others were asked to have a repeat test in 12 months. One athlete had increased serum iron, ferritin and transferrin saturation but was negative on genetic testing for haemochromatosis. Thirteen athletes had increased transferrin saturation combined with increased serum iron. Seven were using iron supplements. Two returned to normal on repeat testing. Two were negative on genetic testing for haemochromatosis. One athlete had been previously investigated and it was felt that high values were normal for her and one athlete was referred for follow-up by her general practitioner but no follow-up results were forwarded to the researcher. Fifty-two athletes (19% of all screenings) had a serum ferritin less than 30 ng/mL. In 12 athletes serum ferritin was below 20 ng/mL. There were no cases of anaemia and no athlete in the low ferritin group had microcytic, hypochromic red cells. Only one athlete had a pathological explanation—–coeliac disease.
Discussion—–male athletes Sixty-nine male subjects were investigated for significant abnormalities. A satisfactory explanation was found in 56 cases. One case of microcytosis and one of an increase in serum iron accompanied by increased transferrin satu-
4
(11.01—12.54) × 109 L−1
ration could not be explained. It is likely that two cases of increased white cell count and two cases of increased platelets were related to training. Five cases of isolated increased serum iron and two cases of isolated increased transferrin saturation were referred, due to time constraints, to the athletes’ local general practitioner for follow-up, with a request to notify the researcher of any abnormal findings. None were notified. Presuming that those followed-up by their general practitioners are normal, screening detected a clinically significant problem in only one male athlete. He was diagnosed with haemochromatosis. Ten athletes had serum ferritin levels less than 30 ng/mL which led to iron supplementation. One of these athletes had a serum ferritin level less than 20 ng/mL which may have had an impact on performance.
Discussion—–female athletes One hundred and thirty three female subjects were investigated for potentially significant abnormalities. A satisfactory explanation was found in 111 cases. One case of an isolated finding of a haemoglobin level greater than 0.5 g/dL below the normal range was presumed to represent dilutional anaemia. Two cases of isolated increased white cell count were probably related to recent training and three were probably related to a laboratory error. Due to athletes being based in other states, 14 cases of an isolated increase in serum iron, 1 of an isolated increase in serum ferritin and 1 with increases in serum iron and transferrin saturation were referred to the athletes’ local general practitioner for follow-up, with a request to notify the researcher of any abnormal findings. None were notified. Presuming that those followed-up by their general practitioners are normal, screening did not detect a clinically significant problem in any female athlete. Fifty-two athletes had serum ferritin levels
334 less than 30 ng/mL which led to iron supplementation. Twelve of these athletes had serum ferritin levels less than 20 ng/mL which may have had an impact on performance.
General comments The results obtained in both males and females in this group of athletes are remarkably similar to those obtained in an earlier study of the screening of 294 elite Australian athletes.2 In aggregate these two studies provide us with data from 870 elite Australian athletes. The only major difference between these studies and data from the only other comprehensive screening study yet published, that of the Israeli Olympic team, was that that study identified four cases of iron deficiency anaemia and two cases of anaemia related to vitamin B12 deficiency in a total of 114 athletes. No cases of such anaemia were found during the current study. The mean, combining the results of 75 males and 39 females, for serum ferritin in the Israeli study, 66.1 ng/mL, was similar to that in this study (males, 84.8 ng/mL; females, 62.6 ng/mL) and in the previous Australian study (males, 84.7 ng/mL; females, 54.4 ng/mL).2 In the present study, when athletes with low serum ferritin levels were included in the analysis the mean values were, for males, 82.3 ng/mL and females, 52.0 ng/mL. In a significant proportion of the athletes screened the abnormalities found were isolated, close to the limits of their normal ranges and not accompanied by symptoms or signs of illness. In general they appeared to be of no clinical significance suggesting that these ‘‘abnormal’’ results are consequent upon the usual method of definition of the normal range. A number of minor abnormalities which have been found in both this and the previous Australian screening study warrant discussion. These include increased MCHC, reduced haemoglobin with or without reduced haematocrit, reduced MCV, reduced transferrin and increased white blood cell count. These were generally isolated findings and none deviated greatly from the normal range. Mean corpuscular haemoglobin concentration is defined as the average concentration of haemoglobin in grams in a decilitre of erythrocytes. It is calculated by dividing the haemoglobin concentration by the haematocrit. The spherocyte is the only erythrocyte which demonstrates hyperchromia (increased MCHC). Spherocytes are present in hereditary spherocytosis, autoimmune haemolytic anaemias, Heinz-body anaemias and in severe burns. Based on the lack of other haemato-
K.E. Fallon logical changes in athletes with an increase in MCHC (reticulocytosis, increased percentage of hyperchromic red cells, reduced mean red cell volume and increased haemoglobin distribution width) it is most unlikely that even the most common of these disorders, hereditary spherocytosis, was present. Based on the findings even a recessive form of this disease is unlikely.7 It is not known if high levels of training have a specific effect on MCHC. Microcytosis (reduced mean red cell volume) in the presence of anaemia is indicative of iron deficiency or thalassaemia but may also be seen in the anaemia of chronic disease. In this group of athletes the isolated nature of this finding (including absence of anaemia and normal iron studies) and the absence of clinical features of relevant illnesses suggest that this finding is of little significance. Again, it is not known if high levels of training have a specific effect on MCV. Reductions in haemoglobin, haematocit or both, in association with normochromic, normocytic red cells and normal iron studies, in athletes, is generally thought to be related to the dilutional anaemia consequent upon training.8 High levels of serum iron can occur as the result of multiple blood transfusions, iron injections and oral supplements, lead poisoning, liver and kidney disease and haemochromatosis. Each of these diagnoses is usually associated with significant findings on clinical history and examination or associated changes in other blood parameters. These diagnoses are highly unlikely in elite athletes with normal histories and examinations and where the increase in serum iron as an isolated finding. It has, however, been shown that young women who are heterozygous for the C282Y mutation for haemochromatosis have significantly higher serum iron, transferrin saturation and haemoglobin when compared with those without such mutations. It has been suggested that this could have a protective effect against iron deficiency.9 Despite the uncertainty of the clinical significance of an isolated increased serum iron level, it may be of interest to assess athletes with minor increases in these parameters for the presence of this mutation. Serum transferrin is a negative acute phase reactant and can be reduced in acute or chronic inflammatory states and infections. It may also be reduced in nephrotic syndrome and thalassaemia. The lack of symptoms and signs of illness and the isolated nature of this finding make the presence of any of these diagnoses unlikely in the subjects of this study. Increases in white blood cell count, in the absence of symptoms and signs of infection or inflammatory disease have been described in a
Screening for haematological and iron-related abnormalities in elite athletes number of studies involving exercise. Significant increases in neutrophils have been demonstrated after relatively mild exercise such as a walk of 30 min duration10 and elevations in total white cell count and neutrophils can persist for at least 6 h following even such short duration exercise.11 Increases in white cell count in otherwise well athletes are therefore likely to be related to training and are very unlikely to represent a pathological process. The true baseline WCC could be assessed after a period of 12 h of rest but this may be difficult in athletes who often train twice a day. Haematological and iron related screening of this group of athletes may underestimate the number of abnormalities that could be found in other athletic groups. Because of the elite nature of the athletes involved, one could expect that they would have been tested previously and that abnormalities identified had been corrected. Screening athletes for haematological and ironrelated disorders is not universal. In a survey of US National Collegiate Athletic Association Division 1-A schools only 43% of schools responding to a survey reported screening female athletes for iron deficiency.12 However, a full blood count and serum ferritin has been previously recommended for athletes of both sexes entering into an elite training program and this suggestion is supported by the somewhat surprising finding of five cases of anaemia in 114 Israeli Olympic athletes when they were screened with a full blood count, iron studies and a biochemical profile. This study supports a recommendation for screening for iron deficiency for athletes involved in an elite training program. In view of the lack of significance of many of the abnormalities found in this study it would be reasonable to assess only haemoglobin and ferritin. For better or worse, however, automated blood analysers produce results in which many other parameters are reported and appreciation of minor abnormalities so demonstrated could lead to over investigation. The issue of general population screening for haemochromatosis remains unresolved13 but the detection of the suggestion of the presence of this disease by assessment of iron-related parameters may be a welcome advantage of screening athletes.
Conclusion Due to the critical nature of the effects of anaemia and low iron stores on some aspects of performance it is reasonable to perform a haemoglobin level and
335
a serum ferritin on both male and female athletes entering an elite training program. Isolated abnormalities which are close to the limits of their normal ranges and not accompanied by symptoms or signs of illness can almost certainly be ignored. This applies particularly to increased MCHC, reduced haemoglobin with or without reduced haematocrit, reduced MCV, reduced transferrin and increased white blood cell count. It may be, however, reassuring to repeat the test several weeks after the initial test as many of the initially abnormal results will be normal on repeat testing.
Practical implications • It is reasonable to perform a haemoglobin level and a serum ferritin on both male and female athletes entering an elite training program. • Isolated abnormalities in the parameters assessed in this study which are close to the limits of their normal ranges, and not accompanied by symptoms or signs of illness, can almost certainly be ignored. • This applies particularly to increased mean cell haemoglobin concentration, reduced haemoglobin with or without reduced haematocrit, reduced mean cell volume, reduced transferrin and increased white blood cell count.
Disclosure Financial support for this research was provided from the Departmental Discretionary Research Budget of the Australian Institute of Sport.
References 1. Eliakim A, Nemet D, Constantini N. Screening blood test in members of the Isreali National Olympic team. J Sports Med Phys Fitness 2002;42:250—5. 2. Fallon KE. Utility of haematological and iron-related screening in elite athletes. Clin J Sport Med 2004;14: 145—52. 3. Lippi G, Schena F, Franchini M, Salvagno GL, Guidi GC. Serum ferritin as a marker of potential biochemical iron overload in athletes. Clin J Sport Med 2005;15:356—8. 4. Mast AE, Blinder MA, Gronowski AM, Chumley C, Scott MG. Clinical utility of the soluble transferrin receptor and comparison with serum ferritin in several populations. Clin Chem 1998;44:45—51. 5. Neilsen P, Nachtigall D. Iron supplementation in athletes. Current recommendations. Sports Med 1998;26:207—16.
336
K.E. Fallon
6. Pitsis GC, Fallon KE, Fallon SK, Fazakerly R. Response of soluble transferrin receptor and iro-related parameters to iron supplementation in elite, iron-depleted, nonanaemic female athletes. Clin J Sport Med 2004;14:300—4. 7. Conway AM, Vora AJ, Hinchliffe RF. The clinical relevance of an isolated increase in the number of circulating hyperchromic red blood cells. J Clin Path 2002;55:841—4. 8. Shaskey DJ, Green GA. Sports haematology. Sports Med 2000;29:27—38. 9. Datz C, Haas T, Rinner H, Sandhofer F, Patsch W, Paulweber B. Heterozygosity for the C282y mutation in the haemochromatosis gene is associated with increased serum iron, transferrin saturation and haemoglobin in young women: a protective role against iron deficiency? Clin Chem 1998;44:2429—32.
10. Nieman DC, Henson DA, Austin MD, Brown VA. Immune response to a 30 min walk. Med Sci Sports Exerc 2005;37:57—62. 11. Mc Carthy DA, Grant M, Marbut M, Watling M, Wade AJ, Macdonald I, et al. Brief exercise indices an immediate and a delayed leucocytosis. Br J Sports Med 1991;24:191— 5. 12. Cowell BS, Rosenbloom CA, Skinner R, Summers SH. Policies on screening female athletes for iron deficiency in NCAA division 1-A institutions. Int J Sport Nut Excer Metab 2003;13:277—85. 13. Qaseem A, Aronson M, Fitterman N, Snow V, Weiss KB, Owens DK. Screening for hereditary haemochromatosis: a clinical practice guideline from the American College of Physicians. Ann Intern Med 2005;143:517—21.
Available online at www.sciencedirect.com