0013-7227/03/$15.00/0 Printed in U.S.A.
The Journal of Clinical Endocrinology & Metabolism 88(1):394 – 401 Copyright © 2003 by The Endocrine Society doi: 10.1210/jc.2002-020037
The Growth Hormone/Insulin-Like Growth Factor-I Axis Hormones and Bone Markers in Elite Athletes in Response to a Maximum Exercise Test C. EHRNBORG, K. H. W. LANGE, R. DALL, J. S. CHRISTIANSEN, P.-A. LUNDBERG, R. C. BAXTER, M. A. BOROUJERDI, B.-A. BENGTSSON, M.-L. HEALEY, C. PENTECOST, S. LONGOBARDI, ´ N, ON BEHALF OF THE GH-2000 STUDY GROUP R. NAPOLI, AND T. ROSE Endocrine Division, Department of Internal Medicine (C.E., B.-A.B., T.R.), Sahlgrenska University Hospital, S-413 45 Go¨teborg, Sweden; Sports Medicine Research Unit, Bispebjerg Hospital (K.H.W.L.), DK-2400 Copenhagen NV, Denmark; Department of Medicine M (Endocrinology and Diabetes), Aarhus University Hospital (R.D., J.S.C.), DK-8000 Aarhus, Denmark; Department of Clinical Chemistry, Sahlgrenska University Hospital (P.-A.L.), S-413 45 Go¨teborg, Sweden; Kolling Institute of Medical Research, University of Sydney, Royal North Shore Hospital (R.C.B.), St. Leonards 2065, New South Wales, Australia; Department of Endocrinology, St. Thomas’s Hospital (M.A.B., M.-L.H., C.P.), SE1 7EH London, United Kingdom; and Department of Clinical Medicine and Cardiovascular Sciences, University Federico II (S.L., R.N.), 80131 Naples, Italy line for IGF-I (P < 0.001, males and females); IGFBP-3 (P < 0.001, males and females); acid-labile subunit [P < 0.001, males; not significant (NS), females], and IGFBP-2 (P < 0.05, females; NS, males). The serum concentrations of the bone markers ICTP (P < 0.001, males; P < 0.05, females) and P-III-P (P < 0.001, males and females) increased in both genders, with a peak value in the direct post-exercise phase and a subsequent decrease to baseline levels or below within 120 min. The osteocalcin and propeptide of type I procollagen values did not change during the exercise test. Specific reference ranges for each variable in the GH/IGF-I axis and bone markers at specific time points are presented. Most of the variables correlated negatively with age. In summary, the maximum exercise test showed a rather uniform pattern, with peak concentrations of the GH/IGF-I axis hormones and the bone markers ICTP and P-III-P immediately after exercise, followed by a subsequent decrease to baseline levels. The time to peak value for total GH and GH22 kDa was significantly shorter for females compared with males. This paper presents reference ranges for each marker in each gender at specific time points in connection to a maximum exercise test to be used in the development of a test for detection of GH abuse in sports. (J Clin Endocrinol Metab 88: 394 – 401, 2003)
The aim of the GH-2000 project is to develop a method for detecting GH doping among athletes. Previous papers in the GH-2000 project have proposed that a forthcoming method to detect GH doping will need specific components from the GH/ IGF-I axis and bone markers because these specific variables seem more sensitive to exogenous GH than to exercise. The present study examined the responses of the serum concentrations of these specific variables to a maximum exercise test in elite athletes from selected sports. A total of 117 elite athletes (84 males and 33 females; mean age, 25 yr; range, 18 –53 yr) from Denmark, the United Kingdom, Italy, and Sweden participated in the study. The serum concentrations of total GH, GH22 kDa, IGF-I, IGF binding protein (IGFBP)-2, IGFBP-3, acid-labile subunit, procollagen type III (P-III-P), and the bone markers osteocalcin, carboxy-terminal crosslinked telopeptide of type I collagen (ICTP), and carboxyterminal propeptide of type I procollagen were measured. The maximum exercise test showed, in both genders, a peak concentration of total GH (P < 0.001) and GH22 kDa (P < 0.001) by the time exercise ended compared with baseline, and a subsequent decrease to baseline levels within 30 – 60 min after exercise. The mean time to peak value for total GH and GH22 kDa was significantly shorter in males than females (P < 0.001). The components of the IGF-I axis showed a similar pattern, with a peak value after exercise compared with base-
D
OPING WITH GH has become an increasing problem in the world of sports, especially since the advent of recombinant GH in the late 1980s (1). It is mainly the anabolic and lipolytic effect of GH that is appreciated by its users, who nowadays are found among elite athletes, nonprofessional body builders, and young people in the gyms. The GH abuse is unwanted for both medical and ethical reasons. There is so far no official method available to discover GH doping, which might partly explain the strong position of GH as a doping agent in elite sports. This study is part of the
GH-2000 project initiated by the GH-2000 Team and funded by the European Union BioMed2 Research Program, with additional support from industry and the International Olympic Committee. The aim of this project is to develop a method for detecting GH doping among athletes. Acute exercise above a certain intensity is one of the most potent stimulators of GH secretion, and the magnitude of the GH response is closely related to the peak intensity of exercise, rather than to total work output (2– 4). Furthermore, 1 yr of endurance training above the lactate threshold has shown an increase in basal 24-h pulsatile GH release (5). Interestingly, subjects training below the lactate level did not show any change in the GH release, indicating that the training intensity may be important in regulating the GH axis as well as fitness. This physiological GH increase to exercise and
Abbreviations: ALS, Acid-labile subunit; bpm, beats per minute; CV, coefficient(s) of variation; HR, heart rate; ICTP, carboxy-terminal crosslinked telopeptide of type I collagen; IGFBP, IGF binding protein; NS, not significant; OC, oral contraceptive(s); PICP, carboxy-terminal propeptide of type I procollagen; P-III-P, procollagen type III; PV, plasma volume.
394
Ehrnborg et al. • GH/IGF-I and Bone Markers in Athletes
J Clin Endocrinol Metab, January 2003, 88(1):394 – 401 395
to other stimuli such as hypoglycemia makes it hard to use measurements of GH itself in blood as a doping marker, because it would be difficult to discriminate a high exercisederived endogenous GH level from that from exogenous GH. In previous papers in the GH-2000 project, Wallace et al. (6 –9) have studied the effects of exercise and supraphysiological GH administration on the GH/IGF-I axis and bone markers in 17 athletic adult males. In summary, acute exercise increased all molecular isoforms of GH, with GH22 kDa constituting the major isoform, with a peak at the end of acute exercise (6). The proportion of non-22-kDa isoforms increased after exercise, due in part to slower disappearance rates of these isoforms. With supraphysiological GH administration, these exercise-stimulated endogenous GH isoforms were suppressed for up to 4 d (7). Also, all components of the IGF-I ternary complex transiently increased with acute exercise, and GH pretreatment augmented these exercise- induced changes (8). Furthermore, acute exercise increased the serum concentrations of the bone and collagen markers bone-specific alkaline phosphatase, carboxy-terminal cross-linked telopeptide of type I collagen (ICTP), carboxy-terminal propeptide of type I procollagen (PICP), and procollagen type III (P-III-P), whereas osteocalcin was unchanged. GH treatment resulted in a augmented response to exercise for the bone markers PICP and ICTP (9). In two other papers from the GH-2000 study, the effects of 4 wk of supraphysiological GH administration to 99 healthy subjects of both genders on the IGF-I axis and bone markers were studied in a double-blind placebo-controlled manner (10, 11). In summary, the increase in IGF-I was markedly higher than that from IGF binding protein (IGFBP)-3 and acid-labile subunit (ALS), whereas the IGFBP-2 response was minor. Furthermore, the bone markers and collagen osteocalcin, PICP, ICTP, and P-III-P all increased on GH treatment, whereas interestingly the levels of P-III-P and osteocalcin were persistently high 8 wk after GH withdrawal (10). The response in both the IGF-I axis and among the bone markers was more pronounced in males compared with females. A forthcoming method to discover GH abuse will probably need the use of specific variables of the IGF-I axis and bone markers, with the prerequisite that these variables are more sensitive to exogenous GH administration than to exercise.
Consistent with this, it is important to closely study how the levels of these variables are influenced by a maximal exercise test in comparison to rest and by other factors such as gender, age, fitness, type of sport, medication, menstrual status, or illness. Thus, in this paper we have studied the effects of the serum concentrations of hormones in the GH/IGF-I axis and among bone markers in 117 elite athletes of both genders and different sports in relation to a maximum exercise test. Subjects and Methods Subjects A total of 117 elite athletes from Denmark, the United Kingdom, Italy, and Sweden (84 males and 33 females; mean age, 25 yr; range, 18–53 yr) competing in 12 different sports were included in the study (Table 1). The sports and the number of subjects in each were: alpine skiing, 21; crosscountry skiing, 23; long-distance cycling, 9; sprint cycling, 3; decathlon, 2; football, 10; rowing, 16; running, 6; swimming, 1; tennis, 3; triathlon, 8; and weight lifting, 15. The subjects were volunteers, and they gave written informed consent in concordance with the Helsinki Declaration II. The athletes were all at the national or international level. One hundred twelve of the athletes were Caucasian, four were black, and one was Oriental. The number and gender of the athletes, sport categories, and demographic data are presented in Table 1.
Methods Sampling was performed during a maximal exercise test performed under laboratory conditions.
Standardized VO2 max test The athletes were tested in different ways according to country and type of sport. Rowing test. Rowers were tested on a Concept II rowing ergometer (Concept II, Inc., Morrisville, VT). The protocol consisted of four 5-min submaximal stages with a 1-min break between the stages. Following the final submaximal stage, subjects were allowed a 10-min rest, after which an “all-out” test (6 min for males and 7 for females) was performed. The submaximal stages corresponded roughly to 55, 65, 75, and 85% of VO2 max, respectively. Ventilation and expiratory O2 and CO2 concentrations were measured by AMIS 2001 System (Innovision, Odense, Denmark). Before each protocol, the gas analyzers were calibrated using commercial gases of known volume. The highest VO2 attained during the test was registered as the VO2 max value. Heart rate (HR) was measured continuously by a HR monitor (Polar Sport tester, Polar Electro Oy, Kempele, Finland).
TABLE 1. Demographic data for 84 male and 33 female athletes in different sports Sport category
Alpine skiing Cross-country skiing Cycling (long distance) Cycling (sprint) Decathlon Football Rowing Running-swimming Tennis Triathlon Weight lifting Data represent mean ⫾
SD.
No.
Gender
Age (yr)
Height (cm)
Weight (kg)
BMI (kg/m2)
11 10 16 7 9 3 2 10 9 7 6 1 3 7 1 8 7
M F M F M M M M M F M F M M F M F
23.1 ⫾ 2.3 23.3 ⫾ 4.5 21.7 ⫾ 1.5 23.6 ⫾ 1.9 24.0 ⫾ 5.6 26.0 ⫾ 6.2 51.5 ⫾ 2.1 24.9 ⫾ 4.4 24.8 ⫾ 2.9 22.0 ⫾ 1.7 30.5 ⫾ 2.4 27.0 23.3 ⫾ 3.5 35.4 ⫾ 6.2 29.0 28.5 ⫾ 6.8 24.0 ⫾ 4.7
183.3 ⫾ 4.3 166.9 ⫾ 5.9 181.4 ⫾ 3.8 168.6 ⫾ 3.5 179.8 ⫾ 4.9 178.7 ⫾ 4.7 170.0 ⫾ 11.3 179.8 ⫾ 7.1 185.6 ⫾ 6.7 179.9 ⫾ 4.9 175.3 ⫾ 11.0 170.0 183.4 ⫾ 3.3 178.3 ⫾ 5.6 174.0 172.3 ⫾ 8.4 162.3 ⫾ 6.6
83.5 ⫾ 5.1 63.1 ⫾ 5.8 74.8 ⫾ 6.0 60.2 ⫾ 5.7 71.3 ⫾ 5.0 77.6 ⫾ 5.3 72.8 ⫾ 11.3 75.5 ⫾ 4.9 75.8 ⫾ 3.7 75.7 ⫾ 5.6 72.6 ⫾ 11.3 64.0 82.5 ⫾ 4.5 75.4 ⫾ 6.5 65.0 83.4 ⫾ 12 64.1 ⫾ 14.1
24.9 ⫾ 1.0 23.2 ⫾ 1.2 22.7 ⫾ 1.5 21.2 ⫾ 2.0 22.1 ⫾ 1.6 24.3 ⫾ 1.8 25.1 ⫾ 0.6 23.4 ⫾ 1.3 22.1 ⫾ 1.7 23.4 ⫾ 0.9 23.6 ⫾ 2.5 22.1 24.5 ⫾ 1.8 23.7 ⫾ 1.4 21.5 28.0 ⫾ 2.1 24.3 ⫾ 5.2
M, Male; F, female.
396
J Clin Endocrinol Metab, January 2003, 88(1):394 – 401
Cycling test (long distance). Long-distance cyclists were tested on a biking ergometer (Ciclotraining Olympionic, Politecnica 80, Padova, Italy). The protocol consisted of four 5-min submaximal stages as described in the rowing test. After a 10-min rest, a 5-min all-out test was performed. HR and VO2 were measured as previously described. Cycling test (sprint). Sprinters were tested on a biking ergometer (Monark, Vansbro, Sweden). The protocol consisted of three 15-sec allout sprints separated by 15-min rest. Weight-lifting test. The weight-lifting tests were performed on a contact carpet (Newtest, Oulu, Finland). The protocol consisted of five sets of three maximal jumps on the carpet: 1) squat jump, 2) counter movement jump, 3) counter movement jump with 50% body weight overload, 4) counter movement jump with 100% body weight overload, and 5) counter movement jump (12). Counter movement is a concentric dynamic muscle contraction that is preceded by an eccentric stretching of the muscle. The subject starts from an upright position. Treadmill. Cross-country and alpine skiers were tested on a treadmill. There was an initial warm-up period of about 10 min on a biking ergometer, with a workload of 200 W and mean HR registration (Polar Sport tester) of 140 –150 beats per minute (bpm) during the last 2 min of warming up. Calibration of airflow and volume before the test was made with a syringe pump. Calibration of sensors for oxygen and carbon dioxide was made with air and a calibrating gas (volume, O2 16% and CO2 6.05%). A facial mask was used to collect the expiratory gases for analysis. The analyses were made with a MetaMax I (Cortex Biophysik GmbH, Leipzig, Germany) analyzer. The test was performed on a treadmill (Spectra Elit, Avesta, Sweden) with an elevation of 3° from start. The speed was adjusted to the running capacity of the athlete with a HR of 160 –170 bpm during 3– 4 min. There was an increase in workload each minute (speed or elevation), starting after 4 min. A rise in HR with 5–10 bpm was aimed for each new level of workload. The number of increases in workload was 8 –12 until VO2 reached a plateau. Respiratory quotient and other parameters were followed during the test. Cycling test (for noncyclists). Tennis players, football players, swimmers, decathlon athletes, triathletes, and two Swedish cross-country skiers performed a cycle test. The test, performed on a biking ergometer (Rodby ergometercyklar, Stockholm, Sweden), followed the same principle as the running test. The athletes started with a workload of 200 W for 4 min, with a rise every minute thereafter monitored by HR. Approximate rise was 25 W/min. HR monitoring and analyses were made with the same equipment and in the same way as for the running test.
Demographic data and body composition Body height and body weight were measured in each participant. In conjunction with the maximum exercise test, data forms were completed, including demographic details concerning age, ethnic origin, sport, level of sport, medication, current illnesses, injuries, and menstrual status.
Maximum exercise test samples The blood sampling at the maximum exercise test was done in a standardized way. Cannulation was done in a forearm vein 30 min before the test, and a baseline sample was taken immediately before the start of the test. Samples were taken at the end of the test and thereafter at 15, 30, 60, 90, and 120 min after exercise. During the 2-h sampling period, no further exercise or food intake was allowed, and according to the protocol a maximum intake of 0.2 liter water was allowed.
Laboratory methods All GH-2000 analyses were performed in centralized laboratories, i.e. in the laboratory of Per-Arne Lundberg at the Sahlgrenska University Hospital (Go¨ teborg, Sweden; total GH, GH22 kDa, IGF-I, osteocalcin, PICP, ICTP, and P-III-P) and in the laboratory of Robert Baxter (Sydney, Australia; IGFBP-2, IGFBP-3, and ALS). All samples were stored in a freezer, ⫺70 C, at each center and then shipped to the laboratories that performed the analyses. The samples were sent at the same time from
Ehrnborg et al. • GH/IGF-I and Bone Markers in Athletes
all of the centers, and all samples in the study were analyzed at the same time. The same batch was used for all countries. The serum concentration of total GH was determined by an immunoradiometric assay (Pharmacia & Upjohn Diagnostics AB, Uppsala, Sweden) with intra-assay coefficients of variation (CV) of 13.2%, 2.8%, and 5.4% at serum concentrations of 0.72, 18.1, and 76.3 mU/liter, respectively. The serum concentration of GH22 kDa was determined by a fluoroimmunoassay (Wallac, Inc. Oy, Turku, Finland), with intra-assay CV of 12.%, 5.1%, and 7.3% at serum concentrations of 0.54, 11.84, and 54.2 mU/liter, respectively. The serum concentration of IGF-I was determined by a hydrochloric acid-ethanol extraction RIA, using authentic IGF-I for labeling (Nichols Institute Diagnostics, San Juan Capistrano, CA), with intra-assay CV of 10.1%, 6.3%, and 5.7% at serum concentrations of 61.5, 340.8, and 776.9 g/liter, respectively. The serum concentration of osteocalcin was measured by a double antibody RIA (International CIS, Gif-sur Yvette, France) with intra-assay CV of 8.16%, 5.60%, and 5.44% at serum concentrations of 3.56, 10.32, and 22.26 g/liter, respectively. The serum concentration of PICP was measured by a RIA (Orion Diagnostica, Espoo, Finland) with intra-assay CV of 6.2%, 11.3%, and 11.3% at serum concentrations of 112.2, 162.7, and 403.4 g/liter, respectively. The serum concentration of the ICTP was measured with a RIA (Orion Diagnostica) with intra-assay CV of 5.57%, 7.22%, and 5.11% at serum concentrations of 5.46, 3.22, and 16.83 g/ liter, respectively. The serum concentration of P-III-P was determined by a RIA (International CIS) with intra-assay CV of 5.72%, 9.12%, and 6.70% at serum concentrations of 0.95, 0.62, and 1.18 kU/liter, respectively. The serum concentration of IGFBP-2 was measured using in-house RIAs and polyclonal antibodies. The intra-assay CV were 2.8%, 2.8%, and 3.2% at serum concentrations of 17, 73, and 330 g/liter, respectively; and the interassay CV were 14.1% and 12.7% at 65 and 775 g/liter, respectively. The serum concentration of IGFBP-3 was measured using in-house RIAs and polyclonal antibodies. The intra-assay CV were 6.2%, 5.5%, and 4.5% at serum concentrations of 2.5, 5.7, and 12.6 mg/liter, respectively, and the interassay CV were 11.9%, 14.5%, and 13.1% at 2.5, 5.7, and 12.6 mg/liter, respectively. The serum concentration of the ALS was measured using in-house RIAs and polyclonal antibodies. The intra-assay CV were 3.4%, 3.3%, and 3.4% at serum concentrations of 60, 245, and 502 nmol/liter, respectively, and the inter-assay CV were 10.5%, 5.4%, and 6.5% at 62, 282, and 676 nmol/liter, respectively.
Statistics All values are presented as mean ⫾ sd. The statistical analyses were done with nonparametric tests in the Statview software package (SAS Institute, Inc., Cary, NC). Comparisons within groups were done with Wilcoxon signed rank test, and comparisons between groups were done with Mann Whitney U test. Correlations were tested with the Spearman correlation test. Results with P values below 0.05 were regarded as significant.
Results Maximum exercise test
The hormonal responses to the maximum exercise test among both the GH/IGF-I axis hormones and the bone markers are shown in Figs. 1 and 2 . Males and females are displayed separately. GH/IGF-I axis hormones. The male and female athletes showed a rather uniform pattern in the serum concentrations of the different components in the GH/IGF-I axis with a peak value at the end of test values, followed by a subsequent decrease to baseline values within 30 – 60 min after exercise. The hormonal responses to the test and the post-exercise period are shown in Fig. 1. The mean serum concentrations (mean ⫾ sd) and the minimum-maximum ranges of the components in the GH/IGF-I axis at baseline, end of test, and ⫹30 min are shown in Table 2. The time from end of exercise to the peak value of both
Ehrnborg et al. • GH/IGF-I and Bone Markers in Athletes
J Clin Endocrinol Metab, January 2003, 88(1):394 – 401 397
FIG. 1. Serum-concentrations (mean ⫾ SD) of components in the GH/IGF-I axis in 84 male and 33 female elite athletes in connection with a maximum exercise test. *, P ⬍ 0.05; **, P ⬍ 0.01; and ***, P ⬍ 0.001, indicate changes compared with baseline.
total GH and GH22 kDa was shorter in females than in males. Thus, the mean time (calculated as the mean time point for each individual’s maximum concentrations) was: total GH, ⫹10.56 ⫾ 9.95 min in males; ⫹0.26 ⫾ 4.24 min in females (P ⬍ 0.001); GH22 kDa, ⫹10.12 ⫾ 10.15 min in males; ⫺1.21 ⫾ 5.87 min in females (P ⬍ 0.001). Bone markers. The serum concentrations of PICP, ICTP, and P-III-P showed in both males and females a uniform pattern for each marker, with a peak at the end of exercise followed by a decrease below baseline values within 120 min. Osteocalcin concentrations showed no significant changes during the test (Fig. 2.) The mean serum concentrations (mean ⫾ sd) and the minimum-maximum ranges of the bone markers at baseline, end of test, and ⫹30 min are shown in Table 2.
Total minimal and maximal values. The minimal and maximal serum concentrations with the 10 –90 percentile ranges of the markers in the GH/IGF-I axis and bone markers based on all of the samples performed in connection with the maximum exercise test are shown in Table 3. Correlations to age, weight, BMI, height, menstrual status, diseases, medication, and ethnic origin of the components in the GH/IGF-I axis and the bone markers
Age. There were no correlations between the serum concentrations of total GH, GH22 kDa, and IGFBP-2 to age in either males or females. However, the concentrations of IGF-I, IGFBP-3, and ALS correlated negatively to age in both males and females at baseline (P ⬍ 0.001, ⫽ ⫺0.5, males; P ⬍ 0.01, ⫽ ⫺0.6, females), end of exercise (P ⬍ 0.001, ⫽ ⫺0.5, males; P ⬍ 0.01,
398
J Clin Endocrinol Metab, January 2003, 88(1):394 – 401
Ehrnborg et al. • GH/IGF-I and Bone Markers in Athletes
FIG. 2. Serum-concentrations (mean ⫾ SD) of components in the bone markers in 84 male and 33 female elite athletes in connection with a maximum exercise test. *, P ⬍ 0.05; **, P ⬍ 0.01; and ***, P ⬍ 0.001, indicate changes compared with baseline.
TABLE 2. The mean (⫾SD) serum concentrations and minimal-maximal ranges in the GH/IGF-I axis and among bone markers at baseline, end of exercise, and ⫹30 min after exercise in 84 males and 33 females in connection with a maximum exercise test Baseline Mean ⫾
Males Total GH (mU/liter) GH22 kDa (mU/liter) IGF-I (g/liter) IGFBP-3 (g/ml) ALS (nmol/liter) IGFBP-2 (ng/ml) Osteocalcin (g/liter) PICP (g/liter) ICTP (g/liter) P-III-P (kU/liter) Females Total GH (mU/liter) GH22 kDa (mU/liter) IGF-I (g/liter) IGFBP-3 (g/ml) ALS (nmol/liter) IGFBP-2 (ng/ml) Osteocalcin (g/liter) PICP (g/liter) ICTP (g/liter) P-III-P (kU/liter)
SD
End of exercise Range
Mean ⫾
SD
30 min after exercise
Range
P
Mean ⫾
SD
Range
P
4.6 ⫾ 8.9 2.9 ⫾ 5.7 285 ⫾ 87.5 3.9 ⫾ 0.6 249 ⫾ 54.2 222 ⫾ 55 12.2 ⫾ 3.1 218 ⫾ 104 4.4 ⫾ 1.38 0.52 ⫾ 0.13
0.0– 42.8 0.0–30.7 123.8–515.1 2.8–5.8 154– 419 119–324 5.60–21.74 100.7– 424.9 2.35– 8.56 0.20– 0.83
38 ⫾ 38.2 25 ⫾ 27.4 319 ⫾ 96.2 4.2 ⫾ 0.6 289 ⫾ 67.7 237 ⫾ 61 12 ⫾ 3.4 212 ⫾ 95 5 ⫾ 1.5 0.6 ⫾ 0.15
0.0–203.4 0.1–132.2 140.2–546.3 3.0–5.7 166– 467 128–325 4.98–21.83 106.1–390.5 2.59–9.24 0.18–1.00
⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 NS NS NS ⬍0.001 ⬍0.001
32.2 ⫾ 29 20 ⫾ 19 288 ⫾ 91 3.9 ⫾ 0.7 239 ⫾ 58 210 ⫾ 56 12.3 ⫾ 2.9 192 ⫾ 79 4.38 ⫾ 1.22 0.49 ⫾ 0.11
1.7–144.7 0.1–90.0 121.6–525.5 2.6–5.5 141– 409 120–296 5.40–20.53 104.5–336.0 2.22–7.84 0.30– 0.75
⬍0.001 ⬍0.001 NS NS NS NS NS NS NS ⬍0.01
17.6 ⫾ 15.1 7.5 ⫾ 9.2 325 ⫾ 106 4.4 ⫾ 0.9 312 ⫾ 104 207 ⫾ 98 9.6 ⫾ 3.7 203 ⫾ 72 4.2 ⫾ 1.01 0.56 ⫾ 0.16
1.2– 46.4 0.2–33.3 187.5–710.0 3.3– 6.5 150–539 42–352 3.25–18.25 127.1–375.7 2.30– 6.24 0.23– 0.88
79 ⫾ 50.5 32 ⫾ 32.7 348 ⫾ 121 4.7 ⫾ 0.8 316 ⫾ 100 212 ⫾ 101 9.2 ⫾ 4.2 211 ⫾ 56 4.5 ⫾ 1.17 0.7 ⫾ 0.17
14.7–202.4 0.4–112.1 184.5–741.7 3.5– 6.4 148–546 48–349 2.51–21.55 135.4–336.2 2.30–7.29 0.41–1.07
⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 NS NS NS NS ⬍0.05 ⬍0.001
36.8 ⫾ 28 14.1 ⫾ 17 314 ⫾ 101 275 ⫾ 89 4.4 ⫾ 0.8 196 ⫾ 95 9.8 ⫾ 3.2 190 ⫾ 55 3.93 ⫾ 1.0 0.55 ⫾ 0.1
4.8–105.1 0.2– 65.2 165.6– 694.6 142– 444 3.1– 6.2 54–313 3.56–18.50 119.6–298.4 1.84– 6.16 0.33– 0.80
⬍0.01 ⬍0.05 NS ⬍0.01 ⬍0.001 NS NS NS ⬍0.05 NS
P indicates changes in mean values compared with baseline.
Ehrnborg et al. • GH/IGF-I and Bone Markers in Athletes
J Clin Endocrinol Metab, January 2003, 88(1):394 – 401 399
TABLE 3. The minimal and maximal serum concentrations with the 10 –90 percentile ranges of the markers in the GH/IGF-I axis and bone markers based on all the samples at any time point, performed in 84 male and 33 female elite athletes in connection with a maximum exercise test Males
Total GH (mU/liter) GH22 kDa (mU/liter) IGF-I (g/liter) IGFBP-3 (g/ml) ALS (nmol/liter) IGFBP-2 (ng/ml) Osteocalcin (g/liter) PICP (g/liter) ICTP (g/liter) P-III-P (kU/liter)
Females
Minimal
Maximal
Range (10 –90 percentiles)
Minimal
Maximal
Range (10 –90 percentiles)
0.0 0.0 104 2.6 138 114 5.0 91.7 2.2 0.16
203.4 139.9 568 6.0 467 325 21.8 425 9.2 1.00
0.3–51.2 0.27–34.4 173– 416 3.2– 4.7 181–334 139–296 8.03–16.1 109–329 3.0– 6.3 0.37– 0.70
0.6 0.1 163 3.0 123 42 2.5 119 1.8 0.23
202.4 112.1 774 6.5 727 365 21.6 376 7.3 1.07
2.0– 84.4 0.5–34.6 200– 429 3.5–5.5 178– 432 65–315 5.4–13.5 131–272 2.5–5.4 0.39– 0.78
⫽ ⫺0.5, females), and at the ⫹15-min sample (P ⬍ 0.001, ⫽ ⫺0.5, males; P ⬍ 0.01, ⫽ ⫺0.5, females). In males, the osteocalcin (P ⬍ 0.001, ⫽ ⫺0.4) and ICTP (P ⬍ 0.001, ⫽ ⫺0.5) concentrations at baseline were negatively correlated to age, and also the ICTP concentrations at the end of the test (P ⬍ 0.001, ⫽ ⫺0.5) and at ⫹15 min (P ⬍ 0.001, ⫽ ⫺0.5). Also, the P-III-P concentrations at the end of the test (P ⬍ 0.05, ⫽ ⫺0.2) and at ⫹15 min (P ⬍ 0.01; ⫽ ⫺0.3) correlated negatively to age. In females, there were negative correlations between the PICP concentrations at baseline (P ⬍ 0.05, ⫽ ⫺0.6) and at the end of the test (P ⬍ 0.05, ⫽ ⫺0.6) and between ICTP concentrations at the end of the test (P ⬍ 0.05, ⫽ ⫺0.4) and at ⫹15 min (P ⬍ 0.05, ⫽ ⫺0.4). Weight and BMI. In males, a negative correlation between serum concentrations of IGFBP-3 and weight at the end of the test (P ⬍ 0.05, ⫽ ⫺0.3) was the only correlation found between weight and serum concentrations of the markers. In the females, the concentrations of IGF-I and weight correlated positively at baseline and at the end of the test (P ⬍ 0.05, ⫽ 0.4; vs. P ⬍ 0.05, ⫽ 0.4). No other correlations between weight and serum concentrations were found in females. Otherwise, no correlations between the components in the GH/IGF-I axis or bone markers to weight or BMI were found. Height. In males, the concentrations of osteocalcin (baseline, P ⬍ 0.05, ⫽ 0.3), ICTP (baseline, P ⬍ 0.05, ⫽ 0.2; end of test, P ⬍ 0.05, ⫽ 0.3) and P-III-P (end of test, P ⬍ 0.05, ⫽ 0.2; ⫹15 min, P ⬍ 0.05, ⫽ 0.2; ⫹30 min, P ⬍ 0.05, ⫽ 0.2; ⫹90 min, P ⬍ 0.01, ⫽ 0.3) correlated to height. Menstrual status and oral contraceptives (OC). A negative correlation was found between the time since the first day of the last period to the day of the test and the serum concentrations of IGFBP-3 at baseline (P ⬍ 0.01; ⫽ ⫺0.5), end of test (P ⬍ 0.05; ⫽ ⫺0.4), ⫹15 min (P ⬍ 0.05; ⫽ ⫺0.4), ⫹30 min (P ⬍ 0.05; ⫽ ⫺0.4), ⫹60 min (P ⬍ 0.05; ⫽ ⫺0.4), and ⫹120 min (P ⬍ 0.01; ⫽ ⫺0.5). The impact of OC could not be ascertained due to the small number of female athletes on OC. Diseases and medication. Any correlation between the concentrations of the components in the GH/IGF-I axis and bone markers to diseases and medications among the athletes could not be performed due to low frequency of diseases and medication in the participants.
Ethnic origin. Because there were too few non-Caucasian participants, no analysis of any differences between ethnic groups could be performed. Discussion
This study encompassing nearly 120 elite athletes describes the hormonal profiles in the GH/IGF-I axis and among bone markers in response to a maximal exercise test. In summary, the hormonal response in the GH/IGF-I axis showed a rather uniform pattern, including a peak value directly after exercise, and a successive decrease to baseline values within 30 – 60 min after exercise, although this pattern was less obvious among the bone markers. Interestingly, the peak value of the total GH and GH22 kDa concentrations occurred significantly earlier in the female athletes compared with males. The reference ranges for components in the GH/IGF-I axis and among bone markers in respective gender, presented in this paper, might be of considerable use in the future development of a test for the detection of GH abuse in sports. This study favorably comprises a large number of elite athletes in a number of specific sports. The subjects were all elite athletes on the national or international level. Although the study was performed on a large number of athletes in four different European countries, the homogeneity of the practical performance of the protocol was high. The maximum exercise tests that were used were specific for the individual sport category, thus rendering familiar exhaustion conditions for the athletes and therefore a possibility to perform a maximal test. We are certain that the participants were driven to exhaustion level. The increase in GH levels in the early post-exercise phase of the maximum exercise test is in accordance with previous findings (4, 13). The neuroendocrine mechanisms of exerciseinduced GH release are still incompletely understood. Thus, it is not yet obvious whether the GH increase is due to increased GHRH stimulation, decreased somatostatin stimulation, or a combination of these (14). The roles of stimulating ␣-adrenergic, inhibiting 2-adrenergic, and stimulating cholinergic pathways on the secretion of somatostatin and GHRH are to be ascertained. We noticed the GH peak to be within 15 min after exercise. Previous studies have shown that the GH release induced by exercise is known to be delayed until 15 min into exercise (15, 16) and to peak by the end of short-term exercise (15, 16) or
400
J Clin Endocrinol Metab, January 2003, 88(1):394 – 401
shortly afterward (17). Furthermore, the GH release is related to the intensity of the exercise (18 –20), thus rendering higher GH release from high-intensity anaerobic work compared with low-intensity anaerobic work, although equal in duration and total work effort (18). Finally, training in itself increases the GH secretion (5). The exercise-induced GH increase was noted in both sexes, although the maximal GH response occurred significantly earlier in the female athletes, in which the peak value was directly after exercise, compared with approximately 15 min after exercise among the males. This is in accordance with a previous study with a constant-load aerobic exercise (21), in which females attained their GH peak significantly earlier than men (24 vs. 32 min after start of exercise). Thus, the present study shows that this phenomenon is valid also during anaerobic conditions among elite athletes. Previous studies in elite athletes have shown no major gender differences in VO2 max-induced GH release. Some authors have noted that females have higher GH levels before exercise, levels that do not return to baseline within 1 h, although the general pattern of exercise response did not differ from the response in males (15, 21, 22). In the female athletes, the different stages of the menstrual phase had no influence on the GH response. This is in accordance with the study of Kanaley et al. (22), although in their study the athletes performed a submaximal test (60% of VO2 max). Otherwise, the GH response at continuous and intermittent exercise is known to be higher in females during the OC phase than the nonuse phase (23), possibly due to elevated levels of total estrogen. However, this could not be shown in our study, because of the low number of OC users among the female athletes. The increase in serum total IGF-I concentrations in the direct post-exercise period, with a subsequent decrease to baseline levels, was noted in both genders and has been observed in other trials (24, 25). Furthermore, we noted the same pattern of response to exercise of the remaining components in the 150 kDa ternary complex, IGFBP-3 and ALS. The maximal response was observed directly after the exercise, in comparison with the maximal GH response, which was observed 15–30 min after exercise. The exact mechanism for the parallel changes in IGF-I, IGFBP-3, and ALS in response to exercise is not known, but reasonably, exercise, with a subsequent GH increase and possibly under the influence of proteases, causes changes such as dissociation of the ternary complex according to the theory presented by Wallace et al. (8). The bone and collagen markers P-III-P and ICTP showed in both sexes, like the components of the ternary complex described above, a peak level directly after exercise, followed by a subsequent decrease to baseline levels within 2 h. The increase in osteocalcin levels was much less obvious, with only a minor increase at the 2-h value. Previous studies have shown divergent results of bone markers in response to exercise (26 –30). Long-term exercise training from 1–18 months has shown increases in the concentrations of osteocalcin and other markers of bone formation, with a slight decline of the levels during the first 4 wk (31–33). The mechanism for the direct post-exercise rise of the P-III-P and ICTP concentrations is probably multifactorial. Metabolic acidosis with high lactate levels is known to stimulate osteoclastic formation and thus increase the ICTP concentrations (34); furthermore,
Ehrnborg et al. • GH/IGF-I and Bone Markers in Athletes
long-term training, including mechanical factors from the exercise performance itself causing microdamage of bone and muscles with leakage to the blood, might also contribute. The present study now presents the actual reference ranges for each marker in the GH/IGF-I axis and among the bone markers at different time courses, i.e. at baseline, end of exercise, and 30 min after exercise. Furthermore, we also present the all-time high and low, including the 10 –90 percentile range values recorded at any time point, in connection with the maximum test, thus giving us the overall description of the total ranges in connection with the test, including baseline. The variables in the ternary complex and most of the bone markers correlated negatively to age, which will have implications when trying to definitely define reference ranges. Although, some variables in the ternary complex had some correlation to weight and BMI at some time points, this does not seem to have any practical consequences. There was no correlation between the phase of the menstrual period and the concentrations of the variables in the GH/IGF-I axis (including the maximal GH response as discussed above) and bone markers, apart from the IGFBP-3 levels, which seemed to decrease during the time of the menstrual cycle. The number of athletes with OC, diseases, medication, and non-Caucasian ethnic origin was too small to make it possible to estimate the influence of these variables on the reference ranges. Most of the markers had a maximum value at about the end of exercise in the maximum exercise test, with a gradual decrease back to baseline values within 30 – 60 min. Therefore, we recommend that blood samples, in a future GHdoping control situation, should be taken not earlier than 30 min after the end of exercise to avoid confusion with this physiological post-exercise increase. We cannot exclude that some of the changes might be due to the effect of hemoconcentration in response to the exercise itself, with a decrease in plasma volume (PV). In this study, the PV changes were estimated by measuring serum albumin concentrations (data not shown), showing a PV decrease estimated to about 20%, which is roughly in accordance with previous studies in which PV reductions from 4 –15% were noted after moderate submaximal exercise and total exhaustion (35). The reference ranges that we have presented for each marker are described according to the actual laboratory results that have been obtained in the actual sampling occasion, regardless of significant hemoconcentration or not, and will in that sense be comparable with the results obtained in the sampling situation of a doping test. In summary, this study encompassing nearly 120 elite athletes describes the hormonal answer in the GH/IGF-I axis and among bone markers in response to a maximum exercise. The hormonal response in the GH/IGF-I axis showed a rather uniform pattern, including a peak value directly after exercise and a successive decrease to baseline values within 30 – 60 min after exercise. This pattern was less obvious among the bone markers. Interestingly, the peak value of the total GH and GH22 kDa concentrations occurred significantly earlier in the female athletes compared with males. The reference ranges for each marker in respective gender, presented in this paper, might be of considerable use in the future development of a test for the detection of GH abuse in sports.
Ehrnborg et al. • GH/IGF-I and Bone Markers in Athletes
J Clin Endocrinol Metab, January 2003, 88(1):394 – 401 401
Acknowledgments We thank all members of the GH-2000 group for their comments and encouragement, including Peter H. So¨nksen (Project Coordinator), Jake Powrie, David Russell-Jones, and Nicola Keay from the St. Thomas’ Hospital, London, UK; Eryl Bassett, Mike Kenward, and Phil Brown from the Mathematics Institute, Kent University, Canterbury, UK; Lena Carlsson from the Sahlgrenska Hospital, Gothenberg, Sweden; Jens-Otto Jorgensen and Hans Ørskov from the Aarhus University Hospital, Aarhus, Denmark; Luigi Sacca and Antonio Cittadini from the University Federico II, Naples, Italy; Laurent Rivier from the Laboratoire Suisse d’Analyze du Dopage, Lausanne, Switzerland; Don Catlin from International Olympic Committee Drug Testing Laboratory, Los Angeles, CA; Jennifer D. Wallace and Ross C. Cuneo from the University of Queensland, Princess Alexandra Hospital, Brisbane, Australia; Benny Larsson, Team Denmark, and Michael Kjaer, Sports Medicine Research Unit, Bispebjerg Hospital, University of Copenhagen, Copenhagen, Denmark; Par Gellerfors and Linda Fryklund from Pharmacia & Upjohn, Stockholm, Sweden; and Anne-Marie Kappelgard from Novo Nordisk, Bagsvaerd, Denmark. We also thank Åke Jo¨ nsson, Hans Ottosson, and Kjell-Erik So¨ der in ¨ stersund, Sweden, for their help with the performance of the Swedish O part of the study.
11.
12. 13. 14. 15. 16. 17. 18.
Received January 14, 2002. Accepted September 13, 2002. Address all correspondence and requests for reprints to: Christer Ehrnborg, Endocrine Division, Department of Internal Medicine, Gro¨ na stråket 8, Sahlgrenska University Hospital, S-413 45 Go¨ teborg, Sweden. E-mail:
[email protected]. This study was supported by grants from the European Union (BIOMED 2 Project Number BMH4 CT 950678) and the International Olympic Committee. Additional financial support and recombinant human GH were provided by Novo Nordisk and Pharmacia & Upjohn. The study was also supported with funds from the Go¨ teborg Society of Medicine and the Swiss Foundation for Research. This study was part of the GH-2000 project, a research program performed within the competitive EU BioMed2 Research Programme, with additional support from the universities of Aarhus, Gothenburg, Naples, and London.
References 1. Macintyre JG 1987 Growth hormone and athletes. Sports Med 4:129 –142 2. Kjaer M, Bangsbo J, Lortie G, Galbo H 1988 Hormonal response to exercise in humans: influence of hypoxia and physical training. Am J Physiol 254: R197–R203 3. Kraemer WJ, Gordon SE, Fleck SJ, Marchitelli LJ, Mello R, Dziados JE, Friedl K, Harman E, Maresh C, Fry AC 1991 Endogenous anabolic hormonal and growth factor responses to heavy resistance exercise in males and females. Int J Sports Med 12:228 –235 4. Felsing NE, Brasel JA, Cooper DM 1992 Effect of low and high intensity exercise on circulating growth hormone in men. J Clin Endocrinol Metab 75:157–162 5. Weltman A, Weltman JY, Schurrer R, Evans WS, Veldhuis JD, Rogol AD 1992 Endurance training amplifies the pulsatile release of growth hormone: effects of training intensity. J Appl Physiol 72:2188 –2196 6. Wallace JD, Cuneo RC, Bidlingmaier M, Lundberg PA, Carlsson L, Boguszewski CL, Hay J, Boroujerdi M, Cittadini A, Dall R, Rosen T, Strasburger CJ 2001 Changes in non-22-kilodalton (kDa) isoforms of growth hormone (GH) after administration of 22-kDa recombinant human GH in trained adult males. J Clin Endocrinol Metab 86:1731–1737 7. Wallace JD, Cuneo RC, Bidlingmaier M, Lundberg PA, Carlsson L, Boguszewski CL, Hay J, Healy ML, Napoli R, Dall R, Rosen T, Strasburger CJ 2001 The response of molecular isoforms of growth hormone to acute exercise in trained adult males. J Clin Endocrinol Metab 86:200 –206 8. Wallace JD, Cuneo RC, Baxter R, Orskov H, Keay N, Pentecost C, Dall R, Rosen T, Jorgensen JO, Cittadini A, Longobardi S, Sacca L, Christiansen JS, Bengtsson BA, Sonksen PH 1999 Responses of the growth hormone (GH) and insulin-like growth factor axis to exercise, GH administration, and GH withdrawal in trained adult males: a potential test for GH abuse in sport. J Clin Endocrinol Metab 84:3591–3601 9. Wallace JD, Cuneo RC, Lundberg PA, Rosen T, Jorgensen JO, Longobardi S, Keay N, Sacca L, Christiansen JS, Bengtsson BA, Sonksen PH 2000 Responses of markers of bone and collagen turnover to exercise, growth hormone (GH) administration, and GH withdrawal in trained adult males. J Clin Endocrinol Metab 85:124 –133 10. Longobardi S, Keay N, Ehrnborg C, Cittadini A, Rosen T, Dall R, Boroujerdi MA, Bassett EE, Healy ML, Pentecost C, Wallace JD, Powrie J, Jorgensen JO,
19. 20. 21. 22. 23. 24.
25. 26. 27. 28. 29.
30. 31.
32.
33. 34. 35.
Sacca L 2000 Growth hormone (GH) effects on bone and collagen turnover in healthy adults and its potential as a marker of GH abuse in sports: a double blind, placebo-controlled study. The GH-2000 Study Group. J Clin Endocrinol Metab 85:1505–1512 Dall R, Longobardi S, Ehrnborg C, Keay N, Rosen T, Jorgensen JO, Cuneo RC, Boroujerdi MA, Cittadini A, Napoli R, Christiansen JS, Bengtsson BA, Sacca L, Baxter RC, Basset EE, Sonksen PH 2000 The effect of four weeks of supraphysiological growth hormone administration on the insulin-like growth factor axis in women and men. GH-2000 Study Group. J Clin Endocrinol Metab 85:4193– 4200 Komi P 1992 Stretch-shortening cycle. In: Komi P, ed. Strength and power in sports, encyclopedia of sports medicine, vol III. Oxford: Blackwell Science; 169–179 Sutton J, Lazarus L 1976 Growth hormone in exercise: comparison of physiological and pharmacological stimuli. J Appl Physiol 41:523–527 Giustina A, Veldhuis JD 1998 Pathophysiology of the neuroregulation of growth hormone secretion in experimental animals and the human. Endocr Rev 19:717–797 Lassarre C, Girard F, Durand J, Raynaud J 1974 Kinetics of human growth hormone during submaximal exercise. J Appl Physiol 37:826 – 830 Raynaud J, Capderou A, Martineaud JP, Bordachar J, Durand J 1983 Intersubject viability in growth hormone time course during different types of work. J Appl Physiol 55:1682–1687 Parkin JM 1986 Exercise as a test of growth hormone secretion. Acta Endocrinol Suppl 279:47–50 Vanhelder WP, Goode RC, Radomski MW 1984 Effect of anaerobic and aerobic exercise of equal duration and work expenditure on plasma growth hormone levels. Eur J Appl Physiol Occup Physiol 52:255–257 Hartley LH, Mason JW, Hogan RP, Jones LG, Kotchen TA, Mougey EH, Wherry FE, Pennington LL, Ricketts PT 1972 Multiple hormonal responses to prolonged exercise in relation to physical training. J Appl Physiol 33:607– 610 Hartley LH, Mason JW, Hogan RP, Jones LG, Kotchen TA, Mougey EH, Wherry FE, Pennington LL, Ricketts PT 1972 Multiple hormonal responses to graded exercise in relation to physical training. J Appl Physiol 33:602– 606 Wideman L, Weltman JY, Shah N, Story S, Veldhuis JD, Weltman A 1999 Effects of gender on exercise-induced growth hormone release. J Appl Physiol 87:1154 –1162 Kanaley JA, Boileau RA, Bahr JA, Misner JE, Nelson RA 1992 Substrate oxidation and GH responses to exercise are independent of menstrual phase and status. Med Sci Sports Exerc 24:873– 880 Bernardes RP, Radomski MW 1998 Growth hormone responses to continuous and intermittent exercise in females under oral contraceptive therapy. Eur J Appl Physiol Occup Physiol 79:24 –29 Bang P, Brandt J, Degerblad M, Enberg G, Kaijser L, Thoren M, Hall K 1990 Exercise-induced changes in insulin-like growth factors and their low molecular weight binding protein in healthy subjects and patients with growth hormone deficiency. Eur J Clin Invest 20:285–292 Cappon J, Brasel JA, Mohan S, Cooper DM 1994 Effect of brief exercise on circulating insulin-like growth factor I. J Appl Physiol 76:2490 –2496 Nishiyama S, Tomoeda S, Ohta T, Higuchi A, Matsuda I 1988 Differences in basal and postexercise osteocalcin levels in athletic and nonathletic humans. Calcif Tissue Int 43:150 –154 Salvesen H, Johansson AG, Foxdal P, Wide L, Piehl-Aulin K, Ljunghall S 1994 Intact serum parathyroid hormone levels increase during running exercise in well-trained men. Calcif Tissue Int 54:256 –261 Virtanen P, Viitasalo JT, Vuori J, Vaananen K, Takala TE 1993 Effect of concentric exercise on serum muscle and collagen markers. J Appl Physiol 75:1272–1277 Kristoffersson A, Hultdin J, Holmlund I, Thorsen K, Lorentzon R 1995 Effects of short-term maximal work on plasma calcium, parathyroid hormone, osteocalcin and biochemical markers of collagen metabolism. Int J Sports Med 16:145–149 Welsh L, Rutherford OM, James I, Crowley C, Comer M, Wolman R 1997 The acute effects of exercise on bone turnover. Int J Sports Med 18:247–251 Menkes A, Mazel S, Redmond RA, Koffler K, Libanati CR, Gundberg CM, Zizic TM, Hagberg JM, Pratley RE, Hurley BF 1993 Strength training increases regional bone mineral density and bone remodeling in middle-aged and older men. J Appl Physiol 74:2478 –2484 Lohman T, Going S, Pamenter R, Hall M, Boyden T, Houtkooper L, Ritenbaugh C, Bare L, Hill A, Aickin M 1995 Effects of resistance training on regional and total bone mineral density in premenopausal women: a randomized prospective study. J Bone Miner Res 10:1015–1024 Eliakim A, Raisz LG, Brasel JA, Cooper DM 1997 Evidence for increased bone formation following a brief endurance-type training intervention in adolescent males. J Bone Miner Res 12:1708 –1713 Frick KK, Jiang L, Bushinsky DA 1997 Acute metabolic acidosis inhibits the induction of osteoblastic egr-1 and type 1 collagen. Am J Physiol 272:C1450 –C1456 Brahm H, Piehl-Aulin K, Ljunghall S 1997 Bone metabolism during exercise and recovery: the influence of plasma volume and physical fitness. Calcif Tissue Int 61:192–198