Overtraining affects male reproductive status

FERTILITY AND STERILITY Copyright c 1993 The American Fertility Society Printed on acid-free paper in U. S. A. Overtraining affects male reproductive status Amy C. Roberts, M.A. t:j: Robert D. McClure, M.D. Richard I. Weiner, Ph.D. II George A. Brooks, Ph.D.t University of California, Berkeley, California, and University of California, San Francisco, San Francisco, California Objective: To substantiate the hypothesis that strenuous exercise disrupts the hypothalamicpituitary-gonadal axis in men. Design: Longitudinal study. Setting: Normal human volunteers in an academic research environment. Patients: Five endurance-trained men (maximum oxygen consumption 65.4 ± 3.6 ml/kg per minute [means ± SEM]) with normal spermatogenic and hormonal profiles. Interventions: Semen and blood samples were collected bimonthly before, immediately after, and 3 months after overtraining, which was defined as twice the previous average weekly training volume with unchanged intensity. Main Outcome Measure: Testosterone, cortisol, and sperm concentration. Results: Basal T levels decreased to 5.37 ± 67 ng/ml from 8.68 ± 93 ng/ml (conversion factor to SI unit, 3.47) immediately after overtraining and basal cortisol levels increas~d to 215.3 ± 31 ng/ml from 145.7 ± 27 ng/ml (conversion factor to SI unit, 2.76). This inverse relationship was highly correlated (r = -0.92). Both cortisol and T levels returned to pretraining values 3 months after resumption of previous training volume. Sperm count (91 ± 23.3 X 10 6 ) decreased significantly by 43% immediately after overtraining (52 ± 6.8 X 10 6 ) and by 52% 3 months after overtraining (44.5 ± 20 X 10 6 ). However, all values remained within normal range and would not be expected to affect fertility. Conclusions: Our results indicate that overtraining reduces T levels, which is highly correlated with an increase in levels of cortisol and possibly a subsequent decrease in sperm concentration 74 days later. Fertil Steril 1993;60:686-92 Key Words: Exercise, T, cortisol, sperm, hypogonadism The effects of physical exertion on reproductive function in women have been extensively studied Received January 25, 1993; revised and accepted June 4, 1993. * Presented at the 37th Annual Meeting ofthe American College of Sports Medicine, Salt Lake City, Utah, May 22 to 25, 1990. t Exercise Physiology Laboratory, Department of Human Biodynamics, University of California, Berkeley. :j: Reprint requests: Amy C. Roberts, M.A., Department of Human Biodynamics, 103 Harmon, University of California, Berkeley, CA 94720. Department of Urology, University of California, San Francisco. II Department of Obstetrics, Gynecology and Reproductive Sciences, University of California, San Francisco. (1, 2). However, there is scant information about the chronic effects of endurance training on reproduction variables in men. "Runner's amenorrhea" is a condition in which a woman ceases menses because of hormonal suppression of ovulation caused by physical exertion and stress (1,2). The complications of this condition include osteopenea (2), which make women prone to bone injury. Strenuous physical exercise alters the regulation of several neuroendocrine systems, resulting in decreased release of hypothalamic GnRH reducing gonadal hormone secretion (1, 3). Whereas alterations in the secretion of GnRH in women are manifested drama~ically and unequivocally as amenorrhea, the signs of hypogonadism in men are not as apparent 686 Roberts et al. Overtraining and reproductive status Fertility and Sterility

(4,5). This may explain why hypothalamic amenorrhea in women has been evaluated so extensively, whereas the effects of chronic physical exertion in men have been addressed in only a few studies, with inconclusive results. We asked whether a situation similar to women's amenorrhea exists in men, whether the implications of this situation are the same, and whether the mechanism of hormonal disruption is similar to that in women. To address these questions, we utilized an overtraining protocol standardized to previous training intensity. Because men do not have an obvious monthly cycle to indicate their hormonal status, blood profiles and semen analysis profiles were evaluated on a longitudinal basis. The purpose of the present investigation was to test the hypothesis that strenuous exercise disrupts the hypothalamic-pituitary-gonadal axis by increasing adrenal secretion of cortisol, which results in a lowering of T levels and a subsequent reduction of spermatogenesis. Subject Selection MATERIALS AND METHODS Five trained men (aged 24.8 ± 1.3 years) volunteered for the study. These subjects were engaged in routine endurance training (i.e., running, swimming, and cycling) for >4 days per week and were interested in the effects of overtraining on their reproductive hormones, body function, physical fitness level, and body composition. Subjects were recruited by posted notice and word of mouth. All subjects underwent a preliminary physical examination' provided a medical history, and were given a full explanation of the experimental procedures before giving informed consent. The experimental protocol was approved by the University of California at Berkeley, Committee for the Protection of Human Subjects (Protocol no. 88-10-49) and by the University of California at San Francisco, Human Subjects Committee (Protocol no. H5038-03700-01). Experimental Methods The duration of the present study was 12 months, a period adequate to analyze three 74-day spermatogenesis cycles. Reproductive variables were compared before a 2-week period of overtraining and immediately and 3 months after overtraining using each subject as his own control. Overtraining was determined from weekly activity diaries in which subjects recorded activity, mileage, and duration of activity. Mileage of activity was doubled, whereas intensity was held constant. At onset of study, maximum oxygen consumption (maximum V0 2 ) and physical examinations were obtained to asses eligibility. Before overtraining, semen samples were obtained bimonthly for 6 months after 2 days of abstinence from sexual activity. The mean of these samples obtained before overtraining served as the control. Samples were again obtained immediately after overtraining and again in 3 months. Blood samples were collected 1 month before overtraining and immediately and 3 months after overtraining. With this design, major changes after overtraining could be detected and minor daily fluctuations could be ruled out. Initial Graded Exercise to Maximum Test The first exercise test was designed to determine the subject's V0 2, resting heart rate, and recovery time and to substantiate a maximum V0 2 of >55 ml/kg per minute. Subjects were tested on a treadmil (model 65; Quinton, Seattle, W A) at a standard protocol that consisted of two 3-minute stages of walking and subsequent 2-minute stages that were increment in mileage. Once a pace of 8 mph was reached, the subsequent stages consisted of 2.5% grade increases. All trained subjects reached at least the 10th stage, which was an 8-mph pace at a 12.5% grade. Cessation of test was voluntary, and all subjects walked during recovery for 3 minutes and sat for a subsequent 3 minutes during which time recovery heart rates were recorded. Respiratory gas exchange was measured on -line using standard open circuit techniques as employed on a computerized system. Inspired flow was measure with a pneumotachometer (Fleich no. 3; Godard, Statham, Holland) and an MP45 pressure transducer and CD15 carrier modulator (Validyne, Northridge, CA). Expired gas was sampled continuously from a 5-L mixing chamber and passed through a heated line into AMETEK oxygen (S- 3A) and carbon dioxide (CD-3A) analyzers (Thermox Instruments Division, Pittsuburgh, PA). Analog signals from the three sensors were digitized on an IBM PC AT computer (Armonk, NY) that was also used for calculation of respiratory gas exchange parameters using the Haldane assumption of V0 2 equivalency in ventilation and data storage and retrieval. Body composition assessment was estimated by skinfold caliper method at six sites. This method has a 2% error variance (6). Roberts et al. Overtraining and reproductive status 687

Target Overtraining and Exercise Logs After qualification of training status, subjects received exercise logs and were advised to record type of exercise, duration, intensity, and estimated mileage of exercise on a daily basis. To standardize training intensities and to make the group more homogeneous, an overtraining protocol was employed as documented in previous reports (7-9). This protocol involved a 2-week overtraining period during which time all subjects doubled their average weekly mileage. After four sample collections at 6- to 8-week intervals, the subjects overtrained. The weekly average was calculated from their daily exercise log, and a personal double mileage protocol was determined for each subject. Second Graded Exercise To Maximum Immediately after overtraining (within 1 to 2 days), subjects had a second maximum oxygen capacity test to substantiate that they were in an overtrained state. The equipment and methodology is the same as was employed during the initial test. The index established to be indicative of an overtrained status includes an elevated resting heart rate, poor recovery time, and possible diminished maximum V0 2 (8). Resting Profile Procedures Subjects reported to our laboratory at approximately 7:00 A.M. after having fasted for 8 hours. A catheter was inserted into a forearm vein, and the subjects were left to rest quietly in the supine position. At 8:00 A.M., 20-mL heparinized samples were withdrawn. All blood samples were collected into sterile tubes. The samples were inverted and then spun at 2,500 X g for 20 minutes at 4 ac. Separated plasma was aliquoted and stored frozen at -80 a C until later analysis. Analytical Methods Spermatogenesis Profiles Sperm concentrations in seminal fluid were determined by a Coulter counter (Coulter Electronics, Hialeah, FL) from a fresh masturbation sample after 2 to 3 days abstinence from sexual activity (10). Hormonal Profiles Hormonal analyses were on plasma and used competitive binding RIAs based on the procedures of Midgley and Jaffe (11). The plasma was extracted with diethylether after the addition of a (1,2,6,7,-3H)-T internal standard, and the extract was chromatographed on a microcolumn of Sephadex-LH20. Standards and unknowns are equilibrated with antiserum to T-3-oxide-bovine serum albumin and the unbound steroid removed with Dextran-coated charcoal. The supernatant containing the bound fraction is then decanted into a vial for liquid scintillation counting. All assays on unknowns and standards were done in duplicate. Testosterone antibody (S205) was purchased from Guy Abrahams, M.D., University of California at Los Angeles Harbor Medical Center, Torrance, CA. Pooled sera were run at the beginning and end of each assay, and these data were used to calculate intra-assay or interassay coefficients of variation. Assays would be considered acceptable if controls were within two SDs of the mean of the duplicate samples. Statistical Analyses The one-factor analysis of variance with repeated measures with subsequent Scheffe' and/or Pearson post-hoc test was used to evaluate statistical significance of differences. Data are reported as means ± SEM. The baseline values obtained in men before overtraining were compared with values obtained immediately after overtraining and 3 months after training. In addition, hormone levels were intercorrelated. Individual data are displayed to illustrate individual variation and similarity in trends. Preliminary Data RESULTS Subjects weighed 77 ± 2.8 kg, of which 6.7 ± 1.4% was body fat, and had a fitness level of 65.4 ± 3.6 ml/kg per minute and a resting heart rate of 55 ± 4 bpm before the onset ofthe study. Individuals doubled the duration of at least two of their weekly training events, whereas intensity remained the same or was increased. All subjects followed this overtraining regimen for a two-week period. Hormone Concentrations Before and After Overtraining Testosterone The mean T level decreased 36% in all subjects immediately after overtraining (8.68 ± 0.93 ng/ml 688 Roberts et al. Overtraining and reproductive status Fertility and Sterility

10 400,=-0.92 8 iii' :z: 0 6 ffie Iii 0. 12" 4 III W t:. 2 300 ::7 ~E 1- a:'" 0" ~ 200 0 PRE 3 MONTHS Figure 1. Mean plasma T concentrations (conversion factor to SI unit, 3.47) before, immediately after, and 3 months after overtraining. *, significant difference between means (pretraining levels versus post-training levels and post-training levels versus 3 months post-training levels; P < 0.05). to 5.37 ± 0.67 ngjml [conversion factor for SI unit, 3.4 7]; P < 0.05). Three months after overtraining T returned to levels not significantly different from preovertraining values. Although there was considerable individual variability among hormone levels, we attempted to minimize these differences by collecting postabsorptive 8-A.M. blood samples via an indwelling catheter after a I-hour period of supine rest. Data are presented as means (Fig. 1). Cortisol The initial mean cortisol level, 144.9 ± 23 ngjml (conversion factor to SI unit, 2.76), increased 30% to 215.4 ± 31 ngjml immediately after overtraining (P < 0.05). Three months after overtraining, cortisol levels decreased to 144.6 ± 27 ngjml, a level not significantly different from preovertraining levels (Fig. 2). 300 200 2 OE!!!D> 1-" II: o ~ 100 O~--------LU~~~--~--- PRE POST 3 MONTHS Figure 2. Mean plasma cortisol concentration (conversion factor to SI unit, 2.76) before, after, and 3 months after overtraining. *, significant difference between the means (pretraining levels versus post-training levels and post-training levels versus 3 months post-training levels; P < 0.05). 3 4 5 6 7 8 [TESTOSTERONE) ng/ml Figure 3. The relationship between cortisol and T concentration immediately after overtraining. This correlation is almost linear (r = -0.92), implying (there may be a causative relationship between cortisol and T) cortisol secretion may affect T production. Relationship Between Cortisol and T Cortisol and T were negatively correlated immediately after overtraining (r = -0.92) (Fig. 3). The correlation between cortisol and T levels before overtraining was not statistically significant. Alteration in Semen Analysis 3 Months After Overtraining Sperm Concentration Because there was great individual variability in sperm concentration, the values before overtraining are the mean of four samples obtained bimonthly for 6 months. The initial mean sperm concentration decreased immediately after overtraining (91 ± 23.3 versus 52 ± 6.8 X 106; P < 0.01) (Fig. 4). Sperm concentrations of all subjects remained decreased 3 months subsequent to overtraining (44.5 ± 20 X 106). This represented a mean 48% decrease from initial value. DISCUSSION After overtraining cortisol and T levels, sperm concentrations, and semen characteristics measured in a group of highly trained athletes changed significantly relative to baseline values. As such, results of our investigation can be interpreted to mean that exercise overtraining exerts significant effects on both hormonal and semen reproductive 9 Roberts et al. Overtraining and reproductive status 689

120 40 20 PRE POST Figure 4. Mean sperm concentration before, after, and 3 months after overtraining. *, significant difference between means (pretraining concentration versus 3 months post-training concentration; P < 0.01). parameters in men. The reason that previous investigations of the effects of exercise on hormone levels in men resulted in contradictory and inclusive results is likely related to variations in experimental design on the intensity of the training parameters. Some studies contrasting trained and untrained men did not use a training protocol (4,5,10, 12-16). Therefore, differences observed may be due to factors other than training. Blood sampling procedures, particularly for cortisol, differ among studies. Often methods of analysis have not accounted for individual variability in representation of data. Overall, our results are similar to those obtained on women, in whom very strenuous endurance exercise regimens frequently lead to menstrual disorders (17). Cortisol Concentration Cortisol levels were significantly elevated after overtraining but were not different from control values by 3 months after overtraining. It is well established that cortisol levels increase in response to exercise (8, 18-21). The autonomic nervous system and the hypothalamic-pituitary-adrenal axis are modulated by stress and exercise and participate in regulating T levels (21). Consequently, Luger and co-workers (13) found elevated evening basal ACTH and cortisol levels in resting male long-distance runners. Androgen levels in these men were not reported; however, in amenorrheic female runners serum cortisol is elevated and cortisol production increased (12). T Levels Testosterone levels decreased after overtraining, but 3 months later levels were not different from control values. Reports of T concentration after endurance exercise have provided inconsistent results. Our results corroborate results of many studies that show T levels to be depressed after prolonged submaximal exercise (6, 10, 15, 21-25). Dessypris et al. (21) reported a 39% reduction, whereas Kuusi et al. (23) noted a 20% decrease in total T in men after a marathon. After an ultramarathon a 32% decrease in T was reported (16). With daily exercise, a 90% reduction of T levels was reported for military recruits during training (24). Also, our findings are in accordance with those of Hackney et al. (12), who found that T levels decreased in the beginning of moderate intensity training but subsequently returned to initial values in the last months of training. Testosterone decreased by 36% immediately after overtraining. Although we did not sample during the two-week overtraining period, we assume because of the intensity of the training during this period that T levels were decreased as well. This assumption is consistent with previous investigation. In a study of a 20-day 1,100 km road race by Schuermeyer et al. (10), T decreased 50% at rest in the morning by day 5 and remained at these reduced levels throughout the race. Wade et al. (16) studied runners who participated in a 15-day foot race and found a 31 % decrease over 15 days. We found T levels to return to normal 3 months after overtraining. Overtraining significantly decreased T levels during and immediately after overtraining. Possible Mechanism of Change in T Levels The mechanism for this decrease in T may be an increase in cortisol. This theory was postulated by Aldecreutz et al. (19), who suggested that a decrease in T is an overtraining syndrome, which in the presence of an increase in the concentration of catabolic steroids results in muscle atrophy. Many studies examined the roles of exercise-induced and psychological stresses and T production. Studies evaluating the effect of psychological stress on T regulation find T levels in army cadets to drop by 32% (24). Cumming et al. (18) studied changes in the levels of T in normal men in response to acute hypercortisolism induced by insulin hypoglycemia or an intravenous bolus of hydrocortisone. Cortisol had an inhibitory effect on testicular LH receptor concentration and T secretion by the Leydig cell. 690 Roberts et al. Overtraining and reproductive status Fertility and Sterility

Thus, hypercortisolism due to acute and chronic stress thereby inhibits testicular secretion by reducing the number of LH receptors on the Leydig cell. It is therefore possible that elevated basal cortisollevels as a result of overtraining may be associated with the altered hormone levels in athletes. This is consistent with our finding of a strong negative correlation between T and cortisol immediately after overtraining (Fig. 3). Because this correlation was absent before overtraining, the relationship is unlikely to be coincidental. It seems possible that the increase in cortisol as a response to stress negatively affected the production of T. However, the possibility that a precursor hormone, such as corticotropin-releasing factor, ACTH, GnRH, LH, or FSH, may.have a primary correlative relationship with this reduction in T is likely and cannot be addressed in the scope of this study. Sperm Concentration and Semen Characteristics All subjects' resting mean sperm concentration was within normal range, and no athlete was oligospermic nor were there any consistent abnormalities of motility or morphology in their semen samples. Immediately after overtraining sperm concentration decreased significantly. This was notable but could not be a result of depressed T levels because spermatogenesis requires 74 days. More interestingly' at 3 months after overtraining the sperm concentration remained significantly depressed. Our research is the first to report an effect of physical exertion on sperm concentration. Limited data are available that show effects of exercise on reproductive function in men. Ayers et al. (4) collected single semen samples from 20 marathon runners and found that the analysis of 18 of the 20 samples were normal. However, a single sperm count may not detect differences because of variation between subjects (25). Bagatell and Bremner (5) compared serial sperm concentrations and semen characteristics between 12 marathon runners and 12 age-matched lean controls over a period of 12 months. They found no differences between groups. Their finding may be due to the design of their study. Although subjects competed in marathons, there was no criteria by which to standardize their training intensity. Individual variability in sperm concentration is very great. If training intensities were not similar, individual variations in sperm concentrations may have been great enough to mask differences of mean values. Consistent with our data, Griffith et al. (9) showed a de- crease in sperm number and T immediately after overtraining. Again in this study the mechanism of the rapid decrease cannot be related to T levels because the hormonal effect on spermatogenesis would not be manifested for 74 days. Semen characteristics also changed after overtraining. There was a greater appearance of immature and nonviable sperm; however, the changes were not consistent among subjects. Our results indicate that very strenuous exercise not only rapidly affects sperm concentration immediately but also for 3 months after overtraining. This latter decrease is more likely the result of the decrease in T production. Possible Mechanism of Change in Sperm Concentration Testosterone is necessary for sperm to complete meiosis (3). Therefore, depressed levels of T may affect the number of mature sperm in the semen several months later. At this time, we found sperm number to be significantly reduced in the presence of normal levels of T. Semen characteristics were also changed: morphology was lessened, immature sperm were present, and there was more agglutination. However, the considerable variation between individual differences may have masked finding any significant changes. Our data are consistent with decreased T levels and lowered sperm concentration. It should be noted, however, that given their regular training regimens all subjects possessed normal to high sperm counts. Therefore, even though sperm concentration changed after overtraining, they declined from high to low values, all within the normal range. REFERENCES 1. Boyden TW, Pamenter RW, Stanforth PR, Rotkis TC, Wilmore JH. Impaired gonadotropin responses to gonadotropin-releasing hormone stimulation in endurance-trained women. Fertil Steril 1984;41:359-63. 2. Riggs BL, Eastell R. Exercise, hypogonadism and osteopenea. J Am Med Rec Assoc 1986; 256:392-3. 3. Hackney AC, Ness RJ, Schrieber A. Effects of endurance exercise on nocturnal hormone concentrations in males. Chronobiol Int 1989;6:341-6. 4. Ayers JWT, Komesu Y, Romani T, Ansbacher R. Anthropomorphic, hormonal, and psychologic correlates of semen quality in endurance-trained male athletes. Fertil Steril 1985;43:917-21. 5. Bagatell J, Bremner WJ. Sperm counts and reproductive hormones in male marathoners and lean controls. Fertil Steril1990;53:688-92. 6. Guglielmini C, Pasolini AR, Conconi F. Variations of serum Roberts et ai. Overtraining and reproductive status 691

testosterone concentrations after physical exercise of different durations. Int J Sports Med 1984;5:246-9. 7. Dressendorfer RH, Wade CE. The muscular overuse syndrome in long-distance runners. Physician Sports medicine 1983; 11:67-72. 8. Brown R, Burke ER. Overtraining of athletes. Physician Sportsmedicine 1983; 11:93-110. 9. Griffith RB, Dressendorfer RH, Fullbright CD. Effects of overwork on testosterone, sperm count and libido. Med Sci Sports Exerc 1988;20:S-39. 10. Schuermeyer T, Jung K, Nieschlag E. The effect of a 1100 km run on testicular, adrenal and thyroid hormones. Int J Androl 1984; 7:276-82. 11. Midgley AR Jr, Jaffe JB. Regulation of human gonadotropins. X. Episodic fluctuation of LH during the menstrual cycle. J Clin Endocrinol Metab 1971;33:962-9. 12. Hackney AC, Sinning WE, Bruot BC. Hypothalamic-pituitary-testicular axis function in endurance trained males. Int J Sports Med 1990; 11:298-303. 13. Luger A, Deuster PA, Gold PW, Loriaux DC, Chrousos GP. Hormonal responses to the stress of exercise. Adv Exp Med Bioi 1988;245:273-80. 14. MacConnie SE, Barkan A, Lampman RM, Schork MA, Beitins IZ. Decreased hypothalamic gonadtropin releasing hormone secretion in male marathon runners. N Engl J Med 1986;315:411-7. 15. Rogol AD, Veldhuis JD, Williams FA, Johnson ML. Pulsatile secretion of gonadotropins and prolactin in male marathon runners. J AndroI1984;5:21-7. 16. Wade CE, Dressendorfer RH, O'Brien JC, Claybaugh FR. Overnight basal findings during a 500 km race over 20 days. J Sports Med Phys Fitness 1982;22:371-6. 17. Lachelin GC, Yen SSC. Hypothalamic chronic anovulation. Am J Obstet Gynecol 1978; 130:825-31. 18. Cumming DC, Quigley ME, Yen SSC. Acute suppression of circulating testosterone levels by cortisol in men. J Clin Endocrinol Metab 1983;57:671-3. 19. Aldercreutz H, Haerkoen M, Kuoppasaimi K, Naveri H, Huhtaniemi I, Tikkanen H, et al. Effect of training on plasma anabolic and catabolic steroid hormones and their response during physical exercise. Int J Sports Med 1986; 7:27-8. 20. Kuoppasalmi K, Naeveri H, Haerkoenen M, Alderceutz H. Plasma cortisol, androstenedione, testosterone and luteinising hormone in running exercise of different intensities. Scand J Clin Lab Invest 1980;40:411-8. 21. Dessypris A, Kuoppasalmi K, Adlercreutz H. Plasma cortisol, testosterone, androstenedione, and luteinising hormone in a noncompetitive marathon run. J Steroid Biochem 1976;7:33-7. 22. Kreuz LE, Rose RM, Jennings JR. Suppression of plasma testosterone levels and psychological stress. Arch Gen Psychiatry 1972;26:479-82. 23. Kuusi T, Kostiainen E, Vartiainen E, Pitkaenen L, Ehnholm C, Korhonen HJ, et al. Acute effects of marathon running on levels of serum lipoproteins and androgenic hormones in healthy males. Metabolism 1984;33(6):527-31. 24. Aakvaag A, Bental 0, Quigstad K, Walstad P, Rinningen H, Fonnum F. Testosterone and testosterone binding globulin in young men during prolonged stress. Int J Androl 1987; 1:22-31. 25. Schwartz DD, Laplanche A, Jouannet P, David G. Withinsubject variability of human semen in regard to sperm count, volume, total number of spermatozoa and length of abstinence. J Reprod FertiI1979;57:391-5. 692 Roberts et al. Overtraining and reproductive status Fertility and Sterility