Influence of exercise duration on post-exercise steroid hormone responses in trained males
The purpose of this study was to systematically evaluate the effect of endurance exercise duration on hormone concentrations in male subjects while controlling for exercise intensity and training status. Eight endurance-trained males (19–49 years) completed a resting control session and three treadmill runs of 40, 80, and 120 min at 55% of VO2max . Blood samples were drawn before the session and then 1, 2, 3 and 4 h after the start of the run. Plasma was analyzed for luteinizing hormone (LH), dehydroepiandrosterone sulfate (DHEAS), cortisol, and free and total testosterone. LH was significantly greater at rest compared to the running sessions. Both free and total testosterone generally increased in the first hour of the 80 and 120 min runs and then showed a trend for a steady decline for the next 3 h of recovery. Dehydroepiandrosterone sulfate increased in a dose-response manner with the greatest increases observed during the 120-min run, followed by the 80-min run. Cortisol only increased in response to the 120-min run and showed a decline across time in all other sessions. The ratios of anabolic hormones (testosterone and DHEAS) to cortisol were greater during the resting session and the 40-min run compared to the longer runs. The results indicate that exercise duration has independent effects on the hormonal response to endurance exercise. At a low intensity, longer duration runs are necessary to stimulate increased levels of testosterone, DHEAS and cortisol and beyond 80 min of running there is a shift to a more catabolic hormonal environment.
KeywordsAndrogens Steroid hormones Running Endurance activity
Exercise is known to be a powerful stimulus of the endocrine system. In males, an acute bout of endurance exercise increases circulating testosterone and dehydroepiandrosterone (DHEA) (Galbo et al. 1977; Ponjee et al. 1994; Sutton et al. 1973) and regular exercise can influence the resting hormone profile (Hackney et al. 1998). Exercise-induced alterations in hormone concentrations and ratios can have a positive influence on muscle hypertrophy and strength (Tarpenning et al. 2001), which would be a desirable adaptation for athletes. However, there is cross-sectional evidence of depressed luteinizing hormone (LH) and testosterone levels in endurance trained males compared to untrained subjects (Hackney et al. 1998; Wheeler et al., 1984). Disruptions in the hypothalamic-pituitary-testicular axis could have negative implications for both musculoskeletal and reproductive health as well as performance (Hackney, 1996). As greater numbers of recreational athletes participate in long distance endurance events like the marathon, further research is needed to determine the effect of long duration endurance exercise on circulating anabolic hormone levels.
The hormone response to exercise is dependent on several factors including the intensity, duration, and mode of exercise and the training status of the subject (Galbo et al. 1977; Sutton et al. 1973; Tremblay et al. 2004). To date, most research has suggested that exercise intensity plays a more important role than duration or mode in determining the magnitude of the endocrine response (Jezova et al. 1985; Kuoppasalmi et al. 1980; Tremblay et al. 2004). Surprisingly, there has been little research attempting to isolate the effect of exercise duration on circulating anabolic and reproductive hormone levels in males. The majority of work examining the hormonal response to long-duration exercise has used a race or competition as the exercise stimulus, which introduces a number of confounding variables such as mental stress, environmental conditions, and fluctuating intensity.
Guglielmini et al. (1984) reported that testosterone levels increased in male subjects following a middle-distance competition and a marathon, but testosterone declined significantly after an ultramarathon. This suggests that there may be a duration threshold beyond which testosterone levels will begin to decline. Ponjee et al. (1994) demonstrated that testosterone, dehydroepiandrosterone sulphate (DHEAS) and cortisol all increased significantly following a marathon in male athletes. In contrast, Lutoslawaski et al. (1991) observed decreased testosterone levels following a 19-km and a 42-km kayak race. Snegovskaya and Viru (1993) found increased levels of cortisol, follicle stimulating hormone and LH in response to both short (7 min) and longer (40 min) bouts of intense rowing; however, they reported no coinciding change in testosterone levels. It is difficult to compare the results of these various studies due to the different types of competition and different modes of exercise (i.e. upper body versus lower body exercise) that were used, which could significantly affect the hormone response.
One of the few studies that examined the effect of duration of continuous exercise on testosterone levels in a laboratory setting was that of Duclos et al. (1996). They found that free testosterone increased in response to intense exercise, regardless of duration. Prolonged (120 min) low-intensity running had no effect on free testosterone or LH while prolonged high-intensity exercise produced increased testosterone levels that returned to baseline values within 30 min post-exercise. Duclos et al. (1996) did not evaluate any other anabolic/androgenic hormones or any exercise durations between 20 (brief) and 120 (prolonged) min.
The purpose of our study was to systematically evaluate the effect of increasing endurance exercise duration on hormone concentrations in male subjects while controlling for exercise intensity and training status. We hypothesized that a shorter bout of endurance exercise would stimulate an increase in anabolic hormone levels and result in a relatively anabolic hormonal milieu while longer duration exercise would result in a more catabolic hormonal environment.
Eight healthy, endurance-trained, males between the ages of 19 and 49 volunteered for this research. All subjects were running a minimum of 75 km per week at the time of testing and their training schedule remained unchanged during the study. Subjects kept a training diary, beginning 4 weeks before the start of testing and were asked to record changes in body mass or any stressful life events during the study period. Body mass was recorded at the start of each session.
All subjects were screened for contraindicating health problems or pharmaceutical use and were cleared for unrestricted physical activity by a physician. The experimental protocol was approved by the Institutional Review Board and all subjects gave written informed consent.
Baseline anthropometric and fitness measurements
Height, body mass, and skinfold thicknesses were taken following the protocols outlined in the Canadian Physical Activity, Fitness and Lifestyle Appraisal (Canadian Society for Exercise Physiology, 1998).
Maximal aerobic power was determined using a progressive, incremental treadmill protocol. Speed was held constant at each subjects’ self-determined pace and the grade of the treadmill (Model 24–72, Quinton Instruments, Seattle Washington) was increased 2% every 2 min for 8 min, then 1% every minute until volitional fatigue. Minute ventilation was determined by a Rayfield Airflow Meter (Rayfield Equipment Ltd., Waitsfield, Vermont) located on the inspired side. Gas analysis was done using Applied Electrochemistry analyzers (S-3A1 oxygen analyzer, CD-3A carbon dioxide analyzer, Ametek Inc., Pittsburgh, PA, USA). Minute ventilation (VE), oxygen consumption (VO2), carbon dioxide production (VCO2), and respiratory exchange ratio (RER) were monitored, as well as cumulative caloric expenditure based on RER and VO2 values.
During session 3, subjects completed a 40-min run at 50–55% of VO2max. The intensity was achieved by increasing level running speed. Expired gases were monitored continuously and cardiorespiratory and metabolic calculations were based on a 60-s sampling period. Heart rate was monitored using a telemetric heart rate monitor (Polar, Kempele, Finland).
80 and 120-minute runs
During sessions 4 and 5, subjects completed an 80-min and 120-min run in random order. The running speed for each individual was established in session 3 as the speed required to achieve 50–55% of their VO2max. The average running speed was 11.1 km hr−1 with a range of 10.1 to 11.5 km hr−1 . Expired gases were not collected for these sessions but heart rate was monitored continuously.
Blood collection and analysis
Figure 1 shows the timing of the blood samples during each session. Blood samples were collected through an indwelling venous catheter inserted in an arm vein, and 10 ml of blood was drawn for each sample. The catheter was maintained with a heparinized saline lock. Initial blood samples were drawn 30 min after catheter insertion. All samples were taken with subjects in a seated position, with the exception of the time two sample during the 80- and 120-min runs which was taken before the run was completed. All samples were drawn in a climate-controlled environment (21°C). Samples were analysed in duplicate for hematocrit, then the samples were centrifuged and the plasma stored at −80°C until assayed.
All plasma samples were analysed in duplicate. Commercial radioimmunoassays (RIAs) were used to analyse total testosterone (ICN Biomedicals Inc., Aurora, OH, USA), free testosterone (Diagnostic Products Corporation, Los Angeles, CA, USA), DHEAS (ICN Biomedicals Inc., Aurora, OH, USA), and cortisol (Kallestad Laboratories Inc., Chaska, MN, USA). Luteinizing hormone was determined by immunoradiometric (IRMA) assay (Diagnostic Products Corporation, Los Angeles, CA, USA). Duplicate samples with a coefficient of variation greater than 5% for the RIAs or 10% for the IRMA were reanalysed. Interassay variation, for the low and high controls respectively, was 5.9 and 0.3% for LH, 10.7 and 9.8% for DHEAS, 6.9 and 5.7% for cortisol, 13.3 and 5.5% for total testosterone, and 27.3 and 5.7% for free testosterone. The high variation in the low testosterone control was noted, however, there were no samples near the concentration of that control. To minimize the effects of inter-assay variation all samples from one subject were analyzed in the same assay.
All statistical analyses were performed using SAS (Version 6, SAS Institute Inc., Cary, NC, USA). Statistical significance was set at P<0.05 and all tests of significance were two-tailed. The statistical analyses were performed using two methods—a repeated measures ANOVA and Area Under the Curve. The repeated measures ANOVA was used to determine if there were significant differences among sessions across time. The area under the hormone-time curve (AUC) was calculated for each subject in each session and was used to examine the potential physiological impact of the series of measurements. A one-way ANOVA was used to look for differences in AUC results among sessions. Total testosterone/cortisol, free testosterone/cortisol and DHEAS/cortisol ratios were calculated and compared in the same manner as individual hormones. Wherever significant main effects were found, a Tukey post-hoc analysis was used.
The mean (SD) physical characteristics of the participants were as follows: age, 31.4 (9.7) years; height, 173.2 (5.5) cm; body mass, 69.9 (9.7) kg; and sum of five skinfolds, 44.0 (16.9) mm. Maximal aerobic power was high (67.1 (8.1) ml kg−1 min−1) as was expected for endurance trained men. Their training schedule was unaltered for the duration of the study and the diaries of stressful life events did not indicate any changes in psychological stress. Body mass did not change throughout the study.
In session 2 (40-min run), running speed was adjusted until the appropriate speed was established for each subject to maintain 50–55% of their maximal aerobic power. This same speed was then used for the subsequent runs of 80 and 120 min. Expired gases were not collected during the longer runs for the comfort of the subjects; however, there were no significant differences in mean heart rate values between sessions. The mean heart rate was 130.0±12.7 bpm during the 40 min run, 128.9±11.5 bpm during the 80 min run and 130.9±12.9 bpm during the 120 min run. The heart rate data confirm that the intensity of the exercise was constant across sessions.
Hormone responses to running
Absolute hormone concentrations during rest and three running sessions (Mean (SD))
LH (IU l−1)
DHEAS (μmol l−1)
Cortisol (nmol l−1)
Total T (nmol l−1)
Free T (pmol l−1)
40 min run
80 min run
120 min run
The results for free testosterone were similar to total testosterone however there were no significant session effects. The only time effect was observed during the 120 min run where time 1 was significantly greater than pre-exercise (time 0).
There was a significant session effect for cortisol as can be seen clearly in Fig. 4. Cortisol levels were greater during the 120 min run compared to all other sessions. This is supported by the AUC results which were significantly greater during the 120 min run compared to the resting session. As can be seen in Fig. 4, cortisol levels decreased over time in all sessions except the 120 min run.
Changes in plasma volume
As would be expected, the hematocrit levels were slightly greater during the longer runs compared to the shorter run and the resting session (mean HCT ± SD: rest=41.6±0.2; 40 min run=42.6±0.4; 80 min run=42.6±0.3; 120 min run=43.3±0.4). The AUC results for hematocrit, however, were not significantly different between sessions. When hormone data were analysed after correction for changes in plasma volume there was only one minor difference in the results (during the resting session cortisol at time 1 is then significantly less than pre-exercise). This indicates that the hormone responses to running observed in this study are not a result of hemoconcentration or hemodilution.
The purpose of this study was to evaluate changes in LH, testosterone, DHEAS and cortisol in response to running sessions of increasing duration, while controlling for exercise intensity and training status. These hormones were measured in eight male endurance athletes at rest and during recovery from a 40 min, 80 min, and 120 min run in a controlled laboratory setting. We used a different blood sampling protocol than is typically found in studies of this nature. By using standardized sampling times in each exercise session, we attempted to provide a “snapshot” of the hormone activities over a 4-h time frame. This protocol resulted in inconsistency of timing of the time 2 blood sample relative to exercise cessation, which we acknowledge may have affected the results. However, the dose-response patterns observed throughout the exercise duration comparisons suggest that our sampling procedures were sensitive to changes in exercise duration.
There was little change in free or total testosterone during the 40 min run or resting session. Both free and total testosterone levels showed an initial increase of about 20% in the first hour of the 80 and 120 min runs, with a subsequent decline that continued throughout recovery. By time 4 of the 120 min run the levels of total testosterone had decreased about 10% below the pre-exercise sample. Increased testosterone in response to running has been previously reported (Guglielmini et al. 1984; Ponjee et al. 1994; Webb et al. 1984) although none of these studies measured testosterone during recovery. Kuoppasalmi et al. (1980) reported a similar pattern and magnitude of testosterone changes in response to a high-intensity 45 min run with an initial increase in testosterone followed by a significant decline during recovery from the run. They found a smaller increase in testosterone after a longer (90 min) lower intensity run and a less-pronounced decline in recovery. The present study isolated the duration effect by comparing different runs of identical intensity and found that testosterone responds to increasing exercise duration in a dose-response fashion as well, with greater changes observed after the 80 and 120 min runs. In contrast, Duclos et al. (1996) found that free testosterone levels did not change in response to either 20 or 120 min of low intensity running. This difference in results is difficult to explain but may be related to the training status of the subjects. Duclos et al. (1996) used only four endurance trained subjects and another group of four sedentary men. The present study used only trained subjects.
The mechanism of the testosterone response to exercise is not clearly understood. Hemoconcentration was ruled out as a possible explanation for the initial increase in the present study, although decreased hepatic clearance is a possibility (Galbo et al. 1977). It has been suggested that elevated catecholamine levels during exercise may stimulate the testis and increase testosterone secretion (Galbo et al. 1977). The decline in testosterone levels that we observed during prolonged running and recovery is consistent with the results of Galbo et al. (1977), who found a decrease in testosterone during the latter part of an 80-min run. Galbo et al. (1977) suggested the decline in testosterone could be explained by a decrease in testicular blood flow as exercise is prolonged. It has also been proposed that an increase in androgen utilization to repair tissues following strenuous exercise might explain a decrease in circulating testosterone during recovery (Cumming 2000).
The LH tended to decrease during recovery from running and the reduction appeared greater with increasing duration. However this result was not significant (P=0.06) and should be interpreted cautiously due the pulsatile nature of LH secretion. Kuoppasalmi et al. (1980) found a dramatic (~45%) decrease in LH after both 45 min and 90 min runs, which is greater than the decline observed in the present study (~20–25%). However, Kuoppasalmi et al. (1980) classified exercise intensity on an absolute basis which makes it difficult to compare to our results. The increase in testosterone we observed during running appears to have occurred independently of LH.
The DHEAS increased significantly in response to running with the greatest increase (26%) occurring during the 120 min run. DHEAS declined during recovery from the 40 and 80 min runs but remained significantly elevated above pre-exercise levels 1 h after the 120 min run. Our previous work has found that the DHEAS response to exercise is intensity-dependent (Copeland et al. 2002; Copeland and Tremblay 2004; Tremblay et al. 2004) and the present results suggest the magnitude of the DHEAS response is also duration-dependent. This is supported by the previous results of Keizer et al. (1989) who found a progressively greater DHEAS response in males after running in progressively longer races. The mechanism for the increased DHEAS in the present study is not known; however, Keizer et al. (1989) believe that the magnitude of the exercise-induced increase in DHEAS is too large to be explained by simply a reduction in metabolic clearance and suggest that the primary mechanism is an increase in glandular secretion.
Cortisol showed a steady decline across time in all sessions except the 120 min run. There was no discernible exercise effect until beyond 80 min of running when a clear increase occurred in the second hour of exercise (Fig. 3). This is an interesting result since cortisol is generally believed to respond to exercise only above an intensity threshold of 60–70% of VO2max (Viru et al. 1996). We observed a considerable cortisol response (22% increase) at a low exercise intensity, but only after a significant volume of running. Viru et al. (1996) proposed that when exercise is performed under threshold intensity, hormonal responses will only occur when a certain amount of work is done. Our results support this theory and furthermore suggest that for cortisol, the duration threshold for adrenal activation is greater than 80 min at low intensity.
The ratios of anabolic hormones (testosterone and DHEAS) to cortisol were greater during the resting session and the 40 min run compared to the longer runs. This suggests the longer runs resulted in a more catabolic hormonal environment, which supports our hypothesis (Fig. 5). A lowered anabolic to catabolic ratio may be beneficial for recovery as it has been suggested that diminished testosterone and elevated cortisol will allow amino acids to be redirected from protein synthesis to gluconeogenesis (Duclos et al. 1996; Nindl et al. 2001).
The present study does include potential confounding variables. The age range of our subjects was relatively large (19–49), however, we do not believe this diminishes our results for a number of reasons. First, the inclusion of a resting session means that each subject essentially served as his own control. Also, we normalized our data which would minimize the effect of any baseline differences and we visually inspected the data prior to analysis and found no outliers that might be evidence of an age effect. It should also be noted the intensity of running was relatively low in this study. The increasing popularity of endurance events such as marathons reinforces the relevance of our study, and the reality is that many recreational athletes complete long- distance events at a low intensity. Therefore, the low intensity selected for this study increases the generalizability of our results.
The physiological consequences of the hormonal changes we observed after various bouts of running are unclear. It is not known if the transient endocrine changes after an acute bout of exercise have any lasting biological effects, although if one repeatedly applies an exercise stimulus (i.e. regular training) it seems likely that the endocrine response mediates some physiological adaptations. Lower testosterone levels and a more catabolic hormonal milieu are typically viewed as negative responses that will disrupt anabolic processes such as muscle growth and minimize training adaptations. In support of this theory, Tarpenning et al. (2001) found that intentionally blunting the cortisol response to exercise was associated with an increase in muscle fibre growth. Spermatogenesis and reproduction could also be negatively influenced by these hormonal changes (Hackney 1996). Conversely, this altered hormone profile following exercise could be seen as advantageous. As previously discussed it may permit the mobilization of fuels for recovery and restoration of glycogen stores. Others have suggested that lower testosterone levels could have a cardioprotective effect (Hackney 2001). Lower testosterone levels and reduced protein synthesis could also be a selective adaptation that is beneficial to endurance athletes by limiting the development of muscle mass (Hackney et al. 2003). Excessive muscle mass would hinder performance in long duration weight-bearing events, such as distance running.
In summary, when the effects of training status and exercise intensity are held constant, exercise duration has an independent, dose-response influence on the hormonal response to endurance exercise in trained males. At a low-intensity, longer duration runs are necessary to stimulate increased testosterone levels, although the response is transient. The adrenal cortex also responds to increasing exercise duration, as DHEAS showed the greatest increase after 120 min of running and cortisol only responded to the longest run. Beyond 80 min of running there is a shift to a more catabolic hormonal environment, although whether this will ultimately have negative or positive consequences is debatable.
- Canadian Society for Exercise Physiology (1998) The Canadian physical activity, fitness, and lifestyle appraisal, 2nd edn. Canadian Society for Exercise Physiology, OttawaGoogle Scholar
- Cumming DC (2000) The male reproductive system, exercise and training. In: Warren MP, Constantini NW (eds) Sports endocrinology. Humana, Totowa, pp 119–131Google Scholar
- Duclos M, Corcuff JB, Rashedi M, Fougere V, Manier G (1996) Does functional alteration of the gonadotropic axis occur in endurance trained athletes during and after exercise? A preliminary study. Eur J Appl Physiol Occup Physiol 73:427–433Google Scholar
- Hackney AC, Szczepanowska E, Viru AM (2003) Basal testicular testosterone production in endurance-trained men is suppressed. Eur J Appl Physiol 89:198–201Google Scholar
- Viru A, Smirnova T, Karelson K, Snegovskaya V, Viru M (1996) Determinants and modulators of hormonal responses to exercise. Biol Sport 13:169–187Google Scholar
- Webb ML, Wallace JP, Hamill C, Hodgson JL, Mashaly MM (1984) Serum testosterone concentration during two hours of moderate intensity treadmill running in trained men and women. Endocrinol Res 10:27–38Google Scholar