Both endurance training and acute endurance exercise seem to have a suppressive effect on the hypothalamic–pituitary–gonadal (HPG) axis in men. For example, lower basal circulating levels of testosterone have been reported in men who have performed chronic endurance exercise for many years.1 Additionally, in an acute long-duration event such as a marathon2^–5 or wrestling tournament, in which athletes undergo extreme physical stress,6 testosterone levels have been shown to decrease. These observed decreases in testosterone are typically seen when events exceed 3 h in duration,7 and remain decreased for up to 48 h.7 However, the cause of the decreased androgen levels is not completely understood. It is possible that these reduced levels of testosterone result from training-induced adaptations in the hypothalamic–pituitary axis at the central (that is, hypothalamus or pituitary) and/or peripheral (that is, disrupted testicular function) levels by alterations in the negative feedback loop that regulates production.8^–10 Testosterone production in the testes is primarily regulated by pituitary luteinising hormone (LH) produced in the pituitary. During marathon running and prolonged exercise, LH levels have been shown both to decrease4 11 and to remain unchanged.2 3 12 13 This discrepancy may be due to the pulsatile release of LH.

Furthermore, during ultra-endurance events stress hormones such as cortisol have been shown to significantly increase above baseline levels14 15 possibly caused by the onset of hypoglycaemia.16 Additionally, cortisol seems to be positively correlated with the duration of exercise.17 Cortisol may interfere with testosterone production, either acutely during an endurance event or chronically as a result of training. Cortisol production is stimulated by interleukin (IL)-6, a cytokine produced by contracting muscles during exercise to induce lipolysis, which may play a role in the testosterone production pathway. Growth hormone (GH) is released from the anterior pituitary gland during aerobic exercise.18 19 Like cortisol, GH release seems to be positively correlated with the duration of exercise.17 Although the primary function of GH is to stimulate growth, it plays an important role during endurance exercise in increasing fat mobilisation and decreasing carbohydrate metabolism.

Environmentally cold conditions present an additional stress to humans in maintaining thermoneutral internal temperatures. Hormones play an important role in thermoregulation.20 Thyroid hormones and noradrenaline are the hormones most responsible for the maintenance of the body’s internal temperature in response to cold conditions,20 but other hormones play a role in the physiological responses to this specific stress. During cold exposure, GH secretion is suppressed21 22 and cortisol secretion is increased if the exposure presents an adequate stress on the body,23^–25 whereas circulating testosterone is not changed.

The combined stress of the duration of an ultra-endurance event and the environmental stress of the cold has previously been shown to reduce serum sodium levels and haematocrit and plasma arginine, vasopressin and serum aldosterone.26 Furthermore, the opioid receptor system seems to regulate the physiological responses to exercise in thermally stressful environments.27 Nevertheless, responses of GH, cortisol and testosterone to these contradicting stresses is unknown and presents a unique summated stress to the athlete.

We examined an ultra-endurance race to evaluate pre-race and post-race responses of these hormones in endurance-trained athletes in the Susitna 100 race (formerly called the Iditasport 100). This is a 160-km human-powered (running or cycling) ultra-endurance race through the frozen wilderness in Alaska spanning an elevation gain of over 2000 m in freezing ambient temperatures while carrying survival gear. Prior research on this event has documented the physiological stress of this event. During this race athletes lose significant amounts of weight, comsume about 30 864 kJ of energy and experience hyponatraemia, decreased serum sodium, ketonuria and proteinuria.26 28 29 Not only does the race put these athletes through extreme physical stress during the race, but training for this extremely long event is rigorous, providing a unique opportunity to evaluate chronic hormonal adaptations and how these hormones respond to such extreme physical and environmental stress.

METHODS

Subjects

All 122 entrants in the 2000 Susitna 100 Human Powered Ultra-Marathon were invited to participate in the study at the mandatory informational meeting held 2 days before the race. In total, 16 male athletes (10 runners and 6 cyclists) from various parts of the USA volunteered to be subjects and signed a written consent document approved by the university internal review board. Each subject had understood the challenges of the race, and owing to the fact that acclimatisation is not a factor, had the clothing necessary for the race and also had travelled to the site before the race to allow for adequate preparation. Each had prepared for this ultra-endurance event. Owing to the field testing nature of the study to determine hormonal (primarily testosterone) levels in a field study, no geographical data or training data were collected.

Design

All pre-race measurements were made 2 days before the race at the informational meeting. Samples were obtained at a meeting the day prior to the race to approximate the same circadian time frame for their finishes. The cyclists and runners completed the same 160-km (100-mile) snow-packed course, which wound through the Alaskan wilderness and included an elevation gain of 2270 m. Ambient temperatures during the race ranged from −8°C to 4°C and wet snow fell a few hours after starting the race. Five checkpoints were located approximately every 15–20 miles (24–32 km), where food and fluid were available. In addition, athletes were required to carry 7 kg (15 lbs) of mandatory equipment at all times, including two litres of fluid in an insulated container and 3000 kcal of food, which was predominately (60%) carbohydrate. Post-race measurements were made within 15 minutes of each athlete completing the race.

Body weight and plasma volume changes

Pre-race and post-race weight was measured using the Tanita Body Fat Monitor/Scale (TBF-622), accurate to ±0.1 kg. Pre-race and post-race blood samples were collected by routine venepuncture, with athletes in a sitting position. Duplicate haematocrits were measured immediately on the samples using standard procedures from which changes in plasma volume were calculated according to the formula of van Beaumont:30

percentage change in plasma volume  = (100/100−hematocrit_pre) × 100 (hematocrit_pre−hematocrit_post)/hematocrit_post,

where hematocrit_pre and hematocrit_post are pre-race and post-race hematocrit samples, respectively.

Hormone analyses

Growth hormone was measured in duplicate using a double antibody ¹²⁵I radioimmunoassay (Nichols Institute Diagnostics, San Juan, Capistrano, California, USA) from serum that was obtained by centrifugation in Vacutainer serum separator tubes (Becton Dickinson and Co, Franklin Lakes, New Jersey, USA), frozen immediately on dry ice and stored at −20°C until thawed for analysis. Intra-assay variance was 4.5 (1.2)%. Circulating testosterone and cortisol were measured in duplicate using commercially available enzyme immunoassays (Diagnostics Systems Laboratories Inc., Webster, Texas, USA) from EDTA-anticoagulated plasma that was obtained by centrifugation in Vacutainer tubes, frozen immediately on dry ice and stored at −20°C until thawed for analysis. Intra-assay variance was 4.8 (1.3)% and 5.2 (1.2) for testosterone and cortisol, respectively. IL-6 was measured in duplicate using a quantitative sandwich enzyme immunoassay technique (R&D Systems Inc., Minneapolis, Minnesota, USA) from EDTA-anticoagulated plasma that was obtained by centrifugation in Vacutainer tubes, frozen immediately on dry ice and stored at −20°C until thawed for analysis. Intra-assay variance was 7.4 (2.2)%.

Statistical analyses

A one-way analysis of variance was used to evaluate whether differences in concentrations of growth hormone, IL-6, testosterone and cortisol existed pre-race between cyclists and runners, and to determine if any significant changes between pre-race and post-race concentrations of growth hormone, IL-6, testosterone and cortisol differed by method of transport. Pearson correlation analysis was used to determine relationships between pre-race cortisol and testosterone and post-race cortisol and testosterone. Using the nQuery Advisor software (Statistical Solutions, Saugus, Massachusetts, USA) the statistical power for the numbers used ranged from 0.75 to 0.92. Statistical significance was set at p⩽0.05.

RESULTS

The physical characteristics and race results for the runners and cyclists are presented in table 1. Subjects lost significant (p = 0.008) body mass between pre-race (mean (SD) 78.52 (8.32) kg) and post-race testing (76.91 (7.60) kg) corresponding to 2.05% body mass loss. There was no significant (p = 0.102) change in plasma volume.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Pre-race and post-race growth hormone and testosterone responses for the cyclists. Dashed line between black triangles, mean value; horizontal dashed line, normal values for men.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2 Pre-race and post-race growth hormone and testosterone responses for the runners. Dashed line between black triangles, mean value; horizontal dashed line, normal values for men.

No significant differences were found in the hormone levels between the cyclists and runners before and after the race. The results pre-race to post-race can be seen in table 2. Figure 3 shows mean pre-race and post race levels of testosterone, GH, IL-6 and cortisol for runners and cyclists. Correlation analysis did not reveal a significant relationship between pre-race cortisol and testosterone (r = 0.114, p = 0.674) and between post-race cortisol and testosterone (r = −0.399, p = 0.126).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3 Mean pre-race and post-race levels of (A) testosterone, (B) cortisol, (C) growth hormone, and (D) interleukin-6, for runners and cyclists.

DISCUSSION

The purpose of this study was to evaluate pre-race and post-race responses of testosterone, cortisol, GH and IL-6 in endurance trained cyclists and runners during the Susitna 100 ultra-endurance race under extreme conditions. Although we have no comparison conditions to determine the effects of such independent variables as temperature, sleep loss and altitude, these were all conditions related to the overall stress of the race, and our data point to the gestalt of the stressful conditions presented by the race. Overall, our data indicate that for both runners and cyclists, pre-race to post-race levels of cortisol, GH and IL-6 increased, whereas testosterone decreased. Additionally, pre-race levels of circulating testosterone were low (runners 12.32 (4.47) nmol/l: cyclists 13.81 (3.19) nmol/l) compared with normal reference values (14 to 28 nmol/l).31 This finding was consistent with previous research, which has shown that men who have performed chronic endurance exercise for many years have lower basal levels of free and total testosterone compared with age-matched sedentary men.1

The cause of these suppressed resting testosterone levels remains speculative. However, it is possible that the negative feedback loop of hypothalamic–pituitary unit is unresponsive to reduced levels of circulating testosterone. Pituitary LH regulates testosterone production, but the present study did not measure simultaneous circulating LH levels. However, Wheeler et al32 found that endurance-trained men with low basal testosterone levels do not seem to have raised circulating LH levels. Thus, the mechanisms of testosterone suppression remains uncertain, and the pulsatile release of LH further challenges this understanding.

Furthermore, Hackney et al33 found that men with low basal circulating testosterone showed a blunted response to an exogenous gonadotropin releasing hormone (GnRH stimulus), while testosterone production from the LH response seemed normal. It could also be possible that the number of testicular LH receptors on the Leydig cells may be reduced, resulting in reduced testosterone production.1 LH receptor number can possibly be reduced by persistent rises in circulating LH, resulting in downregulation of receptor number; by the presence of other hormones that can suppress testicular function; or by the thermic effects of exercise training.1 34 Additionally, raised basal levels of follicle-stimulating hormone (FSH) observed in endurance-trained athletes may provide further evidence of hypogonadism compensation due to intensive chronic training.15

Our data show no significant relationship between pre-race levels of cortisol and testosterone. Furthermore, pre-race cortisol levels of our endurance athletes were not above the normal range.31 Similarly, other researchers have found resting levels of cortisol in endurance-trained athletes to be similar to untrained people.35 36 Nevertheless, some researchers have speculated that raised cortisol levels, as a direct consequence of endurance training37 and more specifically overtraining,38 is a plausible mechanism to explain low basal testosterone. However, in our investigation, it is unlikely that our athletes were overtrained, and the training of these athletes was not a significant source of physical stress inducing raised resting cortisol levels. It has been proposed that these stress hormones may be responsible for the feedback disruption, as strong negative relationships have been observed between testosterone and cortisol.39^–41 Furthermore, it should be noted that there was a large variation in pre-race cortisol. The reason for this is possibly attributable to the pulsatile nature of cortisol release.42 43 Moreover, cortisol release is highly responsive to physiological stress, nutrition and exercise status,44 45 sleep46 and environmental conditions,23 24 which may have posed varying levels of physiological stress for individual athletes.

The race imposed a significant physical and metabolic stress, as indicated by the hormonal responses. Post-race testosterone levels significantly declined, whereas cortisol, GH and IL-6 significantly increased relative to pre-race values. Several feasible mechanisms may explain the decreased post-race levels of testosterone, where either the rate of testosterone utilisation increased to exceed production during the race to preserve protein tissue or the rate of production decreased during the race because of inhibitory mechanisms. At the anterior pituitary (central) level of the HPG axis, the pulsatile release of LH may have decreased or become less frequent because of competing mechanisms while GH production was preferentially increased for immediate exercise effects to provide energy. McColl et al11 showed that exercise induces a general lowering of LH levels but does not inhibit LH pulsatile release. Additionally, blood flow and substrate precursor availability (eg cholesterol, pregnenolone) to the testes may have decreased,1 limiting the rate of testosterone production at the peripheral level of the HPG axis. These plausible alterations in blood flow in the testes may affect β-endorphin and nitric oxide mechanisms related to testosterone secretion.47^–49

Cortisol was possibly raised after the race to maintain plasma glucose level, a response that is commonly observed in endurance exercise.50 Cortisol has been shown to reduce testosterone level when it is directly infused into men. The mechanism for this effect is possibly due either to its direct inhibition of testosterone production by the Leydig cells or to feedback disruption of the HPG regulatory axis. However, our data showed no significant correlations between levels of cortisol and testosterone after the race. Several factors could explain this null finding in hormonal responses, including large variations among the athletes, the time course and/or the magnitude of individual stress responses. Similarly, Daly et al40 showed no significant negative relationships between cortisol and free testosterone in endurance-trained men exercising to exhaustion at 100% of their ventilatory threshold.

GH has been shown to facilitate both the glucose regulatory and protein synthesising actions of cortisol, hence the corresponding increase. Furthermore, other researchers have noted increased GH levels with prolonged running.51 Additionally, IL-6 has been shown to increase in response to exercise52 and low skeletal muscle glycogen stores.53 IL-6 is released from active skeletal muscle to mobilise extracellular substrate oxidation rate (via enhanced lipolysis)54 and/or augment substrate delivery during exercise.55 IL-6 release from the exercising muscle possibly signals the liver to increase its glucose output to preserve blood glucose levels during exercise,56 which plausibly explains the increase seen in these athletes after the race.

In summary, these data provide specific hormonal information on athletes undergoing extreme physiological stress during an actual competition. Data suggest possible suppression of the HPG axis by an ultra-endurance race under such extreme conditions and support the observation that endurance-trained men show lower basal levels of testosterone compared with normal healthy non-endurance-trained males.