18 September 2014

Testosterone and mountain ultramarathon running

This past weekend I put on my running shoes for the first time since Petite Trotte à Léon (PTL) . Despite only jogging slowly for 7 kilometers (4.3 miles) I developed a massive delayed onset muscle soreness in my quadriceps. It felt like I had lost a lot of my muscle mass in my legs, and, thinking about it and looking at my legs I think that might actually been what has happened. I probably literally consumed my own leg muscles due to the catabolic energy state I was in during and immediately after the race. This made me think more about the wasted state I have been in during the past two weeks and what hormonal changes that could lead to this. Looking at my general symptoms of tiredness, poor concentration, lack of vigor and vitality, massive night sweats, occasional hot flushes and sleep disturbance I quickly focused in on low testosterone. Other cardinal symptoms of low testosterone levels include loss of morning erection and reduced sexual desire/loss of libido. Thinking back I have at least to some extent suffered from partly these same symptoms transiently for a while after Tor des Geants (TDG) and other longer ultramarathons I have done in the past. The other day I read John Burton's great race report from Tahoe 200 and in the aftermath he appears to have suffered from some of the same symptoms as well and I have had other runners telling me similar stories.

The suspicion that ultramarathon running might cause low testosterone levels made me read more about the subject and, regretfully, it appears at least from the few studies conducted so far that longer ultramarathon running indeed can lead to testosterone deficiency and even secondary hypogonadism in males. In this post I will review the effect of ultramarathon running, in particular in mountains, on testosterone and partly also other hormones and substances in the hypothalamic-pituitary-gonadal/testicular (HPG/HPT) axis (see figure below). This will be a post focused on the effects in men, mostly as the amount of literature on the effects of training on the female reproductive system is very extensive and clearly requires a separate review.

The hypothalamic-pituitary-gonadal (HPG) axis
Testosterone levels following Western States Endurance Run (WSER)

The first article I stumbled upon was a recent study by Kupchak and colleagues of 12 male runners completing the 161 km (100 miles) race Western States Endurance Run (WSER) (Kupchak et al “The impact of an ultramarathon on hormonal and biochemical parameters in men” Wilderness Environ Med 2014; 25: 278-288). The 12 runners ran an average of 98.7 (81.9 -115.5) km per week before the race and had a mean finish time of 25.08 (22.53 – 27.62) hours. Quite interestingly, 8 out of 12 runners had prerace values of testosterone lower than the reference limits in their assay (< 14 nmol/L). Still the values of both testosterone and luteinizing hormone (LH) decreased even further from these low baseline values during the race and were significantly lower both immediately and one-day after the race compared to the values before the race. Also sex hormone-binding globulin (SHBG) decreased during the race, while cortisol, a hormonal marker of stress, increased significantly leading to a significant and markedly lower Testosterone : Cortisol ratio. One of the problems with the study is that the blood collection for analysis occurred at various times in association with the finishing of the race and hormones such as testosterone exhibit a clear diurnal variation. Also, the authors used an immunoassay rather than today’s golden standard mass spectrometry in measuring the levels of testosterone. Still, the magnitude of change observed was clearly well beyond what could be explained by circadian undulation and the levels were clearly indicative of testosterone deficiency.

Other studies of testosterone levels following marathon and ultramarathon races

Kupchak’s study was clearly not the first in this area as there have been a large number of studies published since the beginning of the 1980’s showing statistically significant reductions of testosterone levels in blood and saliva during a marathon or ultramarathon race compared to the pre-race levels (see table below which include some selected studies I could find).

Distance (km)
Kupchak 2014
161 (WSER)
Wilderness Environ Med 2014; 25:278
Tauler 2014
Appl Physiol Nutr Metab 2014; 39: 560
Kraemer 2008
161 (in cold)
Br J Sports Med 2008; 42: 116
Karkoulias 2008
Eur J Intern Med 2008; 19: 598
Ishigaki 2005
284 (over 8 days)
J Phys Antl Appl Hum Sci 2005; 24: 573
Fournier 1997
Int J Sports Med 1997; 18: 252
Marinelli 1994
42.2 (high altitude)
Horm Res 1994; 41: 225
Dressendorfer 1991
400 (over 15 days)
Med Sci Sports Exerc 1991; 23: 954
Raschaka 1991
1000 (over 20 days)
Z Ernahrungswiss 1991; 30: 276
Pestell 1989
Clin Exp Pharm Physiol 1989; 16: 353
Tanaka 1986
J Endocrin Invest 1986; 9: 97
Kuusi 1984
Metabolism 1984; 33: 527
Schürmeyer 1984
1100 (over 20 days)
Int J Androl 1984;7:276

Testosterone levels in other endurance sports

This is in contrast to shorter runs of less than one hour where there actually appears to be a rise in testosterone levels during the actual run (see for instance Tremblay et al “Influence of exercise duration on post-exercise steroid hormone responses in trained males” Eur J Appl Physiol 2005; 94: 505 and Hackney et al “Testosterone responses to intensive interval versus steady-state endurance exercise” J Endocrinol Invest 2012; 35: 947). It is, however, in line with findings of a decrease in testosterone during longer endurance races in other sports. For instance during adventure racing for over 6 days (Berg et al Scand J Med Sci Sports 2008; 18: 706), cross-country skiing for 75 km and road bicycling for 4 hours (Vasankari et al Acta Endocrinol 1993; 129: 109), an Arctic ski expedition (Bishop et al Acta Astronaut 2001; 43: 261), Ironman triathlon (Neubauer et al Eur J Appl Physiol 2008; 104: 417) and military special operations training for 5 days (Opstad et al Eur J Appl Physiol Occup Physiol 1982; 49: 343) marked decreases in testosterone have been observed. All of these studies in both running and other sports have only looked at hormonal levels and not any symptoms or behavioral changes associated with these changes, like lowered sex-drive, and there is clearly a need for studies investigating this.

Baseline levels of testosterone in runners depending on weekly mileage

There have also been a number of studies performed looking at the normal baseline testosterone levels of runners and some studies have specifically analyzed the correlation between these levels and the weekly running mileage. In an important study of 53 runners published already in 1992, MacDougall and colleagues showed a correlation between weekly running mileage and testosterone levels (MacDougall et al “Relationship among running mileage, bone density, and serum testosterone in male runners” J Appl Physiol 1992;73: 1165). De Souza and colleagues published another similar study in 1994 where they looked at baseline testosterone levels in 11 high mileage runners (108 ± 4.5 km/week), 9 moderate mileage runners (54 ± 3.7 km/week) and 10 sedentary controls of similar age (28.3 ± 1.5 years) (De Souza et al “Gonadal hormones and semen quality in male runners. A volume threshold effect of endurance training” Int J Sports Med 1994; 15: 383). Interestingly they found a dose-response effect with regards to running volume as the levels of both total testosterone and free testosterone were lower in the high mileage group compared to the other groups. This also correlated with decreased sperm quality in the high mileage group. A similar correlation was found by MacKelvie and colleagues studying 5 high mileage runners (> 95 km/week) and 7 moderate mileage runners (64-80 km/week) and comparing them to sedentary controls (MacKelvie et al “Bone mineral density and serum testosterone in chronically trained, high mileage 40-55 year old male runners” Br J Sports Med 2000; 34: 273).Comparing running with endurance training one study showed that it is only the former where low baseline values of testosterone can be found together with subclinical modifications in semen characteristics (Arce et al “Subclinical alteration in hormone and semen profile in athletes” Fertil Steril 1993; 59: 398).

These findings are in contrast to other studies, for instance by Ayers et al in Fertil Steril 1985; 43: 917, Bagatell et al in Fertil Steril 1985; 43: 917 and Cooper et al in Eur J Endocrinol. 1998; 138: 517, where no low baseline levels of at least free testosterone could be observed in runners. However, these studies did not include high mileage runners. A drawback with all of these studies is also that they are retrospective. A recent interesting prospective randomized controlled trial from Iran of 286 “habitual aerobic exercisers”, training for an average of 1.8 h per day 5 days a week before entering the study, looked at the effects of a 60-week training program where the subjects ran for 120 minutes 5 times a week at either ~60% of VO2max or ~80% of VO2max (Safarinejad et al “The effects of intensive, long-term treadmill running on reproductive hormones, hypothalamus-pitutary-testis axis, and semen quality: a randomized controlled study J Endocrinol 2009; 200: 259). In both groups there was marked decrease of testosterone levels and semen quality and these changes were observed already from 12 weeks into the study. That longer period of increase in training appears required for a decrease in testosterone is indicated in a study by Hall et al where 2 weeks of running exercise at 186% of normal training intensity was not enough to lead to any hormonal changes (Hall et al “Effects of intensified training and detraining on testicular function” Clin J Sport Med 1999; 9: 203). Also, it appears again that the training intensity needs to be quite high to lead to testosterone decrease as a prospective study of 24 marathon runners did not show changes in this hormone, despite changes in semen parameters (Jensen et al in Fertil Steril 1995; 64: 1189).

Testosterone levels following tapering and following overtraining/overreaching in runners

Quite interestingly, following cessation of the training program in the Iranian study the testosterone levels rather quickly returned to the same values, or even somewhat higher, than before the start of the training period. This is in contrast to a study by Houmard and colleagues where 10 runners were followed for 4 weeks of normal training (81 ± 5 km/week; 6 days/week) followed by a 3 week tapering period with a 70% training reduction (to 24 ± 2 km/week; 5 days/week) (Houmard et al “Testosterone, cortisol, and creatine kinase levels in male distance runners during reduced training” Int J Sports Med 1990; 11: 41). The low testosterone baseline values observed during the training period were in this study not restored by the reduction in training. I clearly think more studies are needed on the effect of tapering on testosterone and other hormonal levels. Another subject were more studies are needed is the role of testosterone and the testosterone/cortisol ratio in overtraining syndrome (OTS) / overreaching (OR). The current consensus standpoint appear to be that low testosterone levels and a low testosterone : cortisol ratio only indicates actual physiological strain and cannot be used for diagnosis of OTS / OR. While I agree with this, I still think it too early to completely rule out a role for testosterone in the development of OTS /OR and further studies are needed to see if this indeed might be the case. Also, another question I have thought about is whether fluctuations in testosterone levels that might occur after endurance racing and in tapering could influence mood (swings) and well-being, something which in numerous other studies have been shown to be influenced by testosterone. I am sorry to be repetitive when I again ask for studies of this in runners.

Testosterone levels at high altitude

Mountain ultramarathons like Petite Trotte à Léon (PTL) and Tor des Geants (TDG) are not only long in terms of kilometers and race time, but also occurs at comparatively high altitudes and are associated with a marked degree of sleep deprivation in most runners at the later parts of the race (Saugy et al “Alterations of neuromuscular function after the World's most challenging mountain ultra-marathon" PLoS One 2013; 8: e65596). Both of these factors have actually independently been associated with a decrease in testosterone levels. Studies of several expeditions to the Himalayas have revealed markedly decreases in testosterone levels coupled with changes in the sperm quality (see for instance Okumura et al High Alt Med Biol 2003; 4: 349; Benso et al Eur J Endocrinol 2007; 157: 733 and Pelliccione et al Fertil Steril 2011; 96: 28). Interestingly, a study by Mirsepasi and colleagues show a decrease in testosterone after only 30 minutes running at 70% of maximal heart rate already at an altitude of 3250 meters (Mirsepasi et al “Effect of submaximal aerobic exercise at the altitude of 3250 meters on levels of serum cortisol, testosterone and testosterone to cortisol ratio in active young men” Adv Environment Biol 2013; 7: 854).

Testosterone levels following sleep deprivation

A number of studies also show that total sleep deprivation leads to a reduction of testosterone levels (for instance Schmid et al Clin Endocrinol 2012; 77: 749 and Jauch-Chara et al PloS One 2013; 8: e54209). The production of testosterone is dependent on sleep and requires at least 3 hours of deeper so called nonrapid eye moment (NREM)  slow wave sleep (SWS), which normally occurs in the first part of a sleep episode (reviewed in Wittert “The relationship between sleep disorders and testosterone in men” Asian J Andrology 2014; 16: 262). This is interesting as at least I have had the strategy to not sleep for more than 2 hours at each occasion in both TDG and PTL and that would therefore not be enough to stimulate a normal testosterone production.

The Exercise-Hypogonadal Male Condition and the effect of running on male fertility

One of the researchers who has published perhaps most scientific papers on the subject of male reproductive dysfunction following endurance training is Anthony C. Hackney at University of North Carolina in Chapel Hill in the USA. He has coined the syndrome “Exercise-Hypogonadal Male Condition” or EHMC (reviewed in for instance Lane & Hackney “Reproductive dysfunction from the stress of exercise is not gender specific: The “Exercise-Hypogonadal Male Condition” J Endocrinol Diab 2014; 1: 4). That endurance training, and in particular ultra-endurance training, might lead to impaired fertility in males have been discussed also in other review articles, for instance by Arce and colleagues (Acre et al “Exercise and male factor infertility” Sports Med 1993; 15: 146), Brandt and colleagues (Brant et al “Male athletic activities and their effects on semen and hormonal parameters” Phys Sportsmed 2010; 38: 114), du Plessis and colleagues (du Plessis et al “Is there a link between exercise and male factor infertility?” Open Rep Sci J 2011; 3: 105) and Vaamonde and colleagues (Vaamonde DM et al “The impact of physical exercise on male fertility” in the book Male Infertility 2014; 47-60 edited by du Plessis et al). Nevertheless, there are no good controlled studies of this and further studies are clearly needed to determine whether male ultramarathon runners really have impaired fertility in general.

What are the mechanisms behind the low testosterone levels following an ultramarathon?

Today we can only speculate why prolonged endurance activities such as an ultramarathon might lead to lower testosterone levels. Testosterone is produced in the testis and it has been speculated that reduced testicular blood flow or damaging testicular heating might influence the production. Others have advocated the increase in testosterone/androgen utilization in the repair of tissues might lead to increased consumption of testosterone leading to lower blood levels. Also the stress of prolonged endurance activities might lead to an inflammatory reaction acting at various levels of the HPG-axis. The response of the body to an ultramarathon might not only be a reactive response to damaging stimulus, but also a way for the body to protect itself in the longer run and it has been speculated that one positive effect of low testosterone levels in male ultramarathon runners is that it limits development of excessive muscle mass. Also, during the actual race it might lead to mobilization of amino acids for energy consumption through gluconeogenesis, hence leading to use of the muscles for energy. I have encountered some athletes, in particular in adventure racing, advocating  muscle strength training sessions in association with the endurance race or early during the recovery phase and it would be interesting to study whether this could counteract some of these effects and actually raise the testosterone levels.

Negative effects of low testosterone levels in male ultramarathon runners

No studies have been conducted on possible negative medical effects of low testosterone levels in male ultramarathon runners besides some studies showing that, rather surprisingly, low baseline levels of testosterone appear not to be associated with a decrease in bone mineral density (BMD) (see references above to MacDougall et al 1992 and MacKelvie et al 2000). Low testosterone levels and hypogonadism in middle aged and elderly men appear quite common and have been associated with male infertility and the symptoms I mentioned in the beginning of this post (reviewed in Wu et al “Identification of late-onset hypogonadism in middle-aged and elderly men” N Engl J Med  2010; 363: 123; Basaria “Male hypogonadism” Lancet 2014; 383: 1250, Finkelstein et al “Gonadal steroids and body composition, strength and sexual function in men” N Engl J Med 2013; 369: 1011 and Hackett et al “Testosterone deficiency, cardiac health, and older men” Int J Endocrinol 2014; 143763). There are also indications that low testosterone levels might lead to increased incidence of cardiovascular diseases and even higher mortality, but in the studies conducted so far there are a number of co-founding factors such as presence of type 2 diabetes and it is difficult to say what is the hen and the egg in these studies. There is clearly a need to see study if there are any long-term health effects, if any, of low testosterone levels in trained male ultramarathon runners and it will for instance be interesting to see whether there could be a correlation between testosterone levels, weekly mileage and cardiac health in runners. In these studies it would have been interesting to see the effects of testosterone replacement therapy, which is becoming increasingly popular in hypogonadism in elderly men, but this would of course be impossible and not ethical as testosterone in all forms rightly is classified as a prohibited performance enhancing drug and is on all doping lists.


In summary, I think that it has been shown quite convincingly in a number of studies that prolonged endurance activities can lead to a decrease in testosterone levels. In longer mountain ultramarathon races also the high altitude and the sleep deprivation of the runners might further contribute to this decrease. Even though some studies indicate that runners with a high weekly training mileage have lower baseline testosterone levels than the normal aged matched population, I still think further studies are needed to determine whether ultramarathon induced hypogonadism really exists. The long-term health effects of decreased testosterone in ultramarathon runners are also not known today.  


  1. So, what food at aid stations, if any, can I eat to counter this?

  2. Hi Peter,

    Sorry for late reply. I wish I knew, and the influence of nutrition during ultramarathon running on testosterone levels has not been studied to my knowledge. However, I think it reasonable that you can counter some parts of it, at least for "shorter" ultramarathons up to 100 miles, by keeping as good an energy balance as possible as discussed in a previous blog post http://summitrun.blogspot.se/2014/05/nutrition-carbohydrates-and-mountain.html . In speculation, for longer runs >100 miles I am less certain of the value of primarily carbohydrates and it might be so that proteins and amino acids become more important, but, as mentioned, it has not been studied to my knowledge.

  3. Great article! My friend worried me telling me that my running (40 to 60 km per week) for halfmarathon and marathon (once per year) is going to hurt my muscle mass and testosterone... I was freakin. But as I understand from your story this is only case in long distances runners. Right?

  4. Hi Claudio,

    Thanks! And thanks for question. Yes, I think you are correct that your running training and racing volume should not be a problem based on the knowledge available right now referred to in the post.

  5. So 25 miles of intensive running every week is ok right?

  6. Do you think walking would lower t levels like ten miles a day

    1. Probably not - however depends on many factors such as previous training, other training, weight, age etc so is impossible to say for sure.

  7. Hi Peter, great post.

    A couple weeks ago I ran my first ultramarathon (~85k), after ~3 months of training 80-90k a week.
    I've noticed my penis feels slightly smaller, and even my GF noticed it.

    Is this usual in low-testosterone levels? Should I go to the doctor? Or it will pass on its own after reverting to ~40k/week?

    1. First of all, I am not giving any kind of medical advice here and for this you definitively need to go to your doctor.

      Perceptions like yours have been described in association with low testosterone levels. To me it is however unclear whether your running has caused lowering of your levels, if any, and you of course need to measure the levels in order to be certain that they are low. Your mileage is still quite moderate and the time you increased it over is quite long so the body might have adapted well. It is neither not known whether the levels will adjust over time to the new mileage level, something I think would be likely, and more studies are clearly needed here. That would mean that there would be no need to revert to lower mileages.

      In your case I would try different mileages and see how you respond, just bearing in mind that it make take quite long time for the body to adapt to the changes.