Various Factors May Modulate the Effect of Exercise on Testosterone Levels in Men

Exercise has been proposed to increase serum testosterone concentrations. The analysis of existing literature demonstrates a large degree of variability in hormonal changes during exercise. In our manuscript, we summarized and reviewed the literature, and concluded that this variability can be explained by the effect of numerous factors, such as (a) the use of different types of exercise (e.g., endurance vs. resistance); (b) training intensity and/or duration of resting periods; (c) study populations (e.g., young vs. elderly; lean vs. obese; sedentary vs. athletes); and (d) the time point when serum testosterone was measured (e.g., during or immediately after vs. several minutes or hours after the exercise). Although exercise increases plasma testosterone concentrations, this effect depends on many factors, including the aforementioned ones. Future studies should focus on clarifying the metabolic and molecular mechanisms whereby exercise may affect serum testosterone concentrations in the short and long-terms, and furthermore, how this affects downstream mechanisms.


Introduction
Testosterone is the most potent naturally secreted steroid androgenic hormone. It is required for promotion of secondary male-sex characteristics, as well as muscle growth and neuromuscular adaptation [1]. At the muscle level, testosterone is known to exert its anabolic effect via the following two mechanisms: (a) stimulating amino acid uptake and protein synthesis; and (b) inhibiting protein degradation by counteracting cortisol signaling [1]. Age, higher body weight, poor nutritional status, stress, sleep deprivation, and alcohol consumption are known physiological factors leading to lower serum testosterone concentrations. [2,3]. Low plasma testosterone concentration is associated with fatigue, sexual dysfunction, depressed mood, difficulty concentrating, and hot flashes [4]. If left untreated, patients may develop anemia, low bone mass density (i.e., osteoporosis), higher pro-atherogenic lipoprotein-associated changes [5], and muscle wasting [4]. Thus, maintaining physiological levels of testosterone has significant health benefits.
Exercise has significant health-related benefits and it is proposed to increase plasma testosterone concentrations [6]. However, analysis of the existing literature demonstrates a large degree of inter-individual and inter-study variability in hormonal changes during exercise. Age, body weight, and exercise type, together with exercise intensity, volume, and the involved muscle type, were studied as factors modulating these hormonal changes. This review article intends to clarify the factors that contribute to the variability in serum testosterone concentrations during exercise, and the underlying mechanisms. Part 1 will focus on the acute or immediate post-exercise changes in plasma testosterone concentrations, and Part 2 will discuss the changes in basal or resting plasma testosterone concentrations after completion of exercise protocols.
An online search through the Pubmed and Medline databases was initially performed using the combination of the following keywords: "testosterone", "exercise", and "men". Additional exercise description such as "type" and "intensity" were added. "Obesity" and "age" as key words were added during advanced search for population-focused data. The exclusion criteria included: publications written in languages other than English, publications involving subjects with chronic medical conditions, such as congestive heart failure and diabetes, and publications involving subjects on testosterone replacement. Table 1 contains a brief summary of the analyzed data on the effect of exercise on serum testosterone concentrations. Table 1. Summary of analyzed data on the effect of exercise on testosterone response.

Exercise Variable Factor Population (Men) Exercise Description Results Reference
Endurance 5 untrained but active Age 25 ± 1 year 60 RPM, cycle ergometer. Load increased by 25 w/min following the first 3 min at zero load, until the subject was no longer able to continue TT significantly increased at 0, 5, 10, 15, 20, 25 min during exercise, then start declining at 25 min and no difference afterwards [6] Intensity, duration 7 untrained Aged 23-26 years 3 bicycle ergometer tests A, B, C of the same total work output but different intensity and duration. Tests A and B: 3 consecutive exercise bouts, lasting 6 min each, of either increasing (1.5, 2.0, 2.5 W·kg −1 ) or constant (2.0, 2.0, 2.0 W·kg −1 ) workloads, respectively Test C: 2 exercise bouts each lasting 4.5 min, with workloads of 4.0 W All the exercise bouts were separated by l-rain periods of rest Significant increases in TT were seen only in the high intensity exercise group C [7] Resistance, intensity 9 healthy Age 24.66 ± 4.27 years High resistance exercise protocols control for load ( Each group exhibited an increase (p < 0.05) in TT and FT immediately post exercise. middle-aged and older men showed similar relative T responses to those of younger men to a single bout of high-intensity resistance exercise [26] Resistance exercise young vs. middle aged 8 healthy middle aged (49 ± 2 years) and 10 young aged 21.2 ± 2.2 years Acute moderate intensity resistance exercise testing (AMIRE) consisted of 4 sets of 12 repetitions bench press with 120 s of rest between sets performed before and after 8-week progressive resistance training program three times per week TT increased at IP and remained elevated at 30 min post exercise for MI, with a higher level at IP compared with pre-training. Pre-training and post-training, in AMIRE, testosterone concentration in young men was higher at IP and 30P than middle aged men [27]   TT concentrations were increased after resistance training, while there were no significant changes in serum levels of these hormones in control group [38] Resistance 36 Age 22 years Randomized into an resistance training (12 weeks of training, 3/week) or control group (12 weeks no training), each session lasting one hour. The training overload had 3 phases. Phase 1 (weeks 1-2), 2 sets of 12-15 repetitions for each exercise at 100% RM. In phase 2 (weeks 3-7) 3 sets of 8-12 repetitions at 100% RM, phase 3 (weeks 8-12): 6-8 repetitions at 100% RM TT did not change in either group. These changes were noted without weight loss, and in concert with increases in lean body mass and decreasing fat mass [39] Aerobic exercise 16 normal-weight men and 28 overweight/obese men Age 50 ± 1.2 years 12-week aerobic exercise with physical activity levels labeled as (light, moderate, or vigorous) in all participants TT significantly increased in overweight/obese men. In addition, stepwise multivariable linear regression analysis revealed the increase in vigorous physical activity was independently associated with increased serum total testosterone levels [40]  TT: total testosterone IP: immediately post exercise RM: repetition maximum. NOTES: The analysis of published data demonstrated that either plasma total or free testosterone concentrations (T-Testo or F-Testo, respectively) or both are being presented. The majority of the work presents the data on either serum or plasma testosterone, while some report on the salivatory hormone. Unless specified, we will present the data mainly on T-Testo, which was the most frequently reported. Lastly, within the scope of this work, we analyzed only studies conducted in healthy men. As many conditions can alter the plasma concentration of sex hormone binding globulins, interpretation of T-Testo and F-Testo should take these conditions into consideration.

Part 1: Acute or Immediate Post-Exercise Testosterone Response
The changes in plasma testosterone concentrations during exercise may depend on multiple factors. Below we analyze the effect of the type of exercise (i.e., endurance or resistance), intensity, volume (i.e., the amount of muscle involved), obesity, and age on the acute or immediate post-exercise plasma testosterone changes.

Endurance Exercise
Endurance or aerobic exercise refers to any type of cardiovascular conditioning where breathing and heart rates increase for a sustained period of time. Although different types of endurance exercises have been performed, running and ergometer cycling with different protocols were most often used. Regarding the study populations, the majority of research done using running as a study approach was conducted in elite athletes. The study subjects in ergometer cycling studies included elite athletes, moderately active people, sedentary people, and people with obesity. Some studies have been conducted in patients with type 2 diabetes or chronic heart failure; however, these will not be discussed within the scope of this work.
Intensity refers to the load used for a given exercise. There appears to be a relative exercise intensity that must be reached in order to induce changes in serum testosterone concentrations [1]. Jezova et al. [7] compared the plasma testosterone changes during ergometer cycling conducted at three different intensities: high, moderate, and low. Significant increases in the serum testosterone concentrations were seen only in the high intensity exercise group. Kraemer et al. [8] reported that, when the number of repetitions during exercise was kept constant, the intensity determined the degree of acute post-exercise increase in serum testosterone concentrations. In the study of Kraemer et al. [9], well-trained runners underwent a treadmill running exercise with stepwise increase in intensity. The first two steps were at intermittent intensity of 60% and 75% VO2max for 10 min each (i.e., step 1 and step 2, respectively). The last two steps were at high intensity: 90% and 100% VO2max for 5 and 2 min, respectively. T-Testo increased only after 5 min of exercise at 90% VO2 max (Basal vs. 90% VO2max: 18.2 ± 3.4 vs. 24.1 ± 4.6 nmol/L, p < 0.01), and remained elevated at 100% VO2, returning to baseline 1 h into recovery [9]. These data support the notion that intensity is required to alter plasma testosterone concentrations. Galbo et al. [10] reported that, in young healthy men, a significant increase in T-Testo (~31%) was observed after 40 min of maximum intensity (reflected by the individual's highest oxygen uptake) during exhaustive treadmill running. However, other factors may also affect this hormonal change. For example, in the above-mentioned study by Kraemer et al. [9] the increase in T-Testo was at 25 min after the start of the exercise, and thus it is possible that not only the intensity but also the duration of the exercise contributed to these results. Moreover, Maresh et al. [11] studied the same individuals under four different conditions: 70% and 85% VO2max intensities with pre-existing euhydration (i.e., EU70 and EU85) and hypohydration (i.e., H70 and H85). The results of their studies demonstrated that only EU85 resulted in increased T-Testo, suggesting that levels of both intensity as well as hydration are important in determining the outcomes of an exercise intervention. These results suggest that intensity among other factors can play a role in the immediate changes in serum testosterone concentrations with endurance exercise.

Resistance Exercise
Resistance exercise, also known as strength and weight training, involves the voluntary activation of specific skeletal muscles against some form of external resistance. This external resistance is provided either by free weights, or a variety of exercise machines [12]. Examples include heavy weightlifting, jumping, or sprinting.
Multiple studies have shown that resistance exercise can cause acute changes in serum testosterone concentrations. Circulating T-Testo has been shown to increase immediately after a bout of heavy resistance exercise and return to baseline or even decrease beyond that level within 30 min post-exercise [13]. A major determinant for this increase in plasma testosterone concentrations is the muscle mass used. Involvement of a small muscle mass, even when vigorous resistance exercise is performed, does not elevate serum testosterone concentrations above resting levels [14]. In a study of young untrained men [15], unilateral biceps curl exercise alone did not induce a significant change in post-exercise serum testosterone concentrations. However, the addition of bilateral knee extensions and leg press to the biceps curl protocol resulted in a significant elevation of the T-Testo. Shaner et al. [49] evaluated the hormonal changes with similar lower body multi-joint movement free (i.e., squats) or machine weight (i.e., leg press) exercises. Free weight exercises induced a greater increase in plasma testosterone concentrations than did the machine weight exercises. A potential explanation for this finding is that squatting requires balancing on two feet with substantial engagement of stabilizing and core musculature, such as the abdominals and back. Research on muscle activation has also shown that free weight exercise results in a greater muscle activation than machine exercise, likely by inducing a larger overall muscle mass involvement than similar machine-based weight exercises [50]. These findings are supported by studies in junior Olympic-style weightlifters [16]. Weightlifting, an example of resistance exercise involving large muscle mass, resulted in a significant elevation of T-Testo at 5 min after exercise (Basal vs. Post-exercise: 16.2 ± 6.2 vs. 21.4 ± 7.9 nmol/L). Another possible explanation for this hormonal change is the involvement of larger muscle mass, which, in addition to resistance, may be required to induce significant acute changes in plasma testosterone concentrations.
Not much has been reported about the effect of variable vs. constant exercise intensities on serum testosterone concentrations. However, Charro et al. [17] reported that, when the total volume of the load lifted is fixed, both the variable and constant exercise intensities produce similar acute changes in T-Testo in healthy young men. Similar effects were observed in healthy elderly men [51]. Another study investigated the effect of a combination of exercise intensity, muscle volume (i.e., number of sets and repetition per a set), and the duration of the resting period between the sets on the acute hormonal variations [18]. The results demonstrated that a combination of a moderate intensity, higher volume, and shorter resting periods between sets can acutely and significantly increase the post-exercise T-Testo. Interestingly, the testosterone concentrations remained elevated for 48 h after exercise cessation. Similar findings were also reported by Kreamer et al. [52] further confirming the importance of a combination of various factors to mount significant increases in the post-exercise concentrations of serum testosterone.
Lastly, Tremblay et al. [19] evaluated the effect of baseline physical activity status on hormonal changes after an exercise protocol. They studied sedentary, endurance trained, and resistance trained individuals performing endurance and resistance exercise sessions a week apart. T-Testo increased in all groups after both types of exercise sessions, but the increase was more pronounced after the resistance training. Comparing the three groups, resistance-trained individuals had a higher increase in T-Testo, especially after resistance exercise. In summary, resistance exercise appears to be a direct stimulant to testosterone production when sufficient muscle mass load is met, or when a moderate and higher exercise intensity is combined with larger muscle volume and shorter resting periods between the sets.

Sustainability of Post-Exercise Elevated Levels of Testosterone, and Underlying Mechanisms
Both endurance and resistance exercise studies have demonstrated an increase in plasma testosterone concentrations upon exercising; however, these levels were not sustainable beyond several minutes. In the study by Manesh et al. [11], the increase in serum testosterone concentrations was not sustained at 20 min into the recovery. Daly et al. [20] likewise showed that, despite rapid increases in T-Testo after 30 min of endurance running, the levels significantly decreased 90 min into the recovery. Several studies have sought to investigate the potential underlying mechanisms that may explain these outcomes. Cumming et al. [6] showed that both testosterone and luteinizing hormone (LH) synchronously peaked at 20 min of progressive intensity exercise on an ergometer. However, a 45 min physiological lag between the LH pulse and testosterone production was well established [53], and thus LH stimulation may not be the mechanism responsible for the increase in plasma testosterone concentrations with exercise. Aerobic exercise can provide a large physiological stress to the body, resulting in a corresponding response of the neuro-endocrine system. This can be manifested by an initial rise in plasma testosterone concentrations secondary to a catecholamine surge and testicular stimulation, followed by increases in cortisol levels, a hormone that inhibits testosterone production [54]. Others have demonstrated that an increase in serum testosterone concentrations is not secondary to increased production rate [55,56]. Cadoux et al. [54] injected radiolabeled testosterone in men who underwent vigorous aerobic exercise for 50 min. The increase in plasma testosterone concentrations during the exercise was associated with a decrease in estimated hepatic plasma flow, metabolic clearance, and plasma volume. Moreover, infusion of lactate in rats, which is normally increased during exercise, was associated with a simultaneous increase in serum testosterone concentrations, particularly by increasing the testicular cAmp production [57]. Therefore, even though aerobic exercise seems to induce the initial rise in T-Testo, this effect is not sustained due to several factors including (a) sympathetic nervous system stimulation and the subsequent inhibition of testosterone production; and (b) changes in the testosterone metabolism. In contrast, anaerobic exercise-induced stimulation of testosterone production was explained by the effect of the anaerobic glycolytic pathway on the release of gonadotropin releasing hormone (GnRH) and LH [58]. The sustainability of exercise-induced elevation of testosterone concentrations may not vary between endurance and resistance exercise; however, the underlying mechanisms may be different.

Obesity
So far, we have discussed works done in lean, young men; however, body weight and aging are inversely related to serum testosterone concentrations [59]. In this and the following, section, we will discuss how obesity and aging, respectively, affect the exercise-induced changes in testosterone concentrations. The underlying mechanisms of this relation include: (a) attenuated amplitude of LH pulses due to obesity-induced systemic inflammation [60]; (b) increased aromatization of circulating testosterone in the adipose tissue [61]; and (c) higher leptin production by fat cells which has been shown to disrupt testosterone production [62].
Studies investigating the effect of exercise on serum testosterone concentrations in overweight and obese individuals also show conflicting results. Rubin et al. [21] compared resistance exercise-associated serum testosterone changes in physically active lean vs. obese men. Although T-Testo was comparable between obese and lean individuals immediately after exercise, the levels were lower in the obese men (lean vs. obese: 20 vs. 8 nmol/L, p < 0.01) at 30 min into recovery. Despite the initially similar change in both groups, the baseline and integrated concentrations during recovery appeared to be inversely associated with the degree of adiposity. Sheikholeslami-Vatani et al. [22] investigated the acute effect of different resistance exercise orders on serum testosterone concentrations in untrained normal weight and obese men. Although in both groups T-Testo increased acutely post exercise, the increase was higher in the lean individuals, suggesting an obesity-associated blunting in hormonal changes with exercise. Another study by Velasco-Orjuela et al. [23] evaluated the acute effect of high-intensity, resistance, or combined exercise protocols on T-Testo in inactive overweight men. Surprisingly, none of the exercise protocols affected the T-Testo. The authors hypothesized that obesity-associated endothelial dysfunction and impaired vasodilatation in the testes may be, at least partially, responsible for the altered overall endocrine response in this population. Thus, whether exercise can still potentiate testosterone spikes in overweight and obese individuals is uncertain; and if present, they seem to be inferior to those seen in lean/normal weight men.

Age
Aging is associated with declining levels of serum testosterone concentrations in men. This is secondary to the decreasing capacity of aging Leydig cells to produce testosterone in response to LH stimulation [63]. Low plasma testosterone concentrations are associated with a number of adverse health consequences including loss of muscle mass, increased fat mass, reduced physical performance, and increased cardiovascular disease risk [64]. The effect of exercise on serum testosterone concentrations in older men is not clearly understood. Studies in older participants refer to studies in men with an average age of 60 ± 5 years. Zmuda et al. [24] examined the acute effect of moderate physical activity with increasing intensity on T-Testo in elderly (70 ± 4 years), sedentary men. Levels acutely increased during exercise and peaked at higher intensities (basal vs. post-exercise: 11.2 vs. 16 nmol/L, p < 0.01); however, the levels returned to baseline at 60 min into recovery. Kraemer et al. [25] examined the acute effect of heavy resistance exercise on T-Testo in young (29.8 ± 5.3 years), and older (62 ± 3.2 years) men. Both groups showed significant increases in serum testosterone concentrations immediately and 5, 10, and 15 min post-exercise. However, the magnitude of increase was higher in the younger population. Another study evaluating the effect of aging on changes in serum testosterone concentrations included three groups of healthy untrained men: young (20-26 years), middle aged (33-58 years), and older (59-72 years), who performed one session of resistance training. All three groups exhibited an increase in T-Testo post exercise, with middle aged and older men showing similar relative testosterone concentration changes to younger men. Levels returned to baseline 15 min after exercise cessation in all groups [26]. Arazi et al. [27] studied young and middle-aged men who underwent an 8-week-long progressive resistance training program. The serum testosterone concentrations were measured at four separate time points: (1) before any exercise was conducted (baseline); (2) immediately after one bout of exercise, but before the training started (i.e., pre-training immediate; I-preT); (3) immediately after completion of 8-week-long training (i.e., post-training immediate; I-postT); and (4) basal or resting post-training (PT). Compared to the baseline levels, the T-Testo concentrations were increased at I-preT in middle-aged men only. In the same group, the I-postT T-Testo were also increased and remained elevated at 30 min into the recovery. These I-postT values were also higher than the I-preT values in this group. In young men, the plasma testosterone concentrations were higher at I-preT, I-postT, and 30 min into the recovery, when compared to those of the middle-aged men. In this study, no changes were observed in the resting concentrations of serum testosterone in middle-aged men, but resting testosterone concentrations increased in young men post-training. All the analyzed studies confirmed that middle-aged and older men still mount an elevation in plasma testosterone concentrations acutely after physical activity, despite age-related hormonal declines. However, the magnitude of increase can be lower than that seen in younger men.

Endurance and Resistance Exercise
Many studies have addressed the effect of habitual or intervention exercise on basal (resting) serum testosterone concentrations, with no clear effect reported so far. Studies have investigated the associations between the degree of physical activity and basal plasma testosterone concentrations. The 5-year-long NHANES study included 738 participants, who were classified in three tertiles, based on metabolic equivalent of task (MET) score, and according to the compendium of physical activities. No cross-sectional association was found between a greater physical activity and changes in basal plasma testosterone concentrations [28]. Houmard et al. [29] showed that despite increasing endurance exercise's frequency, duration, and intensity over 14 weeks (3-4 days/week, 30-45 min/day), no significant changes in the resting plasma testosterone concentrations were noted. Similarly, White et al. [30] found no change in resting testosterone concentrations with higher training mileage (i.e., 100% increase in the habitual distance run for 12 weeks) in recreational joggers. MacKelvie et al. [31] showed similar basal serum testosterone concentrations between long-distance runners and age-matched sedentary controls. In highly trained swimmers, the basal plasma testosterone concentrations did not differ between periods of intensive training and exercise tapering [32]. Interestingly, some studies have even shown that chronic endurance exercise can correlate inversely with basal serum testosterone concentrations. For instance, professional cyclists tend to have lower basal T-Testo after major competitions compared to baseline [33]. Safarinejad et al. [34] conducted a randomized trial of middle-age men undergoing intensive treadmill running. Throughout the study period, these men had low basal serum testosterone concentrations, which was associated with low follicular-stimulating hormone (FSH) and LH levels. The authors hypothesized that exercise-associated stress induced production of reactive oxygen species that can suppress the hypothalamic pituitary axis and cause hypogonadotropic hypogonadism. Interestingly, the sex hormone binding globulin levels did not decrease with declining T-Testo, reflecting that the serum testosterone changes are not related to the variation in serum binding globulin. Hackney et al. [35] reported that endurance trained men had lower T-Testo than sedentary men. In this study, LH levels were not elevated despite the lower limit values of testosterone, which may indicate HPA axis suppression with long-term endurance exercise. On the other hand, lower basal testosterone concentrations have been reported despite unaltered plasma LH and FSH levels [36]. Thus, although low basal testosterone concentration is likely due to HPA suppression and hypogonadotropic hypogonadism during chronic exercise, additional contributing factors affecting the serum testosterone concentrations in the absence of LH suppression are yet to be determined.
Although studies have proven that resistance exercise can cause significant acute changes in serum testosterone concentrations, similar changes were not observed in basal plasma testosterone levels. Nicklas et al. [37] reported no significant change in basal serum testosterone concentrations after 16 weeks of progressive resistance training program. Moreover, the previously mentioned study by Hansen et al. [15] showed unchanged resting testosterone concentrations during unilateral biceps curl exercise alone or in combination with bilateral knee extensions and leg press. Therefore, independent of exercise type, nature, or intensity, exercise does not seem to increase resting T-Testo. Based on these studies, exercise either decreases or have a neutral effect on T-Testo.

Obesity
The effect of exercise on basal serum testosterone concentrations in obese individuals has been evaluated in multiple studies. Although Moradi et al. [38] reported significant increases in basal serum testosterone concentrations in obese men after 12 weeks of resistance exercise (Basal vs. post-exercise: 23.9 ± 8.3 vs. 28.4 ± 5.9 nmol/L, p = 0.018), this correlation was not found in another study using a similar population and the same exercise type [39]. Kumagai et al. [40] investigated the effect of a 12-week aerobic exercise intervention on circulating serum testosterone concentrations in overweight/obese men. At baseline, T-Testo were significantly lower in overweight/obese men than in normal-weight men (p < 0.01). After the 12-week aerobic exercise intervention, serum testosterone concentrations significantly increased in the overweight/obese men (p < 0.01). In addition, stepwise multivariable linear regression analysis revealed that the increase in vigorous physical activity was independently associated with increased basal T-Testo (p = 0.011). Similarly, Rosety et al. [41] reported that a 16-week-long aerobic training program on a treadmill, with three sessions per week, increased the basal serum testosterone concentrations in obese men (baseline vs. post-test range: 15.1 vs. 16.6 nmol/L, p = 0.036). Khoo et al. [42] reported significant increases in serum testosterone concentrations in individuals with obesity after 24-weeks of high volume moderate-intensity exercise. Thus, most of the studies in overweight/obese men have shown a direct correlation between both aerobic and anaerobic exercise and plasma testosterone concentrations. Surprisingly, these results contradict the results of studies in lean individuals, where even those using strict protocols to stimulate acute testosterone increases failed to change basal testosterone concentrations (see above). One possible explanation for these findings is the weight/fat mass loss effect. Some studies showing direct correlations between exercise and serum testosterone concentrations also showed decreased fat mass and waist circumference in individuals with obesity [25,26]. Whether increase in basal testosterone concentrations is solely due to exercise, or is secondary to weight loss is still to be determined.

Age
The effect of exercise on basal serum testosterone concentrations in older men is not clearly understood. Ari et al. [43] reported that well-trained, athletic older men have significantly higher resting T-Testo than age-matched sedentary men (sedentary vs. athletic: 18.7 ± 5.9 vs. 28.8 ± 4.5 nmol/L p < 0.01). However, other studies were unable to distinguish differences in basal T-Testo between lifelong trained and sedentary elderly men (Hayes et al. [44] and Tissandier et al. [45]).
Some studies have been conducted to assess the changes in serum testosterone concentrations during exercise in elderly men. Hayes et al. [46] examined the impact of 6-week-long supervised exercise training on resting concentrations of serum testosterone in a cohort of lifelong sedentary men, compared to a control group of age-matched lifelong exercisers. The results revealed that only sedentary men experienced a significant exercise-induced increase in resting T-Testo. Another study by Lovell et al. [47] found no significant changes in resting plasma testosterone concentrations after 16 weeks of aerobic or resistance exercise in men aged 70-80 years. Of note, T-Testo increased immediately post sub-maximum exercise in all groups, showing a pattern similar to the post-exercise results in young adults (see above). Sellami et al. [32] conducted a randomized trial to test the effect of exercise on serum testosterone fluctuations in moderately trained young and middle-aged men (average age 20 vs. 40 years, respectively). At rest, lower T-Testo were reported in the middle-aged compared to the younger group. However, after 13-weeks of intensive anaerobic activity, the levels taken 48 h to 7 days post-exercise cessation were significantly increased in the middle-aged group, eliminating the age-associated difference between the groups. It was previously questioned whether the exercise protocols contributed to the variable results in these studies. However, even in studies involving populations with similar age, physical activity status, exercise background, and protocol duration, the change of basal plasma testosterone concentrations during exercise has not been consistent [46,47]. Based on the current literature, no conclusions can be drawn on the effect of exercise on basal serum testosterone concentrations in older men.

Discussion
This review highlights that substantial research has been done on the effect of exercise on (a) acute immediate; and (b) basal or resting post-exercise serum testosterone concentrations in men. Regardless of pre-existing conditioning, body weight, or age, sufficient evidence indicates that resistance exercise, when combined with larger muscle involvement (multi-joint movements), bigger exercise volume, sufficient intensity (moderate/high), and short resting intervals between training sets, may result in optimal acute increases in serum testosterone concentrations. However, the magnitude of this acute hormonal change is lower in older men or those with obesity.
Whether this temporary surge in post-training serum testosterone concentrations has any impact on the extent of muscle anabolism and hypertrophy is widely debated. Multiple studies found a direct link between post-exercise serum testosterone changes and muscular hypertrophic adaptation/increase in lean body mass [48]. A possible explanation is that increases in serum testosterone concentrations mediate an upregulation in acute androgen receptor expression and subsequent increases in myofibrillar protein synthesis, possibly because of enhanced ligand binding capacity or activation of the testosterone-androgen receptor signaling pathway [65]. Post-exercise peak plasma testosterone enhances androgen receptor mRNA translation and increases its half-life. Evidence suggests that acute increases in serum testosterone concentrations during exercise may likely optimize hypertrophic adaptations via enhancing the testosterone-androgen receptor [66]. On the other hand, Wilkinson et al. [67] observed significant gains in strength and hypertrophy in the absence of any measurable changes in F-Testo and insulin growth factor 1. A study by West et al. [68] showed that exposure of muscles to basal or high serum testosterone concentrations with exercise can result in similar muscle adaptations and hypertrophy. Thus, there is no solid evidence that the post-exercise acute plasma testosterone spike has a beneficial effect on muscle hypertrophy. Data on whether exercise induces prolonged testosterone stimulation is still limited, with the majority of studies showing similar resting serum testosterone concentrations in active and inactive individuals. Some promising studies in older men have shown a direct correlation between exercise and basal plasma testosterone concentrations; however, conclusions are still preliminary until a greater depth of literature is available [44]. Similarly, studies showing positive correlations between exercise and increased basal plasma testosterone concentrations in overweight/obese individuals also showed significant associated fat loss. Whether this effect is secondary to weight loss and less aromatization, or solely secondary to exercise, is unclear.
In conclusion, the up-to-date data on the effect of exercise on serum testosterone concentrations in men have significant inter-individual and inter-study variability. This variability can be explained by (a) the use of different types of exercise (e.g., endurance vs. resistance); (b) the other factors of the training (e.g., training intensity or duration of resting periods); (c) the variety in study populations (e.g., young vs. elderly; lean vs. obese; sedentary vs. athletes); and (d) the time points when testosterone was measured (e.g., during or immediately after vs. several minutes or hours after the exercise). It is our conclusion that future studies should focus on clarifying the metabolic and molecular mechanisms whereby exercise may affect testosterone production in the short-and long-term, and furthermore how this release affects downstream mechanisms; such knowledge will be the key to understanding the exercise-testosterone-muscle hypertrophy axis.