Testosterone undecanoate administration prevents declines in fat-free mass but not physical performance during simulated multi-stressor military operations

Abstract

Male military personnel conducting strenuous operations experience reduced testosterone concentrations, muscle mass, and physical performance. Pharmacological restoration of normal testosterone concentrations may attenuate performance decrements by mitigating muscle mass loss. Previously, administering testosterone enanthate (200 mg/wk) during 28 days of energy deficit prompted supraphysiological testosterone concentrations and lean mass gain without preventing isokinetic/isometric deterioration. Whether administering a practical dose of testosterone protects muscle and performance during strenuous operations is undetermined. The objective of this study was to test the effects of a single dose of testosterone undecanoate on body composition and military-relevant physical performance during a simulated operation. After a 7-day baseline phase (P1), 32 males (means ± SD; 77.1 ± 12.3 kg, 26.5 ± 4.4 yr) received a single dose of either testosterone undecanoate (750 mg; TEST) or placebo (PLA) before a 20-day simulated military operation (P2), followed by a 23-day recovery (P3). Assessments included body composition and physical performance at the end of each phase and circulating endocrine biomarkers throughout the study. Total and free testosterone concentrations in TEST were greater than PLA throughout most of P2 (P < 0.05), but returned to P1 values during P3. Fat-free mass (FFM) was maintained from P1 to P2 in TEST (means ± SE; 0.41 ± 0.65 kg, P = 0.53), but decreased in PLA (−1.85 ± 0.69 kg, P = 0.01) and recovered in P3. Regardless of treatment, total body mass and fat mass decreased from P1 to P2 (P < 0.05), but did not fully recover by P3. Physical performance decreased during P2 (P < 0.05) and recovered by P3, regardless of treatment. In conclusion, administering testosterone undecanoate before a simulated military operation protected FFM but did not prevent decrements in physical performance.

NEW & NOTEWORTHY This study demonstrated that a single intramuscular dose of testosterone undecanoate (750 mg) administered to physically active males before a 20-day simulated, multi-stressor military operation increased circulating total and free testosterone concentrations within normal physiological ranges and spared FFM. However, testosterone administration did not attenuate decrements in physical performance across multiple measures of power, strength, anaerobic or aerobic capacity.

INTRODUCTION

Special Operations Forces conducting sustained, multi-stressor military training and combat operations consist predominately of male personnel, who often experience substantial decreases in circulating testosterone below hypogonadal concentrations (≤300 ng/dL or ≤10.4 nmol/L) (1), muscle mass loss, and declines in physical performance (2–7). These negative consequences can manifest during training lasting as little as 72 h (6, 8) and are often attributed, in part, to the collective effects of sleep deprivation, increased exercise-induced energy expenditure, reduced energy intake (decreased appetite, time constraints, and limited food availability), and the resultant energy deficit (the percentage of energy intake required to establish energy balance and maintain body mass) on endogenous testosterone synthesis (9–12).

The anabolic effects of providing exogenous testosterone to hypogonadal young males are well-documented and include increased muscle mass and strength (13, 14). Thus, pharmacological restoration of eugonadal testosterone concentrations in males during sustained military operations may be an effective therapeutic to attenuate performance decrements by mitigating muscle mass loss (15). The potential for testosterone restoration to mitigate muscle mass loss and performance declines during energy deficit was recently examined by our laboratory in the Optimizing Performance for Soldiers I (OPS I) study (16–20). In OPS I, administering testosterone enanthate (200 mg/wk for 4 wk) to healthy, nonobese males promoted gains in lean mass during 28 days of 55% exercise- and diet-induced energy deficit, whereas controls maintained their lean mass while being underfed [mean difference between groups (95% CI), 2.5 kg (3.3 kg, 1.6 kg); P < 0.001] with no negative health-related consequences (16, 17). However, isokinetic/isometric measures of single-joint, single-leg lower-body strength and endurance declined similarly in both groups.

Although the accrual of lean mass observed in OPS I may confer physiological benefits, several experimental aspects in OPS I preclude definitive conclusions regarding the efficacy of administering supplemental testosterone to military personnel during strenuous military operations. For example, the short-acting testosterone ester administered weekly in OPS I (200 mg testosterone enanthate/wk) may not be feasible (i.e., medical oversight), practical (i.e., weekly injections), or safe (supraphysiological increases in circulating testosterone followed by prolonged suppression upon cessation) for military personnel operating in austere environments. The consequences related to the testosterone ester used in OPS I are likely overcome by using a different testosterone formulation and dosing strategy. More specifically, a single intramuscular injection of testosterone undecanoate to hypogonadal males can maintain testosterone concentrations within the normal physiological range for 8–10 wk (21). It does so with less logistical and clinical burden than testosterone enanthate, as the dosing regimen only requires additional injections every few months (initial 750 mg intramuscular injection, followed by another 750 mg at week 4, and 750 mg every 10 wk thereafter) (22). Circulating testosterone concentrations also decline gradually once the use of testosterone undecanoate treatment has ceased (22), reducing the risk of adverse events. As such, this dosing strategy may be feasible for military personnel who experience suppression of endogenous testosterone, muscle mass loss, and performance decrements as part of their deployment cycle.

In addition, OPS I isolated the consequences of exercise- and diet-induced energy deficit from other stressors experienced during sustained military operations, and, as a result, demonstrated that the physiological stress and its consequences during real-life military operations are greater than what could be accounted for by the energy deficit alone. More specifically, the unexpected preservation of lean mass and nonsignificant, 20% reduction in total testosterone observed in OPS I study controls were far less than the declines observed during real-life military operations (11, 23, 24). The 55% energy deficit used in OPS I, which was elicited via increasing total daily energy expenditure by 50% above baseline values and restricting energy intake to 45% of this elevated total daily energy expenditure (16, 17), was similar to the relative energy deficit observed during real-world military operations (24–27). However, the magnitude of energy expenditure induced via discrete exercise bouts, as opposed to prolonged, low-intensity aerobic exercise throughout the day, and the absolute energy deficit (daily energy deficit multiplied by the number of days underfed) were less than those observed in the field (24–27). Furthermore, OPS I study participants were provided with ample time to sleep and recover, which does not occur during real-world training and combat operations. Collectively, these factors may have contributed to the observed outcomes in OPS I, suggesting that the stress imposed was insufficient at replicating the multifaceted stress endured during real-world sustained military operations. We also suspect that the gains in lean mass in individuals administered testosterone and the maintenance of lean mass in OPS I study controls may have been overestimated as a result of increased water retention (28), which could not be discerned from relying on dual-energy X-ray absorptiometry (DXA)-derived estimates of lean mass alone, as these inherently include body water. Using a 4-compartment (4-C) measure of body composition may provide better resolution to assess body composition changes to testosterone restoration during strenuous operations. The 4-C model accounts for biological variation due to hydration, thereby resulting in more accurate estimates of fat-free mass (FFM) and fat mass than DXA alone (29).

Finally, physical performance in OPS I was assessed using single-joint, single-leg isometric and isokinetic dynamometry, which provides an extremely narrow assessment of performance and may not capture potential benefits of testosterone restoration on military-relevant metrics of performance (16, 17). As such, we designed the OPS II study (30) to address these factors. Therefore, the purpose of OPS II was to test the effects of a single dose of testosterone undecanoate (750 mg) on 4-C measures of body composition and broad measures of physical performance in healthy males in response to a simulated, sustained, multi-stressor military operation. The simulated military operation comprised severe energy deficit caused by high daily exercise-induced energy expenditure, resulting from prolonged exercise periods throughout most of the day, with limited recovery and restricted sleep. We hypothesized that a single dose of long-acting testosterone undecanoate would safely and steadily maintain normal testosterone concentrations and mitigate decrements in FFM and militarily-relevant measures of physical performance during a simulated, multi-stressor military operation.

MATERIALS AND METHODS

Study Design and Setting

This three-phase, interventional study used a parallel, randomized, placebo-controlled, double-blind design, as previously described (30). All participant testing occurred at the Pennington Biomedical Research Center (PBRC) in Baton Rouge, LA. The Institutional Review Board (IRB) of the PBRC (Protocol 2019-017) and the US Army Medical Research and Development Command, Human Research Protections Office approved the study protocol and trial documents, and all procedures were in accordance with the ethical standards of the 1964 Helsinki Declaration and its later amendments. The clinicaltrials.gov identifier is NCT04120363.

Eligibility Criteria

Complete inclusion and exclusion criteria and extended details of participant recruitment are described in detail elsewhere (30). Briefly, participants were healthy, physically active (expended at least 300 kcal/day on average through structured aerobic and strength-training activities) males aged 18–35 yr that met age-specific US Army body composition standards (31) and had normal testosterone concentrations (10.4–34.7 nmol/L; 300–1,000 ng/dL) (1).

Intervention

Study overview.

Extended details of the experimental design and testing procedures are reported elsewhere (30). Briefly, participants underwent a three-phase, 50-day study, consisting of 7 days of baseline testing (P1, days 1–7), 20 days of simulated, sustained operations (P2, days 8–27), and 23 days of recovery (P3, days 28–50) (Fig. 1). On day 8, after completing P1, participants were randomized to receive either a single intramuscular injection of testosterone undecanoate (TEST; AVEED, 750 mg testosterone undecanoate in 3 mL) or an iso-volumetric placebo (PLA; sesame oil solution, 3 mL). Details of randomization, treatment allocation, and blinding have been published (30). The 20-day simulated military operation (P2) was highly controlled (participants lived in the inpatient unit at PBRC) and consisted of four consecutive cycles of undulating stress, starting with 2 days of low stress, followed by 3 days of high stress. After completing P2, participants resumed their habitual physical activity routines but consumed a controlled diet in recovery (P3).

Figure 1.

Study design, adapted with permission from Elsevier from Varanoske et al. (30). P1 (days 1–7) and P3 (days 28–50) were run-in and free-living with a standardized diet with energy derived from 15% protein, 55% carbohydrates, and 30% fat. P2 (days 8–27) was a highly controlled, multi-stressor military operation, consisting of four consecutive cycles of undulating stress, starting with 2 days of low stress (∼1,000 kcal/day exercise-induced energy deficit; 8 h/day sleep; denoted by bolded number), followed by 3 days of high stress (∼3,000 kcal/day exercise-induced energy deficit, 4 h/day sleep; denoted by red bolded and underlined number). Participants were randomized to receive either a single intramuscular injection of testosterone undecanoate (TEST; 750 mg) or an iso-volumetric placebo (PLA; sesame oil solution) on day 8. Participants consumed the same total calories and macronutrient distribution in P2, but food was derived from the Meal, Ready-to-Eat, a US combat ration ([MRE] menu 39; Ameriqual, Evansville, IN). Body mass was measured daily. Body composition and physical performance were measured once at the end of each phase. Blood draws were obtained on several occasions throughout each phase to assess circulating endocrine biomarkers. P1, phase 1; P2, phase 2; P3, phase 3; 3-RM, 3-repetition maximum; V̇o_2peak, maximal cardiorespiratory fitness test.

Dietary intake.

Throughout all phases, participants consumed the same amount of total energy, which was prescribed on an individual basis, with the macronutrient distribution fixed based on the composition of the Meal, Ready-to-Eat (MRE) (15%, 55%, and 30% total energy from protein, carbohydrate, and fat, respectively) (30). Individual participant physical activity patterns and exercise-induced energy expenditure before P1 were determined using accelerometry and a physical activity questionnaire (PAR-Q+). Resting metabolic rate was measured by using indirect calorimetry (Deltatrac II Metabolic Cart Sensormedics, Yorba Linda, CA). The 3-day food records and resting metabolic rate measurements were used to calculate total daily energy expenditure and prescribe individual dietary intake to maintain energy balance and body mass within ±2% during P1. For P1 and P3, registered dietitians developed individualized menus for each participant, and compliance was checked daily. Dietary intake during P2 consisted solely of items derived from MREs (menu 39; Ameriqual, Evansville, IN), but individual total intake and macronutrient distribution remained the same as in P1 and P3 (Supplemental Table S1; all Supplemental material is available at https://doi.org/10.6084/m9.figshare.17155544). The MRE is the standard US Department of Defense field ration. Water was consumed ad libitum throughout all phases.

Simulated, sustained military operation.

The 20-day simulated sustained military operation (P2) consisted of four consecutive cycles of undulating stress, starting with 2 days of low stress followed by 3 days of high stress (see Ref. 30 for details). Low- and high-stress days entailed low- and high exercise-induced energy expenditure (∼1,000 and ∼3,000 kcal/day, respectively, above P1 exercise-induced energy expenditure) and adequate and restricted sleep (8 h/day vs. 4 h/day). This level of energy deficit was selected based on a recent meta-regression of military field studies indicating that a total energy deficit of ∼43,380 kcal over 20 days should result in a moderate to large declines in total body mass (approx −9%) and lower-body physical performance (approx −7%) (32). Participants performed multiple exercise sessions each day using a variety of endurance and muscle-loading modalities, which were used to mimic movements typically observed during real-life military operations. Steady-state (i.e., speed and grade) ruck marching while carrying a load of ∼30% of total body mass was the primary exercise modality and was prescribed to account for ∼50% of daily exercise-induced energy expenditure. Other activities included walking, running, cycling, elliptical, field-based operational activities, and stretching. Exercise-induced energy expenditure was determined for each participant and exercise using the Compendium of Physical Activities (33) or according to published equations (34, 35).

Outcome measures.

Anthropometrics and body composition.

Height was measured using a stadiometer (Harpenden Stadiometer, Holtain Company, UK) at the beginning of the study, and metabolic body mass (mass in a gown and with gown mass subtracted) was measured after an overnight fast and morning void daily (GSE Inc. Model 450, GSE Scale Systems, Novi, MI). Body composition [fat, lean soft tissue, bone mineral, visceral adipose tissue (VAT) mass, subcutaneous adipose tissue (SAT) mass, and total body water (TBW)] was measured after an overnight fast, proper hydration, and morning void at the end of each phase (Fig. 1), as previously described (30). Body composition was computed using a 4-C model with DXA (Hologic, Discovery A, Marlborough, MA) and measures of TBW derived from 99% deuterium dilution (D₂O; Sigma Aldrich, St. Louis, MO) and standardized procedures (36–38). Coefficients of variation (CVs) for the spine phantom DXA calibration over the course of the study were as follows: bone mineral density: 0.227%, bone mineral content: 0.255%, area: 0.229%. FFM and fat mass were calculated using standardized equations for 4-C body composition from DXA variables and TBW (39, 40). Appendicular lean soft tissue (ALST) mass as measured by DXA was used as a proxy for total body skeletal muscle mass (41).

Physical performance.

A battery of performance tests was completed during each phase, as previously described (30) (refer to Fig. 1 for timing of the performance tests during each phase). Participants were familiarized with each test several days prior to testing in each phase. The order and timing of the tests were standardized, and participants completed a dynamic warm-up before testing began.

Vertical jump test.

Lower-body power was evaluated using the vertical jump test (Vertec, Jump USA, Sunnyvale, CA). Participants completed a series of three maximal countermovement jumps with a Tendo unit attached around the waist to measure power and velocity (Tendo Sports Machines, Trenchin, Slovak Republic). Participants jumped by flexing their knees and hips, moving downward, and extending their knees and hips rapidly while swinging up their dominant arm to touch the highest possible vane on the Vertec. Jump height from the Vertec, as well as average power, partial average power (measure of average power throughout the initial phase of the jump), peak power, average velocity, peak velocity, and peak force from the Tendo unit were recorded for each jump. Data for the jump with the greatest height were used in the analysis.

3-Repetition maximum deadlift.

A 3-repetition maximum (RM) trap bar deadlift was used to assess total body muscular strength and power in accordance with the US Army Combat Fitness Test (42). Participants stood in the middle of the bar with feet shoulder width apart, and a Tendo unit was attached to the end of the bar. Participants bent at the knees and hips, reached down and grasped the center of the handles, stood up, and lifted the bar by extending the hips and knees until in an upright stance. They paused slightly at the top of the movement, then flexed the hips and knees slowly, and lowered the bar to the ground in a controlled manner. After three warm-up sets, the bar was loaded with ∼85%–90% of estimated 1-RM. Participants were instructed to complete three continuous repetitions, ascending as rapidly as possible. If they failed to complete three repetitions, they retested at a lower weight. If successful, additional weight was added and they retested after 3 min of rest. The 3-RM mass and subjective rating of perceived exertion (RPE; 1–10 Likert scale) were recorded. Average power, partial average power, peak power, average velocity, peak velocity, and peak force for each repetition of the 3-RM from the Tendo unit were also recorded.

Wingate anaerobic cycle test.

The Wingate test was used to measure anaerobic capacity. Participants were positioned on an electronically braked cycle ergometer (Excalibur Sport, Lode, The Netherlands) equipped with software (Lode Ergometry Manager software version 10.11.0, Lode B.V., Lode, The Netherlands) and began pedaling for 5 min at 50 W. On “Go,” participants increased their cadence to 90 rpm, and a fixed resistance was added to the bike. The fixed resistance was determined from body mass, cycle cadence, and a torque factor determined off the participant’s performance during familiarization. The cycle settings and fixed resistance remained the same throughout the entire study, despite a decrease in body mass throughout P2. Participants pedaled maximally for the duration of the 30-s test. Peak power, total work, fatigue index, maximum revolutions per minute (rpm), and maximum load were recorded by the Lode Ergometry Manager software (version 10.11.0, Lode B.V., Lode, The Netherlands).

Treadmill maximal cardiorespiratory fitness test (V̇o_2peak).

Aerobic capacity (i.e., V̇o_2peak) was measured using a graded exercise test and an indirect open circuit respiratory system (ParvoMedics TrueOne 2400, East Sandy, UT) on a treadmill (Track Master TMX425CP, Full Vision, Inc., Newton, KS). Participants began by completing a 5-min warm-up and then ran for 4 min at a pace predetermined during familiarization at a 0% grade. The grade was then increased to 2%, followed by an additional 2% every 2 min thereafter until volitional exhaustion. V̇o_2peak, respiratory exchange ratio (RER), and maximum grade were recorded.

Ruck march time trial.

Load carriage is an essential military task in which Soldiers are expected to carry a standard fighting load of 68.9 lbs (31.3 kg) and move at a rate of 4 km/h in an ideal situation (43). A 4-km (2.5 mile) outdoor timed ruck march (e.g., backpack load carriage) was completed while wearing a 31.3 kg rucksack to assess military-relevant aerobic endurance. Time, heart rate (HR), and RPE were recorded at every half mile.

Endocrine biomarkers.

Blood samples were collected after an overnight fast on several occasions throughout the study to assess endocrine function and stress (Fig. 1), as previously described (30). All assays were conducted by PBRC, which is a laboratory accredited by the Clinical Laboratory Improvement Amendments (CLIA) and The College of American Pathologists (CAP). All blood samples were analyzed for total testosterone and free testosterone [determined by calculation (44)], follicle-stimulating hormone (FSH), estradiol, sex-hormone-binding globulin (SHBG), luteinizing hormone (LH), insulin-like growth factor-1 (IGF-1), growth hormone (GH), insulin, and cortisol according to the manufacturer’s instructions (Siemens Immulite 2000, Siemens Medical Solutions, Malvern, PA). All assays were run with standards and appropriate quality control material. Intra- and interassay CVs for the methods listed in the manufacturer enzyme-linked immunoassays were: total testosterone (Siemens Catalog No. L2KTW2; 9.8% and 9.3%), FSH (Siemens Catalog No. L2KFS2; 3.4% and 4.9%), estradiol (Siemens Catalog No. L2KE22; 6.7% and 7.4%), SHBG (Siemens Catalog No. L2KSH2; 3.1% and 4.6%), LH (Siemens Catalog No. L2KLH2; 4.8% and 4.8%), IGF-1 (Siemens Catalog No. L2KIGF2; 3.5% and 6.0%), GH (Siemens Catalog No. L2KGRH2; 3.7% and 3.9%), insulin (Siemens Catalog No. L2KIN2; 4.0% and 5.7%), and cortisol (Siemens Catalog No. L2KCO2; 6.0% and 5.2%), respectively.

Sample Size Estimation

Physical performance (lower-body muscular strength and power) was the primary outcome measure. The intended total energy deficit during the 20-day simulated military operation was ∼43,380 kcal, which, based on a previous meta-regression of data generated from military field studies, should elicit a 9% reduction in total body mass and 7% decline in lower-body physical performance (45). This magnitude of energy deficit is of greater magnitude than that used in OPS I (17). The anticipated energy deficit and body mass loss also exceed the energy deficit and ∼7% reduction in body mass that occurred during the first 4 weeks of US Army Ranger training concomitantly lowered total testosterone concentrations by >50% (pre-Ranger training, 17.3 nmol/L; week 4 of Ranger training, ∼7.0 nmol/L) (6). We anticipated a similar decline in total testosterone concentrations, and based on the results from OPS I (17), we expected that body mass loss in TEST would be ∼50% less than PLA, such that lower-body muscular strength and power declines from baseline are attenuated by 50% in TEST relative to PLA. Sample size analysis was completed using SAS (SAS Institute version 9.4, Cary, NC) and was based on a t test of changes in total body mass and lower-body physical performance from P1 to P2 between treatment groups. Thus, the sample size necessary to detect differences between groups was 15 per group with 90% power. To account for attrition, 16 participants were randomized to each group.

Statistical Analysis

All data analyses were based on the intent-to-treat principle using SAS (SAS Institute version 9.4, Cary, NC). Between-treatment comparisons for all demographic baseline variables and the area under the curve (AUC) for blood biomarker concentrations during P2 were assessed using independent samples t tests. Race and ethnicity data, as well as incidence of adverse events were analyzed by χ² tests. Total daily energy intake, total daily energy expenditure, energy deficit, and exercise-induced energy expenditure during P2, as well as physical performance, body mass and composition, and blood biomarkers were analyzed using a mixed-effect linear model. Treatment (TEST and PLA), phase (P1, P2, and P3), and phase × treatments interaction were considered fixed effects covariates in the model. For variables that were measured daily or frequently throughout the study (e.g., body mass and blood biomarker concentrations), treatment, day, and day × treatment interactions were considered fixed effects covariates in the model. The random effect included an unstructured covariance matrix to account for correlations within participants over time. Least squares means from the model were used to estimate interaction effects. Family wise error rate was adjusted using the Bonferroni correction when appropriate. All analyses were considered two-tailed, with α = 0.05 considered statistically significant. All data are presented as means ± standard error, unless otherwise noted.

RESULTS

Participants

Physically active males were enrolled (n = 34), randomized (n = 34), and completed the study (n = 32; TEST: n = 16; PLA: n = 16; participant flow chart presented in Fig. 2). No differences were observed at baseline in participant demographics (all P > 0.05; Table 1).

Figure 2.

Participant flow chart. *Dropped after P3 physical performance testing but before body composition testing was completed. ITT, intent-to-treat; P1, phase 1; P2, phase 2; P3, phase 3; PLA, participants randomized to 750 mg sesame oil solution on day 8; PP, per protocol; TEST, participants randomized to 750 mg testosterone undecanoate on day 8.

Table 1.

Participant demographic data at the start of P1 in the intent-to-treat population

	TEST, n = 16	PLA, n = 18	P Value
Age, yr	26.9 ± 4.4	26.2 ± 4.5	0.646
Body mass, kg	76.5 ± 14.7	77.5 ± 10.2	0.817
Height, cm	177.1 ± 8.1	176.3 ± 5.3	0.759
BMI, kg/m2	24.3 ± 3.8	24.9 ± 2.7	0.634
Race, n (%)			0.897
Caucasian	11 (32.4%)	12 (35.3%)
African American	3 (8.8%)	5 (14.7%)
Asian	2 (5.9%)	0 (0%)
Other	0 (0%)	1 (2.9%)
Ethnicity, n (%)			0.530
Hispanic/Latino	3 (8.8%)	2 (5.9%)
Not Hispanic/Latino	13 (38.2%)	16 (47.1%)

Age, body mass, height, and BMI were analyzed using independent-samples t tests, and values are presented as means ± standard deviation. Race and ethnicity data are presented as n (% of total sample) and were analyzed using χ² tests. Significance was set at P < 0.05. P1, phase 1; PLA (n = 18), participants randomized to 750 mg sesame oil solution on day 8; TEST (n = 16), participants randomized to 750 mg testosterone undecanoate on day 8.

Adverse Events

The frequency of adverse events reported throughout P1 (broad categories: P = 0.057, expanded categories: P = 0.067), P2 (broad categories: P = 0.839, expanded categories: P = 0.809), P3 (broad categories: P = 0.359, expanded categories: P = 0.234), and the entire study (broad categories: P = 0.736, expanded categories: P = 0.796) were not different between TEST and PLA (Table 2).

Table 2.

Incidence of adverse events in the intent-to-treat population

	TEST, n = 16				PLA, n = 18				Category Total
	P1	P2	P3	Group Total	P1	P2	P3	Group Total
	Frequency (% of category, % of total for phase, % of overall total)			Frequency (% of category, % of overall total)	Frequency (% of category, % of total for phase, % of overall total)			Frequency (% of category, % of overall total)	Frequency (% of overall total)
Cardiovascular		1 (50%, 0.6%, 0.5%)		1 (50%, 0.5%)		1 (50%, 0.6%, 0.5%)		1 (50%, 0.5%)	2 (1.0%)
Edema		1 (50%, 0.6%, 0.5%)		1 (50%, 0.5%)		1 (50%, 0.6%, 0.5%)		1 (50%, 0.5%)	2 (1.0%)
Dermatological	2 (100%, 12.5%, 1%)	21 (47.7%, 12.1%, 10.1%)	2 (40%, 10%, 1%)	25 (49.0%, 12%)		23 (52.3%, 13.3%, 11%)	3 (60%, 15%, 1.4%)	26 (51.0%, 12.4%)	51 (24.4%)
Blister/Abrasion	2 (100%, 12.5%, 1.0%)	19 (48.7%, 10.9%, 9.1%)		21 (50%, 10.1%)		20 (51.3%, 11.6%, 9.6%)	1 (100%, 5%, 0.5%)	21 (50%, 10.1%)	42 (20.1%)
Bruising							1 (100%, 5%, 0.5%)	1 (100%, 0.5%)	1 (0.5%)
Erythema		2 (50%, 1.2%, 0.9%)	2 (66.7%, 10%, 1.0%)	4 (57.1%, 1.9%)		2 (50%,	1 (33.3%, 5%, 0.5%)	3 (42.9%, 1.4%)	7 (3.4%)
Hematoma						1 (100%, 0.6%, 0.5%)		1 (100%, 0.5%)	1 (0.5%)
Gastrointestinal	1 (50%, 6.3%, 0.5%)	7 (70%, 4.1%. 3.4%)	1 (50%, 5%, 1.0%)	9 (64.3%, 4.3%)	1 (50%, 6.3%, 0.5%)	3 (30%, 1.7%, 1.4%)	1 (50%, 5%, 1.0%)	5 (35.7%, 2.4%)	14 (6.7%)
Constipation		1 (100%, 0.6%, 0.5%)		1 (50%, 0.5%)			1 (100%, 5%, 0.5%)	1 (50%, 0.5%)	2 (1.0%)
Heartburn/Indigestion		3 (60%, 1.7%, 1.4%)		3 (50%, 1.4%)	1 (100%, 6.3%, 0.5%)	2 (40%, 1.2%, 1.0%)		3 (50%, 1.4%)	6 (2.9%)
Nausea/Vomiting	1 (100%, 6.3%, 0.5%)	3 (75%, 1.7%, 1.4%)	1 (100%, 5%, 0.5%)	5 (83.3%, 2.4%)		1 (25%, 0.6%, 0.5%)		1 (16.7%, 0.5%)	6 (2.9%)
Genitourinary			1 (100%, 5%, 0.5)	1 (100%, 0.5%)					1 (0.5%)
Hematuria			1 (100%, 5%, 0.5%)	1 (100%, 0.5%)					1 (0.5%)
Musculoskeletal	3 (100, 18.8%, 1.4%)	50 (57.5%, 28.9%, 23.9%)	1 (25%, 5%, 0.5%)	54 (57.5%, 25.8%)		37 (42.5%, 21.4%, 17.7%)	3 (75%, 15%, 1.4%)	40 (42.6%, 19.1%)	94 (45.0%)
Foot pain	1 (100%, 6.3%, 0.5%)	8 (50%, 4.6%, 3.8%)	1 (50%, 5%, 0.5%)	10 (52.6%, 4.8%)		8 (50%, 4.6%, 3.8%)	1 (50%, 5%, 0.5%)	9 (47.4%, 4.3%)	19 (9.1%)
Muscle/Joint soreness, tears, sprains/strains	2 (100%, 12.5%, 1.0%)	42 (59.2%, 24.3%, 20.1%)		44 (58.7%, 21.1%)		29 (40.9%, 16.8%, 13.9%)	2 (100%, 10%, 1.0%)	31 (41.3%, 14.8%)	75 (35.9%)
Neurological		5 (55.6%, 2.9%, 2.4%)	2 (100%, 10%, 1.0%)	7 (53.9%, 3.4%)	2 (100%, 12.5%, 1.0%)	4 (44.4%., 2.3%, 1.9%)		6 (46.2%, 2.9%)	13 (6.2%)
Dizziness		1 (50%, 0.6%, 0.5%)	2 (100%, 10%, 1.0%)	3 (75%, 1.4%)		1 (50%, 0.6%, 0.5%)		1 (25%, 0.5%)	4 (1.9%)
Headache/Migraine		3 (60%, 1.7%, 1.4%)		3 (42.9%, 1.4%)	2 (100%, 12.5%, 1.0%)	2 (40%, 1.2%, 1.0%)		4 (57.1%, 1.9%)	7 (3.4%)
Night sweats		1 (50%, 0.6%, 0.5%)		1 (50%, 0.5%)		1 (50%, 0.6%, 0.5%)		1 (50%, 0.5%)	2 (1.0%)
Nonspecific	3 (100, 18.8%, 0.01%)	4 (50%, 2.3%, 1.9%)	1 (50%, 5%, 0.5%)	8 (61.5%, 3.8%)		4 (50%, 2.3%, 1.9%)	1 (50%, 5%, 0.5%)	5 (38.5%, 2.4%)	13 (6.2%)
Biopsy pain	3 (100, 18.8%, 1.4%)	1 (20%, 0.6%, 0.5%)		4 (44.4%, 1.9%)		4 (80%, 2.3%, 1.9%)	1 (100%, 5%, 0.5%)	5 (55.6%, 2.4%)	9 (4.3%)
Cramping		2 (100%, 1.2%, 1.0%)		2 (100%, 1.0%)					2 (1.0%)
Fatigue		1 (100%, 0.6%, 0.5%)		1 (100%, 0.5%)					1 (0.5%)
Other			1 (100%, 5%, 0.5%)	1 (100%, 0.5%)					1 (0.5%)
Oral complaints		4 (80%, 2.3%, 1.9%)		4 (66.7%, 1.9%)	1 (100%, 6.3%, 0.5%)	1 (20%, 0.6%, 0.5%)		2 (33.3%, 1.0%)	6 (2.9%)
Tooth pain/mouth ulcers		4 (80%, 2.3%, 1.9%)		4 (66.7%, 1.9%)	1 (100%, 6.3%, 0.5%)	1 (20%, 0.6%, 0.5%)		2 (33.3%, 1.0%)	6 (2.9%)
Respiratory		4 (57.1%, 2.3%, 1.9%)	3 (100%, 15%, 1.4%)	7 (63.6%, 3.4%)	1 (100%, 6.3%, 0.5%)	3 (42.9%, 1.7%, 1.4%)		4 (36.4%, 1.9%)	11 (5.3%)
Nose bleed						1 (100%, 0.6%, 0.5%)		1 (100%, 0.5%)	1 (0.5%)
Sinus/cold/cough		4 (66.7%, 2.3%, 1.9%)	3 (100%, 15%, 1.4%)	7 (70%, 3.4%)	1 (100%, 6.3%, 0.5%)	2 (33.3%, 1.2%, 1.0%)		3 (30%, 1.4%)	10 (4.8%)
Special senses	2 (100%, 12.5%, 1.0%)	1 (100%, 0.6%. 0.5%)	1 (100%, 5%, 0.5%)	4 (100%, 1.9%)					4 (1.9%)
Fainting	2 (100%, 12.5%, 1.0%)		1 (100%, 5%, 0.5%)	3 (100%, 1.4%)					3 (1.4%)
Visual disturbance		1 (100%, 0.6%, 0.5%)		1 (100%, 0.5%)					1 (0.5%)
Total	11 (68.8%, 5.3%)	97 (56.1%, 46.4%)	12 (60%, 5.7%)	120 (57.4%)	5 (31.3%, 2.4%)	76 (43.9%, 36.4%)	8 (40%, 3.8%)	89 (42.6%)	209

Data were analyzed using χ² tests. PLA (n = 18), participants randomized to 750 mg sesame oil solution on day 8; TEST (n = 16), participants randomized to 750 mg testosterone undecanoate on day 8.

Energy Expenditure and Deficit

Average daily exercise-induced energy expenditure during P1 and P3 was similar between groups (TEST, P1: 384 ± 33 kcal/day; P3: 344 ± 31 kcal/day; PLA, P1: 368 ± 30 kcal/day; P3: 349 ± 31 kcal/day). Average daily exercise-induced energy expenditure during P2 was not different between groups (P = 0.177; Table 3; Supplemental Fig. S1), and was ∼1,500 kcal greater on high stress days than on low stress days (high stress P = 0.058; low stress P = 0.739). Ruck marching energy expenditure comprised 49.5% and 50.7% of daily total exercise-induced energy expenditure for TEST and PLA, respectively, on low stress days, and 48.6% and 49.8% for TEST and PLA, respectively, on high stress days, and was similar between groups (high stress P = 0.372; low stress P = 0.383). The mass of the ruck sacks for ruck marching exercises, designated on an individual basis at ∼30% of day 7 body mass, was similar between TEST (23.5 ± 1.0 kg) and PLA (23.6 ± 1.0 kg) (P = 0.935). Energy expenditure elicited via all other modalities was not different between groups. Total energy deficit throughout P2 exceeded 43,000 kcal and was not different between groups (P = 0.939).

Table 3.

Characteristics of energy expenditure, exercise modalities, and energy deficit during P2

	TEST, n = 16	PLA, n = 16	P Value
Overall
Average total daily energy expenditure, kcal/day	4,972 ± 81	5,157 ± 81	0.117
Average total exercise-induced energy expenditure, kcal/day	2,689 ± 20	2,728 ± 20	0.177
Outdoor ruck march, kcal/day	434 ± 17	436 ± 17	0.904
Indoor ruck march, kcal/day	877 ± 23	927 ± 23	0.133
Military activities, kcal/day	115 ± 3	120 ± 3	0.283
Walking, kcal/day	201 ± 13	189 ± 13	0.526
Running, kcal/day	36 ± 13	57 ± 13	0.254
Cycle ergometry, kcal/day	424 ± 25	410 ± 25	0.696
Elliptical, kcal/day	543 ± 30	534 ± 29	0.825
Yoga/stretching, kcal/day	39 ± 1	40 ± 1	0.795
Average absolute energy deficit, kcal/day	2,181 ± 17	2,163 ± 17	0.461
Average relative energy deficit, kcal/kg body mass/day	29 ± 1	29 ± 1	0.701
Low stress day
Average total daily energy expenditure, kcal/day	4,055 ± 103	4,187 ± 103	0.377
Average total exercise-induced energy expenditure, kcal/day	1,750 ± 22	1,761 ± 22	0.739
Outdoor ruck march, kcal/day	382 ± 22	370 ± 22	0.695
Indoor ruck march, kcal/day	478 ± 28	522 ± 28	0.267
Military activities, kcal/day	114 ± 3	119 ± 3	0.255
Walking, kcal/day	126 ± 14	129 ± 14	0.870
Running, kcal/day	10 ± 14	21 ± 14	0.565
Cycle ergometry, kcal/day	268 ± 28	250 ± 28	0.650
Elliptical, kcal/day	321 ± 32	303 ± 31	0.705
Yoga/stretching, kcal/day	40 ± 1	40 ± 1	0.854
Average absolute energy deficit, kcal/day	1,254 ± 35	1,200 ± 35	0.286
Average relative energy deficit, kcal/kg body mass/day	17 ± 1	16 ± 1	0.560
High stress day
Average total daily energy expenditure, kcal/day	5,618 ± 103	5,796 ± 103	0.233
Average total exercise-induced energy expenditure, kcal/day	3,315 ± 21	3,373 ± 21	0.058
Outdoor ruck march, kcal/day	468 ± 19	481 ± 19	0.638
Indoor ruck march, kcal/day	1,143 ± 26	1,196 ± 25	0.148
Military activities, kcal/day	117 ± 3	121 ± 3	0.337
Walking, kcal/day	250 ± 13	229 ± 13	0.270
Running, kcal/day	53 ± 13	80 ± 13	0.158
Cycle ergometry, kcal/day	528 ± 26	517 ± 26	0.764
Elliptical, kcal/day	692 ± 31	688 ± 30	0.924
Yoga/stretching, kcal/day	39 ± 1	40 ± 1	0.757
Average absolute energy deficit, kcal/day	2,806 ± 35	2,803 ± 35	0.955
Average relative energy deficit, kcal/kg body mass/day	38 ± 1	37 ± 1	0.772
Totals
Total energy expenditure, kcal	98,962 ± 2,104	102,898 ± 2,104	0.196
Total exercise-induced energy expenditure, kcal	53,279 ± 799	54,392 ± 799	0.332
Total energy deficit, kcal	43,208 ± 659	43,136 ± 659	0.939

Data were analyzed using a mixed-effect linear model. Values are presented as least squares means ± standard error. P2, phase 2; PLA (n = 16), participants randomized to 750 mg sesame oil solution on day 8; TEST (n = 16), participants randomized to 750 mg testosterone undecanoate on day 8.

Body Mass and Composition

Body mass was significantly decreased from P1 (day 7) by day 10 and continued to decrease throughout the remainder of P2 (Fig. 3). By the end of P2, body mass reached a minimum (mean change from P1 to P2: −4.31 ± 0.44 kg; P < 0.001). Body mass increased during P3 but was still lower than P1 values by the end of P3. There was no difference in body mass between groups (P = 0.977) or day × treatment interaction (P = 0.542).

Figure 3.

Body mass for the intent-to-treat population. Data were analyzed using a mixed-effect linear model. Data are presented as least squares means + standard error. For main effect of day, $significantly different from P1 (day 7) (P < 0.05). P1 values denoted in red outline. P1, phase 1; P2, phase 2; P3, phase 3; PLA (n = 18), participants randomized to 750 mg sesame oil solution on day 8; TEST (n = 16), participants randomized to 750 mg testosterone undecanoate on day 8.

A phase × treatment interaction was observed for 4-C FFM (P = 0.017). FFM decreased from P1 to P2 in PLA (−1.85 ± 0.69 kg; P = 0.011) but did not change in TEST from P1 to P2 (0.41 ± 0.65 kg; P = 0.534) or from P2 to P3 (−0.11 ± 0.60 kg; P = 0.850) (Table 4; Fig. 4A). FFM recovered from P2 to P3 in PLA (1.65 ± 0.66 kg; P = 0.016) and was not different from P1 by P3 (P = 0.827). A phase × treatment interaction was also observed for right arm lean mass, right arm bone mineral mass, and right arm total mass (all P < 0.05). Right arm lean mass was maintained in TEST throughout all phases but decreased throughout P2 in PLA. Right arm bone mineral mass was increased throughout P2 in TEST but remained the same in PLA. Right arm total mass decreased from P1 to P2 in both TEST and PLA, remained suppressed in P3 in TEST, but returned to P1 during P3 in PLA.

Table 4.

DXA and TBW outcomes for each study phase for the intent-to-treat population

	TEST, n = 16			PLA, n = 18			P Value
	P1	P2	P3	P1	P2	P3	Phase	Treatment	Phase × Treatment
4-C
FFM, kg^	59.2 ± 2.0	59.6 ± 2.0	59.5 ± 2.0	59.8 ± 2.1a	57.9 ± 2.1b	59.6 ± 2.1a	0.098	0.898	0.017
Fat, kg^	17.3 ± 1.5	13.0 ± 1.5	15.1 ± 1.5	17.5 ± 1.6	14.4 ± 1.6	15.5 ± 1.6	<0.0014	0.756	0.263
DXA
Bone mineral, kg	2.85 ± 0.12	2.89 ± 0.12	2.88 ± 0.12	2.73 ± 0.11	2.77 ± 0.11	2.76 ± 0.11	<0.0011,3	0.465	0.923
VAT, cm3	264.1 ± 22.9	216.0 ± 23.4	264.5 ± 22.9	286.6 ± 21.6	252.9 ± 22.8	304.4 ± 23.1	<0.0011,2	0.232	0.849
SAT, cm3	664.3 ± 83.2	442.5 ± 83.6	470.3 ± 83.2	740.8 ± 78.5	445.5 ± 79.6	515.4 ± 80.2	<0.0014	0.710	0.240
ALST, kg	27.1 ± 1.1	27.1 ± 1.1	27.3 ± 1.1	27.2 ± 1.1	26.7 ± 1.1	27.5 ± 1.1	0.003 2	0.964	0.092
Trunk
Lean, kg	28.2 ± 1.0	27.3 ± 1.0	28.0 ± 1.0	28.4 ± 0.9	27.3 ± 0.9	28.2 ± 0.9	<0.0011,2	0.906	0.707
Fat, kg	7.1 ± 0.6	5.4 ± 0.6	5.9 ± 0.6	7.4 ± 0.6	5.5 ± 0.6	6.4 ± 0.6	<0.0014	0.768	0.442
Bone mineral, kg	0.78 ± 0.04	0.79 ± 0.04	0.78 ± 0.04	0.72 ± 0.04	0.73 ± 0.04	0.72 ± 0.04	0.103	0.247	0.657
Total, kg	36.1 ± 1.4	33.6 ± 1.4	34.8 ± 1.4	36.5 ± 1.3	33.6 ± 1.3	35.3 ± 1.3	<0.0014	0.854	0.484
Left leg
Lean, kg	9.6 ± 0.4	9.6 ± 0.4	9.7 ± 0.4	9.6 ± 0.4	9.5 ± 0.4	9.7 ± 0.4	0.038 2	0.957	0.280
Fat, kg	3.0 ± 0.3	2.5 ± 0.3	2.6 ± 0.3	3.1 ± 0.3	2.7 ± 0.3	2.8 ± 0.3	<0.0014	0.707	0.205
Bone mineral, kg	0.53 ± 0.02	0.54 ± 0.02	0.54 ± 0.02	0.52 ± 0.02	0.53 ± 0.02	0.53 ± 0.02	<0.0011,3	0.745	0.239
Total, kg	13.2 ± 0.6	12.7 ± 0.6	12.9 ± 0.6	13.2 ± 0.6	12.7 ± 0.6	13.1 ± 0.6	<0.0011,2	0.899	0.453
Right leg
Lean, kg	9.8 ± 0.4	9.8 ± 0.4	9.8 ± 0.4	9.9 ± 0.4	9.7 ± 0.4	9.9 ± 0.4	0.312	0.963	0.195
Fat, kg	3.1 ± 0.3	2.6 ± 0.3	2.7 ± 0.3	3.1 ± 0.3	2.7 ± 0.3	2.9 ± 0.3	<0.0014	0.753	0.147
Bone mineral, kg	0.53 ± 0.02	0.54 ± 0.02	0.55 ± 0.02	0.53 ± 0.02	0.54 ± 0.02	0.53 ± 0.02	0.034 1	0.784	0.424
Total, kg	13.4 ± 0.6	12.9 ± 0.6	13.1 ± 0.6	13.5 ± 0.6	13.0 ± 0.6	13.3 ± 0.6	<0.0014	0.868	0.396
Left arm
Lean, kg	3.8 ± 0.2	3.7 ± 0.2	3.8 ± 0.2	3.7 ± 0.2	3.6 ± 0.2	3.7 ± 0.2	<0.0011,2	0.855	0.992
Fat, kg	1.0 ± 0.1	0.8 ± 0.1	0.9 ± 0.1	1.0 ± 0.1	0.8 ± 0.1	0.9 ± 0.1	<0.0014	0.871	0.456
Bone mineral, kg	0.22 ± 0.01	0.22 ± 0.01	0.22 ± 0.01	0.22 ± 0.01	0.21 ± 0.01	0.21 ± 0.01	0.749	0.712	0.703
Total, kg	5.0 ± 0.2	4.7 ± 0.2	4.9 ± 0.2	4.9 ± 0.2	4.7 ± 0.2	4.8 ± 0.2	<0.0014	0.923	0.945
Right arm
Lean, kg	4.0 ± 0.2a	3.9 ± 0.2a	3.9 ± 0.2a	4.0 ± 0.2a	3.8 ± 0.2b	4.0 ± 0.2a	<0.0011,2	0.888	0.005
Fat, kg	1.0 ± 0.1	0.8 ± 0.1	0.8 ± 0.1	1.0 ± 0.1	0.8 ± 0.1	0.9 ± 0.1	<0.0014	0.892	0.103
Bone mineral, kg	0.23 ± 0.01a	0.24 ± 0.01b	0.23 ± 0.01a	0.23 ± 0.01a	0.23 ± 0.01a	0.23 ± 0.01a	0.044 1	0.854	0.010
Total, kg	5.2 ± 0.2a	5.0 ± 0.2b	5.0 ± 0.2b	5.2 ± 0.2a	4.9 ± 0.2b	5.1 ± 0.2a	<0.0014	0.950	0.007
D2O
TBW (L)^	44.5 ± 1.7	44.9 ± 1.7	44.2 ± 1.7	44.5 ± 1.8	43.7 ± 1.8	46.2 ± 1.8	0.429	0.906	0.066

Data were analyzed using mixed-effect linear models. Values are presented as least squares means ± standard error. P values in bold are significant (P < 0.05). For phase × treatment interactions, data not sharing the same letter superscript within a treatment group are different by phase. For main effects of phase, ¹P1 is different from P2, ²P2 is different from P3, ³P1 is different from P3; ⁴All phases are different; ^smaller sample size due to equipment malfunction (TEST n = 15; PLA n = 14). ALST, appendicular lean soft tissue; 4-C, 4-compartment; D₂O, deuterium; DXA, dual-energy X-ray absorptiometry; FFM, fat-free mass; P1, phase 1; P2, phase 2; P3, phase 3; PLA (n = 18), participants randomized to 750 mg sesame oil solution on day 8; SAT, subcutaneous adipose tissue; TBW, total body water; TEST (n = 16), participants randomized to 750 mg testosterone undecanoate on day 8; VAT, visceral adipose tissue.

Figure 4.

Body composition [fat-free mass (A) and fat mass (B)] for the intent-to-treat population. Data were analyzed using mixed-effect linear models. Circles and squares represent individual data points, and bar graphs represent least squares means + standard error. For main effect of phase, $$significantly different from P1 and P3, $significantly different from P1 (P < 0.05). For phase × treatment interactions, time points not sharing the same letter are different. Figures display all data points collected for the intent-to-treat population. There was a smaller sample size for these variables due to equipment malfunction and missing data points (TEST n = 15; PLA n = 14); thus, figures are presented including all data points collected, and statistical models were analyzed as such. P1, phase 1; P2, phase 2; P3, phase 3; PLA, participants randomized to 750 mg sesame oil solution on day 8; TEST, participants randomized to 750 mg testosterone undecanoate on day 8.

A main effect of phase was observed for 4-C fat mass (P < 0.001; Fig. 4B), as well as most DXA variables examined (see Table 4 for details). 4-C fat mass decreased from P1 to P2 (−3.64 kg ± 0.45 kg; P < 0.001) and increased from P2 to P3 (1.56 kg ± 0.44 kg; P < 0.001), but was still lower than P1 values at P3 (P = 0.001). In general, other lean and fat mass variables decreased, and bone mineral mass increased from P1 to P2, with several variables returning to P1 by P3. No main effect of phase, treatment, or phase × treatment interaction was observed for D₂O TBW (P > 0.05).

Physical Performance

Vertical jump.

A main effect of phase was observed for vertical jump height (P = 0.002) (Fig. 5; Supplemental Table S2), which decreased from P1 to P2 (−4.5 ± 1.3 cm; P = 0.001), increased from P2 to P3, and was not different between P1 and P3. There was no difference in vertical jump height between groups or phase × treatment interaction. No main effect of phase, treatment, or phase × treatment interaction was observed for jump average power, partial average power, peak power, average velocity, peak velocity, or peak force (all P > 0.05).

Figure 5.

Selected physical performance variables [vertical jump height (A), 3-repetition maximum (RM) mass lifted (B), Wingate absolute peak power (C), absolute peak oxygen consumption during exercise (V̇o_2peak) (D), and ruck march total time (E)] for the intent-to-treat population. Data were analyzed using mixed-effect linear models. Circles and squares represent individual data points, and bar graphs are presented as least squares means + standard error. For main effect of phase, ^$$significantly different from P1 and P3, $significantly different from P3 only (P < 0.05). Figures display all data points collected for the intent-to-treat population. P1, phase 1; P2, phase 2; P3, phase 3; PLA (n = 18), participants randomized to 750 mg sesame oil solution on day 8; TEST (n = 16), participants randomized to 750 mg testosterone undecanoate on day 8.

3-RM deadlift.

A main effect of phase was observed for 3-RM total mass lifted (P = 0.011) (Fig. 5; Supplemental Table S3), which decreased from P1 to P2 (−6.8 ± 3.0 kg; P = 0.027), increased from P2 to P3, and was not different between P1 and P3. No main effect of treatment or phase × treatment interaction was observed. Main effects of treatment were observed for 3-RM deadlift peak power, average velocity, and peak velocity across all three repetitions, as well as RPE, such that TEST had greater values compared with PLA (all P < 0.05). However, no main effect of phase or phase × treatment interactions was observed for these variables (all P > 0.05). In addition, no main effect of phase, treatment, or phase × treatment interaction was observed for 3-RM deadlift average power, partial average power, or peak force (all P > 0.05).

Wingate anaerobic cycle test.

A main effect of phase was observed for peak power (absolute and relative), total work (absolute and relative), fatigue index, and maximum RPM (all P < 0.02) (Fig. 5; Supplemental Table S4). Peak power, total work (absolute and relative), and maximum RPM decreased, and fatigue index increased from P1 to P2, recovered from P2 to P3, and were not different between P1 and P3. Relative peak power also decreased from P1 to P2 but was not different in P2 compared with P3. A main effect of treatment was observed for fatigue index, such that TEST had greater values compared with PLA when phases were collapsed (P = 0.028). No phase × treatment interactions were observed for any of these variables (all P > 0.05).

Treadmill maximal cardiorespiratory fitness test (V̇o_2peak).

A main effect of phase was observed for absolute and relative V̇o_2peak and maximum RER (all P < 0.05) (Fig. 5; Supplemental Table S5). Absolute V̇o_2peak increased from P2 to P3 (0.20 ± 0.05 L/min, P < 0.001), but no differences were observed between P1 and P2 or P1 and P3. Relative V̇o_2peak was greater during P3 compared with P1 and P2, but there were no changes from P1 to P2. Maximum RER decreased from P1 to P2, increased from P2 to P3, and was not different between P1 and P3. No main effect of phase was observed for maximum grade. In addition, no main effects of treatment or phase × treatment interactions were observed for any of these variables.

Ruck march time trial.

A main effect of phase was observed for time, RPE, and HR at every half mile (all P < 0.05) (Fig. 5; Supplemental Table S6). Time to complete every half mile increased from P1 to P2, decreased from P2 to P3, and was not different between P1 and P3. HR at every half mile decreased from P1 to P2 and increased from P2 to P3. HR remained lower during P3 at 0.5, 1.5, and 2.5 miles compared with P1. RPE at 0.5, 2.0, and 2.5 miles increased from P1 to P2 and decreased from P2 to P3. RPE at 1.0 and 1.5 miles decreased from P2 to P3. RPE at 1.5, 2.0, and 2.5 miles was also greater during P1 than P3. A main effect of treatment was observed for RPE at 2.5 miles, such that TEST had greater values compared with PLA when all phases were collapsed (P = 0.045), but no other main effects of treatment, and no phase × treatment interactions were observed for any variable.

Endocrine Biomarkers

Day × treatment interactions were observed for changes in total and free testosterone concentrations (both P < 0.001; Fig. 6, A and B). In TEST, total and free testosterone concentrations increased from P1 (day 7) and were greater than PLA by day 12 (P < 0.001), which persisted throughout P2. Free testosterone concentrations in PLA were lower than P1 values on days 28 (P = 0.050) and 29 (P = 0.036), but returned by day 32. However, by day 26, total and free testosterone concentrations in TEST were not different from P1 and were not different between groups on day 29 (day following the end of P2). On days 32 and 38, free testosterone concentrations in TEST were lower than P1 (P = 0.043 and P = 0.019, respectively). By day 44, total and free testosterone concentrations were not different between TEST and PLA. The AUC for both total and free testosterone concentrations was greater in TEST than PLA during P2 (both P < 0.001), but were not different between groups in P3 (both P > 0.05).

Figure 6.

Total (A) and free testosterone (B) concentrations for the intent-to-treat population. Data were analyzed using mixed-effect linear models. Data are presented as least squares means + standard error. For day × treatment interactions, *significant difference (P < 0.05) between TEST and PLA, ^time point for TEST significantly different from P1 (day 7), #time point for PLA significantly different from P1 (day 7). P1 values denoted in red outline. P1, phase 1; P2, phase 2; P3, phase 3; PLA, participants randomized to 750 mg sesame oil solution on day 8; TEST, participants randomized to 750 mg testosterone undecanoate on day 8.

Day × treatment interactions were observed for FSH, LH, and estradiol concentrations (all P < 0.001; Supplemental Fig. S2, A–C). Both FSH and LH concentrations in TEST were lower than P1 and were different than PLA by day 12, which continued to decrease throughout P2. Concentrations began to increase in TEST throughout P3 and reached P1 values by the end of P3. The AUC for both FSH and LH concentrations during P2 and P3 was greater in PLA than TEST (P ≤ 0.001). Conversely, estradiol concentrations in TEST were greater than P1 by day 9 and peaked on day 12. Estradiol concentrations remained unchanged from P1 throughout P3 in TEST and PLA. The AUC for estradiol during P2 was greater in TEST than PLA (P = 0.001) but was not different between groups in P3 (P = 0.874).

Main effects of day were observed for SHBG, IGF-1, GH, insulin, and cortisol concentrations throughout P2 and P3 (all P < 0.001; see Supplemental Fig. S2, D–H for details). Cortisol concentrations were greater in PLA than TEST (P = 0.002; Supplemental Fig. S2H). The AUC for SHBG, IGF-1, GH, insulin, and cortisol concentrations during both P2 and P3 was not different between groups (all P > 0.05), and no day × treatment interactions were observed for these hormones.

DISCUSSION

The primary findings from this study demonstrate that administering one injection of testosterone undecanoate (750 mg) to young, healthy males immediately before a 20-day, simulated, sustained military operation increased circulating testosterone concentrations within the normal range and prevented FFM loss. However, physical performance deteriorated similarly in both groups across metrics of muscular strength and power, anaerobic and aerobic capacity, and ruck march duration. These findings are strengthened by the use of a carefully controlled and designed live-in environment, where exercise, diet, and sleep were precisely monitored. The total energy deficit, which was prescribed to exceed 43,000 kcal over the course of P2 to result in physiological consequences (45), exceeded this amount and produced significant reductions in FFM and physical function. Taken together with our previous work in OPS I (16, 17), administering testosterone immediately before or during a simulated, sustained military operation (21–28 days) increases circulating testosterone and positively affects lean mass but does not affect physical performance.

The anabolic effects of exogenous testosterone administration on FFM are well documented (46–48), but the magnitude of muscle mass gain is dependent on the testosterone dose, frequency and duration of use, formulation, and administration route (48–50). We report a 2.3 kg mean difference in the change in FFM from P1 to P2 between TEST and PLA, such that PLA lost 1.9 kg, and TEST gained 0.4 kg. Similarly, OPS I reported a 2.8 kg mean difference in lean mass from P1 to P2 between groups. This difference in OPS I was attributed primarily to lean mass gain in those receiving testosterone rather than an attenuation of lean mass loss as observed in OPS II. However, changes in extracellular water or glycogen content could conceivably have contributed to the increase in lean mass in OPS I given the use of solely DXA-derived methods of body composition. The use of a 4-C body composition method in OPS II, which was not used in OPS I, reduces the error in fat mass estimation, results in more accurate FFM estimates than DXA, and confirms that FFM loss is attenuated in TEST compared with PLA. In addition, although we observed a significant increase in right arm bone mineral mass in TEST throughout P2, this likely cannot be accounted for with the crude measures of bone mass used in the current study. Future research using more specific techniques for assessing bone quality such as high-resolution peripheral quantitative computed tomography (HR-pQCT) may be necessary to elucidate these findings.

The maintenance of FFM and elevated circulating testosterone concentrations in TEST did not translate to physical performance improvements. Within the larger body of research, the impact of testosterone administration on performance is mixed, as some studies report no effects on performance (17, 51), whereas meta-analyses report improvements only in muscular strength or power, or only on physical function or aerobic capacity (46, 48, 52, 53). The current study used a battery of reproducible and validated performance metrics to ensure that all energy systems were examined. The inability to detect group differences among any of these metrics suggests that factors other than muscle mass contribute to performance sustainment. Muscle strength is not only a factor of its size, as neuromuscular activation, noncontractile tissue content, muscle architecture, metabolism, and fiber type distribution also affect force output (54). These factors, which comprise muscle quality, require further exploration in the context of testosterone administration and physical performance.

The inability to augment performance after only 3 wk may be due to study duration as improvements following testosterone administration may take several months to manifest despite changes in FFM (46, 55). Short-term (≤4 wk) supplementation studies in young, healthy males consistently report increased FFM with no functional improvement (17, 51, 56); however, longer duration protocols with higher testosterone doses may increase both FFM and performance. Bhasin et al. (57) reported increases in FFM, muscle size, and strength in young males (19–40 yr) following 10 wk of testosterone administration (600 mg/wk intramuscular testosterone enanthate). Likewise, 20 wk of testosterone administration (300 or 600 mg/wk of testosterone enanthate) increased FFM and leg press strength in both younger (18–35 yr) and older males (60–75 yr) (49, 58). However, moderate doses (125 mg/wk) increased FFM but did not increase leg press strength, and low doses (25 and 50 mg/wk) did not result in changes to these parameters (49, 58). Taken together, these findings suggest that higher doses provided over a longer duration may be necessary for functional improvement (50, 55).

The lack of performance improvements in TEST also provides valuable information about the limitations of the current investigation and the design of future studies attempting to replicate sustained military operations in laboratory settings. The number of physical performance indices included in this report increases the risk for type II error, and it is possible that this study was underpowered to detect any group differences in performance outcomes. Future studies with larger sample sizes are necessary to determine the effects of testosterone administration on physical performance. In addition, the high frequency of musculoskeletal adverse events limits the findings of this investigation, suggesting that participants were not adequately prepared to endure the sustained military operation, which may have influenced physical performance outcomes. Although participants were physically active before the study, most did not have military experience and were not specifically trained on the exercises used in P2, and future research is necessary in individuals with previous military-type experience. Ruck marching, which comprised ∼50% of total exercise-induced energy expenditure, was reported to be painful for most participants and was likely the source of most musculoskeletal injuries. The lack of a functional improvement in TEST may be due to cumulative injury sustained from 10 to 15 h of exercise per day over 20 days, coupled with inadequate recovery, resulting in an inability to complete performance tests to the maximal ability by the end of P2. Reduced motivation during the performance tests at the end of P2 may have also contributed to the lack of performance benefits. Furthermore, although the intended 43,000 kcal absolute energy deficit throughout the 20-day military operation was achieved and exceeded that elicited in OPS I, study controls lost less total mass (−6.0%) than the anticipated 9% reduction based on a meta-regression of data generated from military field studies (45). Nevertheless, physical performance was significantly impaired across all indices examined, demonstrating that this intervention was sufficient at eliciting sufficient stress to impair physical function.

Previous studies conducted in real-world sustained military operations report robust increases in hypothalamic-pituitary-adrenal axis activity, including sympathetic nervous system and adrenal gland activation and hypothalamic-pituitary-gonadal axis suppression, resulting in a hypogonadal state (2–7). However, in both OPS I and OPS II, control subjects maintained eugonadal total testosterone concentrations. OPS II employed additional physiological and psychological stressors intended to simulate real-life military operations and therefore impact endogenous testosterone to a greater degree than that imposed in OPS I, including sleep restriction, prolonged aerobic activity, and a greater cumulative energy deficit (9–12). However, we were still unable to replicate the hypogonadal state that is characteristic of military personnel during real-life sustained military operations (4, 6, 7, 32), which is a limitation of the current investigation. It is possible that a lack of sufficient psychological stress compared with what is typically observed in field training and combat operations (i.e., live fire, drill sergeants, casualties, etc.) may have contributed to these findings, although these are extremely difficult to replicate in laboratory settings. Nevertheless, other physiological responses observed in control subjects (i.e., decreases in body mass, FFM, fat mass, free testosterone concentrations, and physical performance; increased circulating cortisol and SHBG concentrations) indicate that subjects endured substantial stress, which were comparable with those observed during US Army Ranger School and Special Operations Forces training (2–7). In addition, recent research suggests that free testosterone concentrations should be prioritized over total testosterone in conditions where SHBG concentrations change, as circulating testosterone bound to SHBG is inactive (59, 60). Examination of total testosterone alone may not adequately reflect anabolic status, and a comprehensive examination of multiple blood parameters, specifically free testosterone concentrations, is necessary to provide a clear depiction of the true stress of the simulated military operation. Notably, despite maintenance of total testosterone, the significant decreases in free testosterone throughout P2 in PLA that fell below the normal range for healthy young males (250 pmol/L) (1, 44) suggest diminished anabolic status. Therefore, the inability to reproduce decrements in total testosterone concentrations should not imply that the stress imposed in OPS II was insufficient at eliciting the magnitude of stress typically endured by military personnel. Other physiological indices reported here demonstrate that OPS II adequately replicated the physiological decrements associated with operational stress.

Importantly, the testosterone dose and formulation used in OPS II elicited changes in testosterone concentrations that remained within the Endocrine Society reference range for total testosterone (9.2–31.8 nmol/L) (59) and the Vermeulen reference ranges for total and free testosterone (10.4–34.7 nmol/L and 250–785 pmol/L, respectively) (1, 44, 61) throughout the entire study. Concentrations also returned to P1 values by the end of P3, indicating that endocrine status was maintained in recovery. In contrast, the testosterone formulation used in OPS I elicited increases in total and free testosterone above the normal range, followed by a hypogonadal state in recovery (17). The single dose of testosterone undecanoate used in the current study is also more operationally practical than the weekly intramuscular injections of a supraphysiological dose of testosterone enanthate used in OPS I, as the latter may not be adopted due to possible safety and ethical concerns. This evidence demonstrating a positive effect of clinical testosterone undecanoate on body composition during short-term operations may inform leaders and clinicians when considering the safety and efficacy of various pharmacological strategies to optimize health and performance for Special Operators. Of importance, we do not intend to suggest the use of exogenous testosterone administration for all Soldiers, and these findings are only relevant to male military personnel. Future research is necessary to determine other potential interventions for mitigating muscle mass loss and performance declines in female military personnel.

In conclusion, this study demonstrated that a single dose of testosterone undecanoate (750 mg) before a 20-day simulated, multi-stressor military operation increased circulating testosterone concentrations and did not attenuate performance decrements; however, there was a substantial protective effect on FFM. The findings from OPS I and OPS II also provide important rationale, framework, and suggestions for the design of future research designed to examine testosterone administration for mitigating declines in military performance. Taken together, the ability of testosterone administration to spare FFM loss during severe, multi-stressor military operations as demonstrated in this study and in OPS I may be important for military personnel conducting strenuous field operations.

SUPPLEMENTAL DATA

Supplemental Tables S1–S6 and Figs. S1 and S2: https://doi.org/10.6084/m9.figshare.17155544.

GRANTS

The US Army Medical Research and Development Command, Military Operational Medicine Research Program Grant W81XWH-19-C-0162 (S. M. Pasiakos) funded this research. Supported in part by an appointment to the US Army Research Institute of Environmental Medicine administered by the Oak Ridge Institute for Science and Education through an interagency agreement between the US Department of Energy and the US Army Medical Research and Development Command.

DISCLAIMERS

The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the Army or the Department of Defense. Any citations of commercial organizations and trade names in this report do not constitute an official Department of the Army endorsement of approval of the products or services of these organizations. The funding sources had no role in the study design; collection, analysis, and interpretation of data; in writing the report; and in the decision to submit this article for publication.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

A.N.V., N.M.J., S.B.H., F.L.G., L.M.M., H.R.L., A.A.F., S.M.P., and J.C.R. conceived and designed research; A.N.V., M.N.H., C.H., E.E.H., S.B.H., F.L.G., D.D.C., and J.C.R. performed experiments; A.N.V. and R.A.B. analyzed data; A.N.V. and R.A.B. interpreted results of experiments; A.N.V. prepared figures; A.N.V. drafted manuscript; A.N.V., E.E.H., N.M.J., S.B.H., F.L.G., L.M.M., H.R.L., D.D.C., A.A.F., S.M.P., and J.C.R. edited and revised manuscript; A.N.V., M.N.H., C.H., E.E.H., N.M.J., S.B.H., F.L.G., L.M.M., H.R.L., R.A.B., D.D.C., A.A.F., S.M.P., and J.C.R. approved final version of manuscript.