The combination of male gender, obstructive sleep apnoea (OSA) and obesity magnifies cardiometabolic risk. There has been no systematic study evaluating whether testosterone therapy can improve cardiometabolic health in obese men with OSA by improving body composition, visceral abdominal fat and insulin sensitivity.
To assess body compositional and cardiometabolic effects of testosterone treatment in obese men with severe OSA.
An 18-week randomised, double-blind, placebo-controlled and parallel group trial in 67 men.
Participants (age=49±12 years, apnoea hypopnoea index=39.9±17.7 events/h, BMI=31.3±5.2 kg/m2) were placed on a hypocaloric diet and received i.m. injections of either 1000 mg testosterone undecanoate (n=33) or placebo (n=34) for 18 weeks. Outcomes were the changes in body composition (total muscle mass, total and abdominal fat, total body dual-energy X-ray absorptiometry and computerised tomography (CT)), weight, insulin sensitivity (homeostasis model assessment), abdominal liver fat (CT), arterial stiffness (pulse wave analysis), resting metabolic rate and respiratory quotient (indirect calorimetry) and blood lipids and metabolic syndrome from baseline to week 18.
After 18 weeks, testosterone treatment increased insulin sensitivity (−1.14 units, 95% confidence interval (95% CI) −2.27 to −0.01, P<0.05), reduced liver fat (0.09 Hounsfield attenuation ratio, 95% CI 0.009 to 0.17, P=0.03) and increased muscle mass (1.6 kg, 95% CI 0.69 to 2.5, P=0.0009) to a greater extent than placebo. Other measures of body composition and regional adiposity as well as the number of participants with metabolic syndrome did not change. Testosterone also decreased arterial stiffness (augmentation index) by 3.2% (95% CI −6.01 to −0.46%, P=0.02) and decreased the respiratory quotient (95% CI −0.04, −0.08 to −0.001, P=0.04) after 18 weeks compared with placebo.
Eighteen weeks of testosterone therapy in obese men with OSA improved several important cardiometabolic parameters but did not differentially reduce overall weight or the metabolic syndrome. Longer term studies are required.
Obesity occurs in 32% of all men in the USA (1) and many of these men develop obstructive sleep apnoea (OSA) (2, 3). The resulting intermittent hypoxia amplifies the cardiometabolic risk of obesity by promoting atherogenesis through mechanisms that include inhibition of triglyceride clearance, inactivation of lipoprotein lipase, and induction of insulin resistance (IR) (4, 5, 6). Indeed, all-cause (7, 8) and cardiovascular (8, 9) mortality is amplified when untreated OSA is combined with coexisting obesity and male gender. Understanding the mechanisms and methods by which cardiometabolic risk may be reduced in this specific obese population is therefore important.
Testosterone treatment increases muscle and decreases fat in hypogonadal and eugonadal males (10), as well as in men with relative androgen deficiency related to age (11) or HIV (12). In addition, testosterone therapy decreases visceral abdominal fat (VAF) (13, 14, 15) and improves insulin sensitivity (16, 17, 18, 19, 20) in some but not in all (21, 22, 23, 24, 25, 26) studies. Specific characteristics of the populations studied may contribute to these discrepancies. Improved body composition would be expected to improve cardiometabolic health including insulin sensitivity, VAF, non-alcoholic fatty liver disease (NAFLD) and metabolic syndrome. No single study has assessed the effect of testosterone on both radiographically measured VAF and liver fat (16). Few studies have assessed the impact of testosterone therapy on the metabolic syndrome (16). Conversely, biochemical androgen deficiency is associated with increased cardiovascular events (26) and testosterone therapy has been shown to improve arterial stiffness, an early marker of atherosclerosis, in men with coronary artery disease or advanced age (27, 28).
Biochemical androgen deficiency is common in men with OSA (29, 30), is more severe with increasing hypoxemia and can be partly reversed by continuous positive airway pressure therapy (31). Biochemical androgen deficiency is also common in obese men (29), especially with increasing adiposity (32) and can be partly reversed with weight loss (33, 34). These data suggest that men who are obese or who have OSA have some degree of relative androgen deficiency (29). Whether testosterone therapy is beneficial in these men requires direct evaluation by randomised placebo-controlled trials in this population.
Previous research has suggested that testosterone therapy may improve cardiometabolic risk in older men with chronic heart failure (18), abdominally obese men (14, 15) and men with the metabolic syndrome and/or diabetes (16, 17, 20, 35, 36) but may increase cardiovascular risk in frail older men with limited mobility (37). As the effect of testosterone therapy on cardiometabolic risk has not been systematically studied in obese men with OSA who are at magnified cardiometabolic risk and who have some degree of relative androgen deficiency, we undertook a randomised placebo-controlled trial to directly determine this. We measured change in anthropometry, abdominal visceral fat, total body fat and lean muscle, basal metabolic rate, insulin sensitivity, blood lipids and metabolic syndrome status. These assessments included more precise measures of cardiometabolic risk including radiographically determined liver fat and tonometry-determined arterial stiffness. The effects of testosterone therapy on sleep and breathing in this study have recently been published (38).
Materials and methods
Participants and study design
The entry criteria have been previously described in our report of respiratory effects (38). In brief, eligible participants were adult obese men with OSA, defined as aged 18 years or over, BMI >30 kg/m2 and apnoea hypopnoea index >10 events/h by in-laboratory polysomnography.
This was a randomised double-blind, placebo-controlled and parallel group study. Eligible participants were randomly assigned to receive i.m. injections of either Reandron (testosterone undecanoate 1000 mg in 4 ml castor oil vehicle, Bayer Schering) or 4 ml oil vehicle placebo at 0, 6 and 12 weeks. All participants were concurrently enrolled in a lifestyle modification programme in which a dietitian prescribed a 2500 kJ (600 kcal) daily deficit diet based on a macronutrient content of <30% fat (with emphasis on reducing saturated fat in the diet), 15% protein and at least 55% carbohydrate. Advice on behavioural modification and physical activity was also provided (38). At randomisation, subjects received a number in ascending order that corresponded with a pre-numbered and pre-packaged treatment kit. A computer-generated randomisation list was created using a block size of four. Subjects were assigned to active medication or placebo in a 1:1 ratio. All researchers involved in the conduct of the study were blinded to the treatment allocation for the duration of the study.
All subjects provided written informed consent and both the patient information sheet and study protocol were approved by the Sydney South West Area Health Service Human Research and Ethics Committee (Royal Prince Alfred Hospital Zone). The study is registered with the Australia New Zealand Clinical Trials Network, no. ACTRN12606000404527.
Height and weight as well as waist, hip, neck, mid-arm and mid-thigh circumferences were measured according to anthropometric recommendations (39) at each visit by a single observer.
Body composition included measures of abdominal, liver and total fat and total lean muscle mass by computerised tomography (CT) and dual-energy X-ray absorptiometry (DXA) scanning that were performed at 0 and 18 weeks. Two 10 mm contiguous images of both the upper (T12) and lower (L4) abdomen were obtained by CT (16 slice light speed extra, GE Healthcare, Madison, WI, USA). The volumes of subcutaneous abdominal fat (SCAF) and VAF respectively in the abdomen were quantified using a validated software tool (Hippofat; CNR Institute of Clinical Physiology, Pisa, Italy) (40). Total fat and lean mass were measured by total body DXA (Lunar Prodigy; GE Healthcare).
Liver fat was estimated by liver and spleen Hounsfield unit attenuations (E-FILM, Merge Healthcare, Milwaukee, WI, USA) from unenhanced CT slices in several homogenous areas (≥10 cm2) that were adjacent to the T12/L1 intervertebral space and free of vasculature. The liver to spleen attenuation ratio (L/S ratio) was used to estimate liver fat (41). A higher ratio indicates a lower liver fat content. Arterial stiffness and central blood pressure were measured at 0, 6 and 18 weeks in the supine position by pulse wave analysis (Sphymo-Cor; AtCor Medical, Sydney, NSW, Australia), a method that our group has previously implemented in another study (41). A validated transfer algorithm was used to calculate the aortic augmentation index (AIx), which is the ratio of the augmentation pressure (due to the reflected component of the pulse pressure wave) to pulse pressure, expressed as a percentage. Resting metabolic rate and the respiratory quotient were determined by indirect calorimetry (Delta Trak II Metabolic Monitor, Datex-Ohmeda, Inc., Madison, WI, USA).
Fasting venous blood was collected in the early morning at each visit. Standard biochemistry and haematology panels, insulin, leptin, sex hormone-binding globulin and prostate-specific antigen (PSA) were analysed by platform assays while LDL-cholesterol was calculated by the Friedewald equation (42). Blood glucose was determined by the hexokinase method (Modular P; Roche Diagnostics) from blood collected in sodium fluoride (Vacutainer; Becton Dickinson, Rutherford, NJ, USA). Insulin sensitivity was calculated by the homeostasis model assessment of IR (HOMA-IR) method (43).
LH and FSH concentrations were measured by commercially available Delfia assays (Perkin-Elmer Life Sciences, Rowville, VIC, Australia) and total testosterone by mass spectrometry (API-5000 triple-quadruple; Applied Biosystems/MDS SCIEX, ON, Canada). Free testosterone was calculated using a mass equation (44). Within-assay coefficients of variation was <10% for all assays.
Determination of metabolic syndrome and other measurements
The metabolic syndrome was defined according to international consensus guidelines (45) and the National Cholesterol Education Program's Adult Treatment Program Panel Criteria III using the previously described anthropometric, blood pressure and blood parameters (46). A score was given to each individual criterion and then summed together. Metabolic syndrome was defined by a total score ≥3. At all visits, physical activity was assessed by the physical activity scale for the elderly questionnaire and handgrip strength by dynamometer (JAMAR; JA Preston Co., Jackson, MI, USA) as described previously (47). Lower urinary tract symptoms were assessed at the same time by International Prostate Symptom Score (48). At each visit, subjects were asked to report adverse events, including injection-related adverse events such as bruising, stinging or pain. Adverse events were categorised according to the Medical Dictionary for Regulatory Activities (MedDRA) System Organ Class classification.
The randomisation code was not broken until all data were collected and the database cleaned and locked. The outcome variables for continuous variables were the calculated differences from baseline at 6, 12 and 18 weeks. The between-group differences of all changes were assessed by linear regression. Outcomes with repeated measures were assessed by mixed model analyses and tested treatment, time and the interaction of treatment and time. Analyses were adjusted by factors that, despite randomisation, significantly differed between groups at baseline and were also a priori confounders of the outcome of interest. Normality of residuals was confirmed.
The presence (or absence) of metabolic syndrome was assessed at every visit. The change in the proportion of participants with or without the metabolic syndrome from baseline was analysed separately at each subsequent visit by examining the treatment by syndrome interaction. Repeated measurements were modelled in these analyses that utilised generalised estimating equations that were then confirmed by Bayesian methods.
Additional analysis explored the influence of baseline testosterone on treatment effects in order to assess potential thresholds of relative androgen deficiency that may identify subgroups in whom testosterone may have enhanced effects. Baseline total and free testosterone were included in separate regression models as a dichotomised factor using predefined cut points for total testosterone (8, 11 and 13 nmol/l) and free testosterone (160, 220 and 280 pmol/l), recently shown to be useful in the definition of late-onset hypogonadism (32), which included the definition stated in the current guidelines (total testosterone, 8 nmol/l) (49). The statistical significance of the interaction terms of treatment and each dichotomised variable was examined.
Analyses were performed using SAS version 9.2 (SAS Institute, Cary, NC, USA). Data were considered significantly different at P<0.05 (two sided) and are presented as mean differences±s.e.m., mean±s.d. or mean (95% confidence interval (95% CI)), as indicated.
Eighty-one men were screened of whom 67 were randomised to receive either testosterone (n=33) or matching placebo (n=34). Thirteen participants discontinued leaving 26 men in the testosterone group and 28 men in the placebo group who completed the entire 18-week treatment period. The flow of participants through the study has been previously published (38). Baseline characteristics were comparable except testosterone-treated men had lower waist circumference (Table 1). Pre-existing cardiovascular disease was reported in three participants: acute myocardial infarction (n=1, placebo), rheumatic valvular disease (n=1, placebo) and transient ischaemic attack (n=1, testosterone). There were no differences in the use of lipid-lowering medications between the groups at baseline (Table 1). All participants remained on these medications for the duration of the study and no other participants started lipid-lowering treatment during the course of the study. MedDRA-defined adverse events were similar between the two groups (Table 2).
Baseline participant characteristics (mean±s.d.). Values are mean±s.d. or number (percent) as specified.
|Testosterone (n=33)||Placebo (n=34)||P value*|
|Free testosterone (nmol/l)||0.29±0.02||0.28±0.07||0.81|
|Waist circumference (cm)||115.7±8.8||120.7±11.1||0.04|
|Hip circumference (cm)||115.8±9.2||118.9±10.1||0.19|
|Neck circumference (cm)||43.9±3.0||44.1±2.6||0.79|
|Total lean mass (kg)||63.1±6.4||65.5±7.8||0.20|
|Total fat mass (kg)||37.9±8.3||40.0±7.6||0.33|
|Metabolic syndrome (n (%))||17 (53)||15 (48)||0.80|
|Systolic blood pressure (mmHg)||129.1±12.1||128.0±14.3||0.73|
|Diastolic blood pressure (mmHg)||82.3±11.8||82.9±11.8||0.85|
|Heart rate (b.p.m.)||71.4±10.1||73.3±8.5||0.41|
|Liver fat (HU)||0.9±0.06||0.7±0.3||0.12|
|Hyperlipidemia (n (%))||6 (18)||4 (12)||0.51|
|Statin therapy (n (%))||5 (15)||4 (12)||0.73|
|Total cholesterol (mmol/l)||5.3±1.0||5.2±0.9||0.71|
|Total cholesterol/HDL-C ratio||4.8±0.3||4.6±0.2||0.68|
|Prostate and haematology|
|IPSS (scores between 1 and 7)||5.0±0.68||4.5±0.85||0.63|
*The P values were calculated by Student's t-test or Fisher's exact test as appropriate. SHBG, sex hormone-binding globulin; SCAF, subcutaneous abdominal fat; VAF, visceral abdominal fat; AIx, augmentation index; CAP, central augmentation pressure; PPA, peripheral pulse amplification; REE, resting energy expenditure; RQ, respiratory quotient; HU, Hounsfield units; HOMR-IR, homeostasis model assessment of insulin resistance; ALT, alanine transaminase; AST, aspartate transaminase; PSA, prostate-specific antigen; IPSS, International Prostate Symptom Score; HCT, haematocrit; HB, haemoglobin.
|Number of events|
|Adverse event category||Testosterone (n=33)||Placebo (n=34)||P value*|
|All (n (%))||24 (73)||24 (71)||1.0|
|Administration site (n (%))||7 (21)||4 (12)||0.34|
|Cardiac (n (%))||1 (3)||0 (0)||0.49|
|Respiratory (n (%))||12 (36)||11 (32)||0.80|
|Musculoskeletal (n (%))||7 (21)||7 (21)||1.0|
|Renal and urinary (n (%))||1 (3)||0||0.61|
*P values from Fisher's exact test.
Hormones and cardiometabolic effects
Testosterone therapy increased blood testosterone concentrations and suppressed gonadotropins, often to the detection limit of the assay (Table 3). Arterial stiffness (i.e. AIx) was significantly less in the testosterone-treated group compared with the placebo group overall (Table 3) and at 18 weeks (Fig. 1A). Central augmentation pressure was also improved after testosterone compared with placebo (Table 3). There were no changes in heart rate, systolic or diastolic blood pressure in either group after 18 weeks of treatment (Table 3).
Changes from baseline in testosterone and placebo treated groups.
|Unadjusted mean change from baseline|
|Testosterone (n=33)||Placebo (n=34)||Adjusted between-group mean differences (95% CI)||P value*|
|Hormones and SHBG|
|LH (U/l)||−2.90||0.52||−2.98 (−3.84 to −2.11)||<0.0001|
|FSH (U/l)||−3.53||0.23||−3.37 (−4.43 to −2.30)||<0.0001|
|Testosterone (nmol/l)||3.83||0.69||5.73 (3.63 to 7.84)||<0.0001|
|Free testosterone (nmol/l)||0.11||0.004||0.11 (0.06 to 0.16)||<0.001|
|SHBG (nmol/l)||0.37||1.85||−1.35 (−3.29 to 0.59)||0.17|
|Leptin (ng/ml)||−5.66||−4.45||−1.27 (−4.74 to 2.19)||0.47|
|Weight (kg)||−1.54||−2.15||0.85 (−0.33 to 2.03)||0.16|
|BMI (kg/m2)||−0.50||−0.71||0.30 (−0.10 to 0.69)||0.14|
|Waist circumference (cm)||−3.26||−3.51||0.25 (−1.27 to 1.76)||0.75|
|Hip circumference (cm)||−0.82||−1.42||0.61 (−1.28 to 2.49)||0.52|
|Waist–hip ratio||−0.02||−0.019||−0.002 (−0.02 to 0.02)||0.81|
|Total lean mass (kg)||1.2||−0.4||1.6 (0.69 to 2.5)||0.0009|
|Total fat mass (kg)||−3.1||−2.9||−0.02 (−1.4 to 1.4)||0.98|
|SCAF (cm3)||−70.1||−48.3||−21.8 (−67.1 to 23.5)||0.34|
|VAF (cm3)||−29.6||−44.4||23.1 (−23.4 to 70.0)||0.32|
|Systolic blood pressure (mmHg)||1.7||0.5||2.3 (−3.0 to 7.6)||0.39|
|Diastolic blood pressure (mmHg)||0.64||0.69||0.32 (−4.4 to 5.1)||0.89|
|Heart rate (b.p.m.)||−2.0||−2.6||1.3 (−3.3 to 5.8)||0.58|
|AIx (%)||−0.79||2.20||−3.24 (−6.01 to −0.46)||0.02|
|CAP (mmHg)||−0.27||0.94||−1.28 (−2.56 to −0.004)||<0.05|
|PPA||0.03||−0.02||0.04 (−0.0002 to 0.09)||0.06|
|REE (kcal)||81.6||24.1||42.6 (−69.3 to 154.5)||0.45|
|RQ||−0.04||0.01||−0.04 (−0.08 to −0.001)||0.04|
|Liver fat (HU)||0.11||0.05||0.09 (0.009 to 0.17)||0.03|
|HOMA-IR (units)||−1.30||−0.37||−1.14 (−2.27 to −0.01)||<0.05|
|Insulin (μIU/ml)||−29.1||−9.33||−23.6 (−51.8 to 4.57)||0.10|
|Glucose (mmol/l)||−0.36||−0.05||−0.22 (−0.71 to 0.26)||0.36|
|Total cholesterol (mmol/l)||0.01||−0.10||0.12 (−0.14 to 0.38)||0.35|
|Total cholesterol/HDL-C ratio||0.13||0.001||0.18 (−0.14 to 0.50)||0.26|
|HDL-C (mmol/l)||−0.06||−0.04||−0.02 (−0.09 to 0.05)||0.57|
|LDL-C (mmol/l)||0.21||−0.06||0.27 (0.03 to 0.54)||0.03|
|Triglycerides (mmol/l)||−0.29||−0.16||−0.01 (−0.45 to 0.42)||0.95|
|ALT (U/l)||−2.79||−1.14||−1.65 (−4.87 to 1.57)||0.31|
|AST (U/l)||1.68||1.63||0.04 (−4.20 to 4.28)||0.98|
|Prostate and haematology|
|PSA (ng/ml)||0.21||−0.03||0.24 (0.10 to 0.38)||0.001|
|IPSS (scores 1–7)||0.20||−0.06||0.01 (−1.63 to 1.65)||0.99|
|HCT (l/l)||0.03||−0.0004||0.03 (0.02 to 0.04)||<0.0001|
|HB (g/l)||6.73||0.30||6.48 (3.72 to 9.23)||0.0001|
|Other efficacy outcomes|
|Muscle strength non-dominant arm (kg)||1.6||0.6||0.83 (−1.19 to 2.84)||0.41|
|Muscle strength dominant arm (kg)||2.1||1.0||0.82 (−1.49 to 3.13)||0.48|
|Physical activity (PASE score)||43.7||6.2||−35.0 (−6.54 to 76.5)||0.10|
*Between-group differences (and 95% CIs) and P values are adjusted by waist circumference. Mean differences from baseline in the testosterone and placebo groups are unadjusted. SHBG, sex hormone-binding globulin; SCAF, subcutaneous abdominal fat; VAF, visceral abdominal fat; AIx, augmentation Index; CAP, central augmentation pressure; PPA, pulse pressure amplification; REE, resting energy expenditure; RQ, respiratory quotient; HU, Hounsfield units; HOMR-IR, homeostasis model assessment of insulin resistance; ALT, alanine transaminase; AST, aspartate transaminase; PSA, prostate-specific antigen; IPSS, International Prostate Symptom Score; HCT, haematocrit; HB, haemoglobin; PASE, physical activity scale for the elderly.
Testosterone treatment decreased IR (HOMA-IR; Fig. 1B) and liver fat (Fig. 1C) compared with placebo (Table 3). Testosterone significantly increased LDL-cholesterol compared with placebo (Table 3). There were no changes in any other blood lipids, the total to HDL-cholesterol ratio, or liver transaminases (Table 3). The respiratory quotient significantly decreased with testosterone compared with placebo treatment indicating a shift to fat, rather than carbohydrate, utilisation (Table 3). However, resting energy expenditure did not significantly improve more with testosterone than with placebo (Table 3). Testosterone therapy did not significantly alter the net development of metabolic syndrome at week 6 (P=0.93), week 12 (P=0.33) or week 18 (P=0.96) using a consensus definition of metabolic syndrome (45). Analogous findings were obtained using an alternate definition of metabolic syndrome (46).
Weight, BMI and waist circumference as well as total fat, SCAF and VAF significantly decreased over time due to the weight-loss programme (P<0.0001), but there were no significant differences between groups at any single timepoint or overall (Table 3). In contrast, testosterone therapy significantly increased lean muscle mass compared with placebo (Fig. 1D and Table 3). There were no significant between-group differences in any other anthropometric measurement (Table 3).
Prostate and haematological outcomes
Testosterone increased serum PSA and haematocrit compared with placebo, but no participant met predefined withdrawal criteria and there was no change in self-reported prostate symptoms in either group (Table 3). Haemoglobin also increased with testosterone compared with placebo (Table 3). There were no other significant between-group differences in blood haematology or other routine tests of renal and liver function (data not shown).
Other efficacy outcomes and exploratory analysis
Dominant and non-dominant hand strength increased with time (both P<0.05), but there were no differences between groups (Table 3). Physical activity was significantly increased compared with placebo at week 6 (P=0.03); however, there was no overall between-group difference (Table 3).
Exploratory analyses using total testosterone dichotomised by cut points of 8, 11 and 13 nM and free testosterone dichotomised by thresholds of 160, 220 and 280 pmol/l were performed and the interaction between treatment and the dichotomised variable examined. The interaction term of treatment and any dichotomised factor for testosterone were not statistically significant for any measures of body composition. This indicates that the testosterone treatment effects on body composition were not influenced by baseline testosterone concentrations.
This is the first randomised placebo-controlled study to examine the body compositional and cardiometabolic effects of testosterone treatment in obese men with OSA. Indeed, only two previous studies, one of which we conducted, have examined pharmacological methods to alter body composition in men with OSA (50, 51), and neither study was randomised nor contained a no-treatment control. This study therefore adds to the very limited literature examining mechanisms and methods by which cardiometabolic health and body composition can be altered in men with OSA. We demonstrate that 18 weeks of testosterone therapy does not further reduce weight, BMI or metabolic syndrome but does improve several important cardiometabolic risk factors as well as lean muscle mass and respiratory quotient in obese men with OSA. This is a specific population in whom both relative androgen deficiency and increased cardiometabolic risk have been documented. Recently, published data from this study has shown that testosterone mildly worsens sleep-disordered breathing in a time-limited manner, irrespective of initial testosterone concentrations (38).
We demonstrate that testosterone therapy reduces liver fat and increases insulin sensitivity in obese men with OSA. These changes in liver fat or insulin sensitivity should independently decrease cardiovascular risk (52). To our knowledge, this is the first study of testosterone therapy to accurately measure liver fat by CT imaging, although there are now ongoing studies examining the effect of transdermal testosterone treatment on liver fat in hypogonadal men with (clinicaltrials.gov number: NCT01560546) and without diabetes (NCT00700024). Transaminases did not change after testosterone treatment, which is in agreement with a previous 6-month randomised controlled trial in older men (53), although a non-controlled study did find significant reduction in transaminases after 1 year (54). Imaging techniques for the quantification of liver fat are sensitive and validated methods whereas transaminases may not detect changes in liver fat and can be influenced by a variety of biological circumstances (55). Previous randomised placebo-controlled studies show that testosterone increases insulin sensitivity in three populations of men: eugonadal abdominally obese men, by euglycemic hyperinsulinemic clamp (13, 14, 15); biochemically androgen-deficient men with type II diabetes and/or the metabolic syndrome, by HOMA-IR (17, 35, 56, 57, 58) including a meta-analysis of men with the metabolic syndrome (20); and elderly patients with chronic heart failure, by HOMA-IR (18, 59). Another meta-analysis of diabetic men reported significant reductions in HbA1c, fasting glucose and triglycerides, which may have resulted in an improvement in IR, although this was not directly assessed (36). Over the longer term, these changes would be expected to reverse metabolic syndrome and NAFLD, although we did not show a reduction in metabolic syndrome after 18 weeks. Only two other randomised placebo controlled trials of testosterone therapy have examined the effects on metabolic syndrome, albeit in non-OSA populations (57, 58). Both reported reductions in metabolic syndrome but were of longer duration (1–2 years) and both specifically recruited men with metabolic syndrome. However, no studies thus far have targeted reversal of NAFLD despite an independent association between low testosterone levels and hepatic steatosis, a marker for NAFLD (60, 61). The prevalence of NAFLD is increasing, particularly in patients with cardiometabolic risk factors and is associated with increased risk of liver damage and possibly future CVD (52, 62). Therefore, a reduction in NAFLD by lowering liver fat could have an important impact on disease burden. Our findings in conjunction with analogous data in other populations suggest the utility of performing a longer term study.
Other cardiovascular risk factors were also generally improved after testosterone supplementation. Arterial stiffness as measured by the AIx was significantly reduced in the testosterone group compared with the placebo group. Similar findings have been reported in men with chronic heart failure and low testosterone levels (27) and in men with acquired hypogonadism (28). Augmentation pressure, an additional marker of arterial stiffness, was also significantly improved at 18 weeks in the testosterone group compared with placebo. These measures are validated predictors of cardiovascular risk (63). Testosterone therapy also significantly altered substrate utilisation to favour oxidation of fat, rather than carbohydrate, as shown by a significant reduction in the respiratory quotient. Such changes should promote fat loss if persistent over the longer term, although this was not observed within the 18-week treatment period. There was no change in the total to HDL-cholesterol ratio, a standard measure of coronary heart disease risk (64), although a small increase of uncertain clinical relevance in LDL-cholesterol was observed. In other studies, testosterone therapy has been reported to either decrease (65, 66) or not change LDL-cholesterol (53, 56, 67, 68, 69). There were no changes to lipid-lowering therapy during the course of the study, suggesting that the changes in LDL-cholesterol were not as a result of alterations in concomitant medications.
On balance, the improvements in insulin sensitivity, liver fat and arterial stiffness would seem to outweigh the possible changes in LDL-cholesterol, suggesting that testosterone therapy generally improves cardiometabolic risk in obese men with OSA. Whether this improvement persists beyond 18 weeks would need to be verified with longer term studies. Furthermore, the factors that ultimately prove to be decisive in modulating whether or not cardiovascular events occur in response to testosterone therapy have not yet been determined, but advanced age and immobility are two possibilities (37). However, baseline testosterone concentrations do not appear to be relevant in this population. Such an analysis was possible because participants were specifically selected to have both obesity and OSA and therefore to have relative androgen deficiency (29), rather than recruited on the basis of baseline testosterone concentrations.
Testosterone increased lean muscle mass in this population, which is consistent with previous studies in both eugonadal and hypogonadal men (10, 11). We did not observe a further reduction in total fat mass in contrast with previous randomised controlled studies, but many of these were not performed in obese populations and therefore did not include lifestyle modification (10, 11). Indeed, the differential effect on muscle and fat we observed could also be explained by the lifestyle modification programme administered, which was focussed on diet rather than exercise. Furthermore, as chronic intermittent hypoxia directly influences the uptake of triglycerides in animal models (5), testosterone effects on fat mass may differ between those with and without OSA.
Unsurprisingly, we did not detect a significant reduction in total weight compared with placebo as testosterone would increase muscle and reduce fat resulting in attenuated net changes. This finding is consistent with most (14, 15, 23, 70) but not all (35) previous randomised controlled studies. We did not observe a reduction in VAF with testosterone compared with placebo, which may also be explained by the lifestyle modification programme administered to all participants and/or the influence of chronic intermittent hypoxia. Furthermore, this finding is consistent with the hypothesis that changes in insulin sensitivity occur before changes in VAF (15). Nevertheless, previous randomised placebo-controlled studies using computer imaging to estimate VAF have generally shown no change in VAF after 6–24 months of testosterone treatment (21, 22, 23, 24). One notable exception is a study of non-obese older men with age-related hypogonadism, which reported a between-group reduction in VAF that was driven by an increase in the placebo group (70). Testosterone has also been shown to reduce VAF in abdominally obese men with low to normal testosterone levels after 8 and 9 months of treatment, despite no changes in any other body composition measurement including total lean mass (13, 14, 15). It could be that abdominally obese men may be more susceptible to changes in VAF compared with our population who were selected for generalised obesity as defined by a BMI ≥30 kg/m2.
It is currently unclear whether or not low testosterone levels are causally related to CVD or whether it is simply a marker of illness (71, 72). In this study of obese men with OSA, a population who may be at a greater risk of CVD, testosterone therapy improved some markers of cardiometabolic health despite no reduction in weight, fat mass or OSA parameters. However, longer term studies of the effect of testosterone therapy of hard cardiometabolic endpoints in this particular population are required for confirmation.
There are some possible limitations that should be acknowledged. The duration of testosterone treatment may have been too short to result in changes in some of the outcome measures (73). As all the participants received a weight loss programme, we cannot extrapolate these findings to men not receiving a weight loss programme. However, all obese men with OSA (such as ours) or without OSA ought to be placed on a weight loss programme, so our scientific study design appeared to be the most appropriate to answer a pragmatic clinical question. Our findings also pertain to the use of intramuscular testosterone undecanoate. Different modes of administration may have different results (74). Finally, the cut points used to examine the influence of baseline testosterone levels on the effect of treatment on body composition were developed for late-onset hypogonadism, not necessarily for obese middle-aged men with OSA; nevertheless, these cut points seemed to be a reasonable starting point.
The role of testosterone replacement in obese men with OSA in clinical practice needs to consider all potential risks and benefits, including prostate, haematological and respiratory effects (38, 75). Nevertheless, this mechanistic report of solely cardiometabolic effects shows that 18 weeks of testosterone therapy may potentially improve certain important parameters of cardiometabolic health and provides a basis on which future larger and longer-term studies can be implemented.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Supported by the National Health and Medical Research Council of Australia (NHMRC) through a project grant (512499), a Centre for Clinical Research Excellence in Interdisciplinary Sleep Health (571421) and fellowships to C M Hoyos, C L Phillips, R R Grunstein and P Y Liu (512057, 571179, 202916 and 1025248 respectively). Bayer Schering supplied study drug, matching placebo and $20 000 to date.
The authors thank Roo Killick, Kerri Melehan, Farid Saad and the Data and Safety Monitoring Committee, which consisted of Prof. Michael Hensley (chair), Associate Prof. Val Gebski and Prof. Stephen Twigg. They also thank the men who participated in the study, the sleep physicians and the study coordinators.
DragerLFLiJShinMKReinkeCAggarwalNRJunJCBevans-FontiSSztalrydCO'ByrneSMKroupaOOlivecronaGBlanerWSPolotskyVY. Intermittent hypoxia inhibits clearance of triglyceride-rich lipoproteins and inactivates adipose lipoprotein lipase in a mouse model of sleep apnoea. European Heart Journal201233783–790. (doi:10.1093/eurheartj/ehr097).
IiyoriNAlonsoLCLiJSandersMHGarcia-OcanaAO'DohertyRMPolotskyVYO'DonnellCP. Intermittent hypoxia causes insulin resistance in lean mice independent of autonomic activity. American Journal of Respiratory and Critical Care Medicine2007175851–857. (doi:10.1164/rccm.200610-1527OC).
IsidoriAMGiannettaEGrecoEAGianfrilliDBonifacioVIsidoriALenziAFabbriA. Effects of testosterone on body composition, bone metabolism and serum lipid profile in middle-aged men: a meta-analysis. Clinical Endocrinology200563280–293. (doi:10.1111/j.1365-2265.2005.02339.x).
BhasinSStorerTWJavanbakhtMBermanNYarasheskiKEPhillipsJDikeMSinha-HikimIShenRHaysRDBeallG. Testosterone replacement and resistance exercise in HIV-infected men with weight loss and low testosterone levels. Journal of the American Medical Association2000283763–770. (doi:10.1001/jama.283.6.763).
CaminitiGVolterraniMIellamoFMarazziGMassaroRMiceliMMammiCPiepoliMFiniMRosanoGM. Effect of long-acting testosterone treatment on functional exercise capacity, skeletal muscle performance, insulin resistance, and baroreflex sensitivity in elderly patients with chronic heart failure a double-blind, placebo-controlled, randomized study. Journal of the American College of Cardiology200954919–927. (doi:10.1016/j.jacc.2009.04.078).
MunzerTHarmanSMHeesPShapiroEChristmasCBellantoniMFStevensTEO'ConnorKGPabstKMSt ClairCSorkinJDBlackmanMR. Effects of GH and/or sex steroid administration on abdominal subcutaneous and visceral fat in healthy aged women and men. Journal of Clinical Endocrinology and Metabolism2001863604–3610. (doi:10.1210/jc.86.8.3604).
NairKSRizzaRAO'BrienPDhatariyaKShortKRNehraAVittoneJLKleeGGBasuABasuRCobelliCToffoloGDalla ManCTindallDJMeltonLJIIISmithGEKhoslaSJensenMD. DHEA in elderly women and DHEA or testosterone in elderly men. New England Journal of Medicine20063551647–1659. (doi:10.1056/NEJMoa054629).
WebbCMElkingtonAGKraidlyMMKeenanNPennellDJCollinsP. Effects of oral testosterone treatment on myocardial perfusion and vascular function in men with low plasma testosterone and coronary heart disease. American Journal of Cardiology2008101618–624. (doi:10.1016/j.amjcard.2007.09.114).
Barrett-ConnorEDamTTStoneKHarrisonSLRedlineSOrwollE. The association of testosterone levels with overall sleep quality, sleep architecture, and sleep-disordered breathing. Journal of Clinical Endocrinology and Metabolism2008932602–2609. (doi:10.1210/jc.2007-2622).
WuFCTajarABeynonJMPyeSRSilmanAJFinnJDO'NeillTWBartfaiGCasanuevaFFFortiGGiwercmanAHanTSKulaKLeanMEPendletonNPunabMBoonenSVanderschuerenDLabrieFHuhtaniemiIT. Identification of late-onset hypogonadism in middle-aged and elderly men. New England Journal of Medicine2010363123–135. (doi:10.1056/NEJMoa0911101).
KalinchenkoSYTishovaYAMskhalayaGJGoorenLJGiltayEJSaadF. Effects of testosterone supplementation on markers of the metabolic syndrome and inflammation in hypogonadal men with the metabolic syndrome: the double-blinded placebo-controlled Moscow study. Clinical Endocrinology201073602–612. (doi:10.1111/j.1365-2265.2010.03845.x).
BasariaSCovielloADTravisonTGStorerTWFarwellWRJetteAMEderRTennstedtSUlloorJZhangAChoongKLakshmanKMMazerNAMiciekRKrasnoffJElmiAKnappPEBrooksBApplemanEAggarwalSBhasinGHede-BrierleyLBhatiaACollinsLLeBrasseurNFioreLDBhasinS. Adverse events associated with testosterone administration. New England Journal of Medicine2010363109–122. (doi:10.1056/NEJMoa1000485).
Anthropometric Standardization Reference Maunal
PhillipsCLYeeBJTrenellMIMagnussenJSWangDBanerjeeDBerendNGrunsteinRR. Changes in regional adiposity and cardio-metabolic function following a weight loss program with sibutramine in obese men with obstructive sleep apnea. Journal of Clinical Sleep Medicine20095416–421.
AlbertiKGEckelRHGrundySMZimmetPZCleemanJIDonatoKAFruchartJCJamesWPLoriaCMSmithSCJr. Harmonizing the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity. Circulation20091201640–1645. (doi:10.1161/CIRCULATIONAHA.109.192644).
LiuPYWishartSMHandelsmanDJ. A double-blind, placebo-controlled, randomized clinical trial of recombinant human chorionic gonadotropin on muscle strength and physical function and activity in older men with partial age-related androgen deficiency. Journal of Clinical Endocrinology and Metabolism2002873125–3135. (doi:10.1210/jc.87.7.3125).
BarryMJFowlerFJJrO'LearyMPBruskewitzRCHoltgreweHLMebustWKCockettAT. The American Urological Association symptom index for benign prostatic hyperplasia. The Measurement Committee of the American Urological Association. Journal of Urology19921481549–1557.(discussion 1564).
WangCNieschlagESwerdloffRBehreHMHellstromWJGoorenLJKaufmanJMLegrosJJLunenfeldBMoralesAMorleyJESchulmanCThompsonIMWeidnerWWuFC. Investigation, treatment, and monitoring of late-onset hypogonadism in males: ISA, ISSAM, EAU, EAA, and ASA recommendations. Journal of Andrology2009301–9. (doi:10.2164/jandrol.108.006486).
Emmelot-VonkMHVerhaarHJNakhai PourHRAlemanALockTMBoschJLGrobbeeDEvan der SchouwYT. Effect of testosterone supplementation on functional mobility, cognition, and other parameters in older men: a randomized controlled trial. Journal of the American Medical Association200829939–52. (doi:10.1001/jama.2007.51).
HaiderAGoorenLJPadungtodPSaadF. Improvement of the metabolic syndrome and of non-alcoholic liver steatosis upon treatment of hypogonadal elderly men with parenteral testosterone undecanoate. Experimental and Clinical Endocrinology & Diabetes2010118167–171. (doi:10.1055/s-0029-1202774).
KapoorDGoodwinEChannerKSJonesTH. Testosterone replacement therapy improves insulin resistance, glycaemic control, visceral adiposity and hypercholesterolaemia in hypogonadal men with type 2 diabetes. European Journal of Endocrinology2006154899–906. (doi:10.1530/eje.1.02166).
HeufelderAESaadFBunckMCGoorenL. Fifty-two-week treatment with diet and exercise plus transdermal testosterone reverses the metabolic syndrome and improves glycemic control in men with newly diagnosed type 2 diabetes and subnormal plasma testosterone. Journal of Andrology200930726–733. (doi:10.2164/jandrol.108.007005).
AversaABruzzichesRFrancomanoDRosanoGIsidoriAMLenziASperaG. Effects of testosterone undecanoate on cardiovascular risk factors and atherosclerosis in middle-aged men with late-onset hypogonadism and metabolic syndrome: results from a 24-month, randomized, double-blind, placebo-controlled study. Journal of Sexual Medicine201073495–3503. (doi:10.1111/j.1743-6109.2010.01931.x).
LaurentSCockcroftJVan BortelLBoutouyriePGiannattasioCHayozDPannierBVlachopoulosCWilkinsonIStruijker-BoudierH. Expert consensus document on arterial stiffness: methodological issues and clinical applications. European Heart Journal2006272588–2605. (doi:10.1093/eurheartj/ehl254).
SteidleCSchwartzSJacobyKSebreeTSmithTBachandR. AA2500 testosterone gel normalizes androgen levels in aging males with improvements in body composition and sexual function. Journal of Clinical Endocrinology and Metabolism2003882673–2681. (doi:10.1210/jc.2002-021058).
WangCSwerdloffRSIranmaneshADobsASnyderPJCunninghamGMatsumotoAMWeberTBermanN. Transdermal testosterone gel improves sexual function, mood, muscle strength, and body composition parameters in hypogonadal men. Journal of Clinical Endocrinology and Metabolism2000852839–2853. (doi:10.1210/jc.85.8.2839).
BhasinSCunninghamGRHayesFJMatsumotoAMSnyderPJSwerdloffRSMontoriVM. Testosterone therapy in men with androgen deficiency syndromes: an Endocrine Society clinical practice guideline. Journal of Clinical Endocrinology and Metabolism2010952536–2559. (doi:10.1210/jc.2009-2354).