RUDY M. HADDAD, MD; CASSIE C. KENNEDY, MD; SEAN M. CAPLES, DO; MICHAL J. TRACZ, MD; ENRIQUE R. BOLONA, MD; KOSTANDINOS SIDERAS, MD; MARIA V. URAGA, MD; PATRICIA J. ERWIN, MLS; AND VICTOR M. MONTORI, MD, MSC

From the Knowledge and Encounter Research Unit (R.M.H., C.C.K., S.M.C., M.J.T., E.R.B., K.S., M.V.U., P.J.E., V.M.M.), Division of Pulmonary and Critical Care Medicine (S.M.C.), Mayo Clinic Libraries (P.J.E.), and Division of Endocrinology, Diabetes, Metabolism, and Nutrition (V.M.M.), Mayo Clinic College of Medicine, Rochester, Minn.

This work was funded by a Mayo Foundation scholarship to Dr Montori and supported by the Department of Medicine, Mayo Clinic College of Medicine.

Victor M. Montori, MD, MSc, Division of Endocrinology, Metabolism, Diabetes, and Nutrition, Mayo Clinic College of Medicine, 200 First St SW, Rochester, MN 55905; E-mail: kerunit@mayo.edu.

OBJECTIVE: To conduct a systematic review and meta-analysis of randomized trials that assessed the effect of testosterone use on cardiovascular events and risk factors in men with different degrees of androgen deficiency.

METHODS: Librarian-designed search strategies were used to search the MEDLINE (1966 to October 2004), EMBASE (1988 to October 2004), and Cochrane CENTRAL (inception to October 2004) databases. The database search was performed again in March 2005. We also reviewed reference lists from included studies and content expert files. Eligible studies were randomized trials that compared any formulation of commercially available testosterone with placebo and that assessed cardiovascular risk factors (lipid fractions, blood pressure, blood glucose), cardiovascular events (cardiovascular death, nonfatal myocardial infarction, angina or claudication, revascularization, stroke), and cardiovascular surrogate end points (ie, laboratory tests indicative of cardiac or vascular disease). Using a standardized data extraction form, we collected data on participants, testosterone administration, and outcome measures. We assessed study quality with attention to allocation concealment, blinding, and loss to follow-up.

RESULTS: The 30 trials included 1642 men, 808 of whom were treated with testosterone. Overall, the trials had limited reporting of methodological features that prevent biased results (only 6 trials reported allocation concealment), enrolled few patients, and were of brief duration (only 4 trials followed up patients for >1 year). The median loss to follow-up across all 30 trials was 9%. Testosterone use in men with low testosterone levels led to inconsequential changes in blood pressure and glycemia and in all lipid fractions (total cholesterol: odds ratio [OR], -0.22; 95% confidence interval [CI], -0.71 to 0.27; high-density lipoprotein choles-terol: OR, -0.04; 95% CI, -0.39 to 0.30; low-density lipoprotein cholesterol: OR, 0.06; 95% CI, -0.30 to 0.42; and triglycerides: OR, -0.27; 95% CI, -0.61 to 0.08); results were similar in patients with low-normal to normal testosterone levels. The OR between testosterone use and any cardiovascular event pooled across trials that reported these events (n=6) was 1.82 (95% CI, 0.78 to 4.23). Several trials failed to report data on measured outcomes. For reasons we could not explain statistically, the results were inconsistent across trials.

CONCLUSION: Currently available evidence weakly supports the inference that testosterone use in men is not associated with important cardiovascular effects. Patients and clinicians need large randomized trials of men at risk for cardiovascular disease to better inform the safety of long-term testosterone use.

CI = confidence interval; HDL-C = high-density lipoprotein cholesterol; LDL-C = low-density lipoprotein cholesterol.

There has been increasing interest in the use of testoste-rone in men other than those with classic symptomatic hypogonadism, a trend fueled by the aging of the population and the development of novel drug delivery systems. Clinicians may be interested in using testosterone to treat men with some degree of androgen deficiency who present with suggestive symptoms, osteoporosis, sexual dysfunc-tion, and poor quality of life. However, more widespread use raises concern about the undesirable cardiovascular consequences of testosterone administration.

A few systematic reviews have been published that assessed the effect of testosterone on cardiovascular risk, 2 of which present rigorous meta-analyses: 1 on the effect of intramuscular testosterone on lipids¹ and 1 on the effect of testosterone on intermittent claudica- tion.² Thus, to this day, the extent and direction (ie, beneficial or harmful) of the cardiovascular consequences of testosterone administration remain unclear.

The Endocrine Society established a task force to generate evidence-based clinical practice guidelines about the use of testosterone in men with different degrees of androgen deficiency. To support this effort, we systematically reviewed the best available evidence about the effects of testosterone use on cardiovascular risk in men with different degrees of androgen deficiency.

Methods

We developed a systematic review protocol (available by request) in collaboration with the members of the Endo-crine Society Task Force on Testosterone in Men with Androgen Deficiency. This report adheres to the Quality of Reporting of Meta-analyses standards for reporting sys-tematic reviews of randomized trials.³

ELIGIBILITY CRITERIA

Eligible studies were fully published randomized trials of men with different degrees of androgen deficiency (includ-ing studies of men with normal or low-normal testosterone levels) who were allocated to receive either testosterone therapy alone (any available preparation) or placebo. Eli-gible studies measured major cardiovascular events, surro-gate cardiovascular end points, and cardiovascular risk factors. Major adverse cardiovascular events included cardiac death, myocardial infarction, and other vascular events such as stroke. Surrogate events included timing of exercise- induced electrocardiographic changes, subjective improvement in symptoms of claudication, ankle brachial index, walking distance, tests of muscle blood flow and plethysmography, hospitalization for chest pain, and need for revascularization. Cardiac risk factors included lipid fractions (total cholesterol, low-density lipoprotein cholesterol [LDL- C], high-density lipoprotein cholesterol [HDL-C], and triglycerides), blood pressure control, and glycemic control.

STUDY IDENTIFICATION

To identify eligible studies, we conducted a systematic search of the literature using the electronic databases MEDLINE (1966 to October 2004), EMBASE (1988 to October 2004), and Cochrane CENTRAL (inception to October 2004); reference sections of identified narrative and systematic reviews identified through a MEDLINE search in October 2004 and of each of the eligible primary studies; and contact with task force expert members. An expert reference librarian (P.J.E.) designed and conducted the electronic search strategy with input from an endocri-nologist with expertise in conducting systematic reviews (V.M.M.). The search was updated in March 2005.

Teams of 2 reviewers independently and with substan-tial reliability (chance-adjusted interobserver agreement K=0.7) screened all abstracts and titles, as well as all resulting full-text publications for eligibility. In cases in which disagreement between 2 reviewers existed, another member of the research team not involved in the initial assessment and with both content and methodological expertise (V.M.M.) adjudicated the study as eligible or not, after reviewing the stated reasons for the initial assessment and the full text of the report.

DATA COLLECTION

Working in duplicate and using a standardized data extraction form, we extracted the following descriptive data from every study: year and journal of publication, patient population (degree of androgen deficiency, prior exposure to testosterone, age, testosterone level), treatment (dose and route of administration of testosterone) and control interventions, and the number of patients in exposed and unexposed groups.

We classified reports by the mean testosterone level at baseline: low testosterone level was defined as a total testosterone level of 300 ng/dL or less (10.4 nmol/L).⁴ When this was not reported, we used values below the lower limit of normal for bioavailable or free testosterone levels. When laboratory values were not available, we classified studies by the the type of patients enrolled (ie, patients with previous bilateral orchiectomy). Chance- adjusted interobserver agreement for this classification was almost perfect (K=0.91).

QUALITY ASSESSMENT

To ascertain the validity of eligible randomized trials, pairs of reviewers working independently and with adequate reliability (corresponding chance-adjusted interobserver agreement in parentheses where pertinent) determined the adequacy of randomization (K=1.0) and concealment of allocation (K=0.82); blinding of patients (K=0.7), health care professionals (K=0.7), data collectors (K=0.77), and outcome assessors (K=0.84); and the extent of loss to follow-up (ie, proportion of patients in whom the investigators were not able to ascertain outcomes).

STATISTICAL ANALYSES

Meta-analyses. For the available lipid fraction and blood pressure outcomes in each study, we determined the mean and SD for the testosterone and placebo arms (for either end-of-study or change-from-baseline data at the longest point of most complete follow-up). We determined the effect size of the difference between the treatment and placebo groups by dividing the mean difference by the pooled SD between arms with adjustment for small samples (Hedges g) generating standardized mean differences as implemented in RevMan 4.2 (Cochrane Collaboration). We then conducted meta-analysis using the random-effects method and quantified the extent to which the inconsistency observed corresponded to between-study differences (and not to chance) using the I² statistic. To pool across dichotomous outcomes (ie, cardiovascular events), we calculated a pooled estimate using Mantel-Haenszel methods with a Robind-Brenslow-Greenland variance⁵ using the Sweeting continuity correction.^6,7

Subgroup Analyses. Our a priori hypotheses to explain potential heterogeneity across studies included study quality (particularly loss to follow-up), patient population (level of testosterone at baseline), and interventions (testosterone preparations: transdermal vs intramuscular vs oral; testosterone dose: physiologic vs supraphysiologic).

We explored these subgroups one at a time and tested the difference across subgroups using a test for subgrouptreatment interaction.⁸

Results

STUDY CHARACTERISTICS

Figure 1 shows the results of our systematic search. We found 30 eligible trials that enrolled 1642 men, 808 of whom were treated with testosterone.

Figure 1. Results of the systematic search. HDL-C = high-density lipoprotein cholesterol; LDL-C = low-density lipoprotein cholesterol.

METHODOLOGICAL QUALITY

Table 1 provides the methodological characteristics of the included trials. Overall, the included trials had limited re-porting of methodological features that protect trials from the introduction of bias. All but 6 trials (20%)^{9,12,17,19,24,26} inadequately reported allocation concealment; 2 had inadequate blinding.^25,37 The median loss to follow-up across all 30 trials was 9%; in 7 trials, loss to follow-up exceeded 20% ^{21,24-26,28,31,36}

Table 1. Methodological Quality of Included Trials.

CLINICAL CHARACTERISTICS

Table 2 provides the clinical characteristics of included trials and, when trials had multiple arms, of each of the trial arms included in the meta-analyses. Nine trials enrolled patients with low testosterone levels; the remaining trials enrolled men receiving long-term glucocorticoid therapy,^36,37 men with chronic obstructive pulmonary disease,^35,38 men with coronary artery disease,^14,25,30 and men with lower-extremity peripheral vascular disease.^13,17 The typical trial used usual replacement doses of testosterone for brief periods (ie, only 4 trials ^12,21,22,26 followed up patients for >1 year).

Table 2. Characteristics of the Included Randomized Placebo-Controlled Trials*.

EFFECT OF T ESTOSTERONE ON CARDIOVASCULAR RISK RACTORS

Lipid Fractions in Men With Low Testosterone Levels. Figures 2 and 3 show nonsignificant effects of testosterone preparations on all lipid fractions in men with low testosterone levels. Reporting bias likely affects these estimates because there were 3, 4, 4, and 2 trials that measured total cholesterol, LDL-C, HDL-C, and triglyceride levels, respectively, but reported results only as “not significant”^27,31 or not at all.^33,34 Given the precision of the pooled estimates, these data exclude unfavorable eleva-tions in total cholesterol levels of more than 9 mg/dL (0.23 nmol/L), in LDL-C levels of more than 14 mg/dL (0.36 nmol/L), and in triglyceride levels of more than 7 mg/dL (0.08 nmol/L) and exclude unfavorable reductions in HDL-C levels of greater than 5 mg/dL (0.12 nmol/L).

Figure 2. Standardized mean differences (SMDs) in the effects of testosterone preparations on all lipid fractions in men with low testosterone levels, low-normal or normal testosterone levels, and chronic disease. CI = confidence interval; HDL-C = high-density lipoprotein cholesterol; LDL-C = low-density lipoprotein cholesterol.

Figure 3. Standardized mean differences (SMDs) in the effects of testosterone preparations on all lipid fractions in men with low testosterone levels (A), low-normal or normal testosterone levels (B), and chronic disease (C) by.

Important between-study differences in the meta-analysis regarding total cholesterol were found (I²=50%), but the results for other fractions were consistent across trials (Figure 3). Exploration of our a priori hypotheses to explain heterogeneity did not yield a subgroup of trials with significantly different results.

Lipid Fractions in Men With Low-Normal or Nor-mal Testosterone Levels. Figures 2 and 3 show the effect of testosterone preparations on lipid fractions in men with low-normal or normal testosterone levels. Testosterone re-duced total cholesterol levels by 16 mg/dL (0.41 nmol/L) (95% confidence interval [CI], 6-26 mg/dL [0.15-0.67 nmol/L]); all other lipid fractions were not significantly affected. Reporting bias likely affects these estimates: 3 trials101425 collected data on total cholesterol but did not report them. The same was true for the LDL-C and HDL-C meta-analyses (n=3 trials ^11,14,25) and for the triglycerides meta-analysis (n=2 trials ^11,14). Given the precision of the pooled estimates, these data exclude unfavorable eleva-tions in LDL-C levels of more than 3 mg/dL (0.08 nmol/L) and in triglyceride levels of more than 41 mg/dL (1.06nmol/L) and exclude unfavorable reductions in HDL-C levels of less than 7 mg/dL (0.18 nmol/L).

Important between-study differences in the meta-analyses of all lipid fractions were found, except for HDL-C (Figure 3). Exploration of our a priori hypotheses to ex-plain heterogeneity did not yield a subgroup of trials with significantly different results.

Lipid Fractions in Men With Chronic Disease. Fig-ures 2 and 3 show the effect of testosterone preparations on lipid fractions in men with chronic disease, such as those who use long-term glucocorticoids or who have chronic obstructive pulmonary disease. Only 2 trials reported results,^35,37 whereas 2 other trials ^36,38 reported that testosterone did not significantly affect lipid fractions. One trial³⁵ collected data on LDL-C and triglyceride levels but did not report these. Testosterone reduced HDL-C levels by 11 mg/dL (0.28 nmol/L) (95% CI, 3-19 mg/dL [0.08-0.49 nmol/L]); all other lipid fractions were not significantly affected. The effect of testosterone on total cholesterol and HDL-C was consistent across the 2 trials that reported these outcomes (Figure 3). Given the precision of the pooled estimates, these data exclude unfavorable elevations in total cholesterol levels of more than 13 mg/dL (0.34 nmol/L), in LDL-C levels of more than 30 mg/dL (0.78 nmol/L), and in triglyceride levels of more than 48 mg/dL (1.24 nmol/L). The limited number of studies precluded us from conducting exploratory subgroup analyses.

EFFECT OF TESTOSTERONE SUPPLEMENTATION ON BLOOD PRESSURE

Testosterone preparations had nonsignificant effects on systolic (0.8 mm Hg; 95% CI, -4 to 5 mm Hg) and diastolic (2 mm Hg; 95% CI, -2 to 6 mm Hg) blood pressure that were consistent across trials. Reporting bias likely affects this estimate because 2 trials11,16 measured blood pressure data but did not report these. Overall, the precision of the estimates excludes unfavorable elevations in systolic blood pressure of more than 5 mm Hg or in diastolic blood pressure of more than 6 mm Hg.

EFFECT OF TESTOSTERONE SUPPLEMENTATION ON BLOOD GLUCOSE

One trial that enrolled 12 men with low testosterone levels32 reported a difference of 1 mg/dL between testosterone and placebo (95% CI, -10 to 12 mg/dL). Three trials that enrolled a total of 108 men with normal testosterone lev- els^14,16,20 measured the effect of testosterone on glycemia; 2 of these trials ^14,16 did not report glucose levels despite collecting these data, and 1 trial²⁰ reported this outcome only as “not significant.” Thus, we could not conduct a meta-analysis.

EFFECT OF TESTOSTERONE ON CARDIOVASCULAR EVENTS

Unfortunately, most studies that reported cardiac events had neither strict outcome definitions nor independent and blinded judicial assessors of these outcomes. Thus, the outcomes reported may not represent a complete or unbiased selection of the outcomes that took place. Further-more, the length of follow-up in most of the included trials precluded the accumulation of enough events.

Cardiovascular Events. We sought to summarize the effect of testosterone on cardiovascular death, fatal and nonfatal myocardial infarction, and other cardiovascular events (eg, angina, arrhythmia, revascularization proce-dures, stroke). Six randomized controlled trials^{9,12,14,22,26,38} reported on cardiovascular events with consistent results (Figure 4). There were 14 events (including 5 myocardial infarctions and 1 cardiovascular death) in 161 men who received testosterone and 7 events (including 2 myocardial infarctions and 1 death) in 147 men in the control groups (any cardiovascular event: odds ratio, 1.82; 95% CI, 0.78-4.23; fatal and nonfatal myocardial infarction: odds ratio, 2.24; 95% CI, 0.50-10.02).

Figure 4. Six randomized controlled trials reported on cardiovascular events with consistent results. CI = confidence interval; OR = odds ratio; ∞ = infinity.

One trial³⁶ described 4 patients with myocardial infarctions and 2 patients with a cardiomyopathy without describing which events occurred in the testosterone and placebo groups. Another trial14 only reported that angina frequency did not change in either group. Finally, one trial³⁰ reported a nonsignificant improvement in the Seattle angina questionnaire score with testosterone.

Vascular Events. Only one trial³⁶ reported vascular events (2 patients receiving either testosterone or placebo had ruptured abdominal aortic aneurysm), whereas 2 other trials^13,17 reported no significant improvement on symptoms of claudication with testosterone vs placebo.

Surrogate End Points. Several trials reported on physi-ological or laboratory end points of unclear patient importance or validity as surrogates for the effect of testosterone on important cardiovascular end points.

EXERCISE ELECTROCARDIOGRAPHIC CHANGES

Two trials in men with coronary artery disease (1 in men with normal or low-normal testosterone levels¹⁴ and 1 in men with low testosterone levels³⁰) reported significantly longer time (25 and 74 seconds longer, respectively) to 1-mm ST-segment depression with exercise in the testosterone group vs placebo. Another trial in eugonadal men with coronary artery disease¹⁹ reported significant reductions in the extent of ST depression in anterolateral electrocardio-graphic leads after exercise in men receiving testosterone vs placebo.

VASCULAR FUNCTION

One trial13 found nonsignificant effects of testosterone vs placebo on the metronome walking test, volume of the foot pulse, and change in foot temperature. Two trials^28,39 reported nonsignificant effects of testosterone vs placebo on endothelial function. Finally, one trial¹⁷ reported a nonsignificant effect of testosterone on muscle blood flow.

Discussion

PRINCIPAL FINDINGS

The best available evidence suggests small and clinically negligible effects of testosterone use on lipid fractions, blood pressure, and glycemic control in men with different degrees of androgen deficiency. On the basis of the width of the 95% CI, the pooled data are consistent with both a 1fold decrease and a 4-fold increase in the odds of cardiac events in patients using testosterone.

LIMITATIONS AND STRENGTHS

A key limitation of this review refers to the extent to which authors did not explicitly report on all outcomes measured. To the extent that there is a consistent relationship between outcomes and whether authors report them, bias could affect our pooled estimates. The direction and magnitude of this potential bias are unclear for safety outcomes, such as those summarized herein. The degree of unexplained heterogeneity represents another key limitation of our review. Taken together, the paucity of data, unexplained heterogeneity, and reporting bias weaken the inferences drawn from the best available evidence. Despite our efforts and the active participation of clinical experts, we may have missed eligible trials. However, our systematic review has the strengths appropriate for this study design: a protocol-driven process; clear, explicit, and reproducible eligibility criteria; reproducible judgments about study quality; and systematic data collection and targeted analyses.

OUR REVIEW IN RELATIONSHIP TO OTHER SYSTEMATIC REVIEWS

Our review differs from the meta-analysis by Whitsel et al¹ in scope (searched up to 1999, only interested in the effect of intramuscular testosterone on lipids of hypogonadal men), methods (included observational studies, did not assess the methodological quality of eligible studies), and consequentially results (found a dose-dependent decrease in HDL-C and total cholesterol levels). In contrast, our review searched to 2005, had a broader scope (all forms of administration of testosterone in men with varying degrees of androgen deficiency), had methods that paid attention to bias (eligible studies were placebo-controlled randomized trials only, and we assessed their quality), and reported inconsequential effects of testosterone on lipid fractions. In particular, among hypogonadal men who received intramuscular testosterone, we found nonsignificant effects on total cholesterol and HDL-C levels.

Our results are consistent with the conclusions from a systematic review of placebo-controlled randomized trials in older men by Gruenewald and Matsumoto⁴⁰ that summarized data up to 2001 without conducting meta-analyses. Another narrative review from Wu and von Eckardstein⁴¹ summarized data up to 2002 without conducting metaanalyses. This review included randomized and observational studies; reviewed the putative effects of testosterone on surrogate end points and cardiovascular risk factors, including lipid fractions; and suggested that supraphysiological testosterone use could decrease HDL-C levels. In contrast to these 2 reviews,^40,41 our meta-analyses add 4 and 2 more years of evidence, respectively; have a broader scope (include adult men with varying degrees of androgen deficiency); and report pooled estimates for all outcomes, including cardiovascular events. As in the first⁴⁰ and in contrast with the second meta-analysis,⁴¹ we limited our summary to placebo-controlled trials. Our review methods and results are consistent with those of a recently updated Cochrane review² of the effect of testosterone on intermittent claudication.

Our review results are consistent with the meta-analysis by Isidori et al⁴² despite some differences in study inclusion and search time frame (our study being limited to placebo- controlled trials and searching for 1 additional year). Also, our review includes a meta-analysis of cardiovascular events. A review by Calof et al⁴³ limited its investigation to older men with low testosterone levels at baseline who received replacement doses of testosterone. This focus led to fewer included trials and no meta-analysis of lipid fractions but similar results about cardiovascular events.

Finally, a systematic review performed by Krause⁴⁴ summarized studies of testosterone use in older men but did not report quantitative estimates of treatment effect on each of the lipid fractions or on specific cardiovascular outcomes.

REVIEW IMPLICATIONS

As patients and clinicians consider the use of testosterone in the management of symptoms consistent with hypogo-nadism, they should be concerned about potential adverse consequences of its long-term use. Key among these are the potential effects on cardiovascular risk. Unfortunately, the best available evidence on this matter is inconsistent, imprecise, and poorly reported. As a result, clinicians and policymakers cannot be sure what consequences testoste-rone may have on cardiovascular risk.

CONCLUSION

Currently available evidence weakly supports the inference that testosterone use in men is not associated with important cardiovascular effects. Large randomized trials that enroll men with and without cardiovascular disease and measure cardiovascular end points are needed to better inform the decision to use long-term testosterone for other indications.

We thank Gunjan Y. Gandhi, MD, Theophilus E. Owan, MD, and Laura I. Pelaez, MD, for their assistance with study selection, data collection, and author contact. We are grateful for the ongo-ing input and advice from the Endocrine Society Task Force on Testosterone in Men with Androgen Deficiency. We extend our gratitude to the research team at the Knowledge and Encounter Research Unit at Mayo Clinic College of Medicine.

References

Whitsel EA, Boyko EJ, Matsumoto AM, Anawalt BD, Siscovick DS. Intramuscular testosterone esters and plasma lipids in hypogonadal men: a meta-analysis. Am JMed. 2001;111:261-269. 

Price JF, Leng GC. Steroid sex hormones for lower limb athero-sclerosis.[update of Cochrane Database Syst Rev. 2000;2:CD000188]. Coch-rane Database Syst Rev. 2002;1:CD000188. 

Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D, Stroup DF, QUOROM Group. Improving the quality of reports of meta-analyses of randomised controlled trials: the QUOROM statement. Lancet. 1999;354:1896- 1900. 

Bhasin S, Cunningham GR, Hayes FJ, et al. Testosterone therapy in adult men with androgen deficiency syndromes: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab. 2006 Jun;91:1995-2010. Epub 2006 May 23. 

Robins J, Breslow N, Greenland S. Estimators of the Mantel-Haenszel variance consistent in both sparse data and large-strata limiting models. Biometrics. 1986;42:311-323. 

Sweeting MJ, Sutton AJ, Lambert PC. What to add to nothing? use and avoidance of continuity corrections in meta-analysis of sparse data. Stat Med. 2004;23:1351-1375. 

Bongartz T, Sutton AJ, Sweeting MJ, Buchan I, Matteson EL, Montori V. Anti-TNF antibody therapy in rheumatoid arthritis and the risk of serious infections and malignancies: systematic review and meta-analysis of rare harmful effects in randomized controlled trials [published correction appears in JAMA. 2006;295:2482]. JAMA. 2006;295:2275-2285. 

Altman DG, Bland JM. Interaction revisited: the difference between two estimates. BMJ. 2003;326:219. 

Amory JK, Chansky HA, Chansky KL, et al. Preoperative supraphysio- logical testosterone in older men undergoing knee replacement surgery. J Am  Geriatr Soc. 2002;50:1698-1701. 

Aversa A, Isidori AM, Spera G, Lenzi A, Fabbri A. Androgens improve cavernous vasodilation and response to sildenafil in patients with erectile dysfunction. Clin Endocrinol (Oxf). 2003;58:632-638. 

Clague JE, Wu FC, Horan MA. Difficulties in measuring the effect of testosterone replacement therapy on muscle function in older men. Int J Androl. 1999;22:261-265. 

Copenhagen Study Group for Liver Diseases. Testosterone treatment of men with alcoholic cirrhosis: a double-blind study. Hepatology. 1986;6:807- 813. 

Dohn K, Hvidt V, Nielsen J, Palm L. Testosterone therapy in obliterating arterial lesions in the lower limbs. Angiology. 1968;19:342-350. 

English KM, Steeds RP, Jones TH, Diver MJ, Channer KS. Low-dose transdermal testosterone therapy improves angina threshold in men with chronic stable angina: a randomized, double-blind, placebo-controlled study. Circulation. 2000;102:1906-1911. 

Ferrando AA, Sheffield-Moore M, Yeckel CW, et al. Testosterone ad-ministration to older men improves muscle function: molecular and physi-ological mechanisms. Am J Physiol Endocrinol Metab. 2002;282:E601-E607. 

Harman SM, Blackman MR. The effects of growth hormone and sex steroid on lean body mass, fat mass, muscle strength, cardiovascular endurance and adverse events in healthy elderly women and men. Horm Res. 2003; 60(suppl 1):121-124. 

Hentzer E, Madsen PC. Testosterone in the treatment of arterial insufficiency of the lower limbs. Scand J Clin Lab Invest Suppl. 1967;99:198-206. 

Howell SJ, Radford JA, Adams JE, Smets EM, Warburton R, Shalet SM. Randomized placebo-controlled trial of testosterone replacement in men with mild Leydig cell insufficiency following cytotoxic chemotherapy. Clin Endo-crinol (Oxf). 2001;55:315-324. 

Jaffe MD. Effect of testosterone cypionate on postexercise ST segment depression. Br Heart J. 1977;39:1217-1222. 

Liu PY, Yee B, Wishart SM, et al. The short-term effects of high-dose testosterone on sleep, breathing, and function in older men. J Clin Endocrinol Metab. 2003;88:3605-3613. 

Page ST, Amory JK, Bowman FD, et al. Exogenous testosterone (T) alone or with finasteride increases physical performance, grip strength, and lean body mass in older men with low serum T. J Clin Endocrinol Metab. 2005 Mar;90:1502-1510. Epub 2004 Nov 30. 

Snyder PJ, Peachey H, Berlin JA, et al. Effect of transdermal testosterone treatment on serum lipid and apolipoprotein levels in men more than 65 years of age. Am J Med. 2001;111:255-260. 

Tenover JS. Effects of testosterone supplementation in the aging male. J Clin Endocrinol Metab. 1992;75:1092-1098. 

Wittert GA, Chapman IM, Haren MT, Mackintosh S, Coates P, Morley JE. Oral testosterone supplementation increases muscle and decreases fat mass in healthy elderly males with low-normal gonadal status. J Gerontol A Biol Sci Med Sci. 2003;58:618-625. 

Wu S, Weng X. Therapeutic effect of andriol on serum lipids and apolipoproteins in elderly male coronary heart disease patients. Chinese Med Sci J. 1992;7:137-141. 

Amory JK, Watts NB, Easley KA, et al. Exogenous testosterone or testosterone with finasteride increases bone mineral density in older men with low serum testosterone. J Clin Endocrinol Metab. 2004;89:503-510. 

Dobs AS, Hoover DR, Chen MC, Allen R. Pharmacokinetic characteristics, efficacy, and safety of buccal testosterone in hypogonadal males: a pilot study. J Clin Endocrinol Metab. 1998;83:33-39. 

Kenny AM, Prestwood KM, Gruman CA, Fabregas G, Biskup B, Mansoor G. Effects of transdermal testosterone on lipids and vascular reactivity in older men with low bioavailable testosterone levels. J Gerontol A Biol Sci Med Sci. 2002;57:M460-M465. 

Malkin CJ, Pugh PJ, Jones RD, Kapoor D, Channer KS, Jones TH. The effect of testosterone replacement on endogenous inflammatory cytokines and lipid profiles in hypogonadal men. J Clin Endocrinol Metab. 2004;89:3313- 3318. 

Malkin CJ, Pugh PJ, Morris PD, et al. Testosterone replacement in hypogonadal men with angina improves ischaemic threshold and quality of life. Heart. 2004;90:871-876. 

Sih R, Morley JE, Kaiser FE, Perry HM III, Patrick P, Ross C. Testosterone replacement in older hypogonadal men: a 12-month randomized controlled trial. J Clin Endocrinol Metab. 1997;82:1661-1667. 

Simon D, Charles MA, Lahlou N, et al. Androgen therapy improves insulin sensitivity and decreases leptin level in healthy adult men with low plasma total testosterone: a 3-month randomized placebo-controlled trial [letter]. Diabetes Care. 2001;24:2149-2151. 

Steidle C, Schwartz S, Jacoby K, Sebree T, Smith T, Bachand R, North American AA2500 T Gel Study Group. AA2500 testosterone gel normalizes androgen levels in aging males with improvements in body composition and sexual function. J Clin Endocrinol Metab. 2003;88:2673-2681. 

Tan RS, Pu SJ. A pilot study on the effects of testosterone in hypogonadal aging male patients with Alzheimer’s disease. Aging Male. 2003; 6:13-17. 

Casaburi R, Bhasin S, Cosentino L, et al. Effects of testosterone and resistance training in men with chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2004 Oct 15;170:870-878. Epub 2004 Jul 21. 

Crawford BA, Liu PY, Kean MT, Bleasel JF, Handelsman DJ. Randomized placebo-controlled trial of androgen effects on muscle and bone in men requiring long-term systemic glucocorticoid treatment. J Clin Endocrinol Metab. 2003;88:3167-3176. 

Reid IR, Wattie DJ, Evans MC, Stapleton JP. Testosterone therapy in glucocorticoid-treated men. Arch Intern Med. 1996;156:1173-1177. 

Svartberg J, Aasebo U, Hjalmarsen A, Sundsfjord J, Jorde R. Testosterone treatment improves body composition and sexual function in men with COPD, in a 6-month randomized controlled trial. Respir Med. 2004;98:906-913. 

Ly LP, Jimenez M, Zhuang TN, Celermajer DS, Conway AJ, Handelsman DJ. A double-blind, placebo-controlled, randomized clinical trial of transdermal dihydrotestosterone gel on muscular strength, mobility, and quality of life in older men with partial androgen deficiency. J Clin Endocrinol Metab. 2001;86:4078-4088. 

Gruenewald DA, Matsumoto AM. Testosterone supplementation therapy for older men: potential benefits and risks. J Am Geriatr Soc. 2003;51:101-115. 

Wu FC, von Eckardstein A. Androgens and coronary artery disease. Endocr Rev. 2003;24:183-217. 

Isidori AM, Giannetta E, Greco EA, et al. Effects of testosterone on body composition, bone metabolism and serum lipid profile in middle-aged men: a meta-analysis. Clin Endocrinol (Oxf). 2005;63:280-293. 

Calof OM, Singh AB, Lee ML, et al. Adverse events associated with testosterone replacement in middle-aged and older men: a meta-analysis of randomized, placebo-controlled trials. J Gerontol A Biol Sci Med Sci. 2005;60: 1451-1457. 

Krause W. Testosteronsubstitution beim alternden Mann: Welche Fragen sind beantwortet? Urologe A. 2004;43:1097-1100.