Clinical Care/Education/Nutrition/Psychosocial Research

O R I G I N A L A R T I C L E

Testosterone Replacement in
Hypogonadal Men With Type 2
Diabetes and/or Metabolic Syndrome
(the TIMES2 Study)

L
T. HUGH JONES, MD, FRCP1,2 ANTONIO MARTIN MORALES, MD8 ow serum testosterone is common in
STEFAN ARVER, MD3 MAURIZIO VOLTERRANI, MD, FCCP9 men with type 2 diabetes and/or
HERMANN M. BEHRE, MD4 ANN YELLOWLEES, CSTAT10 metabolic syndrome (MetS) and nu-
JACQUES BUVAT, MD5 JULIAN D. HOWELL, MB, BS, FRCS, MFPM11 merous studies have reported an associ-
ERIC MEULEMAN, MD6 KEVIN S. CHANNER, MD, FRCP12
IGNACIO MONCADA, MD7 TIMES2 INVESTIGATORS* ation between testosterone deficiency and
visceral obesity, insulin resistance (IR)
and dyslipidemia (1–4). In men with
OBJECTIVE—This study evaluated the effects of testosterone replacement therapy (TRT) on
type 2 diabetes, low testosterone is asso-
insulin resistance, cardiovascular risk factors, and symptoms in hypogonadal men with type 2 ciated with a high prevalence of symp-
diabetes and/or metabolic syndrome (MetS). tomatic hypogonadism (3), frequently
due to hypogonadotrophic hypogonad-
RESEARCH DESIGN AND METHODS—The efficacy, safety, and tolerability of a novel ism (5).
transdermal 2% testosterone gel was evaluated over 12 months in 220 hypogonadal men with Small studies have demonstrated that
type 2 diabetes and/or MetS in a multicenter, prospective, randomized, double-blind, placebo- testosterone replacement therapy (TRT)
controlled study. The primary outcome was mean change from baseline in homeostasis model in hypogonadal men with and without
assessment of insulin resistance (HOMA-IR). Secondary outcomes were measures of body com-
position, glycemic control, lipids, and sexual function. Efficacy results focused primarily on
type 2 diabetes is associated with reduc-
months 026 (phase 1; no changes in medication allowed). Medication changes were allowed tions in IR, waist circumference, choles-
in phase 2 (months 6212). terol, glycated hemoglobin (HbA1c), and
fasting plasma glucose (FPG) concentra-
RESULTS—TRT reduced HOMA-IR in the overall population by 15.2% at 6 months (P = tions (6–9). Conversely, withdrawal of
0.018) and 16.4% at 12 months (P = 0.006). In type 2 diabetic patients, glycemic control was TRT in hypogonadal men leads to de-
significantly better in the TRT group than the placebo group at month 9 (HbA1c: treatment creased insulin sensitivity (10). Further-
difference, 20.446%; P = 0.035). Improvements in total and LDL cholesterol, lipoprotein a more, androgen suppression therapy for
(Lpa), body composition, libido, and sexual function occurred in selected patient groups. There
were no significant differences between groups in the frequencies of adverse events (AEs) or
prostate cancer can result in alterations of
serious AEs. The majority of AEs (.95%) were mild or moderate. individual cardiovascular risk factors and
increases the occurrence of incident dia-
CONCLUSIONS—Over a 6-month period, transdermal TRT was associated with beneficial betes, myocardial infarction and sudden
effects on insulin resistance, total and LDL-cholesterol, Lpa, and sexual health in hypogonadal cardiac death (11,12). Epidemiologic
men with type 2 diabetes and/or MetS. studies have reported that low testoster-
one in men is associated with increased
Diabetes Care 34:828–837, 2011 all-cause and cardiovascular mortality
(13).
The TIMES2 (Testosterone replace-
c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c ment In hypogonadal men with either
From 1The Robert Hague Centre for Diabetes and Endocrinology, Barnsley Hospital, Barnsley, U.K.; the MEtabolic Syndrome or type 2 diabetes)
2
Department of Human Metabolism, School of Medicine and Biomedical Science, University of Sheffield, study investigated the effects of transder-
Sheffield, U.K.; the 3Center for Andrology and Sexual Medicine, Karolinska University Hospital and the
Karolinska Institute, Stockholm, Sweden; the 4Center for Reproductive Medicine and Andrology, University
mal TRT on IR, selected cardiovascular
Hospital Halle (Saale), Martin-Luther University Halle-Wittenberg, Halle, Germany; the 5Centre d’Etudes et de risk factors, and symptoms in hypogona-
Traitement de la Pathologie de l’Appareil Reproducteur et de la Psychosomatique (CETPARP), Lille, France; dal men with MetS and/or type 2 diabetes.
the 6Department of Urology, VU University Medical Center, Amsterdam, the Netherlands; the 7Department of The safety and tolerability of a novel,
Urology, Hospital Sanitas-La Zarzuela, Madrid, Spain; the 8Department of Urology, Carlos Haya University metered-dose, transdermal 2% testoster-
Hospital, Malaga, Spain; the 9Department of Cardiology, Istituto di Ricovero e Cura a Carattere Scientifico
(IRCCS) San Raffaele Pisana, Rome, Italy; the 10Quantics Consulting Limited, Tweed Horizons, Newtown one gel were also examined.
St Boswells, Scottish Borders, U.K.; the 11Clinical Research Department, ProStrakan, Galashiels, U.K.; and the
12
Department of Cardiology, Royal Hallamshire Hospital, Sheffield, U.K.
Corresponding author: T. Hugh Jones, hugh.jones@nhs.net.
RESEARCH DESIGN AND
Received 29 June 2010 and accepted 2 January 2011. METHODS—This 12-month, pro-
DOI: 10.2337/dc10-1233. Clinical trial reg. no. ISRCTN457417, isrctn.org. spective, randomized, double-blind,
This article contains Supplementary Data online at http://care.diabetesjournals.org/lookup/suppl/doi:10. placebo-controlled, multicenter study was
2337/dc10-1233/-/DC1. conducted at 36 outpatient centers in
*Additional study investigators are listed in the Supplementary Data.
© 2011 by the American Diabetes Association. Readers may use this article as long as the work is properly Belgium, France, Germany, Italy, the
cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons.org/ Netherlands, Spain, Sweden, and the
licenses/by-nc-nd/3.0/ for details. U.K. between February 2006 and March

828 DIABETES CARE, VOLUME 34, APRIL 2011 care.diabetesjournals.org

 Jones and Associates

2007. The study was conducted in accor- adjustments made as follows: TT .52 population was assessed on a case-by-case
dance with the revised guidelines of the nmol/L, TRT dose reduced to 40 mg/day; basis for each end point. Inclusion was
World Medical Association Declaration of TT ,17 nmol/L, dose increased to 80 allowed if there were no changes in drugs
Helsinki, the International Conference on mg/day. Dummy dose changes were per- during the first 6 months that affected the
Harmonization Standard Operating Pro- formed in the placebo group to maintain end point in question. These were OADs
cedures, and local laws and regulatory blinding. and antiobesity drugs for HOMA-IR (n =
requirements. The study protocol was All laboratory assessments (Sup- 130); OADs, antiobesity drugs, b-blockers,
approved by independent ethics commit- plementary Data) were performed cen- and diuretics for body composition (n =
tees for each center. Participants provided trally by Pivotal Laboratories Ltd., York, 147); lipid-lowering drugs, b-blockers,
written informed consent. U.K. Measurements were taken at base- and a-blockers for lipid profile (n = 139);
Men aged $40 years were eligible to line, then every 3 months for HbA1c, PSA, OADs, antiobesity drugs, b-blockers, and
enter the study if they had confirmed lipids (total cholesterol [TC], HDL cho- diuretics for HbA1c and FPG (n = 140);
hypogonadism (early morning [08002 lesterol, LDL cholesterol, triglycerides, phosphodiesterase-5 inhibitors and
1000 h] total testosterone [TT] #11 lipoprotein a [Lpa]), hematology, bio- b-blockers for AMS (n = 152) and IIEF
nmol/L or free testosterone #255 pmol/L chemistry, testosterone, FPG (no 9-month (n = 154); and antihypertensive drugs for
on two occasions $1 week apart), with at measurement) and insulin (mean of three blood pressure (n = 140).
least two symptoms of hypogonadism samples at 5-min intervals, no 9-month The study was designed to have a
(14) and fulfilled criteria for type 2 diabe- measurement). Free testosterone was power of 95% to detect a mean difference
tes (15) and/or MetS (16) (Supplementary calculated from TT, albumin, and sex in HOMA-IR change from baseline be-
Data). Erectile dysfunction was not an in- hormone binding globulin, using the Ver- tween TRT and placebo of 0.82 (SD 1.7)
clusion criterion. meulen equation (18). Waist circumfer- at 6 months. After completion of an
Exclusion criteria included TRT within ence, BMI, body composition (TANITA interim analysis in 63 patients, the study
6 months of randomization, hormone- BF-300 body fat analyzer; TANITA Cor- power was revised to 80%, giving a final
modulating therapies or topical/systemic poration, Tokyo, Japan), Aging Males target completer sample of 136 patients.
glucocorticoids within 3 months of ran- Symptoms (AMS) score (19), and Interna- All statistical analyses (Chiltern Interna-
domization, or insulin therapy; a history of/ tional Index of Erectile Function (IIEF) tional Ltd; Quantics, Newtown St Boswells,
current prostate or breast cancer; abnor- (20) were assessed at baseline, 6, and 12 U.K.) were conducted using SAS version
mal digital rectal examination suggestive months. Adverse events (AEs) were elic- 8.2 (SAS Institute, Cary, NC) or later.
of prostate carcinoma; severe symptom- ited from the patient at each visit by a non- Descriptive statistics were calculated for
atic benign prostatic hyperplasia or ele- leading question. continuous variables, with 95% CI, SD,
vated age-specific prostate-specific antigen and P values. For comparison of efficacy
(PSA) (17). For PSA levels and other exclu- Statistical analysis outcomes between treatment groups
sion criteria see the Supplementary Data. The primary end point was the difference (treatment difference [TD]), an ANCOVA
Concomitant medications (oral antidia- between treatment groups in homeostasis model was applied, with disease category
betic drugs [OADs], lipid-lowering thera- model assessment (HOMA)-IR from base- as a covariate (MetS and type 2 diabetes;
pies, antihypertensives) were continued if line to months 6 and 12. Secondary end MetS only; type 2 diabetes only). P , 0.05
doses had remained stable for $3 months points were changes from baseline in was considered statistically significant
before randomization. The first 6 months HbA1c, fasting insulin, FPG, lipid param- for all analyses. No adjustment was
of the study, during which no therapy eters, body composition, BMI, waist cir- made for multiple testing as all second-
changes were permitted unless the investi- cumference, AMS and IIEF scores, AEs, ary end points were regarded as explor-
gator deemed it necessary for good clinical and other safety parameters. HOMA of atory. For the primary end point and
practice, was designated Phase 1. The sec- b-cell function (HOMA-B) was deter- HOMA-B, a log transformation was ap-
ond 6-month period, when changes in mined post hoc. plied for the ANCOVA and a last obser-
medication were allowed, was designated The safety population comprised all vation carried forward approach was
Phase 2. randomized patients who received $1 used. Unless otherwise stated, all efficacy
Subjects were randomized (1:1) to dose of study medication. The primary data were analyzed in the ITT popula-
receive either 3 g metered-dose 2% tes- analysis of all end points was based on tion, all safety data were analyzed in the
tosterone gel (60 mg testosterone, Tos- the intention-to-treat (ITT) population safety population, and all results are
tran [also known as Fortigel, Tostrex, (all randomized patients). Two other quoted as mean 6 SD.
Itnogen, Foresta; ProStrakan, Galashiels, populations were also studied: the per-
Scotland, U.K.]) or placebo gel once daily protocol (PP) population (randomized RESULTS—Of 518 patients screened,
and stratified to treatment arms according patients who received $1 dose of study 220 were randomized (testosterone n =
to the presence of MetS only and type 2 medication with no protocol violations 108; placebo n = 112) and received at least
diabetes with or without MetS. All indi- during the first 6 months of the study) one dose of study drug (ITT and safety
viduals involved in monitoring, data and the modified PP (mPP) population. populations). Baseline demographic and
management, or other study aspects This latter population was established clinical characteristics were comparable
were blinded to treatment. post hoc because the PP population was between groups (Table 1). Phase 1 was
Treatment was applied daily substantially reduced in size after exclu- completed by 157 patients (71%), and
(070021000 h) to clean, dry, intact skin sion of patients who took unauthorized Phases 1 and 2 were completed by 118
on the thighs or abdomen. Blood samples concomitant medications from the ITT (54%) (Supplementary Fig. A1). Reasons
were taken 2 h after application. TT was population (Supplementary Data). Pa- for early withdrawal included protocol
measured at 2, 4, and 12 weeks with dose tients’ eligibility for inclusion in the mPP violation (n = 28), withdrawal of consent

care.diabetesjournals.org DIABETES CARE, VOLUME 34, APRIL 2011 829

Testosterone therapy in hypogonadal men

Table 1—Baseline characteristics (safety population)

Testosterone Placebo Total

n 108 112 220
Age (years) 59.9 6 9.1 (37277) 59.9 6 9.4 (39288) 59.9 6 9.3 (37288)
Race
Caucasian 105 (97) 110 (98) 215 (98)
Asian 3 (3) 0 3 (1)
Afro-Caribbean 0 1 (1) 1 (0.5)
Other 0 1 (1) 1 (0.5)
Total serum testosterone (nmol/L) 9.2 6 2.6 (3.8220.1) 9.5 6 3.3 (2.2220.6) 9.4 6 3.0 (2.2220.6)
Free testosterone (pmol) 198.0 6 49.3 (422319) 202.4 6 62.1 (352364) 200.3 6 56.1 (352364)
Type 2 diabetes 68 (6) 69 (62) 137 (62)
MetS 88 (82) 88 (79) 176 (80)
Insulin resistance (HOMA-IR index) 5.9 6 3.8 (1220) 4.9 6 3.3 (1221) 5.4 6 3.6 (1221)
Hematocrit (L/L) 0.43 6 0.04 (0.3420.54) 0.43 6 0.04 (0.3320.51) 0.43 6 0.04 (0.3320.54)
Hemoglobin (g/dL) 14.9 6 1.5 (11.0218.6) 14.9 6 1.3 (11.4218.0) 14.9 6 1.4 (11.0218.6)
PSA (mg/L) 1.6 6 1.8 (0.21213.60) 1.2 6 1.2 (0.2026.38) 1.4 6 1.5 (0.20213.60)
FSH* (IU/L) 12.3 6 13.4 (1.2263.6) 12.5 6 15.7 (1.2278.9) 12.4 6 14.5 (1.2278.9)
FSH*
Low 0 0 0
Normal 77 (71) 79 (71) 156 (71)
High 31 (29) 33 (29) 64 (29)
LH* (IU/L) 5.7 6 5.7 (0.9234.3) 5.3 6 5.4 (0.6236.1) 5.5 6 5.5 (0.6236.1)
LH*
Low 0 2 (2) 2 (1)
Normal 86 (80) 92 (82) 178 (81)
High 22 (20) 18 (16) 40 (18)
Data are mean 6 SD (range) or n (%) unless otherwise indicated. FSH, follicle stimulating hormone; LH, luteinizing hormone. *Screening values (not measured at
baseline). Normal ranges: 0.7 IU/L # FSH # 11.1 IU/L; 0.8 IU/L # LH # 7.6 IU/L.

(n = 27), AEs (n = 26) and loss to follow- placebo at 6 months (Fig. 1). Type 2 di- in all three analysis groups (total popula-
up (n = 9). Protocol was violated by 125 abetic patients demonstrated mean re- tion [TD 20.049 mmol/L (20.094 to
patients, giving a PP population of 95. ductions in HOMA-IR of 16.0% (0–29; 20.004); P = 0.032], MetS [TD 20.058
Hypogonadotrophic or secondary hypo- P = 0.049) compared with placebo. The mmol/L (20.105 to 20.011); P = 0.016],
gonadism, defined as low or normal lutei- difference between treatments in the MetS type 2 diabetes [TD 20.062 mmol/L
nizing hormone, occurred in 82% of patients approached significance (TD (20.123 to 20.002); P = 0.043]). There
subjects whereas 18% demonstrated pri- 13.3% [21 to 26]; P = 0.069) (Fig. 1). were no significant effects of TRT on tri-
mary hypogonadism (Supplementary Post hoc analyses of HOMA-IR in patients glycerides, abdominal obesity, percentage
Table A1). The majority of subjects without type 2 diabetes only, showed no body fat, BMI, or waist circumference.
(80%) had MetS at baseline, and 62% difference between treatment groups In the overall ITT population, TRT
had type 2 diabetes. MetS and type 2 di- (Supplementary Table A2). There were was associated with significantly greater
abetes were both present in 44 and 40% of no significant differences between treat- increases from baseline in total IIEF score
the TRT and placebo groups, respectively. ments in the overall population or the dis- (TD 4.868 [95% CI 0.644–9.092]; P =
Baseline serum TT values were 9.2 6 2.6 ease subgroups in HbA1c (Fig. 1), fasting 0.024) and in the sexual desire domain
nmol/L (TRT) and 9.5 6 3.3 nmol/L (pla- serum insulin, FPG, or HOMA-B (Table 2). (TD 0.800 [0.271–1.329]; P = 0.003)
cebo). The majority of subjects (62.6%) TRT had a number of beneficial (Table 2). Treatment had no significant
qualified for study entry, meeting both TT effects on lipid profile. In the MetS sub- effect on erectile function, orgasmic func-
and free testosterone criteria. Of the sub- group, TRT was associated with signifi- tion, or overall sexual satisfaction, al-
jects, 12.3 and 23.7% qualified on the ba- cantly greater reductions from baseline in though these three measures did show a
sis of TT ,11 nmol/L or free testosterone plasma levels of Lpa (TD 20.31 mmol/L consistent trend for greater improvement
,255 pmol/L, respectively. [95% CI 20.543 to 20.082]; P = 0.008), with TRT versus placebo. Change in
TC (TD 20.336 mmol/L [20.558 to AMS score was similar in both treatment
Efficacy: Phase 1 20.114]; P = 0.003) and LDL cholesterol groups.
At the end of Phase 1 (6 months), serum (TD 20.210 mmol/L [9520.374 to Data for secondary end points in the
TT values in the ITT population had 20.047]; P = 0.012) than placebo (Table mPP and PP populations are shown in
increased by 19.0 6 22.1 nmol/L with 2). TRT also reduced Lpa in the overall Supplementary Tables A3 and A4, respec-
TRT versus 0.1 6 2.9 nmol/L with pla- ITT population (TD 20.235 mmol/L tively. The small size of the PP population
cebo (P , 0.001). [20.431 to 20.039]; P = 0.019). HDL negated statistical power and the differ-
TRT reduced HOMA-IR by 15.2% cholesterol decreased from baseline sig- ence between groups in HOMA-IR re-
(95% CI 3–26; P = 0.018) compared with nificantly more with TRT than placebo duction was not statistically significant

830 DIABETES CARE, VOLUME 34, APRIL 2011 care.diabetesjournals.org

 Jones and Associates

Figure 1—Mean (95% CI) percentage change from baseline in HOMA-IR (ITT population, last observation carried forward) and change from
baseline in HbA1c (ITT population, study completers) among all patients (A and D), patients with type 2 diabetes (B and E), and patients with MetS
(C and F). P values reported for comparisons between groups.

(TRT 20.92 6 2.454, placebo 20.06 6 associated with significant reductions with TRT (Supplementary Table A3).
2.451; P = 0.17). However, reductions from baseline in percentage body fat, There were no significant between-group
from baseline in Lpa, TC, LDL choles- Lpa, TC, and LDL cholesterol in all pa- differences in HDL cholesterol in the mPP
terol, and HDL cholesterol were signifi- tients and in the MetS subgroup population (Supplementary Table A3).
cantly greater in the TRT group than in (Supplementary Table A3). The type 2 di- No other end points were significantly
the placebo group (overall PP cohort). In abetes subgroup showed significant re- different between treatment groups in
the exploratory mPP population, TRT was ductions in Lpa and percentage body fat the PP or mPP populations.

care.diabetesjournals.org DIABETES CARE, VOLUME 34, APRIL 2011 831

Testosterone therapy in hypogonadal men

Table 2—Mean values for primary and secondary end points at baseline and months 6 and 12 (ITT population, last observation
carried forward)

All patients MetS patients Type 2 diabetic patients

Testosterone Placebo Testosterone Placebo Testosterone Placebo
Glucose metabolism
Fasting serum insulin (pmol/L)
Baseline 151.68 6 153.26 137.65 6 147.46 162.17 6 166.29 153.76 6 163.02 145.08 6 158.55 126.19 6 109.17
n 102 105 82 81 64 65
Phase 1: Month 6 124.59 6 98.65 128.20 6 99.76 125.08 6 86.89 135.50 6 104.45 121.61 6 102.01 136.40 6 117.44
n 101 101 83 80 65 63
Phase 2: Month 12 134.25 6 108.57 126.89 6 94.31 135.50 6 100.92 131.33 6 98.47 130.91 6 113.99 131.96 6 104.30
n 103 101 85 80 65 63
FPG (mmol/L)
Baseline 7.94 6 2.99 7.59 6 2.62 7.60 6 2.41 7.60 6 2.71 9.05 6 3.18 8.49 6 2.84
n 104 107 84 83 67 67
Phase 1: Month 6 7.84 6 3.03 7.99 6 2.86 7.67 6 2.82 7.88 6 2.99 8.91 6 3.31 9.09 6 2.97
n 102 101 84 80 65 63
Phase 2: Month 12 7.97 6 3.45 8.14 6 3.19 7.74 6 3.25 8.05 6 3.41 9.18 6 3.75 9.35 6 3.46
n 102 101 84 80 65 63
HOMA-B†
Baseline 104.8 6 77.5 99.9 6 75.8 110.5 6 78.3 104.7 6 74.7 83.3 6 67.0 84.1 6 74.0
n 99 105 80 81 64 66
Phase 1: Month 6 100.7 6 78.9 98.3 6 72.7 103.9 6 79.4 106.0 6 74.9 79.4 6 66.0 75.3 6 58.2
n 105 108 86 84 67 67
Phase 2: Month 12 102.0 6 79.5 99.4 6 70.8 105.1 6 80.1 105.4 6 70.2 79.1 6 65.4 79.3 6 66.5
n 105 109 86 85 67 68
Lipid parameters
Lpa (mmol/L)
Baseline 1.44 6 1.56 1.40 6 1.17 1.50 6 1.64 1.42 6 1.21 1.36 6 1.36 1.37 6 1.28
n 98* 102 81** 79 62 64
Phase 1: Month 6 1.22 6 1.19 1.59 6 1.80 1.22 6 1.20 1.49 6 1.65 1.20 6 1.16 1.61 6 2.15
n 96* 98 80** 79 62 60
Phase 2: Month 9 1.39 6 1.39 1.90 6 1.94 1.37 6 1.43 1.77 6 1.58 1.51 6 1.52 1.96 6 2.34
n 51* 56 43 44 35 35
Phase 2: Month 12 1.26 6 1.22 1.62 6 1.82 1.26 6 1.24 1.49 6 1.38 1.29 6 1.26 1.65 6 2.19
n 99 100 83** 80 63 61
TC (mmol/L)
Baseline 4.65 6 1.18 4.87 6 1.11 4.79 6 1.22 4.98 6 1.09 4.51 6 1.17 4.55 6 1.01
n 106 108 86 84 67 67
Phase 1: Month 6 4.52 6 1.11 4.89 6 1.13 4.56 6 1.11 5.05 6 1.13 4.44 6 1.12 4.57 6 0.97
n 103 102 85** 81 66 63
Phase 2: Month 9 4.55 6 1.24 4.80 6 1.12 4.61 6 1.24 4.95 6 1.10 4.32 6 1.14 4.59 6 1.13
n 62 65 54 52 42 39
Phase 2: Month 12 4.49 6 1.16 4.77 6 1.07 4.52 6 1.17 4.88 6 1.05 4.31 6 1.05 4.52 6 0.95
n 104 102 86 81 66 63
LDL cholesterol (mmol/L)
Baseline 2.78 6 0.99 2.89 6 0.96 2.89 6 1.02 2.98 6 0.95 2.61 6 0.91 2.61 6 0.82
n 106 108 86 84 67 67
Phase 1: Month 6 2.65 6 0.91 2.85 6 0.91 2.70 6 0.95 2.99 6 0.93 2.60 6 0.85 2.55 6 0.70
n 103 102 85* 81 66 63
Phase 2: Month 9 2.67 6 1.01 2.77 6 0.88 2.70 6 1.01 2.90 6 0.90 2.53 6 0.93 2.51 6 0.77
n 62 65 54 52 42 39
Phase 2: Month 12 2.59 6 0.91 2.75 6 0.89 2.62 6 0.93 2.85 6 0.91 2.49 6 0.82 2.50 6 0.70
n 104 102 86 81 66 63
HDL cholesterol (mmol/L)
Baseline 1.19 6 0.30 1.23 6 0.28 1.21 6 0.31 1.21 6 0.26 1.16 6 0.31 1.23 6 0.29
n 106 108 86 84 67 67
Phase 1: Month 6 1.12 6 0.26* 1.21 6 0.25 1.12 6 0.24 1.20 6 0.24 1.09 6 0.27 1.21 6 0.26

832 DIABETES CARE, VOLUME 34, APRIL 2011 care.diabetesjournals.org

 Jones and Associates

Table 2—Continued

All patients MetS patients Type 2 diabetic patients

Testosterone Placebo Testosterone Placebo Testosterone Placebo
n 103 102 85* 81 66* 63
Phase 2: Month 9 1.14 6 0.31 1.20 6 0.27 1.14 6 0.28 1.18 6 0.27 1.12 6 0.33 1.20 6 0.27
n 62 65 54* 52 42 39
Phase 2: Month 12 1.14 6 0.30 1.21 6 0.26 1.14 6 0.28 1.20 6 0.25 1.09 6 0.30 1.21 6 0.27
n 104 102 86 81 66 63
Triglycerides (mmol/L)
Baseline 1.92 6 1.36 2.10 6 1.44 1.94 6 1.42 2.21 6 1.52 2.06 6 1.58 2.05 6 1.30
n 106 108 86 84 67 67
Phase 1: Month 6 1.93 6 1.23 2.04 6 1.28 1.92 6 1.05 2.10 6 1.34 1.98 6 1.31 2.07 6 1.30
n 103 102 85 81 66 63
Phase 2: Month 9 2.00 6 1.21 2.23 6 1.71 2.07 6 1.26 2.29 6 1.76 1.91 6 0.98 2.36 6 1.99
n 62 65 54 52 42 39
Phase 2: Month 12 1.90 6 1.15 2.14 6 1.49 1.91 6 1.06 2.16 6 1.54 1.90 6 1.17 2.19 6 1.49
n 104 102 86 81 66 63
Abdominal obesity and body composition
Body fat (%)
Baseline 33.60 6 6.54 32.24 6 6.33 33.99 6 6.69 32.37 6 6.17 33.58 6 5.43 32.54 6 6.55
n 108 112 88 88 68 69
Phase 1: Month 6 33.01 6 6.83 32.31 6 5.77 33.31 6 6.82 32.43 6 5.59 32.86 6 5.92 32.33 6 5.88
n 94 98 78 78 61 62
Phase 2: Month 12 32.81 6 7.08 32.63 6 5.57 33.08 6 7.18 32.66 6 5.51 32.69 6 6.44 32.68 6 5.71
n 100 98 83 78 63 62
BMI (kg/m2)
Baseline 32.87 6 6.58 31.29 6 5.44 33.27 6 6.80 31.63 6 5.07 32.76 6 6.12 31.56 6 5.87
n 108 112 88 88 68 69
Phase 1: Month 6 33.07 6 6.96 31.45 6 5.26 33.28 6 7.28 31.67 6 5.03 33.16 6 6.35 31.70 6 5.55
n 97 99 80 79 62 62
Phase 2: Month 12 32.90 6 7.01 31.43 6 5.27 33.17 6 7.35 31.53 6 5.07 32.92 6 6.52 31.73 6 5.57
n 102 99 84 79 64 62
Waist circumference (cm)
Baseline 112.7 6 13.22 110.1 6 13.77 113.3 6 13.43 110.8 6 12.65 112.7 6 13.35 111.7 6 15.23
n 108 112 88 88 68 69
Phase 1: Month 6 111.9 6 13.64 110.2 6 13.36 112.0 6 13.62 110.3 6 12.02 113.1 6 13.57 111.8 6 14.46
n 96 99 80 79 61 62
Phase 2: Month 12 111.6 6 13.89 110.6 6 13.96 112.0 6 13.92 110.8 6 12.52 111.9 6 13.79 112.4 6 15.52
n 100 99 83 79 63* 62
Blood pressure
Systolic blood pressure (mmHg)
Baseline 138.6 6 17.30 136.7 6 17.12 — — — —
n 108 112
Phase 2: Month 12 138.7 6 15.15 134.9 6 16.49 — — — —
n 50 58
Diastolic blood pressure (mmHg)
Baseline 82.5 6 10.23 81.6 6 9.50
n 108 112
Phase 2: Month 12 82.8 6 9.72 80.3 6 9.80 — — — —
n 50 58
Sexual dysfunction
IIEF total score
Baseline 32.4 6 18.47 32.8 6 21.24 — — — —
n 94 98
Phase 1: Month 6 41.3 6 19.97 39.3 6 22.12 — — — —
n 82* 81

care.diabetesjournals.org DIABETES CARE, VOLUME 34, APRIL 2011 833

Testosterone therapy in hypogonadal men

Table 2—Continued

All patients MetS patients Type 2 diabetic patients

Testosterone Placebo Testosterone Placebo Testosterone Placebo
Phase 2: Month 12 40.9 6 20.34 37.3 6 22.08 — — — —
n 88* 88
Erectile function domain score
Baseline 12.0 6 9.07 11.9 6 10.02
n 101 103
Phase 1: Month 6 15.6 6 9.53 15.1 6 10.49 — — — —
n 84 89
Phase 2: Month 12 15.2 6 9.71 14.4 6 10.42 — — — —
n 90 93
Orgasmic function domain score
Baseline 4.6 6 3.80 4.7 6 4.20 — — —
n 105 107
Phase 1: Month 6 5.9 6 3.89 5.8 6 4.18 — — — —
n 88 90
Phase 2: Month 12 5.5 6 3.91 5.3 6 4.23 — — — —
n 95 93
Sexual desire domain score
Baseline 5.3 6 2.13 5.6 6 2.47 — — —
n 105 106
Phase 1: Month 6 6.5 6 2.27 5.9 6 2.38 — — — —
n 89** 91
Phase 2: Month 12 6.3 6 2.21 5.8 6 2.55 — — — —
n 94** 95
Intercourse satisfaction domain score
Baseline 4.7 6 4.58 4.7 6 4.84 — — — —
n 106 108
Phase 1: Month 6 6.2 6 4.95 6.3 6 4.99 — — — —
n 88 91
Phase 2: Month 12 6.4 6 5.01 5.7 6 4.89 — — — —
n 94* 92
Overall sexual satisfaction domain score
Baseline 4.7 6 2.46 4.8 6 2.68 — — — —
n 96 99
Phase 1: Month 6 6.0 6 2.62 5.7 6 2.81 — — — —
n 85 89
Phase 2: Month 12 5.9 6 2.70 5.5 6 2.73 — — — —
n 91 94
AMS total score
Baseline 40.6 6 11.44 39.9 6 11.46 — — — —
n 88 96
Phase 1: Month 6 37.1 6 11.72 36.1 6 10.60 — — — —
n 80 86
Phase 2: Month 12 36.4 6 11.16 37.8 6 11.57 — — — —
n 89 92
Data are means 6 SD or n. *P , 0.05. **P , 0.01 between groups. †HOMA-B = (20 x FPI [mIU/L])/(FPG [mmol/L]23.5) (calculated excluding FPG values # 3.5
mmol/L [formula invalid] and atypical insulin values [$249 pmol/L screening values substituted when baseline values unavailable]).

Efficacy: Phase 2 except for HOMA-IR (the primary end decreased slightly by month 9 with TRT
At the end of Phase 2 (12 months), serum point) and sexual function, a measure of but increased with placebo, creating a dif-
TT values in the ITT population were hypogonadism. In the total ITT popula- ference of 20.58% that was significant
significantly higher following TRT tion, HOMA-IR was significantly im- (P = 0.005). This trend continued nonsig-
(19.5 6 17.0 nmol/L) versus placebo proved with TRT versus placebo at 12 nificantly at 12 months (20.49%, P =
(20.6 6 3.3 nmol/L) (P , 0.001). months (TD 16.4% [95% CI 5–26]; P = 0.066).
Changes in concomitant medications 0.006). There was a significant improvement
were allowed in Phase 2. Efficacy results Exploratory analysis of the type 2 in IIEF total score and sexual desire and
are therefore not presented for this phase diabetes mPP subgroup showed HbA1c intercourse satisfaction domain scores in

834 DIABETES CARE, VOLUME 34, APRIL 2011 care.diabetesjournals.org

 Jones and Associates

all patients at 12 months in the ITT men with type 2 diabetes and/or MetS cardiovascular benefits of reducing Lpa
population (Table 2). The positive effect improves IR, the central biochemical are unknown.
of TRT on the erectile function domain in defect associated with these conditions. The safety and tolerability profile of
the ITT population (Table 2) approached This result is of added interest, as many TRT was comparable to placebo with no
significance in the mPP exploratory anal- study participants were already taking significant differences in cardiovascular
ysis (TD 3.225 [95% CI 20.454 to OADs (mainly metformin), which affect events, including blood pressure. Caution
6.904]; P = 0.089). insulin sensitivity. Although an im- has been urged following a recent study
provement in IR could be expected to that reported an increase in cardiovascu-
Safety: Phases 1 and 2 result in better glycemic control, no re- lar events in testosterone-treated, frail,
In the safety population, 137 patients duction in HbA1c was demonstrated af- elderly men (24). However, cardiovascu-
ter 6 months of treatment. Previous lar events were not a planned outcome
(62%) experienced 502 AEs with 224
(45%) considered to be related to study studies have reported that TRT does im- in this study, and verification of events
medication (TRT, 116; placebo, 108).
prove glycemic control (6,9); however, varied among subjects. Moreover, .70%
the current study was not powered to of the men in that trial received higher
The majority of AEs were classified as
detect changes in glycemic control and doses of testosterone (100 or 150 mg/day)
mild or moderate (TRT, 98%; placebo,
included patients with controlled dia- than any of the men in the current study
97%) with no significant differences be-
betes (HbA1c ,6.5%). A significant dif- (40280 mg/day). The current study
tween treatments. The most common AEs
ference in HbA 1c between the TRT showed a trend for cardiovascular events
in patients experiencing $1 AE were
and placebo groups was observed at 9 to occur more frequently in the placebo
erythema (TRT, 4%; placebo, 5%), pru-
months, but this potential treatment ef- group than in the TRT group. A similar
ritis (TRT, 4%; placebo, 4%), and naso-
fect was confounded by the permitted frailty study that used standard doses (tes-
pharyngitis (TRT, 3%; placebo 4%)
(Supplementary Table A5). changes in diabetes-related medications tosterone gel 50 mg/day) in clinical prac-
for ethical reasons. Therefore, no clear tice found no increase in cardiovascular
A total of 17 serious AEs were expe-
conclusion can be made. A larger defin- events (25).
rienced by 13 patients. Of these, 3 in the
itive study with HbA1c as a primary end The frequency of application site re-
placebo group were considered related to
point in hypogonadal men with uncon- actions in this study was higher than has
treatment. Thirteen resolved without se-
trolled diabetes is required to investigate been reported in other trials involving
quelae. One patient (placebo) died as the
this further. transdermal TRT. However, in TIMES2
result of myocardial infarction. A total of
The majority of cross-sectional stud- mandatory application site inspections
30 patients withdrew from the study
ies have shown that hypogonadism is were performed at each study visit, and
because of AEs (TRT, 18; placebo, 12).
associated with dyslipidemia, albeit with the investigators were trained to identify
The most common AEs leading to dis-
continuation were skin and subcutaneous conflicting results (4). There are only lim- and document skin-related reactions.
ited data supporting the effect of testos- This is in contrast to other studies that
tissue disorders (21 AEs; testosterone, 11;
terone gel formulations on lipids. Two relied on patients reporting a dermato-
placebo, 10). Cardiovascular events oc-
curred more commonly with placebo studies reported a small rise in HDL cho- logical AE.
lesterol but no effect on TC or LDL cho- The beneficial effects of TRT on over-
(10.7 vs. 4.6%; P = 0.095).
lesterol in hypogonadal men (9,21). In all sexual function and, importantly, li-
No clinically relevant mean changes
the current study, Lpa was significantly bido (a key symptom of testosterone
in laboratory parameters were noted.
reduced compared with placebo in the deficiency) support the assessment and
After 12 months, hematocrit had in-
overall population after 6 months of ther- treatment of hypogonadism in men with
creased by 0.03 6 0.04 L/L for TRT
apy. Moreover, the subgroup of patients type 2 diabetes and/or MetS who do not
versus a decrease of 20.01 6 0.02 L/L
with MetS at baseline showed significant present specifically with erectile dysfunc-
for placebo, and hemoglobin levels had
reductions from baseline with TRT versus tion.
increased by 1.42 6 1.55 g/dL from base-
placebo, in Lpa, TC, and LDL cholesterol Results from the mPP population
line (TRT) versus 0.06 6 0.7 g/dL (pla-
cebo). There was no significant increase in after 6 months, even though the majority should be regarded as exploratory only,
of these patients were taking statins and since this population was created as part
age-adjusted PSA values. The PSA level
had mean baseline TC concentrations of a post hoc analysis. However, these
exceeded the upper limit of age-adjusted
below 5 mmol/L. A reduction in HDL results have been included because they
normal values in three subjects at baseline
cholesterol has been reported in studies allow assessment of the impact of testos-
(testosterone, 2; placebo, 1). This viola-
using intramuscular testosterone ester terone on individual parameters without
tion of the entry criteria would not have
therapy (4). the potential confounding influence of
affected the primary efficacy end point,
Lpa is a strong independent risk changes in concomitant medication. Al-
and because the patients had been ran-
factor for premature coronary heart dis- though this information is typically
domized and were satisfactorily proceed-
ease (22). Surgical orchidectomy for pros- gained from the PP population, the small
ing in the study they were not withdrawn.
None of these patients underwent pros- tate cancer increases Lpa concentrations, size of the PP population here limited the
whereas administration of testosterone in power of any analyses and precluded
tate biopsy during the study. PSA levels
healthy men reduces Lpa levels (23). In analysis of disease subgroup data (MetS
exceeded normal limits in four subjects at
12 months (TRT, 3; placebo, 1). the current study, reductions from base- and type 2 diabetes). The mPP data pro-
line in serum Lpa concentration were sig- vide supportive evidence that TRT im-
nificantly greater in TRT-treated patients proves glycemic control, lipid profile,
CONCLUSIONS—This study demon- at month 6 in both the overall population central obesity and body composition in
strates that transdermal TRT in hypogonadal and the MetS subgroup. The long-term hypogonadal men with type 2 diabetes

care.diabetesjournals.org DIABETES CARE, VOLUME 34, APRIL 2011 835

Testosterone therapy in hypogonadal men

and/or MetS. In this population, the loss H.M.B., J.B., E.M., I.M., A.M.M., M.V., A.Y., subnormal plasma testosterone. J Androl
of statistical significance for some end and J.D.H. contributed to data analysis and 2009;30:726–733
points between 6 and 12 months is likely interpretation and to the writing and editing 10. Yialamas MA, Dwyer AA, Hanley E, Lee H,
to be due to the smaller size of the of the manuscript. All authors had access Pitteloud N, Hayes FJ. Acute sex steroid
to all data and had final responsibility for withdrawal reduces insulin sensitivity in
population at month 12 and the allowed
the manuscript content; T.H.J. had final re- healthy men with idiopathic hypogonado-
use of concomitant medications after 6 sponsibility for the decision to submit for tropic hypogonadism. J Clin Endocrinol
months (Supplementary Table A3). The publication. Metab 2007;92:4254–4259
delayed effect of TRT on HbA1c could The authors thank the TIMES2 investigators 11. Alibhai SM, Duong-Hua M, Sutradhar R,
be explained by the cumulative effect and their staff for the conduct of this study and et al. Impact of androgen deprivation
on reduction in body fat over several Polly Winter, Euro RSCG Life Medicom therapy on cardiovascular disease and
months. (supported by ProStrakan) for additional diabetes. J Clin Oncol 2009;27:3452–
The high dropout rate is a limitation writing support and assistance with the colla- 3458
of this study. However, there was no tion of author comments. 12. Keating NL, O’Malley AJ, Smith MR. Di-
difference between treatment groups in abetes and cardiovascular disease during
the number or characteristics of the pa- androgen deprivation therapy for pros-
tate cancer. J Clin Oncol 2006;24:4448–
tients who failed to complete the study. 4456
References
This study has shown that TRT in 13. Jones TH. Testosterone deficiency: a risk
1. Ding EL, Song Y, Malik VS, Liu S. Sex
hypogonadal men with type 2 diabetes differences of endogenous sex hormones factor for cardiovascular disease? Trends
and/or MetS improves several cardiovas- and risk of type 2 diabetes: a systematic Endocrinol Metab 2010;21:496–503
cular risk factors, most notably IR, Lpa, review and meta-analysis. JAMA 2006; 14. Nieschlag E, Swerdloff R, Behre HM,
TC, and LDL cholesterol. These findings 295:1288–1299 et al. Investigation, treatment, and moni-
and the therapeutic benefit on sexual 2. Guay A, Jacobson J. The relationship be- toring of late-onset hypogonadism in males:
function strongly support a role for TRT tween testosterone levels, the metabolic ISA, ISSAM, and EAU recommendations.
in these clinical conditions. syndrome (by two criteria), and insulin J Androl 2006;27:135–137
resistance in a population of men with 15. World Health Organization, Department
organic erectile dysfunction. J Sex Med of Noncommunicable Disease Surveil-
2007;4:1046–1055 lance. Definition, Diagnosis and Classification
Acknowledgments—This study was spon- 3. Kapoor D, Aldred H, Clark S, Channer KS, of Diabetes Mellitus and its Complications:
sored and funded by ProStrakan. Investigators Jones TH. Clinical and biochemical as- Report of a WHO Consultation. Part 1: Di-
collected study data and the sponsor moni- sessment of hypogonadism in men with agnosis and classification of diabetes. Geneva,
tored the study conduct. Data collection and type 2 diabetes: correlations with bio- Switzerland. World Health Org., 1999
analysis was performed by Chiltern Inter- available testosterone and visceral adi- 16. International Diabetes Federation. (2006)
national Ltd., U.K.; post-hoc analysis was per- posity. Diabetes Care 2007;30:911–917 The IDF consensus worldwide definition
formed by AY (Quantics Consulting Limited, 4. Jones TH, Saad F. The effects of testos- of the metabolic syndrome. Brussels: In-
U.K.). T.H.J. has served as a consultant to terone on risk factors for, and the mediators ternational Diabetes Federation. Available
ProStrakan and on advisory boards and as a of, the atherosclerotic process. Atheroscle- from http://www.idf.org/webdata/docs/
speaker for ProStrakan, Bayer-Schering Pharma, rosis 2009;207:318–327 IDF_Meta_def_final.pdf. Accessed 25
and Ipsen. K.S.C. has served as a consultant to 5. Dhindsa S, Prabhakar S, Sethi M, March 2010
ProStrakan. S.A. has received lecture fees and Bandyopadhyay A, Chaudhuri A, Dandona 17. Richardson TD, Oesterling JE. Age-specific
served as a consultant to ProStrakan and Bayer, P. Frequent occurrence of hypogonado- reference ranges for serum prostate-specific
and has received research grants from Bayer. tropic hypogonadism in type 2 diabetes. antigen. Urol Clin North Am 1997;24:339–
H.M.B. has received honoraria for lectures or J Clin Endocrinol Metab 2004;89:5462– 351
has served on the advisory boards of Bayer, 5468 18. Vermeulen A, Kaufman JM. Diagnosis of
Ferring Pharmaceuticals, ProStrakan, and 6. Kapoor D, Goodwin E, Channer KS, Jones hypogonadism in the aging male. Aging
Solvay Pharmaceuticals and has received re- TH. Testosterone replacement therapy Male 2002;5:170–176
search grants from Acrux, Bayer, ProStrakan, improves insulin resistance, glycaemic 19. Heinemann LAJ, Zimmermann T, Vermeulen
and Solvay Pharmaceuticals. J.B. has recently control, visceral adiposity and hyper- A, Thiel C, Hummel W. A new ‘Aging male’s
been, or is currently a consultant and/or cholesterolaemia in hypogonadal men symptoms’ (AMS) rating scale. Aging Male
speaker or investigator for Bayer-Schering with type 2 diabetes. Eur J Endocrinol 1999;2:105–114
Pharma, Boehringer Ingelheim, Eli Lilly and 2006;154:899–906 20. Rosen RC, Riley A, Wagner G, Osterloh
Co., Johnson & Johnson, Janssen-Cilag, and 7. Naharci MI, Pinar M, Bolu E, Olgun A. IH, Kirkpatrick J, Mishra A. The inter-
ProStrakan. I.M. has served on the advisory Effect of testosterone on insulin sensitivity national index of erectile function (IIEF):
boards of Janssen-Cilag, Eli Lilly and Co., in men with idiopathic hypogonadotropic a multidimensional scale for assessment
Bayer-Schering, and ProStrakan; as a speaker hypogonadism. Endocr Pract 2007;13: of erectile dysfunction. Urology 1997;49:
for Janssen-Cilag, Eli Lilly and Co., and Bayer- 629–635 822–830
Schering; and as a clinical investigator for Eli 8. Pitteloud N, Mootha VK, Dwyer AA, et al. 21. Wang C, Cunningham G, Dobs A, et al.
Lilly and Co.; and has received research grants Relationship between testosterone levels, Long-term testosterone gel (AndroGel)
from Pfizer. A.M.M. has served on the advisory insulin sensitivity, and mitochondrial treatment maintains beneficial effects on
board of ProStrakan. A.Y. has served as a re- function in men. Diabetes Care 2005;28: sexual function and mood, lean and fat
tained consultant for ProStrakan. J.D.H. is an 1636–1642 mass, and bone mineral density in hypo-
employee of ProStrakan. No other potential 9. Heufelder AE, Saad F, Bunck MC, Gooren gonadal men. J Clin Endocrinol Metab
conflicts of interest relevant to this article were L. Fifty-two-week treatment with diet and 2004;89:2085–2098
reported. exercise plus transdermal testosterone 22. Rosengren A, Wilhelmsen L, Eriksson E,
T.H.J. and K.S.C. conceived and designed reverses the metabolic syndrome and im- Risberg B, Wedel H. Lipoprotein (a) and
the study protocol and produced the initial proves glycemic control in men with coronary heart disease: a prospective case-
draft of the manuscript. T.H.J., K.S.C., S.A., newly diagnosed type 2 diabetes and control study in a general population

836 DIABETES CARE, VOLUME 34, APRIL 2011 care.diabetesjournals.org

 Jones and Associates

sample of middle aged men. BMJ 1990; 24. Basaria S, Coviello AD, Travison TG, et al. strength, physical function, body compo-
301:1248–1251 Adverse events associated with testoster- sition, and quality of life in intermediate-
23. Zmuda JM, Thompson PD, Dickenson R, one administration. N Engl J Med 2010; frail and frail elderly men: a randomized,
Bausserman LL. Testosterone decreases 363:109–122 double-blind, placebo-controlled study.
lipoprotein(a) in men. Am J Cardiol 1996; 25. Srinivas-Shankar U, Roberts SA, Connolly J Clin Endocrinol Metab 2010;95:639–
77:1244–1247 MJ, et al. Effects of testosterone on muscle 650

care.diabetesjournals.org DIABETES CARE, VOLUME 34, APRIL 2011 837