Hypogonadal men with type 2 diabetes benefit from testosterone treatment: 10-year real-life data from a registry study 
 

F Saad, A Haider, KS Haider, G Doros, A Traish



Background: Prevalence of hypogonadism in men with type 2 diabetes mellitus (T2DM) can be as high as 50%. 

Aim: To study long-term effectiveness and safety of testosterone undecanoate injections (TU) in a subgroup of hypogonadal men from an ongoing registry study in a urological setting.

Method: 
A registry of hypogonadal men was established in 2004 in a single urologist’s office. All men had testosterone ≤12.1 nmol/L. In a total 776 patients, 286 (36.9%) had type 2 diabetes which had been diagnosed and is being treated elsewhere (family physician, diabetologist).
Of these, 133 received TU 1000 mg/12 weeks (T-group) following an initial 6-week interval. 153 had opted against TTh and served as controls (CTRL). Measurements were performed 1-4 times a year for up to 11 years and included, anthropometric parameters, glycaemic control and concomitant diabetes medication. 10-year data are reported. Mean changes over time between groups were compared by mixed effects model for repeated measures with random effect for intercept and fixed effects for time, group and their interaction and adjusted for age, weight, waist circumference, fasting glucose, blood pressure and lipids to account for baseline differences between groups. Mean follow-up was 6.9 years in the T-group and 7.3 years in CTRL, median follow-up 7 years in the T-group and 8 years in CTRL. 

Results:
Mean age at baseline was 61.8±5.4 years in the T-group and 64.2±4.6 years in CTRL. 

Mean HbA1c progressively decreased from 8.8±0.9 to 5.9±0.3% after 10 years in the T-group (p<0.0001). The decrease was statistically significant vs. previous year for the first 8 years, then continued without reaching statistical significance. In CTRL, mean HbA1c increased from 7.7±0.6 to 9.5±0.9% (p<0.0001). The estimated adjusted difference between groups was -4.6% (p<0.0001). 

Fasting glucose decreased from 7.6±1.1 to 5.3±0.1 mmol/L (p<0.0001) in the T-group. The decrease was statistically significant vs. previous year for the first 3 years. In CTRL, fasting glucose increased from 6.2±0.6 to 6.9±1.0 mmol/L (p<0.0001). The estimated adjusted difference between groups was -2.2 mmol/L (p<0.0001).

HOMA-IR decreased from 10.2±2.0 to 3.6±1.2 after 10 years (p<0.0001). The decrease was statistically significant vs. previous year for the first 5 years. In CTRL, HOMA-IR increased from 7.4±1.4 to 11.4 (p<0.0001). The estimated adjusted difference between groups was -9.4 mmol/L (p<0.0001).

Fasting insulin decreased from 29.6±4.2 to 15±4.8 µU/mL (p<0.0001) with statistical significance vs. previous year for the first 7 years. In CTRL, fasting insulin increased from 26.5±2.6 to 37 µU/mL (p<0.0001). The estimated adjusted difference between groups was -23.1 µU/mL (p<0.0001).

At baseline, 54 patients in the T-group received insulin at a mean dose of 32.4±12.1 units/day. Mean dose requirement declined to 20.6±11.3 units (p<0.0001) with statistical significance vs. previous year for the first 7 years. In CTRL, 55 patients received insulin at a mean dose of 29.4±5 units/day. Mean dose requirement increased to 39.5±7.2 units (p<0.0001). The estimated adjusted difference between groups was -25.7 units/day (p<0.0001).

In the T-group, 106 (80%) achieved HbA1c <6.5%, and 116 (87%) achieved HbA1c <7% at last measurement. In CTRL, no patient achieved either HbA1c <6.5% or HbA1c <7.0%. All but 1 man had an increase in HbA1c.

In the T-group, 4 patients (3%) died of non-cardiovascular events. There were no non-fatal major adverse cardiovascular events (MACE). In CTRL, 24 patients (15.7%) died. 50 non-fatal MACE occurred (26 myocardial infarctions, 24 strokes).

Since all injections were administered in the doctor’s office and documented, there was a 100% adherence to testosterone treatment. 

Discussion: 
Limitations: The study was not designed to investigate effects of testosterone on type 2 diabetes. Also, patients were not randomised. In a regular medical office, neither randomisation nor placebo-control were an option. After receiving information about testosterone treatment, to undergo testosterone therapy was the patients' own choice. There was no information about adherence to antidiabetic medications. The untreated control patients were approximately 2.5 years older at baseline. However, their baseline diabetes parameters were more favourable than they were in the testosterone-treated patients. 
Strengths: The long duration of follow-up, the excellent medication adherence to testosterone, and close interdisciplinary collaboration with the diabetologists.
Conclusion: long-term treatment with TU in hypogonadal men with T2DM improved glycaemic control compared to untreated controls. In patients in the T-group, insulin dose could be substantially reduced. Correcting hypogonadism in men with T2DM sustainably supports standard diabetes treatment.