Introduction
Mild Leydig cell insufficiency affects a substantial proportion of testicular cancer survivors. Previous studies have not shown a beneficial effect of testosterone replacement therapy, however, with a pronounced interindividual effect. Thus, biomarkers identifying the subgroups that might benefit are wanted. We aimed to determine if insulin-like factor 3 (INSL3), basal and human chorionic gonadotropin (hCG)-stimulated testosterone can predict the effect of testosterone replacement therapy in testicular cancer survivors with mild Leydig cell insufficiency.

Patients and Methods
We randomized adult testicular cancer survivors with mild Leydig cell insufficiency 1:1 to 12 months of transdermal testosterone replacement therapy (Tostran gel 2%) or placebo. INSL3, basal, and hCG-stimulated testosterone were measured at baseline. Outcomes (glucose, insulin, HbA1C, lipids, blood pressure, and body composition) were measured at baseline, 6 and 12 months. We applied a linear mixed-effect model comparing patients receiving testosterone with placebo in subgroups by biomarker.

Results
We included and randomized 69 patients between October 2016 and February 2018. Patients with INSL3 and hCG-stimulated testosterone concentrations below the median had a -1.7 kg (95% CI: -3.1, -0.4) and -2.0 kg (95% CI: -3.5, -0.6) change in fat mass after 12 months of testosterone replacement therapy compared with placebo. This was not the case in patients with INSL3 and hCG-stimulated testosterone above the median. We did not find any effect of these biomarkers on glucose, insulin, HbA1c, or lipids.

Conclusion
Patients with INSL3 and hCG-stimulated testosterone concentrations below the median had decreased fat mass after 12 months of testosterone replacement therapy compared with placebo. It should be evaluated in larger trials if these biomarkers can be used as predictive markers identifying testicular cancer patients with mild Leydig cell insufficiency who might benefit from testosterone substitution.