To determine the rate and time of testosterone (T) recovery in patients (pts) with localised prostate cancer treated with radiotherapy plus 0-, 6-, 18- or 36-month of androgen deprivation therapy (ADT).

In 1230 pts with prostate cancer randomised into two phase III trials, serum T was measured at baseline, then regularly. T recovery rate was compared between normal vs. abnormal baseline T and with ADT duration with Chi-square test or Fisher's exact test. A multivariable logistic regression model to predict the probability of recovering normal T was performed.

Overall, 87.4 % (167/191), 75.9 % (293/386), 54.8 % (181/330) and 43.2 % (80/185) of pts, recovered normal T on the 0-, 6-, 18- or 36-month schedule, respectively (p < 0.001). In patients recovering normal T, the median time to T recovery increased with ADT duration ranging from 0.31, 1.64, 3.06 to 5.0 years for the 0-, 6-, 18- or 36-month schedules, respectively (p < 0.001) and was significantly faster for those with a normal T at baseline (p < 0.001). On multivariable analysis, older age and longer ADT duration are associated with a lower T recovery.

Testosterone recovery rate after ADT depends on several factors including hormonal duration, normal baseline T, age and medical comorbidities. A longer ADT duration is the most important variable affecting T recovery. The data from this report might be a valuable tool to help physicians and patients in evaluating risks and benefits of ADT.