Several series have described the rate of recovery of testosterone levels following AST.5–11
Factors that may influence the rate of recovery include age5,6,8,10
duration of LHRH-A therapy,9,11,17
baseline testosterone levels,5,8,10
type of AST agents used10
and the type of LHRH-A preparation used.10
To minimize the duration of side effects and long-term risks of androgen suppression, an appealing strategy is to use preparations of LHRH-A that could potentially hasten the recovery of testosterone levels after completion of AST. Another strategy is to use non-steroidal anti-androgens alone.18
We are intrigued with the type of LHRH-A preparation used and its effect on rate of testosterone recovery. In a retrospective analysis by Pickles and colleagues,10
49 men who were treated with curative intent with external radiation therapy received treatment with the 1-month preparation of LHRH-A for a median duration of 9 months (range: 3–26 months). Another 68 men who received radiation therapy were given the 3-month preparation of LHRH-A for a median duration 11 months (range: 3–35 months). Additionally, 150 men received cyproterone acetate/stilbestrol (CPA/DES) for a median duration of 7 months (range: 3–34 months). Total testosterone measurements were obtained 6 weeks after completion of radiation, then every 6 months for 3 years and yearly thereafter. Interestingly, the median time to recovery of testosterone levels to the lower limit reference range calculated from the end of AST was 7, 8 and 16 months for the groups receiving CPA/DES, 1 month preparation of LHRH-A and 3 month preparation of LHRH-A, respectively. On multivariable analysis, factors associated with a delay in testosterone recovery were advanced age, low baseline testosterone level and use of the 3-month preparation of LHRH-A. However, there were some imbalances in the distribution of factors between the 2 groups of patients receiving different LHRH-A preparations that may have influenced the results. For example, men who received the 3-month preparation of LHRH-A were more likely to be elderly and receive a longer duration of AST. Thus, the observation that men who received longer preparation of LHRH-A had a slower recovery of testosterone was only hypothesis-generating from this retrospective series; our intent with this present randomized study was to validate these findings.
In this study, the median time to recovery of testosterone to baseline value was twice as long for the L-LHRH-A arm (8 months) compared to the S-LHRH-A arm (4 months). However, this result was not statistically significant. The median time to recovery of testosterone to lower limit of reference range was again twice as long for the L-LHRH-A arm (4 months) compared to the S-LHRH-A arm (2 months). This was not statistically significant, but a trend was noted favouring faster recovery using S-LHRH-A (p = 0.087). A Cox multivariable analysis failed to identify the treatment arm as a significant predictor of time to recovery of testosterone.
There are several possible explanations why we did not observe a statistically significant difference in rate of testosterone recovery, despite a trend observed favouring faster recovery in the patients receiving S-LHRH-A. The study did not reach the planned sample size of 100 and may not have had enough power to detect a statistically significant difference in outcome. Additionally, measurements of testosterone were incomplete in many patients. For example, only 29% and 45% of patients in the short-acting and long-acting preparations, respectively, had full testosterone levels available. This raises concern that the data collected may not necessarily be representative of the entire cohort of men, given that in both arms only a minority of patients actually completed the all the blood testing. Note that the Pearson’s chi-square test of homogeneity did not reveal any imbalance in the percent of completed blood tests between treatment arms, which could have biased results. Perhaps a statistically significant difference in testosterone recovery rates could have been detected, with a larger sample size and a more complete set of blood results. However, one may argue whether the observed difference of 2 months to the lower limit of reference range between treatment arms is clinically significant and worth the additional injections required to give the 1-month preparations. Our testosterone recovery results share some similarities to that of Pickles and colleagues.10
The difference in testosterone recovery to the lower limit of reference range between treatment arms was by a factor of 2. However, there is considerable difference in the absolute time to recovery between the 2 studies. For example, in the Pickles’ retrospective study, the median time to testosterone recovery to the lower limit reference range was 16 and 8 months for the L-LHRH-A and S-LHRH-A groups, respectively. In our study, the corresponding times were 4 and 2 months, respectively. We noted, however, that the testosterone measurements in Pickles study were done 6 weeks post-radiotherapy and then every 6 months, which is considerably less frequent than every 2 months, as in this current study. The longer duration between testosterone testing in the Pickles study may have resulted in prolongation of the recovery rates. This further makes comparisons of our results to Pickles difficult to make. Additionally, it is difficult to compare our recovery rates to those reported by other investigators given the heterogeneity in age distribution of patients, dosing, duration and the type of hormonal agents used between studies and the different definitions used for testosterone recovery. Our results are more congruent with those of Gulley and colleagues who reported a median recovery time to the lower limit of reference range of 4 months, in keeping with our results in the L-LHRH-A arm.8
In that study, men also received 6 months of the 3-month preparation of LHRH-A, along with thalidomide or placebo thereafter. Furthermore, in a study by Oefelein and colleagues, a small series of patients received a single 3-month preparation of LHRH-A ± flutamide prior to prostatectomy or external beam radiotherapy.12
The median time to testosterone recovery was 4 months after cessation of LHRH-A therapy. The discrepancy between the longer median time to recovery to the lower limit of the reference range in Pickles’ study and the shorter recovery times noted in the other studies, including ours, may be explained by the use of external beam radiotherapy to the prostate in Pickles’ study. External beam radiotherapy delivers a higher cumulative dose of radiation to the testes compared to I-125 implants used in this study, which can further suppress Leydig cell function and delay testosterone recovery.19–23
For example, another study by Pickles and colleagues showed that testosterone levels can fall to about 83% of the baseline value 6 months after external beam radiotherapy to the prostate.22
Other limitations of this study include the lack of free or bio-available testosterone measurements, which are the bioactive forms of circulating testosterone that may directly influence the side effects of AST than total testosterone levels. Additionally, we did not record the time of the day at which the testosterone measurements were performed; it is known that testosterone levels should be drawn in the morning as levels tend to decrease later in the day.
We observed between one-fifth to one-third of patients did not have recovery of testosterone levels to baseline at the last follow-up of 18 months following completion of AST (). It is likely, however, that with longer follow-up, a higher proportion of men would have recovered their testosterone levels. In the study by D’Amico and colleagues, 9% of men failed to experience a return of testosterone to baseline after long-term median follow-up of 7.5 years.6
However, only 1 patient in our entire cohort failed to recover their testosterone levels to the lower limit of the reference range. Others have reported full recovery of testosterone levels after AST.7,24,25
Recovery of testosterone levels to the lower limit of the normal range is likely more clinically relevant than recovery to the baseline value. Thus, we conclude that the risk of permanent or prolonged androgen suppression with the use of 6 months of LHRH-A is negligible and acceptable.
Several investigators have reported a shorter time to recovery of testosterone levels in men under the age of 60 to 67 years.5,6,8,10
Acknowledging the same sample sizes, we also reported a faster testosterone recovery time (to baseline) in men under the age of 65 years, however, only in the group that received S-LHRH-A (). The median time to recovery of testosterone levels to baseline was 4 months versus 16 months for the <65 years and ≥65 years of age groups, respectively (log-rank test, p
= 0.038). Although the Cox model hazard ratios for age ≥65 years versus age <65 years showed that the older age group had longer times to testosterone recovery than the younger group, this was not statistically significant.
Fig. 3. Kaplan Meier curves of proportion of patients in the short-acting luteinizing hormone-releasing hormone agonist (LHRH-A) treatment arm with testosterone recovery to baseline level in men < 65 years or (n=12) or ≥ 65 years of age (n=12). (more ...)