Is there still a role for intermittent androgen deprivation therapy for prostate cancer?—Yes.
Intermittent androgen deprivation (IAD) therapy was first described in mid 1980s in a small pilot study of men with symptomatic, metastatic prostate cancer.1 Men underwent androgen deprivation therapy (ADT) with diethylstilbestrol until symptoms resolved, at which time therapy was withheld until symptomatic progression. The authors1 observed that sexual function returned in 9 of 10 patients who had normal erectile function at baseline. The conclusion was that IAD could adequately palliate symptomatic men and, at the same time, result in improved quality of life (QOL) by allowing return of sexual function in the majority with normal function at baseline.
A few years later, Bruchovsky et al2 proposed that reexposure of prostate cancer stem cells to androgens could reinduce an apoptotic potential in their progeny, prolonging the time to androgen resistance. This was demonstrated in the Shionogi mammary androgen-dependent murine model. The feasibility and safety of IAD was subsequently studied in numerous phase 2 trials. These, in combination with preclinical data showing delayed time to castration resistance, prompted 9 phase 3 trials comparing continuous ADT (CAD) vs IAD with respect to overall survival (OS) and QOL. These consistently showed no difference in OS, with QOL benefits. Seven of them were limited by small sample size and/or heterogeneous patient cohorts. The 2 largest phase 3 trials published to date are the PR7 trial, designed by the National Cancer Institute of Canada, in men with nonmetastatic hormone-sensitive disease3 and S9346, a trial initiated by the Southwest Oncology Group (SWOG), in men with hormone-sensitive metastatic disease.4 These trials were intended to be complementary, using the same dose and schedule for IAD. Both were designed as noninferiority studies, with OS as the primary end point in the PR7 trial and as a coprimary end point in the S9346 trial. Quality of life was a secondary end point in PR7 and a coprimary end point in S9346.
In the PR7 trial,3 the median OS of men with nonmetastatic disease on IAD was 8.8 years compared with 9.1 years in CAD arm (hazard ratio [HR], 1.02; 95% CI, 0.86-1.21), meeting criteria for noninferiority. Of the 268 deaths in the IAD arm, 41% were due to prostate cancer and 28% due to non–cancer-related causes compared with 34% and 36% of 246 deaths in the CAD arm, respectively. The median follow-up was 6.9 years.
Scores for symptoms including hot flashes, sexual activity, and urinary symptoms were significantly better in the intermittent arm (P < .001, P < .001, and P < .006, respectively). Quality of life, assessed at fixed time points, trended toward better scores in the IAD than CAD arms but did not reach statistical significance.
Crook5 performed a post hoc analysis of PR7 to look at the impact of Gleason score in this population. Cox regression analysis including Gleason scores of 1 to 6, 7, and 8 to 10 and the treatment arm showed no difference; however, the median OS in those with Gleason scores of 8 to 10 was approximately 14 months longer in the continuous arm (HR, 1.23; P = .31). Crook5 cautioned that although Gleason score was not a planned stratification factor and that the trial was not powered to detect a small or moderate difference, IAD should be probably not be offered to men with Gleason scores of 8 to 10.
The results of the S9346 trial4 yielded inconclusive results. Neither inferiority, noninferiority, nor superiority of either arm could be excluded. However, there was a nonsignificant trend to improvement in median OS in the CAD arm, with a median OS of 5.8 years compared with 5.1 years in the IAD arm (HR, 1.10; 95% CI, 0.97-1.25). Although this represents a 10% relative increase risk of death in the IAD arm, a 20% increase in risk could not be ruled out with 90% confidence.
An unspecified subanalysis stratifying by minimal disease (confined to axial skeleton and pelvis or lymph nodes) vs extensive disease (ribs, long bones, skull, and/or viscera) showed a benefit in the minimal disease group for CAD with a median OS of 6.9 years vs 5.4 years in the IAD arm and no difference in those with extensive disease (median OS, 4.4 years vs 4.9 years, respectively).
The primary comparison in QOL outcomes (erectile dysfunction, libido, vitality, mental health, physical function) was was done at 11 months, that is, 3 months after stopping ADT on the IAD arm or at 11 months of ADT on the CAD arm. Men in the IAD arm reported less erectile dysfunction (P < .001) and better mental health (P = .003), and libido was significant (P < .04) compared with the CAD arm. At 9 months after randomization, 4 of the 5 QOL outcomes favored IAD but by 15 months, only physical functioning was higher in IAD. However, at that time, 78% of the men on the IAD arm had been restarted on ADT, thus confounding this assessment. The overall conclusion was that the effect of IAD on QOL was “marginal” or “modest at best.”6
Median OS was longer than predicted in both the PR7 and S9346 trials, consistent with what is seen in many contemporary phase 3 cancer trials. Both of these large phase 3 trials have statistical issues discussed at length elsewhere.6,7 Both trials were unable to fully capture the QOL changes in the IAD arms owing to the choice of fixed time points for collecting data. However, there was a “signal” in both trials that various domains of QOL were better on IAD than CAD. A recent meta-analysis of IAD vs CAD trials7 demonstrated that IAD was not inferior to CAD with respect to survival and that some QOL criteria were improved. These authors concluded that IAD could be offered as an alternative therapy. Taken together, the results of the 2 large phase 3 trials3,4 and the most recent meta-analysis7 call for taking the individual patient into account when deciding about the option of IAD (Figure). Of note, in the metastatic setting, prostate-specific antigen (PSA) response to ADT is a powerful predictor of survival. Patients who demonstrate an excellent response (PSA level <0.2 ng/mL by 7-8 months of ADT) are ideal candidates for a trial of IAD. In contrast, patients whose PSA fails to nadir less than 4.0 ng/mL are unlikely to benefit.8
aThe trial S9346 initiated by the Southwest Oncology Group established that the absolute prostate-specific antigen (PSA) value after 7 months of ADT was predictive for risk of death: Good risk: a PSA value of 0.2 ng/mL or less; median overall survival (OS), 75 months. Intermediate risk: PSA value greater than 0.2 ng/mL but no greater than 4.0 ng/mL; median OS, 44 months. Poor risk, PSA value greater than 4 ng/mL; median OS, 13 months (P < .001).8
bCounsel regarding post hoc subgroup analyses from the PR7 (nonmetastatic) and S9346 (metastatic) trials. Nonmetastatic: patients with Gleason scores of at least 8 had a 14-month shorter median OS when treated with IAD compared with those treated with CAD. Metastatic: patients with minimal skeletal metastases had a median OS of 6.9 years in the CAD arm vs 5.4 years in the IAD arm. Extensive skeletal metastases: no difference between IAD and CAD for median OS.
cIn men with a preexisting history of cardiovascular disease, use of a gonadotropin-releasing hormone antagonist rather than agonist for ADT should be considered.9 This is especially a consideration for men choosing IAD.10
A recent analysis of the S9346 trial using Medicare claims data to examine long-term health-related effects of IAD vs CAD showed that there was no difference in terms of bone, endocrine, or cognitive events.10 However, there was a decreased 10-year cumulative incidence in ischemic and thrombotic events in the CAD arm (24% vs 33%; HR, 0.69; P = .02). It is known that ADT can increase risk of cardiovascular disease (CVD)-related events and death in men with preexisting CVD, especially in the first 6 months of therapy.11 It is not surprising, therefore, that men on the IAD arm who stop and restart ADT are at greater risk. However, all men in the S9346 trial were treated with a GnRH analog.
Pooled data of trials comparing a GnRH agonist with a gonadotropin-releasing hormone antagonist suggests CVD risk is mitigated by the latter.9 Clinicians should take prior comorbidities into account when prescribing ADT in either schedule.
Intermittent ADT is a viable treatment option for men who experience the adverse effects of ADT. Clinicians who are experienced in administration of IAD, sometimes for decades, can attest to the QOL benefits for the individual patient (and his partner) over time.
Patients treated with IAD therapy see a “light at the end of the tunnel”—the time when their testosterone levels rise, and they feel “more normal,” and, according to their partners, “act more like the person I knew.” Intermittent ADT should be offered in the context of good clinical judgement and should include both the patient and his partner in the deliberations.
Corresponding Author: Laurence Klotz, MD, FRCSC, CM, Division of Urology, Sunnybrook Health Sciences Centre, 2075 Bayview Ave, Ste MG408 Toronto, ON M4N 3M5, Canada (firstname.lastname@example.org).
Published Online: August 25, 2016. doi:10.1001/jamaoncol.2016.3138.
Conflict of Interest Disclosures: Dr Klotz has served on advisory boards for Abbvie, Astellas, Bayer, Sanofi, Ferring, and Tolmar, and has received research support from Ferring, Bayer, and Medivation. Dr Higano has served on advisory boards for Asana Biosciences, AbbVie, Astellas, Bayer, Blue Earth Diagnostics, Churchill Pharma, Clovis, Dendreon, Emergent, Ferring, Medivation, MorphoSys, Orion, and Pfizer; and has received research support from Aragon, AstraZeneca, Bayer, Dendreon, Emergent, Exelixis, Genentech, Medivation, Millennium, and Sanofi. No other disclosures are reported.
Klotz L, Higano CS. Intermittent Androgen Deprivation Therapy—An Important Treatment Option for Prostate Cancer. JAMA Oncol. 2016;2(12):1531-1532. doi:10.1001/jamaoncol.2016.3138