The rationale for combined androgen blockade began with the recognition that suppressing testicular production of androgens did not lead to the complete absence of androgens, and that all tumors ultimately gained the ability to grow in castrate conditions. In animal models, the addition of androgen receptor blockade to castration therapy enhanced antitumor efficacy. Early clinical trials seemed favorable, and two SWOG trials, 8494 and 8894, were completed to test combined blockade against medical or surgical castration monotherapy [14, 15]. The addition of flutamide to medical castration with leuprolide significantly prolonged survival, 35.6 months versus 28.3 months, two-sided p = 0.035 [14]. However the addition of an anti-androgen to surgical castration did not significantly prolong survival (HR 0.91, 95 % confidence interval 0.81–1.01) [15]. A European (EORTC) trial mirroring the leuprolide trial also documented a survival advantage to combined blockade [16]. Several meta-analyses of combined blockade trials have been published; despite methodological differences, there appears to be consensus that combined blockade is associated with a small but detectable improvement in 5-year overall survival, but questions remain about cost-benefit and quality of life trade-offs [17–19].

Continuous testosterone deprivation was initially the standard, given that bilateral orchiectomy was the original means of reducing circulating testosterone, leading to permanent absence of testicular androgen production. After the development of LHRH analogues, the ability to reversibly suppress testosterone production fuelled interest in applying therapy intermittently. Using the Shionogi mouse model, Bruchovsky and colleagues [20] demonstrated that the number of castration-resistant prostate cancer stem cells was reduced by applying intermittent therapy, suggesting that cancer cells adapt to the hormone milieu, and suggesting that castration resistance could be delayed by intermittently re-introducing testosterone. Experiments in which prostate tumors were reduced via castration, then transplanted into non-castrate hosts, then re-transplanted into castrate hosts resulted in a fold-fold prolongation of time to castration resistance compared to tumors which remained in the castrate host [21]. In addition to the physiological rationale, there is a strong incentive to utilize intermittent therapy in order to improve quality of life and reduce adverse effects via testosterone recovery during the “off-treatment” period.

Multiple phase II and small randomized trials of intermittent androgen deprivation have documented the safety of the approach, reviewed by Boccon-Gibod and colleagues [22]. These studies document the intermittent experience across a wide range of induction durations, from as little as 3 months to as long as 48 months, and include studies using diethylstilbestrol, LHRH analogue monotherapy, and combined blockade with an antiandrogen. The triggers for resumption of ADT have been similarly variable, although essentially all have been PSA-based, ranging from any rise in PSA to a 50 % increase in nadir, an absolute PSA level >3 ng/mL to >20 ng/mL, and of course clinical progression being an absolute indication to reinitiate therapy regardless of PSA level. Importantly, nearly all patients respond to re-treatment and testosterone levels typically do recover, bringing with them restoration of sexual function for a majority of men. One small randomized trial suggested progression at 3 years was markedly lower for intermittent therapy [23], however larger studies have not borne this out.

Two large, randomized phase III trials comparing intermittent to continuous ADT have issued at least preliminary reports. The South European Uroncological Group (SEUG) treated 766 men with locally advanced or metastatic prostate cancer with cyproterone acetate plus monthly LHRH analogue for an induction period of 3 months [24]. The 626 men whose PSA declined to <4 ng/mL or to 80 % below the baseline value were randomized to intermittent or continuous therapy; those assigned to intermittent therapy stopped treatment at randomization. The re-treatment trigger was two-tiered; PSA >10 ng/mL for symptomatic patients and 20 ng/mL for asymptomatic patients and for patients whose PSA declined to below 80 % of the initial value, a PSA rise >20 % above the nadir triggered the next cycle of ADT. Survival was similar, with disease progression documented in 127 men from the intermittent arm and 107 men from the continuous arm; while there were more cancer-related deaths in the intermittent arm, there were more cardiovascular deaths in the continuous arm. Quality of life was not significantly different in the two arms, except that men on the intermittent arm had greater sexual function. A second multinational trial of intermittent therapy, JPR7, reported preliminary findings at the 2011 annual meeting of the American Society of Clinical Oncology [25]. One thousand three hundred and eighty-seven men with biochemical recurrence were treated with 8 months of LHRH analogue monotherapy. Those who achieved a PSA <4 ng/mL and did not show signs of castration resistance were randomized to intermittent or continuous therapy. Men on the intermittent arm resumed therapy when PSA reached 10 ng/mL and each treatment cycle was 8 months. For the primary endpoint of overall survival, there was no significant difference, and the p value was 0.009 for non-inferiority; as in the SEUG trial there were more prostate cancer deaths on the intermittent arm but more non-prostate cancer deaths on the continuous arm. Men on the intermittent arm spent only 27 % of their time on treatment, which would be expected to yield cost savings; quality of life data are not yet available. SWOG 9346 has a similar design, except that the treatment comparison is being made in metastatic prostate cancer patients; these data are anxiously awaited.

The ability to block AR activity without depleting the body of testosterone led to the question of whether this “peripheral” blockade could be therapeutic in the absence of castration therapy. The side effect profile may be favorable, however the available data suggest that it may yield inferior disease control compared to castration or combined blockade. A Swedish study which compared bicalutamide 50 mg to orchidectomy in metastatic prostate cancer found that survival was significantly better in the castration group, HR 1.76 (95 % confidence interval 1.27–2.44) [26]. An Italian study which compared bicalutamide 150 mg to goserelin plus flutamide in advanced prostate cancer found no difference in progression-free or overall survival, although combined blockade trended towards superior results in the subgroup of metastatic patients [27]. Some have argued that the addition of 5-alpha reductase inhibitors more completely deprives cancer cells of androgen fuel, and yields deeper PSA nadir than bicalutamide monotherapy, comparable to that achieved with castration [28]. On the other hand, the 150 mg dose of bicalutamide has been associated with some safety concerns, such as a higher death rate when added to active surveillance in the early prostate cancer trialists group study [29], which has led the United States and Canada to recommend against prescribing the 150 mg dose [30]. Overall, at present, most physicians would not include peripheral androgen blockade as a standard first-line treatment option.

No randomized studies to date have proven that the incorporation of chemotherapy prior to the emergence of castration resistance is beneficial. However, in vitro experiments show synergism between taxanes and androgen deprivation, and in mouse models combination therapy was associated with a doubling of time to disease progression compared to androgen deprivation followed by chemotherapy [31]. A European trial tested combined androgen blockade alone or with epirubicin and found that progression-free survival was significantly prolonged 12 months versus 18 months, p = 0.02, and overall survival showed a trend toward improvement 22 months versus 30 months, p = 0.1; in the subgroup of men with >5 bone metastases the survival difference approached significance, 17 months versus 27 months, p = 0.06 [32]. Phase II studies have shown promise for the addition of docetaxel to androgen deprivation therapy at biochemical recurrence [33], but the phase III studies which will definitively answer the question have not yet been reported.

One year of castration therapy is associated with the loss of 4 % of bone mineral density in the hip and 2 % in the spine [Diamond; 45] which corresponds to a fracture risk as high as 20 % at 10 years [35]. Bisphosphonates are effective in counteracting the bone mineral density loss; pamidronate administered every 12 weeks during a year of medical castration offset the bone mineral density loss induced by castration alone [46]. However quarterly administration, and even a single dose of zoledronic acid was subsequently shown to increase bone mineral density during a year of LHRH analogue therapy [36, 47]. Thus, bone mineral density should be evaluated at baseline and periodically during androgen deprivation so that appropriate patients requiring intervention can be identified. Calcium and vitamin D should be part of the approach to support bone mineral density, and weight-bearing exercise can also be recommended. The selective estrogen receptor modulator toremifene was also studied in this arena and was found to increase bone mineral density by 1.6 % in men receiving ADT compared to 0.7 % decrease in the placebo arm [47].

Bone metastases can cause significant morbidity, including pain, fracture, and spinal cord compression which are collectively labelled “skeletal related events”. Unravelling the unique crosstalk between prostate cancer cells and bone marrow stroma have facilitated the development and study of bone-targeted therapies. In castration-resistant metastatic disease, zoledronic acid administered intravenously every 3–4 weeks was shown to reduce the risk of skeletal related events from 44 to 33 %, including fractures, which were reduced from 22 to 13 % [49]. In addition, zoledronic acid was associated with a significant delay in the time to the first skeletal event, from 321 days to 488 days, p = 0.009 [50]. Denosumab, a monoclonal antibody against RANK ligand showed an even stronger reduction and delay in skeletal related events compared to zoledronic acid, with a median time to skeletal-related event of 17.1 months for zoledronic acid compared to 20.7 months for denosumab [51]. The benefits of these agents in castration-sensitive disease have not yet been well described; a CTSU trial C90202 is currently enrolling men with bony metastatic disease and randomizing them to early bisphosphonate therapy versus bisphosphonates at the time of castration resistance with a primary endpoint of skeletal related events. This will inform clinicians of the relative risks and benefits of starting therapy earlier; significant adverse events related to bone-targeted therapies in general include osteonecrosis of the jaw and for bisphosphonates there is also the potential for renal toxicity.

Increases in serum total cholesterol and triglycerides occur during androgen deprivation, though the degree and pattern of changes varies somewhat according to the form of ADT (ex: monotherapy or combined blockade). For instance, LHRH analogue therapy for 1 year has been associated with a 9 % increase in total cholesterol and 26 % increase in triglycerides [37]. Although this has been difficult to correlate with clinical outcomes such as myocardial infarction or cerebrovascular accident, the changes are concerning. Selective modulation of the estrogen receptor, for instance with toremifene, appears to ameliorate these changes, leading to an increase in HDL, decrease in total cholesterol as well as LDL and triglycerides [48].

Androgen deprivation is known to induce insulin resistance, which appears to be a direct consequence of hypogonadism and independent of body mass index changes during ADT [52]. Using Medicare data, investigators documented an increased risk of developing outright diabetes after orchiectomy or LHRH analogue use, with hazard ratios of 1.34 and 1.44, respectively [53]. In addition, abdominal obesity and the metabolic syndrome are more prevalent in men on long-term ADT than in matched prostate cancer and normal controls (55 % versus 22 % and 20 %, respectively) [54]. Clearly this is a serious problem which warrants attention. Lifestyle counseling should be part of the treatment plan when androgen deprivation is employed, though specific interventions have not yet been prospectively evaluated.

Alarm was raised when Medicare database analysis and review of radiation oncology clinical trial data suggested a significant increase in cardiovascular morbidity, namely myocardial infarction and sudden cardiac death, when men received androgen deprivation adjunctive to radiation [53, 55]. This is physiologically plausible, given the known propensity for androgen deprivation to cause weight gain, worsening lipids, and insulin resistance. However, other analyses have failed to find an association between ADT and cardiac mortality, or have found it limited to certain risk groups, such as those over 65 or with serious comorbidities [38]. As such there is no recommendation for routine involvement of cardiology in men undergoing ADT, but rather standard measures to control contributing risk factors such as lipids and blood pressure should apply. Undoubtedly this will continue to be an area of active research moving forward.