Some 65 years after the landmark discovery by Huggins that prostate tissue is androgen regulated, androgen suppression therapy (AST) remains the primary treatment for advanced stage carcinoma of the prostate (CaP). Since this significant discovery, refinement and validation of the efficacy of AST in a variety of clinical situations has occurred. Despite these advances, the adverse events associated with AST have become an important clinical concern to physicians treating men with carcinoma of the prostate. Specifically, reports regarding cancer treatment-induced bone loss (CTIBL, eg, osteoporosis),1 neurocognitive dysfunction,2 sexual dysfunction, and increased cardiovascular morbidity and mortality associated with AST3 have stimulated a dialogue in a complacent, yet important area of urologic oncology—the use and the misuse of androgen ablation.

Of the potentially harmful therapies oncologist prescribe, AST appears to be in the therapeutic “sweet spot”—highly active, with an initial side-effect profile equivalent to male menopause. Health-related quality of life (HRQOL) assessments of men on AST recognize that early hot flush bother is ultimately replaced with chronic weight gain bother.4 Surprisingly, the diminished libido and erectile dysfunction exacerbated by AST in this aged male population is not infrequently considered less of a worry than the concern regarding an increase in their prostate-specific antigen (PSA) level. The morbidity of AST, however, is insidious and cumulative, and the harms must be fully disclosed.

The Harms

  1. Top of page
  2. The Harms

In CaP patients on AST, weight gain bother is the major symptom of a significant metabolic alteration. Central obesity with a large panus and peripheral lean muscle loss are clinically evident, and these findings suggest the metabolic syndrome.5–7 Adipocytes are now recognized as metabolically active through leptins and adiponectins, which alter insulin sensitivity.7, 8 Bone aromatase activity is diminished in the absence of gonadal steroids, which results in the insidious development of osteoporosis.1 This CTIBL ultimately leads to compromised bone health and to clinical events—fragility and pathological fractures.1 AST-associated neurocognitive dysfunction has been less well studied; however, evidence of significant impairment in memory and mood has been observed.2 Cardiovascular-associated adverse events (coronary heart disease, myocardial infarction, prolonged QT interval and sudden cardiac death) have emerged as a more significant risk.3, 9, 10 In the last decade, efforts to link cardiovascular events to drug therapies have been vigorously pursued (eg, rofecoxib). The clinical environment is, therefore, established, and in the controversial realm of whether or not to treat CaP, this concern regarding excess cardiovascular morbidity further inflames the dialogue.

The connection between cardiac events and AST has history in the Veterans Administration Cooperative Urological Research Group (VACURG) trials,9, 11 and more recently in the Early Prostate Cancer Trial (EPCT).10

Between 1960 and 1975, the VACURG conducted a consecutive series of 3 major randomized clinical trials comparing various endocrine treatments for newly diagnosed prostate cancer patients. Six major conclusions concerning hormonal treatment emerged from these studies: 1) there is increased hazard of cardiovascular death after therapy with 5 mg diethylstilbestrol (DES); 2) orchiectomy plus DES is no better than orchiectomy or DES alone; 3) 1.0 and 5.0 mg DES have an equivalent effect on cancer; 4) there is a reduced cardiovascular hazard from therapy with 1.0 mg DES; 5) Premarin and Provera are no better than 1.0 mg DES at the doses studied; 6) decisions about hormone treatment at diagnosis should be dependent on patient characteristics, mainly age and Gleason grade. The last conclusion was based on a post hoc analysis of VACURG I, which found a 5-mg dose of DES may slow disease progression and decrease death from prostate cancer in younger men (aged 60 years or younger) with metastasis-positive (M+), high-grade disease.9

Additional internal findings of the VACURG II trial are worthy of comment. In this trial, 3 dosages of DES (0.2, 1.0, and 5.0 mg/day) were compared. Men receiving the 0.2-mg/day dose (an ineffective dose based on testosterone response)12 had a significantly shorter overall survival than men receiving 5mg/day of DES.9, 11 These observations have frequently been overlooked in the interpretation of the VACURG outcomes; nevertheless, they demonstrate an important concept—namely, ineffective AST, whether in the form of placebo or a DES dose insufficient to result in effective testosterone suppression (0.2 mg/day), leads to a reduced prostate-specific and overall survival.

In the EPCT, the nonsteriodal antiandrogen bicalutamide (150 mg, daily) was assessed as an alternative to castration therapies to preserve HRQOL.10 Unlike the prostate cancer collaborative trialists study, which only included M+ disease, the EPCT addressed only locally advanced and localized disease. Patients with locally advanced disease benefited from antiandrogen therapy; however, patients with localized disease had a reduced survival if on bicalutamide. It is important to emphasize bicalutamide is a nonsteroidal antiandrogen (NSAA) and blocks the androgen receptor. Castrate therapy, conversely, decreases the production of androgens. The antiandrogen mechanism of action may have independent implications on cardiovascular risk, and in lieu of the EPCT results, I believe this issue may confound (eg, NSAA-induced harm) outcome analysis.

The Benefits

Despite these significant concerns, we must acknowledge the benefits of AST, which are numerous, and they are supported by level I evidence derived from randomized controlled trials and published in high-quality peer-reviewed journals. Specifically, overall survival benefit has been recognized for AST in the neoadjuvant,13, 14 the adjuvant,15 and the treatment settings.9, 11

The neoadjuvant setting

Bolla and associates found that maximum androgen blockade (MAB) initiated before external beam radiotherapy for high-risk CaP patients (Gleasons score ≥8, PSA ≥20, stage ≥T3) and continued for 3 years provided an overall survival advantage compared with radiotherapy alone.13 D'Amico and associates refined this observation in men with intermediate-risk disease (Gleason score ≥7, PSA ≥10, stage ≥T2c); however, the use of MAB was limited to 6 months.14 In so doing, an overall survival advantage was found in these patients treated with neoadjuvant MAB (6 months) versus external beam radiotherapy alone. These 2 studies are important, and they have established the standard of care for intermediate- and high-risk CaP.

The adjuvant setting

Messing and coworkers reported the results of a multicenter trial in which CaP patients found to have lymph node-positive disease at the time of radical prostatectomy were randomized to receive immediate AST or observation until clinical progression.15 A significant overall survival advantage was identified in men who receive immediate versus deferred AST. This important trial established efficacy for AST in the adjuvant setting, and it has been cited as evidence supporting the earlier initiation of AST in patients at risk for CaP progression.

The therapeutic setting

The VACURG established that men with metastatic CaP lived longer if they received AST.9, 11 However, whether AST is begun early or delayed until clinical symptoms develop was addressed, and no survival benefit was identified in patients receiving the early administration of AST (in the form of 5 mg DES) over waiting until symptomatic progression.9 The excess cardiovascular deaths in the immediate treatment group were the basis of this finding. This conclusion regarding the timing of initiating AST remains controversial. Subset analysis of the VACURG I study identified a survival advantage for younger men with high-grade disease who received immediate AST.9

The Criticisms

Recent reports criticizing AST as harmful deserve proper scrutiny, given the level 1 clinical trial results previously cited. Specifically, the reports implicating AST are post hoc analyses of clinical trial data or from Surveillance, Epidemiology, and End Results (SEER)-Medicare database inquiries.3 Importantly, some of these reports, but not all,16 conclude AST is associated with an increased risk of cardiovascular events, including death, and the postulated pathological mechanism is through the development of the metabolic syndrome.3, 5–7 D'Amico and coworkers17 acknowledge that discordant results exist within the literature, and they evoke the competing cause hypothesis to explain the incompatible findings. These authors persuasively argue that the predominate cause of death may mask the identification of secondary causes of death. Although the competing cause hypothesis is a significant concern, the level I evidence cited above provides compelling evidence that CaP patients in specific clinical situations live longer if their treatment plan includes AST. Give these paradoxical results, additional explanations should be introduced.

Although the link between AST, cardiac events, and diabetes is derived from carefully conducted evaluations of SEER-Medicare-linked and post hoc clinical trial data analysis, the input information available for analysis was not prospective and is limited.3 Consequently, these facts raise questions regarding cause and effect. The study limitations are most poignant regarding the omission of important cardiovascular risk factors—namely, tobacco usage, body mass index/waist circumference, cholesterol, and blood pressure medication use, as well as oral anticoagulation and NSAA use.3 These are significant flaws in methodology, raising question regarding the conclusions.

Keating and coworkers,3 accessing the SEER-Medicare-linked database, reported an association between gonadotropin-releasing hormone agonist administration and the development of diabetes, coronary heart disease, myocardial infarction, and sudden cardiac death. This provocative study raises several conundrums, which include: 1) bilateral orchiectomy was associated with diabetes but not associated with cardiovascular events; and 2) the timing of the observed cardiovascular events—peaking within the first 5 months to 12 months of initiating AST and subsequently decreasing as the duration of AST increased—is inconsistent with the proposed pathological cascade—specifically, a hypogonadism-induced metabolic syndrome, causing diabetes and subsequent atherosclerosis, and culminating with cardiovascular events.5–7 Lastly, it should be noted that the exposure to ADT was much more common (2-fold higher) in the population of patients most likely to die from CaP (poorly differentiated Gleason score ≥8). Despite the statistical adjustments performed by these careful investigators, these observations, in total, create uncertainty regarding cause and effect, and raise more questions than they provide answers.

Cause of death, especially in a patient with a diagnosis of prostate cancer, is subject to bias and misreporting. Cardiovascular events are frequently implicated as the cause of death with very limited evidence. Alternatively, and not infrequently, thromboembolism occurs in cancer patients.18, 19 Malignancies, in general, and prostate cancer, specifically, have been associated with a hypercoagulable state.18–20 This component of Virchow triad should be acknowledged in the assignment of cause of death. Although the deep science connecting cancer-induced thromboembolism with venous and arterial events is beyond the scope of this article, I recommend to those interested the work of Duggirala and coworkers.20–22 The takeaway message is that prostate cancer may create a hypercoagulable state and results in both venous and arterial thromboembolic events, which may not be appreciated when cause of death is assigned in this setting.

In summary, AST used in the neoadjuvant, adjuvant, and therapeutic settings described above save more lives than it harms. Nevertheless, AST does cause harm; however, this harm must be properly judged. Men with localized, low-risk CaP with multiple comorbidities may experience few benefits from AST. As such, these patients probably would be better served by an active surveillance approach, with delayed AST if evidence of progression develops. A proactive position regarding the potential for AST to unmask existing or occult coronary artery disease or for AST to facilitate development of cardiovascular disease warrants consideration. In the end, don't throw the baby out with the bath water.


  1. Top of page
  2. The Harms