• Future trials on prostate cancer should use consensus definitions for patient enrollment criteria, subgroup characteristics, and trial endpoints, such as those developed by the World Health Organization.

There is a large body of randomized controlled trials on androgen suppression for advanced prostate cancer. However, the ability to compare and combine the results of these trials is hampered by the lack of consistent definitions of outcomes and prognostic groups, inconsistencies in followup intervals and reporting of adverse effects, and the use of a variety of methods for evaluating disease stage, grade, and progression.

The use of consensus definitions and criteria for patient enrollment, subgroup stratification, and study endpoints (Andersson, 1997) will produce a more robust body of cumulative evidence by improving the ability to compare results among trials and increasing the potential for combined analyses. Standardized reporting of adverse effects using the common toxicity criteria jointly developed by the U.S. National Cancer Institute (NCI) and the European Organization for Research and Treatment of Cancer (EORTC, 1996) also is desirable.

• The hypotheses that combined androgen blockade provides a greater benefit than monotherapy either for all men with advanced prostate cancer or for a subgroup of patients with good prognostic factors is not supported by the available evidence and does not merit continued investigation.

The preponderance of evidence finds no statistically significant difference in survival between men treated with combined androgen blockade or monotherapy. This evidence consists of 27 randomized controlled trials, including 7,987 patients. Among these is the most recent and largest trial by the Southwestern Oncology Group (INT 0105, n=1,382), which found no significant difference between combined androgen blockade and monotherapy. Additional trials that address the aggregate population of men with advanced prostate cancer are unlikely to alter the balance of the evidence.

The evidence favoring combined androgen blockade consists of three trials, which report improvements in median survival ranging from 3.7 to 7 months; our meta-analysis of 5-year survival (hazard ratio 0.871; 95 percent CI 0.805 to 0.942); and a prior meta-analysis restricted to studies using nonsteroidal antiandrogens for combined androgen blockade (hazard ratio 0.84; 95 percent CI 0.76 to 0.93) (Caubet, Tosteson, Dong et al., 1997). Although the meta-analyses show a statistically significant difference in survival at 5 years, the magnitude of this difference is of questionable clinical significance.

One hypothesis is that the favorable findings, although few and of somewhat disappointing magnitude, are attributable to a subgroup of patients who benefit substantially from combined androgen blockade. To date, the main subgroup of interest has been patients with minimal disease and good performance status. However, the SWOG trial (INT 0105), the only study prospectively designed and adequately powered to compare outcomes for this subgroup, reported no significant difference in survival between combined androgen blockade and monotherapy in these good-risk patients (Eisenberger, Blumenstein, Crawford et al., 1998). Thus, the available evidence does not appear to justify continued investigation of this hypothesis. In the event that future efforts define characteristics of another patient subgroup that appears to benefit from combined androgen blockade, further research may be warranted.

• Randomized controlled trials are needed to assess the efficacy of various strategies for the timing of androgen suppression. The most urgent priorities for future research include: treatment at biochemical progression, intermittent androgen suppression, and neoadjuvant androgen suppression.

The patients presently being considered for androgen suppression differ from those in most trials we reviewed. This is due to the widespread use of PSA measurement to detect and monitor prostate cancer. Consequently, men now are considered for androgen suppression earlier in the natural history of the disease than was previously the case. Therefore, questions of optimal timing for androgen suppression have become paramount.

Alternative strategies are emerging for this patient population. These include: androgen suppression initiated at biochemical progression after definitive therapy; intermittent therapy, where androgen suppression is begun at PSA rise after definitive therapy and discontinued when PSA returns to nadir; androgen suppression initiated immediately upon discovery of locally advanced disease at the time of prostatectomy; and neoadjuvant therapy initiated prior to either surgery or radiation therapy for clinically localized disease, with the objective of downstaging the tumor for definitive therapy. The evidence we reviewed showing that adjuvant androgen suppression initiated with radiotherapy and continued for several years or more improves survival by more than radiotherapy alone followed by androgen suppression at progression should be viewed within this context. Followup trials are needed to compare radiotherapy plus androgen suppression with androgen suppression alone in patients with locally advanced or asymptomatic metastatic disease.

For each strategy, the threshold question is whether overall or cancer-specific survival is improved. But the trials must also provide adequate information to assess any survival advantages in relation to the impact of prolonged duration of therapy on the quality of life and on the risk and severity of adverse effects. Trials should be designed prospectively to address whether there are subgroups of patients who are more likely to benefit from each strategy for the timing of androgen suppression.

• Evidence collected from patients on the effects of various androgen suppression therapies on the quality of life is urgently needed. The information obtained should be incorporated into patient education materials and used in shared decision making, and its impact on patients' treatment choices should be evaluated.

Scant evidence has been collected from patients using standardized and validated instruments for measuring quality of life. Despite the absence of such data, and in the face of uncertainty with respect to survival advantages, earlier androgen suppression therapy is being adopted into clinical practice and recommended to patients.

Better evidence is needed that compares patient evaluations of the quality of life with various therapies so that patients, with their physicians, can make better informed choices among the available options for androgen suppression therapy. Collecting such evidence and comparing results from separate trials will be facilitated when investigators agree on which disease-specific quality-of-life parameters should be measured and then determine how best to measure them. The reports from two studies that patients and their physicians have markedly different perceptions of the effects of therapy on the quality of life serve to emphasize the urgency of such research. It is also important to evaluate how access to such quality-of-life information affects patients' choices about their treatments.

Although patients overwhelmingly prefer to avoid orchiectomy, it is not known whether patient perceptions when they contemplate this therapy are borne out by the realities of life after surgery. No studies have objectively compared the effects of surgical and medical methods of castration on the quality of life. If quality of life is not adversely affected, orchiectomy would represent a highly cost-effective alternative for androgen suppression. Note, however, that orchiectomy may be viewed rather differently if future research on intermittent androgen suppression makes the reversibility of treatment an important consideration.

• The cost-effectiveness analysis points to additional priorities for future research.

First, as noted above, the quality of life after orchiectomy has a major impact on the cost-effectiveness of this therapy and therefore should be studied using objective measures. Second, DES may be a highly cost-effective strategy for androgen suppression. At the lowest effective dose, 1 mg/d, up to 70 percent of men achieve castrate levels of testosterone, and cardiovascular toxicity is minimized. For the remaining 30 percent, whether a dose of 3 mg/d significantly increases cardiovascular risk merits investigation. The clinical setting has changed markedly since the VACURG trials were conducted. Today's patient population is younger and healthier, and management of cardiovascular risk is vastly improved. If intermittent therapy is demonstrated to be a clinically useful approach, the cost-effectiveness ratio of DES may become even more attractive, as intermittent use may further reduce cardiovascular risk. Third, trials of androgen suppression therapies should collect economic data so that the costs of the available therapies can be directly compared and to provide more complete empirical data for cost-effectiveness analysis.