Eight relevant studies were identified (n=317): two RCTs (n=26 and n=54); one non-RCT with an unmatched control group (n=66); and five before-after studies (n=171, range 20 to 46). Both RCTs were low quality (Jadad scores <3). One RCT had a high loss to follow-up (26%).
Change in prostrate-specific antigen levels: One RCT (n=54) found a significantly lower mean for prostrate-specific antigen in the treatment group versus controls of 3.0ngmL-1 (SD 1.9) versus 9.0ngmL-1 (SD 7.5) (p<0.001). The other RCT (n=26) found no significant difference between the treatment and control groups. For the non-RCT (n=66) there was a significant reduction in mean prostrate-specific antigen after treatment from 10.9ngmL-1 (SD 1.1) to 8.7ngmL-1 (SD 0.9) (p<0.001). These were also significantly lower than the control group endpoint 13.8 ng mL-1 (SD 2.4). Three of the before-after studies also reported statistically significant reductions in prostrate-specific antigen level after the interventions. The remaining two before-after studies did not have significant results.
Evidence of progression from bone scans: There was a significantly higher number of patients with a normal bone scan (complete response) in the intervention versus the control group in one RCT (25% versus 15%, p<0.02) and progressive disease was significantly less common in the intervention group (p<0.02). One before-after study reported a 25% reduction in overall metastatic lesions.
One RCT reported longer overall survival in the intervention group compared to control at 25.5 months (p<0.001). None of the studies that examined toxicity/side effects (six studies) reported any severe toxicity or intolerance related to intervention. Two studies reported improvements in cancer-related symptoms (details in review).