Addition of estramustine to chemotherapy and survival of patients with castration-refractory prostate cancer: a meta-analysis of individual patient data.

BACKGROUND: Estramustine phosphate is a mustard-oestradiol conjugate, and has hormonal and non-hormonal effects. In phase II trials of patients with cancer, response to microtubule inhibitors increases when these drugs are combined with estramustine. We aimed to assess whether combining estramustine with chemotherapy increases survival in patients with castration-refractory prostate cancer.
METHODS: We systematically searched for randomised clinical trials that compared chemotherapy regimens with and without estramustine in patients with histologically-proven prostate cancer and were published between 1966 and 2004. Data from these studies were verified centrally and updated individual patient data were analysed. The primary endpoint was overall survival. Secondary endpoints were prostate-specific antigen (PSA) response, time to PSA progression, and toxicity. A Cox regression model that was stratified by trial and adjusted for covariates at baseline was used.
FINDINGS: The initial search identified seven eligible trials that included 742 patients, from which data from five trials including 605 patients had been collected. Individual patient data from two trials (137 patients) were no longer available. The 605 patients had been accrued between Jan 1, 1993 and Dec 1, 2003 and randomly assigned to chemotherapy plus estramustine or to chemotherapy without estramustine. Chemotherapy (with or without estramustine) consisted of docetaxel, paclitaxel, ixabepilone, and vinblastine. Median follow-up was 2.8 years (range 0.0-3.4), and 510 deaths had occurred by the end of follow-up. Cox regression analysis stratified by trial showed that concentrations of serum haemoglobin (p<0.0001), use of chemotherapy plus estramustine (p=0.008), performance status (p=0.002), and serum PSA concentrations (p=0.04) were associated independently with overall survival. Overall survival was significantly better in patients assigned chemotherapy plus estramustine (adjusted hazard ratio [HR] 0.77 [95% CI 0.63-0.93], p=0.008). Estimated absolute increase in overall survival when estramustine was added to chemotherapy was 9.5% (SE 4.0) at 1 year after randomisation. We did not note a significant association between treatment effect on overall survival and age, concentration of serum haemoglobin, performance status, or serum PSA concentration. Patients who received chemotherapy plus estramustine had a better PSA response than those who received chemotherapy without estramustine (RR 0.53 [0.38-0.72], p<0.0001). Time to PSA progression was significantly longer in patients assigned chemotherapy plus estramustine than in those assigned chemotherapy without estramustine (HR 0.74 [0.58-0.94], p=0.01). Patients assigned chemotherapy and estramustine had more grade 3 or grade 4 thromboembolic events compared with those assigned chemotherapy without estramustine (12 of 271 vs 1 of 275).
INTERPRETATION: In patients with castration-refractory prostate cancer, addition of estramustine to chemotherapy increases time to PSA progression and overall survival compared with chemotherapy without estramustine. However, this benefit should be balanced with the risk of increased thromboembolic events in patients who receive estramustine and chemotherapy in combination compared with chemotherapy without estramustine.