Differences in Trial and Real-world Populations in the Dutch Castration-resistant Prostate Cancer Registry

BACKGROUND: Trials in castration-resistant prostate cancer (CRPC) treatment have shown improved outcomes, including survival. However, as trial populations are selected, results may not be representative for the real-world population. The aim of this study was to assess the differences between patients treated in a clinical trial versus standard care during the course of CRPC in a real-world CRPC population. DESIGN, SETTING, AND PARTICIPANTS: Castration-resistant Prostate Cancer Registry is a population-based, observational, retrospective registry. CRPC patients from 20 hospitals in the Netherlands have been included from 2010 to 2013. OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS: Baseline characteristics, systemic treatment, and overall survival were the main outcomes. Descriptive statistics, multivariate Cox regression, and multiple imputations with the Monte Carlo Markov Chain method were used. RESULTS AND LIMITATIONS: In total, 1524 patients were enrolled of which 203 patients had participated in trials at any time. The median follow-up period was 23 mo. Patients in the trial group were significantly younger and had less comorbidities. Docetaxel treatment was more frequently used in trial patients (85% vs 40%). Despite an observed unadjusted median overall survival difference of 35 mo versus 24 mo between the trial and standard care group, this difference was not retained after adjustment for baseline characteristics and treatment effect. CONCLUSIONS: At CRPC diagnosis, the baseline characteristics of the patients who had been enrolled in trials notably differed from patients who received standard treatment options only. The survival difference between the trial and standard care group could be explained by baseline differences and treatment effects. These results indicate that trial results cannot easily be translated to real-world practice. PATIENT SUMMARY: We observed that patients treated in clinical trials differed from patients who were not. We concluded that this may lead to differential treatment and survival. Caution is warranted when real-world outcomes are compared with trial results.