Radical prostatectomy in Denmark: Survival analysis and temporal trends in clinicopathological parameters with up to 20 years of follow-up.

OBJECTIVES

To describe mortality, cause of death, and temporal trends in clinicopathological parameters with up to 20 years of follow-up in a nationwide cohort of prostate cancer (PCa) patients who underwent radical prostatectomy (RP).

MATERIALS AND METHODS

A total of 6857 patients with PCa treated with RP at six different hospitals in Denmark between 1995 and 2011. Data were extracted from the nationwide DaPCa database. Histopathology reports from the RP specimens were manually reviewed. Date and cause of death were obtained from national registries and cross-checked in patient files. The cumulative incidence of PCa specific mortality (PCSM) was analysed with the Aalen-Johansen method for competing risks with non-PCa death as a competing event. Risk of PCSM was analysed in a multivariate Cox regression model using age, preoperative PSA level, surgical margin status, RP Gleason score (GS), pathological T-category, and N-category as explanatory variables.

RESULTS

The median follow-up was 6.4 years. Significant temporal changes in clinicopathological parameters were observed. During the study period, median age at surgery increased from 61.4 to 64.8 years and median preoperative PSA declined from 12.0 to 8.0 ng/ml. The proportion of men with pT2 PCa increased from 65% to 75% whereas the proportion of pT3 cancers decreased from 28% to 25%. The percentage of men with positive surgical margins decreased from 37% to 20%. During follow-up, 644 patients died, whereof 189 (29.3%) died from PCa. The cumulative incidence of PCSM and other-cause mortality after 15 years was 10.3% (95% CI 8.0-12.7) and 18.2% (95% CI 15.4-20.9), respectively. In a multivariate analysis, RP GS (P ≤ 0.001) and pT-category (P ≤ 0.001) were significantly associated with the risk of PCSM. Compared with GS ≤6, both GS +4 (HR 1.47), GS 4 + 3 (HR 2.32), GS 8 (HR 4.8) and GS 9 or 10 (HR 5.26) significantly increased the risk of PCa death. T3a PCa and T3b/T4 was also a significant predictor of PCSM with an increased risk of PCa death compared with pT2 of 2.24 and 4.5, respectively.

CONCLUSIONS

In a complete national cohort of men treated with RP during a 17-year period, we described the incidence of mortality after RP and predictors of PCSM. We demonstrated that RP GS and pT-category are the most significant predictors of PCa mortality. We found that an increasing proportion of men undergo RP for low-risk PCa suggesting that early detection of PCa is indeed undergoing in Denmark despite national recommendations. The Danish national results seem to concur with findings from international single- and multi-institutional reports.