Abstract: PD03-11
Sources of Funding: None; PcBaSe Sweden is funded by the Swedish Research Council (25-2012-5047).

Introduction

Comorbidities are medical disorders co-existing in patients with prostate cancer (PCa). Comorbidity, like PCa, is age-related and prevalent, influencing treatment choices. Although comorbidity may adversely affect competing-cause mortality in PCa patients, the impact on PCa-specific mortality is not known.

Methods

Using the PcBaSe Sweden composite population-based dataset, 118543 men diagnosed with PCa between 1998 and 2012, and followed up for survival until 2014 were identified. Median follow-up was 8.3 (IQR=5.2-11.5) years to death from PCa or other causes. Patients were categorised by patient (marital status, educational level) and tumour (serum Prostate-specific Antigen [PSA], tumour grade and clinical stage) characteristics, and treatment type (Radical Prostatectomy [RP], Radical Radiotherapy [RT], Androgen Deprivation Therapy [ADT] or Watchful Waiting [WW]). Data were stratified by Charlson Comorbidity Index (CCI) (0, 1, 2, >/=3) and treatment type. Mortality from PCa and other causes were calculated following stabilized inverse probability weighting (SIPW) adjustments for patient and tumour characteristics, and treatment type. Kaplan-Meier estimates and Cox regression were used to calculate hazard ratios. Ethical approval (EPN DnR 2012/499-31/4).

Results

In the complete unadjusted dataset, an effect of increasing comorbidity was observed on PCa-specific and other-cause mortality. Following adjustments for patient and tumour characteristics, the effect of comorbidity on PCa-specific mortality was lost, while maintained for other-cause mortality. Following additional adjustment for treatment type, an effect of comorbidity on PCa-specific mortality was not observed, while present for other-cause mortality.

Conclusions

During the study period in Sweden, comorbidity appeared to affect other-cause, but not PCa-specific, mortality after accounting for patient and tumour characteristics, and treatment type. Increasing comorbidity did not impact on PCa-specific mortality irrespective of radical treatment (RP or RT). Consequently, in population-based comparative PCa treatment effectiveness studies, the differences in oncological outcomes may not be due to the varying distribution of comorbidity among treatment groups, hence comorbidity is unlikely to be a confounder.

Funding

None; PcBaSe Sweden is funded by the Swedish Research Council (25-2012-5047).