MP21-08: OTHER CAUSE MORTALITY IS HIGHER IN HIGH RISK VERSUS LOW AND INTERMEDIATE RISK PROSTATE CANCERS (APL - 2018)

OTHER CAUSE MORTALITY IS HIGHER IN HIGH RISK VERSUS LOW AND INTERMEDIATE RISK PROSTATE CANCERS

View Poster

INTRODUCTION

Prostate cancer (PCa) risk groups and competing risks of mortality are critical components to patient counseling and informed decision-making. Competing risks may include age, demographic risk factors, comorbid health conditions, and treatment related health effects. While PCa-specific survival and overall survival are often reported for PCa risk groups, other cause mortality is not well described.

METHODS

Using SEER 18 Registries Research Data for the years 2004 to 2014, patients with localized PCa were categorized using the D&[prime]Amico classification. Mortality rates per 1000 person years and adjusted rate ratios (RR) were calculated using low risk PCa as referent, controlling for age, race, year of PCa diagnosis and first mode of PCa treatment. Multivariable Cox proportional hazards models assessed the association between D&[prime]Amico classification and overall, PCa-specific and other cause mortality over time.

RESULTS

473,709 localized PCa patients were identified. Overall mortality was 13.2% including PCa-specific mortality of 2.1%. PCa-specific mortality was strongly associated with high risk PCa (RR=12.21 [95% CI 11.09-13.45], p < 0.0001) and intermediate risk PCa (RR=2.73 [95% CI 2.47-3.02], p < 0.0001) compared to low risk PCa. Compared to low risk patients, other cause mortality was higher for high risk (RR=1.71 [95% CI 1.66-1.76], p < 0.0001) and intermediate risk (RR=1.37 [95% CI 1.34-1.41], p < 0.0001) patients. Adjusted PCa-specific and other cause survival curves are significantly different for each risk group and first mode of PCa treatment (Figure).

CONCLUSION

For all patients with localized PCa, the risk of dying of other causes is greater than the risk of dying of PCa, regardless of risk group or the first mode of treatment. The correlation between higher risk group and higher other cause mortality demonstrated in this study is not well understood. This could be the result of later diagnosis due to confounding variables or treatment-related effects. Further research into our findings is warranted, as this is the main driver of overall mortality.

Funding: none