Login to Access Video or Poster Abstract: MP57-05
Sources of Funding: none

Introduction

Debilitating pain in men with castration-resistant prostate cancer (CRPC) is one of the most common, yet understudied morbidities. There is evidence that baseline pain is an important predictor of survival outcome in men with CRPC. Understanding the dynamic nature of pain, we seek to develop a prognostic model of overall survival (OS) based on the longitudinal pain score trajectories in men with CRPC.

Methods

729 patients with metastatic CRPC from the control arms of three phase III, randomized, double-blind, placebo-controlled clinical trials (NCT00519285, NCT00988208, and NCT00638690) were accessed from the Prostate Data Sphere and were analyzed. Joint latent class modeling (JLCM) method was applied to stratify our study cohort into different subpopulations with three different pain score trajectories. Patient reported brief pain inventory (BPI) or the present pain intensity index (PPI) instruments were obtained to characterize longitudinal pain scores. The OS was measured as the time-to-event outcome from randomization to death or censoring.

Results

536 (73.5%) patients were found to have a stable pattern of pain scores. 111 (15.2%) and 82 (11.3%) patients had rapidly increasing and decreasing trends, respectively. Patients with stable pain levels had the most favorable OS while rapidly increasing pain was associated with the worst OS (HR: 11.36, p-value<0.01 when compared to those with stable pain).

Conclusions

Patients having both increasing and decreasing patterns of pain had higher risk of death when compared to those with stable pain levels. Longitudinal pain scores can be used as a prognostic factor for overall survival in patients suffering from metastatic CRPC.

Funding

none