Assessing the External Validity of the Updated Prostate Cancer (PCa) Intervention Versus Observation Trial (PIVOT)
Abdollah F, Dalela D, Akshay S, Alanee S, Keeley J, Peabody J, Rogers C, and Mani M. Assessing the external validity of the updated prostate cancer (PCa) intervention versus observation trial (PIVOT). Eur Urol, Supplements 2018; 17(2):e238-e239.
Introduction & Objectives: The PIVOT study has been instrumental in shaping health policy and public discourse about treatment of PCa in the US. However, concerns have been raised about the external validity of this trial to men diagnosed with PCa in the US. We assessed the study's generalizability (external validity) to men who were diagnosed with PCa using three nationally representative databases. Materials & Methods: We used data of men diagnosed with PCa within the Surveillance, Epidemiology, and End-Results (SEER; population based registry) 2000-2004, National Cancer Database (NCDB; hospital based registry) 2004-2005, and Prostate, Lung, Colorectal and Ovarian (PLCO) trial between 1993-2001. We included only men undergoing surgery or observation who satisfied the inclusion criteria of PIVOT. Within SEER, men were excluded from the comparative analyses if surgery was not recommended (because of pre-existing conditions or otherwise): These men were separately assessed for age at diagnosis and overall mortality. Chi-square test and ANOVA test were used to compare proportions and means, respectively. Results: A total of 2847, 60089, and 63303 PCa patients were included from PLCO, SEER, and NCDB respectively. Mean age (standard deviation) at diagnosis of clinically localized PCa for men in PIVOT was 67.0 (5.2) years, compared to 65.8 (4.5), 61.3 (7.3) and 60.2 (7.0) in PLCO, SEER and NCDB respectively (p<.001). The majority of men (94-96.4%) in PLCO and NCDB had Charlson comorbidity index (CCI) of 0, compared to only 56% for men in PIVOT (p<.001). No comorbidity data were available in SEER. Overall mortality was 64% in PIVOT (over a median 12.7-year follow-up) vs. 8-23% in the three examined cohorts (the latter over a median 12.3-year follow-up). Interestingly, within SEER, the mean age at diagnosis for men who were not recommended surgery was 65.1 years, with overall mortality 50% over a median follow-up of 12.5 years. Conclusions: These findings across multiple large datasets consistently demonstrate that men within PIVOT were older and sicker than their counterparts across the US. Interestingly, the reported overall survival in PIVOT was worse even when compared to men within SEER who were not recommended definitive treatment. These findings suggest a sampling bias, and highlight the lack of generalizability, and thus external validity, of PIVOT findings to men diagnosed with clinically localized PCa in the US. (Table presented).