Generalizability of the Prostate Cancer Intervention Versus Observation Trial (PIVOT) Results to Contemporary North American Men with Prostate Cancer

Abstract

The Prostate Cancer Intervention Versus Observation Trial (PIVOT) concluded that radical prostatectomy (RP) offered no survival benefit compared with observation in men with clinically localized prostate cancer (PCa). We identified patients within the National Cancer Database (NCDB) for the period 2004-2012 who met the inclusion criteria of PIVOT (ie, histologically confirmed PCa, clinical stage T1–2NxM0, prostate-specific antigen <50 ng/ml, age <75 yr, estimated life expectancy >10 yr, and undergoing RP or observation as initial treatment within 12 mo of diagnosis) to confirm the generalizability of the PIVOT results to the US population. Life expectancy was calculated using the US Social Security Administration life tables and was adjusted for comorbidities at diagnosis. Compared with PIVOT, men in the NCDB were younger (mean age 60.3 vs 67.0 yr) and healthier (Charlson-Deyo comorbidity index of 0: 93% vs 56%; both p < 0.001). Furthermore, 42% of men randomized to receive RP in PIVOT harbored D'Amico low-risk PCa, whereas 32% of men undergoing RP in the NCDB had low-risk disease. Our findings were confirmed in a sensitivity analysis including men regardless of life expectancy but satisfying all other inclusion criteria of PIVOT. Given that the NCDB represents nearly 70% of all incident cancers diagnosed in the United States, our data provide further evidence that PIVOT results may not be generalizable to contemporary clinical practice. Patient summary We observed that men diagnosed with clinically localized prostate cancer within the National Cancer Database (2004–2012) were younger, healthier, and more likely to have radical prostatectomy for higher risk disease than men in the Prostate Cancer Intervention Versus Observation Trial (PIVOT), raising questions about the applicability of PIVOT conclusions to the contemporary US population.