Findings question validity of large PCa trial

February 14, 2019

According to the updated report from the Prostate Cancer Intervention versus Observation Trial (PIVOT), radical prostatectomy had no significant benefit over observation for reducing mortality among men with clinically localized disease.

San Francisco-According to the updated report from the Prostate Cancer Intervention versus Observation Trial (PIVOT), radical prostatectomy had no significant benefit over observation for reducing mortality among men with clinically localized disease.

The appropriateness of applying those findings to patients seen in daily practice is questionable, however, considering information from an analysis assessing the external validity of PIVOT, said Firas Abdollah, MD, at the 2018 AUA annual meeting in San Francisco.

To evaluate the generalizability of the PIVOT results, Dr. Abdollah and colleagues compared the characteristics of its patient population with those of prostate cancer patients within three nationwide databases-the Surveillance, Epidemiology, and End-Results (SEER) population-based registry, the National Cancer Database (NCDB) hospital based-registry, and the Prostate, Lung, Colorectal and Ovarian (PLCO) trial.

Relative to all three external populations, men in PIVOT were older and had more comorbidities, the authors found. The differences between cohorts in age and general health might explain why the all-cause mortality rate was also much higher in PIVOT compared with the other prostate cancer patient populations, said Dr. Abdollah, fellow in uro-oncology and robotic surgery, Henry Ford Health System, Detroit.

“Our findings point to a sampling bias in PIVOT and indicate that the men who enrolled in PIVOT are not representative of those we are seeing in clinical practice. Clinically localized prostate cancer has a protracted course, and perhaps the older, sicker men in PIVOT did not have a long enough life expectancy to benefit from treatment,” added Dr. Abdollah, working with Mani Menon, MD, and colleagues.

Men from the external databases were included for the comparisons if they had characteristics matching those used for the PIVOT inclusion criteria: age ≤75 years, T1-T2NxM0 prostate cancer, PSA <50 ng/mL, and life expectancy ≥10 years.

Read: Bone marker data support mCRPC combination Tx

PIVOT included 731 men enrolled between 1994 and 2002. The comparator groups were comprised of 2,847 men enrolled in the PLCO trial between 1993 and 2001, 60,089 men entered into SEER between 2000 and 2004, and 63,303 men registered in the NCDB between 2004 and 2005.

“The databases that we used capture a wide variety of clinical settings and are very representative of men seen in clinical practice,” Dr. Abdollah told Urology Times.

The mean age at diagnosis was 67 years for men in PIVOT, which was older than for the SEER, NCDB, and PLCO cohorts (61.3, 60.2, and 65.8 years, respectively), and the between-group difference was statistically significant comparing PIVOT with the SEER and NCDB populations. Charlson comorbidity index (CCI) was not reported in SEER, but the proportion of men with no comorbidities (CCI of 0) was significantly smaller in PIVOT than in the NCDB and PLCO databases (56% vs. 96.4% and 94%, respectively).

Next: Higher mortality rate in PIVOT cohortHigher mortality rate in PIVOT cohort

The overall mortality rates in the PIVOT, SEER, NCDB, and PLCO cohorts were 64%, 23%, 9%, and 8.1%, respectively. Dr. Abdollah noted that the median follow-up in PIVOT, which was 12.7 years, is similar to that for the SEER database (12.3 years), whereas the median follow-up for men in the PLCO trial and NCDB was only 7.5 and 9 years, respectively.

“The shorter follow-up for the latter two cohorts could only partly explain their lower overall mortality rates,” said Dr. Abdollah.

Also see: Latin-American PCa patients less likely to receive definitive treatment

 

“Interestingly, the overall mortality rate for a subgroup of men from SEER that we excluded from our comparative analysis because they had not been given a recommendation for surgery was 50%. Presumably, these individuals were not recommended definitive treatment because of pre-existing conditions, and yet their mortality rate was still lower than that reported in PIVOT.”

download issueDownload Issue : Vol. 47 No. 2