Among men with clinically localized prostate cancer that had been diagnosed after PSA testing came into practice, our study showed that radical prostatectomy did not reduce all-cause or prostate-cancer mortality, as compared with observation, through at least 12 years of follow-up. Confidence intervals for the effect size indicated that surgery did not reduce all-cause mortality by more than 10% and might have increased mortality by as much as 4%. Differences in all-cause mortality decreased over time, suggesting that longer follow-up would not alter these findings. Only 10% of patients were younger than 60 years of age. Longer follow-up may be important for the minority of men with prostate cancer who were younger than 60 years of age. However, the nonsignificant between-group difference in prostate-cancer mortality and the significant 6% reduction in bone metastases with radical prostatectomy remained fairly constant after 8 years. Our findings add to evidence supporting observation, and possibly active surveillance, for most men who receive a diagnosis of localized prostate cancer, especially those with a low PSA value or low-risk disease.2,3,6,8,15–24
Death due to prostate cancer or treatment occurred infrequently, in 7.1% of patients. Any differences in prostate-cancer mortality between surgery and observation occurred primarily among men whose death was judged as probably due to prostate cancer or treatment. Among men whose death was considered to be definitely due to prostate cancer or treatment, we found almost no difference between surgery and observation. Between-group differences in the time of death due to prostate cancer or treatment did not explain the differences in all-cause mortality in the entire cohort or the subgroups. These findings highlight the limitations of using prostate-cancer mortality as an outcome, even with the use of adjudication committees whose members are unaware of treatment assignments and who are following standardized protocols.25–27
The effect of radical prostatectomy on mortality did not vary according to age, race, self-reported performance status, or coexisting conditions, but our findings suggest that it may vary according to PSA value and possibly tumor risk. Positive results were from multiple subgroup comparisons; the tests of interaction typically approached but did not reach significance and may therefore be due to chance. Among men with PSA levels of 10 ng per milliliter or less, all-cause mortality was slightly lower at 12 years in the observation group than in the radical-prostatectomy group; prostate-cancer mortality in the observation group was 6%, with a nonsignificant absolute reduction of less than 1.0 percentage point in the radical-prostatectomy group. Among men with low-risk disease, observation was associated with a non-significant reduction in all-cause and prostate cancer mortality, with no significant between-group difference in bone metastases. Among men with a PSA value that was greater than 10 ng per milliliter and possibly among those with intermediate-risk or high-risk prostate cancer (as determined according to the PSA value, local histologic findings, and stage), absolute reductions in all-cause mortality with radical prostatectomy ranged from 6.7 to 13.2 percentage points. Reductions were smaller and not significant when central histopathological findings were used, and we found no significant reductions with radical prostatectomy in categories that were derived solely on the basis of higher scores on the Gleason histologic scale or tumor stage. Reductions in prostate-cancer mortality in the radical-prostatectomy group were limited to men with a PSA value that was greater than 10 ng per milliliter and to those with high-risk disease, with absolute reductions of 7.2 to 8.4 percentage points. Absolute reductions in bone metastases of 10.4 and 8.6 percentage points occurred, respectively, in men with a PSA value of 10 ng per milliliter or higher and in those with high-risk disease.
As compared with the Scandinavian Prostate Cancer Group 4 (SPCG-4) trial of radical prostatectomy versus watchful waiting in men with prostate cancer detected before widespread PSA testing,6
PIVOT enrolled a higher percentage of men with nonpalpable tumors (stage T1c, 50% vs. 12%) and with PSA values of 10 ng per milliliter or lower. Treatment adherence was similar in the two trials.6,28
In contrast to the SPCG-4 trial, we did not find a significant reduction in all-cause or prostate-cancer mortality with radical prostatectomy. Our findings are particularly robust among men with a PSA value of 10 ng per milliliter or less, including men with a score of 7 or higher on the Gleason histologic scale, and low-risk tumor — categories that were underrepresented in the SPCG-4 trial. Unlike the SPCG-4 trial, our study did not show that the effect of surgery, as compared with observation, varied according to age. Although hazard ratios indicated that the relative effect of radical prostatectomy on prostate-cancer mortality was similar in PIVOT and the SPCG-4 trial (37% and 38% reduction, respectively), the relative reduction in all-cause mortality in our study was less than half that in the SPCG-4 trial (12% vs. 25%), as were the absolute reductions in all-cause mortality (2.9 percentage points vs. 6.6 percentage points) and prostate-cancer mortality (2.6 percentage points vs. 6.1 percentage points); the overall percentage of men who died from prostate cancer was also lower in our study (7.1% vs. 19.6%). The mortality reductions in our study were not significant and probably reflect the more favorable prognosis for patients with tumors detected by means of PSA testing.
Our study has strengths that enhance the clinical applicability of the findings. The age, health status, PSA value, and tumor-risk characteristics of the men enrolled in this study were similar to those of both men who were eligible but declined to undergo randomization9
and men in the general population who have received a diagnosis of prostate cancer.1–3,8,29
Perioperative morbidity and mortality were similar to those previously reported.28,30
The percentage of men with positive surgical margins was similar to that in earlier studies and lower than that in the SPCG-4 trial.27
The tumor volumes and PSA values in our study population, although higher than in some contemporary series,31–35
are probably representative of those in the general population of men who received a diagnosis of prostate cancer at the time the study was being conducted. Our choice of all-cause mortality as the primary outcome underscores the importance of improving life expectancy with cancer treatment and eliminates the possibility of biased cause-of-death ascertainment.25–27
Our study was conducted in the early era of PSA testing. The current practices of performing repeated PSA testing, using a lower PSA threshold for biopsy, obtaining more tissue-biopsy cores, and performing a repeat biopsy after initially negative findings increase the detection of smaller-volume indolent cancers.15,16
Along with systematically higher assignment of tumor grades (upgrading), these factors increase the likelihood of overdiagnosis and overtreatment.36–38
Among men with a current diagnosis of prostate cancer who undergo radical prostatectomy, the absolute reductions in the risks of metastasis and death will probably be smaller, and the time required to identify a reduction will probably be longer than reported in our study or in the SPCG-4 trial.
Our findings support observation for men with localized prostate cancer, especially those who have a low PSA value and those who have low-risk disease. Up to two thirds of men who have received a diagnosis of prostate cancer have a low PSA value or low-risk disease, but nearly 90% receive early intervention — typically surgery or ra-diotherapy.1,15,16,24
In contrast to observation, active surveillance initiates therapy with curative intent if disease progression is suspected on the basis of repeat PSA testing, digital rectal examinations, and prostate biopsies.3,24
Active surveillance is being compared with surgery or radiotherapy in a randomized trial.39
Informing men of the favorable long-term effects of observation on mortality, bone metastases, urinary and erectile function, and quality of life40–42
and increasing the use of observation may avert the harms of unnecessary biopsies43
while maintaining excellent long-term disease-specific survival.
In conclusion, our study showed that, as compared with observation, radical prostatectomy did not significantly reduce all-cause or prostate-cancer mortality through at least 12 years among men with clinically localized prostate cancer that had been diagnosed in the era of PSA testing. Absolute differences in mortality between the study groups were less than 3 percentage points. Subgroup analyses suggested that surgery might reduce mortality among men with higher PSA values and possibly among men with higher-risk tumors, but not among men with PSA levels of 10 ng per milliliter or less or among men with low-risk tumors.