Radical prostatectomy in older men: survival outcomes in septuagenarians and octogenarians

Authors


Phillip M. Pierorazio, Department of Urology, Marburg 100, Johns Hopkins Hospital, 600 N. Wolfe Street, Baltimore, MD 21287-2101, USA.
e-mail: philpierorazio@jhmi.edu

Abstract

Study Type – Therapy (case series)
Level of Evidence 4

OBJECTIVE

To examine the survival outcomes of septuagenarians and octogenarians (aged ≥80 years) who underwent radical prostatectomy (RP) at our institution, as the US Preventive Services Task Force recently released recommendations that men aged ≥75 years should not be screened for prostate cancer.

PATIENTS AND METHODS

Our institutional RP database (1982–2008) was queried for men aged ≥70 years at the time of surgery to evaluate actuarial survival after RP; 386 aged 70–81 years (median 71) underwent RP. The median (range) follow-up was 6.5 (1–22) years. Clinicopathological characteristics and mortality data were evaluated; mortality data were gathered through Social Security Administration Death Index and causes of death were confirmed with the Center for Disease Control National Death Index information. Kaplan-Meier analysis was used to evaluate cause-specific survival.

RESULTS

Ten patients (2.6%) had clinical stage T1a-b, 213 (55.3%) had T1c, and 143 (37.1%) had T2 prostate cancer. The median (range) preoperative prostate-specific antigen (PSA) level was 6.2 (0.2–49.9) ng/mL, and the biopsy and pathological Gleason sum was 6 (3–9) and 7 (4–9), respectively. Causes of death included prostate cancer (17), other malignancies (14), cardiovascular causes (14), neurological disease (four), pneumonia (two) and accident (one). The prostate cancer-specific survival rate was 97.6%, 94.0% and 90.2% at 5, 10 and 15 years after RP, respectively; the respective cardiovascular survival rate was 99.5%, 97.6% and 92.5%, and the overall survival rate was 93.1%, 82.5% and 68.9%, respectively.

CONCLUSIONS

If appropriately selected, older men have excellent overall and prostate-cancer specific survival after RP. The benefits of surgery should be weighed against the increased risks of surgical and anaesthetic complications.

Abbreviations
RP

radical prostatectomy

USPTF

United States Preventive Task Force

CSS

cancer-specific survival

OS

overall survival

bRFS

biochemical recurrence-free survival

NCCN

National Comprehensive Cancer Network

SSA

Social Security Administration

INTRODUCTION

The United States Preventive Task Force (USPTF) recently released guidelines for prostate-cancer screening, recommending against PSA screening for men aged ≥75 years (or those with a <10-year life-expectancy) [1]. Citing ‘moderate certainty’, the USPTF concluded that minimal incremental benefit is gained from PSA screening, while subjecting these men to invasive treatments puts them at substantial risk of morbidity and harm. Controversy ensued, as the report on screening cited by the USPTF is inconclusive and does not specifically address the age threshold of 75 years [2].

Cancer-specific survival (CSS) rates in several studies indicate that men aged >70 years are at substantial risk of death from prostate cancer [3,4]. In addition, urinary incontinence, impotence, bowel dysfunction, emotional and cognitive deficits are the well-known side-effects of prostate cancer treatments, and older patients are unquestionably at greater risk of dysfunction as performance in these areas declines with age [5]. The combination of CSS, side-effects and complications are therefore important considerations when selecting patients for radical prostatectomy (RP), and highlighted by the USPTF recommendations as the ‘prevention of harm in older men’.

Also, the USPTF equated men aged ≥75 years with those with a life-expectancy of ≤10 years. What might be more insightful than a specific age threshold is a careful consideration of the life-expectancy and medical comorbidities of a particular patient as ‘not all 75-year-old men are created equal.’[6]. The National Comprehensive Cancer Network (NCCN) Guidelines for the Treatment of Localized Prostate Cancer recommend the use of Social Security Administration (SSA) tables for life-expectancy and an adjustment of 50% for those men in the best quartile of health, yielding life-expectancies of 15 years for these healthy men aged 75 years. A competing-risk analysis by Sweat et al.[7] indicated that men aged 70–74 years with no significant comorbidities (Charlson score 0) had equivalent prostate cancer mortality rates to men aged <70 years, while noncancer-related mortality was similar among all men aged >65 years. Adjusting for age and comorbidity in a propensity-scoring model, Tewari et al.[8] found that radiotherapy and RP gave better survival rates than active surveillance for prostate cancer. In a sub-analysis of each treatment they reported similar overall survival (OS) and CSS rates among men aged <60, 60–70 and >70 years, while within each treatment, OS and CSS were markedly worse for men with a Charlson score of >1 than in those with score of 0 or 1. However, the association between CSS and comorbidity is not readily apparent, as poor overall health might mask a patient’s ability to respond to and recover from treatment. Also, basic science and clinical studies indicate that mounting comorbidities might make a patient more susceptible to biochemical processes of inflammation, oxidative stress and subsequent carcinogenesis [9,10]. Finally, comorbidities might increase the risk of adverse events, including perioperative complications and functional outcomes [11,12].

Therefore the objective of the present study was to analyse survival outcomes in septuagenarians (70–79 years) and octogenarians (aged 80–89 years) who underwent RP at our institution. We recognize that a retrospective analysis of a surgical series carries specific limitations of selection bias, and that our conclusions might not apply specifically to recommendations on screening for prostate cancer. Our hypothesis was therefore refined to reflect these limitations, i.e. that selected patients aged ≥70 years benefit from surgical treatment for prostate cancer.

PATIENTS AND METHODS

The Johns Hopkins RP database (1982–2008) was queried for men age ≥70 years at the time of RP; 386 were identified. Clinical and pathological data were evaluated using appropriate comparative tests. The Kaplan-Meier method was used to estimate overall, cardiovascular and prostate cancer-specific mortality. Overall mortality was defined as death from all causes; cardiovascular mortality was defined as death from atherosclerotic disease, specifically myocardial infarction or cerebrovascular accident; or chronic heart failure. Prostate cancer-specific mortality was defined as death due to or as a direct result of treatment for prostate cancer. Mortality data were collected from the SSA Death Index and cause of death was confirmed by the Center for Disease Control National Death Index information. The database uses the International Classification of Diseases Code-9 based, Deyo-Charlson comorbidity index as a proxy for self-rated health status [13]. The RP database and queries of the SSA and Center for Disease Control Death Indices were approved by Institutional Review Boards.

RESULTS

The median age of the men was 71 years; most (55.3%) were diagnosed by an elevated PSA level with impalpable disease (T1c) and the median PSA level was 6.2 ng/mL. The median (range) biopsy and pathological Gleason sum was 6 (3–9) and 7 (4–9), respectively. Nearly half of the men (46.8%) had organ-confined disease at pathological analysis. Of 321 men with preoperative comorbidity information available, 258 (80.4%) were in ‘excellent’ health (Charlson score 0), 58 (18.1%) were considered in ‘fair’ health (Charlson score 1 or 2) and five (1.6%) were considered in ‘poor’ health (Charlson score 3) before RP. The clinical and pathological data are detailed in Table 1.

Table 1. 
Clinical and pathological data, and the cause of death
VariableMedian (range) or n (%)
Age, years 71 (70–81)
Clinical stage 
 T1a-b 10 (2.6)
 T1c213 (55.3)
 T2143 (37.1)
 ≥T3  4 (1.0)
PSA level, ng/mL  6.2 (0.2–49.9)
Biopsy Gleason sum 
 3–6240 (62.3)
 7109 (28.3)
 8–9 27 (7.0)
Pathological Gleason sum 
 3–6171 (44.4)
 7160 (41.6)
 8–9 51 (13.2)
Pathological stage 
 Organ-confined180 (46.8)
 Extraprostatic extension163 (42.3)
 Seminal vesicle involvement 30 (7.8)
 lymph node positive 13 (3.4)
Cause of death (56 of 386 men, who died over 20 years) 
 Prostate cancer 17 (30.4)
 Cardiovascular 14 (25.0)
 Malignancy (not prostate) 14 (25.0)
 Unknown  4 (7.1)
 Neurological disease  4 (7.1)
 Infectious/pneumonia  2 (3.6)
 Accident/trauma  1 (1.8)

In all, 56 patients (14.5%) died over a median (range) follow-up of 6.5(1–22) years. The first three causes of death were prostate cancer (17), cardiovascular (14) and malignancy other than prostate cancer (14); causes of death are also detailed in Table 1. Four patients (1%) had a complication; one each had an nonfatal pulmonary embolus, thrombophlebitis, excessive blood loss and prolonged drainage of the vesico-urethral anastomosis. No patient had a serious perioperative or anaesthesia-related complication; no patients died immediately after surgery.

The biochemical recurrence-free survival (bRFS) rate was 73.3% at 5 years and 59.7% at 10 years. The prostate CSS rate was 97.6%, 94.0% and 90.2% at 5, 10 and 15 years after RP, respectively. The respective cardiovascular survival rate was 99.5%, 97.6% and 92.5%, and the OS rate was 93.1%, 82.5% and 68.9%; the Kaplan-Meier survival curves are shown in Fig. 1. For men in excellent, fair and poor health (based on Charlson score), the CSS was 92.7%, 98.0% and 100% at 5 years (P= 0.32) and the OS was 81.2%, 88.1% and 66.7%, respectively (P= 0.9). The multivariable analysis is detailed in Table 2; notably, when controlling for PSA level, pathological Gleason sum and pathological stage, comorbidity status was not a significant predictor of CSS or OS.

Figure 1.

OS, prostate CSS, cardiovascular and bRFS after RP in men aged ≥70 years.

Table 2.  Multivariable analysis of CSS and OS
VariableHazard ratio (95% CI), P
CSSOS
PSA0.96 (0.9–1.02), 0.21.00 (0.96–1.03), 0.8
Pathological Gleason sum  
 2–6 (ref)
 72.9 (0.56–14.9), 0.21.33 (0.68–2.58), 0.4
 8–107.7 (1.29–45.5), 0.032.05 (0.81–5.2), 0.1
Pathological stage  
 Organ-confined (ref)
 Extraprostatic extension2.5 (0.47–12.8), 0.30.91 (0.47–1.78), 0.8
 Seminal vesicle invasion2.6 (0.33–20.6), 0.41.18 (0.45–3.1), 0.7
 Lymph node invasion9.2 (1.37–61.6), 0.021.74 (0.56–5.46), 0.3
Charlson score  
 0 (ref)
 1–20.32 (0.04–2.54), 0.31.02 (0.5–2.08), 0.96
 31.00 (0.13–7.6), 0.99

DISCUSSION

Most men who underwent RP in our series had excellent preoperative prostate cancer characteristics and overall health. Within the overall cohort, the group of aged men in this study was a minority of the patients operated upon at our institution for prostate cancer, and an even smaller proportion of the general population. However, this study showed men aged >70 years and with favourable preoperative characteristics have long survival after RP. These men are in the ‘best’ quartile of health for their age, i.e. those who could expect a 50% increase in survival based on the NCCN guidelines for the treatment of prostate cancer, and this is reflected in the 70% OS rate in this cohort at 15 years. We therefore argue that surgery be considered a reasonable option in ‘healthy’ men aged ≥70 years, after careful consideration of potential complications, as they exceed the expectations for survival in the general population.

Health status, as defined by the Deyo-Charlson comorbidity index, did not influence CSS or OS. This might reflect the relatively few patients in this study with ‘fair’ and ‘poor’ health status. However, prostate cancer risk is reportedly modulated by markers of chronic inflammation and oxidative stress in several basic science and clinical studies, and therefore men with mounting comorbidites might be at higher risk of developing and dying from prostate cancer [9,10]. Anecdotally, the five patients with ‘poor’ health status had a biochemical recurrence within 5 years of RP, while ≈25% of the rest of the group had a PSA recurrence at the same time point. No prostate cancer-related deaths were reported in the five patients with ‘poor’ health, and although their OS was worse than the rest of the group it was not statistically significant, most probably reflecting the few deaths and in men with many comorbidities. The alternative interpretation is that highly selected and healthy aged men have a high likelihood of being cured of prostate cancer by RP.

In addition to being in the best health, the men in our series had the ‘best’ preoperative prostate cancer characteristics (Table 1). For instance, only a few (7%) had a biopsy Gleason sum of >7, 38% had palpable disease and half had a PSA level of <6 ng/mL, giving rise to organ-confined cancer on the pathological specimen in nearly half (47%) of the men, and the excellent bRFS and OS rates. This might reflect the inherent selection bias of a retrospective RP series, and the belief of the operating surgeon that these men will benefit from surgery. However, the algorithm to decide who will benefit from RP is more complicated than simply comparing life-expectancy independent of prostate cancer with the predicted loss of years due to prostate cancer based on preoperative and pathological prostate cancer characteristics. Life-expectancy is considered, along with the susceptibility to prostate cancer if untreated, and the likelihood and impact of side-effects of prostate cancer treatments. Choosing not to operate on a man with prostate cancer who desires treatment obliges that man to undergo radiotherapy and/or hormone deprivation therapy. Like surgery, radiotherapy might exacerbate already poor urinary or bowel continence, and has been shown to worsen irritative voiding symptoms, while hormone therapy can increase the likelihood of cardiovascular events, especially in older men [5,14–16]. In addition, our low complication rate (1%) highlights that in experienced hands we are subjecting healthy men to a minimal risk of morbidity and mortality from surgery; therefore the potential benefits might outweigh the risks of surgery, although urinary incontinence and impotence outcomes should be considered.

When examining the cause of death after RP, most causes were divided, relatively equally, into three general categories; death from prostate cancer, death from a second malignancy or death from a cardiovascular cause. As survival is examined over time, the Kaplan-Meier curves for cardiovascular and prostate CSS decay in a similar pattern (Fig. 1). The OS (which includes both cardiovascular and prostate CSS) decreases by ≈10% every 5 years after RP until it crosses 70% at 15 years. Therefore, more than half of the healthy men aged ≥70 years who undergo RP can expect to live 15 years and surpass the life-expectancy of 85 years predicted by the SSA. Also, of the men who die within this period, about a third can expect to die from prostate cancer or complications thereof, while an equal proportion will die from a (most likely age-related) secondary malignancy or cardiovascular insult. While the exact number of men ‘saved’ by RP remains unknown in this cohort, death from prostate cancer is significant in this group of men and would most probably be greater given no intervention. Conversely, the conclusion could be drawn that most men are likely to die from a secondary malignancy or cardiovascular cause due to advanced age, emphasising the importance of a careful evaluation of the patients before the treatment recommendation.

The present study has several important limitations. First, this group of men represents a minority of our institutional experience. They were carefully selected while considering both general health and prostate cancer-specific variables. Therefore these data cannot be extrapolated to the general aged population, nor applied to a larger screening population. The recent update of the Scandinavian Prostate Cancer Group study showed a significant reduction in metastases, prostate cancer-specific and overall mortality for men aged <65 years, and not for older men [17]. The contrasting themes from that study and the present highlight the distinction between a screened population and a selected surgical cohort. However, these data and conclusions are still valid; selected older men will fare well after RP. Further studies of this aged population would be strengthened by the use of a comparison group, either of similar, younger men who had RP, or age-matched controls with prostate cancer who were treated with surveillance, primary radiotherapy or hormone treatment. Finally, this study captured only the major comorbidities of RP and did not assess potency or continence data. Quality of life in this cohort is an important consideration when assessing the risks and benefits of surgery in this aged cohort.

In conclusion, men aged >70 years, when appropriately selected for RP, have an excellent OS and CSS of 70% and 90% at 15 years, respectively. However, the benefits of surgery should be weighed against the increased risks of surgical and anaesthetic complications in this population.

ACKNOWLEDGEMENTS

This study was supported by SPORE grant P50CA58236 from the National Institutes of Health and the National Cancer Institute, American Geriatrics Society (MH).

CONFLICT OF INTEREST

None declared.

EDITORIAL COMMENT

The issue with RP in older men is not whether they can tolerate the surgery without excessive perioperative morbidity, but whether they derive any survival benefit from the operation. Undoubtedly, there are some men aged >70 years who have a clinically significant but apparently localized cancer. Those with high-grade tumours might be well served by surgical intervention. However, most of the patients in this series had stage T1C, Gleason grade 6 or 7 cancers. Functional outcomes are not reported but erectile dysfunction and, to some degree incontinence, are more common in older men who undergo RP. Most men aged >70 years whose cancers would be categorized as good risk or possibly intermediate risk, do not derive a survival benefit from surgery and should not undergo RP, even though they can survive the operation with relatively minimal perioperative morbidity. For those with high-grade but apparently localized tumours, surgery might have an important role.

Joseph A. Smith, Jr,

Professor and Chairman, Department of Urologic Surgery, Vanderbilt University, Nashville, TN, USA

Ancillary