CLINICAL INVESTIGATION Prostate

PRIMARY CAUSES OF DEATH AFTER PERMANENT PROSTATE BRACHYTHERAPY

NATHAN BITTNER, M.D., M.S.,y GREGORY S. MERRICK, M.D.,* ROBERT W. GALBREATH, PH.D.,*

WAYNE M. BUTLER, PH.D.,* KENT E. WALLNER, M.D.,yz ZACHARIAH A. ALLEN, M.S.,*

SARAH G. BRAMMER, B.S.,* AND MARK MOYAD, M.D., M.P.H.x

*Schiffler Cancer Center Wheeling Jesuit University, Wheeling, WV; yDepartment of Radiation Oncology, University of Washington,Seattle, WA; zRadiation Oncology, Puget Sound Healthcare Corporation, Department of Veterans Affairs, Seattle, WA; and

xDepartment of Urology, University of Michigan Medical Center, Ann Arbor, MI

Purpose: To evaluate the primary causes of death in low-risk (low-risk), intermediate-risk (intermediate-risk), andhigh-risk (high-risk) patients undergoing permanent prostate brachytherapy with or without supplemental ther-apies.Methods and Materials: From April 1995 through November 2004, a total of 1,354 consecutive patients underwentprostate brachytherapy. All patients underwent brachytherapy >3 years before analysis. Of the patients, 532(39.3%) received androgen deprivation therapy and 703 (51.9%) received supplemental radiation therapy. Themedian follow-up was 5.4 years. Multiple parameters were evaluated as predictors of cause-specific, biochemicalprogression–free, and overall survival.Results: The 10-year cause-specific survival was 97.0% (99.7%, 99.0%, and 90.1% for low-risk, intermediate-risk,and high-risk patients). Overall survival was 76.7% (82.5%, 78.3%, and 67.6% for low-, intermediate-, and high-risk patients, respectively). The cumulative death rate for cardiovascular disease was 11.5% (8.7%, 9.3%, and19.8% for low-, intermediate-, and high-risk patients). The death rate from second malignancies (nonprostate can-cer) was 7.2% and was not substantially different when stratified by risk group. Death from all other causes was6.5% for the entire cohort but 1.3%, 5.0%, and 10.8% for low-, intermediate-, and high-risk patients. In multivar-iate analysis, death from prostate cancer was best predicted by Gleason score and risk group, whereas death fromcardiovascular disease, nonprostate cancer, and all other causes were most closely related to patient age andtobacco use.Conclusions: Although cardiovascular mortality was the predominant cause of death, prostate cancer was respon-sible for approximately 10% of all deaths. In particular, overall survival was poorest in the high-risk group. Al-though high-risk patients were most likely to die of prostate cancer, the divergence in overall survival betweenhigh-risk and lower-risk patients primarily resulted from an excess of cardiovascular deaths. Changes in lifestyleto improve cardiovascular health may improve overall survival in patients with clinically localized prostatecancer. � 2008 Elsevier Inc.

Prostate cancer, Primary causes of death, Brachytherapy, ADT.

Int. J. Radiation Oncology Biol. Phys., Vol. 72, No. 2, pp. 433–440, 2008Copyright � 2008 Elsevier Inc.

Printed in the USA. All rights reserved0360-3016/08/$–see front matter

doi:10.1016/j.ijrobp.2008.02.013

INTRODUCTION

The widespread use of prostate-specific antigen (PSA)

screening has led to earlier detection of prostate cancer and

increased lead time to diagnosis (1–3). In some studies, there

is a suggestion that this early detection has translated into

a fall in the incidence of metastatic disease and a reduction

in prostate cancer–related mortality (4). Of equal importance

are the strides that have been made in prostate cancer treat-

ment. With prostate brachytherapy, in particular, refinement

of implant dosimetry and treatment margins have led to im-

provements in biochemical control (5–7). Together, effective

Note—An online CME test for this article can be taken at http://asro.astro.org under Continuing Education.

Reprint requests to: Gregory S. Merrick, M.D., Schiffler CancerCenter, Wheeling Hospital, 1 Medical Park, Wheeling, WV26003-6300. Tel: (304) 243-3490; Fax: (304) 243-5047; E-mail:

43

screening methods and advances in therapeutic management

have improved cure rates in men with localized prostate

cancer.

With improving prostate cancer cure rates, one would

expect that this population of patients might become increas-

ingly susceptible to death from other causes. There are

already some data to suggest that high-risk patients managed

with brachytherapy and supplemental external beam radia-

tion therapy are more likely to die of cardiovascular disease

and other malignancies than from prostate cancer itself (8).

This likely relates in part to the high prevalence of comorbid-

ities in the prostate cancer population. Some epidemiologic

gmerrick@urologicresearchinstitute.orgConflict of interest: none.Received Nov 30, 2007, and in revised form Jan 29, 2008.

Accepted for publication Feb 8, 2008.

3

434 I. J. Radiation Oncology d Biology d Physics Volume 72, Number 2, 2008

studies estimate a crude prevalence of serious comorbidity

>50% among this patient population (9).

Even before the advent of PSA screening, there were data

suggesting that coexisting disease was an important cause of

death in prostate cancer patients (10). One early study from

the Berkley School of Public Health showed that only 54%

of decedent prostate carcinoma patients died of prostate can-

cer. In the PSA era, large population-based studies continue

to suggest that comorbidity has a decisive influence on the

risk of early death among patients with a diagnosis of prostate

cancer (11). Unfortunately such population studies are often

difficult to interpret, given the broad range of disease stage

and diversity of treatment. There are few published studies

specifically examining the patterns of death among prostate

cancer patients managed with any particular treatment mo-

dality.

Given the paucity of published data, we report the primary

causes of death among low-risk, intermediate-risk, and high-

risk patients treated definitively with prostate brachytherapy

for localized adenocarcinoma of the prostate.

METHODS AND MATERIALS

Between April 1995 and November 2004, a total of 1,354 consec-

utive patients underwent permanent interstitial brachytherapy for

clinical stage T1b to T3a (2002 AJCC) prostate cancer by a single

brachytherapist (G.S.M.). All patients underwent brachytherapy

more than 3 years before analysis. Before the formulation of the

treatment plan, all biopsy slides were reviewed by a single patholo-

gist. Preplanning technique, interoperative approach, and dosimetric

evaluation have been described in detail (12, 13). Patients were clin-

ically staged by medical history, physical examination including

digital rectal examination and serum PSA determinations. Bone

scans and computed tomography (CT) of the abdomen/pelvis

were obtained for patients with higher but not lower risk disease.

No patient underwent seminal vesicle pathologic lymph node stag-

ing. Tables 1 and 2 summarize the clinical treatment in dosimetric

parameters of the patient population stratified by risk group. Of

the 1,354 patients, 475 presented with low-risk disease (PSA #10

ng/ml, Gleason score #6, and clinical stage #T2a), 636 with inter-

mediate-risk disease (one adverse factor: PSA 10.1–19.9 ng/ml, or

a Gleason score of 7 or clinical stage T2b) and 243 presented with

high-risk disease (PSA $20 ng/ml or Gleason score $8 or clinical

stage $T2c or two to three of the intermediate adverse features).

Of the 1,354 patients, 532 (39.3%) received androgen deprivation

therapy (ADT). In all, 361 patients (26.7% of the total) received #6

months of ADT, whereas 171 (12.6%) received >6 months of ADT.

In patients receiving ADT, the ADT was initiated 3 months before

implantation and consisted of a lutinizing hormone-releasing hor-

mone agonist and an antiandrogen. The maximum duration of

ADT was 36 months.

Of the 1,354 patients, 703 (51.9%) received supplemental radia-

tion therapy. In general, radiation therapy consisted of 45 Gy deliv-

ered in 1.8-Gy fractions with 15- to 18-MV photons delivered via

a conformal technique that used four fields (opposed laterals and an-

terior–posterior and posterior–anterior) custom treatment devices, to

spare as much normal tissue as possible. For patients with #10%

risk of pelvic lymph node involvement (14), the target volume con-

sisted of the prostate gland and seminal vesicles. For patients with

>10% risk of pelvic lymph node involvement, the pelvic lymph no-

des were also included in the target volume. A total of 126 patients

received 20 Gy in 2-Gy fractions as part of a clinical trial (15). In all

cases, supplemental radiation therapy was delivered before brachy-

therapy.

The brachytherapy target volume consisted of the prostate gland

with periprostatic treatment margins including the proximal 1.0 cm

of the seminal vesicles (12, 13). The minimum peripheral dose was

prescribed to the target area with margin. Of the 1,354 patients,

1,079 (79.9%) underwent implantation with palladium (Pd)-103

and 275 (23.3%) with iodine-125. Standard implant doses were

used for monotherapy and boost regimens. At implantation, the

prostate gland, periprostatic region and base of the seminal vesicles

were implanted (12, 13).

Patients were monitored by physical examination including

digital rectal examination and serum PSA determinations at 3- to

6-month intervals. Endpoints of analysis include cause-specific

survival, biochemical progression–free survival (definition: PSA #

0.40 ng/ml after nadir), and overall survival (16). No patient

Table 1. Patient characteristics and treatment parameters stratified by risk group

Low risk (n = 475) Intermediate risk (n = 636) High risk (n = 243) Total (n = 1,354)

Variable Median Mean � SD Median Mean � SD Median Mean � SD p Median Mean � SD

Age at implant (y) 65.0 64.3 � 7.3 67.0 66.2 � 7.3 68.0 66.7 � 7.6 <0.001 66.0 65.6 � 7.4Follow-up (y) 5.2 5.8 � 2.6 5.2 5.5 � 2.4 6.2 6.2 � 2.8 0.001 5.4 5.7 � 2.6Pretreatment PSA 5.9 6.1 � 1.9 6.3 7.1 � 3.9 12.8 15.5 � 9.1 <0.001 6.7 8.3 � 5.9Gleason score 6.0 5.9 � 0.4 7.0 7.0 � 0.8 7.0 7.5 � 0.8 <0.001 7.0 6.7 � 0.9% Positive biopsy results 16.7 24.6 � 16.4 37.5 41.9 � 37.5 50.0 58.7 � 27.4 <0.001 33.3 38.7 � 25.4Body mass index 27.4 28.1 � 4.2 27.5 28.2 � 4.7 27.3 27.9 � 4.2 0.620 27.4 28.1 � 4.4Prostate volume (cm3) 35.8 36.4 � 8.7 33.6 34.1 � 9.2 31.1 31.6 � 10.3 <0.001 34.1 34.4 � 9.4Planning volume (cm3) 66.4 66.0 + 12.4 63.7 63.2 � 13.2 56.4 57.3 � 13.2 <0.001 63.5 63.1 � 13.2V100 (%vol) 97.2 95.4 � 5.5 97.9 96.3 � 5.0 97.7 95.6 � 6.2 0.015 97.7 95.9 � 5.4V150 (%vol) 65.8 62.2 � 15.0 71.6 68.2 � 14.7 71.7 66.7 � 15.6 <0.001 69.4 65.8 � 15.2V200 (%vol) 33.3 32.3 � 12.7 40.6 38.9 � 13.1 39.8 37.8 � 13.6 <0.001 38.0 36.4 � 13.4D90 (% Rx dose) 115.2 114.6 � 13.4 120.2 119.5 � 14.6 119.2 118.4 � 16.3 <0.001 118.1 117.6 � 14.7ADT* (mo) 4.0 4.1 � 1.3 4.0 6.7 � 6.0 12.0 15.1 � 10.4 <0.001 4.0 8.2 � 8.0Most recent PSAy <0.04 0.07 � 0.06 <0.04 0.05 � 0.11 <0.04 0.05 � 0.16 0.134 <0.04 0.06 � 0.14

Abbreviations: ADT = androgen deprivation therapy; PSA = prostate specific antigen; Rx = prescribed.* Only those patients who had ADT were included in this analysis (n = 532).y Only those patients who have not experienced a biochemical failure (n = 1,304).

Table 2. Patient characteristics and treatment parameters stratified by risk group

CharacteristicLow risk

(n = 475), count (%)Intermediate risk

(n = 636), count (%)High risk

(n = 243), count (%) pTotal

(n = 1354), count (%)

StageT1b–T2b 475 (100) 629 (98.9) 159 (65.4) <0.001* 1263 (93.3)T2c–T3a 0 (0) 7 (1.1) 84 (34.6) 91 (6.7)

IsotopeI-125 196 (41.3) 61 (9.6) 18 (7.4) <0.001 275 (20.3)Pd-103 1079 (58.7) 575 (90.4) 225 (92.6) 1079 (79.7)

ADT (mo)0 304 (64.0) 421 (66.2) 97 (39.9) <0.001* 822 (60.7)#6 164 (34.5) 160 (25.2) 37 (15.2) 361 (26.7)>6 7 (1.5) 55 (8.6) 109 (44.9) 171 (12.6)

XRTNo 455 (95.8) 171 (36.9) 25 (10.3) <0.001 651 (48.1)Yes 20 (4.2) 465 (73.1) 218 (89.7) 703 (51.9)

HypertensionNo 264 (55.6) 321 (50.5) 117 (48.1) 0.117 702 (51.8)Yes 211 (44.4) 315 (49.5) 126 (51.9) 652 (48.2)

DiabetesNo 264 (55.6) 321 (50.5) 117 (48.1) 0.107 702 (51.8)Yes 211 (44.4) 315 (49.5) 126 (51.9) 652 (48.2)

Tobacco userNever 192 (40.4) 237 (37.3) 78 (32.1) 0.094 507 (37.4)Former 220 (46.3) 298 (46.9) 116 (47.7) 634 (46.8)Current 63 (13.3) 101 (15.9) 49 (20.2) 213 (15.7)

Statin useNone 358 (75.4) 477 (75.0) 194 (79.8) 0.191 1029 (76.0)Non-Lipitor 62 (13.1) 66 (10.4) 23 (9.5) 151 (11.2)Lipitor 55 (11.6) 93 (14.6) 26 (10.7) 174 (12.9)

Abbreviations: ADT = androgen deprivation therapy; XRT = radiation therapy.* Because of the small sample size, for data in some cells Fisher’s exact test was used.

Causes of death after prostate brachytherapy d N. BITTNER et al. 435

underwent postimplantation biopsy. Multiple clinical, treatment,

and dosimetric parameters were evaluated for impact on cause-spe-

cific, biochemical progression–free, and overall survival.

Cause of death was determined for each patient. Patients with

metastatic prostate cancer or hormone-refractory disease without

obvious metastases who died of any cause were classified as having

died of prostate cancer. All other deaths were attributed to the imme-

diate cause of death.

Clinical and treatment variables that were continuous were com-

pared among the three risk groups using a one-way analysis of var-

iance. Patterns of death across the three risk groups were compared

using Fisher’s exact test. Cox regressions were used to determine

univariate and multivariate predictors of death by cardiovascular

disease, prostate cancer, and any other cause. After stratifying pa-

tients by risk, Cox regression was also used to determine univariate

and multivariate predictors of cardiovascular and prostate cancer–

specific deaths. Cardiovascular and cancer-specific Kaplan-Meier

survival and log rank analysis were used to compare survival across

risk groups, isotope, age, tobacco use status, and Gleason score. For

all tests a value of p # 0.05 was deemed statistically significant. Sta-

tistical analysis was preformed with SPSS version 15.0 software

(SPSS Inc., Chicago, IL).

RESULTS

Patient characteristics and treatment-related parametersDisease characteristics and treatment-related parameters of

the evaluated patient population are summarized in Tables 1

and 2. The median age among all evaluable patients was 66

years. Low-risk (n = 475) and intermediate-risk (n = 636)

prostate cancer patients comprised the majority of the study

population; however there were a substantial minority of

high-risk (n = 243) patients. The median D90 (i.e., minimum

percentage of the dose that covered 90% of the target volume)

for all patients was 118.1%. A total of 651 patients (48.1%)

received supplemental external beam radiation therapy in

conjunction with brachytherapy implantation; this was

primarily accounted for by the intermediate- and high-risk

patients. In all, 465 intermediate-risk (73.1%) and 218 high-

risk (89.7%) patients received supplemental external beam

radiation therapy. Androgen deprivation therapy (ADT) was

administered to 532 patients (39.3%). Among those receiving

ADT, 361 received therapy for #6 months and 171 for >6

months.

Cause-specific, biochemical progression–free, and overallsurvival

With a median follow-up of 5.4 years, the 10-year cause-

specific, biochemical progression–free, and overall survivals

for all patients were 97.0%, 95.9%, and 76.7%, respectively

(Fig. 1). When patients were stratified according to prostate

cancer risk category, the 10-year cause-specific, biochemical

progression–free, and overall survivals for low-risk patients

436 I. J. Radiation Oncology d Biology d Physics Volume 72, Number 2, 2008

were 99.7%, 98.4%, and 82.5%, respectively (Fig. 2a). The

10-year cause-specific, biochemical progression–free, and

overall survivals for intermediate-risk patients were 99.0%,

96.8%, and 78.3%, respectively (Fig. 2b). Finally, the

cause-specific, biochemical progression–free, and overall

survivals for high-risk patients were 90.1%, 88.5%, and

67.6%, respectively (Fig. 2c).

Cause of deathIn analyzing mortality among all patients, cardiovascular

disease was the predominant cause of death, accounting for

41.6% of all deaths. Nonprostate cancer accounted for

30.4% of all deaths and prostate cancer only 8.7% (Table

3). The cumulative hazards of death from cardiovascular dis-

ease, nonprostate cancer, prostate cancer, and all other causes

were 11.5%, 7.2%, 3.0%, and 7.4%, respectively (Fig. 3).

Among low-risk patients, the cumulative hazards of death

from cardiovascular disease, nonprostate cancer, prostate

cancer, and all other causes were 8.7%, 5.4%, 0.2%, and

1.3%, respectively (Fig. 4). In intermediate-risk patients,

the cumulative hazards of death from cardiovascular disease,

nonprostate cancer, prostate cancer, and all other causes were

8.8%, 9.3%, 1.0%, and 5.4%, respectively (Fig. 5). In high-

risk patients, the cumulative hazards of death from cardiovas-

cular disease, nonprostate cancer, prostate cancer, and all

other causes were 19.8%, 6.0%, 10.3%, and 11.8%, respec-

tively (Fig. 6).

Factors predicting cause of deathPatient and treatment-related parameters were analyzed to

determine which of these variables were predictive of death

from cardiovascular disease. On univariate analysis, age

(RR = 1.142, p < 0.001), D90 (RR = 0.982, p = 0.024), and

tobacco use status (never vs. former use, RR = 2.525, p =

0.005; never vs. current use, RR = 2.312, p = 0.04) were pre-

dictive of death from cardiovascular disease. Only age (RR =

1.152, p < 0.001) and tobacco use status (never vs. former

121086420Years Since Implant

100

80

60

40

20

0

Su

rv

iv

al (%

)

Overall, 74.8%

Cause-Specific, 97.0%BPFS, 95.9%

Fig. 1. Kaplan-Meier survival curves for cause-specific, biochemi-cal progression–free, and overall survival.

use, RR = 2.285, p = 0.012; never vs. current, RR = 3.053,

p = 0.007) remained significant predictors on multivariate

analysis (Table 4).

121086420Years Since Implant

100

80

60

40

20

0

Su

rvival (%

)

Overall, 82.5%

Cause-Specific, 99.7%BPFS, 98.4%

a. low risk

121086420Years Since Implant

100

80

60

40

20

0

Su

rvival (%

)

Overall, 78.3%

Cause-Specific, 99.0%BPFS, 96.8%

b. intermediate risk

121086420Years Since Implant

100

80

60

40

20

0

Su

rvival (%

)

Overall, 61.9%

Cause-Specific, 90.1%BPFS, 88.5%

c. high risk

Fig. 2. Kaplan-Meier survival curves for cause-specific, biochemi-cal progression–free, and overall survival. (a) Low-risk, (b) interme-diate-risk, and (c) high-risk patients.

Table 3. Cause of death stratified by risk group

VariableLow risk,count (%)

Intermediate risk,count (%)

High risk,count (%) p*

Total,count (%)

Prostate cancer 1 (2.1) 3 (4.2) 10 (24.4) 14 (8.7)Cardiopulmonary disease 28 (58.3) 32 (44.4) 21 (51.2) 81 (50.3)Gastrointestinal cancers 5 (10.4) 10 (13.9) 5 (12.2) 0.011 20 (12.4)Lung cancers 7 (14.6) 7 (9.7) 2 (4.9) 16 (9.9)Other cancers 1 (2.1) 4 (5.6) 0 (0.0) 5 (3.1)Other 6 (12.5) 16 (22.2) 3 (7.3) 25 (15.5)

* Because of the small sample size, for data in some cells Fisher’s exact test was used.

Causes of death after prostate brachytherapy d N. BITTNER et al. 437

Patient and treatment-related parameters were also ana-

lyzed to determine which of these variables were predictive

of death from prostate cancer or from nonprostate cancer.

On univariate analysis, Gleason score (RR = 3.316, p <0.001), percentage of positive biopsy results (RR = 1.030,

p = 0.001), stage (RR = 4.705, p = 0.006), and risk category

(high vs. low, RR = 0.058, p = 0.007; high vs. intermediate,

RR = 0.145, p = 0.001) were predictive of death from prostate

cancer. Only Gleason score (RR = 2.786, p = 0.001) and risk

category (high vs. intermediate, RR = 0.193, p = 0.013) re-

mained significant predictors on multivariate analysis. On

univariate analysis, age (RR = 1.087, p < 0.001) and tobacco

use (never vs. former, RR = 2.620, p = 0.016; never vs. cur-

rent, RR = 3.461, p = 0.007) were predictive of death from

nonprostate cancer. Both age (RR = 1.093, p < 0.001) and to-

bacco use (never vs. former, RR = 2.445, p = 0.025; never vs.

current, RR = 4.093, p = 0.002) remained significant predic-

tors on multivariate analysis.

DISCUSSION

With early detection and use of ever-improving treatment

methods, the likelihood of curing localized prostate cancer is

very good. Long-term outcomes data with brachytherapy

demonstrate excellent biochemical control rates (17). There

is even some evidence to suggest that brachytherapy may of-

fer a therapeutic advantage over other competing treatment

modalities (18, 19). Given these unprecedented cure rates,

12108520Years Since Implant

12

10

8

6

4

2

0

Cu

mu

lative H

azard

(%

)

Diseases of the Heart, 11.5%

Non-prostate cancer, 7.2%

Prostate cancer, 3.0%

Other, 7.4%

Fig. 3. Cumulative hazard of death by select causes. Each curverepresents the same group of patients.

prostate cancer patients would seem to be increasingly sus-

ceptible to death from other causes, especially when one con-

siders that this population is known to have a high incidence

of comorbidities (9).

In our series of 1,354 consecutive prostate cancer patients

treated with transperineal brachytherapy, the 10-year cause-

specific survival was 97.0%. When patients were analyzed

according to risk group, even high-risk patients fared excep-

tionally well, with a cause-specific survival of 90.1% and

a biochemical control rate of 88.5%. Although the observed

cure rates are encouraging, it is nonetheless discardiovascu-

larening to confirm that cardiovascular disease was the pre-

dominant cause of death among the study population. In

fact, cardiovascular disease accounted for >41% of the ob-

served deaths. Even among high-risk patients, the cumulative

hazard of death from cardiovascular disease exceeded that of

prostate cancer. Of interest, among low- and intermediate-

risk patients, nonprostate cancer also accounted for more

deaths than prostate cancer.

Our findings underscore the imperative need for both ap-

propriate patient selection before definitive treatment and

proper clinical management after treatment. It has long

been argued that only patients with a life expectancy >10

years should undergo definitive treatment for prostate cancer

(20, 21). Comorbidity instruments such as the Charlson co-

morbidity index, Index of Co-existent Disease, Kaplan-Fein-

stein Index, and Cumulative Illness Rating Scale have been

validated as reliable predictors of mortality (22, 23). These

comorbidity scales should be used routinely by clinicians

Fig. 4. Cumulative hazard of death by select causes for low-risk pa-tients. Each curve represents the same group of patients.

to select those prostate cancer patients who are most likely to

benefit from aggressive treatment.

All of the patients included in our analysis were appropri-

ately screened and had an estimated life expectancy of $10

years at the time that they received definitive therapy. As

such, the trend in patterns of death is attributable not to an ail-

ing patient population but rather to a highly efficacious treat-

ment modality that has afforded exceptional cause-specific

survival. With the excellent prostate cancer cure rates ob-

served in our analysis, the patterns of death emulated that

of the general population, with cardiovascular disease being

the main cause of mortality.

121086420Years Since Implant

12

10

8

6

4

2

0

Cu

mu

lative H

azard

(%

)

Diseases of the Heart, 8.8%

Prostate Cancer, 1.0%

Non-Prostate Cancers, 9.3%

Other, 5.4%

Fig. 5. Cumulative hazard of death by select causes for intermedi-ate-risk patients. Each curve represents the same group of patients.

438 I. J. Radiation Oncology d Biology d Physics

These findings demand a reassessment of the way that

prostate cancer patients are managed after definitive treat-

ment. In response to this analysis, we have implemented pro-

grams to promote a healthy lifestyle among our patients. This

includes encouragement of a low-fat diet, medical manage-

ment of hypertension/hyperlipidemia, smoking cessation,

and enrollment in a cardiovascular fitness program. With

the relatively high prevalence of nonprostate cancer deaths,

we also continue to encourage age appropriate cancer screen-

ing such as colonoscopy. We believe that these measures will

ultimately improve overall survival among prostate cancer

patients, as they have in the general population (24–31).

121086420

Years Since Implant

20

15

10

5

0

Cu

mu

la

tiv

e H

aza

rd

(%

)

Prostate, 10.3%Other, 11.8%

Non-Prostate Cancer, 6.0%

Diseases of the Heart, 19.8%

Fig. 6. Cumulative hazard of death by select causes for high-risk pa-tients. Each curve represents the same group of patients.

Volume 72, Number 2, 2008

Table 4. Univariate and multivariate analysis examining predictors of death by various causes

Diseases of the Heart Prostate Cancer Non-Prostate Cancers Other

Univariate Multivariate Univariate Multivariate Univariate Multivariate Univariate Multivariate

Variables p RR p RR p RR p RR p RR p RR p RR p RR

Age < 0.001 1.142 < 0.001 1.152 0.528 <0.001 1.087 <0.001 1.093 0.059 0.031 1.062PSA 0.251 0.150 0.983 0.774Gleason Score 0.772 < 0.001 3.136 0.001 2.786 0.927 0.749% Positive

Biopsies0.389 0.001 1.030 0.443 0.718 0.155

Body Mass Index 0.215 0.407 0.057 0.291 0.442Prostate Volume 0.516 0.336 0.815 0.516V100 0.051 0.108 0.797 0.524 0.616V150 0.159 0.399 0.489 0.667V200 0.277 0.405 0.603 0.843%D90 0.024 0.982 0.162 0.641 0.899 0.201PNI 0.141 0.902 0.489 0.365Stagec 0.842 0.006 4.705 0.630 0.848 0.678Isotopec 0.055 0.014 0.533 0.138 0.367 0.694Riskc: 0.216 0.001 0.043 0.345 0.441High vs Low - 0.007 0.058 0.395 - -High vs

Intermediate- 0.001 0.145 0.013 0.193 - -

XRTc 0.581 0.024 0.301 0.596 0.114Hormonesc 0.150 0.318 0.356 0.912Hormone

Durationc0.081 -* 0.123 0.437 0.487

Hypertensionc 0.051 0.075 0.094 0.077 0.308 0.635Diabetesc 0.338 0.429 0.165 0.654Tobacco Usec: 0.017 0.014 0.236 0.019 0.009 0.010 0.004Never vs Former 0.005 2.525 0.012 2.285 - 0.016 2.620 0.025 2.445 0.101 - 0.119Never vs Current 0.040 2.312 0.007 3.053 - 0.007 3.461 0.002 4.093 0.003 4.948 0.002 5.517

Causes of death after prostate brachytherapy d N. BITTNER et al. 439

Although ADT was not identified as a significant predictor

for cardiovascular death in our study, there is growing evidence

to suggest that the use of ADT may be associated with an in-

creased risk of death from cardiovascular causes (32–34).

Use of ADT in conjunction with brachytherapy has not been

evaluated in the setting of a prospective clinical trial; however

the largest retrospective studies fail to show any improvement

in cause-specific survival or overall survival when ADT is

given in combination with brachytherapy (8, 34). Given the

lack of documented benefit and the potential risk of increased

cardiovascular mortality, we currently recommend minimiza-

tion of ADT in our prostate brachytherapy patients.

CONCLUSIONS

In conclusion, death from prostate cancer in this study rep-

resented approximately 10% of all deaths, with cardiovascu-

lar mortality predominating. In particular, overall survival

was poorest in the high-risk group. Although high-risk pa-

tients were most likely to die of prostate cancer, the diver-

gence in overall survival between high-risk and lower-risk

patients resulted primarily from an excess of cardiovascular

deaths. Changes in lifestyle to improve cardiovascular health,

including cessation of tobacco and minimization of ADT use,

may improve overall survival in patients with clinically local-

ized prostate cancer.

REFERENCES

1. Cooner W, Mosley B, Rutherford C, et al. Prostate cancer detec-tion in a clinical urological practice by ultrasonography, digitalrectal examination and prostate specific antigen. J Urol 1990;143:1146–1152.

2. Catalona W, Richie J, Ahmann F, et al. Comparison of digitalrectal examination and serum prostate specific antigen in theearly detection of prostate cancer: Results of a multicenter clin-ical trial of 6630 men. J Urol 1994;151:1283–1290.

3. Gann P, Hennekens C, Stampfer M. A prospective evaluation ofplasma prostate-specific antigen for detection of prostatic can-cer. JAMA 1995;273:289–294.

4. Tarone R, Chu K, Brawley O. Implications of stage-specific sur-vival rates in assessing recent declines in prostate cancer mortal-ity rates. Epidemiology 2000;11:167–170.

5. Stock R, Stone N, Tabert A, et al. A dose–response study for I-125 prostate implants. Int J Radiat Oncol Biol Phys 1998;41:101–108.

6. Orio P, Wallner K, Merrick G, et al. Dosimetric parameters aspredictive factors for biochemical control in patients with highrisk prostate cancer treated with Pd-103 and supplementalbeam radiation. Int J Radiat Oncol Biol Phys 2007;67:342–346.

7. Choi S, Wallner K, Merrick G, et al. Treatment margins predictbiochemical outcomes after prostate brachytherapy. Cancer J2004;10:175–180.

8. Merrick G, Butler W, Wallner K, et al. Androgen deprivationtherapy does not impact cause-specific or overall survival inhigh-risk prostate cancer managed with brachytherapy and sup-plemental external beam. Int J Radiat Oncol Biol Phys 2007;68:34–40.

9. Coebergh J, Janssen-Heijnen M, Post P, et al. Serious co-mor-bidity among unselected cancer patients newly diagnosed inthe southeastern part of The Netherlands in 1993–1996. J ClinEpidemiol 1999;52:1131–1136.

10. Satariano W, Ragland K, Van Den Eeden S. Cause of death inmen diagnosed with prostate carcinoma. Cancer 1998;83:1180–1188.

11. Post P, Hansen B, Kil P, et al. The independent prognostic valueof comorbidity among men aged <75 years with localized pros-tate cancer: A population-based study. Br J Urol Int 2001;87:821–826.

12. Merrick GS, Butler WM. Modified uniform seed loading forprostate brachytherapy. Rationale, design and evaluation.Tech Urol 2000;6:78–84.

13. Merrick GS, Butler WM, Wallner KE, et al. Dosimetry of anextracapsular annulus following permanent prostate brachyther-apy. Am J Clin Oncol 2007;30:228–233.

14. Partin AW, Mangold LA, Lamm DM, et al. Contemporaryupdate of prostate cancer staging nomograms (Partin Tables)for the new millennium. Urology 2001;58:843–848.

15. Wallner KE, Merrick GS, True L, et al. 20 Gy versus 44 Gy sup-plemental beam radiation with Pd-103 prostate brachytherapy:Preliminary biochemical outcomes from a prospective random-ized multi-center trial. Radiother Oncol 2005;75:307–310.

16. Merrick GS, Wallner KE, Butler WM, et al. Brachytherapy inmen aged # 54 years with clinically localized prostate cancer.Br J Urol 2006;98:324–328.

17. Stone N, Stock R, Unger P. Intermediate term biochemical-freeprogression and local control following 125-I brachytherapy forprostate cancer. J Urol 2005;173:803–807.

18. Zelefsky M, Wallner K, Ling C, et al. Comparison of the 5-yearoutcome and morbidity of three-dimensional conformal radio-therapy versus transperineal permanent iodine-125 implantationfor early-stage prostate cancer. J Clin Oncol 1999;17:517–522.

19. Zelefsky M, Yamada Y, Hunt M, et al. Comparison of 7-yearoutcomes between LDR brachytherapy and high dose IMRTfor patients with clinically localized prostate cancer. Int J RadiatOncol Biol Phys 2007;69:S178.

20. Aus G, Abbou C, Bolla M, et al. EAU guidelines on prostatecancer. Eur Urol 2005;48:546–551.

21. Middleton R. The management of clinically localized prostatecancer: Guidelines from the American Urologic Association.CA Cancer J Clin 1996;46:249–253.

22. De Groot V, Beckerman H, Lankhorst G, et al. How to measurecomorbidity. A critical review of available methods. J Clin Epi-demiol 2003;56:221–229.

23. Hall W, Jani A, Ryu J, et al. The impact of age and comorbidityon survival outcomes and treatment patterns in prostate cancer.Prostate Cancer Prostatic Dis 2005;8:22–30.

24. Miller T, Balady G, Fletcher G. Exercise and its role in the pre-vention and rehabilitation of cardiovascular disease. Ann BehavMed 1997;19:220–229.

25. Shepherd J, Cobbe S, Ford I, et al. Prevention of coronary car-diovascular disease with pravastatin in men with hypercholes-terolemia. N Engl J Med 1995;333:1301–1307.

26. Downs J, Clearfield M, Weis S, et al. Primary prevention ofacute coronary events with lovastatin in men and women withaverage cholesterol levels: Results of AFCAPS/TexCAPS.JAMA 1998;279:1615–1622.

27. Sever P, Dahlof B, Poulter N, et al. Prevention of coronary andstroke events with atorvastatin in hypertensive patients whohave average or lower-than-average cholesterol concentrations,in the Anglo-Scandinavian Cardiovascular Outcomes Trial–Lipid Lowering Arm (ASCOT-LLA): A multicentre rando-mised controlled trial. Lancet 2003;361:1149–1158.

28. Yusuf S, Sleight P, Pogue J, et al. Effects of an angiotensin-con-verting-enzyme inhibitor, ramipril, on cardiovascular events inhigh-risk patients. The Cardiovascular Outcomes PreventionEvaluation Study Investigators. N Engl J Med 2000;342:145–153.

440 I. J. Radiation Oncology d Biology d Physics Volume 72, Number 2, 2008

29. Fox K. Efficacy of perindopril in reduction of cardiovascular

events among patients with stable coronary artery disease:

Randomised, double-blind, placebo-controlled, multicentre trial

(the EUROPA study). Lancet 2003;362:782–788.30. Multiple Risk Factor Intervention Trial Research Group. Multi-

ple Risk Factor Intervention Trial: Risk factor changes and mor-

tality results. JAMA 1982;248:1465–1477.31. Hjermann I, Velve Byre K, Holme I, et al. Effect of diet and

smoking intervention on the incidence of coronary cardiovascu-

lar disease. Report from the Oslo Study Group of a randomisedtrial in healthy men. Lancet 1981;2:1303–1310.

32. Tsai H, D’Amico A, Sadetsky N, et al. Androgen deprivationtherapy for localized prostate cancer and the risk of cardiovas-cular mortality. J Natl Cancer Inst 2007;99:1516–1524.

33. Saigal C, Gore J, Krupski T, et al. Androgen deprivation ther-apy increases cardiovascular morbidity in men with prostatecancer. Cancer 2007;110:1493–1500.

34. Lee W. The role of androgen deprivation therapy combinedwith prostate brachytherapy. Urology 2002;60:39–44.