Brachytherapy 9 (2010) 42e49

Factors impacting all-cause mortality in prostate cancer brachytherapypatients with or without androgen deprivation therapy

Al V. Taira1, Gregory S. Merrick2,3,*, Robert W. Galbreath2,3, Wayne M. Butler2,3,Kent E. Wallner4

1Department of Radiation Oncology, University of Washington, Seattle, WA2Schiffler Cancer Center at Wheeling Hospital, Wheeling, WV

3Department of Physics, Wheeling Jesuit University, Wheeling, WV4Department of Radiation Oncology, Puget Sound Healthcare Corporation, Group Health Cooperative, University of Washington, Seattle, WA

ABSTRACT PURPOSE: Certain subsets of patients hav

Received 30 Apr

accepted 12 June 200

* Corresponding

1 Medical Park, Whe

þ304-243-5047.

E-mail address: gm

1538-4721/10/$ e see

doi:10.1016/j.brach

e an increased risk of all-cause mortality whenandrogen deprivation therapy (ADT) is used with definitive radiotherapy. We evaluated the relation-ship between pretreatment serum testosterone, age, and comorbidities on survival after prostatebrachytherapy in men treated with and without ADT.METHODS AND MATERIALS: From October 2001 to September 2005, 803 patients underwentbrachytherapy and 720 had a pretreatment serum testosterone. Comorbidities were prospectivelyrecorded for each patient (body mass indexO30, hypertension, diabetes, current smoker). Medianfollowup was 5.0 years. 34.2% of the patients received ADT. Focus was on subset of men whomight be expected to have more significant side effects associated with ADT.RESULTS: ADT did not significantly impact overall survival (OS) in men !65 years, O65 years,with one or no comorbidities, with more than one comorbidity, or with normal/high testosteronelevel. ADT use in men with low testosterone level was associated with decreased OS (83.6% vs.93.1%, p 5 0.01). The adverse impact of ADT in men with low testosterone level was restrictedto men with low testosterone level and more than one comorbidity (OS of 71.3% vs. 92.8%,p!0.01), with death from cardiovascular diseases accounting for almost all of the excess mortality.The subset of men with multiple comorbidities and normal/high testosterone level did not experi-ence adverse OS with ADT.CONCLUSIONS: Low pretreatment testosterone level may be a marker for men at increased riskof premature death with ADT. The combination of low pretreatment serum testosterone level andmultiple preexisting comorbidities is associated with decreased OS when ADT is incorporated intotreatment. � 2010 American Brachytherapy Society. Published by Elsevier Inc. All rights reserved.

Keywords: Prostate cancer; Testosterone; Overall survival; Androgen deprivation therapy

Introduction

Androgen deprivation therapy (ADT) is commonly usedwith definitive radiotherapy in the treatment of men withprostate cancer. The evidence supporting this for prostatecancer patients treated with external beam is robust, withseveral large randomized trials demonstrating cause-specific

il 2009; received in revised form 11 June 2009;

9.

author. Schiffler Cancer Center, Wheeling Hospital,

eling, WV 26003-6300. Tel.: þ304-243-3490; fax:

errick@urologicresearchinstitute.org (G.S. Merrick).

front matter � 2010 American Brachytherapy Society. Pu

y.2009.06.008

and overall survival (OS) benefits (1e3). The evidence sup-porting ADT use with brachytherapy is more limited. None-theless, excellent results in retrospective brachytherapyseries that include ADT (4) have convinced most practi-tioners to consider androgen suppression for higher-riskimplant patients as well.

The benefit of androgen deprivation is probably bothadditive and synergistic when combined with radiotherapy(5, 6). However, even short-term androgen deprivation haswell-known side effects, including loss of libido, hotflashes, weight gain, lethargy, and occasional cognitiveimpairment. Longer-term androgen suppression canincrease the risk of metabolic syndrome, bone fracture,and cardiovascular disease (7e9).

blished by Elsevier Inc. All rights reserved.

43A.V. Taira et al. / Brachytherapy 9 (2010) 42e49

Recent studies suggest that certain subsets of patientseither may not benefit or may have lower OS when ADTis added to definitive radiotherapy (10e12). Older ageand presence of existing comorbidities, in particular, havebeen implicated as potential risk factors in men receivingshort or longer courses of ADT. In a recent study fromour institution, we identified a trend toward lower OS inmen with low testosterone level receiving primarily short-course ADT (13). In this study, we update our earlier expe-rience, specifically focusing on the relationship betweenpretreatment serum testosterone, age, and comorbiditieson survival after prostate brachytherapy in men treated withand without ADT.

Methods and materials

From October 2001 through September 2005, 803 consec-utive patients underwent permanent interstitial brachyther-apy by a single brachytherapist (G.S.M.). All patientsunderwent brachytherapy for more than 3 years before anal-ysis, and no patient was lost to followup. Preplanning tech-nique, intraoperative approach, and dosimetric evaluationhave been described in detail (14).

Information on patient comorbidities was collectedprospectively for all patients. Comorbidities were definedas body mass indexO30, hypertension, diabetes, andcurrent tobacco use. Two hundred sixteen patients(30.0%) had body mass indexO30 kg/m2, 383 (53.2%)had hypertension, 104 (14.5%) had diabetes, and 111(15.4%) were current smokers. Two hundred thirty-fivepatients (32.6%) had more than one comorbidity.

Of the 803 patients, 720 had a pretreatment serumtestosterone. Patients had testosterone levels drawn beforebeginning any therapy, including ADT. Because of thegeographic diversity of the origin of patients in the studypopulation, pretreatment serum testosterone levels weredrawn from different regions of the United States and fromforeign countries using multiple different assays. Normalvalue ranges for these assays varied by 25e30%. Tonormalize the data for analysis, pretreatment testosteronelevels were assigned by the following criteria: belownormal for the particular assay (n 5 130), low one thirdof the normal range (n 5 344), mid one third normal range(n 5 171), higher one third normal range (n 5 61), andabove normal (n 5 14). All patients were encouraged tohave the serum testosterone level drawn in the morning.Although recent research has suggested that diurnal varia-tion of testosterone is much less than previously thought(15), we continue to encourage but not require morninglaboratory draws. All abnormal (below normal or abovenormal) testosterone levels were repeated to ensureaccuracy.

Of the 720 patients, 474 (65.8%) did not receive ADT,whereas 141 (19.6%) received <6 months of ADT, and105 (14.6%) received more than 6 months of ADT. ADTwas initiated 3 months before implantation and consisted

of a luteinizing hormone-releasing hormone agonist andan antiandrogen. The range of ADT duration was 3e36months. Indications for ADT were prostate cytoreductionbefore implant and higher-risk disease in men. Of the 720patients, 337 (46.8%) received supplemental external beamradiation therapy, generally to a dose of 45 Gy coveringprostate, seminal vesicles, and at-risk pelvic nodes.

Of the 720 patients, 249 presented with low-risk disease(prostate-specific antigen [PSA] < 10 ng/mL, Gleasonscore < 6, and clinical Stage <T2a). Three hundred sixty-three patients presented with intermediate-risk disease (oneadverse factor: PSA 5 10.1e19.9 ng/mL, Gleason score 5 7,or clinical Stage T2b), and 108 patients presented with high-risk disease (PSA > 20 ng/mL, Gleason score > 8, clinicalStage T2c, or two or more intermediate risk factors).

Patients were monitored by physical examination,including digital rectal examination and serum PSA determi-nation at 3- and 6-month intervals. The primary endpoint ofthe analysis was cause-specific survival (CSS) and OS. Bio-chemical progression-free survival (bPFS) was evaluated asa secondary endpoint, and was defined as PSA < 0.40 ng/mLafter nadir (16). Multiple clinical, treatment, and dosimetricparameters were evaluated for impact on survival.

Cause of death was determined for each deceased patient.Patients with metastatic prostate cancer or hormone refrac-tory disease without obvious metastases who died of anycause were classified as dead of prostate cancer. All otherdeaths were attributed to the immediate cause of death.

Clinical and treatment variables that were continuouswere compared across groups using a one-way analysis ofvariance. Categorical variables were compared using a c

2

analysis. The differences in the survival of the patientgroups were compared using KaplaneMeier curves andlog-rank tests. Univariate Cox regression analysis was usedto determine if any of the clinical or treatment variablespredicted survival. Variables with a p value !0.10 wereentered into a multivariate, forward conditional Cox regres-sion. For all tests, a p value <0.05 was considered signifi-cant. Statistical analyses were performed with SPSSVersion 15.0 software (SPSS, Inc., Chicago, IL).

Results

Clinical, treatment, and dosimetric parameters for the720 patients included in the analysis, stratified by pretreat-ment testosterone, is detailed in Table 1. For the group asa whole, 6-year bPFS and CSS were 98.3% and 99.9%,respectively (Fig. 1). On multivariate analysis, onlypretreatment percent positive biopsy cores ( p 5 0.004)were found to be significantly related to bPFS (Table 2).Because only 1 patient died of prostate cancer, CSS predic-tors were not identified. OS at 6 years was 90.3%. Fifty-seven patients died during the study followup period, withonly one dead from metastatic prostate cancer. On multivar-iate analysis, number of comorbidities and patient age weremost highly correlated with OS ( p!0.01).

Table 1

Clinical, treatment, and dosimetric parameters of study population

Continuous

variables

Testosterone

Below

normal (n 5 130)

Lower 1/3 of

normal (n 5 344)

Middle 1/3 of

normal (n 5 171)

Upper 1/3 of

normal (n 5 61)

Above

normal (n 5 14)

p

Total (n 5 720)

Median Mean Median Mean Median Mean Median Mean Median Mean Median Mean

Age (y) 64.5 64.6 66.0 64.6 65.0 65.1 66.0 65.2 63.5 62.9 0.811 65.0 64.7

Followup 5.2 5.1 5.0 5.0 4.8 5.0 5.7 5.6 4.7 4.6 0.062 5.0 5.1

Preimplant PSA 5.5 7.0 5.8 7.3 6.2 8.3 6.7 8.5 6.6 9.2 0.076 5.9 7.6

Gleason score 7.0 6.8 7.0 6.8 7.0 6.8 7.0 6.9 7.0 6.8 0.902 7.0 6.8

Percent positive

biopsies

33.3 36.3 30.0 36.0 25.0 34.2 33.3 41.3 31.7 36.3 0.473 30.0 36.1

Body mass index 29.5 31.0 28.1 28.8 27.5 28.2 26.3 26.7 24.8 24.2 !0.001 28.1 28.8

Prostate volume 33.7 33.9 31.8 32.0 34.2 34.0 32.5 32.9 32.5 32.1 0.096 32.7 32.9

V100 98.3 97.6 98.4 97.8 98.3 97.7 98.2 97.6 98.7 97.6 0.912 98.3 37.7

% D90 121.8 122.0 122.6 122.5 122.2 123.0 123.2 122.9 124.6 124.2 0.890 122.4 122.6

Categorical variables Count % Count % Count % Count % Count % p Count %

Clinical stage

T1beT2b 129 99.2 338 98.3 160 93.6 61 100 14 100 0.015 702 97.5

T2ceT3c 1 0.8 6 1.7 11 6.4 0 0 0 0 18 2.5

Risk group

Low 48 36.9 128 37.2 56 32.7 14 23.0 3 21.4 0.057 249 34.6

Intermediate 72 55.4 163 47.4 83 48.5 36 59.0 9 64.3 363 50.4

High 10 7.7 53 15.4 32 18.7 11 18.0 2 14.3 108 15.0

Isotope

I-125 5 3.8 10 2.9 2 1.2 2 3.3 0 0 0.601 19 2.6

Pd-103 125 96.2 334 97.1 169 98.8 59 96.7 14 100 701 97.4

XRT

Yes 50 38.5 154 44.8 89 52.0 35 57.4 9 64.3 0.032 337 46.8

No 80 61.5 190 55.2 82 48.0 26 42.6 5 35.7 383 53.2

ADT, months

0 89 68.5 230 66.9 111 64.9 35 57.4 9 64.3 0.488 474 65.8

<6 27 20.8 66 19.2 28 16.4 17 27.9 3 21.4 141 19.6

O6 14 10.8 48 14.0 32 18.7 9 14.8 2 14.3 105 14.6

Hypertension

Yes 80 61.5 183 53.2 91 53.2 24 39.3 5 35.7 0.040 383 53.2

No 50 38.5 161 46.8 80 46.8 37 60.7 9 64.3 337 46.8

Diabetes

Yes 25 19.4 49 14.2 25 14.6 4 6.6 1 7.1 0.190 104 14.5

No 104 80.6 295 85.8 146 85.4 57 93.4 13 92.9 615 85.5

Tobacco

Current 9 6.9 48 14.0 32 18.7 16 26.2 6 42.9 0.002 111 15.4

Former 62 47.7 133 38.7 74 43.3 23 37.7 5 35.7 297 41.3

Never 59 45.4 163 47.4 65 38.0 22 36.1 3 21.4 312 43.3

Perineural invasion

Yes 36 27.9 124 36.0 49 28.7 18 29.5 4 28.6 0.335 231 32.1

No 93 72.1 220 64.0 122 71.3 43 70.5 10 71.4 488 32.1

PSA 5 prostate-specific antigen; XRT 5 external beam radiation therapy; ADT 5 androgen deprivation therapy. Entries in bold indicate statistically

significant.

44 A.V. Taira et al. / Brachytherapy 9 (2010) 42e49

Figure 2 displays OS, with patients stratified by age,number of comorbidities, and pretreatment testosteronelevel. Older patients ( p 5 0.001) and patients with more thanone comorbidity ( p 5 0.004) had statistically significantlower OS. There was no difference in OS when patients werestratified by pretreatment testosterone level ( p 5 0.64).

ADT use did not significantly impact OS when menwere stratified by prostate cancererisk group. Among

men with low-risk disease, OS was 94.6% vs. 92.4%( p 5 0.732) for those who did/did not receive ADT. Therealso was no difference in OS in intermediate-risk (80.4%vs. 90.6%, p 5 0.130) or high-risk patients (90.8% vs.100%, p 5 0.239) who did/did not receive ADT.

ADT use did not significantly impact OS when men werestratified by age group. Among men !65 years old, OS was96.4% vs. 95.3% ( p 5 0.74) for those who did/did not receive

Years Since Implant

876543210

Su

rv

iv

al (%

)

100

80

60

40

20

0

n =720 706 696 685 533 365 227 83

Cause-Specific Survival, 99.9%

Biochemical Progression-Free Survival, 98.3%

Overall Survival, 90.3%

Fig. 1. KaplaneMeier curves for cause-specific, biochemical progre-

ssionefree, and overall survival (each curve represents the total study

population).

45A.V. Taira et al. / Brachytherapy 9 (2010) 42e49

ADT. Among men >65 years old, OS was 84.0% vs. 88.1%( p 5 0.50) for those who did/did not receive ADT.

ADT use did not significantly impact OS when menwere stratified by number of comorbidities. Among menwith one or no comorbidities, OS was 94.2% vs. 92.5%( p 5 0.56) for those who did/did not receive ADT. Amongmen with more than one comorbidity, OS was 76.3% vs.89.7 ( p 5 0.065) for those who did/did not receive ADT.

ADT use also did not significantly impact OS in menwith normal/high pretreatment testosterone level, with OSof 95.7% vs. 89.0% ( p 5 0.07) for those who did/did not

Table 2

Univariate and multivariate analysis (Cox regression) for biochemical progression

there was only one prostate cancererelated death)

Variables

Biochemical survival

Univariate Multivariate

p RR p

Age 0.904

PSA 0.015 1.053 0.094

Gleason score 0.067 0.761

% Positive biopsies 0.004 1.028 0.004

Body mass index 0.932

Prostate volume 0.128

Comorbidities (0e4) 0.861

V100 0.498

% D90 0.149

PNIa 0.456

Stage groupa 0.706

Isotopea 0.699

Riska 0.010 0.102

Low vs. intermediate 0.703

Low vs. high 0.017 7.045

XRTa 0.428

Hormonesa 0.578

Hormone duration 0.567

RR 5 relative risk; PSA 5 prostate-specific antigen; PNI 5 perineural invasioa Categorical data. All others were continuous.

receive ADT. However, ADT use was associated withdecreased survival in men with low pretreatment testos-terone level. Among these men, OS was 83.6% vs. 93.1%( p 5 0.01) for those who did/did not receive ADT (Fig. 3).

Because of the trend toward significance of ADT use inmen with multiple comorbidities, we hypothesized thatmen with both low testosterone level and multiple comor-bidities might be particularly at risk for adverse effectsassociated with ADT. In fact, we found that the adverseimpact of ADT use was limited to men with low testos-terone level and more than one comorbidity. In this group,OS was 71.3% vs. 92.8% ( p 5 0.005) in men who did/didnot receive ADT (Fig. 4). Death because of cardiovasculardisease accounted for almost all of the excess mortality(Fig. 5). There was no difference in OS for men with lowtestosterone level and fewer comorbidities, with OS of92.0% vs. 93.2% ( p 5 0.59) for those who did/did notreceive ADT. Furthermore, there was no difference in OSfor men with normal/high testosterone level and more thanone comorbidity, with OS of 91.7% vs. 83.2% ( p 5 0.45)for those who did/did not receive ADT.

There were no statistically significant differences inbPFS among the main comparison groups. bPFS was98.3% vs. 98.2% ( p 5 0.95) for men with fewer vs. morecomorbidities; 98.1% vs. 98.4% ( p 5 0.80) for men !65and >65 years old; 98.3% vs. 98.2% ( p 5 0.95) for menwith higher vs. lower pretreatment testosterone level,respectively. There were no significant differences in CSSin any groups compared.

Table 3 presents the cause of death, stratified by pretreat-ment testosterone level. Cardiovascular death was by far

-free and overall survival (cause-specific survival was not included, because

Overall survival

Univariate Multivariate

RR p RR p RR

!0.001 1.113 !0.001 1.115

0.514

0.036 1.344 0.253

1.028 0.822

0.252

0.391

0.015 1.383 0.009 1.430

0.036 0.899 0.048 0.902

0.188

0.589

0.434

0.525

0.165

d

d

0.183

0.314

0.346

n; XRT 5 external beam radiation therapy.

Fig. 2. Overall survival stratified by (a) age group, (b) comorbidities, and (c) testosterone group.

46 A.V. Taira et al. / Brachytherapy 9 (2010) 42e49

the most common cause of mortality. For patients with lowtestosterone level, cardiovascular death accounted for 41%(16 of 39) of total deaths. For patients with normal or hightestosterone level, cardiovascular death accounted for 27%(5 of 18) of total deaths. Of note, only 1 patient died ofprostate cancer.

Fig. 3. Overall survival of men with low to lower 1/3 testosterone, strat-

ified by hormone treatment.

Discussion

It is well established that low serum testosterone level,even without iatrogenic chemical castration, is associatedwith an increased risk of dyslipidemia, diabetes, vasculardisease, bone fracture, depression and overall mortality(17e21). We hypothesized that men with already low base-line testosterone level might be predisposed to earlier andmore severe adverse effects associated with iatrogeniccomplete androgen deprivation. Laughlin et al. (19)recently evaluated the impact of baseline testosterone onmortality in a prospective cohort of men not undergoingcancer therapy with median 11.8-year followup. Theyfound that low baseline testosterone level was associatedwith a statistically significant increase in cardiovasculardeaths and in deaths from pulmonary disease. A nonsignif-icant increase in cancer death was also noted. Thisincreased mortality risk continued to be significant evenwhen adjusted for the presence of existing comorbiditiesand was not affected if deaths in the first 5 years wereexcluded. These findings suggest that low testosterone levelmay be more than just a marker for existing comorbiditiesand, in fact, may contribute to the development of cardio-vascular and other ailments.

In an earlier report (13), we found a trend towarddecreased OS in men with low baseline testosterone level,who were then treated with ADT. However, we did notexplore the potential confounding role of existing comorbid-ities in that study. The purpose of the current report is to betterassess whether low baseline testosterone level by itself isa risk factor, or if low testosterone level is simply a markerfor existing comorbidities. We also have the benefit in thisreport of a larger number of patients (720 vs. 546).

In our analysis, neither older age nor the presence ofmultiple comorbidities increased all-cause mortality risk

for men treated with ADT. However, the presence of lowpretreatment testosterone level was associated with anincrease in all-cause mortality in men receiving ADT.Although we found that low testosterone level was moreimportant than the existence of comorbidities in predictingmen who were at risk of serious ADT side effects on univar-iate analysis, the relationship between the presence of comor-bidities and the baseline testosterone level seems to be veryimportant. Men with normal or high testosterone levels andmultiple comorbidities did not experience lower OS withADT use nor did those with low testosterone level and oneor fewer comorbidities. It was only the subgroup of men withlow baseline testosterone level and multiple comorbiditeswho had much higher all-cause mortality when ADT wasused. Hence, multiple comorbidities in the absence of lowtestosterone level does not seem to place men at higher riskof premature death, and the same can be said of low testos-terone level in the absence of multiple comorbidities.

Several authors (3, 22, 23) have recently reviewedoutcomes of large randomized trials of external beam radio-therapy from the early 1990s and reported no increase incardiovascular mortality in patients receiving ADT.However, these analyses were not stratified by age,

Fig. 4. Overall survival for patients who have low to lower 1/3 testosterone in addition to (a) one or no comorbidity, stratified by androgen deprivation

therapy (ADT) and (b) two or more comorbidities, stratified by ADT.

47A.V. Taira et al. / Brachytherapy 9 (2010) 42e49

presence of existing comorbidities, and pretreatment testos-terone. In fact, it is consistent with our results that wewould not expect to see a survival detriment associated withADT in unstratified patients. We also found no survivaldetriment with ADT use when considering our entirepatient population. It was only when we stratified usinglow pretreatment testosterone level and comorbidities thatwe identified a particularly vulnerable patient subgroup.Furthermore, prostate cancerespecific mortality was quitehigh in these studies (up to 35%), reflecting the advancedstage of patients included in the trials. As noted by anauthor of one of those reports (23), the adverse impact of

Fig. 5. Cardiovascular-related mortality as percent of men in each group for (a

subset of men with low testosterone stratified by use of androgen deprivation the

comorbidities stratified by use of ADT.

ADT may be more prominent when considering patientswith more favorable cancer-specific prognosis.

Even those studies that have identified an adverse impactof ADT use in certain subsets of men have somewhat con-flicting findings. Thus far, investigators have focused onolder men or those with preexisting comorbidities. D’Ami-co et al. (11) investigated the impact of ADT on the inci-dence of fatal myocardial infarction in men 65 years orolder. No differences were found in the number of fatalcardiac events (18 vs. 16 cumulative events for men whodid/did not receive ADT); but fatal events occurred soonerin the ADT group. In a separate update from a previous

) entire population stratified by normal/high vs. low testosterone; (b) the

rapy (ADT); and (c) the subset of men with low testosterone and multiple

Table 3

Cause of death, stratified by pretreatment testosterone

Cause of death

Testosterone level

Mid-1/3ehigh

(n 5 246)

Lowelower 1/3

(n 5 474)

Total

(n 5 720)

n % n % n %

Prostate cancer 0 0 1 0.2 1 0.1

Cardiovascular 5 2.0 16 3.4 21 2.9

GI malignancies 4 1.6 5 1.1 9 1.3

Bladder cancer 0 0 1 0.2 1 0.1

Lung cancer 3 1.2 5 1.1 8 1.1

Leukemia/lymphoma 0 0 2 0.4 2 0.3

Melanoma 0 0 1 0.2 1 0.1

GBMF 1 0.4 0 0 1 0.1

Head and neck

cancers

0 0 1 0.2 1 0.1

Pulmonary 4 1.6 2 0.4 6 0.8

Trauma 0 0 3 0.6 3 0.4

Sepsis 1 0.4 1 0.2 2 0.3

Alzheimer’s 0 0 1 0.2 1 0.1

Total 18 7.3 39 8.2 57 7.9

GI 5 gastrointestinal; GBMF 5 glioblastoma multiforme.

48 A.V. Taira et al. / Brachytherapy 9 (2010) 42e49

randomized trial (10), D’Amico et al. reported a lack ofsurvival benefit with ADT in a subset of 49 men withmoderate or severe comorbidities, with a trend towarddecreased OS in men receiving androgen deprivation. Formen 70 years and older, Dosoretz et al. (12) recently re-ported in abstract that ADT use was associated with anincreased risk of all-cause mortality in a retrospective bra-chytherapy cohort, in which mean androgen deprivationwas less than 4 months.

However, Tsai et al. (24), using the Cancer of the Pros-tate Strategic Urologic Research Endeavor (CAPSURE)database found no increased risk of cardiovascular deathwith ADT use in radiotherapy patients 65 years and older.Interestingly, this study did identify an increased risk ofcardiovascular death for prostatectomy patients 65 yearsand older, treated with ADT. Efstathiou et al. (25) foundno relationship between cardiovascular mortality and dura-tion of ADT in men older than 70 years enrolled on Radi-ation Therapy Oncology Group 92-02, and Silverman et al.(26) presented in abstract that longer ADT duration wasassociated with improved OS in a retrospective cohort ofexternal beam radiotherapy patients older than 70 years.

Our results suggest that men with the combination oflow testosterone level and multiple comorbidities representa particular group vulnerable to the adverse effects ofandrogen deprivation. Our results also may help to explainthe conflicting conclusions of previous research regardingpotential adverse impact of ADT. For instance, none ofthe earlier negative studies stratified patients by pretreat-ment testosterone levels, which in our analysis was themost predictive univariate risk factor for poor outcomeswith ADT. In addition, several of the negative reports(25, 26) investigated the impact of the duration of ADTuse on overall mortality and concluded that longer vs.

shorter ADT duration did not lead to an increase in cardio-vascular or all-cause mortality. Of course, this still leavesopen the question whether ADT use (vs. none) may leadto worse outcomes in certain subsets of patients.

Our results also may provide context for the mixed find-ings of studies that did identify an adverse impact of ADT.For instance, the D’Amico study focusing on men >65years old (11) and the study focusing on men with preexist-ing moderate/severe comorbidities (10) found an adverseimpact of ADT use in these subgroups but no OS detriment.This is consistent with our finding that older age alone orcomorbidities alone may not by themselves increasemortality risk with ADT use. However, stratification bypretreatment testosterone may help to identify a subset ofpatients, particularly those with low testosterone level andmultiple comorbidities, which may be most impacted.Our hope is that further stratification also will identifya significant portion of older men and those with comorbid-ities who may not have increased risk when ADT is incor-porated into their care.

Limitations

This is a retrospective analysis. Unfortunately, there areno randomized trials evaluating the use of ADT with bra-chytherapy nor is there any analysis from randomized trialsinvestigating the impact of low baseline testosterone levelin men treated with brachytherapy.

Based on the data presented earlier, we postulate a causalrelationship that iatrogenically lowering testosterone tocastration levels in a subset of men with already low testos-terone level and multiple comorbidities leads to higher all-cause mortalitydprimarily cardiovascular mortality. Ofcourse, a retrospective analysis like this cannot provecausality because of potential unknown confounder. Westratified our patients as best as we could to account forknown confounders of older age and comorbidities. Impor-tantly, we found that men with multiple comorbidities andnormal or high testosterone levels did not experiencea survival detriment with ADT use. We found the same inolder men with normal or high testosterone levels. Thus,we were able to isolate the potential impact of theseconfounders on the analysis. Furthermore, like several ofthe studies listed earlier that link ADT to adverse outcomesin specific patient populations, our absolute number of 57events (deaths) is fairly small.

It is beyond the scope of this article to evaluate whichpatients may benefit from combination of ADT with defin-itive brachytherapy. Rather, we attempt to identify a subsetof men who may be at higher risk for adverse outcomeswith total androgen suppression. This potential increasedrisk can then be considered along with the known andpresumed benefits of combined modality when decisionsare made regarding ADT. This may become more relevantas the benefit of short-term androgen deprivation for

49A.V. Taira et al. / Brachytherapy 9 (2010) 42e49

cytoreduction in men undergoing prostate brachytherapy isbeing increasingly called into question (27, 28).

As it was important, we attempted to more specificallyidentify factors that did not seem to increase risk of all-cause mortality with ADT use. Corroboration of our find-ings, ideally as part of a randomized trialdbut at least withadditional retrospective analyses from a range of treatmentcentersdwould be strongly encouraged.

A strength of this study compared with some of thepreviously discussed studies is that it represents a moremodern cohort of patients, with treatment methods andcancer-specific prognosis similar to patients that clinicianswill see in the coming years and about whom they need tomake decisions regarding ADT.

Conclusion

ADT use in men with low pretreatment serum testosteronelevel and multiple comorbidities was associated withincreased all-cause mortality. Neither older age nor presenceof multiple comorbidities, per se, was associated withdecreased survival when ADTwas used. We recommend thatfuture randomized trials incorporate pretreatment testos-terone as a stratification factor to better elucidate the interac-tion between baseline testosterone, comorbidities, andandrogen deprivation.

References

[1] Bolla M, Collette L, Blank L, et al. Long-term results with immediate

androgen suppression and external irradiation in patients with locally

advanced prostate cancer (an EORTC study): a phase III randomised

trial. Lancet 2002;360:103e106.

[2] D’Amico AV, Manola J, Loffredo M, et al. 6-Month androgen

suppression plus radiation therapy vs radiation therapy alone for

patients with clinically localized prostate cancer: a randomized

controlled trial. JAMA 2004;292:821e827.

[3] Roach M III, Bae K, Speight J, et al. Short-term neoadjuvant

androgen deprivation therapy and external-beam radiotherapy for

locally advanced prostate cancer: Long-term results of RTOG 8610.

J Clin Oncol 2008;26:585e591.

[4] Stock RG, Ho A, Cesaretti JA, Stone NN. Changing the patterns of

failure for high-risk prostate cancer patients by optimizing local

control. Int J Radiat Oncol Biol Phys 2006;66:389e394.

[5] Milosevic M, Chung P, Parker C, et al. Androgen withdrawal in

patients reduces prostate cancer hypoxia: implications for disease

progression and radiation response. Cancer Res 2007;67:6022e6025.

[6] Pinthus JH, Bryskin I, Trachtenberg J, et al. Androgen induces adap-

tation to oxidative stress in prostate cancer: Implications for treat-

ment with radiation therapy. Neoplasia 2007;9:68e80.

[7] Hakimian P, Blute M Jr, Kashanian J, et al. Metabolic and cardiovas-

cular effects of androgen deprivation therapy. BJU Int 2008;102:

1509e1514.

[8] Harle LK, Maggio M, Shahani S, et al. Endocrine complications of

androgen-deprivation therapy in men with prostate cancer. Clin Adv

Hematol Oncol 2006;4:687e696.

[9] Isbarn H, Boccon-Gibod L, Carroll PR, et al. Androgen deprivation

therapy for the treatment of prostate cancer: Consider both benefits

and risks. Eur Urol 2009;55:62e75.

[10] D’Amico AV, Chen MH, Renshaw AA, et al. Androgen suppression

and radiation vs radiation alone for prostate cancer: A randomized

trial. JAMA 2008;299:289e295.

[11] D’Amico AV, Denham JW, Crook J, et al. Influence of androgen

suppression therapy for prostate cancer on the frequency and timing

of fatal myocardial infarctions. J Clin Oncol 2007;25:2420e2425.

[12] Dosoretz AM, Chen MH, Salenius SA, et al. Mortality in men age 70

or more with localized prostate cancer treated with brachytherapy

with or without neoadjuvant hormonal therapy. Int J Radiat Oncol Bi-

ol Phys 2008;72:s39.

[13] Taira AV, Merrick GS, Galbreath RW, et al. Pretreatment serum

testosterone and androgen deprivation: Effect on disease recurrence

and overall survival in prostate cancer patients treated with brachy-

therapy. Int J Radiat Oncol Biol Phys 2009;74:1143e1149.

[14] Merrick GS, Butler WM. Modified uniform seed loading for prostate

brachytherapy: Rationale, design, and evaluation. Tech Urol 2000;6:

78e84.

[15] Crawford ED, Barqawi AB, O’Donnell C, et al. The association of

time of day and serum testosterone concentration in a large screening

population. BJU Int 2007;100:509e513.

[16] Bittner N, Merrick GS, Galbreath RW, et al. Primary causes of death

after permanent prostate brachythearpy. Int J Radiat Oncol Biol Phys

2008;72:433e440.

[17] Ding EL, Song Y, Malik VS, et al. Sex differences of endogenous sex

hormones and risk of type 2 diabetes: a systematic review and meta-

analysis. JAMA 2006;295:1288e1299.

[18] Khaw KT, Dowsett M, Folkerd E, et al. Endogenous testosterone and

mortality due to all causes, cardiovascular disease, and cancer in

men: European prospective investigation into cancer in Norfolk

(EPIC-Norfolk) Prospective Population Study. Circulation 2007;

116:2694e2701.

[19] Laughlin GA, Barrett-Connor E, Bergstrom J. Low serum testos-

terone and mortality in older men. J Clin Endocrinol Metab 2008;

93:68e75.

[20] Matsumoto AM. Andropause: Clinical implications of the decline in

serum testosterone levels with aging in men. J Gerontol A Biol Sci

Med Sci 2002;57:M76eM99.

[21] Shores MM, Matsumoto AM, Sloan KL, et al. Low serum testos-

terone and mortality in male veterans. Arch Intern Med 2006;166:

1660e1665.

[22] Bolla M, Collette L, van Tienhoven G, et al. Ten year results of long

term adjuvant androgen deprivation with goserlin in patients with

locally advanced prostate cancer treated with radiotherapy: A phase

III EORTC study. Int J Radiat Oncol Biol Phys 2008;72:s30.

[23] Efstathiou JA, Bae K, Shipley WU, et al. Cardiovascular mortality

after androgen deprivation therapy for locally advanced prostate

cancer: RTOG 85-31. J Clin Oncol 2009;27:92e99.

[24] Tsai HK, D’Amico AV, Sadetsky N, et al. Androgen deprivation

therapy for localized prostate cancer and the risk of cardiovascular

mortality. J Natl Cancer Inst 2007;99:1516e1524.

[25] Efstathiou JA, Bae K, Shipley WU, et al. Cardiovascular mortality

and duration of androgen deprivation for locally advanced prostate

cancer: analysis of RTOG 92-02. Eur Urol 2008;54:816e823.

[26] Silverman JS, Ruth K, Horwitz EM, et al. Radiation therapy for high

risk prostate cancer: Do patients 70 years or older benefit from

combined androgen deprivation therapy. Int J Radiat Oncol Biol Phys

2008;72:s75es76.

[27] Petit JH, Gluck C, Kiger WS, et al. Androgen deprivation-medi-

ated cytoreduction before interstitial brachytherapy for prostate

cancer does not abrogate the elevated risk of urinary morbidity

associated with larger initial prostate volume. Brachytherapy 2007;

5:256e261.

[28] Gibbons EP, Jacobs BL, Smith RP, et al. Dosimetric outcomes in

prostate brachytherapy: Is downsizing the prostate with androgen

deprivation necessary? Brachytherapy 2009;8:304e308.