Received: 6 August 2018 | Revised: 4 April 2019 | Accepted: 8 April 2019

DOI: 10.1111/ecc.13076

ORIGINAL ARTICLE

Health‐related quality of life in long‐term survivors with

localised prostate cancer by therapy—Results from a
population‐based study

Salome Adam1,2 | Lena Koch‐Gallenkamp3 | Heike Bertram4 | Andrea Eberle5 |

Bernd Holleczek6 | Ron Pritzkuleit7 | Mechthild Waldeyer‐Sauerland8 |
Annika Waldmann8,9 | Sylke Ruth Zeissig10 | Sabine Rohrmann2 |
Hermann Brenner3,11,12 | Volker Arndt1
1
Unit of Cancer Survivorship, Division of Clinical Epidemiology and Aging Research, German Cancer Research Center (DKFZ), Heidelberg, Germany
2
Division of Chronic Disease Epidemiology, Epidemiology, Biostatistics and Prevention Institute, University of Zurich, Zurich, Switzerland
3
Division of Clinical Epidemiology and Aging Research, German Cancer Research Center (DKFZ), Heidelberg, Germany
4
Cancer Registry of North Rhine‐Westphalia, Bochum, Germany
5
Bremen Cancer Registry, Leibniz Institute for Prevention Research and Epidemiology – BIPS, Bremen, Germany
6
Saarland Cancer Registry, Saarbrücken, Germany
7
Schleswig‐Holstein Cancer Registry, Lübeck, Germany
8
Hamburg Cancer Registry, Hamburg, Germany
9
Institute of Social Medicine and Epidemiology, University Lübeck, Lübeck, Germany
10
Cancer Registry of Rhineland‐Palatinate, Mainz, Germany
11
Division of Preventive Oncology, German Cancer Research Center (DKFZ) and National Center for Tumor Diseases (NCT), Heidelberg, Germany
12
German Cancer Consortium (DKTK), German Cancer Research Center (DKFZ), Heidelberg, Germany

Correspondence
Volker Arndt, Unit of Cancer Survivorship, Abstract
Division of Clinical Epidemiology and Aging Objective: Several therapies for localised prostate cancer (PC) are available; all yield
Research, German Cancer Research Center
(DKFZ), Im Neuenheimer Feld 280, 69120 similar survival rates. However, each therapy has significant side effects that can
Heidelberg, Germany. influence patients' health‐related quality of life (HRQoL) in the long run.
Email: v.arndt@dkfz-heidelberg.de
Methods: The study sample included 911 survivors with localised PC, 5–15 years
Funding information post‐diagnosis who were identified from the population‐based CAESAR + study in
Deutsche Krebshilfe, Grant/Award Number:
No. 108262 Germany. HRQoL was assessed using the EORTC QLQ‐C30 and EORTC QLQ‐PR25
questionnaires. The association between type of therapy and HRQoL was assessed
with multivariable linear regression and global F-test adjusting for age, time since
diagnosis and comorbidities.
Results: Overall, survivors treated with radical prostatectomy (RP) or radiotherapy (RT)
alone reported the best HRQoL and the lowest symptom burden. Conversely, survivors
treated with androgen deprivation therapy (ADT) (& RP/RT) or RP & RT (in combination)
reported the worst HRQoL and the highest symptom burden. Significant differences
among treatment groups in HRQoL were found for global health status (p = 0.041),

Eur J Cancer Care. 2019;00:e13076. wileyonlinelibrary.com/journal/ecc © 2019 John Wiley & Sons Ltd | 1 of 10
https://doi.org/10.1111/ecc.13076
2 of 10 | ADAM et al.

social functioning (p = 0.007), urinary symptoms (p = 0.035), bowel symptoms (p = 0.017)

and hormonal treatment‐related symptoms (p < 0.001) among other symptoms.
Conclusions: Long‐term localised PC survivors formerly treated with a combination
of RP and RT or with ADT report poorer HRQoL and more symptoms than patients
treated with either RP or RT alone.

KEYWORDS
health‐related quality of life, prostate cancer, therapy, well‐being

1 | I NTRO D U C TI O N it was not possible to pool findings, indicating the need for more re‐
search. Additionally, none of the studies specifically investigated how
In more economically developed countries, prostate cancer (PC) re‐ therapy combinations and/or ADT can affect HRQoL. Thus, we com‐
mains the most prevalent form of cancer in men (American Cancer pared HRQoL by therapy, including patients prescribed a combination
Society, 2015). In those countries, an estimated 754,503 men were di‐ treatment and/or ADT, using data from a German population‐based
agnosed with PC in 2018 (The Union for International Cancer Control's, observational study with a long‐term follow‐up time (≥5 years).
2019). Since the introduction of the prostate‐specific antigen (PSA)
test, most PC cases are diagnosed with an localised tumour (Etzioni et
2 | PATI E NT S A N D M E TH O DS
al., 2002). However, for patients with localised PC, there is currently
no consensus on an optimal therapy strategy (Heidenreich et al., 2014;
2.1 | Study design and study population
National Comprehensive Cancer Network, 2014; Serrell et al., 2017).
Until now, radical prostatectomy (RP) and radiotherapy (RT) (external Participants were included from the multi‐regional, population‐based
beam radiation or brachytherapy) were the most common therapy op‐ CAESAR + study in Germany. The intention of this study was to investi‐
tions for PC patients < 75 years of age (Miller et al., 2016). Other treat‐ gate HRQoL and other survivor aspects among long‐term breast, colo‐
ment strategies include watchful waiting (WW) and active surveillance rectal and prostate cancer survivors. Details of the CAESAR + study
(AS), as well as androgen deprivation therapy (ADT). The latter is either recruitment and data collection design have been described elsewhere
applied prior to or concurrent with RT (National Comprehensive Cancer (Arndt et al., 2017). In short, the CAESAR + study included 6,952 long‐
Network, 2014; Parker, Gillessen, Heidenreich, & Horwich, 2015). Yet, term cancer survivors who were diagnosed between 1994 and 2004
it remains unclear which treatment has the most favourable survival and age 20–75 years at diagnosis. They were identified and recruited
outcome or whether survival differs by treatment at all. Whereas re‐ via six German cancer registries (Bremen, Hamburg, Münster/North
cent studies reveal that curative therapies (RP and RT) do not (or only Rhine‐Westphalia, Rhineland‐Palatinate, Saarland and Schleswig‐
minimally) improve PC‐specific and overall survival in comparison with Holstein). Data collection was conducted between 2009 and 2011 by
AS in patients with localised PC (Hamdy et al., 2016; Wilt et al., 2017), postal questionnaire. Non‐respondents received up to two follow‐up
another study revealed that patients with localised PC treated with RP reminders by mail or phone. Among PC survivors, 5,092 were eligible
have significantly reduced PC‐specific mortality compared to patients for the CAESAR + study, out of which 2,403 returned a completed
on WW (Holmberg et al., 2002). Additionally, prognosis in survivors questionnaire (overall response rate: 47.2%).
with localised PC appears better predicted by pre‐treatment patho‐ Ethical approval for the study was obtained from the inter‐
logic and disease factors than by treatment options (Freedland, 2011). nal review board (ethics committee) of the medical faculty of the
In general, patients with localised PC experience very good five‐year University of Heidelberg and by all review boards accountable for the
relative rates of 90% to 95% in developed countries (National Cancer participating cancer registries. (IRB approval number: S‐438/2008).
Institute, 2018; Robert Koch Institut, 2017) and are more likely to die For the present analysis, we included only survivors with a diag‐
of causes other than the disease (Stattin et al., 2010). Given the lack of nosis of localised PC (T1N0M0/T2N0M0). PC survivors diagnosed
agreement on optimal PC treatment and the limited knowledge regard‐ with another malignancy prior to PC were excluded.
ing long‐term effects in survivors with localised PC, it is increasingly
important to assess PC survivors' long‐term HRQoL (≥5 years after di‐
2.2 | Study measurements
agnosis) (American Cancer Society, 2000) by treatment (Drummond et
al., 2015; Adam, Feller, Rohrmann, & Arndt, 2018).
2.2.1 | European Organisation for Research and
A recent review suggests that HRQoL scores do not seem to differ
Treatment of Cancer (EORTC) QLQ‐C30 and EORTC
by primary therapy in long‐term survivors diagnosed with localised
QLQ‐PR25
PC (Adam et al., 2018). However, the included studies, which investi‐
gated HRQoL by treatment in localised PC, used different comparison HRQoL and PC‐specific symptom burden were assessed with in‐
groups and included a total of four different treatments. As a result, ternationally validated instruments: European Organisation for
ADAM et al. | 3 of 10

Research and Treatment of Cancer (EORTC) QLQ‐C30 (Aaronson et both HRQoL and treatment choice. Moreover, in order to estimate
al., 1993) and the PC‐specific module QLQ‐PR25 (van Andel et al., the independent unimodal treatment effects on HRQoL in compar‐
2008). Scoring of these questionnaires was performed according to ison with survivors who reported no treatment, we decomposed all
the EORTC scoring manual (Fayers et al., 2001). treatment groups into unimodal index variables and employed a lin‐
ear main effects model with the adjustments described above.
In the case of PC progression or relapse, it was impossible to de‐
2.2.2 | Demographic and clinical data
termine conclusively whether a particular treatment was given due
Cancer registries included in this study provided demographic and to this change in disease status or already at baseline. We therefore
clinical information such as date of birth, date of diagnosis and performed sensitivity analyses, which excluded all participants with
cancer stage. Self‐reported demographic and clinical data included disease progression or relapse.
marital status, education, income, comorbidities, treatment (surgery, We employed the Multiple Imputation Chained Equations (MICE)
chemo‐, radiation and hormonal therapy) and disease progression/ procedure with 25 repetitions to a reduce possible bias due to miss‐
relapse (including biochemical and clinical recurrence and metastasis ing values (e.g., on therapy and comorbidities) (Royston & White,
after diagnosis of primary tumour at time of survey). 2011; White, Royston, & Wood, 2011). Differences in mean HRQoL
scores larger than 10 points were considered clinically meaningful
(Osoba, Rodrigues, Myles, Zee, & Pater, 1998). A p‐value < 0.05
2.3 | Statistical analysis
(two‐sided) was considered statistically significant. The p‐values
Therapy was categorised as: (a) RP (alone) (b) RT (alone), (c) RP & RT were not adjusted for multiple testing, so the p‐values refer to the
(in combination), (d) ADT (& RP/RT) and (e) no treatment reported individual tests rather than a global test for differences. All analyses
(NTX report.). We defined our therapy groups with unordered treat‐ were performed using STATA statistical software (version 13.1).
ment combinations, as detailed information regarding treatment
date was incomplete. Therefore, we could not clearly determine the
3 | R E S U LT S
primary therapy in those patients who reported multimodal therapy,
or for whom treatment was given in the case of PC progression or
3.1 | Description of the study population
relapse. As most patients who received ADT also underwent either
RP or RT (86.2%), and exploratory analyses indicated that patients In total, 911 PC survivors participating in the CAESAR + study cat‐
treated with ADT either alone, or in combination with RP and/or RT, egorised as T1N0M0 or T2N0M0 and without diagnosis of another
reported similar HRQoL (data not shown), we created one group in‐ malignant tumour prior to their PC diagnosis were included for this
cluding all survivors treated with ADT. RT comprises both external analysis. The participation rate for this subsample of survivors with
beam radiation therapy and brachytherapy as the questionnaire did localised PC (911 out of 1,944, 46.9%) was similar to the participa‐
not distinguish between these two treatment options. Finally, the tion rate of all PC survivors within the CAESAR + study. In addition,
questionnaire did not include specific questions on the treatment respondents and non‐respondents eligible for this analysis were sim‐
modalities AS and WW. To cover these PC survivors, we created the ilar with respect to age and stage at diagnosis, but non‐respondents
therapy group “no treatment reported.” were more likely to be ≥10 years post‐diagnosis (p = 0.017, Table S1).
For descriptive purposes, we compared PC survivors by treat‐ Among respondents, RP (unimodal) was the most frequent treat‐
ment group with respect to age and stage at diagnosis, time since ment option (57.8%), followed by ADT (& RP/RT) (14.6%), RT & RP
diagnosis, disease progression/relapse at time of survey, comorbid‐ (10.7%) and RT alone (9.6%) (Table 1). Only 7.4% reported no treat‐
ity status at diagnosis, marital status, income and education level at ment at all. Overall, mean age at diagnosis was 65.1 years and varied
survey using parametric tests. Non‐parametric tests were applied significantly by therapy (p < 0.001). PC survivors treated with RP
when normality and homogeneity assumptions were violated. & RT were the youngest at diagnosis (mean: 63.5 years), and survi‐
Adjusted means based on multivariable linear regression mod‐ vors treated with RT alone were the oldest (mean: 67.9 years). Mean
els were calculated to describe and test for differences in HRQoL. time since diagnosis was 7.1 years and varied from 6.6 years (RT) to
Global F‐tests were performed to assess the association between 7.5 years (RP & RT, p = 0.008). The highest number of comorbidities
treatment scheme and adjusted mean HRQoL score. Models were at diagnosis was reported by survivors treated with RT alone (26.8%
adjusted for education, number of comorbidities at diagnosis, years reported at least one comorbidity). Disease progression/relapse also
since diagnosis and age at diagnosis, as they are known to be con‐ varied significantly (p < 0.001) by therapy and was the lowest for
founders of the association of treatment and HRQoL based on cur‐ survivors with no reported treatment (2.3%) and highest for survi‐
rent literature (core model). Other variables including cancer stage, vors treated with RP & RT (in combination) (42.1%). All other charac‐
family status, income and cancer stage were considered as additional teristics did not significantly differ between therapy groups.
potential confounders but not included as they did not improve the Overall, multiple imputation did not substantially alter the distri‐
model fit (maximum likelihood test with p < 0.1). Comorbidities con‐ bution of the different treatment groups (Table S2) or their associ‐
sidered included stroke, cardiac infarction, angina pectoris and heart ation with baseline demographic and clinical characteristics (Tables
failure, as these comorbidities may be independently associated with S3 and S4).
4 of 10 | ADAM et al.

TA B L E 1 Demographic and clinical characteristics of study population by treatment group (after multiple imputation of missing values)

Total RP RT RP & RT ADT (& RP/RT) NTX report.

100.0% 57.8% 9.6% 10.7% 14.6% 7.4%
n = 911 n = 527 n = 87 n = 97 n = 133 n = 67

Characteristic Col% Col% Col% Col% Col% Col% p‐Valuea

Age at diagnosis (years)

<60 14.4 15.0 7.9 25.3 10.4 10.0
60–69 63.6 68.4 48.3 55.4 57.4 69.1
≥70 22.0 16.6 43.8 19.3 32.2 20.9 <0.001
Mean (SD) 65.1 (5.5) 64.6 (5.0) 67.9 (5.6) 63.5 (7.2) 66.5 (5.7) 65.6 (5.1) <0.001
Education (years)
≤10 71.6 69.3 58.2 67.3 72.5 84.5
>10 28.4 30.7 41.8 32.7 27.5 15.5 0.72
Living with spouse 84.1 86.0 79.6 87.3 80.0 81.6 0.52
Income (EURO)
0–2,000 49.8 43.8 30.2 51.7 56.0 62.9
2,001–2,999 30.5 33.1 50.6 28.3 30.3 24.5
≥3,000 19.7 23.1 19.2 20.0 13.7 12.6 0.05
Tumour Stage
T1N0M0 4.6 2.8 8.8 6.7 4.2 7.9
T2N0M0 95.4 97.2 91.2 93.3 95.8 92.1 0.07
Years since diagnosis
5–6 41.6 43.9 51.5 35.6 32.8 36.5
7–9 47.3 46.1 42.1 46.4 52.9 54.2
≥10 11.1 10.0 6.4 18.0 14.3 9.3 0.024
Mean (SD) 7.1 (1.5) 7.0 (1.5) 6.6 (1.2) 7.5 (2.0) 7.2 (1.4) 7.0 (1.4) 0.008
Comorbidities at diagnosis
0 87.0 88.7 73.2 90.4 85.6 89.4
1 8.3 4.1 19.8 6.4 6.6 9.2
≥2 4.7 7.2 7.0 3.2 7.8 1.4 0.023
Disease progression/relapse 13.2 2.5 12.2 42.1 38.3 2.3 <0.001

Note: ADT, androgen deprivation therapy; NTX report., no treatment reported (according to self‐report); RP, radical prostatectomy; RT, radiotherapy.
a
p‐Values indicate the statistically significant difference between treatments.
b
Bold values indicate statistical significant values.

none of these previously described significant differences in HRQoL

3.2 | QLQ‐C30 global and functional scales
were clinically meaningful. Analyses restricted to complete cases
In general, survivors treated with RP or RT alone reported the best revealed no substantial differences from those based on multiple
HRQoL and survivors treated with ADT (& RP/RT) or RP & RT (in imputation (Table S5).
combination) reported the worst. However, significant differences
among adjusted group means of long‐term PC survivors were only
3.3 | QLQ‐C30 symptom scales
reported for social functioning (p = 0.007) and global health status
(p = 0.041) (Figure 1). Fatigue, pain, dyspnoea and insomnia were the symptoms with the
There was no significant difference in HRQoL among PC survi‐ highest reported burden (Figure 2). In contrast, nausea and vomit‐
vors either unimodally treated with RP or RT. However, HRQoL was ing and appetite loss were the symptoms with the lowest reported
significantly lower for survivors treated with RP & RT (in combina‐ burden. Overall, symptom complaints were lower for PC survivors
tion) in the domains social functioning (p < 0.01) and global health treated with RP alone, RT alone or those who did not report a treat‐
status (p < 0.05) when compared to survivors treated with RP only. ment. In contrast, survivors treated with ADT (& RP/RT) experienced
PC survivors treated with ADT (& RP/RT) scored significantly lower significantly higher symptom burden than patients treated with RP
than survivors treated with RP in emotional (p < 0.05) and social alone for fatigue (p < 0.01), nausea and vomiting (p < 0.001), insom‐
functioning (p < 0.05) and in global health status (p < 0.05). However, nia (p < 0.05), diarrhoea (p < 0.05) and financial difficulties (p < 0.05).
ADAM et al. | 5 of 10

F I G U R E 1 Mean scores of EORTC QLQ‐C30 global and functional scales by treatment group (after multiple imputation of missing
values). A high score represents a high/healthy level of functioning; a high score for the global health status/QoL represents a high QoL.
Scores in parentheses represent p‐values for global F‐tests for comparing differences among group means. Models were adjusted for
education, number of comorbidities at diagnosis, years since diagnosis and age at diagnosis. ADT, androgen deprivation therapy; NTX
report., no treatment reported (according to self‐report); RP, radical prostatectomy; RT, radiotherapy. *p < 0.05, **p < 0.01, ***p < 0.001
statistically significant different mean score from RP. I bars represent ± standard errors

F I G U R E 2 Mean scores of EORTC QLQ‐C30 symptom scales by treatment group (after multiple imputation of missing values). A high
score represents a greater symptom burden. Scores in parentheses represent p‐values for global F‐tests for comparing differences among
group means. Models were adjusted for education, number of comorbidities at diagnosis, years since diagnosis and age at diagnosis. ADT,
androgen deprivation therapy; NTX report., no treatment reported (according to self‐report); RP, radical prostatectomy; RT, radiotherapy.
*p < 0.05, **p < 0.01, ***p < 0.001 statistically significant different mean score from RP. I bars represent ± standard errors

and sexual activity (p < 0.001) differed significantly between

3.4 | QLQ‐PR25 scales
therapy groups (Figure 3). Overall, long‐term survivors treated
Adjusted means for urinary symptoms (p = 0.035), bowel symptoms only with RP or RT reported the fewest problems with urinary,
(p = 0.017), hormonal treatment‐related symptoms (p < 0.001) bowel and hormonal treatment‐related symptoms. Additionally,
6 of 10 | ADAM et al.

F I G U R E 3 Mean scores of EORTC QLQ‐PR25 scales by treatment group (after multiple imputation of missing values). A high score
represents a greater symptom burden or better sexual functioning and activity. Scores in parentheses represent p‐values for global F‐tests
for comparing differences among group means. Models were adjusted for education, number of comorbidities at diagnosis, years since
diagnosis and age at diagnosis. ADT, androgen deprivation therapy; NTX report., no treatment reported (according to self‐report); RP,
radical prostatectomy; RT, radiotherapy. *p < 0.05, **p < 0.01, ***p < 0.001 statistically significant different mean score from RP; ‡clinically
meaningful different mean score from RP. I bars represent ± standard errors. 1Smaller sample sizes as the questions referring to these scales
are conditional

these survivors reported the best sexual activity and function‐ No systematic difference with respect to HRQoL was observed
ing. Conversely, survivors treated with RP & RT (in combination) between survivors with no reported treatment (NTX report.) and
reported a significantly higher burden of urinary (p < 0.01) and survivors treated either with RP or RT. PC survivors treated with
bowel symptoms (p < 0.05). Survivors after ADT (& RP/RT) demon‐ ADT reported statistically significant worse emotional functioning
strated a significantly higher burden of bowel symptoms (p < 0.01) and more problems with fatigue, nausea and vomiting, dyspnoea,
and hormonal treatment‐related symptoms (p < 0.001, clinically financial difficulties, urinary bother, bowel symptoms, hormonal
meaningful) and scored much worse on sexual activity (p < 0.001, treatment‐related symptoms and sexual activity than survivors with
clinically meaningful) when compared to survivors treated with RP no reported treatment.
alone. Survivors treated with RT alone reported the highest bur‐
den of negative symptoms related to sexual functioning (p < 0.05,
3.6 | Sensitivity analyses
clinically meaningful).
Restricting our analyses to long‐term PC survivors without disease
progression/relapse did not substantially alter the pattern of our
3.5 | Independent unimodal treatment‐specific
findings (Table S6).
associations
After partition of multimodal treatment groups into unimodal index
variables, the comparison between RP and RT indicated no signifi‐ 4 | D I S CU S S I O N
cant differences, except for hormonal treatment‐related symptoms
(p = 0.027) and sexual functioning (p = 0.01) (Table 2). The comparison In light of the exceptionally high five‐year relative survival rates of
between RP and ADT groups revealed poorer emotional functioning localised PC survivors and the excellent comparable survival rates
(p = 0.033) and a higher prevalence of several post‐ADT side effects. for various treatment methods, it is imperative to understand more
With respect to functional scales, no significant differences were ob‐ about the effect of therapies on HRQoL in long‐term PC survivors.
served between RT and ADT. However, several general and PC‐spe‐ Our population‐based study conducted in Germany shows that
cific symptoms were more prevalent among patients after ADT. long‐term PC survivors treated with RP or RT alone reported the
ADAM et al. | 7 of 10

TA B L E 2 Independent unimodal treatment‐related HRQoL effects (main effects model) (after multiple imputation of missing values)

RP vs. NTX report. RT vs. NTX report. ADT vs. NTX report. Comparison of β‐estimates

RT vs.
β [95% CI] β [95% CI] β [95% CI] RP vs. RT RP vs. ADT ADT

QLQ‐C30 Global and Functional Scalesa

Global health status/ −1.0 [−5.2, 3.1] −3.5 [−7.0, −0.1] −2.8 [−7.2, 1.5] 0.28 0.56 0.81
QoL
Physical functioning 1.0 [−2.7, 4.7] −1.4 [−4.5, 1.6] −2.4 [−6.5, 1.6] 0.22 0.22 0.72
Role functioning 1.2 [−4.0, 6.4] −1.8 [−6.1, 2.6] −1.8 [−7.7, 4.1] 0.30 0.45 0.99
Emotional 1.8 [−2.4, 5.9] 0.6 [−3.0, 4.2] −4.9 [−9.6, −0.2]* 0.62 0.033* 0.10
functioning
Cognitive −1.9 [−5.9, 2.1] 2.1 [−1.4, 5.5] −0.1 [−4.5, 4.4] 0.08 0.53 0.51
functioning
Social functioning −3.0 [−8.2, 2.2] −1.7 [−6.1, 2.8] −5.7 [−11.3, −0.2] 0.64 0.46 0.30
b
QLQ‐C30 Symptom Scales
Fatigue 0.1 [−4.6, 4.8] 0.9 [−3.0, 4.8] 6.0 [0.6, 11.4]* 0.74 0.11 0.18
Nausea and vomiting −0.8 [−2.6, 1.0] −0.8 [−2.3, 0.7] 3.0 [1.1, 4.8]* 0.99 0.005* 0.006*
Pain 1.3 [−4.4, 6.9] 3.1 [−1.6, 7.7] 2.4 [−3.3, 8.1] 0.55 0.77 0.88
Dyspnoea −0.9 [−6.3, 4.5] −1.5 [−5.9, 3.0] 7.1 [1.2, 13.0]* 0.86 0.045* 0.039*
Insomnia 1.3 [−4.5, 7.0] −0.9 [−5.8, 4.0] 9.1 [−2.9, 15.4] 0.51 0.06 0.027*
Appetite loss −2.1 [−5.2, 1.0] −2.6 [−5.1, 1.7 [−1.4, 4.8] 0.78 0.08 0.06
−0.1]*
Constipation −0.0 [−4.7, 4.6] −2.3 [−6.1, 1.5] 2.0 [−2.8, 6.7] 0.36 0.56 0.21
Diarrhoea 1.4 [−2.5, 5.2] 2.7 [−0.4, 5.8] 2.7 [−1.4, 6.7] 0.53 0.62 0.99
Financial difficulties 0.1 [−4.1, 4.3] −2.8 [−6.3, 0.8] 7.8 [3.4, 12.3]* 0.22 0.012* 0.001*
QLQ‐PR25 Scalesc
Urinary symptoms 3.4 [−0.4, 7.3] 3.2 [−0.1, 6.5] 1.2 [−2.8, 5.2] 0.91 0.43 0.50
Urinary botherd 0.0 [−2.5, 2.3] 0.1 [−1.8, 2.1] 2.5 [0.1, 4.9]* 0.92 0.83 0.89
Bowel symptoms 0.6 [−1.7, 2.9] 2.9 [0.9, 4.9]* 2.7 [0.2, 5.2]* 0.81 0.21 0.90
Hormonal treat‐ 4.5 [1.5, 7.4]* 0.8 [−1.6, 3.2] 10.8 [7.8, 13.9]* 0.027* 0.003* <0.001*
ment‐related
symptoms
Sexual activity 1.9 [−3.3, 7.0] 0.3 [−4.1, 4.6] −11.7 [−17.1, −6.4]* 0.58 <0.001* 0.002*
d
Sexual functioning −9.0 [−10.5, 7.5] 0.6 [−0.7, 1.8] 0.1 [−1.7, 1.8] 0.010* 0.08 0.92

Note: Models were adjusted for education, number of comorbidities at diagnosis, years since diagnosis and age at diagnosis.
Abbreviations: ADT, androgen deprivation therapy; NTX report., no treatment reported (according to self‐report); RP, radical prostatectomy; RT,
radiotherapy.
a
EORTC QLQ‐C30 functional scales: A high score represents a high/healthy level of functioning; a high score for the global health status/QoL rep‐
resents a high QoL.
b
EORTC QLQ‐C30 symptom scales/items: A high score represents a greater burden.
c
EORTC QLQ‐PR25 scales: A high score represents a greater burden or better sexual functioning and activity.
d
Smaller sample sizes as the questions referring to these scales are conditional.
*Indicates whether a difference is statistically significant.
f
Bold values indicate statistical significant values.

best HRQoL and lowest symptom burden and survivors treated study PC survivors treated with external beam radiation therapy
with ADT (& RP/RT) or RP & RT (in combination) reported the worst reported statistically significant lower physical functioning and PC
HRQoL and highest symptom burden. survivors treated with brachytherapy reported significantly higher
The similarity of HRQoL scores among survivors unimodally physical, role and social functioning than PC survivors treated with
treated either with RP or RT is comparable to other studies among RP alone. These different findings in comparison with our results
long‐term PC survivors (Donovan et al., 2016; Giberti, Chiono, Gallo, are most likely explained by differences in treatment categories as
Schenone, & Gastaldi, 2009; Mols et al., 2006; Shinohara et al., 2013), the study by Drummond et al. could differentiate between PC survi‐
except for the study of Drummond et al. (2015). For example, in this vors receiving external beam radiation or brachytherapy. Moreover,
8 of 10 | ADAM et al.

distribution of stage at diagnosis and phase of follow‐up were also with ADT. Even though ADT might extend survival of PC survivors
different compared to our study as they also included patients with (Helgstrand, Berg, Lippert, Brasso, & Røder, 2016), this potential gain
stage III and IV and short‐term survivors. in survival might come at the expense of a significant loss in HRQoL.
Of particular interest is the similarity of PC‐specific symptom bur‐ Some limitations must be considered when interpreting our
den scores among PC survivors treated with RP or RT alone. This result results. First, despite the generally good response rate of 46.9%,
is in line with the randomised clinical trial (RCT) conducted by Giberti there is a possibility of selection bias as survivors with longer time
et al. (2009), but conflicts with the results of the RCT conducted by period since diagnosis were less likely to participate. In addition,
Donovan et al. (2016) and the results of four observational studies it has been reported that participants and non‐participants might
(Hoffman et al., 2004; Resnick et al., 2013; Shinohara et al., 2013; differ with respect to health status (de Rooij et al., 2018). Second,
Smith et al., 2009). All of the afore studies mentioned found signifi‐ no baseline HRQoL data from this cohort exist, making it impos‐
cant differences in symptom burden between long‐term PC survivors sible to identify changes over time and to adjust for the baseline
treated either with RP or RT alone. In general, RP had a stronger neg‐ HRQoL. Third, we were not able to unambiguously determine pri‐
ative effect on urinary symptoms and sexual function (Donovan et al., mary treatment, to distinguish between external beam radiation
2016; Resnick et al., 2013; Shinohara et al., 2013; Smith et al., 2009), therapy and brachytherapy, to define the duration of treatment
whereas RT resulted in an increased negative bowel symptom burden and to determine whether patients initially assigned to AS/WW
(Resnick et al., 2013; Smith et al., 2009). The discrepancy between our received ADT, RT and/or RP at a later date. Fourth, our informa‐
results and the afore mentioned studies could be related to a different tion on comorbidities, disease progression/relapse and treatment
length of follow‐up. Most of the studies above‐cited used a follow‐up choice is based on self‐reported data and is therefore at risk of
time of five years (Donovan et al., 2016; Giberti et al., 2009; Hoffman recall bias. However, our treatment information is comparable to
et al., 2004; Mols et al., 2006; Resnick et al., 2013; Shinohara et al., treatment data reported for Germany (Dörr, Hölzel, Schubert‐
2013), whereas the follow‐up time in our study was up to 15 years. Fritschle, Engel, & Schlesinger‐Raab, 2015). Additionally, our over‐
Resnick et al. (2013) support this hypothesis. Despite a difference in all disease progression/relapse rate is comparable to other studies
disease‐specific functional outcomes at two and five years after treat‐ (Han, Partin, Piantadosi, Epstein, & Walsh, 2001). The high disease
ment, they could not observe any differences after 15 years of follow‐ progression/relapse rate in the RP & RT group is probably based on
up. Moreover, our cohort and the cohort of Giberti et al. (2009) were the fact that most patients received salvage RT due to relapse/dis‐
older than the cohorts of the other studies, and under‐reporting of ease progression. Also, other studies reported up to 40% biochem‐
cancer related symptoms may result as part of the “normalisation pro‐ ical recurrence rates after RP for localised PC survivors (Hull et al.,
cess”(Walker, Szanton, & Wenzel, 2015). "Normalisation” may occur 2002; Izawa, 2009; Kupelian, Katcher, Levin, Zippe, & Klein, 1996;
in older cancer survivors perceiving cancer and its side effects as ex‐ Stephenson et al., 2006). However, to account for the bias that we
pected consequences of normal ageing. Moreover, older individuals could not conclusively determine whether a particular treatment
may suffer from additional chronic conditions (such as cardiovascular was given due to this change in disease status or already at baseline,
disease, diabetes and arthritis) that may have a stronger impact on sur‐ we performed sensitivity analyses. In this analysis, the pattern of
vivors’ HRQoL than cancer or treatment‐related issues (Cohen, 2006; results was not substantially altered, suggesting that this potential
Sinding & Wiernikowski, 2008; Walker et al., 2015). bias is not relevant. Fifth, although we included patients with low
Our finding that PC survivors treated with a combination of RP cancer stage, we cannot exclude the possibility of some misclassi‐
and RT report lower HRQoL and a higher PC‐specific symptom bur‐ fication, as we have no data on Gleason scores and PSA values for
den than survivors treated with RP only most likely reflects a cumu‐ these patients. Sixth, although we corrected for a range of demo‐
lative treatment effect rather than an effect of disease progression/ graphic and clinical factors, the risk of residual confounding is pos‐
relapse and subsequent therapy since this finding is also observed sible as we do not have information on factors related to treatment
among progression‐free survivors. However, in the ADT group, dis‐ (e.g., on digital rectal exam results, PSA values, Gleason scores) or
ease progression/relapse might explain the finding that these survi‐ psychological distress (e.g., fear of progression). Seventh, because
vors reported significantly more financial difficulties than survivors this data set includes PC survivors diagnosed from 1996 to 2004,
treated with RP (unimodal). In addition, long‐term PC survivors new treatment options such as robotic surgery, intensity‐modu‐
treated with ADT experience more side effects than PC survivors lated RT and AS are not part of this study. Moreover, management
treated with RP only and may therefore face more financial difficul‐ strategies for localised PC patients changed over the years, leading
ties due to problems staying at or returning to work. Moreover, our to potentially atypical treatment patterns nowadays, for example
study suggests that ADT‐related adverse effects, such as decreases in the groups with multimodal treatment patterns. Despite this lim‐
in bone mineral density, weight gain, hot flashes, decreased libido and ited transferability of our results to current PC patients, the results
sexual dysfunction, gynecomastia and fatigue (Nguyen et al., 2015), of this study augment the knowledge regarding health problems
have a more profound impact on the symptom burden and HRQoL in in the large group of long‐term PC survivors. Finally, due to the
long‐term PC survivors than the side effects of all other therapies or explorative character of this study, we tested multiple regression
therapy combinations. Therefore, healthcare providers and PC sur‐ models without adjusting p‐values. Multiple testing can lead to an
vivors should carefully balance the harms and benefits of treatment increased risk of type I errors, so that some results found in this
ADAM et al. | 9 of 10

study could be false positive. However, this applies mainly to soli‐ Dörr, M., Hölzel, D., Schubert‐Fritschle, G., Engel, J., & Schlesinger‐
tary findings rather than findings with a systematic pattern. Raab, A. (2015). Changes in prognostic and therapeutic param‐
eters in prostate cancer from an epidemiological view over 20
Nevertheless, there are also several strengths of this study. Our
years. Oncology Research and Treatment, 38(1–2), 8–14. https​://doi.
population‐based design, including patient recruitment via multiple org/10.1159/00037​1717
population‐based cancer registries, results in a socio‐demograph‐ Drummond, F. J., Kinnear, H., O’Leary, E., Donnelly, N., Gavin, A., &
ically diverse cohort with patients included who were not exclu‐ Sharp, L. (2015). Long‐term health‐related quality of life of pros‐
tate cancer survivors varies by primary treatment. Results from the
sively treated at large academic centres. Additionally, our study
PiCTure (Prostate Cancer Treatment, your experience) study. Journal
adds valuable information on the effect of ADT on HRQoL for long‐ of Cancer Survivorship, 9(2), 361–372. https​ ://doi.org/10.1007/
term survivors diagnosed with localised PC. Until now, only limited s11764-014-0419-6
information was available for this specific group (Adam et al., 2018). Etzioni, R., Penson, D. F., Legler, J. M., di Tommaso, D., Boer, R., Gann, P.
H., & Feuer, E. J. (2002). Overdiagnosis due to prostate‐specific an‐
Finally, this is the first study of this kind performed in Germany.
tigen screening: Lessons from U.S. prostate cancer incidence trends.
In conclusion, this study suggests that survivors diagnosed with Journal of the National Cancer Institute, 94(13), 981–990. https​://doi.
localised PC and treated either with RP or RT alone experience the org/10.1093/jnci/94.13.981
best HRQoL and lowest symptom burden. Conversely, survivors Fayers, P., Aaronson, N., Bjordal, K., Groenvold, M., Curran, D., &
Bottomley, A. (2001). The EORTC QLQ‐C30 Scoring Manual, 3rd ed.
treated with ADT (without or in combination with RP/RT) have the
(p. 3). Brussels, Belgium: European Organisation for Research and
largest negative impact on HRQoL and highest symptom burden five
Treatment of Cancer.
and more years after diagnosis. Freedland, S. J. (2011). Screening, risk assessment, and the approach to
therapy in patients with prostate cancer. Cancer, 117(6), 1123–1135.
https​://doi.org/10.1002/cncr.25477​
ORCID Giberti, C., Chiono, L., Gallo, F., Schenone, M., & Gastaldi, E. (2009).
Radical retropubic prostatectomy versus brachytherapy for low‐risk
Salome Adam https://orcid.org/0000-0002-1829-919X prostatic cancer: A prospective study. World Journal of Urology, 27,
607–612. https​://doi.org/10.1007/s00345-009-0418-9
Volker Arndt https://orcid.org/0000-0001-9320-8684
Hamdy, F. C., Donovan, J. L., Lane, J. A., Mason, M., Metcalfe, C., Holding,
P., … Neal, D. E. (2016). 10‐Year outcomes after monitoring, surgery,
or radiotherapy for localized prostate cancer. New England Journal
REFERENCES of Medicine, 375, 1415–1424. https​://doi.org/10.1056/NEJMo​
a1606220
Aaronson, N. K., Ahmedzai, S., Bergman, B., Bullinger, M., Cull, A., Duez, Han, M., Partin, A., Piantadosi, S., Epstein, J., & Walsh, P. (2001). Era
N. J., … de Haes, J. C. (1993). The European Organization for Research specific biochemical recurrence‐free survival following radical pros‐
and Treatment of Cancer QLQ‐C30: A quality‐of‐life instrument for tatectomy for clinically localized prostate cancer. The Journal of
use in international clinical trials in oncology. Journal of the National Urology, 166(2), 416–419.
Cancer Institute, 85(5), 365–376. Heidenreich, A., Bastian, P. J., Bellmunt, J., Bolla, M., Joniau, S., Van Der
Adam, S., Feller, A., Rohrmann, S., & Arndt, V. (2018). Health‐re‐ Kwast, T., & Mottet, N. (2014). EAU guidelines on prostate cancer.
lated quality of life among long‐term (≥5 years) prostate cancer Part 1: Screening, diagnosis, and local treatment with curative in‐
survivors by primary intervention: A systematic review. Health tent – Update 2013. European Urology, 65(1), 124–137. https​://doi.
and Quality of Life Outcomes, 16(1), 22. https​://doi.org/10.1186/ org/10.1016/j.eururo.2013.09.046
s12955-017-0836-0 Helgstrand, J. T., Berg, K. D., Lippert, S., Brasso, K., & Røder, M. A.
American Cancer Society (2015). Global Cancer Facts & Figures. American (2016). Systematic review: Does endocrine therapy prolong
Cancer Society, 3(700), 1–57. https​://doi.org/10.1002/ijc.27711​ survival in patients with prostate cancer? Scandinavian Journal
American Cancer Society (2000). Cancer Facts & Figures—2000. Atlanta, of Urology, 50(3), 135–143. https​://doi.org/10.3109/21681​
GA: American Cancer Society. 805.2016.1142472
Arndt, V., Koch‐Gallenkamp, L., Jansen, L., Bertram, H., Eberle, A., Hoffman, R. M., Gilliland, F. D., Penson, D. F., Stone, S. N., Hunt, W. C., &
Holleczek, B., … Brenner, H. (2017). Quality of life in long‐term Potesky, A. L. (2004). Cross‐sectional and longitudinal comparisons
and very long‐term cancer survivors versus population con‐ of health‐related quality of life between patients with prostate carci‐
trols in Germany. Acta Oncologica, 56(2), 190–197. https​ ://doi. noma and matched controls. Cancer, 101(9), 2011–2019. https​://doi.
org/10.1080/02841​86X.2016.1266089 org/10.1002/cncr.20608​
Cohen, H. J. (2006). Keynote comment: Cancer survivorship and age‐ Holmberg, L., Bill‐Axelson, A., Helgesen, F., Salo, J. O., Folmerz, P.,
ing—a double whammy. Lancet Oncology, 7, 882–883. https​://doi. Häggmann, M. … Norlen, B. J. (2002). A Randomized Trial comparing
org/10.1016/j.critr​evonc.2006.06.003.Yabroff radical prostatectomy with watchful waiting in early prostate cancer.
de Rooij, B. H., Ezendam, N. P. M., Mols, F., Vissers, P. A. J., Thong, M. New England Journal of Medicine, 347(11), 781–789.
S. Y., Vlooswijk, C. C. P., … Franse, L. V. (2018). Cancer survivors not Hull, G., Rabbani, F., Abbas, F., Wheeler, T., Kattan, M., & Scardino, P.
participating in observational patient‐reported outcome studies have (2002). Cancer control with radical prostatectomy alone in 1,000
a lower survival compared to participants: The population‐based consecutive patients. The Journal of Urology, 167(2), 528–534.
PROFILES registry. Quality of Life Research, 27(12), 3313–3324. https​ Izawa, J. I. (2009). Salvage radiotherapy after radical prostatectomy.
://doi.org/10.1007/s11136-018-1979-0 Journal of the Canadian Urological Association, 3(3), 245–250. https​://
Donovan, J. L., Hamdy, F. C., Lane, J. A., Mason, M., Metcalfe, C., Walsh, doi.org/10.1200/JCO.2009.24.2396
E., … Neal, D. E. (2016). Patient‐reported outcomes after monitoring, Kupelian, P., Katcher, J., Levin, H., Zippe, C., & Klein, E. (1996). Correlation
surgery, or radiotherapy for prostate cancer. New England Journal of clinical and pathologic factors with rising prostate‐specific anti‐
of Medicine, 375(15), 1425–1437. https​://doi.org/10.1056/NEJMo​ gen profiles after radical porstatectomy alone for clinically localized
a1606221 prostate cancer. Adult Urology, 48, 249–260.
10 of 10 | ADAM et al.

Miller, K. D., Siegel, R. L., Lin, C. C., Mariotto, A. B., Kramer, J. L., Rowland, Smith, D. P., King, M. T., Egger, S., Berry, M. P., Stricker, P. D., Cozzi, P.,
J. H., & Jemal, A. (2016). Cancer treatment and survivorship statis‐ … Armstrong, B. K. (2009). Quality of life three years after diagnosis
tics, 2016. CA: A Cancer Journal for Clinicians, 66(4), 271–289. https​:// of localised prostate cancer: Population based cohort study. British
doi.org/10.3322/caac.21349​ Medical Journal, 339, 12. https​://doi.org/10.1136/bmj.b4817​
Mols, F., Van De Poll‐Franse, L. V., Vingerhoets, A. J. J. M., Hendrikx, A., Stattin, P., Holmberg, E., Johansson, J. E., Holmberg, L., Adolfsson, J., &
Aaronson, N. K., & Houterman, S. … Essink‐Bot, M. L. (2006). Long‐ Hugosson, J. (2010). Outcomes in localized prostate cancer: National
term quality of life among Dutch prostate cancer survivors: Results prostate cancer register of Sweden follow‐up study. Journal of the
of a population‐based study. American Cancer Society, 107(9), 2186– National Cancer Institute, 102(13), 950–958. https​://doi.org/10.1093/
2196. https​://doi.org/10.1002/cncr.22231​ jnci/djq154
National Cancer Institute (2018). Cancer trends progress report. Retrieved Stephenson, A. J., Scardino, P. T., Eastham, J. A., Bianco, F. J. Jr, Dotan,
August 29, 2018, from https​://progr​essre​port.cancer.gov/after/​survi​ Z. A., Fearn, P. A., & Kattan, M. W. (2006). Preoperative nomogram
val#field_most_recent_estim​ates predicting the 10‐year probability of prostate cancer recurrence
National Comprehensive Cancer Network (2014). NCCN Guidelines for after radical prostatectomy. Journal of the National Cancer Institute,
Treatment of Prostate Cancer. Retrieved October 7, 2018, from https​ 98(10), 715–717. https​ ://doi.org/10.1099/mic.0.2008/022293-0.
://www.nccn.org/profe​ssion​als/physi​cian_gls/defau​lt.aspx#site Quant​itative
Nguyen, P. L., Alibhai, S. M. H., Basaria, S., D’Amico, A. V., Kantoff, P. W., The Union for International Cancer Control's (2019). The Global Cancer
Keating, N. L., … Smith, M. R. (2015). Adverse effects of androgen Observatory ‐ GLOBOCAN 2018. Retrieved March 18, 2019, from
deprivation therapy and strategies to mitigate them. European Urology, https​://www.uicc.org/new-global-cancer-data-globo​c an-2018
67(5), 825–836. https​://doi.org/10.1016/j.eururo.2014.07.010 van Andel, G., Bottomley, A., Fosså, S. D., Efficace, F., Coens, C.,
Osoba, D., Rodrigues, G., Myles, J., Zee, B., & Pater, J. (1998). Interpreting Guerif, S., & Aaronson, N. K. (2008). An international field study of
the significance of changes in health‐related quality‐of‐life scores. the EORTC QLQ‐PR25: a questionnaire for assessing the health‐
Journal of Clinical Oncology, 16(1), 139–144. related quality of life of patients with prostate cancer. European
Parker, C., Gillessen, S., Heidenreich, A., & Horwich, A. (2015). Cancer Journal of Cancer, 44(16), 2418–2424. https​://doi.org/10.1016/j.
of the prostate: ESMO Clinical Practice Guidelines for diagnosis, ejca.2008.07.030
treatment and follow‐up. Annals of Oncology, 26, v69–v77. https​:// Walker, R., Szanton, S. L., & Wenzel, J. (2015). Working toward normalcy
doi.org/10.1093/annon​c/mdv222 post‐treatment: A qualitative study of older adult breast and prostate
Resnick, M. J., Koyama, T., Fan, K.‐H., Albertsen, P. C., Goodman, M., cancer survivors. Oncology Nursing Forum, 42(6), E358–E367. https​://
Hamilton, A. S., … Penson, D. F. (2013). Long‐term functional out‐ doi.org/10.1188/15.ONF.E358-E367
comes after treatment for localized prostate cancer. New England White, I. R., Royston, P., & Wood, A. M. (2011). Multiple imputation using
Journal of Medicine, 368(5), 436–445. https​://doi.org/10.1056/ chained equations: Issues and guidance for practice. Statistics in
NEJMo​a1209978 Medicine, 30(4), 377–399. https​://doi.org/10.1002/sim.4067
Robert Koch Institut (2017). Zentrum für Krebsregisterdaten. Retrieved Wilt, T. J., Jones, K. M., Barry, M. J., Andriole, G. L., Culkin, D.,
August 29, 2018, from https​ ://www.krebs​ daten.de/Krebs/​ DE/ Wheeler, T., … Brawer, M. K. (2017). Follow‐up of prostatectomy
Home/homep​age_node.html versus observation for early prostate cancer. New England Journal
Royston, P., & White, I. R. (2011). Multiple Imputation by Chained of Medicine, 377(2), 132–142. https​://doi.org/10.1056/NEJMo​
Equations (MICE): Implementation in Stata. Journal of Statistical a1615869
Software, 45(4), 1–20. https​://doi.org/10.18637/​jss.v069.i12
Serrell, E. C., Pitts, D., Hayn, M., Beaule, L., Hansen, M. H., & Sammon, J.
D. (2017). Review of the comparative effectiveness of radical pros‐ S U P P O R T I N G I N FO R M AT I O N
tatectomy, radiation therapy, or expectant management of localized
prostate cancer in registry data. Urologic Oncology: Seminars and Additional supporting information may be found online in the
Original Investigations, 36, 183–192. https​://doi.org/10.1016/j.urolo​ Supporting Information section at the end of the article.
nc.2017.10.003
Shinohara, N., Maruyama, S., Shimizu, S., Nishioka, K., Abe, T., C‐
Hatanaka, K. … Shirato, H. (2013). Longitudinal comparison of quality
How to cite this article: Adam S, Koch‐Gallenkamp L,
of life after real‐time tumor‐tracking intensity‐modulated radiation
therapy and radical prostatectomy in patients with localized prostate Bertram H, et al. Health‐related quality of life in long‐term
cancer. Journal of Radiation Research, 54(6), 1095–1101. https​://doi. survivors with localised prostate cancer by therapy—Results
org/10.1093/jrr/rrt049 from a population‐based study. Eur J Cancer Care.
Sinding, C., & Wiernikowski, J. (2008). Disruption foreclosed: Older
2019;e13076. https​://doi.org/10.1111/ecc.13076​
women's cancer narratives. Health, 12(3), 389–411. https​://doi.
org/10.1177/13634​59308​090055