A national study of adverse effects and global quality of life among candidates for curative treatment for prostate cancer


Correspondence: Anne E. Kyrdalen, National Resource Centre for Late Effects after Cancer Treatment, Department of Oncology, Oslo University Hospital, P.O. Box 4953 Nydalen, 0424 Oslo, Norway.e-mail: aed@ous-hf.no


What's known on the subject? and What does the study add?

  • Studies comparing adverse effects after different treatment methods are few and lack information on additional treatment and cancer relapse. Some US studies have provided population-based estimates on adverse effects, however, figures from Europe are most often based on mono-institutional experience or multicentre studies from high-volume university-affiliated hospitals with selected patient populations. Few studies have investigated the relationship between presence of urinary, bowel or sexual dysfunction and global quality of life (QoL).
  • This population-based study investigated real-life practice with regard to prostate cancer in men who were potential candidates for curative treatment, based on high-quality national registry data. Treatment groups of recurrence-free men who completed their treatment or who received no cancer treatment were measured as to adverse effects, global QoL and the association between them. The study shows that each treatment method is associated with distinct patterns of urinary, bowel and sexual dysfunction and that irritative–obstructive urinary symptoms are associated with increased risk of low global QoL. This large survey also investigates the use of medication for erectile dysfunction and the relationship between such use and global QoL.


  • To provide population-based estimates of typical adverse effects (AEs), e.g. urinary, bowel and sexual dysfunction, in patients with non-metastatic recurrence-free prostate cancer (PCa) by curative treatment method, including no treatment.
  • To describe associations between typical AEs and global quality of life (QoL) and to study patients' use of medication for erectile dysfunction (EDmed) and the relationship between such use and global QoL.

Patients and Methods

  • In October 2006 a national population-based sample of PCa survivors diagnosed in 2004 was invited to a postal survey focusing on treatment-related AEs and global QoL, 12–32 months after treatment start. All had completed their initial treatment.
  • In the present study, 771 compliers were categorized into four groups of localized or locally advanced PCa related to the treatment they completed: (i) no treatment; (ii) radical prostatectomy (RP); (iii) radiotherapy (RAD) without hormones; and (iv) RAD with hormone therapy of 3–24 months duration.
  • Measurement of AEs was restricted to function, using selected items from the 50-item Expanded Prostate Cancer Index Composite and the Brief Sexual Function Inventory among others, whereas global QoL was measured with the 12-item short-form health survey.
  • National prescription data enabled assessment of adjuvant hormone application and EDmed use.


  • Men who had undergone RP reported more urinary incontinence (24%) than the other treatment groups, but had the lowest level of moderate/severe urinary irritative–obstructive symptoms. Men from the ‘no treatment’ group had the highest level of moderate/severe irritative–obstructive urinary symptoms. Men who had undergone RAD reported higher levels of irritative intestinal symptoms and faecal leakage compared with the RP group and the no treatment group.
  • In all treatment groups, poor sexual drive and poor erectile function were common AEs, with men treated with RP reporting the highest prevalence of poor erectile function (89%).
  • Presence of irritative–obstructive urinary symptoms and poor sexual drive were independently associated with low global QoL in multivariate analyses.
  • Fifty percent of the study group had used EDmed after treatment start, but only 47% of them were still using EDmed at the time of the survey. Use of EDmed was not significantly associated with global QoL.


  • PCa survivors after curative treatment, but also patients without any anticancer therapy, report high levels of urinary and sexual AEs.
  • Irritative–obstructive urinary symptoms and poor sexual drive were significantly associated with low global QoL, whereas erectile function and use of EDmed were not.

adverse effects


prostate cancer


quality of life


erectile dysfunction


medication for ED


radical prostatectomy




Brief Sexual Function Inventory


androgen deprivation therapy


phosphodiesterase type 5


Norwegian Prescription Database


Cancer Registry of Norway


urinary incontinence


physical component summary


mental component summary


18-item Eysenck Personality Questionnaire


Erectile, urinary and bowel dysfunction are typical adverse effects (AEs) after local treatment for prostate cancer (PCa), in the form of radical prostatectomy (RP) and radiotherapy (RAD), although the published estimates vary considerably [1-8]. Missing information about additional androgen deprivation therapy (ADT) and relapse make interpretation of these findings challenging. The prevalence of AEs depends on treatment method, baseline characteristics, time elapsed since treatment and assessment methodology, which eventually should separate function from bother [9]. After RP, most AEs develop immediately after treatment, with gradual improvement during the first year, whereas AEs after RAD develop more slowly. Maximum recovery is usually achieved during the first 2 post-treatment years [3]. Simultaneous use of adjuvant ADT decreases sexual function and leads to increasing fatigue [6]. Further, some of the symptoms described as treatment-related AEs may be experienced by patients with PCa who never had treatment, such as erectile dysfunction (ED) and urinary urgency and frequency [10]. In Europe, reports on typical AEs are usually based on mono-institutional experience or multicentre studies, performed at high-volume university-affiliated hospitals. Estimates published on AEs include patients with recurrence and additional treatment. It is debatable to what extent such studies can serve as a basis for counselling regarding typical AEs in unselected patients.

Global quality of life (QoL) describes physical and mental health status as reported by the patient, separate from typical AEs. Although AEs are reported as bothersome to patients, they are not always associated with reduced global QoL when other factors such as comorbidity and age are taken into account [11]. Further, results from cross-cultural studies indicate that cultural and national differences influence such associations [12, 13]. In Norwegian patients with PCa, for example, sexual function was not significantly associated with global QoL after radiotherapy [14], contrary to the findings in a US study [4]. More information is therefore needed regarding the relationship between global QoL and typical AEs in patients in different cultures, preferably derived from population-based studies.

Lastly, neuroticism is a personality trait that is closely linked to health perception [15] and should be accounted for in this context. Of the personality traits presented in the five-factor model of Costa and McCrae [16], neuroticism has proven to be robustly correlated to several mental and physical health outcomes [17]. An individual's degree of neuroticism develops from early childhood and remains relatively stable after young adulthood. Elderly men facing the physical and mental health challenges from a PCa diagnosis do so with their established degree of neuroticism. In previous studies from our research group, neuroticism has proven to be a relevant variable when assessing symptoms (e.g. sexual bother) and global QoL [18, 19].

Erectile dysfunction is one of the most frequently reported AEs after treatment for PCa. Treatment for ED is available such as the use of phosphodiesterase type 5 (PDE5) inhibitors. In a US study, about half of patients with PCa treated for localized disease tried/received treatment for ED once or more during a 5-year follow-up period [20]. Differences in the use of PDE5 inhibitors may exist between different cultures [21], and may be related to reimbursement possibilities for these expensive drugs. In Norway, medication for ED (EDmed) has to be paid for by the patients themselves. Use of PDE5 inhibitors has been shown to improve functional and psychosocial aspects of sexual life and may therefore be important for global QoL [22]. Thus, the question is open as to what proportions of patients with PCa in a population-based sample use these agents and whether EDmed use is related to global QoL.

With this background our population-based study of patients with non-metastatic PCa who had completed their planned intervention (no treatment, RP, RAD without ADT or RAD with ADT) had two aims: (i) to estimate the prevalence of typical AEs and low global QoL in relation to treatment. Based on the available literature, we hypothesized, at a median period of 23 months after local treatment, that fewer patients in the RAD group than in the RP without ADT group would experience sexual and urinary AEs, but that more of them would report bowel AEs. Further, we expected adjuvant ADT to increase the prevalence of ED, without significant impact on bowel or urinary AEs; (ii) to explore, independently of treatment group, the association between typical AEs and low global QoL. In a sub-analysis we investigated the use of EDmed and the relationship between such use and global QoL.

Patients and Methods

Data Sources

Based on the unique personal identification number assigned to each Norwegian citizen, data were obtained by merging two population-based registries, the Norwegian Prescription Database (NorPD) and the Cancer Registry of Norway (CRN) with its radiotherapy and PCa registries. Candidates for curative treatment were identified among patients diagnosed with PCa in 2004 (T1–3 N0-XM0 category, PSA level ≤100 ng/mL, any Gleason score, age ≤75 years, Eastern Cooperative Oncology Group performance status 0–1, no other cancer and no major pretreatment comorbidity reported on the registration form to the CRN in the optional field for comments) [23].


Information about RP and RAD was obtained from the CRN. The decision on whether or not to apply curative treatment depends upon the patient's preference and the doctor's evaluation of the medical situation, with ‘no curative treatment’ being one of the options. The pattern of care for men who were to receive curative treatment for PCa in Norway in 2004 is summarized as follows based on D'Amico risk groups [23, 24]: low risk: RP or RAD without hormones; intermediate risk: RP or RAD with ADT of variable duration depending on patient preference, but not longer than 2 years; and high risk: RAD with 2–3 years ADT duration.

In October 2006 a national cross-sectional survey of all Norwegian PCa survivors diagnosed during 2004 was conducted. Survivors were invited to complete a questionnaire concerning global QoL, AEs, lifestyle and psychosocial issues. The present study covers responding patients with PCa and their eligibility criteria are shown in Fig. 1.

Figure 1.

Attrition analysis leading from all 2193 complying patients to the 771 eligible candidates for curative treatment (yellow fields). The grey fields cover patients who are ineligible owing to medical or therapeutic reasons. The red fields reflect study-related eligibility criteria.

The following groups were defined based on the patient's local treatment. Group 1: Patients who did not receive any form of anti-cancer treatment (NoTreat group). For this group the date of diagnosis served as ‘treatment start’. Group 2: RP as monotherapy (RP group). Except for the date of RP (‘treatment start’) no information was available about the operating procedure. Group 3: Definitive radiation therapy without any additional treatment (RADNoHT group). Group 4: Definitive radiation therapy combined with (neo-)adjuvant ADT, the last 3-month LHRH depot injection applied at least 6 months before the survey (RADHT group). All irradiated patients received a target dose of ≥70 Gy to the prostate. ‘Treatment start’ was defined as the day of the first session of radiation.


Patients were divided into risk categories using a slightly modified D'Amico risk assessment where T2x was categorized as T2a, and all T3 cancers were allocated to the high-risk group [23]. Age was included as a continuous variable. A paired relationship was defined as patients who were married or cohabiting. Higher education was defined as >12 years of education. Comorbidity at the time of the survey implied the report of at least two of 18 comorbid conditions listed in the questionnaire, such as myocardial infarction and diabetes.

The survey questionnaire was developed in 2005 and, after discussion with experienced oncologists and urologists, it was decided to use selected items from published questionnaires and supplementary study-specific ad hoc questions. For the present study only items concerning function were selected and AEs were divided into clinically meaningful groups.

The IPSS [25] measures irritative–obstructive urinary symptoms. The IPSS categories used were: none/mildly symptomatic (0–7), moderately symptomatic (8–19) and severely symptomatic (20–35) [26], with further dichotomization (0: no symptoms/mildly symptomatic and 1: moderately/severely symptomatic). Internal consistency for IPSS was 0.75.

Urinary incontinence (UI) was defined as ‘leaking urine at least once a day’, categorizing the degree of leakage into ‘dribbling’ vs ‘more than just dribbling’.

Irritative intestinal symptoms, measured with selected questions from the 50-item Expanded Prostate Cancer Index Composite [27], were defined as experiencing one or more of the following problems: (i) defaecation three times a day or more; (ii) about half the time or more often diarrhoea, blood or mucus in stool or painful defaecation; (iii) at least two times a week cramps or faecal urgency.

Presence of faecal leakage was defined as this occurring at least once a week.

Sexual function was evaluated with the validated instrument Brief Sexual Function Inventory (BSFI) [28, 29]. Only the drive and erection ratings from BSFI were used, ranging from 0 to 8 and 0 to 12, respectively. Higher scores indicated better function. Sexual drive rated as ≤3 was defined as ‘caseness/poor’, as was erectile function rated ≤7 [30]. The questionnaire did not specify if sexual function was reported with or without the use of EDmed. Internal consistencies for drive and erection were 0.85 and 0.94, respectively.

Global QoL was measured by the physical (PCS) and mental component summary (MCS) scores of the 12-item short-form health survey [31] validated in a Norwegian population [32]. Internal consistency was 0.77 for PCS and 0.77 for MCS. The scores of PCS and MCS were T-transformed with a mean score of 50 for the general population and a SD of 10. Low global QoL was defined as PCS and/or MCS of ≤40.

As symptom-reporting and physical complaints may be influenced by a person's neuroticism, we also measured this personality trait. Neuroticism describes if a person generally feels anxious or safe. People with high neuroticism are likely to interpret even normal situations as problematic or threatening [15]. An abbreviated version of the 18-item Eysenck Personality Questionnaire (EPQ-18) measures neuroticism [33, 34]. Six items comprise this personality trait, which ranges from feeling safe (score 0) to feeling quite nervous (score 6). Internal consistency for the EPQ-18 neuroticism was 0.73. Three categories were defined: low neuroticism: score 0–1; moderate neuroticism: score 2–3; and high neuroticism: score 4–6.

Prescribed EDmed (PDE5 inhibitors, alprostadil urethral sticks and papaverin injections) were identified using the NorPD database. Patients who had no records of prescription of EDmed between treatment and survey were defined as ‘never users’. Men with any EDmed prescription between treatment start and survey were defined as ‘ever users’, independent of the number and amount of prescribed medication. Ever users were further divided into those who reported use of EDmed during the last 4 weeks before survey (‘still users’) and those who did not (‘discontinued users’). All but eight men (96%) who stated use of EDmed in the self-report questionnaire were identified in the NorPD data.


The dataset was described with mean and SD for continuous, normally distributed variables and with median and range for variables with skewed distributions. Categorical variables were described with proportions and percentages. Crude associations between pairs of variables were assessed using t-tests, Mann–Whitney–Wilcoxon tests, chi-squared tests and one-way ANOVA. Adjusted associations between pairs of variables in different treatment groups were explored with logistic regression analyses. Internal consistencies of scales were given by Cronbach's coefficient α. Factor analyses were not performed.

Univariate and multivariate associations between typical AEs and low global QoL were explored with logistic regression analyses. Owing to limitations of statistical power it was not possible to include all AEs and significant confounders in one logistic regression model. A separate multivariate regression analysis was performed for each AE to investigate its association with low global QoL. Each sub-analysis was adjusted for age, education, comorbidity and level of neuroticism. The associations between AEs and low global QoL were assumed to be similar for all treatment methods, and therefore all treatment groups were combined in the regression analyses (treatment group was not included as a variable).

All tests were two-sided and the level of significance was set at P < 0.01 owing to multiple testing. The dataset was analysed using PASW version 18 (IBM, Chicago, IL, USA).


The study was approved by the Regional Ethics Committee of Southern Norway.


After one reminder, 2193 of 2998 (73%) invited patients responded to the national survey. Of these, 771 patients were eligible for the present study (Fig. 1). At the time of the survey the median (range) age of the total sample was 66 (45–75) years, and a median (range) of 23 (12–32) months had elapsed since start of local treatment. Patients in the RADHT group had used ADT for a median (range) of 6 (3–24) months and had received their last 3-month LHRH depot injection a median (range) of 21 (7–30) months before the survey. Patients in the NoTreat group were significantly older and those in the RP group were significantly younger than the other treatment groups (Table 1). As expected, significantly more men with RADHT belonged to the high-risk group (61%), compared with the other treatment groups. Of the whole study population, 45% reported comorbidity, with significantly lower prevalence in the RP group compared with the NoTreat and RADNoHT groups, the latter two groups displaying the highest prevalence of comorbidity (53%). Neuroticism was evenly distributed among the treatment groups.

Table 1. Demographics and diagnostic characteristics of the study groups and patients' physical and psychosocial condition at survey.
 Group 1 NoTreat N = 180Group 2 RP N = 293Group 3 RADNoHT N = 156Group 4 RADHT N = 142Total N = 771
  1. *At least one significant difference at p < 0.01. 1,2,3,4Indication of groups whose results differ significantly from the group under study.
Demographics and diagnostics     
Median (range) age at survey*70 (50–75)2,3,462 (45–73)1,3,467 (49–75)1,266 (48–75)1,266 (45–75)
Paired relationship, n (%)     
Yes153 (87)266 (92)136 (90)130 (94)685 (91)
No23 (13)22 (8)16 (11)9 (7)70 (9)
Higher education, n (%)     
Yes58 (34)127 (44)65 (43)55 (41)305 (41)
No115 (67)160 (56)85 (57)79 (59)439 (59)
Risk group, n (%)*     
Low risk92 (51)2,3,4157 (54)1,464 (41)1,415 (11)1,2,3328 (43)
Intermediate risk46 (26)100 (34)67 (43)40 (28)253 (33)
High risk41 (23)36 (12)25 (16)87 (61)189 (25)
Physical and psychosocial condition at survey     
Comorbidity, n (%)*     
No84 (47)2186 (64)1,374 (47)278 (55)422 (55)
Yes96 (53)107 (37)82 (53)64 (45)349 (45)
Neuroticism, n (%)     
Low105 (61)178 (63)80 (54)74 (54)437 (59)
Moderate46 (27)71 (25)42 (28)48 (35)207 (28)
High20 (12)35 (12)27 (18)16 (12)98 (13)

Adverse Effects

Significantly fewer RP patients reported moderate/severe irritative–obstructive urinary symptoms (23%) compared with the NoTreat (50%) and RADNoHT (35%) groups, when adjusted for age, risk group and comorbidity. Among patients in the RP group 24% had UI, significantly higher than the other groups (Table 2, Fig. 2A).

Figure 2.

Percentage of typical side effects per treatment group: A, Moderate/severe irritative–obstructive urinary symptoms and urinary incontinence. B, Intestinal symptoms and faecal leakage. C, Poor sexual drive and poor erectile function.

Table 2. Prevalence of adverse effects and global QoL outcomes, stratified per treatment group.
 Group 1 NoTreat N = 180Group 2 RP N = 293Group 3 RADNoHT N = 156Group 4 RADHT N = 142Total N = 771
  1. *At least one significant difference at p < 0.01 when adjusted for age, risk group and comorbidity. 1,2,3,4Indication of groups whose results differ significantly from the group under study.
Adverse effects     
Irritative–obstructive, n (%)*     
None/mild symptoms88 (50)2,4225 (77)1,387 (56)293 (66)1493 (65)
Moderate/severe symptoms87 (50)67 (23)68 (44)49 (35)271 (36)
Incontinence, n (%)*     
No157 (87)2223 (76)1,3,4140 (90)2134 (94)2654 (85)
Yes23 (13)70 (24)16 (10)8 (6)117 (15)
Drops9 (5)40 (14)13 (8)2 (1)64 (8)
More than drops14 (8)30 (10)3 (2)6 (4)53 (7)
Irritative symptoms, n (%)*     
No140 (79)3,4233 (80)3,483 (53)1,282 (58)1,2538 (70)
Yes38 (21)59 (20)73 (47)60 (42)230 (30)
Faecal leakage, n (%)*     
No172 (97)3,4283 (97)3,4133 (86)1,2120 (86)1,2708 (93)
Yes5 (3)8 (3)21 (14)19 (14)53 (7)
Sexual drive (n = 763)*     
Normal sexual drive, n (%)78 (45)2,3,4113 (39)145 (29)144 (31)1280 (36)
Poor sexual drive97 (55)180 (61)109 (71)97 (69)483 (63)
Erectile function (n = 753)*     
Normal erectile function, n (%)41 (24)2,333 (11)1,325 (17)1,223 (17)122 (16)
Poor erectile function131 (76)258 (89)126 (83)116 (84)631 (84)
Global QoL outcomes     
PCS ≤ 40, n (%)37 (21)32 (11)33 (21)32 (23)134 (17)
MCS ≤ 40, n (%)19 (11)33 (11)22 (14)8 (6)82 (11)
Low global QoL, n (%)*49 (27)54 (18) 347 (30)234 (24)184 (24)

Intestinal symptoms and faecal leakage were most common in irradiated patients; the estimates were significantly higher compared with the NoTreat and RP groups (Table 2, Fig. 2B).

Poor sexual drive was common (>50%) in all treatment groups, with significantly lower prevalence (55%) in the NoTreat group. Poor erectile function was also frequently reported in all treatment groups (>75%). The highest proportion of poor erectile function was observed among patients in the RP group (89%), which was significantly higher than the proportion in both the NoTreat group and the RADNoHT group (Table 2, Fig. 2C).

Global QoL

Compared with the other groups, the RP group had the lowest proportion of men with PCS ≤ 40 (11%), the statistical difference disappeared after adjustment for age, risk group and comorbidity. The proportion of men with MCS ≤ 40 was similar in all treatment groups (11% for all groups combined). The prevalence of low global QoL was 18% in the RP group (Table 2), with the RADNoHT group reaching a significantly higher level (30%).

Associations with Global QoL

All typical AEs (moderate/severe IPSS, UI, irritative intestinal symptoms, faecal leakage, poor sexual drive and poor erectile function) were significantly associated with low global QoL in univariate analyses (Table 3B). Low educational level, comorbidity and moderate or high neuroticism were all significantly associated with low global QoL in univariate analyses (Table 3A). No significant associations with global QoL were observed for age, a paired relationship or D'Amico risk group. Age was considered an important confounder and was therefore adjusted for in the multivariate analyses.

Table 3. Univariate associations between possible confounders and low global QoL and univariate and multivariate logistic regression analyses for each AE, with low global QoL as dependent variable.
A, Associations between possible confounders and low global QoL
OR (95% CI)p
Age (continuous)1.02 (0.99–1.05)0.196
Paired relation (yes = ref)1.88 (1.12–3.17)0.017
Education (high = ref)1.66 (1.16–2.36)0.006
Risk group (D'Amico) (low risk = ref)  
Intermediate risk0.97 (0.65–1.44)0.860
High risk1.57 (1.05–2.36)0.028
Comorbidity (no = ref)3.02 (2.14–4.28)<0.001
Neuroticism (low = ref)a  
Moderate3.47 (2.27–5.29)<0.001
High15.26 (9.14–25.47)<0.001
B, Multivariate logistic regression analyses for each AE, with low global QoL as dependent variable
 OR (95% CI)pOR (95% CI)p
  1. aAdjusted for age, education, comorbidity and level of neuroticism.
Moderate/severe irritative-obstructive urinary symptoms (no/mild = ref)3.02 (2.14–4.26)<0.0012.34 (1.57–3.50)<0.001
Urinary incontinence (no = ref)2.32 (1.53–3.51)<0.0011.80 (1.10–2.93)0.019
Intestinal symptoms (no = ref)2.23 (1.58–3.14)<0.0011.59 (1.05–2.40)0.029
Faecal leakage (no = ref)2.93 (1.66–5.18)<0.0012.37 (1.19–4.70)0.014
Poor sexual drive (normal = ref)2.60 (1.76–3.84)<0.0012.03 (1.29–3.20)0.002
Poor erectile function (normal = ref)3.02 (1.65–5.51)<0.0011.96 (0.99–3.86)0.053

In the multivariate analyses (Table 3B) the presence of moderate/severe irritative–obstructive urinary symptoms and poor sexual drive were each significantly associated with increased risk of low global QoL. UI, intestinal symptoms, faecal leakage and poor erectile function did not reach statistical significance in the adjusted analyses. In all six logistic regression analyses, neuroticism was the variable with the highest point estimate of association with low global QoL (data not shown).

Medication for ED

Of 767 evaluable patients, 385(50%) had used EDmed, but only 24% were ‘still users’. ‘Never users’ of EDmed were significantly older than the ‘discontinued’ and ‘still user’ groups and had a higher prevalence of comorbidity compared with ‘still users’ (Table 4). Poor sexual drive was significantly more prevalent among ‘never users’ (71%) than ‘still users’ (50%). The prevalence of poor erectile function was similar in men from the three EDmed groups. Among the three groups of EDmed users, no significant differences emerged for low global QoL when comorbidity and age were adjusted for.

Table 4. Comparison of men not using EDmed with those who had used, but discontinued EDmed, between start of local treatment and survey* and those still using EDmed at survey.
 EDmed Group 1: ‘Never users’ N = 382‘Ever users’ N = 385
EDmed Group 2: ‘discontinued users’ N = 201EDmed Group 3: ‘still users’ N = 184
  1. *The NoTreat group was calculated from the time of diagnosis as a substitute for start of local treatment. At least one significant difference at p < 0.01. Adjusted for age and comorbidity. 1,2,3Indication of groups whose results differ significantly from the group under study.
Median (range) age68 (50–75)2,363 (45–75)163 (48–75)1
Comorbidity, n (%)197 (52)387 (43)64 (35)1
Poor sexual drive†‡266 (71)3124 (62)92 (50)1
Poor erectile function306 (83)167 (84)155 (85)
Low QoL, n (%)102 (27)51 (25)31 (17)

Significantly more men in the RP group were ‘ever users’ of EDmed compared with the NoTreat, RADNoHT and RADHT groups (Table 5). Men in the RP group were prescribed EDmed closer to treatment start compared with men in the other treatment groups.

Table 5. Post-treatment use of EDmed in 771 evaluable men*.
 Group 1 NoTreat N = 180Group 2 RP N = 293Group 3 RADNoHT N = 156Group 4 RADHT N = 142Total N = 771
  1. *Calculated from date of start of local treatment (Groups 2–4) or date of diagnosis (Group 1). At least one significant difference at p < 0.01, adjusted for age and comorbidity. Proportion of ever users. Percentages do not summarize to 100 because of missing categorization of patients. 1,2,3,4Indication of groups whose results differ significantly from the group under study.
Ever users, n (%)35 (19)2,3,4246 (84)1,3,457 (37)1,251 (36)1,2389 (51)
Discontinued users†‡, n (%)19 (54)129 (53)20 (35)433 (66)3201 (52)
Still users, n (%)14 (40)116 (48)37 (65)17 (34)184 (47)
Months to first prescription after treatment start, median (range]6 (0–29]22 (0–19]1,3,47 (0–19]29 (0–24]23 (0–29]


Our population-based study documents that cure from PCa has its price, but that ‘no treatment’ is also associated with symptoms similar to typical AEs. Two types of urinary AEs emerged with different patterns in the treatment groups. While the prevalence of UI was highest after RP (24%) irritative–obstructive symptoms were most often recorded by patients in the RAD or in the NoTreat group (35–50%). Intestinal AEs were more common after RAD compared with RP or no treatment. More than half of the patients in all groups reported poor sexual drive, though it was significantly less often reported by patients in the NoTreat group. Poor erectile function was reported by >75% of men in all treatment groups, most often after RP. Despite having typical AEs, the global QoL of men with discontinued curative treatment was good for 75%. In adjusted analyses irritative–obstructive urinary symptoms and poor sexual drive each approximately doubled the risk of low global QoL, whereas UI, intestinal symptoms and poor erectile function were not significantly associated with low global QoL. EDmed was most often used by patients who had undergone RP. Such use was not significantly associated with global QoL.

The present results should be interpreted with consideration to the background of the treatment policies valid in Norway around 2004. Firstly, today's surveillance strategy was not implemented as a national recommendation. We suspect that high age and chronic pre-treatment comorbidity allocated many men to our NoTreat group. Secondly, the RADNoHT group consisted mainly of patients with intraprostatic tumours whose comorbidity probably did not allow a major surgical procedure, which would have been the treatment of choice in Norway at that time. These policies explain our high rates of post-treatment comorbidity and high proportion of men with PCS ≤ 40 in the NoTreat and RADNoHT groups. The high proportion of PCS ≤ 40 (23%) in the RADHT group, not different from estimates in the NoTreat and RADNoHT groups, is probably attributable to ADT's negative impact on physical health, persisting for several months after discontinuation of ADT.

The duration of adjuvant ADT is a matter of ongoing debate. As ADT often leads to physical and mental AEs, such treatment should be as short-lasting as possible without reduction of its beneficial effect on survival. In practice, planned long-lasting hormone treatment is often prematurely discontinued in a patient with severe ADT-related AEs, possibly explaining the varying duration of adjuvant ADT in our RADHT group.

The present study adds to the growing evidence that irritative–obstructive urinary symptoms represent late AEs, which affect the global QoL of a patient with PCa. Published studies have emphasized the impact of irritative–obstructive urinary symptoms on satisfaction with treatment outcome and global QoL [6, 8]. In the present study men in the NoTreat and RADNoHT groups were more likely to experience these AEs than men who had undergone RP. Similar results were found in the prospective study by Pardo et al. [8] where irritative–obstructive symptoms were reduced in about half of the patients 3 years after RP. The prevalence of moderate/severe irritative–obstructive urinary symptoms was highest in our NoTreat group, probably related to the growth of the prostate gland, although not significantly different from the RADNoHT group.

Urinary incontinence has been the main focus of many studies and, as expected, it was most common after RP, but only 10% of patients in the present study, who had undergone RP, described their UI as ‘more than just dribbling’. Prevalence of UI found in the Prostate Cancer Outcomes Study [4] after 24 months of follow-up was 4.7, 3.3 and 21.5% for watchful waiting, RAD and RP, respectively. Steineck et al. [10] found that 18% of men randomized to RP and 2% of men randomized to watchful waiting reported a moderate-to-severe degree of urinary leakage after a mean follow-up of 4 years. Our high prevalence of UI in the NoTreat group (13%) and severe leakage (8%) must be viewed on the basis of this group's heterogeneity.

In addition to irritative–obstructive urinary symptoms, poor sexual drive was significantly associated with a doubled risk of having low global QoL. Interestingly, and contrary to Penson et al.'s study [4], the association between poor erectile function and low QoL did not reach statistical significance. This may be related to cultural differences in patients' view concerning sexual functioning with increasing age. A response shift must also be considered as a possible explanation [35, 36]. Response shift reflects the gradual acceptance of treatment-related AEs and change of expectations as to global QoL. Based on our results we speculate that response shift related to low global QoL develops less easily with regard to irritative–obstructive urinary symptoms and poor sexual drive. This is probably because these AEs are more disturbing and interfere with the general perception of health.

Age, education, comorbidity and neuroticism were identified as confounding factors which moderated the observed association between AEs and low global QoL. Even though the confounding effect of neuroticism was moderate, level of neuroticism consequently had the strongest association with low global QoL (data not shown) which gives further strength to its importance in QoL research. A person's presence of a moderate or high level of neuroticism was more important for reporting low global QoL than erectile or intestinal dysfunction or UI. This result also implies that personality should be taken into account when informing patients about functional AEs, as nervous men might experience more symptoms concurrent with a low QoL, than less neurotic ones.

Even though treatment for ED is available, only about half of our patients had used such medication after PCa treatment, most men in the RP group (84%) and fewest in the NoTreat group (19%). We cannot decide whether these differences are primarily attributable to varying patients' demands or if they reflect prescription patterns differing between urologists and oncologists. Both explanations are probably relevant. With today's knowledge of the importance of early activation of nervous pathways responsible for erectile function [37], early and more frequent post-treatment management of ED is challenging, also after RAD. About half of ‘ever-users’ report that they had not used EDmed within the 4 weeks before the survey which may indicate that today's EDmed is not effective for many of these patients. Of special interest, and admittedly based on small figures, is the fact that almost two thirds of patients from the RADNoHT group were ‘still users’ as compared with less than half in the other groups. This may indicate that these drugs are particularly effective after RAD as monotherapy. While erectile function was similar in ‘ever-users’ and ‘never-users’, poor sexual drive was significantly more prevalent in the latter group, probably reflecting their lack of motivation to try EDmed. Low global QoL was not significantly associated with EDmed use and the small proportion of ‘still users’ with low QoL is explained by their younger age and fewer comorbid conditions.

Several limitations concerning our cross-sectional study should be mentioned, such as the lack of pre-treatment data. Not recognized intergroup variability in pre-treatment dysfunctions and low global QoL could have introduced a systematic bias. Our choice to select single domains from different instruments for the survey questionnaire instead of one complete validated questionnaire is today questionable. Our solution has to be viewed as a compromise between interests of oncologists and urologists for whom self-report of AEs was a new methodology in 2004. Further, there is uncertainty as to whether or not the questions about ED have been answered disregarding the use of EDmed. Since 85% of ‘still users’ report poor erectile function similarly to ‘never users’ and ‘discontinued users’, the majority have probably reported their function in absence of sexual aids. We propose that future questionnaires should clearly separate the patient's report on erectile function related to the use of EDmed.

The major strengths of the present study are the population-based design and the comparison of three of today's major treatment methods in addition to a ‘no treatment’ group. Further, the described typical AEs are those emerging after completion of planned initial treatment alone, without the use of adjuvant or salvage treatment. Finally, the reported prevalences of typical AEs and low global QoL are probably persistent ones, as several studies have shown that PCa-related QoL stabilizes 6–12 months after treatment [4, 38, 39].

In conclusion, PCa survivors after curative treatment, but also patients without any therapy, report considerable rates of sexual, urinary and intestinal AEs. Irritative–obstructive urinary symptoms and poor sexual drive each approximately double the risk of low global QoL. Use of EDmed was most common among men in the RP group, and was not associated with global QoL.


This project was financially supported by the Norwegian ExtraFoundation for Health and Rehabilitation through EXTRA funds. The study used data from the CRN. The interpretation and reporting of these data are the sole responsibility of the authors, and no endorsement by the CRN is intended, nor should be inferred.

Conflict of Interest

None declared.