In this report, the authors describe the health-related quality of life (HRQL) of long-term prostate cancer survivors 5 to 10 years after diagnosis and compare it with the HRQL of an age-matched, normative sample of the general Dutch population.
The population-based Eindhoven Cancer Registry was used to select all men who were diagnosed with prostate cancer from 1994 to 1998. Nine hundred sixty-four patients received questionnaires (the 36-item Short Form Health Survey [SF-36] and the Quality of Life-Cancer Survivors questionnaire), and 780 of 964 patients responded (81%).
Unselected, long-term prostate cancer survivors reported comparable HRQL scores but worse General Health Perceptions and better Mental Health scores than an age-matched, normative population. Patients who underwent radical prostatectomy had the highest physical HRQL, followed by patients who received ‘watchful waiting,’ and patients who received radiotherapy. Patients who received hormone treatment, in general, had the lowest physical HRQL.
Prostate cancer currently is the most prevalent cancer among older males in Western, industrialized countries.1 Because of the introduction of prostate-specific antigen (PSA) testing in the early 1990s, more prostate cancer is detected at an earlier disease stage at diagnosis.2 The number of patients with prostate cancer also is increasing because of the growing incidence, improved survival, and ageing of the population.3 This has led to increasing numbers of individuals who are either cured of their cancer or living with it as a chronic disease.4 In the south of the Netherlands, the 5-year relative survival rate for patients with prostate cancer is 83%, and the 10-year relative survival rate is 69%.5 This means that the majority of patients with prostate cancer will become long-term survivors. According to the definition of the American Cancer Society, a ‘long-term survivor’ is a patient with cancer who is alive 5 years after initial diagnosis.6
Prostate cancer and its treatment can affect both disease-specific health-related quality of life (HRQL) (e.g., urinary, sexual, and bowel functions) and general HRQL (e.g., energy/vitality and mental and physical health), although the latter effect is not always found.7–9 Many studies have investigated the HRQL of patients with prostate cancer, but only a few focused on long-term survivors.10–13 In addition, the majority of those studies were not population-based, relatively few patients were included, and studies that compare all available treatment options are scarce. The results from a population-based study that includes large numbers of patients who receive different treatments can be generalized more easily to the full population of long-term prostate cancer survivors. The objectives of the current population-based study were to obtain insight into the long-term HRQL effects of prostate cancer and its various treatments and to compare the HRQL of prostate cancer survivors with that of the general population.
MATERIALS AND METHODS
Setting and Participants
A population-based, cross-sectional survey was conducted at the Eindhoven Cancer Registry (ECR). The ECR data on all patients with newly diagnosed cancer in the southern part of the Netherlands, which is an area that has 2.3 million inhabitants, 10 hospitals with 17 hospital locations, and 2 large radiotherapy institutes.5 The ECR was used to select all patients who were diagnosed with prostate cancer between 1994 and 1998. Participants older than 75 years at diagnosis were excluded, because it was expected that they would have difficulty in completing a self-report questionnaire without assistance. To exclude all patients who had died before November 1, 2004, our data base was linked with the data base of the Central Bureau for Genealogy, which collects data on all deceased Dutch citizens through the civil municipal registries. After excluding all patients who had died, data collection was started in November 2004. Approval for this study was obtained from a local certified Medical Ethics Committee.
Urologists sent their (former) patients a letter, which informed them about the study, along with the questionnaire. The letter explained that, by returning the completed questionnaire, the patient agreed to participate and consented with linkage of the questionnaire data with their disease history as registered in the ECR. Patients were reassured that nonparticipation did not have any consequences for their follow-up care or treatment. If the questionnaire was not returned within 2 months, then a reminder letter with an additional copy of the questionnaire was sent.
The ECR routinely collects data on tumor characteristics, including date of diagnosis, tumor grade (Tumor-Lymph Node-Metastasis [TNM] clinical classification14), clinical stage (TNM clinical classification14), and treatment, and patient background characteristics, including date of birth and comorbidity at the time of diagnosis (a slightly adapted version of the Charlson comorbidity index15). Primary treatment was classified as radical prostatectomy (usually retropubic), radiotherapy, primary hormone therapy only, and watchful waiting (including transurethral resection of the prostate). The radiotherapy group only represented patients who received external beam radiotherapy; brachytherapy was not available as a treatment option in the region of the Comprehensive Cancer Center South between 1994 and 1998. Patients who underwent radical prostatectomy followed by any other adjuvant therapy, such as radiotherapy or hormone therapy, were included in the radical prostatectomy group. Patients who received radiotherapy and hormone therapy were included in the radiotherapy group.
The Dutch version of the 36-item Short Form Health Survey (SF-36) questionnaire was used to assess generic HRQL.16 According to standard scoring procedures, all scales were converted linearly to a scale from 0 to 100, with higher scores indicating better functioning. Differences of at least 5 points in the general health dimension,17 6.5 points in the physical dimensions, and 7.9 points in the mental health dimensions were considered clinically meaningful.18 The SF-36 scores from the patient sample were compared with scores from a gender-matched and age-matched, normative sample drawn from a large, random, nationwide sample of adults (n = 1742 individuals) taken from the general Dutch population.19
Generic HRQL survival issues were assessed with the Dutch version of the Quality of Life-Cancer Survivors (QoL-CS) questionnaire,20 including 45 visual analogue scales, each of which ranges from 0 (worst outcome) to 10 (best outcome). These 45 visual analogue scales are grouped into 4 multiitem subscales on well being: physical, psychological, social, and spiritual. It examines issues of particular concern to long-term cancer survivors, such as distress since diagnosis, sexuality, employment, uncertainty about the future, and the role of spirituality and religion.21 It has been demonstrated that the QoL-CS is a valid and reliable instrument when administered to American prostate cancer survivors20, 22, 23 and to the Dutch population of prostate cancer survivors,24 although the Spiritual Well Being subscale showed low reliability and validity in the latter cohort. The developers of the QOL-CS have suggested a 20% change or a 2-point difference on the 0 to 10 scales to indicate clinical significance.
The questionnaire also included questions on sociodemographic data, including marital status, current occupation, educational level, and disease progression and current comorbidity.
All statistical analyses were performed using SAS software (version 9.1 for Windows, SAS institute Inc., Cary NC). Routinely collected data from the ECR on patient and tumor characteristics enabled us to compare the group of respondents, nonrespondents, and patients who were lost to follow-up by using t tests for continuous variables and chi-square analyses for categorical variables. The sociodemographic and clinical characteristics of the patients were analyzed by using chi-square tests for categorical variables. Two hundred thirty-three patients were excluded from the primary analyses, because they exhibited disease progression (39 new primary tumors, 83 metastases, and 111 recurrences), which left 572 patients for the current analysis. Because a number of variables were skewed, nonparametric tests were used when appropriate.
Linear regression analyses were carried out to investigate the association between patient characteristics (age, comorbidity) and tumor characteristics (stage, treatment, time since diagnosis) with the composite and subscale scores of the SF-36 and with the QOL-CS. On the basis of the univariate results, multivariate models were constructed to determine which patient and tumor characteristics were associated independently with HRQL outcomes. We controlled for these variables in the analysis of covariance, which was used to compare mean scores on the SF-36 and the QOL-CS between different treatment subgroups.
The radical prostatectomy and radiotherapy groups included patients who also received additional hormone therapy as part of primary treatment. In the radical prostatectomy group, this hormone therapy usually was combined with additional radiotherapy. Because it was possible that combining patients who did or did not receive adjuvant therapies may have influenced the comparisons between the radical prostatectomy, radiotherapy, and hormone therapy groups, all analyses were performed twice: first for primary treatment groups without additional adjuvant therapies and, second, with all treatment combinations possible.
Seven hundred eighty of the 964 prostate cancer survivors (81%) returned a completed questionnaire (Fig. 1). A comparison of respondents, nonrespondents, and patients with unverifiable addresses indicated that nonrespondents were significantly older, were diagnosed more often with Stage I disease, and were less likely to undergo radical prostatectomy than respondents or patients with unverifiable addresses. Nonrespondents were more often not treated (‘watchful waiting’) than respondents and patients with unverifiable addresses (Table 1).
Table 1. Sociodemographic and Medical Characteristics of Questionnaire Respondents, Nonrespondents, and Patients with Unverifiable Addresses
No. of patients (%)
Respondents (N = 780)
Nonrespondents (N = 184)
Patients with unverifiable addresses (N = 128)
Age at survey
Years since diagnosis
Stage at diagnosis
Differences in sociodemographic and clinical characteristics of prostate cancer survivors, grouped by primary treatment, are presented in Table 2. Patients in the radical prostatectomy group were significantly younger than patients in the radiotherapy, hormone therapy, and watchful waiting groups. Patients in the radical prostatectomy group and the hormone therapy group were diagnosed mainly with Stage II and Grade 2 disease, patients in the radiotherapy group more often had Stage II and Grade 1 or 2 disease, and patients in the watchful waiting group most often were diagnosed with Stage I and Grade 1 disease. Patients in the radical prostatectomy group and the watchful waiting group more often were employed than patients in the radiotherapy group and the hormone therapy group, probably because of their younger age. No significant differences between the different primary treatment groups were observed in years since diagnosis, comorbidity, marital status, current occupation, or education level.
Table 2. Sociodemographic and Clinical Characteristics of Prostate Cancer Survivors without Recurrent Disease or New Primary Malignancies
P values indicate the difference between treatments.
Grade was based on the Tumor-Lymph Node-Metastasis clinical classification system (International Union Against Cancer, 199214). Grade 1 is comparable to Gleason scores of 2–4, Grade 2 is comparable to Gleason scores of 5–7, and Grade 3 is comparable to Gleason scores of 8–10.
Some patients who underwent radical prostatectomy also received additional primary radiotherapy (3% of patients) and additional primary hormone therapy (13% of patients), and some patients who received primary radiotherapy also received additional primary hormone therapy (19% of patients). Separate, additional analyses for each combination of treatments revealed no significant differences in HRQL (data not shown), which indicated that the same significant differences between treatment groups were identified by using combined or separate treatment groups (Table 3). Therefore, patients who underwent radical prostatectomy combined the receipt of any adjuvant therapy were included in the radical prostatectomy group, and patients who received radiotherapy and hormone therapy were included in the radiotherapy group (Table 2).
Table 3. 36-Item Short Form Health Survey Scores among Dutch Prostate Cancer Survivors According to Treatment at Diagnosis
On average, patients in the primary hormone therapy group scored 18 points lower for Physical Functioning (P < .001) and 9 points lower for Vitality (P < .05) than patients in the radical prostatectomy group. The mean score for Physical Functioning among patients in the radiotherapy group, on average, was 11 points lower compared with patients in the radical prostatectomy group (P < .001) (Table 3). Some potentially clinically meaningful differences were observed that were not statistically significant. This may have been caused by large within-treatment group variations.
Because the choice of treatment is associated strongly with age, we also separately analyzed the HRQL of patients younger than 75 years and older than 75 years (Table 3). The choice of separation into 2 age groups was justified by the analyses of variance, which showed differences in HRQL between these age groups for 5 subscales of the SF-36 (data not shown). In general, HRQL scores were higher for younger survivors than for older survivors. In general, patients in the radical prostatectomy group had higher HRQL scores compared with the other treatment groups, especially among patients 75 years and older. Among the younger patients, the watchful waiting group (n = 22 patients) was comparable to the radical prostatectomy group on all subscales, whereas the older patients in the watchful waiting group had constantly lower scores than the older patients in the radical prostatectomy group. There also were between-treatment differences for patients 75 years and older for the subscales Physical Functioning (P < .05) and Social Functioning (P < .05).
In all age groups, prostate cancer survivors reported statistically and clinically significant better Mental Health scores and worse General Health Perceptions scores compared with scores in the normative sample (Fig. 2). In addition, the youngest group of prostate cancer survivors (ages 60–69 years) reported higher Vitality scores (mean score, 68 vs. 62; P < .01) than the age-matched normative sample from the general population. Survivors between ages 70 years and 74 years showed lower scores for Bodily Pain (mean score, 81 vs. 89; P < .05) compared with the normative sample.
A comparison of SF-36 scores between the different treatment groups and the age-matched, normative sample revealed few statistically significant differences (Fig. 3). Patients in the radical prostatectomy, radiotherapy, and watchful waiting groups reported significantly less pain than the normative sample.
Patients in the radiotherapy, hormone therapy, and watchful waiting groups scored statistically significantly lower (P < .001) on the ‘Physical’ subscale of the QoL-CS compared with patients in the radical prostatectomy group (Table 4). Patients who were managed with watchful waiting scored significantly higher (P < .05) on the ‘Psychological’ subscale than patients who received treatment. However, these statistically significant differences were not clinically meaningful.
Table 4. Quality of Life-Cancer Survivors Scores for Each Subscale among Dutch Prostate Cancer Survivors According to Treatment at Diagnosis
SD indicates standard deviation; QoL-CS, Quality of Life-Cancer Survivors.
P values indicate the difference between treatment groups and were adjusted for disease stage, age at diagnosis, tumor grade, years since diagnosis, education, marital status, and comorbidity.
Physical well being
Psychological well being
Social well being
Spiritual well being
Total QoL-CS score
Five to ten years after diagnosis, patients who underwent radical prostatectomy reported the highest physical HRQL, followed by patients who were not treated (watchful waiting), and patients who received radiotherapy. Patients who received hormone treatment patients, in general, reported the lowest physical HRQL. Long-term survivors of prostate cancer had worse scores for general health perception but better scores for mental health compared with an age-matched, normative sample from the general Dutch population. Direct comparisons between treatment groups and the age-matched, normative population revealed that patients in all treatment groups had similar or even better HRQL scores.
The current results confirmed those from other studies in which the generic HRQL of prostate cancer survivors who underwent radical prostatectomy or received radiotherapy was never identified as worse than the generic HRQL of a control or normative sample.8, 10, 11 Because research on long-term cancer survivors is relatively recent, we also compared our results with those from several studies that focused on short-term survivors. In a longitudinal, prospective cohort study of men with localized prostate cancer,11 at a median 52-month follow-up assessment, patients who underwent radical prostatectomy and patients who received radiotherapy scored about the same or even better than the general population on all SF-36 scales. In a large American study, 12 to 48 months after treatment for early prostate cancer, survivors reported similar overall physical and mental health compared with men who did not have prostate cancer (SF-12).8 General HRQL also remained similar to that for a control group in an HRQL study among 709 long-term survivors (at a median of 6.2 years after treatment).10
There is increasing evidence that most individuals not only experience negative effects but also experience certain positive effects after an encounter with a stressor.25–31 This also is known as “benefit finding,” which can be described as the identification of benefit from adversity.32 Furthermore, patients may experience posttraumatic growth,31, 33–37 which has been described as an experience of significant, positive change that arises from struggle with a major life crisis.38 Benefit finding and posttraumatic growth may explain, at least in part, the higher self-reported HRQL in survivors compared with the age-matched norm population. Another possible explanation is that, although there is no prostate cancer screening program available in the Netherlands, most patients presumably are diagnosed by means of a PSA test, and this may be called screening. Socioeconomic status is associated positively with screening,39 and high socioeconomic status is related to high HRQL.40
Nevertheless, when we compared all prostate cancer survivors according to age group, we observed that self-reported General Health Perceptions were worse compared with the general population, normative sample. This is a novel finding that was not reported in previous studies. It may be attributed either to adverse late effects of treatment (e.g., incontinence, impotence, or bowel problems) or to a general feeling of being less healthy compared with population norms because of having had cancer in the past.
Better HRQL after treatment was described previously in patients who underwent radical prostatectomy compared with patients who received either radiotherapy11 or androgen-deprivation therapy.41 However, several studies reported no differences between treatments in terms of HRQL for short-term survivors42–44 and long-term survivors.10, 45 In 2 reports from the population-based American Prostate Cancer Outcomes Study, no association was observed between primary treatment (radical prostatectomy, radiotherapy, hormone-ablation therapy, and watchful waiting) and 2-year generic HRQL outcomes.43, 44 The follow-up on that study, 5 years after diagnosis, did not report a difference in HRQL between patients who received radiotherapy and patients who underwent radical prostatectomy.45 In another study among localized prostate cancer survivors and controls, measures of general HRQL remained similar for each treatment group 4 to 8 years after treatment.10 A study on HRQL after a mean follow-up of 4 years did not identify significant differences in either well being or subjective HRQL between patients who underwent radical prostatectomy and patients in a watchful waiting group.42 However, in that study, subjective HRQL was measured with a single question.
Possible explanations for differences in study outcomes may include the sometimes small patient numbers primarily in the hormone therapy and watchful waiting treatment groups. In addition, 2 studies were conducted less than 5 years after initial diagnosis.43, 44 It also is possible that baseline differences between treatment groups may have been more pronounced in 1 study or the other, influencing the outcomes of short-term and maybe even long-term HRQL follow-up. The current study, to our knowledge, is among the first population-based studies to include all patients who received all available primary therapies from 5 to 10 years ago. Perhaps this specific feature caused the differences we reported compared with other studies. In the literature, we identified only 1 study that included patients who received primary hormone therapy, which was the therapy that generated the most HRQL differences in our study.43 Finally, only 2 of the studies were population-based.43, 44
Despite the finding that most studies did not identify differences between treatments in generic HRQL, all studies did identify differences in disease-specific HRQL outcomes. This may seem contradictory. However, a Dutch study of patients with prostate cancer revealed that, although patients consider sexual, urinary, and bowel functions as problems, they do not take them into account when completing HRQL questionnaires, because they do not view these dysfunctions as aspects of health.46 Another explanation may be the “response shift” phenomenon, which suggests that an individual may change his internal standards or redefine his concept of HRQL over time.47, 48 In the Dutch study mentioned above, many patients accepted the side effects as inevitable consequences of having been treated for prostate cancer, a condition they perceived to be life-threatening.46
Particularly noteworthy are the high HRQL scores (SF-36) observed for the small group of young patients in the watchful waiting group compared with the older patients in the same group. A possible explanation for this difference is that the reasons for choosing watchful waiting probably differed between younger patients and older patients. Another possibility is that younger patients may understand better or may be able to deal better with such an approach.
Several limitations from the current study should be noted. First, although 12% of patients could not be sent a questionnaire because of unverifiable addresses, and 19% of patients who were sent a questionnaire did not respond, we do not know their current health status. Although nonresponders were diagnosed more often with Stage I disease, they also were older (80–85 years) and more often received no therapy. Therefore, it is possible that our results cannot be generalized to very old patients with prostate cancer. Furthermore, it is important to keep in mind that our results can be generalized only to a small percentage of the original group of 2348 prostate cancer patients, while a large group of patients could not be included in our study (they died, their hospital declined to participate, or their addresses could not be verified). Second, although the ECR routinely collects a range of clinical data, no information on either PSA levels during 5 to 10 years of follow-up or the duration of hormone treatment was available. We also do not know whether secondary hormone therapy was started some time after primary treatment. These factors may have an impact on the HRQL of long-term survivors. Third, it is more difficult to draw conclusions from a cross-sectional study than from a longitudinal study. Some baseline differences were identified in the characteristics of prostate cancer survivors (age, disease stage, tumor grade) grouped by primary treatment. These differences probably were caused by the nonrandom treatment allocation. Randomized controlled trials may ensure comparable groups at baseline. However, only a selected group of prostate cancer patients will be eligible for random allocation to treatments. Fourth, patients with disease progression were excluded from our analysis. Therefore, information on the HRQL of these patients is lost. However, we experienced difficulties in reporting HRQL for this heterogeneous subset of patients, because additional tumor characteristics for disease progression (e.g., localization, stage, grade, and time since diagnosis) were unknown. Our results can be generalized only to survivors with the best prognosis, namely, those who do not show any signs of disease progression. Fifth, another limitation of our study is the possibility that some findings may have been the result of chance. This is because of multiple testing and because we used a P value of .05; using a P value of .01 would reduce this possibility.
Despite these limitations, the results of the current study form an important contribution to the limited information available on HRQL for the growing group of long-term prostate cancer survivors, especially in patients who receive hormone treatment or for whom a watchful waiting policy was applied. The large number of participants in this study and the high response rate allowed us to extrapolate to the broad population of long-term prostate cancer survivors without disease progression.
In conclusion, the results of this study suggest that the long-term HRQL of prostate cancer survivors can vary significantly as a function of the type of primary treatment. Because baseline differences between treatment groups cannot be excluded as part of the explanation for these differences, the current findings need to be verified in longitudinal studies.
We thank all patients for their participation in the study. We also thank Rotary Tilburg Triborch and the Comprehensive Cancer Center South for their financial supporting of the study.