Quality of life of 1276 elderly patients with prostate cancer, starting treatment with a gonadotropin-releasing hormone agonist: results of a French observational study

Abstract

This French observational, longitudinal, prospective study described the health-related quality of life (HRQoL) of elderly men (≥75 years old) with prostate cancer after initiating gonadotropin-releasing hormone (GnRH) agonist therapy. At baseline and 3–6 months after baseline, European Organisation for Research and Treatment of Cancer quality of life questionnaire-core 30 (QLQ-C30) and prostate-specific (QLQ-PR25) questionnaires were completed by patients. Data from 1276 patients were analyzed. At baseline, mean (±SD) age was 80 (±4.1) years, 29.1% of patients had Gleason scores ≥8 and 24.9% had metastases. At baseline, increasing age, presence of metastasis and presence of comorbidity had a negative impact on QLQ-C30 and QLQ-PR25 scores. At follow-up, improvement in emotional-functioning (2.8; p < 0.001), social-functioning (1.7; p = 0.011), global HRQoL (1.6; p = 0.029), sleep-disturbance (−2.1; p = 0.011), appetite-loss (−4.0; p < 0.001) and pain (−4.1; p < 0.001) QLQ-C30 scores were observed. In addition, there was a worsening in treatment-related symptom (8.6; p < 0.001), sexual-activity (−5.5; p < 0.001) and sexual-functioning (−22.6; p < 0.001) QLQ-PR25 scores, and an improvement in urinary symptoms (−3.7; p < 0.001) and incontinence aid (−2.9; p = 0.023) QLQ-PR25 scores. This study shows that, apart from the expected impact on sexual functioning domains, HRQoL is not adversely affected by 3–6 months of GnRH agonist therapy in older men with prostate cancer.

• Elderly
• GnRH agonist
• prostate cancer
• quality of life

Introduction

Prostate cancer is the most common cancer diagnosed in men, with approximately 65 000 cases diagnosed in France in 2008 [1,2]. The incidence of prostate cancer strongly correlates with age and over 45% of all cases of prostate cancer are diagnosed in patients over 75 years of age [3]. Owing to increasing life expectancy and the introduction of more sensitive diagnostic techniques, the number of men with diagnosed prostate cancer and the use of prostate cancer therapies are expected to rise in the coming years [4]. In addition, prostate cancer mortality has steadily decreased over the last 20 years, due to advances in treatment and early detection of the disease [5,6]. Particular attention must be paid, therefore, to patient health-related quality of life (HRQoL) data in treatment decisions.

Androgen-deprivation therapy (ADT), most often achieved by administering a gonadotropin-releasing hormone (GnRH) agonist, is currently used as single-drug therapy for patients with metastases and as adjuvant therapy to radiotherapy for patients with locally advanced disease or intermediate- to high-risk localized disease [7–9]. It may also be used for patients with rising prostate-specific antigen (PSA) levels after prostatectomy. ADT has proven effectiveness in reducing disease-specific symptoms of advanced prostate cancer, leading to reductions in bone pain, pathological fractures, spinal cord compression and urethral obstruction [8]. Increased survival has also been reported in men with locally-advanced disease treated with ADT in conjunction with radiotherapy, but this benefit may be most pronounced in men with Gleason score >7 [10,11]. The optimal duration of treatment with ADT remains undetermined.

As prostate cancer is generally a disease of elderly men, many patients present with significant comorbidities that have a detrimental impact on life expectancy [12,13]. As a result, the potential benefits of ADT must be weighed against its potential side effects and subsequent impact on HRQoL [14].

Various prospective studies have reported the occurrence of side effects with ADT including hot flashes, decreased libido and erectile dysfunction [15]. Other complications include osteoporosis, with increased risk of fractures, and worsening of comorbidities such as diabetes, cardiovascular disease and metabolic syndrome, as well as physical, functional and cognitive impairment [8,16,17].

A better understanding of the impact of ADT on HRQoL and on daily activities of elderly men with prostate cancer is necessary in order to provide a more individualized treatment approach. The purpose of this observational study was to describe HRQoL in elderly patients before and after 3–6 months GnRH agonist treatment for prostate cancer.

Methods

A prospective, multicenter, longitudinal, observational study was implemented in France in 2009. Overall, 395 urologists, oncologists and radiotherapists were approached for participation in the study. The trial was conducted in accordance with the relevant ethical standards and was approved by local ethical committee (CCTIRS) in France. Informed consent was obtained from the patients prior to enrolment into the study.

During the inclusion period, each participating physician was asked to consecutively include the first five patients meeting the selection criteria seen at routine visits.

Demographic data and medical history were collected at the enrolment visit (baseline). HRQoL questionnaires (QLQ-C30 and QLQ-PR25) were completed by the patients at baseline and at the follow-up visit routinely planned approximately 3–6 months after starting GnRH agonist treatment.

Patients

The selection criteria was patients ≥75 years old with diagnosed prostate cancer starting GnRH agonist therapy, and untreated with a GnRH agonist for the previous 12 months. As this was a study, participation should not influence the physicians’ clinical judgment or drug prescribing practices, and patient visits were to be scheduled visits and not instigated by the study.

Data collected

The physician completed the Case Report Form, which consisted of demographic characteristics, disease history, clinical characteristics, the Karnofsky score, which is a measure of performance status ranging from 100 (“perfect” health) to 0 (death) [18], the routine biological parameters collected at the visit and the therapeutic management. Each physician also completed an identification form consisting of demographic data, location and type of practice.

HRQoL was assessed with two validated questionnaires:

• The European Organisation for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire-Core 30 (QLQ-C30) [19,20]. A generic questionnaire for cancer, including nine multi-item scales: global health and HRQoL scale, functional scales (physical, role, cognitive, emotional and social) and symptom scales (fatigue, pain, nausea and vomiting, dyspnea, insomnia, appetite loss, constipation and diarrhea). Several single-item symptom measures are also included. Domain scores range from 0 to 100, 100 score corresponding to the best HRQoL for functional and global health and HRQoL scales and the worse HRQoL for symptom scales.

• The Prostate Cancer Module of the QLQ-C30 and the QLQ-PR25 [20,21]. A more specific validated prostate cancer questionnaire which assesses urinary, bowel, sexual symptoms and sexual function including the side effects of hormonal treatment. Domain scores range from 0 to 100, with 100 corresponding to the worse HRQoL for symptom scales and the best HRQoL for sexual activity and function scales.

Given the observational nature of the study, any adverse reactions reported by the patient during the study were reported to the relevant pharmacovigilance center for the GnRH agonist concerned.

Statistical analysis

SAS® software 8.2 (SAS Institute, Cary, NC) was used for statistical analysis. Descriptive analyses were performed on all data collected. Quantitative variables were described by their mean, standard deviation (SD), median, quartiles 1 and 3, minimum/maximum values and quantity of missing data. Qualitative variables were described by the absolute and relative (%) frequency of each class and quantity of missing data. Where relevant, 95% confidence intervals (CIs) were calculated.

Mann–Whitney–Wilcoxon and Kruskal–Wallis non-parametric tests were used to compare quantitative variables (two subgroups and more than two subgroups, respectively). All statistical tests were two-sided and the α-level for statistical significance was 0.05. The effect of GnRH agonist therapy on change in HRQoL from baseline was further investigated by mixed model analyses (a method of multivariate analyses) which controlled for the potential effects of concurrent illnesses and treatments on HRQoL scores, so that the impact of GnRH on HRQoL could be isolated from the impact of these other factors on HRQoL. Normality assumptions were checked for each model through residuals, and convergence criteria were verified.

Results

Overall, 395 physicians participated in the study, and 252 physicians had at least one patient included (88.9% urologists and 11.1% oncologists/radiotherapists).

Patient distribution and follow-up

Of 1322 patients enrolled in the study, data from 1276 patients (96.5%) were included; 46 patients were excluded from the analysis because they did not meet the inclusion criteria. At baseline, 1071 eligible patients completed at least one evaluable inclusion questionnaire (QLQ-C30 and/or QLQ-PR25; response rate 83.9%). At the follow-up visit, 784 eligible patients completed at least one evaluable follow-up questionnaire (QLQ-C30 and/or QLQ-PR25; response rate 73.2%). Both QLQ-C30 and QLQ-PR25 had satisfactory psychometric properties in terms of clinical validity and internal consistency reliability – for example, at baseline and follow-up, QLQ-C30 questionnaires were completed with no missing information for any question by 85.9% and 85.7%, respectively.

Mean follow-up was 123.6 d (SD ± 69.1 d). On initiation of GnRH agonist therapy, it was combined with another therapy in 79.5% of the patients: antiandrogen (62.4%) and radiotherapy (18.2%). At the follow-up visit, 4.2% of the patients were no longer receiving GnRH agonist treatment, antiandrogens were prescribed in 32.3% and radiotherapy was prescribed in 5.6%.

Baseline patient characteristics

Baseline patient characteristics are summarized in Table 1. The mean age of patients was 80.0 years (SD ± 4.1 years), the mean time since prostate cancer diagnosis was 23.0 months (SD ± 39.3 months), 29.1% of patients had Gleason scores ≥8 and 24.9% of patients presented with metastatic involvement.

Table 1. Baseline patient characteristics.

Variable	Baseline value
Age (years), mean (±SD)	80.0 (±4.1)
75–80 years	53.8%
>80–85 years	30.9%
≥85 years	15.6%
Time since diagnosis (months), mean (±SD)	23.0 (±39.3)
Karnofsky score
≤80	18.3%
>80	81.7%
Sexually active	26.3%
Tumor stage
T0	1.7%
T1	9.2%
T2	28.1%
T3	50.0%
T4	8.7%
Tx	2.3%
N0	42.4%
N1+	9.9%
Nx	47.8%
M0	45.8%
M1	24.9%
Mx	29.2%
PSA (µg/l), median (Q1; Q3)	15.2 (8.0; 37.0)
Gleason score
<7	23.1%
7	47.9%
>7	29.1%

Approximately 53% of patients presented with at least one co-morbidity: diabetes, 37.6%; cardiac failure, 29.5% and chronic obstructive pulmonary disease, 21.2%. Approximately 26% of patients were sexually active 4 weeks before enrolment in the study (Table 1), but sexual activity declined with increasing age (36.9% of patients between 75 and 80 years of age, 16.7% of patients between 80 and 85 years of age and 9.1% of patients ≥85 years old).

Approximately 34% of the patients had already undergone a specific prostate cancer treatment at baseline. These included external radiotherapy (14.9%), a GnRH agonist treatment which ceased >12 months before inclusion in the study (11.7%), radical prostatectomy (8.2%) or an antiandrogen treatment (6.9%).

HRQoL at baseline

Mean baseline scores and SDs for the QLQ-C30 domains are shown in Table 2. Patients with at least one comorbidity had poorer HRQoL than patients with no comorbidities – mean baseline global health status scores were 58.8 [95% CI, 57.1; 60.5] and 65.0 [95% CI, 63.1; 66.8] for those with and without comorbidities, respectively. Baseline scores were also worse in patients with at least one comorbidity for all functional domains, most of the symptom parameters (fatigue, pain, dyspnea, insomnia and appetite loss), and financial difficulties. QLQ-C30 global health status score was worse in patients with metastasis than patients with no metastasis (51.0 [95% CI, 48.4; 53.5] and 65.1 [95% CI, 63.7; 66.5], respectively), as were scores on most of the functional domains and the fatigue, pain, dyspnea, insomnia, appetite loss domains. Likewise, QLQ-C30 global health status score was worse in patients receiving antiandrogens than patients not receiving antiandrogens (59.8 [95% CI, 58.2; 61.5] and 65.0 [95% CI, 63.0; 67.0], respectively) as were scores on the physical domain and the fatigue and appetite loss domains. In addition, mean baseline scores for the physical, role and fatigue domains of QLQ-C30 were worse with increasing age.

Table 2. Responses to QLQ-C30 questionnaire among elderly men starting GnRH therapy.

	Baseline			Follow up (3–6 months)			Change
	N	Mean	SD	N	Mean	SD	Mean [95% CI]	p‡	p¶
Global quality of life*
Global health status	1049	61.7	20.8	774	63.5	19.0	1.6 [0.2; 2.9]	0.034	0.029
Functional scales*
Physical	1071	79.7	20.9	782	80.3	20.0	0.3 [−0.8; 1.4]	0.845	0.541
Role	1066	78.6	25.8	781	80.1	24.0	0.6 [−1.0 2.2]	0.483	0.456
Emotional	1069	79.7	20.5	782	83.0	18.6	2.8 [1.6; 4.1]	<0.001	<0.001
Cognitive	1070	78.0	21.6	783	78.8	20.7	0.9 [−0.4; 2.2]	0.137	0.248
Social	1064	84.0	22.5	781	86.2	20.1	1.7 [0.3; 3.2]	0.006	0.011
Symptom scales†
Fatigue	1067	27.6	23.8	781	26.6	22.1	−0.6 [−2.1; 0.9]	0.465	0.644
Nausea/vomiting	1071	3.0	10.0	782	3.6	9.6	0.8 [−0.1; 1.7]	0.012	0.151
Pain	1069	16.9	23.5	783	12.4	18.7	−4.1 [−5.7; −2.4]	<0.001	<0.001
Dyspnea	1062	21.5	26.0	779	19.6	24.5	−1.6 [−3.3; 0.1]	0.069	0.022
Insomnia	1065	21.9	26.1	779	20.2	25.1	−2.1 [−3.9; −0.3]	0.044	0.011
Appetite loss	1063	13.8	22.6	780	9.6	17.9	−4.0 [−5.6; −2.4]	<0.001	<0.001
Constipation	1067	15.5	25.2	773	13.8	23.4	−0.9 [−2.5; 0.8]	0.353	0.159
Diarrhea	1064	5.6	15.4	777	5.6	14.8	0.9 [−0.3; 2.1]	0.266	0.359
Economical impact†
Financial difficulties	1055	5.4	15.0	774	4.4	13.8	−0.7 [−1.8; 0.5]	0.283	0.255

*Higher scores indicate better global QOL and functioning.

†Higher scores indicate worse symptoms.

‡p Value from univariate analysis.

¶p Value from mixed model.

Table 3 shows the QLQ-PR25 mean scores and SDs of urinary, bowel, treatment-related and sexual symptoms domains. Patients with at least one comorbidity had poorer baseline QLQ-PR25 urinary-symptom domain scores than those with no comorbidities (22.6 [95% CI, 21.1; 24.2] and 19.2 [95% CI, 17.6; 20.8], respectively). Urinary-symptom, bowel-symptom and sexual functioning domains were worse in patients with metastasis than in patients with no metastasis. QLQ-PR25 scores were similar in patients receiving or not receiving antiandrogen treatment at baseline. In addition, mean baseline scores for the urinary symptoms and sexual activity domains of QLQ-PR25 were worse with increasing age.

Table 3. Responses to QLQ-PR25 questionnaire among elderly men starting GnRH therapy.

	Baseline			Follow-up (3–6 months)			Change
	N	Mean	SD	N	Mean	SD	Mean [95% CI]	p¶	p||
Urinary, bowel and treatment related symptom scales*
Urinary symptoms	1052	21.1	18.3	773	16.8	15.2	−3.7 [−4.8; −2.5]	<0.001	<0.001
Incontinence aid‡	292	20.3	29.0	172	15.5	25.8	−2.9 [−7.6; 1.8]	0.122	0.023
Bowel symptoms	992	5.8	10.4	740	5.4	9.0	0.3 [−0.4; 1.1]	0.131	0.538
Treatment related symptoms	1054	9.8	12.7	776	17.7	14.5	8.6 [7.5; 9.6]	<0.001	<0.001
Sexual domains†
Sexual activity	1028	13.3	20.9	757	8.0	16.7	−5.5 [−6.9; −4.1]	<0.001	<0.001
Sexual functioning§	282	58.3	25.0	107	39.0	24.7	−22.6 [−28.8; −16.3]	<0.001	<0.001

*Higher scores indicate worse symptoms.

†Higher scores indicate better activity and functioning.

‡Patients who wear incontinence aids only.

§Patients with sexual activity within the last 4 weeks.

¶p Value from univariate analysis.

||p Value from mixed model.

Evolution of HRQoL between baseline and the 3–6 month follow-up visit

Analysis of the QLQ-C30 score changes showed that mean emotional (2.8; p < 0.001 mixed model), social (1.7; p = 0.011) and global health status (1.6; p = 0.029) scores improved significantly from baseline after 3–6 months of GnRH agonist treatment (Table 2). In addition, significant decreases in mean pain (−4.1; p < 0.001), insomnia (−2.1; p = 0.011) and appetite-loss (−4.0; p < 0.001) scores from baseline were reported (Table 2). Patients with metastases at baseline had significantly higher improvement in social (change of 5.9 [95% CI, 2.8; 9.0] and 0.4 [95% CI, −1.3; 2.0], respectively) and pain (change of −12.0 [95% CI, −16.0; −8.0] and −1.5 [95% CI, −3.2; 0.1], respectively) scores than patients without metastases. Significantly higher improvements were also found in pain (change of −6.6 [95% CI, −8.9; −4.2] and −1.7 [95% CI, −4.0; 0.5], respectively) and appetite-loss (change of −6.6 [95% CI, −9.0; −4.3] and −0.9 [95% CI, −3.0; 1.3], respectively) scores in patients with at least one comorbidity as compared with patients without comorbidities, respectively. QLQ-C30 score changes were not significantly different between patients treated with combined androgen blockade and patients without anti-androgen treatment, or between the different age groups.

The main QLQ-PR25 changes from baseline were a significant worsening of the sexual functioning and sexual activity scores (p < 0.001; Table 3). The prevalence of sexual activity during the 4 weeks before the follow-up visit was 13.6% (17.6% in patients between 75 and 80 years, 9.2% in patients between 80 and 85 years and 8.1% in patients ≥85 years old). Results also showed a deterioration of treatment-related symptoms and an improvement in urinary-symptom scores (p < 0.001; Table 3). In the mixed model, the improvement in incontinence aid was also significant (p = 0.023; Table 3). Analysis of subgroups showed significant impact of age and metastasis involvement on several QLQ-PR25 domain changes. Patients ≥85 years old had significantly lower deterioration from baseline to follow-up of treatment-related symptom scores than patients <85 years old (change of 3.0 [95% CI, 0.1; 5.9], 9.2 [95% CI, 7.2; 11.3] and 9.7 [95% CI, 8.4; 11.0], in the ≥85 years, 80–84 years and 75–79 years age groups, respectively). Patients with metastasis at baseline had both significantly higher improvement of urinary-symptom scores (change of −7.1 [95% CI, −9.4; −4.8] and −2.5 [95% CI, −3.9; −1.1], respectively) and greater deterioration of treatment-related symptom scores (change of 11.6 [95% CI, 9.4; 13.8] and 7.6 [95% CI, 6.4; 8.8], respectively) than patients without metastasis. QLQ-PR25 score changes were not significantly different between patients with or without comorbidities, and between patients treated with combined androgen blockade and patients without anti-androgen treatment.

Discussion

This prospective observational study was conducted on a large cohort of more than 1000 patients, with the participation of approximately 20% of all French urologists. The purpose of this study was to investigate HRQoL of elderly patients starting treatment with GnRH agonists for their prostate cancer and the evolution of their quality of life during the treatment. As would be expected, ADT had a negative impact on sexual functioning domains of HRQoL questionnaires, but other domains of HRQoL are improved or unchanged with GnRH agonist therapy.

Since the work of Huggins et al. in 1941, androgen deprivation has been an important component of the therapeutic strategy of prostate cancer [22]. However, androgen deprivation is associated with significant side effects including the loss of libido, erectile dysfunction, osteoporosis, hot flashes, and cognition and mood disorders [8,23]. Metabolic complications (hyperglycemia, insulin resistance and metabolic syndrome), possibly associated with increased cardiovascular risks, have also been reported [24–26]. All these side effects have a potential impact on the HRQoL of patients.

In this observational study, HRQoL baseline mean scores were generally in line with data provided by the EORTC scoring manual for >70-year-old patients with prostate cancer [20]. Any differences can be explained by the older mean age and lower symptomatic carcinoma in this study. Most domains of HRQoL were reduced with higher age, advanced stage of the disease, the presence of comorbidity and, to a lesser extent, the treatment by combined androgen blockade. Borghede et al. previously described this impact of aging on HRQoL in elderly patients with prostate cancer [27], and Pinkawa et al. reported a strong effect of age and comorbidities on HRQoL, in particular, on urinary function and sexual function, for patients with prostate cancer [28]. In our study, aging was also associated with a decrease in sexual activity at baseline. This decrease in sexual activity with age is consistent with data for the general population [29]. Furthermore, as the prostate is considered a sexual organ, the diagnosis of prostate cancer in itself may have a psychological impact in terms of sexual body image, identity and the sexual functioning of patients [30]. The baseline data from our study therefore confirm and add to what is known about the impact of prostate cancer on HRQoL in elderly men.

Recommendations regarding the interpretation of QLQ-C30 score changes have been published by Cocks et al. [31]. Mean QLQ-C30 changes in scores of below 4 points may be defined as clinically not significant, mean changes of 4–10 points as a small change, mean changes of 10–20 as a medium change and mean changes greater than 20 as a large change. To our knowledge, there are no equivalent guidelines for QLQ-PR25. Based upon these findings, the statistically significant improvements of mean QLQ-C30 scores in global health status, emotional and social functioning, and sleep disturbances 3–6 months after starting GnRH agonist therapy were not clinically significant. However, the improvements in appetite loss and pain after initiating GnRH agonist therapy were small but clinically significant. It is not known if the improvement in the pain domain of the QLQ-C30 may have influenced, for example the emotional functioning improvements. This prompt effect of GnRH agonist therapy on HRQoL domains may be the result of the sudden decrease in testosterone levels.

It is interesting to compare these findings with previously reported effects of GnRH agonist therapy on HRQoL. A prospective, multicenter study, including 104 patients with prostate cancer treated with leuprolide acetate investigated patient HRQoL using QLQ-C30 and QLQ-PR25 questionnaire after 12 weeks. Consistent with our results, patients experienced significant improvement in global health status, appetite loss and a decline in sexual function. However, a significant worsening in physical and role functioning was also observed with leuprolide [32]. This is one of the few studies the authors are aware of that uses the same scales and similar timelines to our observational study. The inclusion of a considerably larger population in this observational study provides useful additional information on the short-term impact of GnRH agonist therapy on HRQoL in everyday practice.

Several studies have been conducted in patients with prostate cancer using other HRQoL questionnaires. A longer-term prospective, observational study measured HRQoL in 144 patients with locally advanced prostate cancer or PSA failure after local treatment, with a questionnaire that combined the EORTC Prostate Cancer Quality of Life Questionnaire with the Intrusion subscale of the Impact of Event Scale and Selby’s QL Uniscale [33]. After 12-months, patients receiving ADT reported lower overall HRQoL (increased fatigue, emotional distress and decreased physical functioning), with a greater adverse effect on HRQoL for patients treated with combined androgen blockade than leuprolide alone [33]. This trend was not seen in our study after 3- to 6-month follow-up, but potentially could become evident with longer follow-up.

Published data indicate that GnRH agonists may negatively influence cognitive functions of patients such as memory and attention [34]. Several HRQoL studies [33,35] have also shown a link between androgen deprivation and worsening of fatigue or energy levels. In contrast, we found no effects in cognitive or fatigue domains in our study, possibly due to the relatively short follow-up in the current study.

Responses to the QLQ-PR25 questionnaire showed a notable worsening of sexual activity and sexual functioning scores, reflecting the decrease of testosterone concentrations. These results are consistent with the decrease in the prevalence of sexually active patients between baseline (26.1%) and follow-up visits (13.6%). This is consistent with several other studies which previously demonstrated a decline in sexual function and sexual satisfaction in patients receiving hormone therapy [36–38].

As recommended [39], the psychometric properties of the QLQ-C30 and QLQ-PR25 questionnaires were assessed in order to confirm that the questionnaires could provide scientifically sound results in this study. The QLQ-C30 and QLQ-PR25 questionnaires were well-accepted by patients as determined by the high completion rate. Satisfactory psychometric properties in terms of clinical validity and consistent internal reliability were also recorded for both QLQ-C30 and QLQ-PR25. Therefore, both QLQ-C30 and QLQ-PR25 can be considered as reliable and valid instruments, supporting their use to measure patients’ health status in the study.

There were a number of limitations in our study. First, our study was observational and the participating physicians and patients were volunteers, which represented a classical non-response selection bias for this type of study. As a result of the observational design, the number of patients responding at baseline and follow-up were considerably different, and the impact of missing data at follow-up cannot be assessed. Other details such as the reasons for patients stopping GnRH therapy in the first 3–6 months of treatment were not recorded. The impact of radiation therapy at baseline and follow-up on HRQoL is not known and a separate sub-group analysis of these patients was not conducted. A further limitation comes from the QLQ-PR25 scoring. As this instrument is still not completely finalized, the results were presented for certain single items, but no global score could be derived for each global symptom scale, thereby limiting the scope of treatment-related symptoms analysis. Despite these limitations, our study represents one of the first large multicenter, prospective studies on HRQoL and treatment-related symptmos in ≥75-year-old men treated for prostate cancer with commonly used therapeutic strategies.

In conclusion, the study highlights the potential impact of GnRH agonist therapy in everyday practice on global health status and several functional and symptom domains in a large population of patients with prostate cancer. The reduction of testosterone levels during ADT has a predictable impact on sexual functioning domains of HRQoL questionnaires, but this study shows that other domains of HRQoL are not adversely affected by GnRH agonist therapy (and some are improved) in older men with prostate cancer.

Declaration of interest

The FOCUS study was funded by Ipsen Pharma.

T.L. has acted as a consultant for Ipsen, Astellas, Amgen and Novartis, and has been a Principal Investigator in trials funded by Ipsen. S.C has acted as a Principal Investigator in trials or contributed to trials funded by Amgen, Janssen, Merck, Pierre Fabre and Sanofi, and as a speaker for Astellas, Bayer, Ferring, Ipsen, Janssen, Pfizer and Sanofi. J.-L.D. has received funding from Ipsen, Ferring, Sanofi, Astellas and Janssen. C.H. has provided consultancy for Ipsen, Astellas, Takeda, Janssen, Sanofi, Amgen and Roche. J.-P.M has no conflict of interest to declare. J.-L.M. has served in advisory boards for Ipsen, Sanofi Pasteur, Ferring and Lilly. D.R. has received funding from Ipsen, Astra Zeneca, Lilly, Bouchara recordati, Sanofi and Takeda. M.Z. has received funding from Ipsen, Ferring, Sanofi, Janssen, Abbott and Recordati. A.M. was an Ipsen employee at the time of the study. IL has participated in advisory boards for Ipsen, Sanofi, Takeda, Ferring, Amgen and Astellas.

Acknowledgements

The authors thank all investigators in the FOCUS study. We also thank Yann Bourhis of MAPI, Lyon, France for statistical input. The authors take full responsibility for the content of the paper but thank Martin Gilmour of ESP Bioscience, Crowthorne, UK (supported by Ipsen) for editorial assistance in the preparation of the manuscript.