European Urology

1. Introduction

The standard treatment for advanced prostate cancer (PCa) has been androgen-deprivation therapy (ADT). Well-known side effects of ADT are hot flushes, night sweats, erectile dysfunction (ED), decreased libido, fatigue, depression, and gynaecomastia. In addition, decreased haemoglobin levels, changes in fat and lean body mass, changes in plasma lipoproteins, increased insulin levels, and osteoporosis occur [1].

The FinnProstate Study VII, a randomised open clinical multicentre trial, was planned to compare intermittent androgen deprivation (IAD) and continuous androgen deprivation (CAD) in the treatment of advanced PCa [2]. One of the objectives was to evaluate any differences in health-related quality of life (QoL) between treatment arms. We present our results comparing changes in QoL and the prevalence of adverse drug reactions (ADRs) and adverse events (AEs) in the IAD and CAD arms.

2. Materials and methods

Between May 1997 and February 2003, 852 patients with locally advanced PCa and prostate-specific antigen (PSA) >20 ng/ml or with metastatic PCa at any PSA level were prospectively enrolled to receive ADT for 24 wk, goserelin acetate 3.6 mg subcutaneously every 28 d, and antiandrogen cyproterone acetate 100 mg twice daily during the first 12.5 d to minimise flare reaction. Our 554 patients, whose PSA decreased to <10.0 ng/ml or at least by 50% (baseline PSA: <20.0 ng/ml), were randomised 1:1 to IAD or CAD. In the CAD arm, 280 patients continued with goserelin acetate or underwent bilateral orchiectomy. In the IAD arm with 274 patients and one patient refusing the randomised intermittent treatment, ADT was withheld after randomisation, resumed as initially for at least 24 wk whenever PSA increased >20.0 ng/ml or above the baseline, and withheld again by the same criteria as for randomisation. The treatment cycle was defined as time off treatment (TOFF) plus time on treatment (TON). Our results concerning time to progression, overall and PCa-specific survival, and time to treatment failure appeared in June 2012. We could not show any significant differences between treatment arms, but risk analysis showed a mild advantage to IAD [2].

QoL was monitored using a validated and self-administered 30-item questionnaire addressing 10 domains (Appendix 1) [3]. Patients continued to the last domain if they answered yes to question 27. The sum of the numerical values of answers in each domain was recorded. For statistical analysis, answers for questions 8, 10, 13, and 27 were renumbered in reverse. In summary, lower scores indicated better health in the domains of pain, activity limitation, bed disability, physical capacity, and sexuality. Higher scores indicated a favourable outcome in the domains of social functioning, emotional well-being, vitality, overall health, and sexual functioning.

Serum testosterone and PSA were measured systematically every 3 mo. QoL questionnaire scores, testosterone, and PSA were analysed and summarised at the end of each TOFF and TON in the IAD arm and approximately at the same point of time in the CAD arm. The approximate time point was defined by calculating the mean durations of previous cycles and the mean duration of the present TOFF or cycle. Patients in the CAD arm were selected by taking into account the visit closest to this point. QoL was analysed by the Mann-Whitney U test, 0.5 standard deviation (SD) rule [4], and repeated measures analysis of variance. Any ADR, AE, or serious adverse event (SAE) was inquired about at each visit. These were summarised by the preferred term and primary body system (system of organ classes), according to the Medical Dictionary for Regulatory Activities [5]. Differences in the prevalences of ADRs and (S)AEs were estimated by the chi-square test. All statistical tests were two sided at a 5% significance level.

3. Results

Treatment arms were comparable regarding advancement of PCa, differentiation grade, performance status, PSA, testosterone, and QoL at entry and at randomisation (Table 1). They were also balanced regarding concurrent diseases that might predispose for cardiovascular (CV) events or fractures. Eight patients in the IAD arm (2.9%) and 11 in the CAD arm (3.9%) used bone-specific treatment during the trial (p = 0.51). The median follow-up time was 65 mo, maximum 11.6 yr, with one patient reaching the 14th treatment cycle in the IAD arm. The number of patients decreased from cycle to cycle, with mean and median durations of each TOFF also decreasing over time (Table 2). The response rates for the QoL questionnaire domains 1–9 were 86–92% at entry and at randomisation, and it was lowest at 73% in the IAD arm and 69% in the CAD arm during the first five cycles (Table 3).

The most frequently detected significant differences in QoL between treatment arms related to activity limitation, physical capacity, and sexual functioning, favouring IAD (Fig 1 and Fig 2). A nonsignificant trend in favour of IAD was also found in other domains, except sexuality. The response rate for the last domain (sexuality) was low. The proportion of respondents in domain 9 (sexual functioning) that continued to the last domain was 48.8% in the IAD and 40.1% in the CAD arm at entry but decreased in both arms thereafter. In the IAD arm, the proportion of respondents was clearly higher at the end than at the beginning of each TOFF. In the CAD arm, response rate was approximately 20% (Fig. 3).

Fig. 1 Differences in quality of life between treatment arms according to (a–d) Mann-Whitney U test (*p < 0.05) and (e–h) the 0.5 standard deviation rule in the domains of activity limitation, physical capacity, sexual functioning, and sexuality. Lower scores indicate better health in activity limitation, physical capacity, and sexuality; higher scores indicate better sexual functioning. SD = standard deviation; TOFF = treatment-off phase; TON = treatment-on phase.

Fig. 2 Differences in quality of life between treatment arms in the domains of activity limitation (n = 48 [IAD] vs 44 [CAD]; p = 0.06], physical capacity (n = 48 vs 44; p = 0.07), and sexual functioning (n = 46 vs 41; p = 0.003) analyzed by repeated measures analysis of variance with covariant of visit 3 (randomisation). Lower scores indicate better health in activity limitation and physical capacity; higher scores indicate better sexual functioning. CAD = continuous androgen deprivation treatment arm; IAD = intermittent androgen deprivation treatment arm; TOFF = treatment-off phase; TON = treatment-on phase.

Fig. 3 Proportion of respondents in domain 9 (sexual functioning) who continued to domain 10 (sexuality) in intermittent androgen deprivation (IAD) and continuous androgen deprivation (CAD) treatment arms.

No significant differences emerged in the prevalence of AEs. CV events were the most prevalent, with 154 in the IAD and 162 in the CAD arm (Table 4). Table 5 shows the numbers of patients experiencing any CV AE or bone fracture and those withdrawn from the trial or who died because of SAEs. In the final survival analysis, 78 patients of 554 (14.1%) died from any CV cause (20% of all 392 deaths): 35 in the IAD (12.8%) and 43 in the CAD arm (15.4%) (p = 0.38). In both treatment arms, 19 fractures were recorded. In the CAD arm, one patient had four and one patient had two fractures (Table 6). The most common ADRs were hot flushes or night sweats. ED and depressed mood were reported more often in the IAD arm (Table 7).

Mean and median plasma testosterone remained at a low level (<1.0 nmol/l) in the CAD arm after randomisation but showed recovery at the end of each TOFF in the IAD arm. Testosterone was decreasing during TONs and increasing during TOFFs but did not reach the same level as at the end of the previous TOFF. At entry, 81.2% of patients in the IAD arm had a testosterone level ≥10 nmol/l, but only 47.4% at the end of the 10th TOFF.

4. Discussion

The present trial showed IAD to offer some benefit in QoL when compared with CAD, especially in activity limitation, physical capacity, and sexual functioning. Although the number of AEs was higher in the CAD arm, no significant differences emerged between treatment arms.

A discrepancy emerged between the domains of sexual functioning and sexuality, with sexual functioning better in the IAD but sexuality better in the CAD arm. This may be due to low response rates for questions 28–30. The last domain was answered only if question 27 (“Have you tried to engage in any type of sexual activity including masturbation or intercourse during the past month?”) was answered yes. Because of that and because of the decreasing number of patients within the trial over time, the response rate was low. Cleary et al. reported generally excellent and consistent response rates for individual questions, except for sexual interest and sexual functioning, and they speculated either misinterpretation or acceptability of these types of questions. They reported the measures of sexual interest and functioning as showing low and inconsistent correlations with general health perceptions, suggesting these are less important determinants of patients’ health perceptions than many of the other domains [3]. Boccardo et al, using the same kind of questionnaire in their study in patients with advanced PCa, also found results concerning sexual interest and sexual functioning somewhat confounding. They suggested that preservation of normal testosterone levels is not the only factor required to maintain sexual function in elderly men [6]. Our study supports this finding. The response rate for the last domain was approximately 20% in the CAD arm despite continuous ADT. In the IAD arm, more patients answered yes to question 27 at the end than in the beginning of each TOFF, implying improvement in sexual functioning during TOFF.

Only a few published phase 2 [7], [8], [9], [10], and [11] and phase 3 trials [12], [13], and [14] concerning IAD evaluated QoL systematically with a formulated questionnaire. Commonly used is the European Organisation for Research and Treatment of Cancer QoL questionnaire QLQ-C30 [15], chosen for the SEUG trial 9401. They reported fewer major side effects of hot flushes and gynaecomastia in the IAD arm. Surprisingly, the QoL figures, except for sexual quality, were slightly lower in the IAD arm [12]. Mottet et al. did not find any clinically relevant differences in QoL between IAD and CAD in patients with metastatic PCa [14].

Interpretation of the significance of changes in QoL scores is not easy. Osoba et al. developed a subjective significance questionnaire to determine the importance of numerical changes in the QLQ-C30 scores that the subjects themselves considered important. Despite all these efforts, patients may not interpret the terminology in the way it was intended [16]. Later evidence-based guidelines appeared to help determine sample sizes in clinical trials and to interpret differences in QLQ-C30 scores, making this questionnaire more applicable in future trials [17]. The questionnaire we used was chosen because it was designed for multinational use and to explore the value of ADT for advanced PCa. To our knowledge, there are no definitions for clinically important differences in our questionnaire, which is one of the limitations in our study. The way to define the time point for analysis in the CAD arm was somewhat arbitrary and may have caused bias. In the Mann-Whitney U test with a higher number of patients, the significant differences did not emerge constantly during TOFFs but also sometimes during TONs, suggesting that even short interruptions of ADT compared with CAD may have a beneficial effect on QoL over time. We tried to find any minimally important differences between treatment arms using the 0.5 SD rule. The changes were mostly <0.5 SD when compared with the baseline SD but similar with the results of the Mann-Whitney U test. In repeated measures analysis of variance with identical results, only the first five cycles were included due to the decreasing number of patients over time and 95–103 patients analysed at the end of the fifth TON. The domain of sexuality could not be evaluated due to the low number of patients.

Hot flushes and night sweats were our most frequent ADRs. The only significant differences emerged in the number of patients reporting ED and mood depression that were more common in the IAD group, which was unexpected. This may be due to the way ADRs were assessed by questioning: “Has anything bothered you since your last visit?” At entry, only 48.8% in the IAD and 40.1% in the CAD arm continued to the last domain of the QoL questionnaire, meaning that <50% of patients were sexually active when starting the trial. Patients may have grown accustomed to their symptoms and no longer felt bothered, especially with CAD. We did not analyse the relief of ADRs during TOFF but relied on QoL analysis in this respect.

Several studies have demonstrated ADT having an adverse effect on CV risk factors such as body fat and lean body mass, serum lipoproteins, insulin sensitivity, and obesity [18], [19], and [20], thus raising the odds of CV morbidity by up to 20% [21]. However, association between ADT and CV mortality is still controversial [22], [23], and [24]. CV events were the most common SAEs in our trial. No statistically significant differences emerged in the prevalence of (S)AEs, the number of patients experiencing CV SAEs, or the incidence of deaths caused by any SAE or CV event.

ADT is associated with loss of bone mineral density at multiple skeletal sites, progressive with increasing duration of ADT but most significant during the first years [25] and [26]. ADT is also associated with an excess risk for fractures or hospitalisation as a consequence of a fracture [27], [28], and [29]. In the present trial, no significant differences emerged in the incidence of fractures, which may be due to a relatively small number of events.

Testosterone recovery is considered essential for improvement in QoL and relief from ADT-induced ADRs [30]. In our study, testosterone levels showed recovery at the end of each TOFF but did not reach the same level as at the end of the previous TOFF. The figure demonstrating changes in the proportion of respondents in the last domain of QoL questionnaire showed an association between recovering plasma testosterone and increasing sexual activity during TOFFs. Despite testosterone recovery during TOFFs in the IAD arm, we found no statistically significant differences in AEs, SAEs, or their consequences between treatment arms. The explanation may be that despite intermittent dosing and shorter exposure time for ADT in the IAD arm, the TOFF duration decreased from cycle to cycle, and patients returned to ADT sooner and sooner repeatedly.

5. Conclusions

Intermittent administration of ADT in patients with advanced PCa is advantageous in respect to QoL when compared with continuous castration, especially seen in activity limitation, physical capacity, and sexual functioning. Incidences of SAEs, fractures, CV events, or CV deaths were not significantly lower with IAD, however.

Author contributions: Arto J. Salonen had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Salonen, Tammela, Ala-Opas.

Acquisition of data: The FinnProstate Group including Salonen, Taari, Ala-Opas, Viitanen, Lundstedt, Tammela.

Analysis and interpretation of data: Salonen, Tammela.

Drafting of the manuscript: Salonen, Tammela, Taari.

Critical revision of the manuscript for important intellectual content: Salonen, Tammela, Taari.

Statistical analysis: Salonen.

Obtaining funding: None.

Administrative, technical, or material support: None.

Supervision: Salonen, Tammela, Taari.

Other (specify): None.

Financial disclosures: Arto J. Salonen certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: Arto J. Salonen has received speaker honoraria from Astellas, Astra Zeneca, and Pfizer; consultant honoraria from UpViser; taken part in congresses with support from Astellas, GSK, Novartis, Sanofi-Aventis, Abbott, and Astra Zeneca. Kimmo Taari has received speaker honoraria from GSK; consultant honoraria from Astellas, GSK, Ferring, and Amgen; taken part in congresses with support from Sanofi-Aventis, Pfizer, and Astellas. Martti Ala-Opas has received speaker honoraria from Astra Zeneca and GSK; consultant honoraria from Amgen, Astra Zeneca, and Novartis; taken part in congresses with support from Abbot, Amgen, Astra Zeneca, and Ferring. Jouko Viitanen has received speaker honoraria from Astra Zeneca, Astellas, and Pfizer; consultant honoraria from Amgen; taken part in congresses with support from Amgen, Astra Zeneca, Pfizer, Astellas, and Abbott. Seppo Lundstedt has received speaker honoraria from Astellas, GSK, and Pfizer; consultant honoraria from Photocure; taken part in congresses with support from Abbott. Teuvo L.J. Tammela has received speaker honoraria from Abbott and GSK; consultant honoraria from Amgen, Astellas, Orion, GSK, and Sanofi-Aventis; taken part in congresses with support from Amgen and GSK.

Funding/Support and role of the sponsor: Astra Zeneca helped manage the data.

Acknowledgement statement: The authors acknowledge both Jyrki Ollikainen and Hanna L. Koskinen from the University of Tampere who performed the statistical analysis, and the urologists and nurses in the trial centres.