Nicolas Mottet, Clinique Mutualiste, 3 rue Le Verrier, BP 209, FR-42013 Saint Etienne Cedex 2, France. e-mail: email@example.com
Study Type – Therapy (RCT)
Level of Evidence 1b
What's known on the subject? and What does the study add?
Intermittent androgen deprivation therapy (ADT) involves cycling ADT, allowing hormonal recovery during off-treatment periods. This could lead to a better quality of life during off-treatment periods and could delay progression to castration resistance. Safety and feasibility of intermittent ADT have been shown but tolerability and health-related quality of life improvement have been suggested but questioned by others. Results from randomized trials, including relapsing or mixed populations, have suggested intermittent ADT to be as effective as continuous ADT.
In this study of only metastatic patients, no statistical difference in either overall survival or progression-free survival was shown between intermittent and continuous ADT and suggests that intermittent might be as safe as continuous castration. It could be an option in highly responding and well-informed metastatic patients even if no clear benefit in health-related quality of life was shown. This intermittent modality could be of interest in metastatic patients with significant treatment-induced side-effects.
• To compare intermittent androgen deprivation therapy (ADT) and continuous ADT after 6 months of induction of ADT in patients with metastatic prostate cancer (PCa).
PATIENTS AND METHODS
• This is an open-label randomized multi-centre study conducted in 58 centres in Europe.
• Patients with metastatic PCa and prostate-specific antigen (PSA) level >20 ng/mL at selection were randomized after 6 months of induction of ADT (leuprorelin and flutamide) if PSA level had decreased below 4 ng/mL.
• Patients received either continuous or intermittent ADT. All patients were treated until signs of disease progression under treatment or until study end with a monthly central PSA determination and follow-up visits were performed every 3 months.
• The primary endpoint was overall survival. Secondary endpoints included progression-free survival, health-related quality of life (QLQ C30 questionnaire) and safety criteria.
• Of 383 selected patients, 173 had a PSA level below 4 ng/mL after 6 months of induction of ADT and were randomized. Median overall survival (52 vs 42 months, P= 0.75) and median progression-free survival (15.1 vs 20.7 months, P= 0.74) were not significantly different between continuous and intermittent ADT.
• Although some differences in quality of life were observed, most of the functional and symptom scales showed no significant difference between the two groups.
• Significantly fewer treatment–emergent adverse events occurred in the intermittent group (P= 0.042), with the incidence of headache and hot flushes also lower.
• This first randomized trial comparing continuous with intermittent ADT in metastatic PCa suggests that intermittent ADT might be as safe as continuous ADT.
• It could be an option in highly responding and well-informed patients even if no clear benefit in health-related quality of life was shown.
Prostate cancer (PCa) is one of the principal oncological problems facing the male population . Continuous androgen suppression is the standard treatment for metastatic situations. Castration escape is constant after a median 36–40 months . Intermittent androgen deprivation therapy (ADT) involves cycling ADT with treatment cessation, allowing hormonal recovery during the off-treatment periods . This could lead to a better health-related quality of life (HRQL) during off-treatment periods and could maintain the hormonal responsiveness, delaying progression to castration resistance. The rationale of intermittent ADT was developed using models, notably the Shionogi murine model, which showed a threefold increase in androgen-sensitive period [3,4]. Safety and feasibility of intermittent ADT have been shown in prospective cohorts . Tolerability and HRQL improvement (including a reduced impact on sexual activity) have been suggested [6–8], but questioned by others . Results from phase 3 trials, including relapsing patients [10,11], or mixed populations of patients with locally advanced and metastatic disease [12–15], have suggested intermittent ADT to be as effective as continuous ADT.
This study aimed to evaluate the effects of intermittent ADT in patients presenting with metastatic PCa with a PSA level >20 ng/mL which normalized within 6 months of initial ADT. The results were compared with those obtained after continuous ADT.
PATIENTS AND METHODS
The protocol was approved by an independent ethics committee at a national level in France and at each investigational site in Germany, Bulgaria, Czech Republic and Slovakia. Eligibility criteria included a histologically confirmed metastatic PCa (Tx Nx M1 ≠ M1a), with bone or measurable visceral metastases, requiring first-line ADT; PSA level at least fivefold higher than normal (i.e. PSA ≥20 ng/mL, Hybritech radioimmunoassay); Eastern Cooperative Oncology Group (ECOG) performance status ≤2; liver enzymes <2.25-fold higher than upper normal levels; and a life expectancy greater than 9 months. Any previous medical or surgical castration, a bilateral adrenalectomy or hypophysectomy, another cancer (except basocellular carcinoma) in the previous 5 years, a serious unstable progressive disease or any experimental treatment in the 3 months before inclusion were all exclusion criteria.
The study was designed as an open-label, comparative, randomized, prospective, multicentre study conducted in Europe. The study was divided into two therapeutic phases. During the first phase, eligible patients received complete androgen suppression (leuprorelin SR 3.75 mg, one subcutaneous injection every 28 days and flutamide, one 250 mg tablet, three times daily). After 6 months of induction therapy eligible patients with no sign of clinical progression and a PSA level <4 ng/mL entered the second phase and were randomized (1:1, central randomization) to either continuous or intermittent ADT.
In the continuous arm, the same ADT treatment continued until disease progression or study end. In the intermittent arm, ADT treatment was discontinued at randomization and resumed for consecutive 3-month periods as soon as the PSA level exceeded 10 ng/mL or there was clinical progression. The treatment was stopped again for the next cycle using the same criteria as for the induction period. A treatment cycle was composed of the ‘off-period’ and the following ‘on-period’. The ‘on-period’ was defined as the date of the first injection of the cycle until the 30th day after the last ADT injection. The ‘off period’ was defined from the 30th day after the last injection until the date of the next injection.
Level of PSA was centrally determined every month and follow-up visits were scheduled at a 3-month interval until progression or study end.
When the PSA level exceeded 10 ng/mL during an ‘off’ treatment period in the intermittent arm, a special visit was scheduled to resume the ADT. Subsequent visits were resumed at a 3-month interval from the time of the special visit.
During each visit, HRQL, ECOG performance status, Lukacs questionnaire on sexual function (visual analogue scale) , symptoms and tolerance to treatment were collected. Centralized (per country) laboratory blood tests were performed before each visit (blood count, testosterone and liver function). In both groups, imaging analyses (bone scan, liver ultrasound, chest X-ray) were performed before pre-assessment, at study entry and during on-therapy progression.
After progression under treatment patients were treated according to local clinical practice. Subject or subject's family was to be contacted every 6 months until end of study for life status.
The primary efficacy endpoint was overall survival (OS) after end of the induction phase. Secondary efficacy endpoints included progression-free survival (PFS) (either biochemical, clinical or withdrawal because of progression) and overall HRQL (EORTC-QLQ C-30 version 2) .
There was no consensus on definition of biochemical progression when the trial was built in 1996 , also biochemical progression under therapy was defined as a PSA level >4 ng/mL and increasing in three successive tests 1 month apart .
Safety variables included treatment-emergent adverse events, vital signs (heart rate, blood pressure and weight), and biological evaluations.
The protocol was designed to detect a minimal increase from 50%  to 70% in the 30-month OS rate, corresponding to an expected 20% improvement in the intermittent arm. To detect this difference, 85 patients in each treatment group had to be followed until a total of 88 deaths had occurred (α= 5%, β= 15%). With an estimated 50% non-inclusion rate at 6 months, a total of 360 patients had to be selected.
Event rates are expressed as the percentage of events per follow-up year, taking into account the censoring of follow-up data. Survival curves and 95% CI of median survivals were estimated using the Kaplan–Meier technique. All comparisons were made by means of a two-sided log-rank test with a 0.05 significance level. Data were analysed according to the intention-to-treat principle.
The HRQL functional and symptoms scales were calculated as needed . Each scale and change from baseline was summarized by visit and compared between treatment groups using analysis of covariance with baseline result as covariate. For the intermittent group, the HRQL scales were compared between time ‘off’ and ‘on’ treatment within each cycle.
All analyses are based on observed data with the assumption that missing data were missing completely at random.
Statistical analyses were conducted with the use of SAS software, version 9.1 (SAS Institute, Cary, NC, USA).
Between December 1996 and September 2005, 383 patients were pre-included after providing written informed consent. All the patients (n= 42) from two sites were excluded from the efficacy analysis after a quality audit, because they were treated outside their assigned treatment arm, and without any clear information on the treatment effectively received. Therefore, 341 subjects were pre-included in the non-randomized induction phase, which lasted for 6 months. This included 10 subjects who were pre-included but withdrawn because of lack of information concerning their treatment as they did not come back to following visits. There were 173/341 (50.7%) subjects randomized from 58 sites (26 sites in France, 27 sites in Germany and five sites in Eastern Europe). Overall, 173 patients were included in the intention-to-treat analysis (Fig. 1). Reasons for non-randomization included PSA level ≥4 ng/mL after 6 months of treatment (51.3%), disease progression during treatment (24.7%), adverse events (13.9%), other reasons (9.5%) and lost to follow-up (0.6%).
The two groups were similar in terms of age, weight, testosterone, PSA level, ECOG score and mean Gleason score at treatment induction. All patients were classified as metastatic; most had extended metastatic disease based on the Crawford classification  (Table 1).
Table 1. Summary of patient characteristics at Month 0 (Visit 0): intention-to-treat analysis set
Continuous ADT(n= 83)
Intermittent ADT (n= 86)
Total randomized (n= 173)
Non- randomized (n= 158)
Data are given as mean ±sd unless indicated as n (%). ADT, androgen deprivation therapy; ECOG, Eastern Cooperative Oncology Group.
69.2 ± 9.4
68.7 ± 9.1
68.8 ± 9.3
69.6 ± 8.55
76.6 ± 11.8
76.7 ± 11.6
76.8 ± 11.6
74 ± 12.97
ECOG performance status, 0; n (%)
Gleason histological classification
7.2 ± 1.4
6.9 ± 1.6
7.1 ± 1.5
7.3 ± 1.34
M1b, n (%)
Disease classification – Crawford
Minimal disease, n (%)
Severe disease, n (%)
Month 0 (V0)
3.94 ± 1.71
4.37 ± 1.67
4.15 ± 1.68
3.855 ± 1.85
Month 6 (V2)
0.25 ± 0.24
0.24 ± 0.15
0.25 ± 0.20
Prostate-specific antigen, ng/mL
Month 0 (V0)
633.1 ± 1377.2
558.7 ± 1134.5
610.8 ± 1251.1
835.8 ± 1121
Month 6 (V2)
0.8 ± 1.0
0.6 ± 0.8
0.7 ± 1.0
The efficacy evaluation showed a mean (median) follow-up of 3.9 (3.7) years.
The Kaplan–Meier estimate of median OS in the continuous ADT group was 52.0 months (95% CI 47.9–67.4 months) compared with 42.2 months (95% CI 37.4–62.3 months) in the intermittent group (P= 0.75) (Fig. 2). There were 45 deaths in the continuous treatment group and 49 in the intermittent treatment group.
The median overall PFS was 15.1 months (95% CI 12.1–22.7 months) in the continuous treatment group and 20.7 months (95% CI 13.9–25.4 months) in the intermittent treatment group (P= 0.74) (Fig. 2). There were 67 events in the continuous group and 70 in the intermittent group. No clinical progression without progression of PSA was observed.
Although the analysis of the HRQL scores showed some significant differences between the two ADT groups at some visits, no difference was clinically relevant and no general trend was observed (Fig. 3). Furthermore, no difference was observed during the off-therapy and on-therapy periods (cycles 1–10, Fig. 4).
Lukacs' questionnaire scores were similar in the two therapy groups, with no significant differences shown for most comparisons. However, any significant differences found tended to favour the intermittent ADT group. In particular it should be noted that ‘strong sexual desire’, ‘satisfied with your erections’, and ‘satisfied with your sex life’ scores were significantly better in the intermittent group at certain visits.
For cycle 1 of intermittent ADT, the mean ‘off-therapy’ duration was 126 days, representing a mean percentage of ‘off-therapy’ cycle duration of 54.6%. At the first visit after randomization, around 3 months later, the continuous therapy group showed testosterone levels remaining at randomization levels (median change 0 ng/mL), whereas the intermittent therapy group showed an increase in testosterone level (median change 4.1 ng/mL, to approximately baseline levels). Mean testosterone value was 4.83 ng/mL during ‘off-therapy’ and 0.29 ng/mL during ‘on-therapy’ (Table 2). Cycles 1–7 are summarized in Table 2. A progressive decrease in the off-period duration was apparent with an increase in the number of cycles, up to cycle 7.
Table 2. Intermittent androgen deprivation therapy group: summary of ‘off-period’ for cycles 1 to 7
Number of days OFF therapy
Percentage of days OFF therapy
Testosterone OFF therapy
Testosterone ON therapy
Based on all the available data up to cycle 15. Mean and median values calculated per cycle for patients whose testosterone level was more than (for OFF therapy) or less than or equal to (for ON therapy) the required testosterone level (i.e. 0.5 ng/mL) during that cycle. Fewer than 10 patients reached cycle 8 and above.
When evaluating safety, the incidence of treatment-emergent adverse events in the continuous treatment group (88/94 patients; 93.6%) was significantly higher than in the intermittent treatment group (81/96 patients; 84.4%) (P= 0.042). In the continuous ADT group, the most frequently reported adverse events were hot flushes (63.8% of patients), headache (46.8%), lumbar pain (13.8%) and joint pain (13.8%). In the intermittent ADT group considered as a whole, the most frequently reported adverse events were hot flushes (60.4% of patients), headache (32.3%), bone pain (13.5%) and lumbar pain (12.5%). During the induction phase, mean scores for headache and hot flushes were similar in the two groups. However, at the first post-randomization visit (Visit 3), headache (6.9 vs 3.2) and hot flush (28.3 vs 8.7) scores tended to be lower in the intermittent group. This trend continued at subsequent visits until Visit 10 for hot flushes and Visit 11 for headache, when scores were similar.
Serious adverse events were reported in 28 (29.8%) patients in the continuous group and 30 (31.3%) patients in the intermittent group. Most (96%) of the serious adverse events were considered to be unrelated to study therapy. The ADT was discontinued because of adverse reactions in nine (9.6%) patients in the continuous treatment group and seven (7.3%) patients in the intermittent treatment group.
To our knowledge, this is the first designed trial devoted to only patients with metastases . In this study, no statistical difference in either OS (42.2 vs 52 months) or PFS (20.7 vs 15.1 months) was shown between intermittent ADT and continuous ADT. The non-significant difference does not preclude a difference favouring the continuous arm. In fact, the present study was underpowered as expected differences were truly overestimated at the time of the study design in 1996. Furthermore the inclusion of only 341 patients compared with the 360 patients initially planned also reduces the statistical power of the study. Nevertheless, the two OS curves cross each other several times (Fig. 2), making differences in OS between both therapeutic modalities unlikely. These results agree with those of all other published trials with available results on the subgroup of metastatic patients. In none of those trials did intermittent ADT prolong median time to tumour progression [10,14,15], PFS , OS [24,25], or time to death . On the other hand, and in contrast with the present study, those trials show no trend in favour of intermittent ADT treatment. Results from the Southwest Oncology Group (SWOG) JPR7 study in relapsing non-metastatic patients were presented in 2011 and did not show either any significant difference in median OS (9.1 vs 8.8 years in continuous and intermittent ADT groups, respectively, hazard ratio 1.02, 95% CI 0.86–1.21) in relapsing patients . A definitive answer on the difference in survival and benefits between continuous and intermittent ADT in metastatic patients is expected from the large SWOG 9346 trial with 1500 patients . It might be of interest to highlight that in the SWOG JPR7 trial , as in the South European Uroncological Group trial , the OS was not significantly different between the two treatment modalities. In both trials however, there was a non-significant trend favouring continuous ADT for specific survival, and a non-specific trend favouring intermittent ADT for the non-prostate cancer death (either cardiovascular or non-cardiovascular).
So far no available trial has ever been associated with a decreased OS with intermittent ADT. Although intermittent ADT is not deleterious, it is not associated with the expected time to progression benefit expected from studies in animals and in vitro. This clearly highlights the limitations of animal models.
Besides, patients who did not achieve a normalized PSA level (i.e. <4 ng/mL) after 6 months of ADT had worse prognosis compared with those achieving a low PSA [5,26,27] and were not randomized in the present study. Therefore, in metastatic PCa, intermittent ADT could be considered only in patients with a good initial PSA response.
Intermittent ADT led to an intermittent castration with the recovery of testosterone during off-therapy periods in most patients. Leuproreline was administered monthly in the present study as the 3-month formulation was not the standard of care in all countries at the time of study initiation. The lack of a clear survival benefit of combined androgen blockade  makes its use in the present trial questionable. However, it must be acknowledged that this combined modality was commonly used in most previously reported series .
Although tolerability and HRQL have been shown to favour intermittent ADT in some reports [14,15], the present study indicated generally no difference in HRQL between the groups. This could be related to the short off-therapy periods and consequently short normal testosterone level periods. Less advanced PCa could have led to more prolonged off-therapy periods , and longer periods of normal testosterone levels. Prolonged periods with normal testosterone levels could lead to either an improved HRQL during the off-therapy periods or a decrease in long-term adverse events, especially those related to bone or metabolic impact. The publication of specific trials, such as the SWOG JPR7 , will address this question in a population that could represent the best target for intermittent ADT.
A lower incidence of treatment-emergent adverse events after intermittent ADT was observed in the present study and hot flushes and headache (two of the most common side-effects of ADT [29,30]) were less prevalent during intermittent ADT. Therefore metastatic patients with a good biochemical response and suffering from ADT side-effects should be considered for intermittent ADT. Other potential benefits, suggested by some authors, were not addressed in this trial, such as the libido and erection benefits of intermittent ADT . The relative bone protection of the intermittent modality was suggested several years ago . It has been recently clarified , with a decrease in bone mineral density under ADT, but to a lesser extent in the intermittent treatment group compared with the continuous treatment group. The reality of such a benefit associated with short ‘off’ periods remains highly theoretical in these advanced situations.
Finally, the possible overall cost saving of intermittent ADT compared with continuous ADT should be specifically analysed. Fewer drugs are used with intermittent ADT, but more biological examinations and consultations are needed, especially in advanced PCa. Cost savings might be higher in patients with less advanced disease, such as relapsing patients .
In conclusion, this randomized trial comparing intermittent and continuous ADT in metastatic PCa suggests that intermittent therapy might be as safe as continuous castration. It could be an option in highly responding and well-informed patients even if no clear benefit in HRQL was shown. This intermittent modality could be of interest in patients with significant adverse events induced by the treatment. So far this is the only small available observed benefit.
This study was supported by Laboratoires Takeda, France and Takeda Pharma Germany. We thank Marion Goussard for her tremendous help and B. Darné for editorial assistance. Finally, we would like to thank all the physicians from the TAP22 investigator group and ultimately the patients. TRIAL REGISTRATION NCT00817739 (http://clinicaltrials.gov/).
CONFLICT OF INTEREST
Nicolas Mottet is a paid consultant and a board member for Takeda Pharmaceuticals. He has also received research funding from Takeda Pharmaceuticals.
BULGARIA: K NEYKOV, National Hospital for Treatment of Haemotological Diseases, Sofia; V TABAKOV, National Hospital for Treatment of Haemotological Diseases, Sofia. CZECH REPUBLIC: J HYNCICA, Zlin; O KOHLER, Central Military Hospital Prague, Strešovice. FRANCE: C C ABBOU, Hôpital Henri Mondor, Créteil; J-P ANSIEAU, Hôpital E. Muller, Mulhouse; C AVANCES, C.H.U. de Nîmes, Nîmes; J BENCHETRIT, Clinique Croix Saint-Michel, Montauban; P BERLIZOT, Hôpital du Val de Grace, Paris; T BILLEBAUD, CHI Créteil, Créteil; L BOCCON-GIBOD, Hôpital Bichat, Paris; P BONDIL, CH de Chambéry, Chambéry; O BOUCHOT, CHU-Hôpital Hôtel-Dieu, Nantes; F BOUCHOU, Tours; C CASSÉ, Polyclinique Du Parc, Cholet; P COLOBY, Centre Hospitalier René Dubos, Pontoise; P COLOMBEL, Hôpital Le Mans, Le Mans; L CORBEL, Clinique Ste-Jeanne D'Arc, Saint-Brieuc; J-P COSSON, Cabinet Médical, Périgueux; O DELBOS, Les Bureaux du Polygone, Montpellier; J-L DESCOTES, CHU-Hôpital Michallon, Grenoble; J GUITER, CHRU-Hôpital la Peyronie, Montpellier; F IBORRA, Les Bureaux du Polygone, Montpellier; J-L JUNG, Hôpitaux Civils de Colmar, Colmar; M LEPELLEY, Auch; J-M LHEZ, Clinique Pasteur, Toulouse; S LOULIDI, Auch; N MOTTET, Clinique Mutualiste, Saint-Etienne; F MOULINIER, Polyclinique Saint-Come, Compiègne; R SOULIÉ, Clinique du Château, Toulouse; P TEILLAC, Hôpital Saint-Louis, Paris; J-M TURBLIN, Centre Hospitalier Général, Saint-Quentin; J VAN DAMME, CHG de Lons le Saunier, Lons Le Saunier. GERMANY: M BEINTKER, Universität Jena, Jena; H BERTERMANN, Städt. Krankenhaus, Kiel; K-H BICHLER, Klinikum der Eberhard-Karls-Universität, Tübingen; U EICKENBERG, St Franziskus Hospital Bielefeld, Bielefeld; FEHL, Butzbach; P-J FUNKE, Ev.Jung-Stilling-Krankenhaus, Siegen; W HÄGELE, Kall; P HAMMERER, Univ.-Klinik u. Poliklinik, Hamburg; E HAHN, Allgemeines Krankenhaus, Hagen; H-J HEINRICHS, Städt. Krankenhaus Martha-Maria, Halle/S; HEYNEMANN, Martin-Luther-Univ. Halle-Wittenberg, Halle; LAHM, Heilbronn; LEITENBERGER, Zentral-Krankenhaus St Jürgen-Straße, Bremen; V MOLL, Augsburg; MUSCHTER, Diakonie-KH, Rotenburg; U REBMANN, Anhaltischen Diakonissenanstalt, Dessau; RIEDEL, Klinikum Neubrandenburg, Neubrandenburg; S ROTH, Lehrstuhl für Urologie der Universität, Wuppertal; C RÜSSEL, Borken; SCHUBERT, Universität Jena, Jena; SEITER, Urologische Klinik der Universität Rostock, Rostock; SIMSON, Lauenburg; M STÖCKLE, Kinderurologie der Universitätskliniken des Saarlendes, Homburg/Saar; U-W TUNN, Städtische Kliniken Offenbach, Offenbach; H VAN AHLEN, Städt. Kliniken Osnabrück, Osnabrück; W WIELAND, Caritas-Krankenhaus St Josef, Regensburg, SCHAUB, Bocholt; M-P WIRTH, AkademieCarl Gustav Carus, Dresden; J-M WOLFF, Einrichtungen der RWTH Aachen, Aachen. SLOVAKIA: F GONCALVES, Faculty Hospital Bratislava, Brastilava Kramare.