MAI4
`
`lilntarnfide Versus Predniscne in Patients With itrostatc
`Cancer‘ fiymntornatically Progressing Alter fitndrogen-«
`rlshlative Therapy: .51 Phase EH Study of the European
`flrganization for Research and “treatment cl" flancer
`Genfttonrinary fironn
`
`By S. D. F0556, P. H.Tl'1. Slee, M. Brousi, S. Horenblos, R. R. Hall, J. W. Helheringlon, N. Aaronson, L. de Pl'llCl(, and L. Collette
`
`Purpgge: Time to progression £TTP), overall survival,
`and quality of life (GL1 were compared in patients with
`hormone-resistant prostate cancer {HRPCJ treated with
`prednisone (5 mg orally, fourtimes a day) or Flutumide
`(250 mg orally, three times a day).
`Patients and Methods: Symptomatic patients were
`randornlzed to receive either prednisone {I01 patients)
`or flutomicle (100 patients}. Subjective response was
`assessed based on performance status, the use of an-
`algesics, and the need to apply alternative palliative
`treatment. Prostote—specific antigen lPSA)~l:msed blo~
`chemical response (2 50% reduction of baseline PSA]
`was recorded. At baseline and at 6-week intervals
`during follow-up, patients completed the European On-
`ganizotion for Research and Treatment of Cancer Quol-=
`ity of Life Questionnaire C-30.
`Ewgsyfi: There was no difference between the
`groups in median TIP (prednisone, 3.4 months; flut-
`amide, 2.3 months} or overall survival (prednisone,
`10.6 months; flutamide, 11.2 months).
`In the pred—
`
`nisone group, 56% of the patients experienced cl sub~
`[active response, compared with 45% in the Flutarnicle
`group [P 2 .18). The median response duration was 4.3
`‘months for prednisone and 4.2 months for Hutamide. A
`biochemical responsewas observed in 21% and 23% of
`the prednisone and flutomldc groups, respectively.
`Gastrointestinal toxicity was the reason for trial discon-
`tinuation in seven patients receiving flutomide and two
`patients receiving prednisane. The QL assessment pa-
`rameters favored the use of prednisone with stal'isti-
`cctlly significant differences in pain, Fatigue, role func-
`tioning, appetite loss, gastrointestinal distress, and
`overall QL.
`In symptomatic HRPC, treatment with
`Conclusion:
`prednisone or Flutamide leads to similar rates of Tl'P
`and overall survival and no difference in subiectlve or
`biochemical response. The Ql. results favor the use of
`lowrcost prednisone in patients with HRPC.
`J Clin Oncol
`l9::'52~7I.
`ca 2001 by American
`Society of Clinical Oncology.
`
`FPROXIMATELY 70% to 80% of patients with
`advanced prostate cancer respond initially or remain
`stable when treated by medical or surgical casEr£1tion,but in
`20% to 30% of the patients,
`the rnalignancy progresses
`despite primary androgen deprivation. In addition, disease
`
`From the D.r.{paI'ImerIr of Medical Oncology and Radiotherapy,
`.‘\lor'wegiw1 Rcrrlium Hospital, Oslo, Norway,‘ .5‘! /lnlmrius Hospital,
`Nicuwegcin, and Nelhcrlattds Ctmccr Hospital, llllltltfl van l.ecmven~
`lroel: Hospital, Amslerdam. rlre Nctlterlands.‘ San Iiajjfizcle. Milan, Italy,‘
`FI'r:eI1-rrrrl H0.i'pi!ul, N¢:wr:a.)'.rle. and }”rr'nce.rS Royal
`/"10.5‘pl‘ic.'l,
`.l'.h{ll,
`Unired Kr'ng.:1'mn,- mm‘ European 0»-gaztizcrionfor Researclz and Track
`men! of Cancer Data Center, Brussels, Helgizzm.
`Submitted March 24, 2000,‘ cccr~,;med July 1'7, 2000.
`Supported by grants no.
`5U.l0 0111488-H through 5Ul0
`C/1ll'4S«!l-29from the National Cancer lmrliirlie, Btztltcscla. MD.
`The c0r1m7t.r cfllzis arriclc are solely the rc.rp0n.rl'bi!'ity aftlw auiltm-.5
`and do not necessarily I2-'pm?scnI the official views of the Nmlonal
`Cancer inmlmc.
`Acldrers rcprir.-r :'cr,u:c.vr.r In Sophie D. Fos.~.~fi. MD, Departnmrzt of
`Oncology. Norwegian Radirlm Hmpllcl, Ullc:~nc}:aa:s.see:a .70. 0310
`Ol.?0. No:-way.' crrmil 5“.clfas.t-c@kli1m1eriuio.no,
`Q: 200.’ by .4llm-:r:‘r::.rn Sac'i°r:!‘y of Clinical Oncology.
`0?32-llS’3X/(J!/J90!-62
`
`will progress in 60% to 80% of the responding patients
`during the first 3 years after the start of treatment.” in a
`recent meta-analysis,
`the addition of an antiandrogen to
`initial androgen suppression (total. androgen blockade
`[TAl3]) was shown to have a limited effect, if any at all.‘
`If the malignancy progresses despite androgen ablation
`after castration,
`three biologically different subgroups of
`lronnoneqcsistant prostate cancer (I-IRPC) can be identified.
`1. HRPC with residual androgen .9cnsz'1z‘w‘zy. The cancer
`cells are still sensitive to residual circulating sndrogens
`produced mainly in the adrenal glands, and the malignancy
`may respond beneficially if the effect of these remaining
`androgcns is removed. This can be acltieved by medical
`suppression of adrenal corticosteroid production or, if not
`used previously, by the application of antiandrogens which
`block the androgen receptors of the cancer cells.
`2. Iulormor.-c—sen.n'lir=e HRPC.‘ The disease is no longer
`androgemsensitivc, but
`it may still be influenced by
`horrnones such as rncdroxy~progcstcronc acetate or high-
`dose cstrogcns.
`3. Andrcgem and ltorm0r1e«refi*acmry HRPC: '1"he disease
`has become completely hormone-insensitive. Chemotherapy
`or investigational treatment modalities may be considered.
`
`62
`
`Journal of Clinical Oncology, Vol l9, No l (January l), 2001: pp 62-7i
`
`Downloaded from jco.ascopubs.org on February 14, 2014. For personal use only. No other uses without permission.
`Copyright © 2001 American Society of Clinical Oncology. All rights reserved.
`
`WCK1017
`Page 1
`
`WCK1017
`Page 1
`
`

`
`FLUTAMlDE v PREDNlSONE FOR ADVANCED PROSTATE CANCER
`
`63
`
`In the individual patient with HRPC, most often it mixture
`of these three cell populations is present
`in unknown
`proportions. Patients with prostate cancer who progress
`after primary androgen deprivation present with increasing
`somatic and psychologic distress,
`including pain due to
`bone metastases, anemia, and fatigue, Ten percent to 20% of
`the patients develop micturition problems caused by a
`growing primary tumor. Radiotherapy and analgesics can
`relieve local symptoms. but effective systemic therapies are
`needed to slow down or reverse the progressive develop-
`ment of the malignancy.
`Several end points can be considered during the treatment
`of HRPC: overall survival,
`time to progression (TTP),
`objective response, physlciamasscssed subjective response,
`and quality of life (QL).
`The evaluation of objective response according to the
`World Health Organization (WHO) criteria5 is problematic
`because only 10% to 20% of the patients have easily
`measurable metastatic lesions. Objective response based on
`assessment of bone scans is also difficult due to frequent
`interobserver and interexamination variations Relief of
`
`metastatic bone pain and improvement of the patient’s
`general condition are the most
`important parameters of
`subjective response in patients with HRPC and should be
`recorded routinely. During recent years, patient-based men»
`itoring of QT. has been introduced into clinical oncology,
`and appropriate questionnaires have been. developed? Bio-
`chemical response based on measurements of serum pros-
`tate—specific antigen (PSA) should be monitored as a sepa«
`rate entity. A recent consensus meeting has published
`guidelines for the evaluation of PSA-based response, thereby
`enabling more unifonn reporting of observed changes.8 In
`addition to the above methodologic problems, trials of chemo»
`therapy to treat HRPC have been hampered by relatively
`frequent
`toxicity problems in the elderly prostate cancer
`patients who often present. with major comorbidity. In this
`situation,
`it seems reasonable to influence the disease with
`hormones as long as possible, because hormonal manipulation
`is easily applied and has limited toxicity.
`Surgical and medical adrenalectomy, the latter by hydro-
`cortisone or prednisone, has been used in the treatment of
`HRPC for many years°'"’ to suppress the adrenal production
`of androstenedione and dehydroepiandrostcrone. Up to the
`early 1990s, however, only a few well-designed phase II or
`III studies were published that evaluated medical adrenal-
`ectomy in HRPC.
`Flutarnide exhibits antiandrogenic effects by binding to
`the cellular androgen receptors and thus reducing the cell’s
`androgen uptake. This drug has been used extensively in
`previously untreated patients, as a part of TAB or as
`monotlierapy.l‘3‘” In limited series, fiutamide has also been
`
`evaluated in the treatment of HRPC, with subjective re-
`sponse rates of 15% to 30%.”’”
`In 1990.
`the European Organization for Research and
`Treatment of Cancer (EORTC) Genitourinary Group initi-
`ated a phase III study to compare the effectiveness of
`prednisone and flutamide as secondary hormone manipula-
`tion in patients with metastatic HRPC. At that time, the
`expectation was that flutamide would be more effective than
`prednisone because of its specific activity in the cancer cell.
`The present report represents the final analysis of this study.
`
`PATIENTS AND METHODS
`
`Patients with histologically confirmed prostate cancer were eligible
`for the trial if they fulfilled the following criteria: (1) presence of
`symptomatic metastatic disease that had progressed after medical
`castration with lutcinizing horrnoneareleasing hormone (LHRH) ana-
`iogs (not estrogens) or bilateral crchiectomy. The pretrial serum
`testosterone level had to be within the range of the institution’s
`castration levels. In the present study, symptomatic disease implied
`cancer-induced deterioration of the patients general condition and/or
`painful, progressive metastatic disease with or without the use of
`analgesics, with or without complete pain relief; (2) WHO performance
`status of 0 to 3; (3) no previous use of prednisonc, fiutamicle, or any
`other oral aritiandrogen, but patients were eligible if they had received
`an antiandrc-gen transiently (for a maximum of 4 weeks) during their
`Ll-IRH treatment in order to prevent a Hate reaction; (4) no previous
`systemic anticancer treatment, except
`the above primary hormonal
`manipulation; and (5) certainty of clinical disease progression after
`prior surgery or previous radiotherapy. The patients were not allowed
`to receive radiotherapy at the time of trial entry.
`Patients with a second primary tumor (except basal cell skin cancer).
`serious cardiovascular problems, or insulirvdcpendcnt diabetes mellitus
`were ineligible for the trial, as were those who were unable to comply
`with regular follow-up.
`The trial was approved by the institutions’ local ethical committee,
`and patients provided written informed consent before randomization.
`The trial was open for patient entry from January l992 to March 1998.
`In October W95, an independent data-monitoring committee approved
`continuation ofthe trial without modification. At the time ofthe present
`analysis, the median follow-up vtas 330 days.
`
`Trial Design
`
`Patients were randomized to receive either fiutamide 250 mg orally
`three times a day (the F group) or predniscnc 5 mg orally four times a
`day {the P group). Patients receiving LHRI-I analogs continued with this
`treatment.
`
`All patients were examined for acute toxicity 3 weeks after trial
`entry. Response was evaluated at 6-week intervals from the start of
`treatment. Patients had to remain in the trial for at least 6 weeks to be
`assessable for response. They were otherwise included in the analysis
`as “non-assessable." Patients who progressed during the first 6 weeks
`were included in the progression category. Patients remained on the
`trial until subjective progression or unacceptable toxicity was recorded
`or until
`they wished to discontinue participation for any reason.
`Therapeutic interventions in patients who had gone off protocol
`treatment were chosen by the individual clinical
`investigator. All
`patients were followed until death.
`
`Downloaded from jco.ascopubs.org on February 14, 2014, For personal use only. No other uses without permission.
`Copyright © 2001 American Society of Clinical Oncology. All rights reserved.
`
`WCK1017
`Page 2
`
`WCK1017
`Page 2
`
`

`
`64
`
`Foss/St ET AL
`
`trial entry and at each follow-up visit, patients underwent a
`AI:
`clinical examination, including assessment of pain using a five«point
`scale (level 0, analgesics not required; level i, nonnarcotic analgesics
`occasionally required;
`level 2, nonnarcotic analgesics regularly re-
`quired; level 3, oral or parenteral narcotic analgesics occasionally
`required;
`level 4, oral or parenteral narcotic analgesics regularly
`required). Types and doses of the prescribed analgesics were recorded.
`A chest xvray and radioisotope bone scan were mandatory; other
`radiologic examinations were optional. “Sttperscan" was defined as 2
`75% metastatic involvement of the central skeleton.
`Blood samples were taken for analysis of hemoglobin, WBC count,
`and thrombocytes, together with the detennination of PSA, alkaline
`phosphatase, creatinine, liver enzyme, and testosterone levels.
`Clinical cxaininations and blood tests were repeated at each fol-
`low—up visit, Patients’ perfonnance status, weight, and degree of
`vomiting and diarrhea were recorded using the WHO criteria For
`toxicity? The perfonnance ofother tests was left to the discretion ofthc
`clinical investigator.
`
`Quality oflflife
`
`QL, as assessed by the patient, was a secondary end point of the
`study, but
`there was no a priori stated hypothesis. Thus,
`the QL
`evaluation was exploratory, with global QL representing the primary
`variable.
`
`At trial entry and at each follow-up visit, patients were asked to
`complete the EORTC Quality of Life Questionnaire (QLQ) C-30
`(version l.0).7 The QLQ C~30 is a 30-item questionnaire that was
`developed to assess a range of physical, emotional, and social health
`issues relevant
`to a broad spectrum of cancer patients. It has been
`shown to be reliable and valid in a wide range of patient populations
`and treatment settings and is currently being used in a large number of
`oncology clinical trials. The questions are organized into a number of
`multidtcm scales and single-item symptom measures (five functioning
`scales (physical, role, cognitive, emotional, and social], three symptom
`scales (fatigue, nausea and vomiting, and pain], and six single items
`[assessing dyspnea, sleep disturbance, appetite loss, constipation,
`diarrhea, and financial impactl). The last two questions ask patients to
`rate their overall health and QL. The QLQ C-30 was supplemented by
`three questions pertaining to analgesic use (Did you take any medica-
`tion for pain‘? If so, how much did it help‘? Have you had pain despite
`the use of analgesics?) All scales and single-item measures were
`linearly transformed to a 0 to 100 scale. ‘4 For the functioning scales
`and the global QL scale, a higher score represents a higher level of
`functioning.'QL; for the symptom measures, a higher score corresponds
`to a greater degree of symptoms.
`
`Response Criteria
`
`Objective response was not assessed. On the basis of the physician‘s
`evaluation,
`three categories for subjective response were defined:
`rcsponse, no change, and progression. No minimum duration of
`response was required.
`Response. At least one of the following three conditions had to be
`fulfilled: (1) reduction ofthe pain score (WHO criteria) by at least one
`level, with no deterioration of performance status; (2) unchanged pain
`level and reduction ofthe prescribed daily dose of analgesics by at least
`25% as compared with the pretreatment situation, with no deterioration
`of performance status; and (3) improvement of the WHO performance
`status by at least one level without either an increase ofthe daily dose
`of analgesics by .>. 25% or an increase in the pain level.
`
`“No change“ was defined as an unchanged pain score,
`No change.
`with less than a 25% reduction in the prescribed daily analgesic dose as
`compared with the pretreatment situation, and unchanged performance
`status.
`
`Progression was evaluated relative to the best condi«
`Progre.rsz'on.
`tion, observed at start of treatment or obtained during treatment.
`Progression was determined to have occurred if patients met at least
`one of the following conditions: increase of the pain score by at least
`one level, increase ofthe daily analgesic dose by at least 25%, any need
`to give additional pain treatment, such as radiotherapy, and WHO
`performance status deterioration by at least. one level.
`Duration of subjective response was calculated from trial entry to the
`date of progression. Biochemical response was defined as a decrease of
`the serum PSA level by 2 50% as compared with the baseline
`value."‘” However, no duration was required for biochemical response.
`
`Statistics
`
`The main end points for this trial were Ti‘? and duration of survival.
`Since virtually all patients entered onto the trial were expected to
`progress and die during follow—up, either of these end points could be
`chosen for calculating the sample size. A total of H92 patients followed
`until death were required in order to detect a. difference of 50% in the
`median duration of survival between the two treatment arms (from 9
`months with prednisone to 13.5 months with fiutamide), using a
`two~sided log-rank test (alpha = 0.5, beta = 0.20). Two hundred
`patients were sufficient to detect a. difference of 20% in the response
`rate in the two arms (alpha = 005, beta = 0.20).
`Given an anticipated median survival time of 8 to 10 months (based
`on the published literature on l-IRPC) and the number of available
`observations at each subsequent assessment point, the QL analysis was
`restricted to the 6~month period following entry onto the study. Means
`and confidence intervals were calculated for the QL scores of both
`treatment groups at each assessment point, yielding a series of’
`descriptive profiles that could be displayed in graphic form. In order to
`adjust for multiple comparisons over time, 99% confidence intervals
`were calculated to maintain an overall 95% confidence interval for each
`QL outcome. A linear mixed model analysis of variance was used that
`accounts for serial correlations between observations, as well as for
`intermittent missing forms. The main effects of treatment and time
`were tested on a reduced model (without an interaction term) whenever
`the interaction effect was found not to be statistically significant.
`The intention—to-treat principle was followed in all statistical analy-
`sis (ie, including ineligible and nonassessable patients in the analysis
`and considering patients in the treatment group they were allocated to
`by randomization).
`
`RESULTS
`
`PaIiem.‘s
`
`A total. of 201 patients were randomized to receive
`prednisone (101 patients) or flutarnide (100 patients, Table
`l). Presumed prognostic factors and cotnorbldities were
`equally distributed between the two treatment groups (P >
`.05). Almost all patients used analgesics, with approxi-
`mately 25% regularly using narcotics for pain level 4. The
`median number of hot spots on bone scans was 12 in both
`groups, and approximately 25% of the patients displayed
`supcrscans. The initial PSA level was elevated to more than
`
`14, 2014. For personal use only. No other uses without permission.
`Downloaded from jco.asccpubs.org on February
`Copyright © 2001 American Society of Clinical Oncology. All rights reserved.
`
`WCK1017
`Page 3
`
`WCK1017
`Page 3
`
`

`
`FLUTAMIDE v PREDNISONE FOR ADVANCED PROSTATE CANCER
`
`65
`
`Table 1. Patient Demographics and Clinical Data
`Preclnisone
`Flularnide
`Group
`Group
`
`No. of patients
`ineligible patients, n
`Median age, years
`Median time since castration, months
`lnitial pain score
`0
`l
`2
`3
`4
`WHO performance status
`0
`i
`2
`3
`Median no. oi hot spots
`lnllial PSA level, median, ng/mL
`S 2.5 X N
`2.5l~l0 XN
`ll-25 X N
`26400 X N
`7" ll}O X N
`Missing
`Abbreviation: N, normal.
`
`W1
`2
`72
`T9
`
`6
`29
`35
`5
`26
`
`l4
`59
`24
`4
`l2
`50
`6
`7
`l7
`29
`29
`lti
`
`W0
`2
`72
`21
`
`9
`22
`42
`ti
`23
`
`l3
`57
`23
`7
`l2
`37
`l l
`l3
`ll}
`33
`24
`9
`
`Tolol
`
`20l
`4
`72
`l‘?
`
`l5
`5l
`77
`9
`49
`
`27
`l l6
`47
`ll
`l2
`43
`l7
`2'0
`27
`él
`53
`23
`
`100 times the upper limit of the normal range in 29% ofthc
`P—group patients and 24% of the F-group patients.
`Two patients in each group were finally deemed to be
`ineligible (prior treatment with prednisone, 1; no metastases
`at
`trial entry,
`I; palliative radiotherapy at
`the time of
`randomization, 1;
`testosterone level above the castration
`range, 1).
`
`Treatment Compliance and Side Eflécts
`
`No attempt was made to regularly monitor a patient’s
`daily intake of prednisone or flutamide, except for the
`recording of dose reductions, prescribed at
`the 6-week
`follow-up visits. Treatment-related modifications of drug
`doses (without drug discontinuation) were recorded in two
`patients from the P group (reduction to 10 mg daily) and in
`eight patients from the F group. Serious nonhematologic
`side effects (grade 3 or 4) contributing significantly to trial
`treatment discontinuation were recorded in seven patients
`
`from the P group and nine patients from the F group (Table
`2). In the F group, gastrointestinal side effects (diarrhea,
`nausea) were most frequent. Two patients had a myocardial
`infarction while receiving flutamide treatment. Two patients
`in the P group refused trial participation immediately after
`randomization and were given fiutamide. In three patients
`from the P group, cardiovascular side effects (edema, fluid
`
`Table 2. Reasons for Going Off Study and Type of Clinical Toxicity
`Prednlsone
`Flulamicle
`Group
`Group
`ln = l0ll
`ln 3 l00l
`
`Total
`(N ‘-'— 2lJll
`
`Reason
`
`Progression only
`Toxicity only
`Progression + toxicity
`Other
`lneilgible
`Unknown
`Death (due to inlercurrem‘ disease)
`
`86
`3
`A
`2‘
`2
`l
`3
`
`86
`7
`2
`l l
`2
`D
`2
`
`172
`l O
`6
`3
`4
`l
`5
`
`NOTE. ln the prednisone group, the clinical toxicities were gastrointestinal
`(dyspepsia, peptic ulcer) ln -- 2), cardiovascular fliulcl retention, deep venous
`thrombosis) (n = 3), Cusl1ing'5 syndrome ln = ll, and unspecified (n : ll. ln
`the llulamlde group, they were gastrointestinal (nausea, vomiting, and dior-
`rheol (R = 7) and cardiovascular (myocordioc inlarclionl (n m 2).
`‘Patients relused trial participation immediately alter randomization.
`’rBiochr.-micol PD, protocol violation.
`
`retention, venous thrombosis) were observed, and in two
`patients, increasing dyspepsia or peptic ulcer caused treat-
`ment discontinuation.
`
`No grade 3 or grade 4 lcukopenia or thrombocytopenia
`was observed, except in one patient in the F group. This
`patient developed grade 4 thrombocytopenia, which most
`probably was related to disease progression. There was no
`difference between the two groups in liver function during
`the triai period.
`
`Response
`
`Subjective response was not assessable in 10 patients. In
`56% of the available patients from the P group and 45% of
`the F group, a subjective response was recorded at least at
`one foilow-up visit (P = .18; Table 3). In l6% of the
`patients in the P group and 21% of the F group, disease had
`progressed at the time of their first response evaluation. If a
`minimum response duration of6 weeks had been required in
`the original protocol,
`the subjective response rate would
`have been 49% for the P group and 34% for the F group.
`Subjective responses lasted for 4.8 months (median) in the
`I’ group and for 4.2 months in the F group (P: 0.21).
`A biochemical response was observed in 21% of the
`patients in the P group and 23% of patients in the F group.
`Responding patients from the P group had significantly
`higher baseline i’SA values than those from the F group
`(Table 4). in nine of the patients in the P group and 10 in the
`F group, the 2 50% PSA reduction was confirmed after 4
`weeks. Both subjective and biochemical responses were
`available in 180 patients.
`When both groups were combined, there was a significant
`association between PSA response at any time and subjec-
`tive response at any time (Table 5) (P < .00l). However,
`
`14, 2014. For personal use only. No other uses without permission.
`Downloaded from lco.ascopubs.org on February
`Copyright © 2001 American Society of Clinical Oncology. All rights reserved.
`
`WCK1017
`Page 4
`
`WCK1017
`Page 4
`
`

`
`66
`
`Fossfi. ET AL
`
`Table 3. Efficacy Evaluation
`
`Table 5. Subiective and Biochemical Response (at any lime)
`
`Pammmer
`
`Respunsent
`No chongel‘
`Pmgressionl’
`Eorly death, malignant disease
`Early cleolh, other cause
`Not assessable
`ineligible
`
`Pradnison
`Group
`{H : ‘ml
`so
`21
`lo
`1
`1
`A
`2
`
`Flulomicle
`Group
`(H 2 mo}
`45
`2‘?
`21
`0
`l
`2
`2
`
`48*
`3 4*
`105*
`
`Duration oi response, monlhs
`Pmgres5.m_FrEe WW0‘ monthsgi
`0%,“ survival, mm,,.5'
`‘P = .18 lor response lg’ lest).
`l5ublsclive response.
`i‘Meclion.
`§1TP lrom lriol enlry in all patients with subiective response.
`
`4.2
`2 3
`H2
`
`I
`Talc
`(N 1 20”
`W]
`50
`37
`1
`2
`6
`:3
`
`11.6
`2 9
`199
`
`33% of the patients had a subjective response without a
`biochemical response, and 5% of the patients had 3 50%
`decrease from baseline PSA without a subjective response.
`Most often, subjective responses preceded biochemical
`responses or were recorded simultaneously with biochemi-
`cal responses. In only 1.8% of the available cases was a
`subjective response observed after a biochemical response
`had been reported.
`
`Q12
`
`At baseline, QLQ C-30 data were available for 90% of
`patients in both treatment arms. Patients without baseline
`QL data were excluded from further serial analyses. During
`the subsequent 24~week period, compliance with QL assess~
`merit decreased to approximately 70%, equally in both arms
`
`Table 4. Baseline of Serum PSA Levels and Number of Biachemicully
`Responding Pcslienls (PEA 2. 50% reduction]
`Biochemical Response
`Flulumicle Group
`Preclnlsone Group
`ln=lOUl
`{rim ion
`
`Total
`(N=20l)
`
`No. oi responding patients
`Baseline PSA, gig/L
`Median
`
`21
`
`70?
`
`23
`
`775
`
`44
`
`K35
`
`Range
`PSA level
`5 2.5 X N
`151-10 X N
`ll-25 X N
`25-100 X N
`l00XN
`Missing
`Tolol
`
`02-2350
`
`02-2039
`
`02-2350
`
`-4
`2
`2
`
`"'
`
`l3
`
`21 l2l%l
`
`6
`3
`3
`6
`4
`1‘
`23 93%}
`
`10
`5
`5
`6
`l7
`1
`44122701
`
`‘Missing baseline PSA, but PSA within normal range during iollow-up.
`
`No, 01 Polienls With Bioclwrnicol Response
`No 0; Pufiems way“
`"*"“""“’""’*"“""'”’"”"’""""“
`‘.
`.
`No
`Yes
`Tolol
`Sublecilve Response
`7‘
`1°
`3‘
`N”
`'55
`34
`99
`Yes
`136
`44
`180
`Tami
`NOTE. P = .00? lor lhe correlation.
`
`(Table 6). Progression of the disease and/or death was the
`-
`-
`.
`main reason for noncompliance in both arms.
`tended to
`Overall, scores on the functioning scales
`decline just before dropout, whereas scores on the symptom
`scales tended to increase, indicating a nonrandom pattern of
`missing data. This held for both treatment arms of the trial.
`The pattern of QL scores over time for the P group and
`F group are presented graphically in Figures 1 and 2.
`Statistically significant treatment effects favoring the P
`group were noted for pain (P < .005), nausea and
`vomiting (P < .001), and diarrhea (P < .001). Patients
`receiving prednisonc reported significantly less constipa-
`tion (data not shown), probably as a result ofhaving used
`less narcotic analgesics. In addition, as compared with
`the F group,
`the P group reported significantly less
`fatigue at 6 weeks (P < .05) and significantly better role
`functioning (P < .01) and less appetite loss (P < .01) at
`3 months. Because there was a trend for a significant
`group difference in overail Q1. at baseline, change scores
`were used to evaluate group differences in overall QL
`over time. The results indicated a statistically signi-
`ficant treatment effect over time that favored the P group
`(P < .01). This difference in overall QL was observed
`primarily from week 3 to week 12.
`
`Survival
`
`At the time of the analysis, four patients from the 1’ group
`and five patients from the F group were still alive. There
`was no differerice in TTP or overail survival between the
`
`two groups, with the respective median duration times of
`3.4 and 10.6 months in the P group and 2.3 and 11.2 months
`in the F group (Table 3 and Fig 3).
`When both groups were combined, a biochemical re-
`sponse was associated with a hazards ratio of 0.42 for the
`risk of death (95% confidence interval, 0.29 to 0.60; P :-
`.00O1) ie, the risk of death for patients with a biochemical
`response was 0.42 that of patients without. The comparable
`figure For the association between subjective response and
`risk of death was 0.55 (95% confidence interval, 0.4] to
`0.75; P = .0001).
`
`14, 2014. For personal use only. No other uses without permission.
`Downloaded from jco.ascopubs.org on February
`Copyright © 2001 American Society of Clinical Oncology. All rights reserved.
`
`WCK1017
`Page 5
`
`WCK1017
`Page 5
`
`

`
`FLUTAMIDE v PREDNlSONE FOR ADVANCED PROSTATE CANCER
`
`67
`
`Table 6. Compliance With Q1. Assessment
`Baseline
`3 Wechs
`6 Weeks
`
`l2 Weeks
`
`18 Weeks
`
`24 Weeks
`
`Flutornide group
`No. of questionnaires received
`Percent of received questionnaires compared
`with expected at that time point
`Prednisone group
`No. of questionnaires received
`Percent of received questionnaires compared
`with expected at that time point
`
`8‘?
`8‘?
`
`92
`91
`
`60
`77
`
`57
`6‘?
`
`32
`73
`
`Al
`63
`
`21
`70
`
`29
`69
`
`19
`70
`
`25
`741
`
`DISCUSSION
`
`the biockadc of adrenal hormone
`In the late I980s,
`production was a valid option in the treatment of HRPC by
`the application of corticosteroids (hydrocortisonc or pred-
`nisone) with or without aminoglutethimiclcg"°"6'2” Mono-
`thcrapy with prednisone or hydrocortisone leads to the
`reduction of adrenal androgens, at least in a subgroup of
`patients? Responses are more likely to occur in patients
`with reduced amounts of adrenal androgens. The present
`study was planned with the knowledge of this adrenosup-
`pressivc effect of prednisone.
`In addition,
`low doses of
`corticosteroids have an unspecific, beneficial, anabolic, and
`fatigue-reducing effect?‘
`Due to its specific activity in the prostate cancer ceil,
`fiutamidc was, in 1990, expected to be more effective than
`prednisonc;
`investigators hoped that
`tlulamide treatment
`would result in higher response rates and improved survival.
`Recent research, however, has shed some doubt on whether
`the application of oral antiandrogcns is beneficial
`in all
`cases. Mutations of the androgen rcccptor(s) occur in the
`castrated patient.” As a result, a prostate cancer cell may
`become sensitive to a stimulating effect of the oral antiso-
`drogcns. The “antiandrogen withdrawal effcct”23 is ob-
`served in 20% to 30% of patients whose disease progresses
`during TAB.
`In HRPC, subjective relief is an important goal of
`treatment,“ and patients need to be assessed systematically.
`In this study, physician-evaluated subjective response was
`based on performance status and pain level (WHO criteria)
`combined with prescribed analgesic dosage. Progression
`also included the clinical parameter “need to start another
`major palliative treatment, as for example, radiotherapy.”
`The use of analgesics was, however, not systematically
`monitored on a daily basis, which would be prcferabie with
`today’s standard. Furthermore, today one would rety more
`on a patient’s self-assessment of pain for
`the overall
`evaluation of subjective response.
`In the present study,
`patient-based QL assessment, including the patient’s eval-
`uation of pain, was analyzed separately from physician-
`
`limitation of our
`based subjective response. A further
`response criteria is the lack of a defined minimum duration
`of pain relief, as used in other investigations. However, the
`post hoc inclusion of a minimum duration of subjective
`response of at least 6 weeks did not significantly alter the
`differences in response. rates between the two trial arms. The
`validity of our criteria of subjective response is also sup-
`ported by the statistical association between subjective and
`biochemical response.
`Biochemical response was defined retrospectively using
`the currently accepted cutoff point of 50% reduction from
`the baseline value.” Such PSA reduction has previously”
`and in the current soidy been shown to be associated with
`increased survival and (weakly) with subjective response.
`In general, however, one has to be reluctant to claim a
`life-prolonging treatment effect if responding patients sur-
`vive longer than rtonrcsponding patients. More research
`is needed to understand the biologic significance of
`PSA reduction.
`
`Our biochemical response rates of 22% and 23% for the
`P and F groups, respectively, are below those to be expected
`if hormone treatment is combined with chemotherapy (see
`Oh and Kantoff” for review) but are consistent with the
`observations of Kantoff et aim and Tannock at 211.2“ In
`general, hormonal manipulation alone in patients whose
`disease progresses despite androgen ablation leads to PSA
`reduction of 2 50% in 20% to 30% the patients (estramus—
`
`tine phosphate is regarded as a strong estrogen in this
`context). When hormones were combined with paclitaxel,
`doxorubicin, or vinblastine, biochemicai response rates of
`50% to 60% were observed in small, singloinstitution phase
`II studies of HRPC. In the two published phase III stud»
`ics,‘9’2° PSA reductions by 2 50% were seen in 33% and
`38% of patients. Whether these increased rates of PSA
`reduction are due to a true additive or synergistic effect
`between hormones and cytostatic agents is not known.
`Further research is warr