Treatment of Metastatic Prostntic Cancer With Low-Dose
`Prednisone: Evaination of Pain and Quality oi Life
`as Pragmatic lndices oi" Response
`
`By lcm Tannock, Mory Gospodo rowicz, William Meokln, Tony Ponzorello, Lesley Stewart, and Waiter Rider
`
`Thirty-seven men with symptomatic bone metastases
`from prostate cancer that had progressed following
`earlier trootment with estrogen: and/or orchidectemy
`were treated with low‘-dose prerinisone (7.5 to 10 mg
`daily). The rationale for this treatment was that some
`patients might still have hormone-sensitive disease
`that was stimulated by weak androgen: of adrenal
`origin, and that these nndrogens could be suppressed
`by prednisona through its negative feedback on
`secretion of udreno-cortlcotrophlc hormone (ACTH).
`Response to treatment was assessed by requirement
`for analgesics, by the McGill-Melzaclc pain question-
`naire, and by a series of
`i7 linear analog, self=
`assessment
`(LASA) scales relating to pain and to
`various aspects of quality of life. Fourteen patients
`(38%) had improvement in inclices used to assess pain
`at
`1 month after starting ptednisane, and seven
`patients (19%) maintained this improvement for 3 to
`30 montl-ts(meclic1n, 4 months). Reduction in pain was
`
`BOUT 75% of patients with symptomatic
`prostate cancer will report relief of symp-
`toms (primarily bone pain) following orchidecto-
`my, or after initial treatment with estrogcns, or
`luteinizing horrnonc—releasing hormone
`(LHRH) analogs. These measures reduce serum
`testosterone to castrate levels, thus removing the
`major source of‘ androgen stimulation. The dura-
`tion of response to primary ancirogcn ablation is
`variable, with median values of about
`i year.”
`When symptoms recur,
`this might be due to
`selection of proslatic cancer celis that are hot-
`moncwindependent, or to thc growth of cells that
`are stimulated by weak androgens of adrenal
`origin? Thus, secondary responses are sometimes
`observed with anti~androgcns such as fiutamide
`or cyprotcrone. acetate,‘ following adrcr1alecto«
`my,‘ or
`following adrcnai
`suppression with
`
`From the Princess: Margaret Hospital,’ the University of
`Toronto; and the Ontario Cancer Treatment and Research
`Foundation, Toronto.
`Submitted July 5}, 1988; accepted December 22, 1988.
`Address reprint reque.rr.r to Ian Ta.-mock. MD, Princess
`Margaret Hospital, and University of Toronto. 500 .S'lter—
`bourne St, Toronto, Ontario, Canada M41 IKE).
`G9 I989 by American Society of Clinical Oncology.
`0732-183X/39/0705-001433.00/0
`
`associated with improvement in other dimensions of
`quality of life, and in the stain for overall wall-being.
`Prednisone treatment led to a decrease in the concen-
`tration of serum testosterone in seven of nine patients
`where it was not initially suppressed below 2 nmol/L,
`and caused a decrease in serum levels of undrostene-=
`dime and del-cydroepiondresterone sulfate in more
`than 50% of patients. Symptomatic response was
`associated with a decrease in serum concentration of
`adrenal cmdrogans. We conclude that (I) low-«close
`prednisone may cause usetul relief at pain in some
`patients with udvnncecl prestatic cancer; (2) relief of
`pain was associated with suppression of adrenal
`undrogans; and (3) measures at pain and quality of
`life can be used to assess possible benefits of systemic
`therapy in patients with metastatic prostate cancer.
`J Clin Dncol }‘':590-597. to I989 by.4men'can Socierycn‘
`Clin icaf Urtcology.
`
`aminoglutcthirnide and hydrocortisonc.‘ An al»
`ternative approach is to proscribe low~dosc corti-
`costeroids” with the aim of producing negative
`feedback on the pituitary gland to inhibit
`secretion of adrcnocorticotrophic hormone
`{ACTH}. This might in turn lead to inhibition of
`the synthesis of the androgens androstenedione
`and clehydrocpiandrosterone (DHEA) and its
`sulfate (DHEAS), which is thought to be stimu-
`lated by ACTH. These relatively weak andro-
`gcns can undergo metabolism to produce small
`amounts of testosterone.3"° Adrenal sources of
`and rogcns may account for up to 20% of activity
`associated with the normal testis.)
`Assessment of response to systemic therapy of
`patients with prostatic cancer has been both
`cliliicult and unreiiable."'” Most patients do not
`have measurable soft
`tissue metastases, and
`serum markers such as acid and alkaline phos»
`phatase are not consistent
`indiccs of disease
`activity. Many patients may report dramatic
`improvement in symptoms after initial hormone
`therapy without improvement in x—rays or bone
`scan. When the aim of treatment is palliation,
`cficctivcncss should be assessed optimally by
`reproducible indicos of symptom control.
`In the current study we first reviewed the
`
`590
`
`Journal of Clinical Oncology, Vol 7, No 5 (May), 1939: pp 59059?
`
`Downloaded from jco.asc:opubs.org on February
`12, 2014. For personal use only No other uses without permission.
`rights reserved;
`Copyright © ‘I989 American Society of Clinical Oncology. Al
`
`WCK1028
`Page 1
`
`WCK1028
`Page 1
`
`

`
`
`
`
`
`QUAUTY OF LIFE AND PROSTATE CANCER
`
`591
`
`charts of all patients with symptomatic meta~
`static prostate cancer who had received predni—
`sone treatment between 1976 and 1980.
`
`Although this retrospective review revealed that
`some patients had major improvement in pain,
`there was no objective assessment of pain control
`(other
`than requirement
`for analgesics), and
`serum levels of androgens were not measured
`routinely. We therefore designed a prospective
`study with the following aims:
`(1) to develop
`methods for assessment of pain and other symp»
`toms in patients with metastatic prostate cancer;
`(2) to document the probability of subjective
`relief of symptoms in patients receiving predni-
`sone following progression after primary an-
`drogen ablation; and (3) to determine whether
`response to prednisone was correlate with a
`decrease in the serum levels of androgens.
`
`METHODS
`
`Patients
`
`The retrospective chart review revealed that 28 patients
`who had undergone prior orchidectomy. or who had pro-
`gressed on cstrogens. were started on prednisonc as treatment
`for symptomatic bone metastases. The charts of
`these
`patients were reviewed initially for evidence of response.
`Thirty»seven patients were then entered into the prospec-
`tive study. Ali patients had biopsy~proven prostate cancer
`and progressive symptomatic bone metastases despite
`estrogen treatment or previous orchideetomy. Patients were
`ineligible if they had received systemic therapy other than
`estrogens, or if they had life—tbrcatening complications such
`as cord compression or hypcrcalcemia. They had to under—
`stand English (in order to complete pain and quality~ef-life
`questionnaires). Patients with a history of diabetes or peptic
`ulcer disease, and those who had received corticosteroids
`since diagnosis of prostate cancer were excluded.
`
`Therapy
`Most patients in the retrospective series, and in the early
`part of the prospective study, received prednisone in a dose of
`5 mg in the morning and 2.5 mg in the evening; subsequently.
`patients received 5 mg twice daily. Those patients who were
`receiving estrogen therapy continued this medication at the
`same dose, usually diethylstilbestrol (DES) I mg three times
`daily.
`
`Assessment of Patients
`In the retrospective series patients were classified as
`responding to prednisone if all of the following conditions
`were satisfied for a minimum period oi‘ 2 months: (1) the
`clinical record gave clear indication of improvement in pain;
`(2) there was reduced intake of analgesics; and (3) no
`additional therapy was required (other than continuation of
`estrogens).
`All patients entered in the prospective study were assessed
`
`initially with a complete history and physical examination.
`bone scan and radiographs of painful areas. A complete blood
`count (CBC), and serum levels of acid and alkaline phospha~
`tase, calcium,
`testosterone. androstenedione, and DHEAS
`were to be measured monthly. Hormone levels were mea-
`sured by radioimmunoassay.
`Three methods were used to assess pain prior to initiation
`of prednisone therapy and at monthly intervals thereafter.
`First, the patients were asked about their average daily intake
`of analgesics during the week before their clinic visit. This
`information was used to generate an analgesic score, rcpre—
`scnting the total daily analgesic intake. Doses of morphine (5
`ms). codeine (30 mg), hydrornorphonc (2 mg), nnilcridine
`(25 mg). or mixed formulations containing codeine (30 mg)
`or oxycodone [l5 mg) were assigned two points. Nonnarcotie
`analgesics such as aspirin (325 mg) or acetaminophen (325
`mg) were assigned 1 point. The above doses may not accu-
`rately reprcsent analgesic potency, but this does not introduce
`error in the present study since patients did not change
`medication. but merely increased or decreased the dosage.
`Secondly, patients were asked to complete the McGill-
`Mclzack pain qucstionnairem‘ under the direction of the
`study nurse. This questionnaire presents 20 groups of verbal
`descriptors relating to sensory, affective, evaluative, and
`miscellaneous aspects of pain (Fig la). Within each group
`the patient was asked to select one word (if any) that best
`described the pain that he had been experiencing during the
`prior week. These words were ranked in each group, and the
`rank values of the words selected were summed to provide the
`“pain rating index.“ A simpler assessment of pain “the
`present pain intensity“ was selected from a 6-point verbal
`scale (0 - no pain,
`i
`-- mild, 2 - discornforting, 3 -«
`distressing, 4 -« horrible, 5 as excruciating).
`The third method for assessing pain was part of a series of
`17 linear analog sell’-assessment (LASA) scales""“ (Fig lb).
`For each scale the patient was presented with a l0—cm line
`that is anchored at its left end by the worst possible scenario
`(eg, extremely severe pain) and at its right end by the best
`possible scenario (eg, no pain at all). The patient was then
`asked to place a vertical mark on the line that represented his
`state (in relation to the end descriptors) during the preceding
`24 hours. The current series of LASAs were adapted from
`scales that were developed and validated for use in patients
`with breast cancer.“ They included scales related to general
`health and to symptoms of disease. We also included a global
`scale relating to overall well—being, which was anchored by
`the descriptors “extremely ill" and “I feel well.” Each scale
`was measured in centimeters (to the nearest 0.5 cm) from its
`left end. so that higher scores represent less symptoms.
`
`RESU LT5
`
`Characteristics ofPan'enrs
`
`Characteristics of patients who were started
`on prednisone therapy are summarized in Table
`1. Most of the patients had received extensive
`prior treatment, including a median of two hor-
`monal treatments (orchidectomy and/or a vari-
`ety of estrogens). Many patients had also
`
`Downloaded from jco.ascopubs.org on February 12, 2014. Forpersonal use only._No other uses without permission.
`Copyright © 1989 American Society of Cli|"llCEl Oncology. All rights reserved.
`
`WCK1028
`Page 2
`
`WCK1028
`Page 2
`
`

`
`592
`
`TANNOCK ET AL
`
`a. segment of ma Mcflitl-Malzack Paln Quutlonnolrn
`
`
`more
`1 5
`Innoytnu
` culling
`trouhluornu
`
`
`
`
`locoraun
`mlurnhtu
`cinching
`lnlinu
`5
`
`
`prqnlnn
`uunnlm
`
`
`gnurlng
`lpvuxtlng
`
`
`enmpln
`rndlulen
`
`
`ponstultng ‘
`I F in
`
`
`
`
`11
`
`
`
`b. Eurnptu ot Linear Analog sot! Asuurntnt Sculls
`Plum piece a mark on the limb below to
`indicate your status durlng tho put 24 hours.
`
`Fatigue
`
`extremely
`tired
`
`
`
`not mad
`at
`Ill
`
`B.
`
`Sonia: Lil-. Meeting and Destiny with People Outnidu the Fun“?
`
`utrlrnoly
`unsatis-
`tnctory
`
`
`
`eomplatlly
`utlstactory
`
`Fig 1.
`Illustration of (o) the McGi|l-Melted: pain
`quostionnoire" and (la) LASA scales that were used to
`_
`.
`_
`mu.-ss pom and quality of lrfa.
`
`received one or more courses of radiation to bony
`metastases.
`
`Symptomatic Response to Prednisone
`
`Tolerance of Prednisone Therapy
`
`One patient discontinued prednisone because
`he claimed that it caused nausea, and a second
`
`patient experienced gastrointestinal pain. There
`was no observable toxicity in the remaining
`patients. In particular there was no gastrointesti-
`nal bleeding, and patients did not become
`Cushingoid with the low doses that were used.
`
`Of the 28 patients who were reviewed retro»
`spectively, seven had improvement in pain with a
`reduced requirement for analgesics for a median
`duration of 5 months (range, 2 to 11 months).
`Pretreatment vaiues of the four indices used to
`
`assess pain in patients treated prospectively were
`as follows (mean 3; SEM): daily analgesic score.
`l0.2 1 2.2; linear analog scale, 4.3 2 0.5; pain
`rating index, 21 : 2; present pain intensity, 2.0 :
`
`Table 1. Characteristics of 37 Patients Treated Prospuctively
`62 (46-76) yr
`
`Median age {range}
`Previous therapy to primary
`Radical prosrotactomy
`Radiation to primary
`Median no. of prior endocrine moneuw.-rs* (range)
`Median no. of irradiated rnolostotic sites (range)
`Median interval from diagnosis (rouge)
`Median interval from initial endocrine therapy‘ (ronge)
`Subjective response to initial endocrine therapy
`‘(:5
`No
`Not ossessoblej’
`initial serum concentration (range) oh’:
`Acid phosphatase
`Aikoline phosphatase
`‘includes orchidaclomy and dillerenl aslrogans.
`tsome patients had endocrine therapy when they were osymplomcxtlc.
`iNorrnul voiues lot serum concentration of acid and clikoline phosphorus: are < 0.8 and 120 lU/L respectively.
`
`0
`72
`2 (141)
`1 (0-43
`27 (6-11?) mo
`18 (4-1 17) mo
`
`25
`A
`3
`
`2.? (0.3-89.6) iU/L
`3 T 3 (/54 ,D00) 1U/L
`
`12, 2014. For personal use only. No other uses without permission.
`Downloaded from jco.a5copub5.org on February
`Copyright © ‘3989 American Society of Clinical Oncology. Ali rights reserved.
`
`WCK1028
`Page 3
`
`WCK1028
`Page 3
`
`

`
`QUALITY OF LIFE AND PROSTATE CANCER
`
`593
`
`0.2. Thus, most of the patients were symptomatic
`from pain. Of 37 patients who received predni-
`sonc in the prospective study, 14 (38%) had
`improvement in each of the three scales that were
`used to assess pain, and a decreased or stable
`requirement for analgesics for a minimum of 1
`month. Five patients became free of pain and
`required no analgesics.
`In seven patients
`improvement in each of the pain indices and
`decreased analgesic requirement persisted for 3
`to 30 months (median, 4 months). Of 14 patients
`with initial relief of pain on prednisone, ten had
`responded subjectively to their initial hormonal
`treatment with estrogens and/or orchidectomy,
`one had not responded, and three could not be
`assessed for response.
`The overall effect of prednisone on pain expe-
`rienced by the entire patient population was
`assessed by comparing pain indices after one
`month of treatment with predniscne, with those
`recorded at the start of treatment. Statistical
`evaluation using the twc~tailed paired r-test
`showed significant improvement in the median
`values of pain rating index and present pain
`intensity (P = .009 and .011, respectively) and
`nonsignificant
`trends toward improvements in
`analgesic score or LASA evaluation of pain
`(P = .24 and .12, respectively). The alternative
`statistical method of using the nonparametric
`sign test to compare pm» and posttreatment pain
`indices gave similar conclusions.
`LASA scores for 13 of the 14 patients who had
`
`improvement in all indices of pain are summa-
`rized in Table 2. Initial scores were not available
`for one patient, but his subsequent
`testing
`showed normal or near normal scores (2 9) for
`all dimensions of quality of life. At the time of‘
`maximum improvement
`in pain, 46% of the
`scales had improved. 11% had deteriorated, and
`43% were unchanged. Most of the unchanged
`scores reflected a normal state (before and after
`treatment) for some of the quality—of-life dimen-
`sions that were assessed. Large improvements
`were seen in the LASA scale, which reflected
`overall well-being by asking patients “How do
`you feel?” (Table 2), and all but two patients
`with improved pain indices had improvement in
`this scale. Overall, the panel of indices that were
`used to assess pain and quality of life provide
`consistent evidence for at least transient relief of
`
`symptoms in 12 of the 37 patients (32%) who
`received prednisone prospectively.
`
`Biochemical Response to Prednfsone
`
`Values of alkaline and acid phosphatase were
`elevated initially in 31 and 22 of the 3? patients,
`respectively. Decreases of > 10 IU/mi. in alka-
`line phosphatase were seen in l6 patients during
`prednisone therapy. but only eight of these had
`consistent improvement in pain. Nine patients
`had a decrease of > 1 IU/mL in acid phospha-
`tase, and only four of these patients had improve-
`ment in pain. Thus there is little evidence for
`
`Table 2. Changes in 16 LASA Scale: Reflecting Different Dimensions of Quality of Life For Patient:
`was Had Improvement in lndices of Pain
`
`Pcrliont
`
`improved
`
`Delariorcrlad
`
`..._........¢.:M--o~om'~.ID-uu.rswr-)-—-
`
`Total
`
`_..._...
`
`OLA!-l&L7IO-*O'O‘OC-"‘-‘
`
`91
`
`-<.ooo-::sru----uoo
`
`21
`
`Unrhunged
`5
`3
`3
`5
`5
`5
`5
`5
`4
`10
`‘I2
`9
`13
`
`34
`
`Change in Score for Ovumli
`Wall-Being
`+3.5
`+4.0
`+1.0
`+5.5
`+3.5
`+2.0
`+4.0
`+2.0
`-- l .5
`+6.5
`+10
`-— 3.0
`+4.0
`
`"For dimensions, see Table 4. Some dimensions lag, employment) were not relevant to some individuals and were not scared. initial [ASA
`scores were not available for one patienl who had complete resolution of pain at ‘l month.
`
`Downloaded from l'co.ascopubs,org on February 12. 2014. For personal use only. No other uses without permission.
`Copyright (:3 1989 American Society of Clinical Oncology. All rights reserved.
`
`WCK1028
`Page 4
`
`WCK1028
`Page 4
`
`

`
`594
`
`TANNOCK ET AL
`
`correlation between relief of symptoms and these
`biochemical markers.
`
`Only five patients with pain relief had reas»
`sessment with skeletal x—rays and/or bone scan.
`Two showed improvement (one after initial dete-
`rioration), one was unchanged, and two had
`deterioration. All three of the patients without
`symptomatic improvement who had reassess-
`ment showed deterioration in bone scan and/or
`skeletal x—rays.
`
`Changes in Hormone Levels
`Initial serum concentration of the androgens
`testosterone, androstenedione, and DHEAS, and
`their changes in response to prcdnisone are sun1~
`marized in Table 3. There were no diiferences in
`the initial
`levels of these hormones between
`
`patients who did or did not have subsequent
`improvement in symptoms on prednisone.
`Nine patients had serum concentration of tes-
`tosterone > 2 nmol / L which has been regarded
`as an upper limit for castrate men; eight of these
`men had previous orchidectomy and one patient
`was receiving DES. Seven of these patients had a
`decrease in serum testosterone concentration
`
`after initiation of prednisone. Three of these
`patients, and one whose testosterone level
`remained unchanged, experienced relief of pain;
`all four responding patients also experienced a
`decrease in serum concentration of androstene-
`(lions and DHEAS, whereas those who had ele~
`vatecl testosterone initially and did not respond
`
`symptomatically experienced no decrease in
`adrenal androgens.
`Symptomatic relief on prednisone was asso-
`ciated with suppression of serum levels of andro-
`stenedione and /or DHEAS (Table 3). Serial
`assessment of one or both of these hormones was
`
`13 patients who experienced
`available for
`improvement in pain; ten of these had a decrease
`in one or both hormones, and in two others they
`were initially present in low concentration. This
`contrasts with a decrease in one or both hormone
`
`levels in only six of 16 patients without pain relief
`who had serial assessment of hormone reduction
`(P = .06, for hormone reduction; P.- .006 for
`hormone reduction or initial suppression, Fish-
`er‘s exact test, two—sided).
`
`Evaluation of Methods
`for Assessing Symptoms
`
`Consistency of the four methods that were
`used to evaluate pain was assessed by calculating
`correlation coefficients between the two McGill—
`Melzaclt. scales,
`the analgesic score, and the
`LASA scale for the entire data set. These corre-
`lation coefiicients were in the range of 0.44 to
`0.68 indicating significant correlation between
`them (P < .01 for all comparisons, Spearman‘s
`rank correlation coeilicient).
`The 16 LASA scores representing various
`dimensions of quality of life were also compared
`with each of the four measures of pain. There
`
`Table 3, Initial Serum Concentration of Hormones, and Change: After Receiving Pradnisono
`Palianis Wnitom
`Patients Mlh
`Entire Series
`improvement of Pain (N =- 23}
`(N " 37)‘
`Improvement of Pain (N -« 14}
`
`Tes1ns'lan‘Jna1'
`Median (range)
`Decrease (by > 1 nmol/L)
`increase or no change
`Initial value < 2 nmol/1
`Androstenedionef
`Median (range)
`Decrease {by 2 1 nmcrll L}
`Increase or no change
`initial value < 1 nmol/L
`DHEAST
`Median (range)
`Decrease (by 2 lnmol/L)
`Increase or no change
`initial value < l Jlrnol/L
`
`1.2 (0-5.6)
`7
`2
`27
`
`3.9 (O.ls8.0)
`34
`12
`3
`
`2.3 (0.341?)
`13
`8
`4
`
`1.3 (04.5)
`3
`l
`to
`
`3.9 to.r«s.o)
`3
`3
`2
`
`2.3 ro.s:s.2)
`8
`1
`3
`
`1.1 ((15.6)
`4
`1
`17
`
`3.7 (03.7.7)
`s
`9
`1
`
`2.2 (o.3s.7)
`5
`7
`1
`
`‘Hormone levels were available or fallow-up visits for 36 patients (testosterone), 29 patients (androstenedlone), and 25 patients (DHEAS).
`fblormul values: testosterone, 12 in 3] nrnolf L (castrate, < 2 nrnol/L); undroslenedione, 2.010 3.7 nmol/L; DHEAS, 2.7 to l().‘? umol/L.
`
`Downloaded fromlco.ascopubs.org on February 12, 2014. For personal use only. No other uses without permission.
`Copyright © 1989 American Society of Clinical Oncology. All rights reserved.
`
`WCK1028
`Page 5
`
`WCK1028
`Page 5
`
`

`
`QUAUTY OF LlFE AND PROSTATE CANCER
`
`595
`
`was a significant correlation for 29 of these 64
`comparisons (P < .05. Spearman’s rank correla-
`tion coefficient, corrected for multiple compari-
`sons by the Bonferroni method). The correlation
`coefficient averaged over the four pain scores is
`shown in Table 4. Note that the correlation is
`
`particularly strong for the overall scale indicat-
`ing sense of well-being.
`Few patients reported abnormalities in the
`categories of self-care, speaking or writing, or
`mental function, so that scores were clustered
`close to 10, ie, at the end of the scale representing
`normal status (Table 4). Therefore, these scales
`give little useful information, and the clustering
`of scores leads to a weaker association with
`
`indices of pain control. A high proportion of
`patients were retired, but the LASA scale re1at~
`ing to employment was completed by 46% of
`patients; this scale also showed a poorer correla-
`tion with indices of pain.
`in an attempt to reduce the number of LASA
`scales without losing information, we examined
`correlations between scaies where similar factors
`
`might influence the ability to function. Spear-
`man rank correlation coefficients (r,) > 0.45
`were recorded for all associations within the
`
`groups: (1) mobility, physicial activity, recre-
`ation, social life, and housework, (2) fatigue and
`sleeping (rs == 0.53), and (3) mood and anxiety
`(r, = 0.61). Thus, in future trials at single LASA
`
`scale might be used to represent each of these
`groupings.
`
`DISCUSSION
`
`Response to Prednisone
`We have demonstrated that about 30% of
`patients may have improvement
`in symptoms
`after treatment with low-close prednisone, when
`they are no ionger
`responding to primary
`androgen ablation with orchidectomy and/or
`estrogens. Although this response is usually tran-
`sient the treatment is nontoxic and inexpensive,
`and a few patients have prolonged relief of pain.
`As with other systemic therapies for prostate
`cancer, there was little evidence that prednisone
`caused a consistent improvement in serum levels
`of acid and alkaline phosphatase, or in x-rays or
`bone scans. These indices have sometimes been
`referred to as objective, in contrast to subjective
`improvement in symptoms. We believe that the
`words “object:ive" and “subjective” are used here
`inappropriately. The aim of treatment is pallia-
`tion, which can be defined as improvement
`in
`either the duration or quality of survival. There is
`no convincing evidence that systemic therapy
`improves the survival of patients with metastatic
`prostate ca ncer," and measures of symptom con-
`trol should therefore be used to assess palliation.
`Whereas radiologic or biochemical assessment of
`
`Table 4. LASA Scales Used to Evaluate Difioron! Features of Quality of Life‘
`
`Absolute Average
`(lot-relation
`Coolficiom‘
`No.
`Mean
`Median
`With 4 Puln
`
`Dlmerulnn
`Completed
`Score
`Score
`Range
`lndicat
`
`8
`7.4
`1323
`Mobility
`8
`6.1
`56
`Employment
`l 0
`9.4
`123
`Self-care
`3.5
`8.2
`123
`Physical activity
`7
`5.8
`l l 6
`Recreation
`9
`7.5
`WI!
`Social life
`H}
`8.5
`l23
`Farniiy relationships
`10
`7.0
`92
`Housework
`ND
`9.8
`123
`Speaking/writing
`5
`5.7
`123
`Fatigue
`9
`7.6
`123
`Appetite
`8.5
`7.45
`T23
`Sleep
`10
`9.7
`123
`Mental ‘function
`7.5
`7.0
`123
`Mood
`7
`6.5
`123
`Anxiety
`6
`6.l
`123
`Sense of well-being
`‘A score of 0 corresponds to the worst possible state and a score of it) indicates ca normal state.
`
`l—‘l0
`(MO
`2-10
`1-10
`0-10
`0-10
`0- 'l 0
`0-10
`6—l O
`0-l D
`[H D
`l-i0
`6-10
`0-10
`0-10
`(HO
`
`0.36
`OJ 9
`0.2l
`0.36
`0.31
`0.28
`0.30
`0.31
`0.20
`0.29
`0.28
`0.26
`0.15
`0.18
`0.24
`0.52
`
`Downloaded from jco.ascopubs.org on February 12, 2014. For personal use only._No other uses without permission.
`Copyright © 1989 American Society of Clinical Oncology. All rights reserved.
`
`WCK1028
`Page 6
`
`WCK1028
`Page 6
`
`

`
`
`
`
`
`596
`
`TANNOCK ET AL
`
`patients is known to be poorly correlated with
`clinical well-being of the patient, our study has
`shown that various methods of assessing pain and
`quality of life can be consistent. Thus assessment
`of symptoms is not only the most relevant end
`point, but also may be the most objective method
`for assessing response to systemic therapy in
`prostatic cancer.
`Comparison of nonrandomized series of
`patients is fraught with difiiculty because of
`variability in selection factors (eg, extent of prior
`treatment) and in criteria for judging response.
`However,
`the probability of improvement
`in
`symptoms after secondary treatment with pred-
`nisone appears to be within the same range as
`when patients progressing after orchidectomy or
`estrogens are treated with anti—androgens (cy-
`proterone acetate or fiutamide) or with amino-
`glutethimide and hydrocortisone.
`Indeed,
`the
`observations of Plowman et all‘ and Dowsett et a1°
`that adrenal androgens were suppressed less with
`arninoglutcthimide plus hydrocortisone than
`with hydrocortisone alone, suggests that hydro-
`cortisone rather than aminoglutethimide may be
`the active drug when this combination produces
`subjective response. Moreover, corticosteroids
`such as dexamethasone are used frequently as
`antiernetics
`in patients receiving cytotoxic
`chemotherapy, and may contribute to any
`improvement in symptoms that is observed in
`such patients.
`In the absence of randomized
`trials demonstrating superiority of other drugs, a
`trial of low~dose prednisone can be recommended
`for most patients with progressive symptomatic
`prostate cancer who have undergone orchidccto-
`my, or who are receiving estrogens or LHRH
`agonists.
`
`Mechanism 0fAr:l.'i0r1 0fPredm‘.rcme
`
`The antiinflamrnatory properties of predni«
`some might lead nonspecifically to improvement
`in pain and other symptoms. However,
`this
`mechanism seems unlikely to explain the magni-
`tude of responses seen in the present study, since
`the low dose of prednisone used was comparable
`to the physiological production of corticosteroids
`by the adrenal gland. It seems probable that most
`of the benefit was mediated through hormonal
`effects on the adrenal gland.
`‘I-I-labeled androstenedione or DHEAS has
`been shown to be incorporated into dihydrotes-
`
`tosterone, the major intranuclear androgen in
`prostatic cells.” Although incorporation of
`androstenedionc and DHEAS was much less
`
`than testosterone, these weak androgens proba-
`bly can stimulate prostatic tissue. Prednisone
`treatment suppressed the serum concentration of
`androstenedione and DHEAS in more than 50%
`
`of patients with initial levels above 1 nmol / L and
`l prnol/L, respectively, as previously reported
`for smaller series of patients.” All but one of the
`patients who responded symptomatically either
`had suppression of adrenal androgens, or had
`initial low levels of these hormones, suggesting
`that this may have been a mechanism leading to
`response.
`Two mechanisms may be expected to lead to
`progression of prostatic cancer in patients who
`have previously undergone orchidectomy or are
`taking estrogens: (1) cells that are no longer
`responsive to hormones may have been selected
`from a heterogeneous initial population, or (2)
`some of the cells may remain androgen-respon»
`sive and are stimulated by adrenal androgens or
`their derivatives. For the latter patients, therapy
`with prednisone (or other measures to suppress
`or counter adrenal androgens) may be elfective
`palliation. Unfortunately, all patients uitimately
`progress to a state of hormone nonresponsive—
`ness.
`
`Evaluation ofSymproms in Prostatic Cancer
`
`We have used the present study to evaluate
`methods for assessment of pain and quality of
`life. The methods used for assessment of pain
`were well correlated, and the use of multiple
`indices decreases
`the probability of
`falsely
`assessing response. Three of these indices, the
`LASA scale, the analgesic score, and the simpler
`6-point verbal scale of the McGil1~Melzaclc
`method (present pain intensity) could be applied
`rapidly and were easily understood by all
`patients. The more complex pain rating index of
`the McGill-Melzack questionnaire required con»
`siderable linguistic ability and was particularly
`difficult for patients whose first language was not
`English.
`Several studies have demonstrated that qual-
`ity of life is multifactorial, and our study used 17
`LASA scales (including that for pain) to assess
`various dimensions of health. These scales were
`
`adapted from a series that had been used to
`
`Downloaded from jco.ascopubs.org on February 12. 2014. For personal use only. No other uses without permission.
`Copyr‘zght© 1989 American Society of Clinical Oncology. At! rights reserved.
`
`WCK1028
`Page 7
`
`WCK1028
`Page 7
`
`

`
`QUAUTY OF LIFE AND PROSTATE CANCER
`
`597
`
`Table 5. Simplified Instrument for Assessing Symptoms
`in Patients With Metastatic Prostate Cancer
`Nidlfiliic SCDN‘
`EASA Scale: for:
`
`Present pain intensity scale of
`McGill»Melzock method, is,
`0 - no po'm
`l - mild
`2 as discornforting
`3 - distressing
`4 w horribie
`5 ~ excruciating
`
`Pain
`Physical activity
`Fatigue
`Appgfifa
`Bowel tuncilon
`Family relationships
`Mood
`Overall well~being
`
`assess quality of life in patients with metastatic
`breast cancer where they were shown to be
`reproducible and valid.“ Some of the scales
`assessed related functions, and scores for these
`dimensions were highly correlated; others tested
`function that was only rarely abnormal, and gave
`little useful information; an additional scale to
`assess bowel function is suggested by the fre-
`quent side effect of constipation in patients who
`are taking narcotic analgesics. Based on this
`experience we suggest that adequate information
`could be obtained by using a smaller series of
`
`eight LASA scales that would assess pain, physi-
`cal activity, fatigue, appetite, bowel function,
`family relationships, mood, and well—being. The
`combination of the analgesic score, the present
`pain intensity, and eight LASA scales (Table 5)
`could provide a simple and useful instrument for
`assessing response to all types of systemic ther»
`apy in metastatic prostate cancer, but will
`require testing for validity and reproducibility in
`a larger series of patients.
`In conclusion, we have documented that some
`patients with symptomatic prostatic cancer may
`respond to prednisone as second—line hormonal
`treatment;
`that response is usually associated
`with suppression of adrenal androgens; and that
`assessment of pain and other features of quality
`of life is a feasible, pragmatic method for assess-
`ing benefit from systemic treatment.
`
`ACKNOWLEDGMENT
`
`We thank our coiieagucs for referring patients for this
`study, and Drs A. Matldn, and E. Mobbs for their construc-
`tive comments.
`
`REFERENCES
`
`1. Resnick Ml, Grayhack .11‘: Treatment of stage IV
`carcinoma of the prostate. Urol Clin North Am Zzl-ill-l6l,
`1975
`2. The Leuprolide Study Group; Leuprolidc versus dieth~
`ylstilbestrol for metastatic prostate cancer. N Engl J Med
`31l:128l-1286,1984
`3. Geller J, Albert JD: Adrenal androgen blockade in
`relapsed prostate cancer. Eur J Cancer Ctin Oncol 21.-1127»
`i131, 1985
`4. Stellar B, Albert DJ: SCH 1.3521 in the treatment of
`advanced carcinoma of the prostate. J Urol 3303-807. 1974
`5. Bhanalaph T, Varkarakis MJ. Murphy GP: Current
`status of bilateral sdrenalectorny for advanced prostatic
`carcinoma. Ann Surg 179:17-23, 1974
`6. Murray R, Pitt P: Treatment of advanced prostatic
`cancer, resistant to conventional therapy. with aminoglute-
`thirnide. Eur J Cancer Clin Oncol 2l:453—45S. 1985
`in
`7. Miller GM, Hinman Jr F: Cortisone treatment
`advanced carcinoma of the prostate. J Urol 72:485-496.
`1954
`8. Plowman PN, Perry LA, Chard T: And