Efficacy and Safety of Nonsteroidal Antiinflammatory
Drugs for Cancer Pain: A Meta-Analysis
By Elan Eisenberg, Catherine S. Berkey, Daniel B. Carr, Frederick Mosteller, and Thomas C. Chalmers
Purpose: To assess the efficacy and safety of nonste-
roidalantiinflammatory drugs (NSAIDs) in the treatment
of cancer pain by meta-analyses of the published ran-
domized control trials (RCTs).
Patients and Methods: Twenty-five studies met inclu-
sion criteria for analysis. Of these, 13 tested a single-
dose effect, nine multiple-dose effects, and three both
single- and multiple-dose effects of 16 different NSAIDs
in a total of 1,545 patients. Baseline pain intensity (when
provided) of moderate or higher was indicated in 81%
of patients.
Results: Single-dose NSAID studies found greater an-
algesic ecacy than placebo, with rough equivalence to
5 to 10 mg of intramuscular morphine. Pain scores dif-
fered insignificantly for aspirin versus three other
NSAIDs. Analgesic responses to low- and high-dose
NSAIDs suggested a dose-response relationship, but this
was not statistically significant. Recommended and su-
pramaximal single doses of three NSAIDs produced com-
parable changes in pain scores, which indicates a ceiling
N ONSTEROIDAL antiinflammatory drugs (NSAIDs)
are universally accepted as part of the treatment of
cancer-related pain. According to World Health Organi-
zation (WHO) guidelines on cancer pain relief,1' 2 NSAIDs
(or other nonopioid analgesics) are recommended as the
sole treatment of mild to moderate pain, or combined
with weak opioids for moderate to severe pain. When
needed, strong opioids can be added to NSAIDs for more
severe pain. (Although WHO terminology has recently
evolved away from the terms weak opioid and strong
opioid in favor of opioid for mild to moderate pain and
opioid for moderate to severe pain, in this report we use
the earlier terms because they are concise and still in
daily use by nearly all clinicians.) This WHO-ladder ap-
From the Departmentof Neurology, Massachusetts GeneralHospital
and HarvardMedical School; Technology Assessment Group, Harvard
School of Public Health; Department of Anesthesia and Division of
Clinical Care Research, Department of Medicine, New England Medi-
cal Center and Tufts University School of Medicine, Boston, MA; and
Pain Relief Clinic, Rambam Medical Center, Haifa, Israel.
Submitted March 4, 1994; accepted July 25, 1994.
Supportedby the Office of the Forumfor Quality and Effectiveness
in Health Care, Agency for Health Care Policy and Research, De-
partment of Health and Human Services, Bethesda, MD.
The views expressed herein do not necessarily represent the posi-
tion of the Department of Health and Human Services.
Address reprint requests to Daniel B. Carr, MD, Department of
Anesthesia, New EnglandMedicalCenter,Box 298, 750 Washington
St, Boston, MA 02111.
© 1994 by American Society of Clinical Oncology.
0732-183X/94/1212-0034$3.00/0
2756 Journalof Clinical Oncology, Vol 12, No 12 (December), 1994: pp 2756-2765
Downloaded from ascopubs.org by DFG Deutsche Forschungsgemeinschaft on July 22, 2019 from 141.002.140.067
Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
analgesic effect. Common side effects included upper
gastrointestinal symptoms, dizziness, and drowsiness.
The incidence of side effects showed a trend to increase
with dose, without a ceiling effect, and to increase with
multiple doses. Single or multiple doses of weak opioids
(WO) alone or in combination (WO/C) with nonopioid
analgesics did not produce greater analgesia than
NSAIDs alone. Single doses of WO/C analgesics pro-
duced more side effects than NSAIDs alone, although
both side effect incidence and patient dropout rates were
equal when multiple doses were administered.
Conclusion: These findings question whether the tra-
ditional World Health Organization (WHO) second anal-
gesic step (addition of a weak opioid when pain is inade-
quately treated by a nonopioid analgesic alone) is
warranted. A lack of comparable studies precluded test-
ing the hypothesis that NSAIDs are particularly effective
for malignant bone pain.
J Clin Oncol 12:2756-2765. © 1994 by American So-
ciety of Clinical Oncology.
proach has been adopted by clinicians worldwide and
is currently a fundamental principle of pharmacologic
management of cancer pain.3 9 Adherence to WHO treat-
ment recommendations will result in only a minority of
patients with terminal cancer experiencing severe pain,
and most will have minimal to moderate pain.10-14
NSAIDs are widely held to be especially effective for
management of malignant bone pain.5-7' 9 They are as-
sumed to produce dose-dependent analgesia and to have
a ceiling analgesic effect in patients with cancer pain.5 ' 9
However, because specific clinical data about analgesic
responses to NSAIDs in cancer patients are sparse,"' 15
many of the assumptions that guide current oncologic
teaching and practice are based on either clinical impres-
sions or trials in nonmalignant pain.
We examined the scientific evidence on efficacy and
safety of NSAIDs in the treatment of cancer-related pain,
by conducting a meta-analysis of data from published
randomized controlled trials (RCTs). Meta-analysis in-
volves the quantitative synthesis of data from multiple
clinical trials and has become a scientific discipline with
well-described features and methods.8"" It uses statistical
methods to integrate individual study results and can in-
crease the strength of findings from individual studies
that, uncombined, may be less significant because of
small sample size or ambiguous results.
PATIENTS AND METHODS
Literature Search and Inclusion Criteria
A literature search was conducted through BRS Colleague and
Docline (National Library on Medicine) for the years 1966 to 1992.
NSAIDs FOR CANCER PAIN 2757
Table 1. Study Characteristics
No. of No. of
First Author/Year Patients Centers Study Design Type of Cancer No. of Doses Type of NSAID

Carlson/1990' 75 M P NR S+ M Ketorolac
Epstein/l 98621 75 S P Oropharyngeal M Benzydamide
Estope/199022 40 S P NR S + M Ketorolac
Ferrer-Brechner/1 984"2 30 S C NR S Ibuprofen (as combination)
Kim/i198524 67 S P Oropharyngeal M Benzydamide
2
Levick/1988 " 100 M P NR M Naproxen sodium
26
Lomen/1986 26 S C Breast M Flubiprofen
Martino/ 197627 24 S C NR S Indoprofen; ASA
Martino/1 97828 18 S C Mixed S Indoprofen; naproxen
Minotti/l 98929 99 S P Mixed M Diclolenac
30
Moertel/1974 100 S C NR S ASA
Pellegrini/1983"3 12 S C Mixed S Indoprofen
Puglisi/198932 45 S P NR S Piprofen
Stambaugh/i198033 170 M P Mixed M Zomepirac
4
Stambaugh/1981" 40 S C NR S Zomepirac
Stambaugh/19823" 45 S C NR S Zomepirac
Stambaugh/1988a31 160 S P NR S Ketoprofen
7
Stambaugh/1988b3 30 S P NR M Ibuprofen (in combination)
Staquet/I98938 126 S P NR S Ketorolac
Sacchetti/ 98439 36 S C Mixed S Ketoprofen; ASA
Sunshine/19884o 123 S P NR S Ketoprofen
Tonachella/1 9854' 20 S C NR M Diclofenac
Turnbull/l 98642 28 S C Mixed S+ M Naproxen; ASA
Ventafridda/1 97443 24 S C Mixed S Indoprofen; ASA
Ventafridda/1990"' 72 S C NR M ASA; indomethacin; piprofen diclofenac;
ibuprofen; naproxen; suprofen; sulindac

Total (N = 25) 1545 S-21; P-12; Mixed, 7; S, 16;

M-4 C-13 NR, 15; M, 12
1 organ, 3

Abbreviations: C, crossover; M, multiple; P, parallel; S, single; NR, not reported; ASA, acetylsalicylic acid.

The search terms were antiinflammatory-agents-nonsteroidals, neo-
plasms, and pain. Full-length reports (ie, not simply abstracts) in the
English language were selected, provided they were double-blind,
randomized, controlled studies related to NSAID use in cancer pain
in humans. References listed in these reports were reviewed and
considered for inclusion according to these criteria. Lack of a precise
description of the randomization process was not considered an ex-
clusion criterion, be.cause it was unavailable in the majority of the
reports.
Data regarding n umber of patients, type of cancer, bone metasta-
ses, type of NSAID used, number of doses (single v multiple), length
of observation, ana Igesic efficacy, side effects, dropout rates, and

Table 2. Analgesic Efficacy of NSAIDs Versus Placebo (single dose)

Effect
No. of No. of No. of No. of NSAIDs Placebo
Scale Studies Comparisons Drugs Patients (%) (%) RD 95% CI

PPID 8 14 6 401/395 48 29 .19 .12-.25*

SPID 6 12 5 289/281 31 15 .16 .08-.22*
PPAR 5 11 5 277/269 60 36 .24 .16-.32*
TOPAR 6 12 5 377/369 43 26 .17 .11-.24*

NOTE. In this and all following tables, the number of patients shown to the left of the slash received the first treatment named in the table heading.
Drugs used: PPID--ketorolac, indoprofen, piprofen, ketoprofen, ASA, zomepirac; SPID-ketorolac, indoprofen, ketoprofen, ASA, zomepirac; PPAR-
ketorolac, indoprofen, ketoprofen, ASA, zomepirac; TOPAR-ketorolac, indoprofen, ketoprofen, ASA, zomepirac. Reports used: PPID-Moertel (1974),30
Carlson (1990),2 Ventafridda (1974),'3 Staquet (1989)," Puglisi (1989),32 Pellegrini (1983),31 Stambaugh (1981,3, 1988a3); Spid-Carlson (1990),20
2
Ventafridda (1974),43 Staquet (1989),38 Pellegrini (1983),31 Stambaugh (1981,"3 1988a" ); PPAR-Carlson (1990),20 Ventafridda (1974),13 Staquet
(1989),38 Stambaugh (1981,24 1988a"6); TOPAR-Moertel (1974),30 Carlson (1990),20 Ventafridda (1974),'3 Staquet (1989),38 Stambaugh ( 19 8 1 ,'
1988a36).
Abbreviations: RD, pooled rate difference; CI, confidence interval.
*Significant at .05 level.

Downloaded from ascopubs.org by DFG Deutsche Forschungsgemeinschaft on July 22, 2019 from 141.002.140.067
Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
2758 EISENBERG ET AL

Table 3. Analgesic Efficacy of NSAIDs Versus Weak Opioid or WO/C (single dose)
Effect
No. of No. of No. of No. of NSAID WO/C
Scale Studies Comparisons Drugs Patients (%) (%) RD 95% CI
PPID 6 9 5/7 471/472 49 54 -. 04 -. 10-.01
TOPAR 5 8 4/7 458/463 44 45 -. 02 -. 07-.05
NOTE. Drugs used: PPID-ASA 650 mg v ASA 650 mg + oxycodone 9.76 mg, ASA 650 mg + pentazocine 25 mg, ASA 650 mg + propoxyphene
100 mg; ketorolac 10 mg v acetaminophen 600 mg + codeine 60 mg; ketorolac 10 mg v pentazocine 50 mg; piprofen 600 mg v pentazocine 67.5 mg;
zomepirac 100 mg v oxycodone 5 mg + ASA 224 mg + acetaminophen 160 mg + caffeine 32 mg; ketoprofen 100 or 300 mg vASA 650 mg + codeine
60 mg. TOPAR-ASA 650 mg v ASA 650 mg + oxycodone 9.76 mg, ASA 650 mg + pentazocine 25 mg, ASA 650 mg + propoxyphene 100 mg;
ketorolac 10 mg v acetaminophen 600 mg + codeine 60 mg; ketorolac 10 mg v pentazocine 50 mg; zomepirac 100 mg v oxycodone 5 mg + ASA 224
mg + acetaminophen 160 mg + caffeine 32 mg; ketoprofen 100 or 300 mg vASA 650 mg + codeine 60 mg. Reports used: PPID-Moertel (1974),30
Carlson (1990),20 Estape (1990),22 Puglisi (1989),32 Stambaugh (1981,3A 1988a36); TOPAR-Moertel (1974),30 Carlson (1990),20 Estape (1990),22
Stambaugh (1981,3" 1988036).

global evaluation of the drugs used were extracted from text, tables,
and figures. Analgesic efficacy data were extracted only from pain
intensity or pain relief scores, provided these were based on patient
self-report with the use of a visual analog or similar quantitative
scale. Patient overall satisfaction or global evaluation of a drug was
not used as a measure of analgesic efficacy. The most frequently
reported pain scores that were later used for the analysis were as
follows: peak pain intensity difference (PPID), the difference be-
tween baseline pain and the minimal pain intensity reported at any
time during the postdrug observation period; summed pain intensity
difference (SPID), the sum of differences between baseline pain and
pain intensity at each time point during the postdrug observation
period (the area above the curve in a pain intensity v time graph);
and peak pain relief (PPAR) and total pain relief (TOPAR), the
maximal and the summed pain relief, respectively, during the post-
drug observation period.
StatisticalAnalysis
Since different scales were used by different investigators to mea-
sure pain intensity or pain relief, the data were transformed into a
common scale, ie, a percentage scale. In the case of PPID and PPAR,
in %, where n is the number of measurements taken within the
observation period. For both formulas, higher values (in %) indicate
larger analgesic effects. Studies were included in a specific analysis
only when data for comparable observation periods could be ex-
tracted. Thus, single-dose studies were analyzed separately from
multiple-dose studies. Six-hour observation periods in the single-
dose, and 10- to 14-day periods in the multiple-dose studies, were
most commonly reported and were therefore chosen to be analyzed.
Many reports did not provide broad statistical information (eg,
only mean values, but no standard deviations). To avoid exclusion
of additional studies, only mean values were extracted and used in
the analysis.
The n-weighted mean effect, which was used throughout this
meta-analysis, allows us to calculate the mean effect of a treatment
from several studies that vary in their sample size and to weight
each study according to its relative sample size. Consequently, larger
studies are assigned greater weight than smaller ones. The n-
weighted mean was calculated separately for each treatment (using
PPID as an example) according to the following formula:
k
I N, x PPID,

the transformation into the percentage scale used the following for-
mula: (presented value [PPID or PPAR])/(maximal value - minimal
value [on scale]) x 100 = value in %. APPID of 3 on a scale of 0
to 5, for example, was transformed to a PPID of 60%. In the case
of TOPAR and SPID, the sum of pain intensity difference or pain
relief depends on both the length of the observation period and the
number of measurements taken within it, both of which often differed
from one study to another. Thus, transformation of the SPID and
TOPAR values into the percentage scale was performed according
to the following formula: (presented values [SPID or TOPAR])/
([maximal value - minimal value (on scale)] X n) x 100 = value
k •= n-weighted mean,
I Ni
where PPID, is the mean PPID from the ith study out of k studies,
and Ni is the number of patients in the ith study.
The DerSimonian and Laird Random Effects Model"9 was used
in the present study to estimate differences between treatments in
efficacy and side effects. The combined estimates are presented (in
tables) with confidence intervals, and approximate significance is
indicated at the .05 level. The DerSimonian and Laird Model was
used for the following analyses: single-dose efficacy of NSAID ver-

Table 4. Analgesic Efficacy of NSAIDs Versus Morphine (single dose)

Effect

No. of No. of No. of No. of NSAIDs Morphine

Scale Studies Comparisons Drugs Patients (%) (%) RD 95% CI

SPID 3 5 3 156/151 36 29 .07 -. 03-.17

TOPAR 2 4 2 144/139 52 45 .07 -. 04-.19
NOTE. Drugs used: SPID-indoprofen, ketoprofen, zomepirac v morphine 5-10 mg intramuscularly; TOPAR--ketoprofen, zomepirac v morphine 5 to
10 mg intramuscularly. Reports used: SPID-Pellegrini ( 1 9 8 3 ),3' Stambaugh (1982),3" Sunshine (1988);40 TOPAR-Stambaugh (1982),"s Sunshine
(1988).40

Downloaded from ascopubs.org by DFG Deutsche Forschungsgemeinschaft on July 22, 2019 from 141.002.140.067
Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
NSAIDs FOR CANCER PAIN 2759

Table 5. Analgesic Efficacy of Nonaspirin NSAIDs Versus Aspirin (single dose)

Effect
No. of No. of No. of No. of NSAIDs ASA
Scale Studies Comparisons Drugs Patients (%) (%) RD 95% CI

PPID 4 7 3 124/124 40 35 -. 05 -. 20-.04

SPID 4 7 3 124/124 35 28 -. 07 -. 18-.03

NOTE. Drugs used: PPID and SPID-indoprofen, ketoprofen, naproxen. Reports used: PPID and SPID-Ventairidda (1974),41 Sacchetti (1984),"3 Martino
(1976),27 Turnbull (1986).42

sus placebo, WO/C, and morphine; single-dose efficacy of low ver-
sus high or supramaximal NSAID doses; single-dose efficacy of
aspirin versus other NSAIDs; single-dose side effects of NSAIDs
versus placebo, weak opioid-NSAID combinations (WO/C), and
morphine; single-dose side effects of low versus high or supramaxi-
mal doses of NSAID; single-dose side effects of aspirin versus other
NSAID; multiple-dose side effects of NSAID versus WO/C; and
patient dropout rates during NSAID versus WO/C multiple-dose
studies.
RESULTS
Twenty-five studies met inclusion criteria 2"" (Table
1). In one study,"3 data regarding analgesic efficacy could
not be extracted because of mismatch between the pain
scales reported in figure legends and the actual scales
used in the figures. This study was used only for side
effect analysis. Three other studies that met inclusion
criteria did not provide combinable data and were not
21 24
included in the formal analysis. Two of them ' each
found benzydamide to be superior to placebo in the treat-
ment of radiation-induced mucositis pain, but their results
could not be integrated. The third23 demonstrated that the
combination of methadone plus ibuprofen provides more
analgesia than methadone plus placebo, but did not com-
pare ibuprofen alone with placebo and, therefore, was not
combined with the other trials.
Twenty-one of 25 studies were single-center and four
were multicenter. Twelve had parallel and 13 crossover
designs (Table 1). Thirteen studies tested a single-dose
effect, nine a multiple-dose effect, and three both single-
and multiple-dose effects of 16 different NSAIDs.
The studies provided data on 1,545 patients. Specific
cancer diagnosis was reported for 515 patients (33%).
Most studies enrolled patients with a variety of malignan-
cies but did not report their outcomes separately. Conse-
quently, no analysis of the efficacy of NSAID according
to type of cancer could be conducted.
Baseline pain intensity was rated moderate or higher
in 81% of patients in whom this measurement was re-
ported (69% of all patients). Only one study 2' indicated
mild baseline pain in all patients enrolled. That study
focused on radiation-induced mucositis, and mild pain
intensity was reported before initiation of radiation ther-
apy. Four other studies 24 40,'
42 43
, did not provide baseline
pain intensity measures.
Analgesic Efficacy
Although all 25 trials reported analgesic efficacy, only
the single-dose studies were combinable for analgesic
efficacy analysis. Better pain relief was obtained from
NSAIDs than placebo in three scores (SPID, PPAR, and
TOPAR) based on five or six studies, and in another score
(PPID) based on eight studies (Table 2). Single doses of
placebo produced a 15% to 36% rate of analgesia,
whereas NSAIDs resulted roughly twice as much analge-
sia (31% to 60%). All differences between NSAIDs and
placebo comparisons were statistically significant.
The peak (PPID) and the sum (TOPAR) effects of
single doses of NSAIDs were compared with WO/C (ie,
pentazocine, acetaminophen plus codeine, etc). The two
scores indicated approximately 50% efficacy of both the
NSAID and the WO/C, with insignificant differences be-
tween them (Table 3). Additionally, three NSAIDs (indo-
profen, ketoprofen, and zomepirac) were compared with
5 to 10 mg of intramuscular morphine. The NSAIDs had
slightly, insignificantly superior analgesic efficacy com-

Table 6. Analgesic Efficacy of Low Versus High Single Doses of NSAIDs

Effect
No. of No. of No. of No. of Low-Dose High-Dose
Scale Studies Comparisons Drugs Patients (%) (%) RD 95%CI

PPID 3 5 3 80/77 39 51 -. 12 -. 27-.02

SPID 3 5 3 80/77 32 40 -. 08 -. 23-.06

NOTE. Drugs used: PPID and SPID-ASA 600 v 1,000 mg, indoprofen 100 v 200 mg, ketorolac 10 v 30 mg. Reports used: PPID and SPID-Martino
(1976),27 Ventafridda (1974),43 Staquet (1989).38

Downloaded from ascopubs.org by DFG Deutsche Forschungsgemeinschaft on July 22, 2019 from 141.002.140.067
Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
2760 EISENBERG ET AL

Table 7. Analgesic Efficacy of Recommended Versus Supramaximal Single Doses of NSAIDs

Effect
No. of No. of No. of No. of Recommended Supramaximol
Scale Studies Comparisons Drugs Patients Dose (%) Dose (%) RD 95% Cl
PPID 3 4 2 125/124 50 48 .02 -. 09-.15
SPID 5 6 3 197/196 39 39 .00 -. 09-.09
PPAR 3 4 2 133/132 65 69 -. 02 -. 13-.09
TOPAR 4 5 3 173/172 48 46 .02 -. 08-.12
NOTE. Drugs used (recommended dose/scales used): ketoprofen intravenously (IV) 100 v 400 mg (IV 100-300 mg/PPID, SPID); ketoprofen orally (PO)
75 v 225 mg (PO 50-100 mg/SPID, PPAR, TOPAR); ketoprofen PO 100 v 300 mg (PO 50-100 mg/PPID, SPID, PPAR, TOPAR); ketorolac intramuscularly
(IM) 10 v 90 mg (IM 10-30 mg/PPID, SPID, PPAR, TOPAR); ketorolac intramuscularly (IM) 30 v 90 mg (IM 10-30 mg/PPID, SPID, PPAR, TOPAR); zomepirac
36
PO 100 v 200 mg (PO 100 mg/SPID, TOPAR). Reports used: PPID-Sacchetti (1984),"3 Staquet (1989),3 Stambaugh (1988a) ; SPID-Sacchetti (1984),39
36
Staquet (1989),38 Stambaugh (1982,35 1988a3), Sunshine (1988);Ao PPAR-Staquet (1989),38 Stambaugh (1988a), Sunshine (1988);40 TOPAR-
36
Staquet (1989),38 Stambaugh (1982,35 19880 ), Sunshine (1988).40

pared with morphine. This was measured by a SPID of
36% for the NSAIDs versus 29% for morphine and by a
TOPAR of 52% versus 45%, respectively (Table 4).
Four studies compared the single-dose analgesic effi-
cacy of aspirin with three other NSAIDs (indoprofen,
ketoprofen, and naproxen). Meta-analysis of these trials
showed a PPID of 35% for aspirin and 40% for the other
NSAIDs. Corresponding SPID values were 28% and
35%, respectively. This analysis was based only on 124
patients in each treatment group. It suggests slightly supe-
rior pain relief from NSAIDs compared with aspirin, but
without statistical significance (Table 5).
The dose-response relationship was evaluated by com-
paring the analgesic efficacy of 600 mg versus 1,000 mg
of oral aspirin, 100 versus 200 mg of oral indoprofen,
and 10 versus 20 mg of intramuscular ketorolac. Three
single-dose studies allowed analysis of 157 patients (80
received low doses and 77 high doses). The two scores
used (PPID and SPID) were approximately 1.3 greater in
the high-dose patients than in the low-dose patients, but
the differences did not attain statistical significance (Table 6).
The ceiling analgesic effect was tested by comparing
pain scores for recommended and supramaximal single
doses of three NSAIDs: ketoprofen 100 versus 300 mg
and 75 versus 225 mg orally, zomepirac 100 versus 200
mg orally, and ketorolac 10 and 30 mg versus 90 mg
intramuscularly. The recommended and supramaximal
doses showed similar efficacy according to all four scores
used (PPID, SPID, PPAR, and TOPAR), which indicates
a ceiling effect (Table 7).
The multiple-dose trials were not comparable to permit
evaluation of a possible dose-response relationship or
ceiling effect of multiple doses of NSAIDs. Only two
multiple-dose analgesic efficacy analyses was allowed,
and both showed no significant differences in either
PPAR or TOPAR between one NSAID and two WO/C
drugs (Table 8). However, these analyses were based on
a small number of patients (20 to 29 in each group) and
on ketorolac as the sole representative of the NSAID
group.
Pain was related to bone metastases in seven studies
(Table 1). Four studies enrolled patients with either ma-
lignant bone pain, non-bone malignant pain, or both, but
results were not reported separately for bone-related and
25 37 39
non-bone-related pain in any study. Three studies , ,
examined the analgesic efficacy of NSAIDs specifically
for malignant bone pain, but not for other types of malig-
nant pain. Analgesic efficacy data were extractable from
only two of these studies, but were not combinable be-
cause one was a single-dose trial 39 and the other a multi-
ple-dose trial. 2 5 The single-dose study reported a mean
NSAID-induced PPID of 40% to 55% and SPID of 34%
to 45%. The NSAID PPID in the multiple-dose study
was 23% to 33%. Although not formally compared, these
scores show similarity to those for the malignant non-
bone pain (Tables 2 through 7).

Table 8. Analgesic Efficacy of NSAIDs Versus Opioid or WO/C (multiple-dose)

Effect

No. of No. of No. of No. of NSAID WO/C

Scale Studies Comparisons Drugs Patients (%) (%) RD 95% CI

PPAR 2 2 1/2 29/20 62 55 .07 -. 20-.35

TOPAR 2 2 1/2 29/20 49 50 -. 01 -. 30-.27

NOTE. Drugs used: PPAR and TOPAR-ketorolac 10 mg v acetaminophen 600 mg + codeine 60 mg; ketorolac 10 mg v pentazocine 50 mg. Reports
used: PPAR and TOPAR-Carlson (1990),20 Estape (1990).22

Downloaded from ascopubs.org by DFG Deutsche Forschungsgemeinschaft on July 22, 2019 from 141.002.140.067
Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
NSAIDS FOR CANCER PAIN 2761
Table 9. Common Side Effects of NSAIDs (episodes per 100 patients)
Source No. of Studies No. of Patients Upper GI Dizziness Drowsiness Other* Total

Single-dose 10 612 34 (5) 10(2) 16(3) 28 (4) 88 (14)

Multiple-dose 7 291 44 (14) 27 (9) 12(4) 89 (31) 172 (59)
NOTE. Single-dose studies used: Moertel (1974),'o Martino (1976,27 197821), Puglisi (1989),32 Stambaugh (1981,"3 1982," 198817), Staquet (1989),38
Sunshine (1988),4 Ventafridda (1974).•3 Multiple-dose studies used: Carlson (1990),20 Estape (1990),22 Minotti (1989),29 Tonachella (1985),1 Turnbull
37
(1986),42 Stambaugh (1988b), Ventafridda (1990).44
Abbreviation: GI, gastrointestinal.
*Other side effects include constipation, dry mouth, insomnia, itches, agitation, and allergic reactions.

Side Effects throughout the entire observation period in six of the

Common side effects of NSAIDs include upper gastro-
intestinal upset, dizziness, and drowsiness (Table 9). Mul-
tiple doses generally produced more side effects than
single doses (Table 9). No patients included in the single-
dose studies received additional analgesics during the ob-
servation period. The only exception was a study in which
methadone plus ibuprofen was compared with methadone
plus placebo.23 This study was excluded from the analysis
of single-dose side effects, the results of which are listed
in Table 10. The incidence of side effects, according to
seven studies of a total of 412 patients treated with single
doses of NSAIDs, was 15 episodes per 100 patients,
which is the same rate as for patients treated with placebo.
The incidence of side effects was insignificantly higher in
patients given single doses of 5 to 10 mg of intramuscular
morphine than those given NSAIDs (15 v nine episodes
per 100 patients). WO/C produced a significantly higher
incidence of side effects than NSAIDs (20 v 14 episodes
per 100 patients). A meta-analysis based on a limited
number of patients shows a trend (P > .05) for the single-
dose side effect incidence to increase in a dose-related
fashion (Table 11). Supramaximal doses of NSAIDs tend
to produce more side effects compared with recommended
doses (19% v 11% incidence; P > .05) (Table 12).
Seven studies that included 291 patients allowed the
analysis of multiple-dose side effects (Table 9). All anal-
gesics except for the tested drug were discontinued
seven trials. In the seventh study,37 ibuprofen or placebo
was given in addition to acetaminophen/oxycodone. In
this study, the investigators classified side effects as re-
lated or definitely not related to the study drug. Therefore,
all seven studies were included in the multidose side
effect analysis. The overall side effect incidence during
repeated dosing was 59 episodes per 100 patients. Upper
gastrointestinal symptoms, dizziness, and drowsiness
were the most common side effects, although many others
have been reported (Table 9). Four studies permitted a
comparison of the incidence of side effects during treat-
ment with repeated doses of NSAIDs to WO/C (Table
13), and showed no differences (63% and 62%, respec-
tively).
Patient Satisfaction and Dropout Rates
Comparisons of patient satisfaction measured by global
evaluations of drugs by patients were limited to single-
doses of NSAIDs. Global evaluation of two NSAIDs was
superior to that of morphine (5 to 10 mg intramuscularly),
but not significantly (65% v 57%) (Table 14). No signifi-
cant differences between satisfaction with single recom-
mended and supramaximal doses of two NSAIDs were
demonstrated in another analysis (Table 15).
Analysis of patient dropout during 7 to 10 days of
administration of either NSAIDs or WO/C showed an
insignificant trend in favor of NSAIDs over the WO/C

Table 10. Side Effects: NSAIDs Versus Placebo, Morphine, Weak Opioid, or WO/C (single dose; episodes per 100 patients)
No. of Side Effect
No. of No. of
NSAID v Studies Comparisons Patients NSAIDs Others RD 95% CI

Placebo 7 12 412/422 65 (15)/71 (15) .02 -. 03-.06

Morphine 2 4 154/149 14 (9)/22 (15) .04 -. 01-.10
Weak opioid or WO/C 6 12 574/564 79 (14)/11 1 (20) .05 .01-.08*

NOTE. Drugs used: v placebo-NSAID: ASA, indoprofen, piprofen, zomepirac, ketoprofen, ketorolac; v morphine--NSAID: ketoprofen, zomepirac;
morphine: 5-10 mg IM; v weak opioids/WO/C drugs--NSAID: ASA, piprofen, zomepirac, ketoprofen; WO/C: pentazocine, ASA + codeine, ASA +
oxycodone, ASA + propoxyphene, ASA + pentazocine. Studies used: v plocebo-Moertel (1974)," Martino (1976),27 Puglisi (1989),32 Stambaugh
(1981,`4 1988a36), Staquet (1989),"8 Ventafridda (1974)43; v morphine--Stambaugh (1982),3" Sunshine (1988)4°; vWO/C-Moertel (1974),30 Puglisi
(1989),32 Stambaugh (1981,34 1988a"36), Staquet (1989),38 Ventafridda (1974).43
*Significant at .05 level.

Downloaded from ascopubs.org by DFG Deutsche Forschungsgemeinschaft on July 22, 2019 from 141.002.140.067
Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
2762
Table 11. NSAID Side Effects: Low Versus High Single Doses
(episodes per 100 patients)
No. of No. of No. of Side Effects
Studies Comparisons Patients Low High RD 95% CI
2 3 56/53 5 (9)/13 (24) -. 17 -. 39-.03
NOTE. Drugs used: ASA 600 v 1,000 mg, indoprofen 100 v 200 mg,
ketorolac 10 v 30 mg. Reports used: Ventafridda (1974),43 Staquet
(1989).38
drugs, with dropout rates of 27% and 30%, respectively
(Table 16).
DISCUSSION
Our meta-analyses indicate that NSAIDs have analge-
sic efficacy in patients with cancer pain. NSAIDs pro-
duced a peak pain relief incidence of 60% even though
baseline pain intensity was moderate to severe in the
majority of the patients. Compared with placebo, NSAIDs
produced significantly more analgesia as measured by
pain intensity difference and pain relief scales. It is im-
portant to emphasize that the differences between
NSAIDs and placebo are both statistically and clinically
significant, with NSAIDs being 1.5 to two times more
effective than placebo. One may criticize these results
since they are derived from single-dose studies. However,
WHO has recommended avoiding the use of placebo in
patients with cancer pain, so it might be regarded as
unethical to conduct multiple-dose placebo-controlled an-
algesic trials in these patients. The results of this meta-
analysis support the WHO position by providing evidence
that although placebo produces some analgesia, true anal-
gesics have a significantly higher analgesic efficacy in
patients with cancer pain.
The comparison of the analgesic efficacy of one
NSAID to another was somewhat limited. In only seven
trials was an attempt made to conduct such a comparison.
Four of the trials were combinable and allowed compari-
son of non-aspirin NSAIDs to aspirin. Although in the
original reports the NSAID tested was superior to aspirin
in at least one score (PPID or SPID) in each trial, reanaly-
sis of the results according to the DerSimonian and Laird
method failed to show statistically significant differences
Table 12. NSAIDs Side Effects: Recommended Versus Supramaximal Single Doses (episodes per 100 patients)
No. of Studies No. of Comparisons No. of Patients
4 5 190/189
NOTE. Drugs used (recommended doses): ketoprofen PO 100 v 300 mg (PO 50-100 mg); ketoprofen PO 75 v 225 mg (PO 90-100 mg); ketorolac IM
10 v 90 mg (IM 10-30 mg); ketorolac IM 30 v 90 mg (IM 10-30 mg); zomepirac PO 100 v 200 mg (PO 100 mg). Reports used: Stambough (1982,"5
1988a36), Sunshine (1988),40 Staquet (1989).38
Downloaded from ascopubs.org by DFG Deutsche Forschungsgemeinschaft on July 22, 2019 from 141.002.140.067
Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
EISENBERG ET AL
in favor of any of the NSAIDs versus aspirin or of all of
them as a group versus aspirin. This may be attributed to
the fact that many of the trials did not provide the standard
deviations required by our methods, so we had to use
rough approximations that could affect the sensitivity of
the analysis. In a different analysis, when compared with
intramuscular injections of approximately 5 mg of mor-
phine, NSAIDs were found to produce equivalent analge-
sia. This finding is consistent with previously published
data indicating that 650 mg of aspirin is equianalgesic to
5 to 10 mg of intramuscular morphine.
The dose-response analgesic efficacy of NSAID in can-
cer pain has not been determined yet and the present
meta-analysis results are still only suggestive. The 8% to
12% differences in pain scores with high- versus low-
dose NSAIDs seem to have clinical significance, and the
lack of statistical significance can be attributed to the
small sample size (- 80 patients in each dose group).
The comparison of recommended to supramaximal single
doses, on the other hand, confirms the analgesic ceiling
effect of NSAIDs, since not one of the four analgesic
scores evaluated showed either a clinical (0% to 4%) or
statistically significant difference.
The incidence of 14 episodes of side effects per 100
patients induced by single doses of NSAIDs is the same
as during placebo treatment, and insignificantly lower
than with morphine (5 to 10 mg intramuscularly), which
indicates that single doses of NSAIDs are relatively safe.
However, not unexpectedly, the incidence increases dra-
matically to 59 episodes per 100 patients when repeated
doses are given. Gastrointestinal symptoms are consid-
ered a common side effect of NSAIDs.4 5 Clinically sig-
nificant neurologic side effects are unusual with
NSAIDs,4 and the relatively high incidence of dizziness
and drowsiness reported in the present studies is not en-
tirely clear. A possible explanation is that patients en-
rolled were monitored more closely than in general prac-
tice. In contrast, renal and hepatic side effects of NSAIDs
were not reported, presumably since these usually occur
following the administration of these drugs on a long-
term basis. Further, many trials examined herein excluded
patients with preexisting medical conditions other than
cancer (and, even if not explicitly stated, patients with
Side Effects
Recommended Supro-maximal RD 95%CI
21 (11)/36 (19) -. 07 -. 15-.01
NSAIDS FOR CANCER PAIN 2763

Table 13. Side Effects: NSAIDs Versus Weak Opioids or WO/C
(multiple doses; episodes per 100 patients)
No. of Side Effects
No. of No. of
Studies Comparisons Patients NSAIDs WO/C RD
4 4 107/113 67 (63)/69 (62) -. 01
NOTE. Drugs used: NSAIDs-diclofenac, ketorolac; WO/C-pentazo-
cine, ASA + codeine, acetaminophen + codeine. Reports used: Carlson
(1990),20 Estape (1990),22 Minotti (1989),29 Tonachella (1985).41
preexisting hepatic and renal failure were likely to be
excluded).
The single-dose side effect incidence showed a nonsig-
nificant trend to increase dose-dependently and following
supramaximal compared with recommended doses. How-
ever, previously presented analyses indicate that the side
effect incidence during repeated supramaximal doses of
NSAID is much higher, without analgesic superiority
compared with multiple recommended doses. These find-
ings therefore support the concept that in cancer patients,
as in patients with other types of pain, high doses of
NSAIDs are likely to be associated with a high incidence
of side effects but with no additional analgesia.
The WHO cancer pain analgesic ladder for cancer pain
treatment recommends the use of NSAIDs or other non-
opioid analgesics as the first step when pain is mild to
moderate. In the next step, for moderate to severe pain
that persists after taking the first step, weak opioids are
added or substituted. The current meta-analysis compared
analgesic efficacy of single doses of the drugs recom-
mended for steps 1 and 2. As shown in Table 3, WO/C
did not provide analgesic efficacy superior to NSAIDs.
The meta-analysis included a relatively large number of
patients (> 450 in each treatment) and was based on
eight or nine comparisons of four or five NSAIDs versus
seven WO/Cs. It seems therefore to indicate strongly that
the second-step medications as a group do not provide
more analgesia than the first-step drugs. Multiple-dose
analysis of the same groups of medications, although
based on a small number of patients, further supports
these findings: NSAIDs (ketorolac in this case) continue
to have analgesic efficacy rates of 49% to 62% over a
period of 7 days, which is comparable to the effect of
WO/Cs (Table 8).
Table 14. Patient Satisfaction: Global Evaluation
(NSAID v morphine, single dose)
Effect
Table 15. Patient Satisfaction: Global Evaluation
(recommended v supramaximal single doses of NSAID)
Effect
95% CI No. of No. of No. of No. of Recommended Supramaximal
Studies Comparisons Drugs Patients (%) (%) RD 95% CI
-. 38-.37
3 3 2 112/112 62 62 -. 01 -. 16-.12
NOTE. Drugs used (recommended dose): ketoprofen orally 75 v 225 mg
(orally 50-100 mg); ketoprofen orally 100 v 300 mg (orally 50-100 mg);
zomepirac orally 100 v 200 mg (orally 100 mg). Reports used: Stambaugh
(1982,35 1988a"6), Sunshine (1988).40
Furthermore, a large number of patients (n = 1,138)
were included in the analysis that found a significantly
lower incidence of side effects induced by single doses
of NSAIDs compared with WO/C drugs (14 v 20 epi-
sodes/100 patients). A smaller scale analysis of multi-
ple doses studies demonstrated a markedly higher yet
almost identical incidence of side effects for both
groups of medications (63 v 62 episodes per 100 pa-
tients). Thus, although on a long-term basis the use of
NSAIDs in cancer patients seems to be associated with
a high incidence of side effects, it is still comparable
to that of WO/C drugs. Additionally, the dropout rate
analysis of patients treated with multiple doses of
NSAIDs over 7 to 10 days showed no difference from
that of the WO/C group.
In conclusion, these meta-analyses provide evidence
that NSAIDs reduce cancer-related pain significantly
more than placebo, and support the WHO recommenda-
tion to avoid the use of placebo in treating cancer pain.
NSAID-induced analgesia has a trend to be dose-depen-
dent up to a ceiling. The incidence of side effects also
tends to have a dose-response relationship, but with-
out a ceiling effect. The incidence of side effects in-
creases dramatically with multiple dosing over 7 to 10
days.
The present meta-analyses compared the first with
the second steps of the WHO drug therapy ladder from
several points of view. The results demonstrate that
both single and multiple doses of WO/Cs do not pro-
duce more analgesia than NSAIDs alone. Single doses
Table 16. NSAIDs Versus Opioids or WO/C:
Study Dropouts Over 7 to 10 Days
Dropouts (%)
No. of No. of No. of
Studies Comparisons Patients NSAIDs WO/C RD 95% CI

No. of No. of No. of No. of NSAIDs Morphine 5 6 217/233 69 (27)/70 (30) -. 05 -. 15-.04
Studies Comparisons Drugs Patients (%) (%) RD 95% CI
NOTE. Drugs used: NSAIDs-diclofenac, ketorolac, zomepirac; WO/
2 4 2 144/139 65 57 .08 -. 07-.25
C-pentazocine, ASA + codeine, acetaminophen + codeine, ASA + oxy-
NOTE. Drugs used: ketoprofen, zomepirac v morphine (5-10 mg IM). codone. Reports used: Carlson (1990),20 Estape (1990),22 Minotti (1989),29
Reports used: Stambaugh (1982),3" Sunshine (1988)." Stambaugh (1980),13 Tonachella (1985).41

Downloaded from ascopubs.org by DFG Deutsche Forschungsgemeinschaft on July 22, 2019 from 141.002.140.067
Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
2764
of WO/Cs produce significantly more side effects than
NSAIDs, although both the side effect incidence and
the dropout rates are equal when multiple doses are
administered. These findings raise the question whether
the WHO second analgesic step is an optimal treatment
protocol or whether should it be modified so as to pro-
ceed directly from NSAIDs to strong opioids in the
face of unrelieved pain.
Finally, a lack of comparable studies precluded us
from conducting an analysis to examine the analgesic
efficacy of NSAIDs specifically in malignant bone pain.
Further well-designed analgesic trials in which bone
pain or pain due to other specific cancer-related syn-
dromes is assessed separately from pain not of bony
origin are needed to answer this question. Related is-
REFERENCES
1. World Health Organization: Cancer Pain Relief. Geneva, Swit-
zerland, World Health Organization, 1986
2. World Health Organization Expert Committee: Cancer pain
relief and palliative care. Geneva, Switzerland, World Health Orga-
nization, 1992
3. American Pain Society: Principles of Analgesic Use in the
Treatment of Acute Pain and Cancer Pain. Skokie, IL, IASP, 1992
4. Portenoy RK: The management of cancer pain. Compr Ther
16:53-65, 1990
5. de Jong RH: Drug therapy in cancer pain. J Med Assoc Ga
80:95-300, 1991
6. Ashburn MA, Lipman AG: Management of pain in the cancer
patient. Anesth Analg 76:402-416, 1993
7. Olsen KD, Creagan ET: Pain management in advanced carci-
noma of the head and neck. Am J Otolaryngol 12:154-160, 1991
8. Inturrisi CE: Management of cancer pain. Cancer 63:2308-
2320, 1989
9. Kanner RM: Pharmacological management of pain and symp-
tom control in cancer. J Pain Sympt Manage 2:S19-22, 1987
10. Grond S, Zech D, Schung SA, et al: Validation of World
Health Organization guidelines for cancer pain relief during the last
days and hours of life. J Pain Sympt Manage 6:411-422, 1991
11. Schug SA, Zech D, Dorr U: Cancer pain management ac-
cording to WHO analgesic guidelines. J Pain Sympt Manage 5:27-
32, 1990
12. Takeda F: Results of field-testing in Japan of the WHO draft
interim guidelines on relief of cancer pain. Pain Clin 1:83-89, 1986
13. Ventafridda V, Tamburini M, Caraceni A, et al: A validation
study of the WHO method for cancer pain relief. Cancer 59:850-
856, 1987
14. Walker VA, Hoskin PJ, Hanks GW: Evaluation of WHO
analgesic guidelines for cancer pain in a hospital-based palliative
care unit. J Pain Sympt Manage 3:145-149, 1988
15. Twycross RG, Lack SA: Symptom Control in Far Advanced
Cancer: Pain Relief. London, England, Pitman, 1983
16. L'Abbe KA, Detsky AS, O'Rourke K: Meta-analysis in clini-
cal research. Ann Intern Med 107:224-233, 1987
17. Sacks HS, Berrier J, Reitman D: Meta-analysis of randomized
controlled trials. N Engl J Med 316:450-455, 1987
18. Thacker SB: Meta-analysis. A quantitative approach to re-
search integration. JAMA 259:1685-1689, 1988
19. DerSimonian R, Laird NM: Meta-analysis in clinical trials.
Control Clin Trials 7:177-188, 1986
Downloaded from ascopubs.org by DFG Deutsche Forschungsgemeinschaft on July 22, 2019 from 141.002.140.067
Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
EISENBERG ET AL
sues to be addressed in these and other future studies
include the efficacy and safety of NSAID therapy dur-
ing chronic dosing, possible benefits and risks of using
NSAIDs that spare platelet or other organ function, and
cost-effectiveness of NSAIDs versus opioids.
ACKNOWLEDGMENT
This study was performed in connection with development of
clinical practice guidelines on cancer pain management47 by a panel
cochaired by Dr Carr and Ada Jacox, RN, PhD, Independence Foun-
dation Professor, Johns Hopkins University School of Nursing, Balti-
more, MD. Primary literature searches undertaken by lone Auston
(National Library of Medicine), Janice Ulmer, RN, PhD, and Donna
Mahrenholz, RN, PhD, Johns Hopkins University School of Nursing.
We thank Dr Richard Kitz for continued support and encouragement.
Evelyn Hall provided expert secretarial assistance.
20. Carlson WR, Borrison AR, Harvey BS: A multiinstitutional
evaluation of the analgesic efficacy and safety of ketorolac trometha-
dine, acetaminophen plus codeine, and placebo in cancer pain. Phar-
macotherapy 10:211-216, 1990
21. Epstein JB, Stevenson-Moore P: Benzydamine hydrochloride
in prevention and management of pain in oral mucositis associated
with radiation therapy. Oral Surg Oral Med Oral Pathol 62:145-148,
1986
22. Estape J, Vinolas N, Gonzales B, et al: Ketorolac, a new non-
opioid analgesic: A double-blind trial versus pentazocine in cancer
pain. J Intern Med Res 18:298-304, 1990
23. Ferrer-Brechner T, Ganz P: Combination therapy with ibupro-
fen and methadone for chronic cancer pain. Am J Med 77:78-83,
1984
24. Kim JH, Chu F, Lakshmi V, et al: A clinical study of benzyd-
amine for the treatment of radiotherapy-induced mucositis of the
oropharynx. Int J Tissue React 7:215-218, 1985
25. Levick S, Jacobs C, Loukas DF, et al: Naproxen sodium in treat-
ment of bone pain due to metastatic cancer. Pain 35:253-258, 1988
26. Lomen PL, Samal BA, Lamborn KR: Flubiprofen for the
treatment of bone pain in patients with metastatic breast cancer. Am
J Med 80:83-87, 1986
27. Martino G, Ventafridda V, Parini J: A controlled study on
the analgesic activity of indoprofen in patients with cancer pain, in
Bonica JJ, Albe-Fessard DG (eds): Advances in Pain Research and
Therapy. New York, NY, Raven, 1976, pp 573-578
28. Martino G, Emanueli A, Mandelli V: A controlled study of
the analgesic effect of two nonsteroidal anti-inflammatory drugs in
cancer pain. Drug Res 28:1657-1659, 1978
29. Minotti V, Patoia L, Roila F: Double-blind evaluation of
analgesic efficacy of orally administered diclofenac, nefopam, and
acetylsalicylic acid (ASA) plus codeine in chronic cancer pain. Pain
36:177-183, 1989
30. Moertel CG, Ahmann DL, Taylor WF: Relief of pain by
oral medications. A controlled evaluation of analgesic combinations.
JAMA 229:55-59, 1974
31. Pellegrini A, Massidda B, Pellegrini P: Effect of iv indoprofen
on cancer pain and serum prolactin and growth hormone levels: A
controlled pharmacologic study vs im morphine and placebo. Int J
Clin Pharmacol Ther Toxicol 21:483-486, 1983
32. Puglisi G, Garagiola U: Piprofen and pentazocine in the treat-
ment of cancer pain. Curr Ther Res 45:333-338, 1989
NSAIDs FOR CANCER PAIN
33. Stambaugh JE, Tejada F, Trudnowski RJ: Double-blind com-
parisons of zomepirac and oxycodone with APC in cancer pain. J
Clin Pharmacol 20:261-270, 1980
34. Stambaugh JE, Sarajian C: Analgesic efficacy of zomepirac
sodium in patients with pain due to cancer. J Clin Pharmacol 21:501-
507, 1981
35. Stambaugh JE: Analgesic efficacy, safety and acceptability
of zomepirac sodium in comparison to morphine sulphate in the
treatment of pain secondary to malignancy. Curr Ther Res 31:922-
929, 1982
36. Stambaugh JE, Drew JA: A double-blind parallel evaluation
of the efficacy and safety of a single dose of ketoprofen in cancer
pain. J Clin Pharmacol 28:S34-39, 1988
37. Stambaugh JE, Drew J: The combination of ibuprofen and
oxycodone/acetaminophen in the treatment of chronic cancer pain.
Clin Pharmacol Ther 44:665-669, 1988
38. Staquet MJ: A double-blind study with placebo control of
intramuscular ketorolac tromethadine in the treatment of cancer pain.
J Clin Pharmacol 28:1031-1036, 1989
39. Sacchetti G, Camera P, Rossi AP: Injectable ketoprofen vs.
acetylsalicylic acid for the relief of severe cancer pain: A double-
blind crossover trial. Drug Intell Clin Pharm 18:403-406, 1984
40. Sunshine A, Olson NZ: Analgesic efficacy of ketoprofen in
Downloaded from ascopubs.org by DFG Deutsche Forschungsgemeinschaft on July 22, 2019 from 141.002.140.067
Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
2765
post partum, general surgery, and chronic cancer pain. J Clin Phar-
macol 28:S47-54, 1988
41. Tonachella R, Gallo Curcio C, Grossi E: Diclofenac sodium
in cancer pain. A double-blind within patients comparison with pen-
tazocine. Curr Ther Res 37:1130-1133, 1985
42. Turnbull R, Hill LJ: Naproxen versus aspirin as analgesics in
advanced malignant disease. J Palliat Care 1:25-28, 1986
43. Ventafridda V, Martino G, Mandelli ScD: Indoprofen, a new
analgesic and anti-inflammatory drug in cancer pain. Clin Pharmacol
Ther 17:284-287, 1974
44. Ventafridda V, De Conno F, Panerai AE: Nonsteroidal anti-
inflammatory drugs as the first step in cancer pain therapy: Double-
blind, within patient study comparing nine drugs. J Int Med Res
18:21-29, 1990
45. Soll AH, Weinstein WM, Kurata J: Nonsteroidal anti-in-
flammatory drugs and peptic ulcer disease. Ann Intern Med 114:307-
319, 1991
46. Lewis JH: Hepatic toxicity of nonsteroidal anti-inflammatory
drugs. Clin Pharm 3:128-138, 1984
47. Carr DB, Jacox A, Payne R, et al: Management of Cancer
Pain. Clinical Practice Guideline No. 9. Rockville, MD, Agency for
Health Care Policy and Research, US Department of Health & Hu-
man Services, Public Health Service, 1994 (AHCPR pub. no. 94-
0592)