close

Вход

Забыли?

вход по аккаунту

?

Estimates of Doses of Antiinflammatory Drugs in Man by Testing for Analgesic Potency.

код для вставкиСкачать
1381
ESTIMATES OF DOSES OF
ANTIINFLAMMATORY DRUGS IN
MAN BY TESTING FOR ANALGESIC
POTENCY
I. I-ISOPRPYL-4 PHENYL-7-METHYL-2 (1H) QUINAZOLONE VERSUS
ASPIRIN
T. G.
KAdfbR;
A. STREEM, and E. LASKA
Dosage estimates of antiinflammatory drugs in human arthritis Phase I1 trials are difficult to obtain and
prolong such trials unnecessarily. Antiinflammatory drugs
almost always have analgesic properties in man and good
dose estimates for analgesic activity can be obtained.
In 140 patients with surgical pain, 300, 600, and
1200 mg of aspirin were compared to 75,150, and 300 mg
of 43-715 ( 1-isopropyl-4 phenyl-7-methyl-2 ( 1H) quinazolone), an antiinflammatory quinazolone derivative, for
analgesia in a double-blind trial using subjective response
methodology. The test drug was shown to be analgesic at a
level four times more potent, milligram for milligram,
than aspirin, an estimate that should be useful for later
definitive Phase I1 trials in arthritis.
As an experimental model for determining the
dose range of a new antiinflammatory drug, rheumatoid
arthritis has many shortcomings. Except in multiclinic
trials, the organization of which require considerable
expense and extraordinary managerial skill, it is difficult
Supported in part by USPHS Grant #DAO1679 and The
Sandoz Foundation.
T. G. Kantor, M.D.: N Y U School of Medicine, 550 First
Avenue, New York. New York; A. Streem, R.N.: N Y U School of
Medicine; E. Laska. Ph.D.: Rockland State Hospital, Orangeburg,
New York.
Address reprint requests to Thomas G . Kantor, M.D., Department of Medicine, N Y U Medical Center, 550 First Avenue, New
York, New York 10016.
Submitted for publication April 4, 1977; accepted April 16,
1977.
Arthritis and Rheumatism, Vol. 20, No. 7 (September-October 1977)
to obtain large numbers of patietits with similar disease
activity for parallel, single-treatmedt comparative studies. Thete is often a wide range in degree of activity, by
any parameters of measurement, among patients in even
the largest clinical facilities. The usual counter to this
situation is the use of cross-over siudies, but the vagaries
of the disease in any particular patient over a reasonable
period of time, plus the long-term effects of many previous treatments, make such studies difficult to interpret. In addition, people with the disease are usually
seen as out-patients. They have often become pharmacological sophisticates, faced with a disease of unknown
etiology and uncertain treatment, and they frequently
interfere with clean, double-blind treatments schedules
because of their ready access to known effective agents
containing salicylates. I f salicylates are forbidden during
the course of a trial, a high percentage of drop-outs
occurs. If salicylate levels in blood or urine are monitored, the trial statistician is uncertain as to the usefulness of the data if “cheating” is found. If liberal access
to salicylates is allowed, a numerical count of their use
may form a measurement parameter, but questions of
variable interaction with the test compound must be
answered and individual differences in salicylate levels
are common.
A possible answer to this problem is initial examination of a more easily defined model so that dosage,
duration of effect, and bioassay comparison to known
standards can more readily be determined. Postoperative pain may form such a model. The pain is
KANTOR ET A L
1382
43-715
Fig. 1. I-luopropyl-4phenyl-7-methyl-2
( I H ) quinazolone.
clearly associated with a well-defined inflammation, and
large numbers of subjects are easily available and well
suited to parallel, single-treatment trial designs. In addition, the trial in each patient is of short duration. Although such a model does not answer questions of longterm toxicity, development of tolerance, or possible
drug interaction, it might produce dosage estimates useful for late Phase I1 and Phase 111 human trials.
Most authorities consider aspirin to be analgesic
( I ) , and there is abundant evidence that it has antiinflammatory properties in humans as well (2,3). The
relationship between the two activities is still obscure,
but definitive experiments suggest that aspirin reduces
pain through an action at the periphery and not in the
central nervous system (4). I n the same experiments,
narcotics were shown to cause their analgesic effects
centrally and not in the peripheral tissues. Other drugs
considered exclusively antiinflammatory have been
shown to have analgetic effects in experimental models
that are ordinarily considered to demonstrate analgesia
in man (5-7).
If we consider the above, it is reasonable to test
drugs that have shown an almost exclusive antiinflammatory effect in animal models for analgesic activity in man. Despite the evidence in man, neither indomethacin nor phenylbutazone can be easily shown to
have analgesic effects in conventional animal models.
Nor, for that matter, can many analgesic agents be
shown to have antiinflammatory activity, but at least
there is some agreement between animal and human
models in this regard.
Analgesic testing of new drugs as a part of subjective response trial methodology has become well established. It is possible to perform bioassay comparisons with known analgetic and antiinflammatory
compounds which give reasonable numerical estimates
of comparative analgesic potency (8), as long as their
plasma half-lives are reasonably close.
In order to explore this situation further, 43-7 15,
a quinazolone with well characterized antiinflammatory
effect in animals models was compared with aspirin for
analgesic effect in man. This paper dekcribes a trial using
oral aspirin as a standard analgesic-antiinflammatory
agent. The test drug (43-71 5), I-isopropyl-4 phenyl-7methyl-2 ( 1 H) quinazolone (Figure I ) , is well characterized in the laboratory as an anti-inflammatory agent
with some analgesic properties (Product Information
Bulletin, Pre-Clinical Data, Sandoz, Basel, 1968). Preliminary testing in normal human volunteers has suggested a low level of toxicity over a fairly wide dose
range and a plasma half-life about that of moderate
doses of aspirin (Product Information Bulletin, Sandoz,
Department of Clinical Research, 1973).
METHODS
Seven treatment groups, composed of a total of 140
postsurgical or posttraumatic patients from Bellevue Hospital’s
Surgical and Trauma Services, were treated with single doses,
randomly assigned, but balanced among the seven groups.
Double-blind conditions were imposed. Three doses of aspirin,
300, 600, and 1200 mg, formed three of the treatment groups
as did three doses of 43-715, 75, 150, and 300 mg. A placebo
group was added. Consenting patients complaining of moderate or severe pain were admitted to the study and were monitored at half-hour intervals for the first hour and hourly
thereafter for 6 hours after dosing. The variety of conditions
treated are listed i n Table I . As has been done arbitrarily in
Table 1. Pain Diagnoses
Condition
Fractures
Limbs
Spine and pelvis
Orthopedic or peripheral
neurosurgery
Plastic surgery
General surgery
(incisions)
TOTAL
No.
31
13
27
7
20
Condition
Gun shot and stab
wounds
Vascular (gangrene or
venous ulcers)
Bone cancer
Gynecologic pain
Traumatic contusions
Other
No.
II
15
4
I
5
2
136
DOSES O F ANTIINFLAMMATORY DRUGS IN MAN
previous studies, patients not wishing to continue up to I’hour
were removed from the study and treated with conventional
analgetic regimens. Those continuing for 2 hours or more and
requesting further relief were also conventionally medicated
but their data were retained. These were assumed t o remain in
the same categories of relief and pain score for the remainder
of the 6-hour period and the data were so recorded. In this
study, n o patients were dropped within the first hour. Four
patients were dropped when they were inadvertently medicated with analgesics by other hospital personnel (Table I ) .
As in our previously reported trials (9,10), pain categories were scored as “0” for no pain, “ I ” for slight, “2” for
moderate, and “3” for severe pain at the observation periods.
At each interval, patients were also asked whether or not at
least 50% of their pain had been relieved. Any movement
required to elicit pain at zero time was repeated exactly at each
observation interval.
From these raw data, half-hour and hourly pain intensity difference (PID) scores were derived with reference to
the zero time score, the mean value, and the standard error for
each drug category determined, and the sum of the pain intensity difference (SPID) mean scores over the 6 hour observa-
1383
tion period was calculated. The mean percentage of positive
answers to the relief question and the standard error for each
drug category was also determined as were the total hours of
relief (TOTAL) for 6 hours. The resultant calculations produced 22 separate values for each drug category (Table 2 ) and
these were plotted as time-effect curves (Figures 2, 3, and 4 ) .
The two summary values, SPID and TOTAL, form a numerical expression of the areas under the individual drug category
curves depicted (Figures 3 and 4). These in turn were plotted
as log dose-effect curves as in Figures 5 and 6. Under the
assumption that the log dose response curves are parallel. their
intercepts and slopes may be estimated by the least squares
method. The distance between the two lines forms a numerical
estimate of the relative potency of the two drugs. Placebo is
not plotted because no log dose value can be assigned to it.
Statistical evaluation for each of the 24 measurements
is based on an analysis of variance to test for possible drug
effects among the dosages. T tests are also calculated to test for
drug difference between each possible drug pair. The bioassay
calculations are based on classic techniques described by Finney ( I I ). Briefly, the three mathematical assumptions made in
the analysis are:
Table 2. Summary of Means and Standard Errors (in Hours)
Placebo
(N = 20)
Variable Name
Mean
S.E.
Aspirin,
300 mg
(N = 20)
Mean
S.E.
Aspirin,
600 mg
(N = 20)
Mean
S.E.
Pain Zero
Pain One-half
Pain One
Pain Two
Pain Three
Pain Four
Pain Five
Pain Six
Relief One-half
Relief One
Relief Two
Relief Three
Relief Four
Relief Five
Relief Six
PID One-half
PID One
PID Two
PID Three
PID Four
PID Five
PID Six
SPID
2.550
1.750
1.200
1.550
1.500
1.800
1.950
2.050
0.150
0.500
0.500
0.300
0.300
0.250
0.250
0.800
1.350
1.050
1.100
0.800
0.650
0.550
5.225
0.11 I
0.198
0.219
0.229
0.229
0.230
0.193
0.180
0.080
0.112
0.112
0.102
0.102
0.097
0.097
0.182
0.237
0.229
0.233
0.207
0.146
0.150
0.950
2.400 0.110
1.800 0.182
1.750 0.171
1.750 0.185
1.750 0.171
1.950 0.150
1.900 0.172
1.900 0.186
0.0
0.0
0.250 0.097
0.250 0.097
0.300 0.102
0.250 0.097
0.150 0.080
0.200 0.089
0.600 0.164
0.650 0.128
0.650 0.162
0.650 0.146
0.500 0.1 12
0.550 0.150
0.550 0.165
3.525 0.706
2.300
1.500
1.450
1.450
1.450
1.500
1.650
1.800
0.200
0.400
0.350
0.500
0.450
0.450
0.350
0.800
0.950
0.900
0.950
0.900
0.750
0.600
4.975
TOTAL
1.925 0.463
1.275 0.377
2.400 0.555
* PID = Pain intensity differences.
t SPID = sum of pain intensity differences (PIDS).
$ TOTAL = Total hours of 50% or more relief of pain.
0.102
0.180
0.229
0.229
0.240
0.240
0.227
0.219
0.089
0.110
0.107
0.112
0.1 1 I
0.11 I
0.107
0.182
0.165
0.199
0.180
0.186
0.185
0.179
0.989
Aspirin,
1,200mg
(N = 19)
Mean
2.474
1.895
1.316
1.211
1.OOO
1.158
1.21I
1.316
0.211
0.632
0.684
0.789
0.684
0.684
0.684
0.632
1.158
1.316
1.526
1.368
1.316
1.21I
7.632
S.E.
0.1 15
0.165
0.198
0.218
0.235
0.250
0.253
0.269
0.094
0.11 1
0.107
0.094
0.107
0.107
0.107
0.170
0.201
0.224
0.228
0.238
0.236
0.242
1.217
3.947 0.521
43-715,75mg
(N = 19)
Mean
2.316
1.632
1.421
1.211
1.737
1.737
1.842
1.947
0.158
0.421
0.421
0.316
0.316
0.263
0.158
0.737
0.895
1.105
0.684
0.684
0.526
0.421
4.237
S.E.
0.107
0.200
0.172
0.191
0.221
0.221
0.186
0.203
0.184
0.113
0.113
0.107
0.107
0.101
0.084
0.146
0.181
0.222
0.211
0.211
0.172
0.172
0.981
1.763 0.477
43-715, 150mg 43-715.300mg
(N = 20)
(N = 18)
Mean
2.500
1.700
1.650
0.950
1.150
1.500
1.750
1.850
0.150
0.550
0.800
0.650
0.450
0.350
0.300
0.800
0.850
1.550
1.350
1.OOO
0.800
0.700
6.225
S.E.
0.112
0.160
0.191
0.165
0.215
0.194
0.21 1
0.203
0.080
0.11 I
0.089
0.107
0.1 1 1
0.107
0.102
0.134
0.146
0.193
0.227
0.187
0.195
0.160
0.909
2.900 0.452
Mean
2.444
1.500
1.389
1.167
1.333
1.333
1.389
1.500
0.389
0.444
0.611
0.556
0.444
0.444
0.389
0.944
1.056
1.278
1.167
1.167
1.1 I I
1.000
6.722
S.E.
0.1 17
0.239
0.210
0.212
0.236
0.261
0.262
0.264
0.1 15
0.1 17
0.115
0.1 17
0.1 17
0.117
0.1 15
0.215
0.184
0.189
0.212
0.226
0.234
0.236
1.087
2.861 0.554
KANTOR ET AL
1384
50% Relief
Mean Pain
2 75
Nz20
W Placebo
20 0 300mg ASA
20 A 600mg ASA
19
*
l2OOmgASA
19 0 75mg43-715
20 0 150mg43-715
18 A 300mg 43-715
2 25
60 VI
2 50-
-
a
.+
.-c
b
1.75 -
a 40-
Y
0
a
0
* 30-
-
C
b
r" 1.25 -
20 -
i
75
10-
I
1
1
20 0 iSOm~43-715
18 A 300mg 43-715
I
5
1/2
1
2
3
6
4
Hours
Fig. 2. Nutnerical values on verrical axis refer
10
data on Table 2 .
I . Each response variable is a linear function of the dose.
2. The response variable for both test and standard drugs can
be represented by parallel plots on a log dose scale.
3. The statistical errors are not correlated and are normally
distributed.
The validity of the analysis requires that the doses of test and
standard be in an approximately equianalgesic range, because
otherwise extrapolations cannot be made with confidence. The
relative potency is the dose of test preparation which would
produce the same response as a given dose of the standard
preparation for each of the parameters. Within the analysis of
variance for each of the parameters, tests for parallelism of the
two curves, slopes different from zero, and difference between
the preparations are performed. Feiler's theorem is then used
as the basis for calculation of fiducial (confidence) limits of the
estimated relative potency for each of the parameters (12)
(Table 3).
RESULTS
Simple inspection of the data shows that separation of the low doses of either test drug (43-715) and
standard (ASA) from placebo has not been accomplished. However, T tests show that in 9 of the 24
Pain Intensity Difference
8
m
al
V
C
150
-
f
Y
'c
0 100-
x
.c
_
v)
C
-
al
t
C
-
c
.-
0
a
5019
19 0
20 0
18 A
1200m~ASA
75mg 23-715
150mg43-715
300mg43-715
DOSES O F ANTIINFLAMMATORY DRUGS IN MAN
Parallel
Line Regression
(Potency 3.98)
43-715
6’0 1
5.25
4.50
1
0
ASA
rn 43-715
1385
for all parameters are shown in Table 3. These are found
to be 4.69 for SPID and 3.98 for TOTAL with respective
95% confidence limits of 2A4-12.4 (SPID) and 1.98-7.97
(TOTAL)or, as depicted in the table, at 99% confidence
limits of 1.3-30.4 (SPID) and 1.3-1 1.8 (TOTAL).A slight
reversal of slope is noted for the higher two doses of 43715 for the global parameter TOTAL. In general, SPID
is considered to be the more sensitive index in studies of
this type.
Adverse effects were almost non-existent in this
single dose study.
DISCUSSION
0
’0 3.00
The relative effectiveness of placebo in this study
is not unexpected and inspection of Figures 2, 3, and 4
shows the rapid onset and rapid decline of effect which
are its usual characteristics in such studies. Because of
its “potency,” with 25-40% of patients finding adequate
relief of even the most severe pain, we use it simply as a
t
2.25
1.50
.75
Parallel
Line Regression
(Potency 4.69)
0
43 -715
Log 10 (Dose)
12.00-
Fig. 5 . Numerical values on vertical axis refer to data on Table 2.
Potency is relative potency value ( R H O on Table 3).
parameters, aspirin at 1200 mg is significantly more
analgesic than placebo at the 95% level of confidence.
Similar findings obtained for 43-715 when compared to
placebo for 3 of the 24 parameters at the 150-mg dose
level and 9 of 24 at the 300-mg level. At zero time, there
were no significant differences between the various drug
groups with respect to the mean pain category.
There appears to be some difference in onset of
analgesia between the high doses of aspirin (1200 mg)
and 43-715 (300 mg) with respect to both onset of response and time to peak effect. In both instances,43-715
is faster and the two high doses also maintain a higher
degree of analgesic effect throughout the 6-hour study
period. In no instance did any of the drug parameters,
including placebo, return to baseline. This is in keeping
with the transient, ever-improving nature of most postsurgical pain.
The analysis of variance for SPID and TOTAL
indicates that the tests for parallelism, positive slope,
and no difference between preparations are such that we
may assume a valid assay. Relative potency estimates
10.50-
9.00-
0
ASA
43-715
7.50
-
0
$, 6.00
6.00 ~
4.50 /
0
3.00-
/./,
/ O
1.500
,
.5
/
1
1.5
2
,
2.5
3
,
3.5
Log 10 (Dose)
Fig. 6 . Numerical values on vertical axis refer to data on Table 2 .
Potency is relalive potency value ( R H O on Table 3).
KANTOR ET A L
1386
Table 3. Summary of Bioassay Results (in Hours)
Variable Name
Pain Zero
Pain One-half
Pain One
Pain Two
Pain Three
Pain Four
Pain Five
Pain Six
Relief One-half
Relief One
Relief Two
Relief Three
Relief Four
Relief Five
Relief Six
PID One-half
PID One
PID Two
PID Three
PID Four
PID Five
PID Six
SPID
TOTAL
Slope of
Standard
Slope of
Test
Potency
0.118
0. I47
-0.724
-0.897
- I .244
-1.317
-1.142
-0.965
0.325
0.633
0.718
0.812
0.72 1
0.888
0.802
0.058
0.845
1.103
1.452
6.160
1.089
6.804
0.217
-0.214
-0.046
0.080
-0.682
-0.671
-0.149
-0.739
0.380
0.043
0.325
0.405
0.216
0.301
0.385
0.344
0.263
0.298
0.815
0.804
0.971
0.961
4.152
4.433
1.842
1.441
I .267
Finite
CL*
Lower
CL
4.427
2 1.969
4.033
4.152
3.349
3.107
7.163
5.583
9. I99
3.759
3.078
3.029
2.464
22.216
4.248
13.1 13
4.305
4.272
3.557
3.360
4.691
60.0
10.0
75.0
80.0
95.0
99.0
95.0
95.0
95.0
90.0
99.0
99.0
95.0
99.0
99.0
50.0
90.0
95.0
99.0
99.0
99.0
99.0
99.0
1.410
0.000
0.01 I
7.127
1.035
0.006
0.573
0.090
2.878
1.364
3.327
0.830
0.47 1
0.444
0.263
6.579
0.786
4.416
1.106
1.165
0.684
0.333
1.306
3.978
99.0
1.326
(RHO)
-
Upper
CL
--t
163.674
-
16.079
-
10.985
13.986
108.611
491.516
-
13.994
9.043
8.803
6.133
-
31.613
-
21.608
19.164
12.589
13.925
30.438
I 1.798
* CL = confidence limits.
t - = infinite.
reference marker in this type of trial. A more reasonable
criterion for the study’s validity is the fact that positive
dose response curves are noted for both active compounds.
In comparison wih other clinical analgesic trials.
the confidence limits achieved for the relative potency
values in the present study are quite good, at least sufficient to suggest an effective dose range for 43-715 in
subsequent clinical trials. Although only analgesia has
been examined here, the nature of the pain strongly
suggests association with inflammation. Among nonnarcotic analgesic drugs, only a few are considered to
show a dissociation between analgesia and antiinflammatory effect ( 1 ).
Inferentially, 43-71 5 should prove to have useful
antiinflammatory as well as analgesic properties when
applied to other clinical models such as rheumatoid
arthritis. This will require experimental verification.
Suggested dose levels would be 150 mg, qid.
CONCLUSIONS
I . A bioassay study for analgesic activity in man has
been performed comparing an antiinflammatory
quinazolone (43-715) with aspirin.
2. The study was done to determine a dose range for
late Phase I1 clinical studies of antiinflammatory effect.
3. It is estimated that 43-715 has approximately four
times the analgesic potency of aspirin on a milligram
per milligram basis.
4. The usefulness of this estimate remains to be seen in
clinical trials now underway on rheumatic patients.
REFERENCES
1. Beaver WT: Mild analgesics. A review of their clinical
pharmacology. A m J Med Sci 251:577-604, 1965;
252:576-599, 1966
DOSES OF ANTIINFLAMMATORY DRUGS I N MAN
2. Fremont-Smith K, Bayles TB: Salicylate therapy in rheumatoid arthritis. J Am Med Assoc 192:103-106, 1965
3. Boardman PL, Hart FD: Clinical management of the antiinflammatory effects of salicylates in rheumatoid arthritis.
Br Med J 4:264-268, 1967
4. Guzman F, Braun C, Lim RKS, et al: Narcotic and nonnarcotic analgesics which block visceral pain evoked by
intra-arterial injection of bradykinin and other algesic
agents. Arch Int Pharmacodyn 149:571-589, 1964
5. Sunshine A, Laska E, Meisner M, et al: Analgesic studies
of indomethacin as analyzed by computer techniques.
Clin Pharmacol Ther 5:699-707, 1964
6. Domenjoz R: The pharmacology of phenylbutazone analogues. Ann NY Acad Sci 86:263-291. 1960
7. Kimsey LR: Treatment of superficial thrombophlebitis
with phenylbutazone. J A M A 172:229-230, 1960
1387
8. Sunshine A, Laska E, Slafta J, et al: A comparative analgesia study of propoxyphene hydrochloride, propoxyphene napsylate and placebo. Tox Appl Pharmacol
19:512-518, 1971
9. Kantor T G , Sunshine A, Laska E, et al: Oral analgesic
studies: pentazocine hydrochloride, codeine, aspirin and
placebo and their influence on response to placebo. Clin
Pharmacol Ther 7:447-454, 1966
10. Kantor TG, Laska E, Streem A: An apparent algesic effect
of meprobamate. J Clin Pharmacol 13:152-1 59 1973
1 I . Finney DT: Statistical Method in Biological Assay. London, Charles Griffen & Co, Ltd, 1952
12. Laska E, Gormley M, Sunshine A, et al: A bioassay computer program for analgesic clinical trials. Clin Pharmacol
Ther 8:658-669, 1967
Primer on the Rheumatic Diseases
Work will soon begin on a new edition (eighth) of the Primer on
the Rheumatic Diseases. The editor of this work welcomes suggestions
from the readers of Arthritis and Rheumatism for new subjects which
they would like to see included in the Primer, and any other advice they
may wish to offer concerning this publication. Please address correspondence to: Dr. Gerald Rodnan, 985 Scaife Hall, Department of Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania 15261.
Документ
Категория
Без категории
Просмотров
0
Размер файла
466 Кб
Теги
potency, testing, drug, estimates, dose, antiinflammatory, analgesia, man
1/--страниц
Пожаловаться на содержимое документа