5
Pain, 41 (1990) 5-13 Elsevier
PAIN 01574
Clinical Section
Clinical analgesic assay of repeated and single doses of heroin and hydromorphone Stanley L. W~lenstein,
Raymond
W. Houde, Russell Portenoy, and Kathleen M. Foley
Jeanne
Lapin, Ada Rogers
Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10021 (U.S.A.) (Received
26 May 1989, revision received
1 December
1989, accepted
5 December
1989)
A direct comparison of the analgesic activities of heroin and hydromo~hone was carried out in cancer patients with s==n=Y postsurgical pain. Intramuscnlar doses of 5 and 10 mg of heroin were compared with 1 and 2 mg of hydromorphone in a randomized, double-blind, 4-point parallel group assay. Design innovations in the study provided that about half the patients would receive prior repeated doses of the same drug as the test medication, and half would receive the alternate medication. Both test drugs were found to be potent, relatively short acting analgesics with similar profiles of action. Hydromorphone was about 5 times as potent as heroin on a ~~~arn basis. The comparison of those patients who had repeated doses of the same treatment prior to the test dose and those who had repeated doses of the alternate drug demonstrated no significant effect on the relative potency estimates. Side effect occurrence was similar for both drugs, with sleepiness the most prominent effect. The study supports the view that hydromorphone and heroin produce similar clinical effects, and that either drug may adequately substitute for the other. Covariate analysis indicated that time since last analgesic was positively related to analgesia, and amount of prior opioid had a negative relationship. To a lesser extent, increase in patient age was associated with an increase in analgesic scores. Taking these covariates into account served to increase the sensitivity of the analysis. Key wordsr
Heroin;
Diacetylmorphine;
Diamorphine;
Hydromorphone;
Introduction The pharmacological profile of heroin (diacetylmorphine, diamorphine) is quite similar to that of another morphine congener, hydromorphone (Dilaudid) [4-71. Both drugs have a rapid onset of action and produce shorter durations of
Correspondence to: Dr. Russell Portenoy, Kettering Cancer 10021, U.S.A.
0304-3959/90/$03.50
Center,
1275 York Avenue,
Memorial SloanNew York, NY
0 1990 Elsevier Science Publishers
Analgesia;
Potency;
Cancer;
Pam
analgesia than morphine, and at doses estimated to be equianalgesic, both drugs appear to have comparable side effects. In clinical analgesic potency assays, parenteral heroin hydrochloride was found to be twice as potent as morphine sulfate on a milligram basis [7], whereas hydromorphone hydrochloride was demonstrated to be 8 times as potent [5]. Both drugs are more soluble than morphine (heroin about 8 times and hydromorphone about 5 times) [lo], so that large doses of either heroin or hydromorphone can be given in a small volume of fluid.
B.V. (Biomedical
Division)
Heroin has been illegal in the United States since 1924. ~ydromo~hone~ in contrast, is used extensively in this country for the control of severe pain. Proponents of special legislation that would provide for increased availability of heroin in the United States contend that this drug may have unique properties which would make it the analgesic of choice in certain clinical situations [9], while those opposed to this view suggest that hydromorphone can substitute for heroin due to its similar pharmacologic profile [ll]. Resolution of this controversy has been hindered by the lack of direct comparisons of the analgesic and moodaltering effects of these two drugs in terms of either acute effects or long-term use. The present study was undertaken to compare directly the analgesic efficacy and side effects of heroin and hydromorphone. The specific objective was to establish the relative analgesic potency of heroin and hydromo~hone in cancer patients with postoperative pain under conditions in which each drug is assayed following repeated doses of itself or of the alternate drug. This study design also allows for direct comparison of the effects of the drugs on mood, sedation and other side-effects. The influence of selected patient variables (e.g., age, prior opioid usage, and severity of pain) on the analgesic actions of the two drugs was also examined.
Methods Approval of the study was obtained from the Institutional Review Board of Memorial SloanKettering Cancer Center. The methodology employed a randomized, double-blind parallel group comparison of 5 and 10 mg of intramuscular heroin and 1 and 2 mg of intramuscular hydromorphone in patients with postoperative pain. The study population comprised consenting adult inpatients with moderate to severe postsurgical pain with no medical contraindications to opioid treatment. Among patients placed on study, there were 162 abdominal procedures, 33 thoracotomies, 8 orthopedic procedures and 6 excisions of soft tissue cancers. On the first postoperative day, patients were
randomly assigned either open label heroin or hydromo~hone as their postoperative analgesic. Doses of the drug were individualiy adjusted by the nurse observer to provide adequate pain relief. and patients received this medication on an as needed schedule (q 3 h prn) for moderate or severe pain overnight. On the second postoperative day. the patient received a dose of test medication when moderate or severe pain was reported at least 3 h after their last analgesic. The test dose was either 5 or 10 mg of intramuscular heroin, or I or 2 mg of intramuscular hydromorphone. These doses were administered double-blind according to a randomized schedule that was designed to ensure that each dose of test medication followed the same or the other medication an approximately equal number of times. Study medications were administered by a nurse observer who assessed drug effects at 15 and 30 min after the test treatment, then hourly for up to 6 h. Assessments were terminated earlier if pain returned to the pretreatment level and a subsequent analgesic was requested and administered. At each assessment period, patients reported pain severity and perceived relief from the test dose using standard categorical scales for these endpoints [16]. The Memorial Pain Assessment Card (MPAC) was also administered at each interview period; this instrument includes an g-point Extended Pain Score (EPS) and visual analog scales (VAS) for pain, pain relief and mood [2,13,15]. The EPS was derived from a pain scale employed by Tursky in experimental pain studies 1121. A VAS for sedation was also employed.
Duta anaiysis Differences between pain intensity at 0 h and at each assessment thereafter were calculated for both categorical and VAS pain intensity scales. These pain intensity difference (PID) values were employed to determine peak effect (peak PID) and a sunny value for total pain intensity difference (SPID). The percent of maximum pain intensity differences (% PID) was calculated as follows: % PID = 100 x PID/PI(O
h).
This formula determines the percent of the maximum possible reduction in pain severity for each patient in terms of the pain level at the time of administration of the test drug. This adjustment serves to equate scores for different starting pain levels (i.e., the % PID for a report of no pain after initial pain reports of either moderate or severe would both be lOO%, as contrasted with scores of 3 or 4, respectively, in terms of the unadjusted PID). The 5%PID values were employed to determine peak 5%PID and % SPID for both categorical and VAS scales (81. Consistent with our findings that the EPS pain categories are related to the VAS Pain Intensity scale as a power function (a log-log relationship implying unequal distances between the categories on the scale) [14], the logarithms of the 8 ordinal categories in the scale were used as the basic pain measures for calculating peak EPS PID and EPS SPIT). Categorical and VAS Pain Relief scales were employed directly to calculate peak relief, and the summary value for total relief (TOTPAR). Both VAS Mood and VAS Sedation were evaluated as direct hourly measures for up to 3 h after receiving the test medication. Duration of analgesia was measured in terms of the time to remedication in hours. For the calculation of SPID and TOTPAR values when patients were remedicated during the observation period, it was assumed that the pain had returned to the premeditation level and that there was no subsequent relief of pain for the remainder of the observation period. Results were evaluated by analysis of variance. Relative potency and associated confidence limits were calculated according to the method of Finney [l]. The effects of age, prior opioid experience, and pain intensity at 0 h were evaluated in a covariance analysis. Two aspects of prior opioid experience were measured separately: the amount of opioids given in the 48 h prior to administration of the test drug was calculated in terms of morphine equivalents [4]; and the length of time on opioids was calculated employing an arbitrary 9-point scale extending from 0 for no prior opioid experience to 8 for use of opioids for over 6 months. An additional covariance factor, VAS Pain Intensity at 0 h was considered in the analy-
sis of the categorical and VAS Pain Relief measures. Finally, the data were evaluated to determine whether differences in the relative potency estimates of the test drugs would become evident if the test doses followed the same or the alternate drug. This provided information about relative potency at approximate steady state, as well as following acute administration, and would also potentially identify any presumptive evidence of acute tolerance.
Results Of 208 patients assigned to the study who had received open label heroin or hydromorphone, 161 received test treatments and completed the study. Of the 47 dropouts, 27 were the result of side effects experienced on the first postoperative day. These effects, some of which may not have been drug related, occurred in 17 patients after the priming doses of hydromo~hone and in 10 patients after heroin. Side effects after hydromorphone included dizziness, nausea, itching, blurred vision, restlessness, sedation, tightness in chest, rash, swelling extremities, bad dreams and palpitations. Those observed after heroin included dizziness, nausea, itching, grogginess, disorientation and sleepiness. One patient refused to continue because of inadequate relief on hydromorphone. The remaining 19 patients failed to complete for non-study-related reasons. Ten had insufficient pain, 3 refused to answer questions, 2 had difficulty answering the questions, 1 refused to take a double-blind dose, and 3 patients were discontinued because of change in medical status (1 spontaneous respiratory arrest due to a mucous plug, 1 unstable cardiac status and 1 death). Demographic information for the patients completing the study is summarized in Table I. The proportion of men to women was greater in the 2 hydromorphone groups than in the heroin groups. The patients receiving the 1 mg dose of hydromorphone were on average younger than the patients on the other treatments, but the range of ages for all 4 treatments was roughly similar. No significant group differences were observed in prior opioid experience as calculated in terms of
‘I-ABLE I DEMOGRAPHIC
‘TABLE 11
DATA FOR PATIENTS
RELATIVE POTENCY DROMORPHONE
ASSAY
OF
COMPLET:TING HEROIN AND HY-
Hydr*mo~h~~e
Heroin
5w
10 mg
1 mg
2mfs
i.m.
i.m.
i.m.
im.
N
40
Male Female
24 16
38 22 16
41 32 9
42 31 11
Age bfean (SE.) Range
52.4 (2.35) 19-74
52.5 (2.77) 19-75
47.2 (2.36) 18-73
54.3 (2.24) 20-73
Surgery Abdo~n~ Thoracotomy Orthopedic Soft tissue
30 7 2 1
32 4 2 0
29 8 2 2
32 8 0 2
Antecedent opioid medication *
11.6
11.9
13.2
10.6
Baseiine Pam VAS Mood VAS Sedation
2.43 54.5 41 .h
2.32 49.9 46.4
2.34 52.5 38.7
2.57 60.9 34.8
* In mean morphine
equivalents
for 48 h before study drug.
morphine equivalents administered in the 48 h prior to receiving the study medication, nor were there any significant differences among the groups in the length of time on opioids. Group differences in baseline pain intensity, mood or degree of sedation were also not statistically significant. Table II summarizes the analgesic data obtained. Relative potency estimates calculated for total effect (an estimate of the area under the time-effect curves) varied from 4.794 (TOTPAR) to 5.898 (VAS SPID). ‘I%le relative potency in terms of peak effect ranged from 3.646 for Peak Relief to 5.846 for VAS peak PID. For the most part, 95% confidence limits of these estimates were relatively narrow. Lambda 131,a measure of assay sensitivity, varied from 0.36 for VAS SPID to 0.768 for Pet VAS peak PID. Lambda is caiculated as the ratio of common slope of the assay to the standard error (s/B). The lower the value of
RELATIVE POTENCY, CONFIDENCE LIMITS AND LAMBDA ESTIMATES FOR THE ANALGESIC ASSAY OF HYDROMORPHONE AND OXYMORPHONE WITH 95% CONFIDENCE LIMITS CALCULATER AFTER COVARIANCE * ANALYSIS Variable
R (rel pot)
TOTPAR Peak rel VAS TOTPAR VAS peak rel SPID Peak PID VAS SPlD VAS peak PID Pet SPID Pet peak PID Pet VAS SPID Pet VAS peak PID Log SPID EPS Log peak PID EPS
4,794 3.646 5.630 4.301 5.064 4.987 5.898 5.846 5.073 4.477 5.325 4.147 4.668 4.605
95% limits 3.606-5.131 1.442-5.085 3.X45-5.325 1.918-5.21X 3.360-5.240 2.891-5.282 4.444-5.470 2.414-6.355 3.273-5.258 2.403-5.209 3.989-5.208 1.64-5.203 2.889-5.204 2.019-5.313 ____I--
Lambda 0.369 0.598 0.501 0.629 0.514 0.584 0.360 0.588 0.531 0.595 11.398 0.768 0.343 0.441
* Covariates are: age. time since last analgesic. prior opioid experience in the last 48 b. duration of prior opioid experience. For pain relief estimates, pain severity at 0 h was also employed as a covariate.
lambda, the greater the assay sensitivity, and those obtained in this st.udy are in the low range for clinical analgesic assays, which in our experience
Fig. I. Dose-effect curves for intramuscular heroin (circles, and intramuscular hydromorphone (squares) in a parallel group assay in 161 cancer patients with postoperative pain. Data points are plotted both before (solid Enes) and after (broken lines) correction for the effects of the covariates: age, am5ut-d of opioid consumption in the 48 h prior to receiving the study drug, length of time on opioids, and VAS pain intensity at 0 h. VAS TOTPAR is platted against the lop of the dose.
9
commonly varies from 0.5 to 0.8 for assays of parenteral opioids in cancer patients. Thus, consistent results were obtained across the various analgesic measures. The analyses for VAS TOTPAR and VAS Peak Relief parameters are illustrative (Figs. 1 and 2). The analysis of variance for VAS TOTPAR in Table III is presented as a prototype, since generally similar results were obtained with the other parameters employed in the study. This analysis demonstrates that the assay was sensitive to increased effect with increased dose (note that the pooled regression of analgesic effect on dose yields a highly significant F ratio), while differences in the individual drug slopes and the effect levels of the 2 preparations were minimal (reported as Parallelism and Preparations, respectively in the table). A relatively constant ratio of doses for any measured effect level without undue extrapolation is thus observed, fulfilling the major criteria for a valid assay.
TABLE
+ Cm e-oti3oinIn e--aHeroln WITHMVfRIAiES ~Hydro It4 m--rttydro + Cm
tEKJIN--
56,
I I
0
0.25
I I
1 t
6.5
0.5
1
Log dose (mgs) Fig. 2. Dose-effect curves for intramuscular heroin (circles) and intramuscular hydromorphone (squares) in a parallel group assay in 161 cancer patients with postoperative pain. Data points are plotted both before (solid lines) and after (broken lines) correction for the effects of the covariates age, amount of opioid consumption in the 48 h prior to receiving the study drug, length of time on opioids, and VAS pain intensity at 0 h. VAS Peak Relief is plotted against the log of the dose.
III
ANALYSIS OF VARIANCE FOR VAS TOTPAR BY TREATMENTS WITH COVARIATES, AND RELATIVE ESTIMATES (R) FOR VAS TOTPAR AND VAS PEAK RELIEF WHEN TEST DOSES FOLLOW REPEATED EITHER THE SAME OR ALTERNATE DRUG Source of variation
df
Mean square
F
Signif. P
Covariates Age Last analgesic Opioid history 48 h duration VASPIOh
5 1 1
270390.437 51214.212 132205.452
16.124 3.054 7.884
< 0.001 0.083 0.006
1 1 1
216606.690 5592.996 24390.767
12.917 0.334 1.454
< 0.001 > 0.1 > 0.1
Treatments Pooled regression Parallelism Preparations
3 1 1 1
90594.341 252408.337 1623.292 10910.176
5.402 15.011 0.097 0.649
0.001 < 0.001 z 0.1 > 0.1
Error
152
16769.425
R*
VAS TOTPAR 5.630 (3.844-5.325)
VAS Peak Relief 4.301 (1.918-5.218)
4.631 (2.008-5.217) 6.092 \_ .:“7-5.875)
4.172 (0.792-5.221) 5.134 (2.562-5.719)
After repeated Heroin Hydromorphone * R = relative potency
of hydromorphone
to heroin.
POTENCY DOSES OF
10
Significant covariance effects were identified for the amount of antecedent opioid medicatiun and time since the last analgesic. with the effect of age approaching, but not reaching significance. Taken as m group, the covariates produced a significant regression on the analgesic scores and accounted for a considerable amount of variation in the study, These effects were consistently observed with the other parameters of analgesic effect, as well as with VAS TOTPAR. When the data were analyzed in terms of the drug administered in repeated on-demand doses in the period prior to the test dose, the relative patency of hydromorphone to heroin was numerically greater after repeated doses of heroin. This difference, however, was not statistically significant, leaving to conjecture whether the difference was the result of a real phenomenon based on the possibility of differential cumulative drug effects, or merely random variability in the data. Similar but smaller differences were observed in terms of the peak effects of the two preparations. Time-effect curves for analgesia were similar for both drugs, as demonstrated by the curves for VAS Pain Relief (Fig. 3), and for pain intensity scores as measured by EPS categories (Fig. 4). The time-effect curves discriminated between the doses of both drugs, and both preparations were relatively rapid-acting, reaching mean peak effects
k4 B
I
I !
1
2
I
I
3 Tome in
3
/
5
w 3.5+--i----+-------
--+ -..-_-.-_+“. _..- ..-
j
Fig. 4. Time-effect cuwes for 5 mg (circles) and IO mg (‘triangles) of intramuscufar heroin (solid fines) and f mg (squares) and 2 mg (crosses) of intramuscular hydromorphone (broken lines) in 161 cancer patients with postoperative pain. Extended Pain Score is plotted against time in hours.
appro~mate~y OS h after ad~nistrat~on. Significant differences (P c 0.05) between the 2 doses of heroin were obtained at 30 min and at I, 3, 4 and 5 h. and between the 2 doses of hydromorphone at 30 min and at 1-4 h after drug in terms of Extended Pain Scores. Analysis of the time-effect curves for the other analgesic parameters produced similar results. The time-effect curves for mood and sedation (Figs. 5 and 6) show little discrimination among the doses in terms of these parameters. VAS Mood
-4 6
Hours
Fig. 3. Time-effect curves for 5 mg (circles) and 10 mg (triangles) of intramuscular heroin (solid lines) and 1 mg (squares) and 2 mg (crosses) of intramuscular hydromorphone (broken lines) in 161 cancer patients with postoperative pain. VAS Pain Relief is plotted against time in hours.
Fig. 5. Time-effect curves for 5 mg (circles) and 10 mg (triangles) of intramuscular heroin (solid lines) and 1 mg (squares) and 2 mg (crosses) of intramuscular hydromorphone (broken lines} in 161 cancer patients with postoperative pain. VAS Sedation is platted against time in hours.
11 TABLE
IV
SIDE EFFECT
OCCURRENCE Hydromorphone
Heroin
Sleepy, drowsy
d--0
I
1
1
2
I 3
licne in Hours Fig. 6. Time-effect curves for 5 mg (circles) and 10 mg (triangles) of intramuscular heroin (solid lines) and 1 mg (squares) and 2 mg (crosses) of intramuscular hydromorphone (broken lines) in 161 cancer patients with postoperative pain. VAS Mood is plotted against time in hours.
was better after the 2 mg ~dromo~hone dose, and VAS Sedation was greater in the patients receiving 10 mg of heroin, but none of these observed differences was statistically significant. These time-effect curves are illustrated for only the first 3 h after medication since over half of the patients were remedicated after that time and there is rapid decline in the number of patients included in the subsequent measurements. Fig. 7 illustrates the hourly decline in the percent of patients remaining in study as determined
First
Second
lhu-d
Fourth
Fifth
Sixth
Hours Fig. 7. Percentage of patients on study at each hour who had not received a rescue analgesic because of return of pain during the 6 h observation period. The percentages for each of the 4 study treatments are plotted for the 6 h observation period.
Dizzy, groggy Shaky Lightheaded Weak Heaviness Itching Sweating Flushed, warm Nausea, bloated Dry mouth Headache Blurred vision Burning eyes Difficulty concentrating Floating feeling Disoriented Crying Visual hallucination Pain af irtjection site Number of patients Number of side effects Patients with side effects
5 mg i.m.
10 mg i.m.
(1 mg i.m.
2 mg i.m.
23 2
21 7
22 7
2
2 1
22 3 1 2 1 1 2
2 4
1
2 1
4 1 1 1 1 1
41 33
41 45
42 43
1 1 I 42 39
27
30
27
26
by the time to reme~cation for each treatment. Greater differences were observed between the dose levels of the drugs than between the drugs themselves, with about half the patients requiring additional medication 3 h after receiving the lower dose of either medication, compared to 4 or 5 h after the upper doses. Overall side effect occurrence was similar for both drugs (Table IV). In terms of observed or reported side effects, sleepiness was by far the most common effect produced by either drug, followed by dizziness, grogginess, li~theade~ess and nausea. Itching was more common after heroin and was apparently dose-related, while there were more headaches after hydromorphone than heroin. These differences were relatively minor, and the general configurations of side effects for both drugs were opioid-like and similar.
Discussion
Both heroin and hydromorphone are effective analgesics in cancer patients with postoperative pain. Hydromorphone is approximately 4.5-5.5 times as potent as heroin. At equianalgesic doses. the 2 drugs are quite similar. Both are relatively short acting analgesics with rapid onset of action. Peak effects for both are observed at about 0.5 h after administration, and the relative potencies of the 2 drugs in terms of peak and total effects are also similar. There is little difference between the sedative actions of the 2 drugs, except that a trend toward greater sedation in the early hours after the upper dose of heroin was observed. Similarly, no significant differences in mood were observed, although a trend toward better mood among those treated with hydromorphone 2 mg was noted. Two innovations in analgesic drug study methodology were employed in this study, one more successful than the other. The study design was modified to include a period of analgesic drug administration prior to administration of the challenge test dose. This was done to test the hypothesis that an observable degree of acute tolerance might develop to the test drug and would be more complete than cross-tolerance to an alternate drug. Patients were thus administered either heroin or hydromorphone on demand for pain relief for a 24 h period prior to administration of the test drug. This design, however, has serious limitations which were dictated by the practicality of the clinical situation in these postoperative patients. A steady state is not reached in on-demand drug administration, but rather unpredictable and variable blood levels. How this might affect acute tolerance is not well understood. Furthermore, partitioning the study into those who had repeated administration of the test drug and those who did not, served to cut the effective study numbers in half and significantly reduce the sensitivity of the study when analyzed for that purpose. The effects of the test drug in those patients with no prior exposure to the drug were only minimally different from the effects in patients who had received repeated prior administrations of the same drug. The possibility of reduced effect on repeated administration remains an issue awaiting a definitive study. Never-
theless, the study does serve, with limitations, as a measure of the stability of relative potency estimates under the specified conditions of repeated versus non-repeated doses. The second methodological innovation involved an attempt to determine which demographic factors served to influence the patients’ analgesic responses to drugs. Covariance analysis was employed to evaluate the effects of patient age, the amount of opioid taken in the previous 48 h (calculated as morphine equivalents), the length of time on opioids. and pain severity at time of administration of the test drug. In our cancer patients with postoperative pain, a significant positive correlation was obtained in terms of time since last analgesic. Antecedent opioid medication as measured by opioid consumption in the 48 h period prior to administration of the test drug was negatively correlated with analgesia. To a lesser degree, a positive correlation with patient age was observed. Length of time on opioids and pain intensity at the time of administration of the test medication were less significant as covariates. Statistical analysis taking the appropriate covariante factors into account can reduce error variance and improve assay efficiency. These associations with analgesic response also serve to cast new light on some aspects of patients’ pain behavior. In addition, the significant influence of prior opioid experience serves as a salient factor relevant to analgesic study design: that even in noncross-over studies. drug interaction with prior analgesics can well exist, and this problem is not solved by doing parallel group assays rather than cross-over studies. Indeed, some measure of precision may be lost in parallel group as compared with cross-over studies. It can be concluded from this study that heroin and hydromorphone are potent and effective analgesics with similar profiles of activity. Although acute single-dose assays, or short-term repeated dose studies, cannot exclude the possibility of the emergence of distinguishing clinical effects after repeated dosing, the data here presented provide supporting evidence for the view that, at equianalgesic doses, either drug can potentially substitute for the other.
13
Acknowledgements This project was supported by National Cancer Institute Grant CA38297. Dr Portenoy’s research is supported by American Cancer Society Grant JFRA-244. We are grateful to Judith Weiss, RN and Lorraine Scimone, RN for their diligence and care in collecting the data for this study.
References Finney, D.J., Statistical Method in Biological Assay, 2nd edn., Hafner, New York, 1964, pp. 98-117. Fishman, B., Pastemak, S., Wallenstein, S.L., Houde, R.W., Holland, J.C. and Foley, K.M., The Memorial Pain Assessment Card: a valid instrument for the evaluation of cancer pain, Cancer, 60 (1987) 1151-1158. Gaddum, J.H., Bioassays and mathematics, Pharmacol. Rev., 5 (1953) 87-134. Houde, R.W., Systemic analgesics and related drugs: opioid analgesics. In: J.J. Bonica and V. Ventafridda (Eds.), Advances in Pain Research and Therapy, Vol. 2, Raven Press, New York, 1979, pp. 263-273. Houde, R.W., Clinical studies of hydromorphone. In: K.M. Foley and C.E. Intunisi (Eds.), Advances in Pain Research and Therapy, Vol. 8, Raven Press, New York, 1986, pp. 129-135. Inturrisi, C.E., Max, M.B., Foley, K.M., Schultz, M., Shin, S.U. and Houde, R.W., The pharmacokinetics of heroin in patients with chronic pain, New Engl. J. Med., 310 (1984) 1213-1217.
7 Kaiko, R.F., Wallenstein, S.L., Rogers, A.G., Grabinski, P.Y. and Houde, R.W., Analgesic and mood effects of heroin and morphine in cancer patients with postoperative pain, New Engl. J. Med., 304 (1981) 1501-1505. 8 Laska, E.M., Sunshine, A., Zighelboim, I., Roure, C., Marrero, I., Wanderling, J. and Olsen, N., Effects of caffeine on acetaminophen analgesia, Clin. Pharmacol. Ther.. 33 (1983) 498-509. 9 Levine, M.N., Sackett, D.L. and Bush, H.. Heroin vs morphine for cancer pain. Arch. Intern. Med., 146 (1986) 353-356. 10 Merck Index, 10th Edition, Merck and Co., Rahway, NJ, 1983. 11 Sawynok, J., The therapeutic use of heroin: a review of the pharmacological literature, Can. J. Physiol. Pharmacol., 64 (1986) l-6. 12 Tursky, B., The development of a pain perception profile: a psychological approach. In: W. Weisenberg and B. Tursky (Eds.), Pain: New Perspectives in Therapy and Research, Plenum Press, New York, 1976, pp. 171-194. 13 Wallenstein, S.L., Measurement of pain and analgesia in cancer patients, Cancer, 53 (1984) 2260-2264. 14 Wallenstein, S.L., Scaling clinical pain and pain relief. In: B. Bromm (Ed.), Pain Measurement in Man. Neurophysiological Correlates of Pain, Elsevier Science Publishers, Amsterdam, 1984, pp. 389-396. 15 Wallenstein, S.L., Heidrich. III, G., Kaiko. R. and Houde, R.W., Clinical evaluation of mild analgesics: the measurement of clinical pain, Br. J. Pharmacol., 10 (1980) 319S3273. 16 Wallenstein, S.L. and Houde, R.W., The clinical evaluation of analgesic effectiveness. In: S. Ehrenpreis and A. Neidle (Eds.), Methods in Opioid Research, Marcel Dekker, New York, 1975, pp. 127-145.
Pain, 41 (1990) 5-13 Elsevier
PAIN 01574
Clinical Section
Clinical analgesic assay of repeated and single doses of heroin and hydromorphone Stanley L. W~lenstein,
Raymond
W. Houde, Russell Portenoy, and Kathleen M. Foley
Jeanne
Lapin, Ada Rogers
Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10021 (U.S.A.) (Received
26 May 1989, revision received
1 December
1989, accepted
5 December
1989)
A direct comparison of the analgesic activities of heroin and hydromo~hone was carried out in cancer patients with s==n=Y postsurgical pain. Intramuscnlar doses of 5 and 10 mg of heroin were compared with 1 and 2 mg of hydromorphone in a randomized, double-blind, 4-point parallel group assay. Design innovations in the study provided that about half the patients would receive prior repeated doses of the same drug as the test medication, and half would receive the alternate medication. Both test drugs were found to be potent, relatively short acting analgesics with similar profiles of action. Hydromorphone was about 5 times as potent as heroin on a ~~~arn basis. The comparison of those patients who had repeated doses of the same treatment prior to the test dose and those who had repeated doses of the alternate drug demonstrated no significant effect on the relative potency estimates. Side effect occurrence was similar for both drugs, with sleepiness the most prominent effect. The study supports the view that hydromorphone and heroin produce similar clinical effects, and that either drug may adequately substitute for the other. Covariate analysis indicated that time since last analgesic was positively related to analgesia, and amount of prior opioid had a negative relationship. To a lesser extent, increase in patient age was associated with an increase in analgesic scores. Taking these covariates into account served to increase the sensitivity of the analysis. Key wordsr
Heroin;
Diacetylmorphine;
Diamorphine;
Hydromorphone;
Introduction The pharmacological profile of heroin (diacetylmorphine, diamorphine) is quite similar to that of another morphine congener, hydromorphone (Dilaudid) [4-71. Both drugs have a rapid onset of action and produce shorter durations of
Correspondence to: Dr. Russell Portenoy, Kettering Cancer 10021, U.S.A.
0304-3959/90/$03.50
Center,
1275 York Avenue,
Memorial SloanNew York, NY
0 1990 Elsevier Science Publishers
Analgesia;
Potency;
Cancer;
Pam
analgesia than morphine, and at doses estimated to be equianalgesic, both drugs appear to have comparable side effects. In clinical analgesic potency assays, parenteral heroin hydrochloride was found to be twice as potent as morphine sulfate on a milligram basis [7], whereas hydromorphone hydrochloride was demonstrated to be 8 times as potent [5]. Both drugs are more soluble than morphine (heroin about 8 times and hydromorphone about 5 times) [lo], so that large doses of either heroin or hydromorphone can be given in a small volume of fluid.
B.V. (Biomedical
Division)
Heroin has been illegal in the United States since 1924. ~ydromo~hone~ in contrast, is used extensively in this country for the control of severe pain. Proponents of special legislation that would provide for increased availability of heroin in the United States contend that this drug may have unique properties which would make it the analgesic of choice in certain clinical situations [9], while those opposed to this view suggest that hydromorphone can substitute for heroin due to its similar pharmacologic profile [ll]. Resolution of this controversy has been hindered by the lack of direct comparisons of the analgesic and moodaltering effects of these two drugs in terms of either acute effects or long-term use. The present study was undertaken to compare directly the analgesic efficacy and side effects of heroin and hydromorphone. The specific objective was to establish the relative analgesic potency of heroin and hydromo~hone in cancer patients with postoperative pain under conditions in which each drug is assayed following repeated doses of itself or of the alternate drug. This study design also allows for direct comparison of the effects of the drugs on mood, sedation and other side-effects. The influence of selected patient variables (e.g., age, prior opioid usage, and severity of pain) on the analgesic actions of the two drugs was also examined.
Methods Approval of the study was obtained from the Institutional Review Board of Memorial SloanKettering Cancer Center. The methodology employed a randomized, double-blind parallel group comparison of 5 and 10 mg of intramuscular heroin and 1 and 2 mg of intramuscular hydromorphone in patients with postoperative pain. The study population comprised consenting adult inpatients with moderate to severe postsurgical pain with no medical contraindications to opioid treatment. Among patients placed on study, there were 162 abdominal procedures, 33 thoracotomies, 8 orthopedic procedures and 6 excisions of soft tissue cancers. On the first postoperative day, patients were
randomly assigned either open label heroin or hydromo~hone as their postoperative analgesic. Doses of the drug were individualiy adjusted by the nurse observer to provide adequate pain relief. and patients received this medication on an as needed schedule (q 3 h prn) for moderate or severe pain overnight. On the second postoperative day. the patient received a dose of test medication when moderate or severe pain was reported at least 3 h after their last analgesic. The test dose was either 5 or 10 mg of intramuscular heroin, or I or 2 mg of intramuscular hydromorphone. These doses were administered double-blind according to a randomized schedule that was designed to ensure that each dose of test medication followed the same or the other medication an approximately equal number of times. Study medications were administered by a nurse observer who assessed drug effects at 15 and 30 min after the test treatment, then hourly for up to 6 h. Assessments were terminated earlier if pain returned to the pretreatment level and a subsequent analgesic was requested and administered. At each assessment period, patients reported pain severity and perceived relief from the test dose using standard categorical scales for these endpoints [16]. The Memorial Pain Assessment Card (MPAC) was also administered at each interview period; this instrument includes an g-point Extended Pain Score (EPS) and visual analog scales (VAS) for pain, pain relief and mood [2,13,15]. The EPS was derived from a pain scale employed by Tursky in experimental pain studies 1121. A VAS for sedation was also employed.
Duta anaiysis Differences between pain intensity at 0 h and at each assessment thereafter were calculated for both categorical and VAS pain intensity scales. These pain intensity difference (PID) values were employed to determine peak effect (peak PID) and a sunny value for total pain intensity difference (SPID). The percent of maximum pain intensity differences (% PID) was calculated as follows: % PID = 100 x PID/PI(O
h).
This formula determines the percent of the maximum possible reduction in pain severity for each patient in terms of the pain level at the time of administration of the test drug. This adjustment serves to equate scores for different starting pain levels (i.e., the % PID for a report of no pain after initial pain reports of either moderate or severe would both be lOO%, as contrasted with scores of 3 or 4, respectively, in terms of the unadjusted PID). The 5%PID values were employed to determine peak 5%PID and % SPID for both categorical and VAS scales (81. Consistent with our findings that the EPS pain categories are related to the VAS Pain Intensity scale as a power function (a log-log relationship implying unequal distances between the categories on the scale) [14], the logarithms of the 8 ordinal categories in the scale were used as the basic pain measures for calculating peak EPS PID and EPS SPIT). Categorical and VAS Pain Relief scales were employed directly to calculate peak relief, and the summary value for total relief (TOTPAR). Both VAS Mood and VAS Sedation were evaluated as direct hourly measures for up to 3 h after receiving the test medication. Duration of analgesia was measured in terms of the time to remedication in hours. For the calculation of SPID and TOTPAR values when patients were remedicated during the observation period, it was assumed that the pain had returned to the premeditation level and that there was no subsequent relief of pain for the remainder of the observation period. Results were evaluated by analysis of variance. Relative potency and associated confidence limits were calculated according to the method of Finney [l]. The effects of age, prior opioid experience, and pain intensity at 0 h were evaluated in a covariance analysis. Two aspects of prior opioid experience were measured separately: the amount of opioids given in the 48 h prior to administration of the test drug was calculated in terms of morphine equivalents [4]; and the length of time on opioids was calculated employing an arbitrary 9-point scale extending from 0 for no prior opioid experience to 8 for use of opioids for over 6 months. An additional covariance factor, VAS Pain Intensity at 0 h was considered in the analy-
sis of the categorical and VAS Pain Relief measures. Finally, the data were evaluated to determine whether differences in the relative potency estimates of the test drugs would become evident if the test doses followed the same or the alternate drug. This provided information about relative potency at approximate steady state, as well as following acute administration, and would also potentially identify any presumptive evidence of acute tolerance.
Results Of 208 patients assigned to the study who had received open label heroin or hydromorphone, 161 received test treatments and completed the study. Of the 47 dropouts, 27 were the result of side effects experienced on the first postoperative day. These effects, some of which may not have been drug related, occurred in 17 patients after the priming doses of hydromo~hone and in 10 patients after heroin. Side effects after hydromorphone included dizziness, nausea, itching, blurred vision, restlessness, sedation, tightness in chest, rash, swelling extremities, bad dreams and palpitations. Those observed after heroin included dizziness, nausea, itching, grogginess, disorientation and sleepiness. One patient refused to continue because of inadequate relief on hydromorphone. The remaining 19 patients failed to complete for non-study-related reasons. Ten had insufficient pain, 3 refused to answer questions, 2 had difficulty answering the questions, 1 refused to take a double-blind dose, and 3 patients were discontinued because of change in medical status (1 spontaneous respiratory arrest due to a mucous plug, 1 unstable cardiac status and 1 death). Demographic information for the patients completing the study is summarized in Table I. The proportion of men to women was greater in the 2 hydromorphone groups than in the heroin groups. The patients receiving the 1 mg dose of hydromorphone were on average younger than the patients on the other treatments, but the range of ages for all 4 treatments was roughly similar. No significant group differences were observed in prior opioid experience as calculated in terms of
‘I-ABLE I DEMOGRAPHIC
‘TABLE 11
DATA FOR PATIENTS
RELATIVE POTENCY DROMORPHONE
ASSAY
OF
COMPLET:TING HEROIN AND HY-
Hydr*mo~h~~e
Heroin
5w
10 mg
1 mg
2mfs
i.m.
i.m.
i.m.
im.
N
40
Male Female
24 16
38 22 16
41 32 9
42 31 11
Age bfean (SE.) Range
52.4 (2.35) 19-74
52.5 (2.77) 19-75
47.2 (2.36) 18-73
54.3 (2.24) 20-73
Surgery Abdo~n~ Thoracotomy Orthopedic Soft tissue
30 7 2 1
32 4 2 0
29 8 2 2
32 8 0 2
Antecedent opioid medication *
11.6
11.9
13.2
10.6
Baseiine Pam VAS Mood VAS Sedation
2.43 54.5 41 .h
2.32 49.9 46.4
2.34 52.5 38.7
2.57 60.9 34.8
* In mean morphine
equivalents
for 48 h before study drug.
morphine equivalents administered in the 48 h prior to receiving the study medication, nor were there any significant differences among the groups in the length of time on opioids. Group differences in baseline pain intensity, mood or degree of sedation were also not statistically significant. Table II summarizes the analgesic data obtained. Relative potency estimates calculated for total effect (an estimate of the area under the time-effect curves) varied from 4.794 (TOTPAR) to 5.898 (VAS SPID). ‘I%le relative potency in terms of peak effect ranged from 3.646 for Peak Relief to 5.846 for VAS peak PID. For the most part, 95% confidence limits of these estimates were relatively narrow. Lambda 131,a measure of assay sensitivity, varied from 0.36 for VAS SPID to 0.768 for Pet VAS peak PID. Lambda is caiculated as the ratio of common slope of the assay to the standard error (s/B). The lower the value of
RELATIVE POTENCY, CONFIDENCE LIMITS AND LAMBDA ESTIMATES FOR THE ANALGESIC ASSAY OF HYDROMORPHONE AND OXYMORPHONE WITH 95% CONFIDENCE LIMITS CALCULATER AFTER COVARIANCE * ANALYSIS Variable
R (rel pot)
TOTPAR Peak rel VAS TOTPAR VAS peak rel SPID Peak PID VAS SPlD VAS peak PID Pet SPID Pet peak PID Pet VAS SPID Pet VAS peak PID Log SPID EPS Log peak PID EPS
4,794 3.646 5.630 4.301 5.064 4.987 5.898 5.846 5.073 4.477 5.325 4.147 4.668 4.605
95% limits 3.606-5.131 1.442-5.085 3.X45-5.325 1.918-5.21X 3.360-5.240 2.891-5.282 4.444-5.470 2.414-6.355 3.273-5.258 2.403-5.209 3.989-5.208 1.64-5.203 2.889-5.204 2.019-5.313 ____I--
Lambda 0.369 0.598 0.501 0.629 0.514 0.584 0.360 0.588 0.531 0.595 11.398 0.768 0.343 0.441
* Covariates are: age. time since last analgesic. prior opioid experience in the last 48 b. duration of prior opioid experience. For pain relief estimates, pain severity at 0 h was also employed as a covariate.
lambda, the greater the assay sensitivity, and those obtained in this st.udy are in the low range for clinical analgesic assays, which in our experience
Fig. I. Dose-effect curves for intramuscular heroin (circles, and intramuscular hydromorphone (squares) in a parallel group assay in 161 cancer patients with postoperative pain. Data points are plotted both before (solid Enes) and after (broken lines) correction for the effects of the covariates: age, am5ut-d of opioid consumption in the 48 h prior to receiving the study drug, length of time on opioids, and VAS pain intensity at 0 h. VAS TOTPAR is platted against the lop of the dose.
9
commonly varies from 0.5 to 0.8 for assays of parenteral opioids in cancer patients. Thus, consistent results were obtained across the various analgesic measures. The analyses for VAS TOTPAR and VAS Peak Relief parameters are illustrative (Figs. 1 and 2). The analysis of variance for VAS TOTPAR in Table III is presented as a prototype, since generally similar results were obtained with the other parameters employed in the study. This analysis demonstrates that the assay was sensitive to increased effect with increased dose (note that the pooled regression of analgesic effect on dose yields a highly significant F ratio), while differences in the individual drug slopes and the effect levels of the 2 preparations were minimal (reported as Parallelism and Preparations, respectively in the table). A relatively constant ratio of doses for any measured effect level without undue extrapolation is thus observed, fulfilling the major criteria for a valid assay.
TABLE
+ Cm e-oti3oinIn e--aHeroln WITHMVfRIAiES ~Hydro It4 m--rttydro + Cm
tEKJIN--
56,
I I
0
0.25
I I
1 t
6.5
0.5
1
Log dose (mgs) Fig. 2. Dose-effect curves for intramuscular heroin (circles) and intramuscular hydromorphone (squares) in a parallel group assay in 161 cancer patients with postoperative pain. Data points are plotted both before (solid lines) and after (broken lines) correction for the effects of the covariates age, amount of opioid consumption in the 48 h prior to receiving the study drug, length of time on opioids, and VAS pain intensity at 0 h. VAS Peak Relief is plotted against the log of the dose.
III
ANALYSIS OF VARIANCE FOR VAS TOTPAR BY TREATMENTS WITH COVARIATES, AND RELATIVE ESTIMATES (R) FOR VAS TOTPAR AND VAS PEAK RELIEF WHEN TEST DOSES FOLLOW REPEATED EITHER THE SAME OR ALTERNATE DRUG Source of variation
df
Mean square
F
Signif. P
Covariates Age Last analgesic Opioid history 48 h duration VASPIOh
5 1 1
270390.437 51214.212 132205.452
16.124 3.054 7.884
< 0.001 0.083 0.006
1 1 1
216606.690 5592.996 24390.767
12.917 0.334 1.454
< 0.001 > 0.1 > 0.1
Treatments Pooled regression Parallelism Preparations
3 1 1 1
90594.341 252408.337 1623.292 10910.176
5.402 15.011 0.097 0.649
0.001 < 0.001 z 0.1 > 0.1
Error
152
16769.425
R*
VAS TOTPAR 5.630 (3.844-5.325)
VAS Peak Relief 4.301 (1.918-5.218)
4.631 (2.008-5.217) 6.092 \_ .:“7-5.875)
4.172 (0.792-5.221) 5.134 (2.562-5.719)
After repeated Heroin Hydromorphone * R = relative potency
of hydromorphone
to heroin.
POTENCY DOSES OF
10
Significant covariance effects were identified for the amount of antecedent opioid medicatiun and time since the last analgesic. with the effect of age approaching, but not reaching significance. Taken as m group, the covariates produced a significant regression on the analgesic scores and accounted for a considerable amount of variation in the study, These effects were consistently observed with the other parameters of analgesic effect, as well as with VAS TOTPAR. When the data were analyzed in terms of the drug administered in repeated on-demand doses in the period prior to the test dose, the relative patency of hydromorphone to heroin was numerically greater after repeated doses of heroin. This difference, however, was not statistically significant, leaving to conjecture whether the difference was the result of a real phenomenon based on the possibility of differential cumulative drug effects, or merely random variability in the data. Similar but smaller differences were observed in terms of the peak effects of the two preparations. Time-effect curves for analgesia were similar for both drugs, as demonstrated by the curves for VAS Pain Relief (Fig. 3), and for pain intensity scores as measured by EPS categories (Fig. 4). The time-effect curves discriminated between the doses of both drugs, and both preparations were relatively rapid-acting, reaching mean peak effects
k4 B
I
I !
1
2
I
I
3 Tome in
3
/
5
w 3.5+--i----+-------
--+ -..-_-.-_+“. _..- ..-
j
Fig. 4. Time-effect cuwes for 5 mg (circles) and IO mg (‘triangles) of intramuscufar heroin (solid fines) and f mg (squares) and 2 mg (crosses) of intramuscular hydromorphone (broken lines) in 161 cancer patients with postoperative pain. Extended Pain Score is plotted against time in hours.
appro~mate~y OS h after ad~nistrat~on. Significant differences (P c 0.05) between the 2 doses of heroin were obtained at 30 min and at I, 3, 4 and 5 h. and between the 2 doses of hydromorphone at 30 min and at 1-4 h after drug in terms of Extended Pain Scores. Analysis of the time-effect curves for the other analgesic parameters produced similar results. The time-effect curves for mood and sedation (Figs. 5 and 6) show little discrimination among the doses in terms of these parameters. VAS Mood
-4 6
Hours
Fig. 3. Time-effect curves for 5 mg (circles) and 10 mg (triangles) of intramuscular heroin (solid lines) and 1 mg (squares) and 2 mg (crosses) of intramuscular hydromorphone (broken lines) in 161 cancer patients with postoperative pain. VAS Pain Relief is plotted against time in hours.
Fig. 5. Time-effect curves for 5 mg (circles) and 10 mg (triangles) of intramuscular heroin (solid lines) and 1 mg (squares) and 2 mg (crosses) of intramuscular hydromorphone (broken lines} in 161 cancer patients with postoperative pain. VAS Sedation is platted against time in hours.
11 TABLE
IV
SIDE EFFECT
OCCURRENCE Hydromorphone
Heroin
Sleepy, drowsy
d--0
I
1
1
2
I 3
licne in Hours Fig. 6. Time-effect curves for 5 mg (circles) and 10 mg (triangles) of intramuscular heroin (solid lines) and 1 mg (squares) and 2 mg (crosses) of intramuscular hydromorphone (broken lines) in 161 cancer patients with postoperative pain. VAS Mood is plotted against time in hours.
was better after the 2 mg ~dromo~hone dose, and VAS Sedation was greater in the patients receiving 10 mg of heroin, but none of these observed differences was statistically significant. These time-effect curves are illustrated for only the first 3 h after medication since over half of the patients were remedicated after that time and there is rapid decline in the number of patients included in the subsequent measurements. Fig. 7 illustrates the hourly decline in the percent of patients remaining in study as determined
First
Second
lhu-d
Fourth
Fifth
Sixth
Hours Fig. 7. Percentage of patients on study at each hour who had not received a rescue analgesic because of return of pain during the 6 h observation period. The percentages for each of the 4 study treatments are plotted for the 6 h observation period.
Dizzy, groggy Shaky Lightheaded Weak Heaviness Itching Sweating Flushed, warm Nausea, bloated Dry mouth Headache Blurred vision Burning eyes Difficulty concentrating Floating feeling Disoriented Crying Visual hallucination Pain af irtjection site Number of patients Number of side effects Patients with side effects
5 mg i.m.
10 mg i.m.
(1 mg i.m.
2 mg i.m.
23 2
21 7
22 7
2
2 1
22 3 1 2 1 1 2
2 4
1
2 1
4 1 1 1 1 1
41 33
41 45
42 43
1 1 I 42 39
27
30
27
26
by the time to reme~cation for each treatment. Greater differences were observed between the dose levels of the drugs than between the drugs themselves, with about half the patients requiring additional medication 3 h after receiving the lower dose of either medication, compared to 4 or 5 h after the upper doses. Overall side effect occurrence was similar for both drugs (Table IV). In terms of observed or reported side effects, sleepiness was by far the most common effect produced by either drug, followed by dizziness, grogginess, li~theade~ess and nausea. Itching was more common after heroin and was apparently dose-related, while there were more headaches after hydromorphone than heroin. These differences were relatively minor, and the general configurations of side effects for both drugs were opioid-like and similar.
Discussion
Both heroin and hydromorphone are effective analgesics in cancer patients with postoperative pain. Hydromorphone is approximately 4.5-5.5 times as potent as heroin. At equianalgesic doses. the 2 drugs are quite similar. Both are relatively short acting analgesics with rapid onset of action. Peak effects for both are observed at about 0.5 h after administration, and the relative potencies of the 2 drugs in terms of peak and total effects are also similar. There is little difference between the sedative actions of the 2 drugs, except that a trend toward greater sedation in the early hours after the upper dose of heroin was observed. Similarly, no significant differences in mood were observed, although a trend toward better mood among those treated with hydromorphone 2 mg was noted. Two innovations in analgesic drug study methodology were employed in this study, one more successful than the other. The study design was modified to include a period of analgesic drug administration prior to administration of the challenge test dose. This was done to test the hypothesis that an observable degree of acute tolerance might develop to the test drug and would be more complete than cross-tolerance to an alternate drug. Patients were thus administered either heroin or hydromorphone on demand for pain relief for a 24 h period prior to administration of the test drug. This design, however, has serious limitations which were dictated by the practicality of the clinical situation in these postoperative patients. A steady state is not reached in on-demand drug administration, but rather unpredictable and variable blood levels. How this might affect acute tolerance is not well understood. Furthermore, partitioning the study into those who had repeated administration of the test drug and those who did not, served to cut the effective study numbers in half and significantly reduce the sensitivity of the study when analyzed for that purpose. The effects of the test drug in those patients with no prior exposure to the drug were only minimally different from the effects in patients who had received repeated prior administrations of the same drug. The possibility of reduced effect on repeated administration remains an issue awaiting a definitive study. Never-
theless, the study does serve, with limitations, as a measure of the stability of relative potency estimates under the specified conditions of repeated versus non-repeated doses. The second methodological innovation involved an attempt to determine which demographic factors served to influence the patients’ analgesic responses to drugs. Covariance analysis was employed to evaluate the effects of patient age, the amount of opioid taken in the previous 48 h (calculated as morphine equivalents), the length of time on opioids. and pain severity at time of administration of the test drug. In our cancer patients with postoperative pain, a significant positive correlation was obtained in terms of time since last analgesic. Antecedent opioid medication as measured by opioid consumption in the 48 h period prior to administration of the test drug was negatively correlated with analgesia. To a lesser degree, a positive correlation with patient age was observed. Length of time on opioids and pain intensity at the time of administration of the test medication were less significant as covariates. Statistical analysis taking the appropriate covariante factors into account can reduce error variance and improve assay efficiency. These associations with analgesic response also serve to cast new light on some aspects of patients’ pain behavior. In addition, the significant influence of prior opioid experience serves as a salient factor relevant to analgesic study design: that even in noncross-over studies. drug interaction with prior analgesics can well exist, and this problem is not solved by doing parallel group assays rather than cross-over studies. Indeed, some measure of precision may be lost in parallel group as compared with cross-over studies. It can be concluded from this study that heroin and hydromorphone are potent and effective analgesics with similar profiles of activity. Although acute single-dose assays, or short-term repeated dose studies, cannot exclude the possibility of the emergence of distinguishing clinical effects after repeated dosing, the data here presented provide supporting evidence for the view that, at equianalgesic doses, either drug can potentially substitute for the other.
13
Acknowledgements This project was supported by National Cancer Institute Grant CA38297. Dr Portenoy’s research is supported by American Cancer Society Grant JFRA-244. We are grateful to Judith Weiss, RN and Lorraine Scimone, RN for their diligence and care in collecting the data for this study.
References Finney, D.J., Statistical Method in Biological Assay, 2nd edn., Hafner, New York, 1964, pp. 98-117. Fishman, B., Pastemak, S., Wallenstein, S.L., Houde, R.W., Holland, J.C. and Foley, K.M., The Memorial Pain Assessment Card: a valid instrument for the evaluation of cancer pain, Cancer, 60 (1987) 1151-1158. Gaddum, J.H., Bioassays and mathematics, Pharmacol. Rev., 5 (1953) 87-134. Houde, R.W., Systemic analgesics and related drugs: opioid analgesics. In: J.J. Bonica and V. Ventafridda (Eds.), Advances in Pain Research and Therapy, Vol. 2, Raven Press, New York, 1979, pp. 263-273. Houde, R.W., Clinical studies of hydromorphone. In: K.M. Foley and C.E. Intunisi (Eds.), Advances in Pain Research and Therapy, Vol. 8, Raven Press, New York, 1986, pp. 129-135. Inturrisi, C.E., Max, M.B., Foley, K.M., Schultz, M., Shin, S.U. and Houde, R.W., The pharmacokinetics of heroin in patients with chronic pain, New Engl. J. Med., 310 (1984) 1213-1217.
7 Kaiko, R.F., Wallenstein, S.L., Rogers, A.G., Grabinski, P.Y. and Houde, R.W., Analgesic and mood effects of heroin and morphine in cancer patients with postoperative pain, New Engl. J. Med., 304 (1981) 1501-1505. 8 Laska, E.M., Sunshine, A., Zighelboim, I., Roure, C., Marrero, I., Wanderling, J. and Olsen, N., Effects of caffeine on acetaminophen analgesia, Clin. Pharmacol. Ther.. 33 (1983) 498-509. 9 Levine, M.N., Sackett, D.L. and Bush, H.. Heroin vs morphine for cancer pain. Arch. Intern. Med., 146 (1986) 353-356. 10 Merck Index, 10th Edition, Merck and Co., Rahway, NJ, 1983. 11 Sawynok, J., The therapeutic use of heroin: a review of the pharmacological literature, Can. J. Physiol. Pharmacol., 64 (1986) l-6. 12 Tursky, B., The development of a pain perception profile: a psychological approach. In: W. Weisenberg and B. Tursky (Eds.), Pain: New Perspectives in Therapy and Research, Plenum Press, New York, 1976, pp. 171-194. 13 Wallenstein, S.L., Measurement of pain and analgesia in cancer patients, Cancer, 53 (1984) 2260-2264. 14 Wallenstein, S.L., Scaling clinical pain and pain relief. In: B. Bromm (Ed.), Pain Measurement in Man. Neurophysiological Correlates of Pain, Elsevier Science Publishers, Amsterdam, 1984, pp. 389-396. 15 Wallenstein, S.L., Heidrich. III, G., Kaiko. R. and Houde, R.W., Clinical evaluation of mild analgesics: the measurement of clinical pain, Br. J. Pharmacol., 10 (1980) 319S3273. 16 Wallenstein, S.L. and Houde, R.W., The clinical evaluation of analgesic effectiveness. In: S. Ehrenpreis and A. Neidle (Eds.), Methods in Opioid Research, Marcel Dekker, New York, 1975, pp. 127-145.
