Anaesthesia, 199 1, Volume 46, pages 541-544
A double-blind study of the speed of onset of analgesia following intramuscular administration of ketorolac tromethamine in comparison to intramuscular morphine and placebo
A. S. C. RICE, J. LLOYD, C. G. MILLER, R. E. BULLINGHAM A N D G. M. O’SULLIVAN
Summary
A double-blind, randomised, parallel group, placebo-controlled study was performed in 85 patients to compare the speed of onset of analgesiafollowing the intramuscular administration of a single dose of 30 mg of ketorolac tromethamine, 10 mg of morphine or placebo. A new, sensitive, method was used to measure the latency of analgesia. The onset of analgesia was defined by the time taken for the pain intensity score to reach a specijiedpercentage of the baseline value. Twenty-five percent of patients achieving a 50% reduction in baseline pain intensity score appears to be the most appropriate parameter to assess the speed of onset of analgesia of ketorolac and morphine in the postoperative setting. Paired comparison demonstrated that ketorolac had a significantly faster onset of analgesia ( p = 0.03) when compared to placebo, whilst comparison of morphine to placebo analgesic latency (p = 0.06) just failed to reach signij?cance. There was no SigniJicant difference between the analgesic onset time of ketorolac and morphine ( p = 0.73). Intramuscular ketorolac and intramuscular morphine have comparable analgesic onset times in the postoperative pain context. However, the sensitive method of measuring onset of analgesia described, highlights the slow onset of analgesia when analgesics of known efficacy are given by the intramuscular route in the postoperative period. More attention should be given to the speed of onset of analgesia in future assessments of analgesics. Key words
Analgesics; morphine. Anti-inflammatory agents; nonsteroidal. Pain; measurement. Ketorolac is a drug which displays potent analgesic, antiinflammatory and antipyretic properties. The efficacy of ketorolac when given by the intramuscular route in postoperative pain of moderate intensity has been demonstrated.’-3 This study aimed to compare the onset time of analgesia after intramuscular administration of equipotent3 doses of ketorolac and morphine. The time of onset of analgesia is generally derived by interpolation of pain scores made at the time of injection and 30 minutes later. The object of this study was to record the pain scores at more frequent intervals and thereby to obtain a more accurate assessment of analgesic latency.
Methods
Local ethics committee approval was obtained before the start of this double-blind, placebo-controlled, randomised study, with three parallel groups. Pregnant or lactating
females were not studied, nor were patients suffering from peptic ulcers or ulcerative colitis. Patients with significant cardiovascular, renal, hepatic, respiratory or nervous system dysfunction were not considered for entry into the study. Known sensitivity to aspirin or other prostaglandin synthetase inhibiting drugs and the presence of a haemorrhagic diathesis also excluded patients from the study. On the day before surgery the study was carefully explained to, and written consent was obtained from, 85 patients scheduled to undergo elective surgery. They were instructed on the use of the pain intensity rating scale and tests of cognitive function which included the ability to correctly name the day of the week and the repeated subtraction of 7 from 99. All patients received an oral premedication of temazepam (20 mg) 2 hours before anaesthesia. Anaesthesia was induced by an intravenous injection of thiopentone (4-5 mg/kg). General anaesthesia was maintained with nitrous oxide, a volatile agent and fentanyl (1.5-2 pg/kg). In the
A.S.C. Rice, FFARCS, Research Fellow, G.M. O’Sullivan, MD, FFARCS, Consultant, Department of Anaesthetics, St. Thomas’ Hospital, Lambeth Palace Road, London SEI 7EH, J. Lloyd, MSc, C.G. Miller, PhD, R.E. Bullingham, MSc, FFARCS, Syntex Research, Maidenhead, Berks. Correspondence to Dr G. OSullivan. Accepted 10 December 1990. 0003-2409/91/070541+ 04 S03.000jO
@ 1991 The Association of Anaesthetists of Gt Britain and Ireland
541
542
A.S.C. Rice el al. 70
r
60 -
50
-
I
Ln
40
-
32.5
5 E
a,
+5
30-
20
-
10 " n-
"
Ketorolac
Morphine
Plocebo
Fig. 1. The time (minutes) after administration of analgesia by which 25% of patients had achieved a 50% reduction in pain intensity. Only 15% of patients in the placebo group achieved a 50% reduction in the pain intensity score in 60 minutes.
cases where nondepolarising muscle relaxants had been used, any residual paralysis at the conclusion of surgery was reversed with neostigmine (0.06 mg/kg) and glycopyrronium (0.0125 mg/kg). Patients experiencing pain of at least moderate severity within 2 hours of the completion of surgery were allowed to continue in the study. The cognitive function tests were performed before the study started. If satisfactory, patients were then randomly allocated to receive intramuscular ketorolac 30 mg, intramuscular morphine 10 mg or intramuscular placebo; each were identical in appearance and administered in a 1-ml volume. Pain intensity scores on a scale of 0-100 (the extremes representing 'no pain' and 'the worst pain you could possibly imagine') were obtained before injection, every 2 minutes for the first 10 minutes following injection, and every 5 minutes for the next 50 minutes. Rescue analgesia was given if required, but not earlier than 30 minutes after the start of the study, unless the patient's state required earlier treatment, in which case the study was discontinued. Vital signs were measured every 15 minutes from the start of the study. Patients were interviewed at the time of the study and 24 hours later. Possible side effects, complaints elicited by indirect questioning and drug usage were recorded. In order to calculate the time of onset of analgesia it was necessary to smooth the pain intensity score/time curve so that an accurate estimate could be made. However, it was first necessary to construct a common pain intensity scale for all subjects. This was performed by converting each score into a percentage of the baseline score for each patient. Each data point was then, in turn, replaced by a weighted mean which was derived by using a three-point moving average technique. This takes three consecutive points and calculates their mean, giving most weight to the central point. The technique also allows for differences in elapsed time between each data point. This method removes the minor alterations in the pain intensity score/ time curves, but not the underlying trends. The formula used for this calculation is shown in appendix 1 .
Ketoroloc
Morphine
Plocebo
Fig. 2. The time (minutes) after administration of analgesia by which 25% (B)and 50% (W) of patients had achieved a 25% reduction in pain intensity. Only 46% of patients in the placebo group achieved a 25% reduction in the pain intensity score in 60
minutes.
The time of onset of analgesia was then set as the time at which the smoothed curve first decreased to 25% and 50% of the baseline pain intensity score. The cumulative percentage of patients reaching these criteria as a function of time then gave a frequency curve for onset of analgesia. The time to onset of analgesia and time to the next analgesic were analysed by survival analysis. The results for each treatment group were summarised using the Kaplan-Meier e ~ t i m a t o r For . ~ analgesic latency and time to next analgesic, groups were compared using the generalised Wilcoxon test (two-tailed with a 5% level of significance).
Results Eighty-five patients participated in the trial, 28 each in the morphine (10 male, 18 female) and ketorolac (7 male, 21 female) groups and 29 (10 male, 19 female) in the placebo group. Three patients, all in the placebo group, failed to perform the pain intensity scoring adequately after operation and were excluded from further analysis; this left 26 patients in this group. The age range was 20 to 77 years with median ages (years) of 40.5 (morphine), 48 (ketorolac) and 43.5 (placebo). The median weights (kg) were 70.3 (morphine), 67.3 (ketorolac) and 72 (placebo), with a range of 48-113.7 kg. The most common operations were abdominal hysterectomy and hernia repair. Together, these accounted for 57% of all operations. The median times for duration of surgery were similar in all groups. There was little difference between the groups in the prestudy pain intensity scores (mean, SD): ketorolac 59.8 (19.55), morphine 57.7 (20.34) and placebo 62.8 (23.99). The time by which 25% of patients achieved a 50% reduction in pain intensity score is shown in Figure 1. In no group did 50% of patients achieve a 50% reduction in pain intensity within the 60 minutes of the study. The times by which 25% and 50% of patients achieved a 25% reduction in pain intensity scores are shown in Figure 2.
Double-blind study of analgesia following administration of ketorolac tromethamine l6
r
2.28
Placebo
Fig. 3. The time (hours) following administration of study analgesic by which the next analgesic drug was given. This is shown for 25% (m), 50% and 75% (El) of patients in each
(a)
study group.
When a criterion of 25% reduction in pain intensity scores was used to define onset of analgesia (Fig. 2) there was no statistically significant difference between any pairs of groups, neither for 25% nor for 50% of patients to reach this level of analgesia (Wilcoxon test). When a criterion of 50% reduction in pain intensity score was used then in none of the groups did 50% of the patients reach the defined level of analgesia within the 60-minute study period. When a criterion of 25% of patients achieving a 50% reduction in pain intensity score was used to measure analgesic latency (Fig. 1) then a paired comparison (Wilcoxon test) to placebo showed a significant reduction for ketorolac (p = 0.03) and a smaller difference for morphine, which just failed to reach significance (p = 0.06). There was no difference when the analgesic latency of morphine and ketorolac were compared (p = 0.73). The time from the initial dose of analgesia to the time at which the next analgesic was administered is shown in Figure 3. Ketorolac was superior to morphine, in this respect. There was no major morbidity in any of the groups. The most frequent adverse event was vomiting, which occurred seven times in the morphine group and six times in each of the other two groups. No patients withdrew from the study because of an adverse event. A total of 29 patients withdrew from the study because of insufficient analgesia: eight in the morphine group, nine in the ketorolac group and 12 in the placebo group. Discussion
The time to onset of analgesia is generally derived from the interpolation of pain intensity differences derived from pain intensity scores taken before treatment and at 15 or 30 minutes after Linear interpolation between points does not allow for accurate assessment of onset times, because the onset of effect is not necessarily linear. Measures of onset time, as usually quoted, do not compare the value for an analgesic with that for placebo, but refer only to decreases from the baseline within a group. Such an onset time cannot then be claimed as drug related, but
543
could just be spontaneous regression of pain. Finally, most studies of analgesics are not primarily concerned with onset times and merely obtain this information as secondary data. Few time points are recorded and there is insufficient data to ensure that outlying data points do not distort the derived onset time. Vignoni et a1.' have taken pain scores at 5-minute intervals following drug administration, but this work was hampered by the use of an insensitive three-point pain scoring system. This study was primarily concerned with the speed of onset of analgesia and therefore a method was designed which could accurately measure this and address the above concerns. The definition of onset time is not straightforward since patients do not have the same pain intensity at the time of analgesic administration and do not all achieve the same degree of reduction in pain from the same dose of an analgesic. Onset of analgesia must therefore specify two parameters to define a time of occurrence; the percentage of patients reaching a satisfactory level of analgesia and the percentage level of pain relief from baseline that a set percentage of patients achieve. If these parameters are set too close to zero then any difference between the groups is masked by the noise of the system. Alternatively, if the criteria are set too high then very few patients will achieve this level and there would be insufficient data to give a meaningful result. The inclusion of a placebo group allowed the validity of the set point parameters to be precisely confirmed, since this determined both the noise of the system and its sensitivity. Frequent measurements using a 100 point scale were used to increase the experimental sensitivity and using this method the speed of onset of analgesia of equipotent doses3 of intramuscular morphine and ketorolac could be compared. Fifty percent pain intensity reduction for 50% of patients appears to be too severe a criterion to measure analgesic onset time in this postoperative pain model. However, it could be claimed that 10 mg of morphine is not a sufficiently efficacious analgesic, but this is probably not a valid criticism since the expected degree of side effects from this analgesic dose were seen. Twenty-five percent of patients reaching a 50% reduction in pain intensity seems an appropriate parameter to measure analgesic onset time in the model studied here. The efficacy of intramuscular ketorolac compared to opioids has been well d~cumented'-~.~.'' and we have confirmed this; intramuscular ketorolac is at least as efficacious as intramuscular morphine and better than intramuscular placebo. The time to next analgesic data, analysed by survival methods, indicate that there is no difference between ketorolac, morphine and placebo for the time by which the first 25% of patients required the next dose of analgesic. Fifty percent of patients receiving ketorolac required the next analgesic sooner than 50% of those patients receiving morphine, and ketorolac was not superior to placebo in these patients. However, when the time to next analgesic data for 75% of patients is examined it appears that 25% of patients who receive ketorolac experience a much longer period (1 3.25 hours) before requiring another analgesic than did those receiving morphine or placebo, although morphine was superior to placebo in this respect, The reasons for this long duration of action of ketorolac could include the nature of the surgery performed and a beneficial effect of the use of nonsteroidal anti-inflammatory analgesics in the early post-
544
A.S.C. Rice et al.
operative period, in a susceptible population. Thus ketorolac may compliment morphine when the two are used in combination in the immediate postoperative period. The side effects experienced by all the groups, including placebo, were similar in frequency and nature, suggesting that none of the drugs shows any particular advantage in this respect. In this study the time to onset of analgesia of intramuscular ketorolac, morphine and placebo has been measured by a sensitive method and the time to onset was similar for ketorolac and morphine, but both ketorolac and morphine had a faster onset than placebo although this difference only reached statistical significance in the case of ketorolac. The literature provides little information regarding the onset of analgesia from an intramuscular administered analgesic in the postoperative period. The sensitive method described here reveals the relatively slow onset of a well established analgesic such as morphine and an equipotent dose of a nonsteroidal anti-inflammatory analgesic (ketorolac). This study demonstrates that not only should future studies examine the quality of analgesia in the early postoperative period, but that onset of analgesia should also be measured by a sensitive method as described here. Acknowledgments
We thank Mr R. Taylor of Syntex Research for his invaluable statistical expertise and the recovery ward nursing staff at St. Thomas’ Hospital for their cooperation.
Time of preceding data point. Time of point to be smoothed (middle). Time of next data point. Preceding pain intensity score. Pain intensity score to be smoothed (middle). Next pain intensity score.
References 1. YEE J, BROWN,ALLISON,C, WILDMAN K. Analgesia from intramuscular ketorolac tromethamine compared to morphine (ms) in ‘severe’ pain following ‘major’ surgery. Clinical Pharmacology and Therapeutics 1985; 37: 239. 2. YEE J, BRADLEYR, STANSKI D, CHERRY C. A comparison of analgesic efficacy of intramuscular ketorolac trornethamine 3. YEEJP, KOSHIVER JE, ALLBON C, BROWNCR. Comparison of intramuscular ketorolac tromethamine and morphine sulfate for analgesia of pain after major surgery. Pharrnacotherapy 1986; 6 253-61. D, In: Analysis of survival data. London: 4. COX DR, OAKES Chapman and Hall, 1984. 5 . TXGERSTEDT I, JANHUNEN L, TAMMISTO T. Efficacy of diclofenac in a single prophylactic dose in postoperative pain. Annals of Clinical Research. 1987; 19: 18-22. 6. MARSALA F, CAVRINI, P, BUFALINO L, GARDINI, F. Treatment of acute pain of ureteral and biliary colic with naproxen sodium administered by the parenteral route. Infernational Journal of Clinical Pharmacology Research 1986; 6 495-500. 7. MCQUAY HJ, CARROLL D, POPPLETON P, SUMMERFIELD RJ, MOORE RA. Fluradoline and aspirin for orthopedic postoperative pain. Clinical Pharmacology and Therapeutics 1987; 41: 531-6. 8. VIGNONIA, FIERRO A, MORESCHINIG, CAUM, AGOSTINO A, DANIELE E, FOTIG. Diclofenac sodium in ureteral colic: a double blind comparison trial with placebo. Journal of International Medical Research 1983; 11: 303-7. 9. O’HARADA, FRACENRJ, KINZERM, PEMBERTON D. Intramuscular ketorolac tromethamine as compared to morphine sulfate in postoperative pain. Anesthesiology 1986; 6 5 A187. 10. FRAGENRJ. A comparison of intramuscularly injected ketorolac and morphine in postoperative pain. Clinical Pharmacology and Therapeutics 1987; 41: 212. and rneperidine in postoperative pain. Clinical Pharmacology and Therapeutics 1986; 3 9 237.
A double-blind study of the speed of onset of analgesia following intramuscular administration of ketorolac tromethamine in comparison to intramuscular morphine and placebo
A. S. C. RICE, J. LLOYD, C. G. MILLER, R. E. BULLINGHAM A N D G. M. O’SULLIVAN
Summary
A double-blind, randomised, parallel group, placebo-controlled study was performed in 85 patients to compare the speed of onset of analgesiafollowing the intramuscular administration of a single dose of 30 mg of ketorolac tromethamine, 10 mg of morphine or placebo. A new, sensitive, method was used to measure the latency of analgesia. The onset of analgesia was defined by the time taken for the pain intensity score to reach a specijiedpercentage of the baseline value. Twenty-five percent of patients achieving a 50% reduction in baseline pain intensity score appears to be the most appropriate parameter to assess the speed of onset of analgesia of ketorolac and morphine in the postoperative setting. Paired comparison demonstrated that ketorolac had a significantly faster onset of analgesia ( p = 0.03) when compared to placebo, whilst comparison of morphine to placebo analgesic latency (p = 0.06) just failed to reach signij?cance. There was no SigniJicant difference between the analgesic onset time of ketorolac and morphine ( p = 0.73). Intramuscular ketorolac and intramuscular morphine have comparable analgesic onset times in the postoperative pain context. However, the sensitive method of measuring onset of analgesia described, highlights the slow onset of analgesia when analgesics of known efficacy are given by the intramuscular route in the postoperative period. More attention should be given to the speed of onset of analgesia in future assessments of analgesics. Key words
Analgesics; morphine. Anti-inflammatory agents; nonsteroidal. Pain; measurement. Ketorolac is a drug which displays potent analgesic, antiinflammatory and antipyretic properties. The efficacy of ketorolac when given by the intramuscular route in postoperative pain of moderate intensity has been demonstrated.’-3 This study aimed to compare the onset time of analgesia after intramuscular administration of equipotent3 doses of ketorolac and morphine. The time of onset of analgesia is generally derived by interpolation of pain scores made at the time of injection and 30 minutes later. The object of this study was to record the pain scores at more frequent intervals and thereby to obtain a more accurate assessment of analgesic latency.
Methods
Local ethics committee approval was obtained before the start of this double-blind, placebo-controlled, randomised study, with three parallel groups. Pregnant or lactating
females were not studied, nor were patients suffering from peptic ulcers or ulcerative colitis. Patients with significant cardiovascular, renal, hepatic, respiratory or nervous system dysfunction were not considered for entry into the study. Known sensitivity to aspirin or other prostaglandin synthetase inhibiting drugs and the presence of a haemorrhagic diathesis also excluded patients from the study. On the day before surgery the study was carefully explained to, and written consent was obtained from, 85 patients scheduled to undergo elective surgery. They were instructed on the use of the pain intensity rating scale and tests of cognitive function which included the ability to correctly name the day of the week and the repeated subtraction of 7 from 99. All patients received an oral premedication of temazepam (20 mg) 2 hours before anaesthesia. Anaesthesia was induced by an intravenous injection of thiopentone (4-5 mg/kg). General anaesthesia was maintained with nitrous oxide, a volatile agent and fentanyl (1.5-2 pg/kg). In the
A.S.C. Rice, FFARCS, Research Fellow, G.M. O’Sullivan, MD, FFARCS, Consultant, Department of Anaesthetics, St. Thomas’ Hospital, Lambeth Palace Road, London SEI 7EH, J. Lloyd, MSc, C.G. Miller, PhD, R.E. Bullingham, MSc, FFARCS, Syntex Research, Maidenhead, Berks. Correspondence to Dr G. OSullivan. Accepted 10 December 1990. 0003-2409/91/070541+ 04 S03.000jO
@ 1991 The Association of Anaesthetists of Gt Britain and Ireland
541
542
A.S.C. Rice el al. 70
r
60 -
50
-
I
Ln
40
-
32.5
5 E
a,
+5
30-
20
-
10 " n-
"
Ketorolac
Morphine
Plocebo
Fig. 1. The time (minutes) after administration of analgesia by which 25% of patients had achieved a 50% reduction in pain intensity. Only 15% of patients in the placebo group achieved a 50% reduction in the pain intensity score in 60 minutes.
cases where nondepolarising muscle relaxants had been used, any residual paralysis at the conclusion of surgery was reversed with neostigmine (0.06 mg/kg) and glycopyrronium (0.0125 mg/kg). Patients experiencing pain of at least moderate severity within 2 hours of the completion of surgery were allowed to continue in the study. The cognitive function tests were performed before the study started. If satisfactory, patients were then randomly allocated to receive intramuscular ketorolac 30 mg, intramuscular morphine 10 mg or intramuscular placebo; each were identical in appearance and administered in a 1-ml volume. Pain intensity scores on a scale of 0-100 (the extremes representing 'no pain' and 'the worst pain you could possibly imagine') were obtained before injection, every 2 minutes for the first 10 minutes following injection, and every 5 minutes for the next 50 minutes. Rescue analgesia was given if required, but not earlier than 30 minutes after the start of the study, unless the patient's state required earlier treatment, in which case the study was discontinued. Vital signs were measured every 15 minutes from the start of the study. Patients were interviewed at the time of the study and 24 hours later. Possible side effects, complaints elicited by indirect questioning and drug usage were recorded. In order to calculate the time of onset of analgesia it was necessary to smooth the pain intensity score/time curve so that an accurate estimate could be made. However, it was first necessary to construct a common pain intensity scale for all subjects. This was performed by converting each score into a percentage of the baseline score for each patient. Each data point was then, in turn, replaced by a weighted mean which was derived by using a three-point moving average technique. This takes three consecutive points and calculates their mean, giving most weight to the central point. The technique also allows for differences in elapsed time between each data point. This method removes the minor alterations in the pain intensity score/ time curves, but not the underlying trends. The formula used for this calculation is shown in appendix 1 .
Ketoroloc
Morphine
Plocebo
Fig. 2. The time (minutes) after administration of analgesia by which 25% (B)and 50% (W) of patients had achieved a 25% reduction in pain intensity. Only 46% of patients in the placebo group achieved a 25% reduction in the pain intensity score in 60
minutes.
The time of onset of analgesia was then set as the time at which the smoothed curve first decreased to 25% and 50% of the baseline pain intensity score. The cumulative percentage of patients reaching these criteria as a function of time then gave a frequency curve for onset of analgesia. The time to onset of analgesia and time to the next analgesic were analysed by survival analysis. The results for each treatment group were summarised using the Kaplan-Meier e ~ t i m a t o r For . ~ analgesic latency and time to next analgesic, groups were compared using the generalised Wilcoxon test (two-tailed with a 5% level of significance).
Results Eighty-five patients participated in the trial, 28 each in the morphine (10 male, 18 female) and ketorolac (7 male, 21 female) groups and 29 (10 male, 19 female) in the placebo group. Three patients, all in the placebo group, failed to perform the pain intensity scoring adequately after operation and were excluded from further analysis; this left 26 patients in this group. The age range was 20 to 77 years with median ages (years) of 40.5 (morphine), 48 (ketorolac) and 43.5 (placebo). The median weights (kg) were 70.3 (morphine), 67.3 (ketorolac) and 72 (placebo), with a range of 48-113.7 kg. The most common operations were abdominal hysterectomy and hernia repair. Together, these accounted for 57% of all operations. The median times for duration of surgery were similar in all groups. There was little difference between the groups in the prestudy pain intensity scores (mean, SD): ketorolac 59.8 (19.55), morphine 57.7 (20.34) and placebo 62.8 (23.99). The time by which 25% of patients achieved a 50% reduction in pain intensity score is shown in Figure 1. In no group did 50% of patients achieve a 50% reduction in pain intensity within the 60 minutes of the study. The times by which 25% and 50% of patients achieved a 25% reduction in pain intensity scores are shown in Figure 2.
Double-blind study of analgesia following administration of ketorolac tromethamine l6
r
2.28
Placebo
Fig. 3. The time (hours) following administration of study analgesic by which the next analgesic drug was given. This is shown for 25% (m), 50% and 75% (El) of patients in each
(a)
study group.
When a criterion of 25% reduction in pain intensity scores was used to define onset of analgesia (Fig. 2) there was no statistically significant difference between any pairs of groups, neither for 25% nor for 50% of patients to reach this level of analgesia (Wilcoxon test). When a criterion of 50% reduction in pain intensity score was used then in none of the groups did 50% of the patients reach the defined level of analgesia within the 60-minute study period. When a criterion of 25% of patients achieving a 50% reduction in pain intensity score was used to measure analgesic latency (Fig. 1) then a paired comparison (Wilcoxon test) to placebo showed a significant reduction for ketorolac (p = 0.03) and a smaller difference for morphine, which just failed to reach significance (p = 0.06). There was no difference when the analgesic latency of morphine and ketorolac were compared (p = 0.73). The time from the initial dose of analgesia to the time at which the next analgesic was administered is shown in Figure 3. Ketorolac was superior to morphine, in this respect. There was no major morbidity in any of the groups. The most frequent adverse event was vomiting, which occurred seven times in the morphine group and six times in each of the other two groups. No patients withdrew from the study because of an adverse event. A total of 29 patients withdrew from the study because of insufficient analgesia: eight in the morphine group, nine in the ketorolac group and 12 in the placebo group. Discussion
The time to onset of analgesia is generally derived from the interpolation of pain intensity differences derived from pain intensity scores taken before treatment and at 15 or 30 minutes after Linear interpolation between points does not allow for accurate assessment of onset times, because the onset of effect is not necessarily linear. Measures of onset time, as usually quoted, do not compare the value for an analgesic with that for placebo, but refer only to decreases from the baseline within a group. Such an onset time cannot then be claimed as drug related, but
543
could just be spontaneous regression of pain. Finally, most studies of analgesics are not primarily concerned with onset times and merely obtain this information as secondary data. Few time points are recorded and there is insufficient data to ensure that outlying data points do not distort the derived onset time. Vignoni et a1.' have taken pain scores at 5-minute intervals following drug administration, but this work was hampered by the use of an insensitive three-point pain scoring system. This study was primarily concerned with the speed of onset of analgesia and therefore a method was designed which could accurately measure this and address the above concerns. The definition of onset time is not straightforward since patients do not have the same pain intensity at the time of analgesic administration and do not all achieve the same degree of reduction in pain from the same dose of an analgesic. Onset of analgesia must therefore specify two parameters to define a time of occurrence; the percentage of patients reaching a satisfactory level of analgesia and the percentage level of pain relief from baseline that a set percentage of patients achieve. If these parameters are set too close to zero then any difference between the groups is masked by the noise of the system. Alternatively, if the criteria are set too high then very few patients will achieve this level and there would be insufficient data to give a meaningful result. The inclusion of a placebo group allowed the validity of the set point parameters to be precisely confirmed, since this determined both the noise of the system and its sensitivity. Frequent measurements using a 100 point scale were used to increase the experimental sensitivity and using this method the speed of onset of analgesia of equipotent doses3 of intramuscular morphine and ketorolac could be compared. Fifty percent pain intensity reduction for 50% of patients appears to be too severe a criterion to measure analgesic onset time in this postoperative pain model. However, it could be claimed that 10 mg of morphine is not a sufficiently efficacious analgesic, but this is probably not a valid criticism since the expected degree of side effects from this analgesic dose were seen. Twenty-five percent of patients reaching a 50% reduction in pain intensity seems an appropriate parameter to measure analgesic onset time in the model studied here. The efficacy of intramuscular ketorolac compared to opioids has been well d~cumented'-~.~.'' and we have confirmed this; intramuscular ketorolac is at least as efficacious as intramuscular morphine and better than intramuscular placebo. The time to next analgesic data, analysed by survival methods, indicate that there is no difference between ketorolac, morphine and placebo for the time by which the first 25% of patients required the next dose of analgesic. Fifty percent of patients receiving ketorolac required the next analgesic sooner than 50% of those patients receiving morphine, and ketorolac was not superior to placebo in these patients. However, when the time to next analgesic data for 75% of patients is examined it appears that 25% of patients who receive ketorolac experience a much longer period (1 3.25 hours) before requiring another analgesic than did those receiving morphine or placebo, although morphine was superior to placebo in this respect, The reasons for this long duration of action of ketorolac could include the nature of the surgery performed and a beneficial effect of the use of nonsteroidal anti-inflammatory analgesics in the early post-
544
A.S.C. Rice et al.
operative period, in a susceptible population. Thus ketorolac may compliment morphine when the two are used in combination in the immediate postoperative period. The side effects experienced by all the groups, including placebo, were similar in frequency and nature, suggesting that none of the drugs shows any particular advantage in this respect. In this study the time to onset of analgesia of intramuscular ketorolac, morphine and placebo has been measured by a sensitive method and the time to onset was similar for ketorolac and morphine, but both ketorolac and morphine had a faster onset than placebo although this difference only reached statistical significance in the case of ketorolac. The literature provides little information regarding the onset of analgesia from an intramuscular administered analgesic in the postoperative period. The sensitive method described here reveals the relatively slow onset of a well established analgesic such as morphine and an equipotent dose of a nonsteroidal anti-inflammatory analgesic (ketorolac). This study demonstrates that not only should future studies examine the quality of analgesia in the early postoperative period, but that onset of analgesia should also be measured by a sensitive method as described here. Acknowledgments
We thank Mr R. Taylor of Syntex Research for his invaluable statistical expertise and the recovery ward nursing staff at St. Thomas’ Hospital for their cooperation.
Time of preceding data point. Time of point to be smoothed (middle). Time of next data point. Preceding pain intensity score. Pain intensity score to be smoothed (middle). Next pain intensity score.
References 1. YEE J, BROWN,ALLISON,C, WILDMAN K. Analgesia from intramuscular ketorolac tromethamine compared to morphine (ms) in ‘severe’ pain following ‘major’ surgery. Clinical Pharmacology and Therapeutics 1985; 37: 239. 2. YEE J, BRADLEYR, STANSKI D, CHERRY C. A comparison of analgesic efficacy of intramuscular ketorolac trornethamine 3. YEEJP, KOSHIVER JE, ALLBON C, BROWNCR. Comparison of intramuscular ketorolac tromethamine and morphine sulfate for analgesia of pain after major surgery. Pharrnacotherapy 1986; 6 253-61. D, In: Analysis of survival data. London: 4. COX DR, OAKES Chapman and Hall, 1984. 5 . TXGERSTEDT I, JANHUNEN L, TAMMISTO T. Efficacy of diclofenac in a single prophylactic dose in postoperative pain. Annals of Clinical Research. 1987; 19: 18-22. 6. MARSALA F, CAVRINI, P, BUFALINO L, GARDINI, F. Treatment of acute pain of ureteral and biliary colic with naproxen sodium administered by the parenteral route. Infernational Journal of Clinical Pharmacology Research 1986; 6 495-500. 7. MCQUAY HJ, CARROLL D, POPPLETON P, SUMMERFIELD RJ, MOORE RA. Fluradoline and aspirin for orthopedic postoperative pain. Clinical Pharmacology and Therapeutics 1987; 41: 531-6. 8. VIGNONIA, FIERRO A, MORESCHINIG, CAUM, AGOSTINO A, DANIELE E, FOTIG. Diclofenac sodium in ureteral colic: a double blind comparison trial with placebo. Journal of International Medical Research 1983; 11: 303-7. 9. O’HARADA, FRACENRJ, KINZERM, PEMBERTON D. Intramuscular ketorolac tromethamine as compared to morphine sulfate in postoperative pain. Anesthesiology 1986; 6 5 A187. 10. FRAGENRJ. A comparison of intramuscularly injected ketorolac and morphine in postoperative pain. Clinical Pharmacology and Therapeutics 1987; 41: 212. and rneperidine in postoperative pain. Clinical Pharmacology and Therapeutics 1986; 3 9 237.
