Secondary Logo

Journal Logo

The Effect of Intravenous Ketorolac on Capsaicin-Induced Deep Tissue Hyperalgesia

Kumar, Keeran, MD; Polston, Greg R., MD; Wallace, Mark S., MD

doi: 10.1213/01.ane.0000229716.39977.fa
Analgesia: Research Report

Preclinical and clinical studies have emphasized that persistent small afferent input will induce a state of central facilitation that can be attenuated by systemically administered nonsteroidal antiinflammatory drugs. However, these studies have been performed using cutaneous models of hyperalgesia. In this study we evaluated the effects of IV ketorolac on an experimental model of deep tissue hyperalgesia using IM capsaicin. We used a double-blind, placebo-controlled, crossover design. Ten subjects received 60 mg of ketorolac or placebo in 2 sessions separated by 1 wk. Capsaicin (100 μg in 10 μL) was then injected into the flexor carpi ulnaris muscle of the left forearm. After injection, spontaneous pain scores, pressure pain scores, gripping pain, pain distribution, and pain quality were recorded at 0, 5, 10, 15, 20, and 25 min. Cutaneous allodynia and dysesthesia were then mapped and thermal and mechanical thresholds were measured. The IM injection of capsaicin resulted in a reliable report of pain, hyperalgesia, and referred pain. Ketorolac had no effect on spontaneous pain, elicited pain, pain distribution, or secondary hyperalgesia induced by capsaicin. The findings of this study support the feasibility of further pharmacological studies using the IM capsaicin pain model.

IMPLICATIONS: Human experimental pain models have been used to study drug effects on pain. The IM injection of capsaicin produces a reliable muscular hyperalgesia that is unresponsive to a single dose of IV ketorolac.

From the Department of Anesthesiology, University of California San Diego, San Diego, California.

Accepted for publication May 16, 2006.

Reprints will not be available.

Supported, in part, by the NIH Summer Research Grant Program.

Address correspondence to Mark S. Wallace, MD, 9500 Gilman Drive #0924, La Jolla, CA 92093-0924. Address e-mail to

Musculoskeletal pain is of great clinical significance and chronically affects 15%–20% of the adult United States population (1). However, the functional mechanisms of muscle pain and muscle hyperalgesia remain largely obscure, and treatment options for patients are suboptimal. Nonsteroidal antiinflammatory drugs (NSAIDs) are commonly used analgesic drugs for musculoskeletal pain. The analgesic efficacy of NSAIDs has been clinically demonstrated in numerous cutaneous experimental pain models (2,3). However, the pharmacology of NSAIDs in experimental muscle pain has not been fully addressed. This is the first study to examine the effect of an NSAID on human experimental muscle pain.

It has been demonstrated that muscle pain can be experimentally induced in humans by injecting the algesic agent, capsaicin (4). In microneurographic recordings from muscle nerves in humans, muscle nociceptors with high mechanical thresholds were found that could be activated by IM injections of capsaicin (5). Although the exact mechanisms of muscle hyperalgesia are not fully understood, it is thought to result from both peripheral and central nociceptive mechanisms. It has been shown that substances released from pathologically altered deep tissue increase the sensitivity of nociceptors to various stimuli such as mechanical pressure and heat, whereby a reduction in the receptor threshold enables responses from weaker stimuli (6). This sensitization of muscle nociceptors is the best established peripheral mechanism explaining local tenderness and pain on movement in pathologically altered muscle. Similarly, central mechanisms of muscle pain exist at the spinal and medullar level, such that persistent input from peripheral nociceptors leads to changes in the function of dorsal horn neurons (7–10). Prostaglandins have been shown to cause a hyperexcitability of spinal neurons after chronic afferent input. In patients this central hyperexcitability is speculated to cause more pain in response to a noxious stimulus (hyperalgesia) or pain in response to a non-noxious stimulus (allodynia).

NSAIDs typically have been thought to act by inhibition of peripheral prostaglandin formation. More recently, a central action of NSAIDs has also been suggested via inhibition of prostaglandin-mediated sensitization in dorsal horn pathways, and animal models have shown NSAIDs to be more effective when administered centrally (11,12). IM capsaicin has been shown to produce a reliable report of local muscle pain (peripherally mediated) and referred muscle pain (centrally mediated) and therefore is a logical experimental model to use in testing NSAIDs' analgesic efficacy.

Because the pharmacology of intradermal and topical capsaicin has been extensively evaluated (13–15), IM capsaicin would allow comparisons between the pharmacology of cutaneous and deep tissue experimental pain. These efforts could have 3 potential impacts; 1) it would validate the use of preclinical models to study drug efficacy before human studies are undertaken; 2) it could allow the development of human phase I efficacy trials (diminishing the likelihood that a drug's inactivity is determined only after the time and money spent in a phase II trial); 3) it would provide an important tool for the development of surrogate biological markers for pain (e.g., functional magnetic resonance imaging, positron emission tomography scanning, intrathecal hormones).

The aim of this study was to determine the effect of a single IV dose of ketorolac on the intensity, distribution, and quality of pain in volunteers after a single dose of IM capsaicin.

Back to Top | Article Outline


The study was performed according to the declaration of Helsinki and approved by the institutional human investigation committee. Written informed consent was obtained from all participants. Eleven healthy pain-free subjects participated in the study (10 men and 1 woman; mean age, 24 yr; range, 23–32 yr). The study used a double-blind, placebo-controlled, crossover design in which both investigator and subject were blinded as to the study drug (placebo or ketorolac) and sequence of crossover between the two sessions. Each subject received 60 mg of IV ketorolac or placebo control in two separate sessions separated by 1 wk. Before the delivery of the study drug, the following baseline measurements were performed over the site to be injected: 1) pressure pain, 2) pain with gripping of the left hand, 3) thermal thresholds, and 4) mechanical thresholds. The study drug was then administered over 20 min through a 22-gauge IV cannula placed in a vein of the right arm. After the study drug was administered, a venous blood sample was obtained for ketorolac assay. Capsaicin (100 μg in 10 μL) was then injected into the mid-portion of the flexor carpi ulnaris muscle of the left arm using a 30-gauge needle and microsyringe. After the injection, spontaneous pain scores, pressure pain scores, pain with gripping of the left hand, a McGill Short form, and pain distribution were recorded at time intervals of 0, 5, 10, 15, 20, and 25 min. At 25 min cutaneous allodynia and dysesthesia were mapped and the thermal and mechanical thresholds repeated.

Before testing, the subjects underwent several trials to become familiar with each of the tests.

Spontaneous pain was measured using a visual analog scale (VAS). This consisted of a 100 mm line with “no pain” written at one end and the “worst imaginable pain” written at the other end. The patient was asked to place a mark along the line that corresponded with their pain. The distance, in mm, from the “no pain” end to the location of the mark gives a measurement of the pain.

Pain with movement was measured using a JAMAR hand dynamometer. The JAMAR hand dynamometer measures grip strength by isometric muscle contraction and was performed in the sitting position, with the shoulder adducted, elbow flexed to 90 degrees, and forearm and wrist in neutral positions. The subject was asked to perform a grip strength test of the left hand. Subjects were instructed to grip the device as tight as they could or when pain occurred (whichever came first). The test was performed three times, and the results were averaged to give the final result. Results were recorded as kilograms of force. If pain occurred, pain intensity was measured using the VAS described above.

Pressure pain (deep mechanical stimulation) was measured using a Waggner FPK manual pressure algometer. This is a hand-held device with a 1-cm rubber-tipped plunger mounted to a calibrated spring that determines the minimum force (in grams) required to produce just detectable pain. The device was placed over the injection site and gradual, increasing pressure was self-applied until the subject reported pain. The test was performed three times, and the results were averaged to give the final result. Results are recorded as kilograms of force. If pain occurred, pain intensity was measured using the VAS described above.

Pain quality was assessed using the McGill Pain Questionnaire Short Form. This form consists of 15 descriptors assessing the sensory (11 descriptors) and effective (4 descriptors) component of pain. Each descriptor is rated as none (0), mild (1), moderate (2), or severe (3). Subjects were asked to rate each descriptor as it correlated with their pain. Final scores were calculated for the total (sum of all scores), sensory (sum of the 11 sensory scores), and affective (sum of the 4 affective scores) scores.

Thermal thresholds were measured using a Thermal Sensory Analyzer (Medoc Advanced Medical System, Minneapolis, MN). Tests performed included i) warm and cold sensation; ii) warm and cold pain. The Thermal Sensory Analyzer consists of a thermode measuring 46 × 29 mm. The temperature of the thermode can either increase or decrease (at a rate of 1.2°C/s for warmth and cold; 3°C/s for hot and cold pain) depending on the direction of current flow through the device. The stimulator was applied the left forearm injection site and the reference temperature was set to 32°C. The patient held a switch that was pressed at the first sensation of temperature change (for warmth and cold) or pain (for hot and cold pain). Pressing the switch reverses the temperature change, returning to a neutral temperature of 32°C. All results were recorded in degrees °C.

Mechanical thresholds were measured using Von Frey hairs. Calibrated Semmes-Weinstein monofilaments sizes (3.61, 3.84, 4.08, 4.17, 4.31, 4.56, 4.74, 4.93, 5.07, 5.18). (Von Frey hairs; Stoelting, Wood Dale, IL). The filaments were selected at random and applied perpendicularly to the area of the left forearm injection site. Three successive stimuli were applied for 1.5 s at 5-s intervals per filament applied in an ascending pattern. The subjects were instructed to close their eyes and report if the stimulus was felt. Thresholds were expressed in mN and measured as the middle between the strongest stimulus not felt in 3 trials and the weakest stimulus felt. At the stimulus intensity evoking a report of sensation (sensory threshold) or discomfort (pain threshold), the next stimulus is one unit lower. This stimulus reversal was repeated twice and the average reversal intensity defined as the threshold.

Pain distribution was determined by asking the subject to outline the pain distribution on a homunculus of the body. Areas of pain distribution were measured with a planimeter.

Allodynia was mapped on the patient's body by gently stroking the painful area with a foam brush. The brush was gently stroked from normal skin inward from 8 directions toward the injection site until the patient reported pain.

Hyperalgesia (cutaneous stimulation) was mapped with a 5.18 von Frey filament. The filament was applied outside the area of pain and moved toward the painful area from 8 directions in 1-cm increments until the patient reported pain. The junction between the abnormal sensations and normal sensations identified the lateral edge of the hyperalgesic area. This was mapped onto a schematic homunculus. Once the general distribution of allodynia and hyperalgesia was established, thermal and mechanical testing was performed at the center.

Data are expressed as the mean ± sd. Baseline sensory thresholds and pain scores before drug and placebo infusion were compared using a one-factor repeated-measures analysis of variance to ensure that there were no differences before onset of treatment. Subsequently, data for each sensory threshold measure, pain score, McGill pain score, and pain distribution were analyzed using a two-factor repeated measures analysis of variance with both drug treatment (drug versus placebo) and titrated drug level as within-subjects factors. Follow-up comparisons of placebo to drug conditions at individual titrated drug levels consisted of paired Student's t-tests; such follow-up comparisons were conducted only if a significant main effect of drug, or drug X titrated drug level interaction was observed in the overall repeated-measures analysis of variance analysis. Pain scores post-capsaicin administration were analyzed using two-factor repeated measures analysis of variance with drug treatment (drug versus placebo) and time post-capsaicin as within-subjects factors. Allodynic areas (post-capsaicin) were analyzed by a paired Student's t-test.

Back to Top | Article Outline


Eleven subjects were initially involved in the study. One subject withdrew because of intolerable nausea. The IM injection of capsaicin reliably resulted in a report of spontaneous pain in all subjects as measured on the VAS. The spontaneous pain scores peaked at the time of injection (ketorolac 56.5 ± 21.4; placebo 63 ± 14.9) and remained significantly elevated for 25 min. At the end of 25 min, 6 of 10 subjects reported no spontaneous pain scores under both placebo and ketorolac conditions. Ketorolac produced no significant decrease in spontaneous pain scores as measured on the VAS compared to placebo (Fig. 1).

Figure 1.

Figure 1.

The IM injection of capsaicin reliably resulted in a decrease in pressure pain threshold, defined as the pressure that the subject first experienced as painful (ketorolac 2.0 ± 1.9 kg; placebo 2.6 ± 1.5 kg) and decrease in grip strength required to elicit pain (ketorolac 28.6 ± 15.4 kg; placebo 29.9 ± 13.1 kg) compared with baseline. The decreases in pressure pain threshold and grip pain threshold peaked at 10 min (Fig. 2). The pain scores elicited by pressure and grip peaked at 5 min and remained significantly elevated for 20 min (pressure: ketorolac 59.2 ± 16.9 kg; placebo 56.8 ± 17.3 kg) (grip: ketorolac 59.3 ± 22.2 kg; placebo 49.1 ± 18.5 kg) (Fig. 3).

Figure 2.

Figure 2.

Figure 3.

Figure 3.

Ketorolac produced no significant decrease in pressure pain or grip pain as measured by threshold and pain scores compared to placebo.

The IM injection of capsaicin resulted in the quality of pain correlating with more sensory descriptors than affective descriptors. Ketorolac produced no significant difference in the pain quality as compared to placebo (Fig. 4).

Figure 4.

Figure 4.

The IM injection of capsaicin resulted in a consistent referred pain pattern proximal and distal to the injection site that peaked at 10 min (Fig. 5). All subjects experienced a cutaneous hyperalgesia to Von Frey Hair stimulation (ketorolac 51.4 ± 51.6 cm2; placebo 59.4 ± 50.8 cm2) measured 25 min post-injection. A cutaneous flare was present in 6 subjects measured at 25 min (ketorolac 3.5 cm2; placebo 1.54 cm2).

Figure 5.

Figure 5.

There was no cutaneous allodynia to brush stroke in any subject. The IM injection of capsaicin resulted in no measurable change in thermal (both painful and nonpainful) or mechanical thresholds (nonpainful). There was a significant decrease in mechanical pain to Von Frey hair stimulation after capsaicin injection (Fig. 6). Ketorolac produced no significant difference in any mechanical or thermal threshold compared to placebo.

Figure 6.

Figure 6.

Back to Top | Article Outline


Most human experimental pain models involve the induction of cutaneous pain. More recently, as in this study, experimental pain models have been developed eliciting deep muscle pain that may more closely resemble many clinical pain conditions. Muscle pain differs in several aspects from cutaneous and visceral pain. Subjectively, muscle pain is perceived as aching and cramping. It is often difficult to localize and exhibits referral to other deep somatic tissues such as fascia and joints (16). Objectively, muscle nociceptors process differently in the central nervous system (CNS). For example, muscle pain has a special relay in the mesencephalon, and cortical imaging studies have shown that muscle pain activates different areas of the cortex (17). These differences have clinical relevance and underscore the importance of developing novel experimental models of deep tissue pain.

In the current study, administration of IM capsaicin produced a pain sensation that manifested as changes in 5 measurable categories; 1) spontaneous pain, 2) allodynia (pain evoked by a previously nonpainful stimulus such as brush), 3) hyperalgesia (increased pain or decreased pain threshold evoked by a previously painful stimulus), and 4) neurogenic inflammation (the area of redness and flare extending beyond the site of injection), and 5) referred pain (pain distant and subjectively distinct from pain at the injection site). In this regard, this deep tissue model is similar to previous intradermal studies but there are differences. Intradermal capsaicin, in contrast to IM, has previously demonstrated higher levels of spontaneous pain scores without significant referred pain to deep tissue (4). The exact reason for differences between cutaneous and deep tissue referred pain is unclear but is likely a combination of peripheral and central nociceptive processing. Classic suggestion for mechanisms of referred pain is convergence of cutaneous, muscular, and visceral afferents onto common neurons in the spinal dorsal horn or supraspinally (6,7). Our results suggest that there is this convergence because the IM injection of capsaicin resulted in a cutaneous hyperalgesia and a cutaneous flare response (6 subjects). It is possible that some capsaicin was deposited in the skin; however, we used a 30-gauge needle and any capsaicin at the tip of the needle would be negligible. The higher frequency of referred pain from muscle versus skin may simply be that muscle has a larger proportion of convergent neurons. A peripheral mechanism alone would be unlikely to explain referral patterns because of the distance between local and referred pain exhibited in this study. In addition, at least 3 subjects in this study described an increase in referred pain to the skin at least 15 minutes after capsaicin injection, a phenomenon which has been described previously. This delay in referred pain is consistent with the study by Graven-Nielson et al. (18) and may be explained by unmasking of previously latent converging connections induced by C fiber input from the local pain area.

This IM capsaicin model produced brush-stroke allodynia, hyperalgesia to Von-Frey cutaneous stimulation, and local and distal referred pain, which are consistent with previous models of muscle pain. A suggested mechanism for these phenomena is central spinal facilitation, which is an interaction between afferent fiber input and spinal cord output to the brain. Although both Aδ and C fiber afferents project to dorsal horn cells in the spine, repetitive low frequency C-fiber activation can increase the cellular response in both magnitude and direction. In addition to this conditioning effect on C-fiber input, spinal sensitization also results in a larger receptive field site (i.e., locally referred pain), such that previously inactive afferents elicit a prominent activating response (19). Consistent with the current and previous models, even low threshold tactile stimulation (i.e., Von Frey pressure) could be effective at activating the cell, thereby causing pain. This central mechanism could explain the referred pain and hyperalgesia exhibited in this model. Peripheral sensitization of afferent neurons could also contribute to hyperalgesia, but only in combination with a central facilitation, as the observed distance and size of the referral is simply too large to be explained exclusively by a peripheral process.

Ketorolac was chosen as an analgesic candidate for this study because it has a mechanism of action that has been suggested to attenuate both a peripheral and central facilitated pain state and is one of the most widely used NSAIDs in the inpatient setting (11). NSAIDs and ketorolac at least partially exhibit a mechanism of analgesia through blockade of cyclooxygenase (COX) enzymes that are responsible for the production of eicosanoids and prostaglandins, which then go on to mediate local inflammatory reactions and augment pain response. Our original hypothesis suggested that IV ketorolac might decrease muscle pain and associated hyperalgesia from IM capsaicin by peripheral and central COX suppression, thereby preventing or diminishing a facilitative process of hyperalgesia. This essentially relies on ketorolac being delivered in a sufficient dose with adequate time for equilibration and receptor binding and ketorolac being able to penetrate the CNS in a sufficient quantity to augment a central nociceptive process. Evidence suggests that NSAIDs may possess a central nociceptive action when delivered intrathecally, but previous studies have been limited in demonstrating only peripheral reductions in prostanoid synthesis (20). In contrast, few studies have demonstrated the ability of ketorolac to penetrate the CNS or inhibit prostaglandin formation centrally. In a study by Rice et al. (21), extrapolation from pharmacokinetic data suggested that a central mechanism of analgesia for ketorolac may be unlikely. In this study, after administration of ketorolac to 29 subjects, plasma and cerebrospinal fluid (CSF) concentrations were obtained for up to 277 minutes, and the authors found no correlation between plasma concentration and appearance of ketorolac in the CSF. In addition, the concentrations of ketorolac found in the CSF after a 30-mg dose were insufficient to cause inhibition of microsomal prostaglandin synthesis in vitro. The author's conclusion suggested a peripheral role of COX suppression without central involvement after systemic administration of ketorolac. The lack of effect seen in the current study thus leaves open the possibility that 1) ketorolac was not present in the CNS in sufficient quantity to attenuate a central nociceptive process, 2) insufficient time and/or dose led to inadequate CNS penetration, or 3) prostaglandins and/or eicosanoids have no role in capsaicin-induced muscle pain. In our study, IV ketorolac (60 mg) delivered over 20 minutes before capsaicin injection had no statistical effect on primary or secondary hyperalgesia. Ketorolac was delivered systemically in a dose that has been shown to produce clinical analgesia in randomized controlled trials and in multiple observational settings (22). Given the measured plasma concentrations of ketorolac, insufficient dosing is an unlikely explanation for the lack of an observed effect, although this does not preclude the possibility that higher levels may have resulted in analgesia.

Although there was no observable effect of ketorolac in our study, we believe that IM administration of capsaicin (100 μg/10 μL) results in an effective deep tissue hyperalgesic model of pain in humans that can be used for further analgesic investigation. Its characteristic properties as a valid experimental pain stimulus are: 1) it is relatively noninvasive and produces no lasting tissue damage, 2) it is specific in its production of pain and no other sensations, 3) it is sensitive enough to measure a pain response that is ethically acceptable and physiologically relevant (all but one subject completed the study), 4) the pain stimulus shows a measurable relationship between stimulus and pain intensity in previous studies, and 5) the reported response of IM capsaicin was reproducible. Given these properties of IM capsaicin and its ability to produce lasting hyperalgesia and neurogenic inflammation, IM capsaicin may be a good candidate to assess the impact of additional pharmacologic analgesics on deep tissue hyperalgesia in human volunteers.

Back to Top | Article Outline


1. Joint Commission on Accreditation of Healthcare Organizations. Pain assessment and management: an organizational approach. Oakbrook Terrace, Illinois: Joint Commission on Accreditation of Healthcare Organizations; 2000.
2. Schaffler K, Reitmeir P, Gschanes A, Eggenreich U. Comparison of the analgesic effects of a fixed-dose combination of orphenadrine and diclofenac (Neodolpasse) with its single active ingredients diclofenac and orphenadrine: a placebo-controlled study using laser-induced somatosensory-evoked potentials from capsaicin-induced hyperalgesic human skin. Drugs R D 2005;6:189–99.
3. Burian M, Tegeder I, Seegel M, Geisslinger G. Peripheral and central antihyperalgesic effects of diclofenac in a model of human inflammatory pain. Clin Pharmacol Ther 2003;74:113–20.
4. Whitting N, Svensson P, Gottrup H, et al. Intramuscular and intradermal injection of capsaicin: a comparison of local and referred pain. Pain 2000;84:407–12.
5. Hwang SW, Hawn C, Jiyeon K, Uhtaek O, et al. Direct activation of Capsaicin receptors by products of lipoxygenases: Endogenous capsaicin-like substances. Proc Nat Acad Sci U S A 2000;97:6155–60.
6. Mense S. The Pathophysiologic basis of muscle pain syndromes. Phys Med Rehab Clin N Am 1997;8:23–53.
7. Sorkin LS, Wallace MS. Acute pain mechanisms. Surg Clin North Am 1999;79:213–29.
8. Wallace MS, Laitin S, Licht D, Yaksh TL. Concentration-effect relations for intravenous lidocaine infusions in human volunteers: effect on acute sensory thresholds and capsaicin-evoked hyperpathia. Anesthesiology 1997;86:1262–72.
9. Ando K, Wallace MS, Schulteis G, Braun J. Neurosensory finding after oral mexiletine in healthy volunteers. Reg Anesth Pain Med 2000;25:468–74.
10. Wallace MS, Ridgeway B, Leung A, et al. Concentration-effect relationships for intravenous alfentanil and ketamine infusions in human volunteers: effects upon acute thresholds and capsaicin-evoked hyperpathia. J Clin Pharmacol 2002;42:70–80.
11. Yaksh TL, Dirig DM, Malmberg AB. Mechanisms of action of nonsteroidal antiinflammatory drugs. Cancer Invest 1998;16: 509–27.
12. Shin SW, Sandner-Kiesling A, Eisenach JC. Systemic, but not intrathecal ketorolac is antinociceptive to uterine cervical distension in rats. Pain 2003;105:109–14.
13. Petersen KL, Jones B, Segredo V, et al. Effect of remifentanil on pain and secondary hyperalgesia associated with the heat– capsaicin sensitization model in healthy volunteers. Anesthesiology 2001;94:15–20.
14. Wallace MS, Grubbs D. Effects of oral desipramine on capsaicin induced hyperalgesia. Anesth Analg 2002;95:973–8.
15. Dirks J, Petersen KL, Rowbotham MC, Dahl JB. Gabapentin suppresses cutaneous hyperalgesia following heat-capsaicin sensitization. Anesthesiology 2002;97:102–7.
16. Graven-Nielsen T, Mense S. The peripheral apparatus of muscle pain: evidence from animal and human studies. Clin J Pain 2001;17:2–10.
17. Mendell LM. Physiologic properties of unmyelinated fiber projections to the spinal cord. Exp Neurol 1966;16:316–32.
18. Graven-Nielsen T, McArdle A, Phoenix J, et al. In vivo model of muscle pain: quantification of intramuscular chemical, electrical, and pressure changes associated with saline-induced muscle pain in humans. Pain 1997; 69:137–43.
19. Mendell LM, Wall PD. Responses of single dorsal cord cells to peripheral cutaneous unmyelinated fibers. Nature 1965;206:97–9.
20. Mroszczak E, Jung D, Yee J, et al. Ketorolac tromethamine pharmacokinetics and metabolism after intravenous, intramuscular, and oral administration in humans and animals. Pharmacotherapy 1990;10:33–49.
21. Rice AS, Lloyd J, Bullingham RES, O'Sullivan P. Ketorolac penetration into the cerebrospinal fluid of humans. J Clin Anesthesia 1993;5:459–62.
22. Power I, Noble DW, Douglas E, Spence AA. Comparison of i*m* ketorolac trometamol and morphine sulphate for pain relief after cholecystectomy. Br J Anaesth 1990;65:448–55.
© 2006 International Anesthesia Research Society