| ||||||||||||||
|
|
|||||||||||||

*Consultation de la douleur, Service de pharmacologie clinique, CHU, Clermont-Ferrand Cedex, France; and
Division dAnesthésiologie, Département APSIC, Hôpitaux Universitaires de Genève, Genève, Switzerland
Address correspondence and reprint requests to Dr. M. R. Tramèr, Division of Anaesthesiology, Department APSIC, Geneva University Hospitals, CH-1211 Geneva 14, Switzerland. Address e-mail to martin.tramer{at}hcuge.ch
| Abstract |
|---|
|
|
|---|
Implications: IV lidocaine (0.5 mg/kg) should be given with a rubber tourniquet on the forearm, 30 to 120 s before the injection of propofol; lidocaine will prevent pain in approximately 60% of the patients treated in this manner.
| Introduction |
|---|
|
|
|---|
| Methods |
|---|
|
|
|---|
Both authors independently read each report that could possibly meet the inclusion criteria and scored them for inclusion and methodological validity using the three-item, five-point, Oxford scale (4). We then reached a consensus by discussion. The scale takes into account proper randomization, double-blinding, and reporting of withdrawals and drop-outs. The minimum score of an included randomized controlled trial is one, the maximum score is five.
We noted information about patients (adults or children), size and site of venous cannulation, speed of injection of propofol, and analgesic interventions from each included report. Different scores of pain measurement (for instance, visual analog or verbal rating scales) were used in these trials. Combining these data was impossible. We therefore decided to extract dichotomous data on complete absence of pain. This dichotomous hurdle may be unnecessarily high. An experimental intervention that does not completely prevent pain may alleviate most symptoms. Such an intervention may, of course, be very useful. However, to extract homogeneous data and to minimize the risk of bias caused by different definitions of endpoints, we decided to concentrate on a clearly defined hard endpoint. Complete absence of pain is such an endpoint.
We calculated relative "benefit" as relative risk with 95% confidence interval (CI) (5), using a fixed-effect model to combine data (6). As an estimate of the clinical relevance of the analgesic efficacy, we calculated the number needed to treat (NNT) (7) using the weighted means (weighted by the sample size) of the event rates of active and control interventions. A positive NNT indicated how many patients had to be exposed to an active intervention to prevent pain in one who would have had pain had they all received control (i.e., propofol without the analgesic intervention). A 95% CI around the NNT point estimate was obtained by taking the reciprocals of the values defining the 95% CI for the absolute risk reduction (8). A statistically significant treatment effect was assumed when the 95% CI around the relative benefit excluded 1; the 95% CI around the NNT would then contain positive numbers only.
| Results |
|---|
|
|
|---|
We analyzed data from 56 randomized, controlled trials (6264 patients) (17,23,2578). Average trial size was 111 patients (range, 28 to 368). The median quality score was 2 (range, 1 to 4). Three reports were published as letters (44,63,73); in one a pseudo-randomization method (allocation according to medical record number) was used (27), and in one randomization was unclear (63). These five reports were included in the analysis. Six trials (11%) reported an appropriate method of blinding (identical ampoules, for instance). In 18 trials (33%), there was no attempt for any blinding. Authors of four trials acknowledged support from the manufacturer (35,44,48,56). Three studies were performed in children (31,34,74), two in both adults and children (32,70); all others were in adults only. In all trials, propofol was injected into the upper limb.
Subgroup Analyses
We performed three subgroup analyses to test the propriety of pooling data. This was done with data from control patients only. In seven trials, propofol was injected into an IV catheter "on the forearm" (41,46), in the cephalic vein (45,72,73), or in both a vein on the back of the hand or the forearm (44,49): 261 of 385 controls (68%; range 24% to 80%) reported pain on injection with propofol. In 49 trials, IV catheters were placed exclusively on the back of a hand: 1156 of 1674 controls (69%, range 10% to 100%) reported pain on injection. The difference between these two subgroups was not statistically significant, relative risk 1.02 (95% CI 0.94 to 1.10, P > 0.05). In 43 trials, propofol was injected on the dorsum of a hand, and the diameter of the catheter (median 21-gauge; range 17- to 23-gauge) was reported. In 45 trials, IV catheters were placed on the back of a hand, and the speed of injection of propofol (average 0.6 mL/s; range 0.125 to 2.0 mL/s) was reported. Graphically, there was no evidence of any relationship between the size of the catheter or the speed of injection and the likelihood of pain on injection with propofol (Figure 1). Thus, we pooled efficacy data according to experimental interventions and doses, whenever appropriate.
|
|
|
Opioids. In two trials, some patients received fentanyl or alfentanil immediately before the injection of propofol (32,34). We considered this method a priori as insufficiently analgesic in this context; these data were, therefore, not analyzed further. With alfentanil, fentanyl, or meperidine, given minutes before the injection of propofol, NNTs were between 3 and 4 (Table 1, D). When meperidine 40 mg was given with a tourniquet, the NNT was 1.9. Alfentanil 10 µg/kg, tested in 51 adults (75), had an NNT of 4.3 (95% CI 2.5 to 15); alfentanil 20 µg/kg, tested in 20 children (34), had an NNT of 1.4 (95% CI 1.1 to 2). With fentanyl 100 µg or 150 µg (adults only) (23,28,44,49), the NNT was 4.
Metoclopramide. The NNT of IV metoclopramide 5 or 10 mg, given before the injection of propofol, was 2.7 (Table 1, E). When injected with a tourniquet, the NNT decreased to 2.2.
Temperature
Cooled (i.e., 4°C) propofol and warmed (i.e., 37°C) propofol had no statistically significant analgesic effect (Table 1, E).
Other Interventions
IV thiopentone, lidocaine 60% tape, and nitroglycerine ointment were each tested in two trials. For thiopentone (42,50) and nitroglycerin (61,76), the respective trials produced contradictory results. With lidocaine tape (72,78), pain of both insertion of IV lines and of injection of propofol was decreased. All other pharmacological or physical interventions were each tested in one trial: IV ondansetron, droperidol, nafamostat mesilate, ketamine, aspirin, ketorolac, prilocaine, or morphine; premedication with oral diazepam or IM ketorolac; iontophoresis with lidocaine; dilution of propofol with homologous blood or dextrose; speed of injection of propofol or of carrier; long chain triglycerides; tourniquet; double or single lumen IV sets; and site of injection. No meaningful conclusions could be drawn. In one trial, pretreatment with IV fentanyl was used concomitantly with a propofol-lidocaine mixture (49); this was more analgesic than either treatment alone. In another trial, none of 40 patients receiving pretreatment with fentanyl plus cold propofol mixed with lidocaine reported any pain (17). Reports on adverse effects were sparse. No data on costs were retrieved.
| Discussion |
|---|
|
|
|---|
We generated comparisons between treatment options (Figure 2). Conclusions derived from such a "league table" of relative efficacy must be interpreted cautiously, because confounding variables cannot be excluded. The data suggested that, for best efficacy, it may not be worthwhile, compared with the lidocaine-tourniquet method, to give lidocaine IV before the injection of propofol or to mix it with propofol. Of 100 the patients treated with lidocaine 20 mg, 53 patients (NNT 1.9) will not have any pain when a tourniquet is used, 42 (NNT 2.4) will be pain-free when the lidocaine is mixed with propofol, and only 25 (NNT 4.0) will profit, when the lidocaine is given before propofol (Table 1, AC).
IV opioids showed less efficacy compared with the lidocaine-tourniquet technique. Meperidine looked promising, when given as a Biers block (23,64). This may be regarded as further evidence of meperidines local-anesthetic properties (79). In one trial, the concomitant use of naloxone did not reduce meperidines efficacy, suggesting that its peripheral analgesic effect is not mediated by opioid receptors (23). Fentanyl and alfentanil were also given as Biers blocks (64,77), although with less success compared with meperidine. Systematic review was unable to confirm any relevant peripheral analgesic efficacy with Biers block with opioids other than meperidine (80).
Biers block with metoclopramide decreased propofol-induced pain (Table 1, E). Metoclopramide has been shown to possess weak local anesthetic properties (81); its chemical structure is an analog of procaine.
Early trials reported significant analgesic efficacy when propofol was cooled to 4°C immediately before injection (29,55) or, oppositely, when it was warmed to 37°C (37). These results could not be confirmed in subsequent studies (52,63,65,73). The combined analysis suggested that temperature has no relevant effect on propofol-injection pain. There was a lack of data for all the other analgesic interventions to allow meaningful conclusions.
There were two further interesting findings. First, there was no evidence of any relationship between catheter size and the incidence of pain on injection (Figure 1). Thus, as expected, catheter size per se is of no importance. No relationship could be established between injection pain and size of veins, because original reports did not provide relevant data. There was evidence from two randomized trials that the incidence and severity of pain with propofol can be reduced when the drug is injected into a vein in the antecubital fossa (56,67). It is, however, unlikely that anesthetists will choose the antecubital fossa vein routinely to avoid propofol-injection pain. The second additional finding was unexpected. There is a widely held view that slow injection of propofol may increase the likelihood of pain. This assumption refers to an early publication in which 15 patients had been randomized to slow injection of propofol (67). In these trials, a wide range of injection speeds were tested (i.e., 0.125 to 2 mL/s); there was no evidence of any impact of speed of injection on the incidence of pain (Figure 1).
Several combination therapies were tested. However, it may be overoptimistic to try to further improve the degree of analgesic efficacy as seen with the lidocaine-tourniquet method; incidences of pain were very low (Table 1, C). The best NNT which can be achieved for efficacy is 1. All control patients would have to report pain on injection with propofol, and none who receives the active intervention; this is unlikely with any analgesic intervention. Also, combination therapies may increase cost and the risk of adverse drug reactions, and they may be circumstantial in daily clinical practice.
Some doubt remains concerning the scientific validity of some of these trials. Surely, almost 20 years after the advent of propofol, it is difficult to accept that the injection of this innovative and widely used IV anesthetic still causes pain, and that the mechanisms of that pain are still obscure. The lack of sponsorship from the manufacturer for most of these trials may be a result of a lack of interest. Perhaps as a consequence, a research program was not obvious, although some trials were designed to study peripheral pain mechanisms. More than 6200 patients have been randomized in 56 trials during the past 18 years. According to the instrument of critical appraisal we used (4), most trials were of rather poor quality. Blinding, for instance, was often inadequate, leaving the trials open to the risk of observer bias. Numerous pharmacological treatments, different doses and combinations, alternative methods of administration, and physical interventions were tested, often without a clear biological basis. Propofol has been warmed or cooled, injected faster or more slowly, with or without a tourniquet, diluted or not. Local anesthetics, opioids, nonsteroidal antiinflammatory drugs, ketamine, metoclopramide, droperidol, and other chemical substances have been tested. The lidocaine-tourniquet method is undeniably effective and simple to perform. This begs the question as to the necessity of clinical studies that may identify yet another intervention with some analgesic efficacy to prevent pain on injection with propofol.
In conclusion, for best prevention of pain on injection with propofol, lidocaine 0.5 mg/kg should be given with a rubber tourniquet before the propofol injection; of 100 treated patients, approximately 60 have no pain.
| Acknowledgments |
|---|
We thank Dr. Y. Kobayashi who responded to our enquiry. We thank Mr. Daniel Haake from the documentation service of the Swiss Academy of Medical Science (DOKDI) for his help in searching electronic databases.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
H. J. Kwak, S. K. Min, J. S. Kim, and J. Y. Kim Prevention of propofol-induced pain in children: combination of alfentanil and lidocaine vs alfentanil or lidocaine alone Br. J. Anaesth., June 19, 2009; (2009) aep163v1. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. M. Cornett, J. A. Matta, and G. P. Ahern General Anesthetics Sensitize the Capsaicin Receptor Transient Receptor Potential V1 Mol. Pharmacol., November 1, 2008; 74(5): 1261 - 1268. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Rochette, A. F. Hocquet, C. Dadure, D. Boufroukh, O. Raux, J. F. Lubrano, S. Bringuier, and X. Capdevila Avoiding propofol injection pain in children: a prospective, randomized, double-blinded, placebo-controlled study Br. J. Anaesth., September 1, 2008; 101(3): 390 - 394. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. A. Matta, P. M. Cornett, R. L. Miyares, K. Abe, N. Sahibzada, and G. P. Ahern From the Cover: General anesthetics activate a nociceptive ion channel to enhance pain and inflammation PNAS, June 24, 2008; 105(25): 8784 - 8789. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Yavas, D. Lizdas, N. Gravenstein, and S. Lampotang Interactive Web Simulation for Propofol and Fospropofol, a New Propofol Prodrug Anesth. Analg., March 1, 2008; 106(3): 880 - 883. [Abstract] [Full Text] [PDF] |
||||
![]() |
O. Canbay, N. Celebi, O. Arun, A. H. Karagoz, F. Saricaoglu, and S. Ozgen Efficacy of intravenous acetaminophen and lidocaine on propofol injection pain Br. J. Anaesth., January 1, 2008; 100(1): 95 - 98. [Abstract] [Full Text] [PDF] |
||||
![]() |
J.-R. Lee, C.-W. Jung, and Y.-H. Lee Reduction of pain during induction with target-controlled propofol and remifentanil Br. J. Anaesth., December 1, 2007; 99(6): 876 - 880. [Abstract] [Full Text] [PDF] |
||||
![]() |
Z. Tabboush Ketamine Versus Lidocaine for the Pain of Propofol Injection Anesth. Analg., August 1, 2007; 105(2): 539 - 540. [Full Text] [PDF] |
||||
![]() |
D. Wax Ketamine for Reducing Propofol-Induced Pain Anesth. Analg., August 1, 2007; 105(2): 540 - 540. [Full Text] [PDF] |
||||
![]() |
M. T. Aouad, S. M. Siddik-Sayyid, A. A. Al-Alami, and A. S. Baraka Multimodal Analgesia to Prevent Propofol-Induced Pain: Pretreatment with Remifentanil and Lidocaine Versus Remifentanil or Lidocaine Alone Anesth. Analg., June 1, 2007; 104(6): 1540 - 1544. [Abstract] [Full Text] [PDF] |
||||
![]() |
S.-W. Koo, S.-J. Cho, Y.-K. Kim, K.-D. Ham, and J.-H. Hwang Small-Dose Ketamine Reduces the Pain of Propofol Injection Anesth. Analg., December 1, 2006; 103(6): 1444 - 1447. [Abstract] [Full Text] [PDF] |
||||
![]() |
Y. Nyman, K. Von Hofsten, C. Palm, S. Eksborg, and P. A. Lonnqvist Etomidate-(R)Lipuro is associated with considerably less injection pain in children compared with propofol with added lidocaine Br. J. Anaesth., October 1, 2006; 97(4): 536 - 539. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. K. Dubey and A. Kumar Pain on Injection of Lipid-Free Propofol and Propofol Emulsion Containing Medium-Chain Triglyceride: A Comparative Study Anesth. Analg., October 1, 2005; 101(4): 1060 - 1062. [Abstract] [Full Text] [PDF] |
||||
![]() |
N. C. H. Sun, A. Y. C. Wong, and M. G. Irwin A Comparison of Pain on Intravenous Injection Between Two Preparations of Propofol Anesth. Analg., September 1, 2005; 101(3): 675 - 678. [Abstract] [Full Text] [PDF] |
||||
![]() |
Y. Nyman, K. von Hofsten, A. Georgiadi, S. Eksborg, and P. A. Lonnqvist Propofol injection pain in children: a prospective randomized double-blind trial of a new propofol formulation versus propofol with added lidocaine Br. J. Anaesth., August 1, 2005; 95(2): 222 - 225. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Yamakage, S. Iwasaki, J.-I. Satoh, and A. Namiki Changes in Concentrations of Free Propofol by Modification of the Solution Anesth. Analg., August 1, 2005; 101(2): 385 - 388. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. Schaub, C. Kern, and R. Landau Comparison of Propofol with Lidocaine Pretreatment Versus Propofol Formulated with Long- and Medium-Chain Triglycerides or Confounding Effect of Tourniquet Anesth. Analg., July 1, 2005; 101(1): 301 - 302. [Full Text] [PDF] |
||||
![]() |
W. S. Yew, S. Y. Chong, K. H. Tan, and M. H. Goh The Effects of Intravenous Lidocaine on Pain During Injection of Medium- and Long-Chain Triglyceride Propofol Emulsions Anesth. Analg., June 1, 2005; 100(6): 1693 - 1695. [Abstract] [Full Text] [PDF] |
||||
![]() |
N. Ahmad, C. Y. Choy, E. A. Aris, and S. Balan Preventing the Withdrawal Response Associated with Rocuronium Injection: A Comparison of Fentanyl with Lidocaine Anesth. Analg., April 1, 2005; 100(4): 987 - 990. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. C. Minogue and D. A. Sun Bacteriostatic Saline Containing Benzyl Alcohol Decreases the Pain Associated with the Injection of Propofol Anesth. Analg., March 1, 2005; 100(3): 683 - 686. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. Schaub, C. Kern, and R. Landau Pain on Injection: A Double-Blind Comparison of Propofol with Lidocaine Pretreatment Versus Propofol Formulated with Long- and Medium-Chain Triglycerides Anesth. Analg., December 1, 2004; 99(6): 1699 - 1702. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Adam, J. van Bommel, M. Pelka, M. Dirckx, D. Jonsson, and J. Klein Propofol-Induced Injection Pain: Comparison of a Modified Propofol Emulsion to Standard Propofol with Premixed Lidocaine Anesth. Analg., October 1, 2004; 99(4): 1076 - 1079. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Agarwal, M. Raza, S. Dhiraaj, R. Pandey, D. Gupta, C. K. Pandey, P. K Singh, and U. Singh Pain During Injection of Propofol: The Effect of Prior Administration of Butorphanol Anesth. Analg., July 1, 2004; 99(1): 117 - 119. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Agarwal, M. F. Ansari, D. Gupta, R. Pandey, M. Raza, P. K. Singh, Shiopriye, S. Dhiraj, and U. Singh Pretreatment with Thiopental for Prevention of Pain Associated with Propofol Injection Anesth. Analg., March 1, 2004; 98(3): 683 - 686. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. von Elm, G. Poglia, B. Walder, and M. R. Tramer Different Patterns of Duplicate Publication: An Analysis of Articles Used in Systematic Reviews JAMA, February 25, 2004; 291(8): 974 - 980. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. McCluskey, B. A. Currer, and I. Sayeed The Efficacy of 5% Lidocaine-Prilocaine (EMLA) Cream on Pain During Intravenous Injection of Propofol Anesth. Analg., September 1, 2003; 97(3): 713 - 714. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. W. Park, E.-S. Park, S.-C. Chi, H. Y. Kil, and K.-H. Lee The Effect of Lidocaine on the Globule Size Distribution of Propofol Emulsions Anesth. Analg., September 1, 2003; 97(3): 769 - 771. [Abstract] [Full Text] [PDF] |
||||
![]() |
L.-H. Tan and N.-C. Hwang The Effect of Mixing Lidocaine with Propofol on the Dose of Propofol Required for Induction of Anesthesia Anesth. Analg., August 1, 2003; 97(2): 461 - 464. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Pellegrini, C. Lysakowski, L. Dumont, A. Borgeat, and E. Tassonyi Propofol 1% versus propofol 2% in children undergoing minor ENT surgery Br. J. Anaesth., March 1, 2003; 90(3): 375 - 377. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. A. Cheong, K. S. Kim, and W. J. Choi Ephedrine Reduces the Pain from Propofol Injection Anesth. Analg., November 1, 2002; 95(5): 1293 - 1296. [Abstract] [Full Text] [PDF] |
||||
![]() |
H. Majedi, M. Rabiee, Z. Hussain Khan, and B. Hassannasab A Comparison of Metoclopramide and Lidocaine for Preventing Pain on Injection of Diazepam Anesth. Analg., November 1, 2002; 95(5): 1297 - 1299. [Abstract] [Full Text] [PDF] |
||||
![]() |
B. Walder, M. R. Tramer, and M. Seeck Seizure-like phenomena and propofol: A systematic review Neurology, May 14, 2002; 58(9): 1327 - 1332. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. Basaranoglu, V. Erden, and H. Delatioglu Reduction of Pain on Injection of Propofol: A Comparison of Fentanyl with Remifentanil Anesth. Analg., April 1, 2002; 94(4): 1040 - 1041. [Full Text] [PDF] |
||||
![]() |
P. Kranke, L. H. Eberhart, N. Roewer, and M. R. Tramer Pharmacological Treatment of Postoperative Shivering: A Quantitative Systematic Review of Randomized Controlled Trials Anesth. Analg., February 1, 2002; 94(2): 453 - 460. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|