Meta-analysis of the efficacy of extradural clonidine to relieve postoperative pain: an impossible task

Transcription

1 British Journal of Anaesthesia 1998; 81: Meta-analysis of the efficacy of extradural clonidine to relieve postoperative pain: an impossible task S. ARMAND, A. LANGLADE, A. BOUTROS, K. LOBJOIT, C. MONRIGAL, R. RAMBOATIANA, A. RAUSS AND F. BONNET Summary Clonidine, an alpha 2 adrenoceptor agonist, has anti-hypertensive and anti-nociceptive effects. It is commonly used in association with local anaesthetics and opioids to enhance the quality and duration of extradural analgesia in the postoperative period, and to decrease the incidence of side effects. As a sole analgesic, it has seldom been used to relieve postoperative pain. The dose of extradural clonidine to achieve good pain relief without deleterious side effects remains undetermined. In order to address this problem, we performed a computer search via two well-known databases, Medline and Excerpta Medica, covering the period from 1985 to September One hundred and fifty-nine articles were retrieved of which 38 dealt with extradural clonidine and postoperative pain. All but 16 studies suffered from serious design flaws, such as lack of controls and/or randomization, or inadequate statistical analysis. The data from these studies were difficult to interpret because of the tremendous variation in variables, especially dose of clonidine, level of extradural injection, time of administration, type of anaesthesia, type of surgery, and reference and rescue drugs. The simultaneous extradural use of local anaesthetics and opioids further hindered data interpretation, and precluded any meta-analysis. Proposals for a standard study design are made to help comparison between studies involving extradural clonidine and postoperative pain. (Br. J. Anaesth. 1998; 81: ) Keywords: analgesic techniques, regional; anaesthetic techniques, extradural; analgesics non-opioid, clonidine; analgesia, postoperative; pain, postoperative Clonidine, an alpha 2 adrenoceptor agonist, is a wellknown anti-hypertensive drug. Its antinociceptive properties have been well documented since the early 1970s 1 and widely used since the first report on extradural clonidine in humans by Tamsen and Gordh in Clonidine appeared devoid of some of the side effects of both local anaesthetics and opioids (that is, motor block, tachyphylaxis, neurological toxicity, respiratory depression, pruritus, and urinary retention) although hypotension, bradycardia and sedation are recognized side effects of clonidine. It is probably used mainly in association with local anaesthetics and/or opioids and via various routes of administration (spinal, intravenous, intramuscular, oral, transdermal, peripheral block) 3 11 in an attempt to decrease the incidence of side effects. Clonidine has less frequently been used alone. As a sole analgesic, at least two areas of controversy remain: the doses required to achieve good pain relief and the extent of side effects. The aim of our meta-analysis was to assess the efficacy of extradural clonidine in the relief of postoperative pain, to determine whether anti-nociceptive effects and side effects were dose-related, and to what extent the latter were deleterious. Methods The literature search aimed to be as comprehensive as possible. The period ranging from 1985 to September 1997 was searched via two widely used computer databases, Medline and Excerpta Medica, using the headings regional analgesia, extradural clonidine, postoperative analgesia, and postoperative pain. To avoid the risk of overlooking any relevant article, citations from articles were cross-referenced with those of the assessors bibliography and those of the articles under review. Each study was first reviewed independently by two assessors who filled out a structured form to give a common frame for future discussion (table 1). Subsequently the assessors collectively classified the articles into 1 of 4 grades according to the type of study design. Grade 1 articles dealt with randomized, controlled, double-blind trials with clearly defined objectives and adequate statistical analysis. Randomized trials lacking control or blinding were classified as grade 2; randomized trials yielding questionable results, or non randomized trials were classified as grade 3, and grade 4 articles consisted of non-randomized or non-controlled trials, case reports, editorials, letters to journal editors, abstracts, and retrospective evaluations. Only articles belonging SUZANNE ARMAND*, MD, Service d Anesthésie-Réanimation, Hôpital Cochin, Paris, France. AGNÈS LANGLADE, MD, FRANCIS BONNET, MD, Service d Anesthésie-Réanimation, Hôpital Tenon, Paris, France. ACHRAF BOUTROS, MD, Service d Anesthésie- Réanimation, Centre Hospitalier InterCommunal, Créteil, France. KARIN LOBJOIT, MD, Service d Anesthésie-Réanimation, Hôpital Général, Saint-Germain-en-Laye, France. CHRISTINE MONRIGAL, MD, Service d Anesthésie-Réanimation, Centre Hospitalier, Angers, France. RODIN RAMBOATIANA, MD, Service d Anesthésie-Réanimation, Centre Hospitalier, Blaye, France. ALAIN RAUSS, MD, Société Roussel UCLAF, Romainville, France. Accepted for publication: February 13, *Address for correspondence: Hôpital Cochin, 27 rue du Faubourg-St-Jacques, 75014, Paris, France.

2 Efficacy of extradural clonidine to relieve postoperative pain 127 Table 1 Data collection form for meta-analysis References: title, first author, review, volume, year, pages, and medical institution Main objective of the study Type of study: randomized, single or double-blind, controlled, parallel or cross-matched, cohort, and case report Population: number of patients, demographic characteristics (age, gender, weight, height, and ASA physical status), homogeneity of the population and statistical tests used, type of surgery, and inclusion and exclusion criteria Type of anaesthesia: general, regional or both, type and dose of drugs used, especially opioids, hypnotics, and local anaesthetics Postoperative analgesia: clonidine dose and mode of administration (bolus injection and/or continuous infusion), level of extradural injection (lumbar/thoracic), total dose of clonidine; associated analgesics: type, dose, and mode of administration (intravenous, extradural, subcutaneous); and duration of analgesia Analgesic efficacy assessment: scales, timing, and results Side effects assessment: definition, timing, and severity, with special regard to hypotension, bradycardia, sedation and respiratory depression Type of statistical tests and reasoning, number of patients excluded and lost for follow-up. to the first grade of classification were deemed eligible for meta-analysis. In order to bring together the different studies, comparisons were made on doses of clonidine, mode of administration (bolus injection or continuous infusion), time of administration, type of anaesthesia and surgery, reference drug, assessment of pain intensity and relief (for example, visual analogue scale, verbal scale), and timing of the assessment. Escape analgesics were also noted. Results The search retrieved 159 articles, of which 38 dealt with extradural clonidine, and 16 met the selection criteria The number of patients per group ranged from 20 to 100, with a total of 681 patients. It was not clear in the two studies by Carabine and colleagues whether the patients belonged to two different populations. Clonidine was injected in various ways, bolus injection alone, bolus injection followed by continuous infusion, bolus injection followed by patient-controlled analgesia infusion, 3 bolus injection of a mixed solution followed by a continuous infusion of clonidine, a mixed solution combining clonidine with fentanyl, local anaesthetics, 20 or two or more of these drugs 22 and continuous infusion of a mixture of clonidine and morphine 23 (table 2). Clonidine-bolus doses ranged from 75 g to 800 g, and from 1 g kg 1 to 8 g kg 1, and continuous infusion ranged from 0.3 g kg 1 h 1 to 2 g kg 1 h 1, and from 10 g h 1 to 50 g h 1. Administration took place either intraoperatively, a few minutes after induction of anæsthesia ; at the beginning of the surgical procedure, min before the end of the surgical procedure or postoperatively, on arrival in the recovery room h after surgery or at the patients first complaint of pain. Side effects were defined in various manners or not defined at all (table 3). The main side effects recorded were hypotension, bradycardia and sedation. Authors often noted that hypotension and bradycardia should they occur were either readily treated (most of the time by an association of ephedrine and volume loading) or necessitated no treatment. When sedation was mentioned, it was not always clear how this was measured. For example, some authors reported only that patients had marked sedation, or that excessive sedation was not a problem, 19 or that sedation was similar in all groups. 13 Extradural catheters were inserted either thoracically, or at the high and/or the low lumbar level (table 4). Pain intensity and relief was assessed by visual analogue score alone, or visual analogue score and verbal scale. The visual analogue scores ranged from 3 to 6 points or were not described. Table 4 also displays when visual analogue scoring was used (only at rest, at rest and at cough, at rest and after mobilization, and at rest, at cough and after mobilization). Additional analgesics consisted of i.v. patient-controlled analgesia with morphine, pethidine or sufentanil, s.c, i.m. or i.v. morphine, extradural morphine associated with i.v. pethidine, i.v. paracetamol, i.m. piritramide, i.v. ketoprofen, and continuous extradural bupivacaine-fentanyl mixture (table 4). Clonidine alone at a dose of 3 g kg 1 was found to be as efficacious as placebo in one study after thoracic surgery. 19 In orthopaedic and rectal surgery a lower dose (2 g kg 1 ) was found to be more efficacious than placebo. 12 In the remaining studies, the efficacy of clonidine in combination with other drugs was almost always superior to that of each drug alone (table 3). Discussion Although our literature search might not have been as comprehensive as it should, it is unlikely that relevant articles were overlooked. Main articles are always published in main journals which are covered by the two selected computer databases, Medline and Excerpta Medica. Moreover, the references of all articles reviewed were also checked. The number of articles dealing with our selection criteria was smaller than expected considering the great number of articles dealing with postoperative pain and clonidine Only two dose-response studies were found. This type of study should be able to address some of the unresolved problems concerning the efficacy and safety of clonidine. The most impressive evidence of this review was the lack of at least two study designs enabling direct comparison. We found as many study designs as there were articles. For instance, where bolus doses of clonidine were used, the type of anaesthesia was either general, 13 extradural, 12 or both, 19 and where general anaesthesia alone was used, clonidine was administered by bolus injection, 26 continuous infusion, 23 or both 14 (table 2). Moreover, when considering only one type of surgery (for example, major abdominal surgery, eight studies) (tables 5, 6 and 7), it was clear that direct comparison between variables was not possible. The most serious shortcoming of all published studies dealing with postoperative extradural clonidine was the l ack of a common background. Consequently, meta-analysis could not be applied to the selected trials.

3 128 British Journal of Anaesthesia British Journal of Anaesthesia 1998; 81: Table 2 Main data from 16 randomized double-blind studies dealing with extradural clonidine and postoperative pain. n* total number of patients; PCA patient-controlled analgesia; PCEA patient-controlledextradural analgesia; Cont. infus. continuous infusion; EA extradural anaesthesia; GA general anaesthesia; i.v. intravenous; ED extradural; CL clonidine; F fentanyl; M morphine; B bupivacaine; THR total hip replacement; * 9 patients not responding to the study drug within 60 min postoperatively were excluded; in parentheses, total number of patients left during the postoperative period after 13 patients had been excluded intraoperatively because of inadequate anaesthesia Year Authors n* Dose of clonidine Mode of administration Time of administration Reference drug Duration of study (h) Type of anaesthesia Type of surgery 1988 Gordh 20 3 g kg 1 Bolus injection 1st request Placebo 12 EA GA Thoracotomy 1989 Bonnet 20 2 g kg 1 Bolus injection 1st request Placebo 6 EA and/or GA Orthopaedic and rectal 1990 Motsch g in 24 h Contin. Infus. CL M 1st request Morphine 72 GA Abdominal 1990 Mendez g 10 g h 1 Bolus inj. Cont. Infus. Recovery room Placebo 24 EA C-section 800 g 20 g h Rostaing g CL 100 g F Bolus inj. CL F 1st request Fentanyl 12 EA GA Abdominal aorta 1991 Van Essen g Bolus injection 1 h after surgery Morphine 6 GA Menisectomy 1992 Carabine g Bolus inj. Intraoperativ. Bupivacaine 12 GA THR 1992 Carabine g 25 g h 1 Bolus inj. Cont. infus. Intraoperativ. Morphine 24 GA THR 150 g 50 g h Huntoon g 40 g h 1 Bolus inj. Cont. infus. 1st request Placebo 24 EA C-Section 800 g 40 g h Mogensen g g 1 h 1 B M Bolus inj. Cont. infus. Intraoperativ. B M Saline 12 EA GA Abdominal hysterectomy 1993 De Koch 40 4 g kg 1 2 g kg 1 h 1 Bolus inject. Cont. infus. Intraoperativ. and postoperativ. i.v. CL 36 GA Abdominal 1993 Delaunay g kg 1 h 1 Fentanyl Contin. infus. CL + F Recovery room Fentanyl 72 EA GA Colorectal 1995 Bernard 24 8 g kg 1 30 g/15 mn Bolus injection PCEA 1st request i.v. CL i.v. PCA CL 10 GA Scoliosis repair 1995 De Koch 40 4 g kg 1 2 g kg 1 h 1 Bolus injection Cont. Infus. 1st request ED sufentanil 12 GA Intestinal resection or reanastomosis 1995 Rockemann 45* 8 g kg 1 Bolus injection 1st request Morphine 10 EA GA Pancreatic 4 g kg 1 M 1997 De Koch 50 2 g kg g kg 1 h 1 Bolus injection Cont. infus. Intraoperativ. and postoperativ. (36) 4 g kg 1 1 g kg 1 h 1 8 g kg -1 2 g kg 1 h 1 / 12 GA Abdominal

4 Efficacy of extradural clonidine to relieve postoperative pain 129 Efficacy of extradural clonidine to relieve postoperative pain 127 Table 3 Main side-effects and analgesic variables from 16 randomized double-blind studies dealing with extradural CL and postoperative pain. * Analgesic duration of CL alone, in min, values are mean (SD); analgesic efficacy variously evaluated (VAS score and/or verbal score and/or total consumption of supplemental analgesics or percent reduction in pain); : presence of data; : absence of data; NM not mentioned and/or no definition of the side-effect available; CL clonidine; M morphine; F fentanyl; Suf sufentanil; Pl placebo; B bupivacaine; Chl chloroprocaine; ED extradural; i.v. intravenous; a time from when PCA was connected (30 min after CL injection) to the first self-administered dose; b values are means SEM; c time from CL injection to first use of supplemental morphine in the bupivacaine group; d time from CL injection to first use of supplemental morphine in the chloroprocaine group, e no definition of hypotension; f no definition of bradycardia; g time from tracheal extubation to first analgesic demand in the extradural group; h time from tracheal extubation to first analgesic demand in the intravenous group; i less morphine requirements in the ED CL group, but same VAS scores; j same VAS scores between groups, but PCEA dose less than IV dose of CL; k time from extubation to first analgesic demand; l reduction in heart rate, no definition of bradycardia available Treatment of Year Authors Sedation Hypotension Bradycardia Sedation Hypotension Bradycardia Duration of analgesia* Analgesic efficacy of clonidine 1988 Gordh 11.3 (12.55) a CL Pl 1989 Bonnet Drowsiness 210 (87) CL Pl 1990 Motsch CL M M NM CL M M 1990 Mendez dose-dependent 270 (400 g CL) CL800 g 400 g Pl 300 (800 g CL) 1991 Rostaing NM NM 543 (183) b CL F F 1991 Van Essen NM NM NM NM CL M M CL 1992 Carabine 179 (17) b CL CL B B CL B CL B 1992 Carabine 144 (99) 50 g 25 g CL M CL M 286 (112) 1992 Huntoon dose-dependent 210 (400 g CL) c B 800 g CL B g CL 330 (800 g CL) c >Chl 800 or 400 g CL 30 (400 g CL) d 48 (800 g CL) d 1992 Mogensen NM NM NM NM NM CL B M B M 1993 De Koch e f 103 (115) g ED i.v. CL i 83 (106) h 1993 Delaunay F CL CL NM F CL F 1995 Bernard NM ED i.v. CL j 1995 De Koch 321 (332) k ED CL ED Suf (6 first h), ED CL ED Suf (beyond 6th h) 1995 Rockemann NM l NM NM 33 6 (118) (CL) M CL M CL 585 (217) (CL M) 1997 De Koch not dose-dependent 30 (21) (low-dose) 251 (237) (medium-dose) High medium low-dose CL 369 (256) (high-dose)

5 130 British Journal of Anaesthesia British Journal of Anaesthesia 1998; 81: Table 4 Main characteristics of study design of 16 randomized double-blind studies dealing with extradural clonidine and postoperative pain. LA* local anaesthetic, heart rate (bpm); ED extradural; EA extradural anaesthesia; GA general anaesthesia; NM not mentioned; # extradural catheter was inserted 1 h after surgery; i.v. PCA intravenous patient-controlled analgesia; i.m. intramuscular; cont. inf. continuous infusion; VS verbal scale: VAS visual analogue scale; Epineph epinephrine; B bupivacaine; Priloc = prilocaine; F fentanyl; M morphine; MAP mean arterial pressure; SAP systolic arterial pressure; a it is not clear whether pain was evaluated at rest, motion or during cough; b no patients received opioids during GA; c lidocaine concentration not mentioned; d study results are in mean (SEM); e study results are in mean (SD); f study results are in median values and ranges Year Authors Level of ED catheter Test dose for EA: LA*/volume Local anaesthetic used for EA:Type/volume Maintenance of GA (drugs) Sedation scale (degrees) Definition of Evaluation of pain: how/ when Hypotension Bradycardia Rescue drug 1988 Gordh T4 2% priloc epineph /4 ml 2% prilocaine /3 6 ml F enflurane 6 Pethid consumption /NM a 100 mm Hg(?) NM i.v. PCA pethidine 1989 Bonnet Lumbar NM 2% lidocaine NM NM b NM VAS formula NM NM NM 5 mg s.c./m 1990 Motsch L1-2 or L % plain B/4 ml 0 F isoflurane 4 Analgesic consumption VAS/NM! 30% MAP or SAP NM ED 2 mg M, then i.v. pethid 1990 Mendez L3-4 or L4-5 2% lidocaine/2 5 ml 0.5% B/<50 ml points VS/NM! 30% AP 50 i.v. PCA M 50 i.v. paracetamol 1991 Rostaing T12-L1 or L1-2 NM 0.5% B/NM F (for intubation) then isoflurane 1991 Van Essen L2-3 # /2 ml Lidoc c Epineph 0 2% halothane, no opioids 1992 Carabine d L1-2 or L % plain lidocaine/3 ml 1992 Carabine e L1-2 or L % plain lidocaine/3 ml 1992 Huntoon L3-4 or L4-5 2% lidoc/2 5 ml or 3% 2-Chl/2 5 ml % isoflurane, no opioids % isoflurane, no opioids 3% 2-Chl or 0.5% B/ 50 ml 1992 Mogensen f T % plain B/3 ml 0.5% B 5 ml + 1 mg M, cont. inf. (0.125%B M) 1993 De Koch L Propofol infusion alfentanil 1993 Delaunay Lower thoracic NM 2% lidoc % B as required 1995 Bernard Caudal route, tip at lumbar level NM VAS formula/nm! 25% MAP or SAP 90 mm Hg NM VAS/NM NM NM i.m. piritramide 3 VAS/NM SAP 85 mm Hg and! MAP (?) 3 4-point VS VAS total i.v. M/NM SAP 85 mm Hg and! MAP (?) mg h 1 ED M i.v. M i.m. M 45 i.v. PCA M point VS/NM! 30% MAP 50 i.v. PCA M Isoflurane, no opioids NM 5-point VS VAS/rest, cough, mobilization 4 4-point VS VAS rest, mobilization SAP 80 mm Hg NM i.m. M NM NM i.v. PCA M Isoflurane 4 VAS SAP 90 mm Hg 45 i.v. paracetamol 0 (radio control) 0 Isoflurane 5 VAS/NM MAP 55 mm Hg 45 i.v. ketoprofen 1995 De Koch L1-2 NM 0 Propofol infusion sufentanil 3 VAS/rest, mobilization! 30% AP NM i.v. PCA Suf Rockemann T7 10 day 1% mepivacaine/5 ml 0.25% B/10 15 ml Enflurane 3 VAS/rest NM NM ED 0.25% B F before surgery 1997 De Koch T /2 ml Propofol infusion bolus propofol i.v. bolus lidocaine 1 mg kg 1 opioids 4 VAS/rest, cough! 30% SAP 40 ED 0.125% B

6 Efficacy of extradural clonidine to relieve postoperative pain 131 Efficacy of extradural clonidine to relieve postoperative pain 127 Table 5 Some variables from eight randomized double-blind studies dealing with extradural clonidine to relieve postoperative pain after major abdominal surgery. n* total number of patients; CL clonidine; F fentanyl; M morphine; B bupivacaine; Contin. infus. continuous infusion; ED extradural; EA extradural anaesthesia; GA general anaesthesia; i.v. intravenous; Intraoperativ.: clonidine administered at the beginning of the procedure; * nine patients not responding to the study drug within 60 min postoperatively were excluded; in parentheses, the total number of patients left during the postoperative period after 13 patients had been excluded intraoperatively because of inadequate anaesthesia Year Authors n* Dose of clonidine Mode of administration Time of administration Reference drug Duration of study (h) Type of anaesthesia Type of surgery 1990 Motsch g / 24 h Contin. infus. CL M 1st request Morphine 72 GA Abdominal 1991 Rostaing g CL 100 g F Bolus inj. CL F 1st request Fentanyl 12 EA GA Abdominal aorta 1992 Mogensen g g 1 h 1 B M Bolus inj. Contin. infus. Intraoperativ. B M Saline 12 EA GA Abdominal hysterectomy 1993 De Koch 40 4 g kg 1 2 g kg 1 h 1 Bolus inject. Contin. infus. Intraoperativ. and postoperativ. i.v. CL 36 GA Abdominal 1993 Delaunay g kg 1 h 1 Fentanyl Contin. infus. CL F Recovery room Fentanyl 72 EA GA Colorectal 1995 De Koch 40 4 g kg 1 2 g kg 1 h 1 Bolus injection Contin. infus. 1st request ED sufentanil 12 GA Intestinal resection or reanastomosis 1995 Rockemann 45* 8 g kg 1 Bolus injection 1st request Morphine 10 EA GA Pancreatic 4 g kg 1 M 1997 De Koch 50 2 g kg g kg 1 h 1 Bolus injection Contin. infus. Intraoperativ. and postoperativ. / 12 GA Abdominal (36) 4 g kg 1 1 g kg 1 h 1 8 g kg 1 2 g kg 1 h 1 Table 6 Some variables from eight randomized double-blind studies dealing with extradural clonidine to relieve postoperative pain after major abdominal surgery. * Values are in minutes and mean (SD); =analgesic efficacy variously evaluated (VAS score and/or verbal score and/or total consumption of supplemental analgesics or percent reduction in pain); : presence of data; : absence of data; NM not mentioned and/or no definition of side effect available; CL clonidine; B bupivacaine; M morphine; F fentanyl; Suf sufentanil; Pl placebo; ED extradural; i.v. intravenous; a values are means (SEM); b no definition available; c time from tracheal extubation to first analgesic demand in the extradural group; d time from tracheal extubation to first analgesic demand in the i.v. group; e less morphine requirements in the ED CL group, but same VAS scores; f time from extubation to first analgesic demand; g reduction in heart rate, no definition of bradycardia available Treatment of Year Authors Sedation Hypotension Bradycardia Sedation Hypotension Bradycardia Duration of analgesia* Analgesic efficacy of clonidine 1990 Motsch CL M M NM CL M M 1991 Rostaing NM NM 543 (183) a CL F F 1992 Mogensen NM NM NM NM NM CL B M B M 1993 De Koch b b 103 (115) c ED IV Cl e 83 (106) d 1993 Delaunay F CL CL NM F CL F 1995 De Koch 321 (332) f ED CL ED Suf (6 first h), ED CL ED Suf (beyond 6th h) 1995 Rockemann NM g NM NM 336 (118) (CL) M CL M CL 585 (217) (CL M) 1997 De Koch not dose-dependent 30 (21) (low-dose) 251 (237) (medium-dose) High medium low-dose CL 369 (256) (high-dose)

7 132 British Journal of Anaesthesia 128 British Journal of Anaesthesia Table 7 Some variables from eight randomized double-blind studies dealing with extradural clonidine to relieve postoperative pain after major abdominal surgery. LA* local anaesthetic; heart rate (bpm); NM not mentioned; ED extradural; EA extradural anaesthesia; GA general anaesthesia; B bupivacaine; F fentanyl; M morphine; Suf sufentanil; Pethid. pethidine; i.v. PCA intravenous patient-controlled analgesia; i.m. intramuscular; Cont. inf. continuous infusion; VS verbal scale; VAS visual analogue scale; MAP mean arterial pressure; SAP systolic arterial pressure Year Authors Level of ED Catheter Test dose for EA: LA*/ volume Local anaesthetic used for EA: Type/volume Maintenance of GA (drugs) Sedation scale (degrees) Definition of Evaluation of pain how/when Hypotension Bradycardia Rescue drug 1990 Motsch L1-2 or L % plain B / 4 ml 1991 Rostaing T12-L1 or L1-2 NM 0.5% B/NM F (for intubation) then isoflurane 1992 Mogensen T % plain B / 3 ml 0 F isoflurane 4 Analgesic consumption VAS/NM 0.5% B 5 ml 1 mg M, cont. infus. of 0.125% B with M 1993 De Koch L Propofol infusion alfentanil 1993 Delaunay Lower thoracic NM 2% lidoc. 0.5% B/ as required 1995 De Koch L1-2 NM 0 Propofol infusion ± sufentanil 1995 Rockemann T7 10 day before surgery 1% mepivacaine / 5 ml 1997 De Koch T8-9/2 ml 0 0 Propofol infusion ± bolus propofol ± i.v. bolus lidoc. 1 mg kg 1 ± opioids! 30% MAP or SAP NM VAS formula/nm! 25% MAP or SAP 90 mm Hg Isoflurane, no opioids NM 5-point VS VAS/rest, cough, mobilization 4 4-point VS VAS/rest, mobilization NM ED 2 mg M, then i.v. pethid 50 i.v. paracetamol SAP 80 mm Hg NM i.m. M NM NM i.v. PCA M Isoflurane 4 VAS/NM SAP 90 mm Hg 45 i.v. paracetamol 3 VAS/rest, mobilization! 30% AP NM i.v. PCA Suf. 0.25% B/10 15 ml Enflurane 3 VAS/rest NM NM ED 0.25% B F 4 VAS/rest, cough! 30% SAP 40 ED 0.125% B

8 Efficacy of extradural clonidine to relieve postoperative pain 133 Meta-analysis has increased in popularity during the past 10 years, along with literature reviews. Unlike the latter, which is a qualitative analysis, the former is a quantitative analysis of data from multiple clinical trials, a systematic reviewing strategy for addressing research questions that is specially useful when results from several studies disagree with regard to magnitude or direction of effect. 31 Controversy exists regarding the way to perform meta-analysis. Some assume that all available data should be taken into account, including unpublished data, abstracts, case reports, and personal data, because published data may favour positive results. Others would select only well-designed randomized controlled trials However, the need for well-designed controlled trials to assess the analgesic efficacy of clonidine has been emphasized in a recent editorial. 34 The author stated that clonidine alone was ineffective and should not be used as a sole analgesic agent for postoperative or chronic pain, and as the occurrence of side effects was unavoidable and dose-related, high doses of clonidine should not be used. Surprisingly, the author backed up this statement with data from two studies involving a small number of patients (n 10) and small doses of clonidine (150 g) Moreover, cross-over studies, such as these, may have many shortcomings, and are not considered as a gold standard for clinical trials. There are well-designed studies from which the author could have assumed, for example,that a dose of clonidine of 150 g or less is constantly ineffective whatever the type of surgery, or if there is some evidence that so small a dose can be effective it may be because of in part the concomitant administration of other analgesic drugs such as local anaesthetics and/or opioids, 26 or to other variables such as the probable effect of benzodiazepines used for premedication. Premedication also varied greatly from one study to another. The drugs used were diazepam 10 mg orally 1 h before the procedure 19 ; diazepam 0.2 mg kg 1 with time and route of administration unknown 12 ; oral flunitrazepam 1 mg time unknown 12 ; 0.75 mg kg 1 meperidine and 0.5 mg i.m. atropine 45 min before 23 ; 0.5 mg i.m. atropine 30 min before 26 ; temazepam 20 mg min before ; lormetazepam 2 mg sublingually the night and 1 h before the procedure ; flunitrazepam 1 mg or hydroxyzine 100 mg orally 90 min before, and no premedication. Researchers should be encouraged to perform randomized dose-response placebo-control studies, involving many patients in order to increase the power. Authors should take into account previous published analgesic regimens, avoid references such as abstracts, and adopt rigorous exclusion criteria (no pretreatment involving clonidine or any other antihypertensive drug, benzodiazepines, tricyclic antidepressants, or non-steroidal drugs), as well as rigorous inclusion criteria (patients aged years, ASA I or II, probably more than patients per group) and a strict study design. In conclusion, despite the frequent use of clonidine perioperatively, the dose requirements to provide effective analgesia after a surgical procedure remains a matter of personal choice. A meta-analysis of the numerous articles dealing with extradural clonidine and postoperative pain could not be carried out because of the tremendous variety of the study designs. The need for well-designed randomized dose-response controlled studies based on previous studies is clear, so that an optimal extradural clonidine dose regimen can be determined for postoperative pain relief. Side effects should be strictly defined, and their occurrence clearly reported. In addition, patient satisfaction should be evaluated. Acknowledgements This study was supported and funded by La Société Francophone d Etude de la Douleur (SOFRED), Hôpital Lariboisiere, Paris, France. References 1. Paalzow L. Analgesia produced by clonidine in mice and rats. The Journal of Pharmacy and Pharmacology 1974; 27: Tamsen A, Gordh T. Epidural clonidine produces analgesia. Lancet 1984; ii: Bernard JM, Kick O, Bonnet F. Comparison of intravenous and epidural clonidine for postoperative patient-controlled analgesia. Anesthesia Analgesia 1995; 81: Bonnet F, Boico O, Rostaing S, Loriferne JF, Saada M. Clonidine-induced analgesia in postoperative patients: epidural versus intramuscular administration. Anesthesiology 1990; 72: Eisenach JC, DuPen S, Dubois M, Miguel R, Allin D, The Epidural Clonidine Study Group. Epidural clonidine analgesia for intractable cancer pain. Pain 1995; 61: Fogarty DJ, Carabine UA, Milligan KR. Comparison of the analgesic effects of intrathecal clonidine and intrathecal morphine after spinal anaesthesia in patients undergoing total hip replacement. British Journal of Anaesthesia 1993; 71: Gaumann D, Forster A, Griessen M, Habre W, Poinsot O, Della Santa D. Comparison between clonidine and epinephrine admixture to lidocaine in brachial plexus block. Anesthesia Analgesia 1992; 75: Racle JP, Benkhadra A, Poy JY. Prolongation of isobaric bupivacaine spinal anesthesia with epinephrine and clonidine for hip surgery in the elderly. Anesthesia Analgesia 1987; 66: Segal IS, Jarvis DJ, Duncan SR, White PF, Maze M. Clinical efficacy of oral-transdermal clonidine combinations during the perioperative period. Anesthesiology 1991; 74: Siddall PJ, Gray M, Rutkowski S, Cousins MJ. Intrathecal morphine and clonidine in the management of spinal cord injury pain: a case report. Pain 1994; 59: Striebel WH, Kœnigs DI, Krämer JA. Intravenous clonidine fails to reduce postoperative meperidine requirements. The Journal of Clinical Anesthesia 1993; 5: Bonnet F, Boico O, Rostaing S, Saada M, Loriferne JF, Touboul C, Abhay K, Ghignone M. Postoperative analgesia with extradural clonidine. British Journal of Anaesthesia 1989; 63: Carabine UA, Milligan KR, Moore J. Extradural clonidine and bupivacaine for postoperative analgesia. British Journal of Anaesthesia 1992; 68: Carabine UA, Milligan KR, Mulholland D, Moore J. Extradural clonidine infusions for analgesia after total hip replacement. British Journal of Anaesthesia 1992; 68: De Koch M, Crochet B, Morimont C, Scholtes J-L. Intravenous or epidural clonidine for intra- and postoperative analgesia. Anesthesiology 1993; 79: De Koch M, Famenne F, Deckers G, Scholtes J-L. Epidural clonidine or sufentanil for intraoperative and postoperative analgesia. Anesthesia Analgesia 1995; 81: De Koch M, Wiederkher P, Laghmiche A, Scholtes J-L. Epidural clonidine used as the sole analgesic agent during and after abdominal surgery. Anesthesiology 1997; 86: Delaunay L, Leppert C, Dechaubry V, Levron JC, Liu N, Bonnet F. Epidural clonidine decreases postoperative requirements for epidural fentanyl. Regional Anesthesia 1993; 18: Gordh T Jr. Epidural clonidine for treatment of postoperative pain after thoracotomy. A double-blind placebo-controlled study. Acta Anaesthesiologica Scandinavica 1988; 32:

9 134 British Journal of Anaesthesia 20. Huntoon M, Eisenach JC, Boese P. Epidural clonidine after cesarean section. Appropriate dose and effect of prior local anesthetic. Anesthesiology 1992; 76: Mendez R, Eisenach JC, Kashtan K. Epidural clonidine analgesia after cesarean section. Anesthesiology 1990; 73: Mogensen T, Eliasen K, Ejlersen E, Vegger P, Nielsen IK, Kehlet H. Epidural clonidine enhances postoperative analgesia from a combined low-dose epidural bupivacaine and morphine regimen. Anesthesia Analgesia 1992; 75: Motsch J, Gräber E, Ludwig K. Addition of clonidine enhances postoperative analgesia from epidural morphine: a double-blind study. Anesthesiology 1990; 73: Rockemann MG, Seeling W, Brinkmann A, Goertz AW, Hauber N, Junge J, Georgieff M. Analgesic and hemodynamic effects of epidural clonidine, clonidine/morphine, and morphine after pancreatic surgery A double-blind study. Anesthesia Analgesia 1995; 80: Rostaing S, Bonnet F, Levron JC, Vodinh J, Pluskwa F, Saada M. Effect of epidural clonidine on analgesia and pharmacokinetics of epidural fentanyl in postoperative patients. Anesthesiology 1991; 75: Van Essen EJ, Bovill JG, Ploeger EJ. Extradural clonidine does not potentiate analgesia produced by extradural morphine after menisectomy. British Journal of Anaesthesia 1991; 66: Eisenach JC, Lysak SZ,Viscomi CM. Epidural clonidine anal- gesia following surgery: Phase I. Anesthesiology 1989; 71: Chalmers TC, Smith H Jr, Blackburn B et al. A method for assessing the quality of a randomized control trial. Controlled Clinical Trials 1981; 2: Sacks HS, Berrier J, Reitman D, Ancona-Berk VA, Chalmers TC. Meta-analyses of randomized controlled trials. New England Journal of Medicine 1987; 316: Thacker SB. Meta-analysis. A quantitative approach to research integration. Journal of the American Medical Association 1988; 259: L Abbé KA, Detsky AS, O Rourke K. Meta-analysis in clinical research. Annals of Internal Medecine 1987; 107: Gerbarg ZB, Horwitz RI. Resolving conflicting clinical trials: guidelines for meta-analysis. Journal of Clinical Epidemiology 1988; 41: Jenicek M. Meta-analysis in medicine. Where we are and where we want to go. Journal of Clinical Epidemiology 1989; 42: Macdonald R. Extradural clonidine the need for well designed controlled trials. British Journal of Anaesthesia 1994; 72: Carroll D, Jadad A, King V, Wiffen P, Glynn C, McQuay H. Single-dose, randomized, double-blind, double-dummy cross-over comparison of extradural and i.v. clonidine in chronic pain. British Journal of Anaesthesia 1993; 71: