Adjunctive analgesia with intravenous propacetamol does not reduce morphine-related adverse effects

Transcription

1 British Journal of Anaesthesia 90 (3): 314±19 (2003) DOI: /bja/aeg076 Adjunctive analgesia with intravenous propacetamol does not reduce morphine-related adverse effects F. Aubrun 1 *, F. Kalfon 1, P. Mottet 1, A. Bellanger 2, O. Langeron 1, P. Coriat 1 and B. Riou 3 1 Department of Anaesthesiology and Critical Care, 2 Department of Pharmacy and 3 Department of Emergency Medicine and Surgery, Groupe Hospitalier PitieÂ-SalpeÃtrieÁre, Assistance Publique-HoÃpitaux de Paris (AP-HP), Universite Pierre et Marie Curie, Paris, France *Corresponding author: DeÂpartement d'anestheâsie-reâanimation, CHU PitieÂ-SalpeÃtrieÁre, 47 Boulevard de l'hoãpital, F Paris Cedex 13, France. frederic.aubrun@psl.ap-hop-paris.fr Background. Propacetamol is widely used in the management of postoperative pain. It decreases morphine requirements but its effect on the incidence of morphine-related adverse effects remains unknown. Methods. Patients (550) were randomly assigned to receive propacetamol or a placebo over the rst 24 h after operation in a blinded study. Intravenous morphine titration was performed, after which morphine was administered s.c. every 4 h according to their pain score. Pain was assessed using a visual analogue scale (VAS). The primary end-point was the incidence of morphine-related adverse effects. The main secondary end-points were morphine requirements and VAS score. Results. After morphine titration, the VAS score and the number of patients with pain relief did not differ between groups. Morphine requirements were decreased in the propacetamol group (21 vs 14.5 mg, P<0.001) but the incidence of morphine-related adverse effects did not differ between groups (42 vs 46%, not signi cant). In patients with moderate pain (n=395), morphine requirements decreased by 37% (P<0.001) and the percentage of patients requiring no morphine was greater (21 vs 8%, P=0.002) in the propacetamol group. In patients with severe pain (n=155), morphine requirements decreased by 18% (P=0.04) in the propacetamol group and the number of patients who did not require morphine (3 vs 8%) did not differ signi cantly. Conclusions. Although propacetamol induced a small morphine-sparing effect, it did not change the incidence of morphine-related adverse effects in the postoperative period. Moreover, no bene t could be demonstrated in patients with severe postoperative pain. Br J Anaesth 2003; 90: 314±19 Keywords: analgesics non-opioid, propacetamol; analgesics opioid, morphine; pain, postoperative Accepted for publication: October 17, 2002 Morphine is the most widely used drug to control postoperative pain; it is ef cacious but is associated with frequent adverse effects (nausea, vomiting, urine retention, pruritus and respiratory depression). Consequently, it has become very common to use non-opioid analgesics in association with morphine 1 to obtain a morphine sparingeffect and, hopefully, a decrease in morphine-related adverse effects. This concept is supported by several studies that emphasize the role of multimodal analgesia. 23 Numerous studies have demonstrated a morphine sparingeffect of several non-opioid analgesics, including propacetamol, 4±7 but signi cant decreases in morphinerelated adverse effects have been shown only for nonsteroid anti-in ammatory drugs (NSAIDs). 8 Propacetamol has become very popular in Europe. However, the i.v. form is relatively expensive. Ripouteau and colleagues 9 have recently reported reduced propacetamol costs by an early change from i.v. to oral acetaminophen. However, we considered that administration of propacetamol is not based on suf cient evidence as no study has yet demonstrated that propacetamol signi cantly decreases morphine-related adverse effects. Ó The Board of Management and Trustees of the British Journal of Anaesthesia 2003

2 Propacetamol and postoperative pain Therefore, we designed a prospective, randomized, blinded study of propacetamol in the postoperative period. In this study, in contrast to previous studies, the main endpoint was the incidence of morphine-related adverse effects and not morphine consumption only. Patients and methods The study was approved by an institutional review board (CCPPRB PitieÂ-SalpeÃtrieÁre, Paris, France) and all patients gave written informed consent. Enrolment began in April 1999 and ended in December Patients were eligible if they were aged 18 yr or older and were undergoing surgery in one of three units (each considered as a centre) of the Department of Anaesthesiology of PitieÂ-SalpeÃtrieÁre hospital (orthopaedic surgery, abdominal and general surgery, gynaecological surgery). Criteria for exclusion were as follows: age above 70 yr, ASA status III or IV, known allergy to propacetamol, emergency surgery. Patients with delirium or dementia, or who were not French speaking, were also excluded. Anaesthetists were asked to exclude patients who were expected to have no postoperative pain and those who were expected to have very severe postoperative pain requiring prolonged epidural and/or spinal postoperative analgesia or prolonged postoperative sedation and/or patient-controlled analgesia (PCA). Adjuvant postoperative regional anaesthesia, such as femoral nerve block, was authorized. The use of i.v. NSAIDs in the postoperative period was also authorized but not the use of other non-opioid analgesic drugs. Ketoprofen (50 mg i.v. every 6 h for 2 days; Profenid â ; Laboratoire Aventis, Paris, France) was used because it is the only i.v. NSAID available in France. 8 The use of clonidine was not authorized. During the preoperative visit, performed at least 48 h before surgery, the anaesthetist explained to the patient the protocol and the pain management, including the use of the visual analogue scale (VAS) score for pain. The expected postoperative pain was classi ed as moderate or severe according to the Consensus Conference of the SocieÂte FrancËaise d'anestheâsie et de ReÂanimation. 10 Moderate postoperative pain included pain which was expected to be moderate and to last less than 48 h, whereas severe postoperative pain included pain which was expected to be either severe (lasting less or more than 48 h) or moderate but to last more than 48 h. The study was a randomized, blinded, placebo-controlled trial. Randomization was strati ed according to centres. Patients were studied over the rst 24 h after operation and were assigned randomly to receive a placebo (saline 125 ml) or propacetamol (Pro-Dafalgan â ; Laboratoires UPSA, Rueil-Malmaison, France) in the postoperative period. Each vial was prepared immediately before administration by a nurse who was not involved in the care or pain assessment of the patient. Vials containing 2 g of propacetamol (yielding 1 g of acetaminophen) or saline were administered i.v. over 15 min. The rst administration was performed at the beginning of skin closure in the operating room. The dose regimen was propacetamol 2 g every 6 h. All nurses in the postanaesthesia care unit (PACU) and in the wards had been trained to assess pain using the VAS (0±100 mm, hand-held slide-rule type). 11 When patients had dif culty in manipulating the VAS, nurses were allowed to use a numerical rating scale (0±100 mm). 12 A strict protocol of i.v. morphine titration had been implemented in the PACU previously As soon as the patients were awake, they were questioned about the presence of pain (at least every 15 min) and asked to rate pain intensity. When the pain increased to more than 30 mm, i.v. morphine was given every 5 min in 3 mg increments (2 mg in patients weighing less than 60 kg). Pain relief was de ned as VAS less than 30 mm. Clinical monitoring included arterial blood pressure, heart rate, pulse oximetry (Sp O2 ), respiratory rate (RR) measurements, and sedation according to the Ramsay score. 13 When the patient was asleep (Ramsay score >2) no attempt was made at arousal, and the patient was considered as having adequate pain relief. Morphine titration was stopped if the patient had an RR lower than 10 bpm and/or a Sp O2 lower than 95%, and/or experienced a serious adverse event related to morphine administration (allergy, vomiting, severe pruritus). If there was severe ventilatory depression (RR less than 10 bpm), naloxone (i.v. bolus of 0.04 mg) was administered until RR was greater than 12 bpm. The decision for the patient to leave the PACU was taken by the anaesthetist after the nurse had checked that the patient ful lled the Aldrete criteria 14 and was not suffering from emesis, severe pain or major postoperative bleeding. Patients did not receive antiemetic prophylaxis. Postoperative nausea and vomiting were treated by the anaesthetist using metoclopramide 10±50 mg and ondansetron 4 mg in case of failure of high-dose metoclopramide. After i.v. morphine administration, the patients received morphine administration s.c. every 4 h, beginning as early as 2 h after the end of i.v. morphine titration either in the PACU or on the ward The dose of subcutaneous morphine was determined by a strict protocol based on the VAS at rest and the weight of the patient. Patients with pain relief (VAS less than 30 mm) did not receive morphine. Patients with a VAS between 30 and 60 mm received morphine at either 5 mg (<60 kg body weight) or 7.5 mg (>60 kg body weight); patients with VAS greater than 60 mm received either 7.5 mg (<60 kg body weight) or 10 mg (>60 kg body weight). The primary end-point of this study was the incidence of morphine-related adverse effects. The following adverse effects were noted over the study period: ventilatory depression de ned as a RR below 10 bpm and/or need for naloxone administration, nausea and vomiting, pruritus, retention of urine requiring drainage, bronchospasm and 315

3 Aubrun et al. Table 1 Patient characteristics. Data are mean (SD) [range], median (95% con dence interval) or number (%). No signi cant differences between groups. ² Only in patients receiving general anaesthesia Placebo Propacetamol Age (yr) 45 [18±72] 44 [18±85] Men 116 (42%) 126 (46%) Women 159 (58%) 149 (54%) Weight (kg) 68 (12) 69 (12) Type of surgery Orthopaedic 148 (54%) 145 (53%) Abdominal or pelvic 51 (19%) 54 (20%) Cervical 76 (28%) 76 (27%) Expected postoperative pain Moderate 196 (71%) 199 (72%) Severe 79 (29%) 76 (28%) Duration of surgery (min) 90 (80±90) 90 (85±90) Pre- and/or perioperative urine drainage 20 (7%) 16 (6%) ASA class I 185 (67%) 185 (67%) II 90 (33%) 90 (33%) Type of anaesthesia General 242 (88%) 247 (90%) Regional 32 (12%) 28 (10%) Both 1 (0.4%) 1 (0.4%) Dose of peroperative sufentanil (mg) ² 45 (18) 45 (24) cutaneous rash or other allergic events. Postoperative nausea and vomiting was graded on a three-item scale: minor (nausea only); moderate (vomiting either in the PACU or in the ward); severe (vomiting both in the PACU and in the ward). The secondary end-points were the total dose of morphine, the number of patients who did not receive morphine at all, VAS measurements, the number of patients with pain relief, and the satisfaction of the patient concerning pain management, using a numerical rating scale from 0 (worse) to 100 (best); satisfaction was de ned as a value of at least 80. All data were assessed retrospectively for potential minor protocol violations. These were classi ed as follows: error in a dose of i.v. or s.c. morphine (over- or underestimation) according to the weight and/or pain of the patient, error in administration of a dose of the study drug (cancelled or delayed), and administration of another non-opioid non- NSAID analgesic drug. Statistical analysis We assumed that morphine-related adverse effects would occur in 45% of patients in the placebo group. Using the Casagrande and Pike method, 17 we estimated that at least 265 patients would be needed in each group for the study to be able to detect a 30% difference (i.e. a decrease from 45 to 31%) with 90% certainty (1±b) and a two-sided 5% signi cance level (a). Therefore, we decided to include up to 550 patients, expecting a 5% drop-out rate. An intention-to-treat analysis was performed. Data are expressed as mean (SD) or median and 95% con dence interval (time delay, duration, morphine doses), unless speci ed otherwise. Student's t-test and repeated measures analysis of variance were used to compare means. The Mann±Whitney U-test was used to compare pairs of medians. Fisher's exact test was used for categorical variables. In a post hoc analysis, the subgroups of patients who were expected to experience severe postoperative pain and those expected to experience moderate postoperative pain were analysed separately. All comparisons were twotailed and a P value of less than 0.05 was required to rule out the null hypothesis. Statistical analysis was performed using NCSS 6.0 software (Statistical Solutions, Cork, Ireland). Results A total of 567 patients consented to participate at the preoperative anaesthesia visit and were randomly assigned to one of the two study groups. Seventeen (3%) patients were excluded because they did not undergo surgery (n=4), did not ful l the criteria for inclusion and exclusion (n=6), refused to participate in the study during the postoperative period (n=2), or experienced severe surgical postoperative complications (n=5). Thus, 550 patients were considered for analysis, 275 in the placebo group and 275 in the propacetamol group. The groups were well balanced with regard to patient characteristics, type and duration of surgery, expected postoperative pain, ASA status and type of anaesthesia (Table 1). The VAS measurement was signi cantly lower just before morphine titration in the PACU in the propacetamol group [44 (32) vs 50 (29) mm, P=0.02]. After morphine titration and during the whole study period, the VAS and the number of patients with pain relief did not differ signi cantly between the two groups. Morphine requirements were lower by 31% (21 vs 14.5 mg, P<0.001), the number of patients who did not require morphine at all was signi cantly greater (Table 2), and the percentage of satis ed patients was signi cantly greater (72 vs 61%, P<0.009) in the propacetamol group. However, the number of morphine-related adverse effects was not signi cantly different between the two groups (Table 2). When patients receiving an NSAID were excluded, the number of patients with morphine-related adverse effects was not signi cantly different between the two groups (44 vs 43%). Minor protocol violation occurred in 80 (15%) patients and there were no signi cant differences between the two groups. We performed a subgroup analysis according to the expected postoperative pain (moderate vs severe). Surgery was expected to induce moderate pain in 395 (72%) patients and severe pain in 155 (28%). In patients expected to experience severe pain, the duration of surgery was greater [135 (68) vs 87 (41) min, P<0.001], the perioperative dose of sufentanil was higher [56 (33) vs 41 (13) mg, P<0.001], the initial VAS was greater [57 (29) vs 43 (33) mm, P<0.001], the total postoperative dose of morphine was greater [29 (25.5±31.5) vs 14.5 (13±15) mg, P<0.001] and the percentage of patients requiring no morphine was lower 316

4 Propacetamol and postoperative pain Table 2 Postoperative variables and morphine-related adverse effects according to treatment. Data are mean (SD), median (95% con dence interval) or number (%). *P<0.05 vs placebo group. ² Number of patients with at least one adverse effect. One patient may have suffered several different adverse effects Placebo Propacetamol Dose of morphine (mg) I.V. titration 9 (9±9) 8 (6-9) Subcutaneous 10 (7.5±15) 7.5 (5±7.5)* Total 21 (16.5±24) 14.5 (12±16)* Patients receiving no morphine 17 (6%) 48 (17%)* Delay to titration in PACU (min) 35 (30±40) 35 (30±40) PACU duration (min) 160 (150±180) 150 (145±155) Titration stopped 5 (2%) 6 (2%) Postoperative NSAID 59 (21%) 54 (20%) Postoperative block 2 (1%) 4 (1%) Morphine adverse effects Nausea/vomiting 107 (39%) 94 (34%) Minor 38 (14%) 34 (12%) Moderate 56 (20%) 45 (16%) Severe 13 (5%) 15 (5%) Ventilatory depression 4 (1%) 4 (1%) Pruritus 2 (1%) 3 (1%) Urinary retention 27 (10%) 26 (10%) Allergy 1 (0.4%) 0 Total ² 127 (46%) 115 (42%) NSAID=non-steroidal anti-in ammatory drug. (8 vs 14%, P=0.002) compared with patients expected to experience moderate pain. In patients expected to experience severe pain (79 in the placebo group, 76 in the propacetamol group), morphine requirement was only 18% less in the propacetamol group [25.5 (28±35.5) vs 31 (28±35.5) mg, P=0.04]. The number of patients who did not require morphine (3 vs 8%) and the number of patients with morphine-related adverse effects (46 vs 56%) did not differ signi cantly between groups. Moreover, the VAS score and the number of patients with pain relief did not differ signi cantly between groups (Fig. 1). The percentage of satis ed patients was not signi cantly different between groups (56 vs 44%). In patients expected to experience moderate pain (196 in the placebo group, 199 in the propacetamol group), morphine requirements decreased by 37% in the propacetamol group [10 (8±13) mg vs 16 (15±18) mg, P<0.001]. The number of patients who did not require morphine (21 vs 8%, P<0.001) was signi cantly greater in the propacetamol group, whereas the number of patients with morphinerelated adverse effects (40 vs 42%) did not differ signi cantly between groups. The VAS score and the number of patients with complete pain relief were signi cantly lower in the propacetamol group (Fig. 2). The number of satis ed patients was signi cantly greater in the propacetamol group (76 vs 65%, P=0.03). Discussion In this study, we demonstrated that propacetamol did not decrease the incidence of morphine-related adverse effects Fig 1 Patients with expected severe postoperative pain. Comparison of (A) visual analogue pain scale (VAS; mean and 95% con dence interval) and (B) the percentage of patients with complete pain relief (VAS <30 mm) in the placebo (n=196) and propacetamol (n=199) groups. Pain relief was de ned as VAS <30 mm. VAS is the maximum value recorded just before morphine administration (I.V.=intravenous; S.C.= subcutaneous; every 4 h, from 1 to 6 doses). Ons=onset of morphine titration; End=end of morphine titration; PACU=postanaesthesia care unit; NS=not signi cant (between-group comparisons). during the rst 24 h after operation. Moreover, no bene t could be demonstrated in patients with severe postoperative pain. The protocol of i.v. morphine titration used in this study has been reported to provide rapid and reliable relief of postoperative pain The subcutaneous administration of morphine was authorized as soon as 2 h after morphine titration, as reported previously Subcutaneous morphine could be administered every 4 h and the dose was adjusted to the patient's pain. We did not use PCA, because, in our hospital, PCA is usually reserved for patients whose postoperative pain is expected to be very severe, and such patients were excluded from our study. However, a subcutaneous regimen of morphine is considered as ef cient as PCA, 18 provided the dose is adapted to the patient's pain. In the present study, administration of an NSAID was authorized because this type of drug is used routinely in our 317

5 Aubrun et al. Fig 2 Patients with expected moderate postoperative pain. Comparison of (A) visual analogue pain scale (VAS; mean and 95%con dence interval) and (B) the percentage of patients with pain relief (VAS <30 mm) in the placebo (n=79) and propacetamol (n=76) groups. Pain relief was de ned as VAS <30 mm. VAS is the maximum value recorded just before morphine administration (I.V.=intravenous; S.C.=subcutaneous, every 4 h, from 1 to 6 doses). P<0.001 between-group global comparison; *P<0.05 vs placebo group. Ons=onset of morphine titration; End=end of morphine titration; PACU=postanaesthesia care unit. practice. Since NSAIDs are thought to potentiate the analgesic effects of morphine, 19 they could have interfered with the effects of propacetamol, but the number of patients receiving NSAID was not signi cantly different between groups and exclusion of these patients did not modify our results. However, because the administration of NSAID was not randomized, our study does not enable us to explore the effect of propacetamol and NSAID coadministration. We con rmed that propacetamol induces a morphine sparing-effect (31%). We also observed that the degree of this effect depends on the intensity of postoperative pain: the value was 37% in patients with moderate postoperative pain vs 18% in patients with severe postoperative pain. The morphine-sparing effect of propacetamol has been reported as 46% after total hip replacement, 5 37% after knee ligamentoplasty 4 and 26% after vertebral disc surgery. 20 This morphine-sparing effect strongly suggests that the protocol was appropriately conducted and that it enabled us to observe the expected analgesic effect of propacetamol. However, in our opinion, such an effect cannot be a primary end-point per se, notably because of the low cost of morphine. Propacetamol failed to decrease morphine-related adverse effects. This result may suggest that the multimodal analgesia concept is erroneous, 1 at least when considering the adverse effects of morphine, and that clinical research should focus on the prevention and treatment of these adverse effects (mainly nausea and vomiting) rather than on the decrease of their incidence through a decrease in morphine requirements. A pharmacoeconomic approach of the problem of prevention and treatment of postoperative nausea and vomiting has been recently proposed and might be an appropriate solution. 21 Most of morphine-related adverse effects recorded in the present study were nausea and vomiting. Ventilatory depression occurred in only eight (1%) patients and severe ventilatory depression requiring naloxone administration was not observed. In contrast, urine retention occurred relatively frequently [53 (10%) patients]. Although the power of our study did not enable us to draw conclusions about the potential effects of propacetamol on the incidence of ventilatory depression and even urinary retention, it should be pointed out that no such trend towards an effect was observed. In patients with moderate postoperative pain, propacetamol was not devoid of clinically signi cant effects, although it did not signi cantly modify the incidence of morphinerelated adverse effects. Indeed, propacetamol increased the number of patients who did not require morphine at all, decreased the initial VAS score, improved pain relief (Fig. 2) and increased the patient's satisfaction. In contrast, in patients with severe postoperative pain we failed to demonstrate any bene t from the administration of propacetamol. Because of the large sample size of the present study, we suggest that administration of propacetamol in these patients is not helpful, notably because of its high cost. Our primary end-point was the incidence of morphinerelated adverse effects, and we must recognize that these adverse effects may be at least partly attributable to opioids given intraoperatively and not only to postoperative morphine, and that some of them, mainly nausea and vomiting, may also be because of factors other than opioid administration. 22 This does not alter our conclusion that propacetamol is not able to decrease signi cantly the occurrence of morphine-related adverse effects, but rather provides some explanation why the concept of multimodal analgesia failed in the postoperative period. Furthermore, our study did not include patients with very severe postoperative pain. However, the intensity of pain in these patients frequently required other techniques, including postoperative regional anaesthesia, to control pain. The poor results of propacetamol in our patients with severe pain compared with patients with moderate pain does not support its use in these clinical conditions. Also, the analgesic effect 318

6 Propacetamol and postoperative pain of propacetamol depends on the rate and amount of active drug reaching the central nervous system, and several authors have recently recommended the use of a loading dose and/or a larger dose of propacetamol to achieve therapeutic concentrations more rapidly and more completely However, larger doses may necessitate a caveat in patients with pre-existing renal and hepatic dysfunction 25 and those receiving microsomal cytochrome P450-inducing medications, 26 and it may induce platelet dysfunction. 27 Acknowledgements We thank David Baker (Department of Anaesthesiology, CHU Necker- Enfants Malades, Paris) for reviewing the manuscript. Other investigators who participated in the present study were D. Heitz, G. Daas, M. Dupuy (Department of Anaesthesiology) and E. Juhel (Department of Pharmacy), all at Groupe Hospitalier PitieÂ-SalpeÃtrieÁre, Paris, France. References 1 Dahl V, Raeder JC. Non-opioid postoperative analgesia. Acta Anaesthesiol Scand 2000; 44: 1191±203 2 Michaloliakou C, Chung F, Sharma S. Preoperative multimodal analgesia facilitates recovery after ambulatory laparoscopic cholecystectomy. Anesth Analg 1996; 82: 44±51 3 Brodner G, Pogatzki E, Van Aken H, et al. A multimodal approach to control postoperative pathophysiology and rehabilitation in patients undergoing abdominothoracic esophagectomy. Anesth Analg 1998; 86: 225±7 4 Delbos A, Boccard E. The morphine-sparing effect of propacetamol in orthopedic postoperative pain. J Pain Symptom Manage 1995; 10: 279±86 5 Peduto VA, Ballabio M, Stefanini S. Ef cacy of propacetamol in the treatment of postoperative pain. Morphine-sparing effect in orthopedic surgery. Italian Collaborative Group on Propacetamol. Acta Anaesthesiol Scand 1998; 42: 293±8 6 Cobby TF, Crighton IM, Kyriakides K, Hobbs GJ. Rectal paracetamol has a signi cant morphine-sparing effect after hysterectomy. Br J Anaesth 1999; 83: 253±6 7 Viel E, Langlade A, Osman M, Bilbault P, Eledjam JJ. Le propaceâtamol: des donneâes fondamentales aá l'utilisation clinique. Ann Fr Anesth ReÂanim 1999; 18: 332±40 8 Basto ER, Waintrop C, Mourey FD, Landru JP, Eurin BG, Jacob LP. Intravenous ketoprofen in thyroid and parathyroid surgery. Anesth Analg 2001; 92: 1052±7 9 Ripouteau C, Conort O, Lamas JP, Auleley GR, Hazebroucq G, Durieux P. Effect of multifaceted intervention promoting early switch from intravenous to oral acetaminophen for postoperative pain: controlled, prospective, before and after study. Br Med J 2000; 321: 1460±3 10 ConfeÂrence de Consensus. Prise en charge de la douleur postopeâratoire chez l'adulte et l'enfant. Ann Fr Anesth ReÂanim 1998; 17: 445±61 11 Huskisson EC. Measurement of pain. Lancet 1974; 9: 1127±31 12 Jensen MP, Karoly P, Braver S. The measurement of clinical pain intensity: a comparison of six methods. Pain 1986; 27: 117±26 13 Ramsay MA, Savege TM, Simpson BR, Goodwin R. Controlled sedation with alphaxalone-alphadolone. Br Med J 1974; 2: 656±9 14 Aldrete JA, Kroulik D. A postanesthetic recovery score. Anesth Analg 1970; 49: 924±34 15 Aubrun F, Monsel S, Langeron O, Coriat P, Riou B. Postoperative titration of intravenous morphine. Eur J Anaesthesiol 2001; 18: 159±65 16 Aubrun F, Monsel S, Langeron O, Coriat P, Riou B. Postoperative titration of intravenous morphine in the elderly. Anesthesiology 2002; 96: 17±23 17 Casagrande JT, Pike MC. An improved approximate formula for calculating sample sizes for comparing two binomial distributions. Biometrics 1978; 34: 483±6 18 ChoinieÁre M, Rittenhouse BE, Perreault S, et al. Ef cacy and costs of patient-controlled analgesia versus regularly administered intramuscular opioid therapy. Anesthesiology 1998; 89: 1377±88 19 Fletcher D, Benoist JM, Gautron M, Guilbaud G. Isobolographic analysis of interactions between intravenous morphine, propacetamol, and diclofenac in carageenin-injected rats. Anesthesiology 1997; 87: 317±26 20 Fletcher D, NeÁgre I, Barbin C, et al. Postoperative analgesia with i.v. propacetamol and ketoprofen combination after disc surgery. Can J Anaesth 1997; 44: 479±85 21 Hill RP, Lubarsky DA, Phillips-Bute B, et al. Cost-effectiveness of prophylactic antiemetic therapy with ondansetron, droperidol, or placebo. Anesthesiology 2000; 92: 958±67 22 Apfel CC, LaÈaÈraÈ E, Koivuranta M, Greim CA, Roewer N. A simpli ed risk score for predicting postoperative nausea and vomiting: conclusions from cross-validation between two centers. Anesthesiology 1999; 91: 693± Schug SA, Sidebotham DA, McGuinnety M, Thomas J, Fox L. Acetaminophen as an adjunct to morphine by patient-controlled analgesia in the management of acute postoperative pain. Anesth Analg 1998; 87: 368±72 24 Heavner JE, Shi B, Diede J, Racz G. Acetaminophen (paracetamol) use and blood concentration in pain patients. Pain Digest 1996; 6: 215±18 25 Thummel KE, Slattery JT, Ro H, et al. Ethanol and production of the hepatotoxic metabolite of acetaminophen in healthy adults. Clin Pharmacol Ther 2000; 67: 591±9 26 Dong H, Haining RL, Thummel KE, Rettie AE, Nelson SD. Involvement of human cytochrome P450 2D6 in the bioactivation of acetaminophen. Drug Metab Dispos 2000; 28: 1397± Niemi TT, Backman JT, SyrjaÈlaÈ MT, Viinakka LU, Rosenberg PH. Platelet dysfunction after intravenous ketorolac or propacetamol. Acta Anaesthesiol Scand 2000; 44: 69±74 319