Extradural morphine gives better pain relief than patient-controlled i.v. morphine after hysterectomy

Transcription

1 British Journal of Anaesthesia 1997; 78: Extradural morphine gives better pain relief than patient-controlled i.v. morphine after hysterectomy M. ERIKSSON-MJÖBERG, J.-O. SVENSSON, O. ALMKVIST, A. ÖLUND AND L. L. GUSTAFSSON Summary We examined if patient-controlled analgesia (PCA) with i.v. morphine provided comparable postoperative analgesia after hysterectomy as extradural morphine, without increasing the incidence of side effects. The study (n 40) was randomized and double-blind. An extradural catheter was inserted before surgery and anaesthesia was standardized. The extradural group received extradural morphine 0.06 mg kg 1 by the end of surgery and a second dose 6 h later. The i.v. group received an i.v. infusion of morphine 0. mg kg 1 after surgery. PCA with morphine 0.04 mg kg 1 i.v. was used in both groups. Pain relief (VAS), side effects and cognitive functions were evaluated for 18 h. Plasma samples were obtained for analysis of morphine concentrations. Mean consumption of PCA morphine was.4 mg h 1 for the i.v. group and 1 mg h 1 for the extradural group. Despite unlimited access to morphine, the i.v. group had higher VAS scores as the extradural group (P 0.001). Plasma concentrations of morphine varied 8 10-fold in both groups. In the i.v. group itching, tiredness, blurred vision and vertigo correlated with cumulative consumption of i.v. morphine whereas in the extradural group this correlation existed only for tiredness. Both groups showed reduced ability to perform tests of cognitive function, indicating a central effect of both i.v. and extradural morphine, despite markedly lower plasma morphine concentrations in the extradural group. (Br. J. Anaesth. 1997; 78: 10 16) Key words Analgesia, postoperative. Analgesics opioid, morphine. Analgesia, patient-controlled. Anaesthetic techniques, extradural. A single dose of extradural morphine provides better and longer lasting pain relief after major surgery than a single dose of i.v. or i.m. morphine. 1 3 However, the clinical use of extradural morphine is hampered by side effects such as pruritus, vomiting, delayed gastric emptying and, most importantly, respiratory depression. 4 Respiratory depression usually occurs within 1 h, although it may occur later, up to 4 h after administration of extradural morphine. 1 4 This delay necessitates supervision of patients for 1 4 h 4. I.v. morphine can also produce respiratory depression. 5 It is well known that the need for i.v. opioids in the postoperative period varies between patients undergoing similar types of surgery even though the same anaesthetic techniques are used. 6 In the postoperative period, patient-controlled analgesia (PCA) is a valuable method for providing pain relief. 5 7 The PCA device allows patients to administer i.v. opioids according to individual needs. Consequently, after major surgery it may be possible to evaluate if the consumption of i.v. morphine gives equivalent or superior pain relief compared with single doses of extradural morphine. This has been examined in some controlled studies with contradictory results Loper and Ready found that patients given repeated doses of extradural morphine after arthrotomy reported significantly lower pain scores than those using PCA with i.v. morphine. 10 In contrast, Weller and co-workers 7 could not find any significant differences in pain ratings between patients receiving single doses of morphine by the extradural route and those who received i.v. morphine using PCA after orthopaedic surgery. A third study reported that after Caesarean section, extradural morphine gave superior pain relief than PCA with i.v. morphine. 8 In that study, the PCA doses were 1 mg and the lockout time was 6 min. 8 The low doses may have reduced the full potential of the PCA technique. Plasma concentrations of morphine and their relation to reported side effects are seldom investigated after i.v. and extradural administration in the postoperative period. We have found no study comparing the efficacy of repeated doses of extradural morphine with higher doses ( 4 mg) of i.v. morphine administered by PCA. In this study, we have compared the effects of PCA administration of i.v. morphine with extradural morphine during the initial 18 h after hysterectomy. The first hypothesis was: patient-controlled administration of i.v. morphine with an initial bolus dose of i.v. morphine can MARIANNE ERIKSSON-MJÖBERG, MD (Department of Anaesthesiology and Intensive Care); JAN-OLOF SVENSSON, LARS L. GUSTAFSSON, MD, PHD (Department of Clinical Pharmacology); OVE ALMKVIST, PHD (Department of Geriatric Medicine); ANDERS ÖLUND, MD, PHD (Department of Obstetrics and Gynaecology); Huddinge University Hospital at Karolinska Institute, S Huddinge, Sweden. Accepted for publication September 4, Correspondence to M. E.-M.

2 Extradural vs patient-controlled i.v. morphine 11 give as good or better pain relief as single doses of extradural morphine combined with PCA administration of i.v. morphine. The second hypothesis was: PCA i.v. morphine administration does not result in more side effects than extradural morphine. Patients and methods The study was approved by the Huddinge Hospital Ethics Committee. The study was double-blind and patients were allocated randomly to one of two groups: i.v. or extradural (table 1). PATIENT RECRUITMENT We examine 84 consecutive patients undergoing elective hysterectomy (table ). Exclusion criteria included backache, obesity, psychiatric disease and age more than 60 yr. Five patients who had agreed to join the study were not included because of administrative problems such as lack of beds in the postoperative unit. We calculated body mass index (BMI, weight (kg) divided by height (m)) and considered 5 kg m to be normal. 11 Patients with a BMI value less than 3 kg m were accepted. We studied 40 female patients who fulfilled the inclusion and exclusion criteria. All patients met the anaesthetist (M. E.-M.) the day before surgery and were given full verbal and written information on the study. They were instructed on how to use the visual analogue scale (VAS) for pain rating and the PCA device was demonstrated. We also performed baseline cognitive tests (see neuropsychological assessments), bloodgas analysis of arterial blood, and venous blood samples were obtained as a zero reference for plasma concentrations of morphine. ANAESTHESIA AND SURGICAL PROCEDURES The anaesthetic procedure was standardized. Patients were given diazepam 10 mg orally in the ward, 1 h before the start of anaesthesia. All patients had an extradural catheter inserted at lumbar segments L3 L4 or L L3. Mepivacaine 0 mg ml l with adrenaline was given extradurally as two consecutive doses. After a test dose of 3 ml, another 1 ml were given to women younger than 50 yr and 9 ml to women older than 50 yr. When the anaesthetic level for pinprick reached T10 1, patients were anaesthetized with thiopentone 5 mg/kg body weight and fentanyl 0. mg. After pancuronium 1 mg, intubation was facilitated with suxamethonium 1 mg/kg body weight. Isoflurane 0.5% was Table Patient recruitment Considered for enrolment 84 patients Refused extradural anaesthesia 10 Administrative problems 5 Excluded 9 History of: Back pain 15 Psychiatric disease Easily induced vomiting after opioid administration High body mass index 5 Poor fluency in Swedish 5 Included in study 40 given in a mixture of nitrous oxide and oxygen to maintain anaesthesia. A single dose of pancuronium 3 mg was given before the start of surgery and additional 1-mg doses when needed. A top-up dose of mepivacaine 8 ml with adrenaline 10 mg ml l was given during surgery, 1 h after the first dose. Administration of isoflurane was discontinued when the skin was closed, and neuromuscular block was antagonized with neostigmine.5 mg. All patients underwent abdominal hysterectomy with or without salpingo-oophorectomy via a Pfannenstiel incision (39 patients). A midline incision was used in one patient only. STUDY DESIGN Blinded solutions of extradural and i. v. morphine, and saline were prepared by the hospital pharmacy and stored in glass bottles. At closure of the peritoneum, patients in the extradural group received a single dose of extradural morphine 0.4 mg ml l, corresponding to 0.06 mg/kg body weight (table 1). The mean dose given was 4 (SD 0.5) mg. Patients in the i. v. group received the same volume of 0.9% saline. After surgery patients were awakened and the trachea extubated. After a mean time of 15 min after the end of surgery, patients were brought to the recovery room where they were monitored for a minimum of 18 h. On arrival in the recovery room (time 0) patients in the i.v. group received an i.v. infusion of morphine 0. ml/kg body weight over 0 min. Mean dose was 1.4 (range ) mg. Patients in the extradural group received the same i.v. volume of physiological saline solution. Patients were connected to the PCA device (Prominject, Pharmacia) which was set to deliver a bolus dose of i.v. morphine 0.04 mg/kg body weight whenever the patient pushed the button. The amount of morphine given in a single dose varied from 3.5 mg. A lockout time of 10 min was used. Table 1 Design of the study I.v. group Extradural group 1 Extradural NaCl 0.15 ml kg 1 as bolus dose when peritoneum is closed Extradural morphine 0.15 ml kg 1 (0.06 mg kg 1 ) as bolus dose when peritoneum is closed Intravenous morphine 0. ml kg 1 (0. ml kg l ) I.v. NaCl 0. ml kg 1 after awakening after awakening 3 PCA 0.04 mg kg 1 bolus dose. Lockout time 10 min PCA 0.04 mg kg 1 bolus dose. Lockout time 10 min 4 Extradural NaCl 0.15 ml kg 1 6 h after first dose Extradural morphine 0.15 ml kg 1 (0.06 mg kg 1 ) 6 h after first dose

3 1 British Journal of Anaesthesia A second dose of extradural morphine 0.06 mg/kg body weight was given 6 h after the initial dose, and a blinded dose of extradural saline to the i.v. group. The mean cumulative dose of extradural morphine after the second dose was 7.9 (1.) mg. Plasma samples were obtained and pain ratings were assessed before the first three episodes of selfadministration. These usually occurred during the first h after arrival in the recovery room. Postoperative pain was assessed using a 100-mm visual analogue scale (VAS) where 0 no pain and 100 unbearable pain. Additionally, VAS scores were assessed and arterial blood samples obtained at specific times (0, 0.5, 1,, 4, 8, 1 and 18 h, and at 1,, 4, 8 and 18 h, respectively, after arrival in the recovery room). Patients spent the night in the recovery room where mobilization is not routine. To evaluate the effects of i.v. and extradural morphine, frequent VAS measurements were made and evaluation of side effects and influence on cognitive functions. VAS assessment on movement and during coughing were not measured as the patients were not mobile. ASSESSMENT OF ADVERSE EVENTS At 4, 8 and 18 h after arrival in the recovery room, staff asked the patients if they experienced any adverse events. If so, they were questioned specifically about the presence of nausea, vomiting, tiredness, vertigo, dry mouth, pruritus, blurred vision and dyspnoea. They were asked to grade symptoms from 0 to 3, where 0 no symptoms, 1 mild, moderate and 3 severe. COGNITIVE FUNCTIONS Two tests of cognitive function were used: the grammatical reasoning test 1 and the letter cancellation test. 13 The grammatical reasoning test evaluates reasoning ability and the letter cancellation test assesses visual search ability, which is measured by speed (number of items covered in 3 min) and correctness (percentage correct) of responses. Both tests were presented to all subjects on four occasions: 1 day before surgery and at, 8, and 18 h after arrival in the postoperative unit. The ability to perform both tests varied considerably as a result of incapacity (tiredness), which might be evaluated by compliance rate (percentage of individuals making efforts to solve the tests). CLINICAL CARE We monitored heart rate, systolic and diastolic arterial pressures, and ventilatory frequency every 30 min. Arterial blood-gas tensions were measured at, 8 and 18 h after arrival in the recovery room. Patients were given i.v. glucose and balanced electrolyte solutions, approximately 3 litre, over 18 h. ANALYTICAL METHODS Blood samples were centrifuged within 1 h. Plasma was transferred into plastic tubes and stored at 0 C until analysis. Plasma concentrations of morphine, morphine-3-glucuronide (M3G) and morphine-6-glucuronide (M6G) were assayed in duplicate using high pressure liquid chromatography, as described previously. 14 The method has an STD (inter-day) variability of 7.0%, 4.% and 3.3% when assaying plasma concentrations of morphine, M3G and M6G, respectively, at therapeutic concentrations. The area under the plasma morphine concentrations vs time curve (AUC) (0 18 h) was calculated for each individual. STATISTICAL METHODS The results were tested with Student s t test for parametric data and the Mann Whitney U test for nonparametric data. Results are given as mean (SD) or median (range) according to distribution. For repeated measurements the Friedman test was used. For analysis of correlation the Spearman rank test was used. P 0.05 was considered significant. The possible differential deleterious effects on cognitive function from the two methods of postoperative pain relief were analysed using a (groups) ε 4 (test occasions) ANOVA for each measure. Results There were no differences in age, weight, height or duration of surgery between the groups (table 3). There was a difference in BMI (P 0.05) and in perioperative bleeding (P 0.05). There were three patients with excessive bleeding. One patient in the i.v. group bled 100 ml and two patients in the extradural group bled 1400 ml and 1500 ml, respectively. One patient in the i.v. group and six patients of those receiving extradural morphine bled ml. All other patients lost less than 500 ml. Patients in the extradural group reported lower pain scores than the i.v. group at 4, 8, 1 and 18 h after arrival in the postoperative unit (P 0.001) (fig. 1). At 4 and 18 h, mean VAS scores were 9 (SD 11) mm and 9 (13) mm, respectively, in the extradural group. Corresponding values in the i.v. group were 51 (3) mm and 8 (18) mm, respectively. Even though extradural morphine was administered on average 30 min before arrival in the postoperative unit, these patients had as high VAS scores as women treated with i.v. morphine during the initial 1 h (fig. 1). Patients reported no pain on awakening. At 8 h, only two of 0 patients in the i.v. group were pain free (VAS 0) compared with 15 of 0 patients Table 3 Characteristics of patients (mean (SD) or range). *P 0.05 I.v. (n 0) Age (yr) 45 (3 5) 46 (36 56) Weight (kg) 63 (11) 67 (8) BMI (kg m ) 3 (3) 5 (3)* Surgery (min) 105 (31) 107 (37) Bleeding (ml) 368 (430) 50 (390)* Mepivacaine (mg) 378 (14) 378 (14) Extradural group (n 0)

4 Extradural vs patient-controlled i.v. morphine 13 Figure 1 Mean pain scores (VAS) in the extradural (, n 0) and i.v. (, n 0) morphine groups during the initial 18 h after operation. Time 0 time of arrrival in the recovery room. in the extradural group. The corresponding values at 18 h were two of 0 and 10 of 0, respectively. There was a difference between groups in the frequency and timing of demand for PCA morphine (fig. ). The extradural group took only a few doses (mean 6.6 (5.9) (range 0 3) doses) (fig. ), mostly within h. The i.v. group took morphine repeatedly over 18 h (mean 18 (7.3) (range 7 30) doses) (fig. ). This difference was significant (P 0.001, Mann Whitney). Consequently, the cumulative dose of morphine at the end of the 18-h study period differed between the two groups (fig. 3). The cumulative dose was mg in the i.v. group and 0 55 mg in the extradural group (P 0.001, Mann Whitney). Mean morphine consumption in the i.v. group was 44 (0.8) mg corresponding to.4 mg h 1 over 18 h. The values were 17. (16) mg and 1 mg h 1 for the extradural group (fig. 3). In the i.v. group we found a 10 0-fold variability in plasma morphine concentrations when patients took their three initial i.v. bolus doses of PCA morphine (table 4). Plasma concentrations of M6G before the first, second and third PCA doses showed less variation but increased significantly with time (P 0.001, Friedman test) (table 4). Plasma concentrations of morphine in the extradural group were markedly lower at the initial three PCA doses. Interestingly, mean plasma morphine concentrations, before the initial three doses, were consistent in patients in the i.v. group. Plasma concentrations of morphine at fixed intervals in the two groups are presented in figure 4. Concentrations were highest at 1 and h when pain scores were high. There were large inter-individual variations in plasma concentra- Table 4 Plasma morphine and morphine-6-glucuronide (M6G) concentrations and pain ratings in the i.v. and extradural groups before self-administration of i.v. morphine (mean(sd) (range)). ***P (Friedman test) Before 1st PCA dose Before nd PCA dose Before 3rd PCA dose Figure Accumulated successive demands for PCA in the extradural (, n 0) and i.v. (, n 0) groups during the 18 h after operation. Values are mean, SD. VAS I.v. group 70 (1) (5 100) 71 (1) (30 100) Extradural group 57 (3) 67 () (0 100) (30 100) Plasma morphine concentration (nmol litre 1 ) I.v. group 169 (158) 130 (113) (9 637) (45 59) Extradural group 19 (9.9) 5 (46) (0 3) (.7 167) Plasma M6G concentration (nmol litre 1 ) I.v. group 100 (37) 149 (46) (11 168) Extradural group 19 (8.3) (0 34) (69 56) 31 (19) (1.7 68) 73 (4) (38 100) 63 (6) (6 100) 108 (6) (40 309) 67 (43) (8 155) 16 (40)*** (78 35) 58 (9) (14 94) Figure 3 Cumulative PCA morphine consumption over 18 h in individual patients in the extradural ( ) and i.v. ( ) groups (n 0 patients in each group), ranging from the lowest to the highest morphine consumption. Figure 4 Plasma concentrations of morphine at 1, 4, 8 and 18 h in the extradural (, n 0) and i.v. (, n 0) groups. Values are mean, SD.

5 14 British Journal of Anaesthesia tions, especially in the i.v. group. Morphine concentrations were lower in the extradural group during the whole postoperative period, resulting in mean 18 h AUC values of 1504 (106) (range ) nmol h for the i.v. and 497 (577) (range ) nmol h for the extradural group, res pectively (P 0.001, Mann Whitney). We found no correlation between AUC of plasma morphine concentrations and the AUC of VAS in either group over 18 h or any correlation between cumulative consumption of morphine at 18 h and AUC of VAS at 0 18 h. SIDE EFFECTS All patients in each group reported adverse events. Tiredness and dryness of the mouth were frequent in both groups of patients. At and 8 h, all patients in the i.v. group reported mild to severe tiredness and at 18 h only one patient reported no tiredness. In the extradural group (n 0), patients reported mild to severe tiredness at, 8 and 18 h. At 18 h the frequency of itching was 4% and 37% in the extradural and i.v. groups, respectively (ns). The risk of morphine-induced urinary retention could not be measured as all patients had urinary catheters for more than 4 h. At 18 h after arrival in the recovery room (time 0), cumulative PCA morphine in the i.v. group correlated with intensity of each of the following side effects: tiredness, itching, blurred vision and vertigo (n 0, P 0.05 Spearman rank; 0 no symptoms, 1 mild, moderate and 3 severe). In the extradural group tiredness at 18 h correlated with cumulative morphine (n 0, P 0.05, Spearman rank). Preoperative P a CO was less than 5 kpa in both groups. At 8 h, mean P a CO was 5.7 kpa in the extradural and 5.6 kpa in the i.v. group, with similar variability within each group. Both groups had a significantly increased P a CO compared with preop- erative values (P 0.001, Friedman test), but there was no difference between groups (Mann Whitney). One patient in the extradural group experienced respiratory depression with increasing Pa CO levels up to a maximum of 8.3 kpa at 10 h after the first dose of extradural morphine and 4 h after the second extradural dose. A total of 7.3 mg of extradural morphine had been given to this patient. She had self-administered i.v. morphine on two occasions and had received a total dose of 4.8 mg of i.v. morphine at 10 h. This patient also hypoventilated with decreasing ventilatory frequency but had no signs of apnoea. Her ventilatory frequency was less than 10 bpm on several occasions, but she did not require ventilatory support and administration of naloxone was not necessary. COGNITIVE FUNCTIONS Some patients were too tired to complete the tests (compliance). At h after arrival in the recovery room, 45% in the extradural and 70% in the i.v. group were too tired to complete the tests. The effect of group was not statistically significant for any of the cognitive measures or compliance. But the effect of test occasion was significant for both compliance (P ), speed (P ) and correctness (P 0.05) of letter cancellation and for compliance (P ) and correctness of grammatical reasoning (P 0.001), but not for speed of grammatical reasoning. There was no interaction between group and test occasion in any measure. There were no significant differences between groups. CLINICAL CARE There were no differences between the extradural and i.v. groups in arterial pressure, heart rate or ventilatory frequency. In both groups several patients had a ventilatory frequency less than 10 bpm but no patient had apnoea. Discussion Our results confirm previous reports on excellent postoperative analgesia from extradural morphine, and support the opinion that the quality of pain relief from extradural morphine is superior to that generated by i.v. morphine administered on demand by patients. Two doses of extradural morphine gave very good analgesia. During the 18 h after operation, only eight of 0 patients required more than 15 mg of i.v. PCA morphine. However, four of 0 patients required more than 30 mg of i.v. PCA morphine in addition to extradural morphine. Sjöström, Hartvig and Tamsen 16 have shown that a minority of patients take remarkably high PCA doses of extradural morphine in the postoperative period. In their study, hourly consumption of extradural morphine via PCA varied from 0.19 to 1.04 mg. It is not yet known what mechanisms are behind this variability among patients in the need for opioids after extradural morphine administration. During the first 1 h in the recovery room, patients in the extradural group had VAS scores of mm (fig. 1). The slow onset of analgesia in the extradural group was related to the hydrophilic nature of morphine causing slow diffusion of drug to the site of opioid receptors localized in the dorsal horn of the spinal cord. 17 Despite individualized administration of i.v. morphine, patients in the i.v. group had consistently higher VAS ratings than the extradural group. It seems unlikely that they could have taken more morphine as all patients except one complained of tiredness. A few patients hesitated to use additional PCA doses because of fear of side effects such as vomiting and nausea. This fear proved correct as we found an association between high doses of i.v. opioids and side effects. An infusion longer than 1 min was not feasible to reduce side effects as some patients complained of incomplete pain relief from PCA, especially during the first h after surgery. These patients, in contrast, found the lockout time of 10 min too long. Nevertheless, the i.v. group had mean VAS scores greater than 30 mm during the whole postoperative period. In spite of the i.v. loading dose and our relatively high PCA doses, pain relief was inadequate.

6 Extradural vs patient-controlled i.v. morphine 15 Why did i.v. morphine provide unsatisfactory pain relief after hysterectomy in this study? It has been suggested that various forms of pain (or pain components) are differentially sensitive to the route of administration of opioids. In one study, patients with visceral pain that was inadequately relieved by systemic opioids had pain relief from extradural morphine. 18 As hysterectomy is an abdominal operation there is a somatic (cutaneous and muscular) and a visceral pain component. I.v. morphine and i.v. pethidine had little effect on the visceral components of labour pain whereas the somatic type (back pain) was relieved We suggest that failure of i.v. morphine to provide as good analgesia as extradural morphine can be explained by the fact that i.v opioids are less efficient in treating visceral pain. We found nine-fold variation in the amount of self administered morphine in the i.v. group over the 18 h after operation. This variation is in accord with results reported by previous investigators. 6 1 Patients in the i.v. group required a mean of.4 mg h 1 of morphine. This value is close to that of.6 mg h 1 reported by Tamsen and colleagues in middleaged patients (30 50 yr) undergoing major colon surgery. 5 We also observed marked inter-individual variability in plasma morphine concentrations (more than 10-fold). Mean plasma morphine concentrations decreased slightly from the first to the third self-administered dose (ns). Our results indicate that there are inter-individual differences in sensitivity to morphine. The reason is unknown but could be genetically influenced. There was a significant increase in concentrations of M6G which could explain the decreasing concentrations of mean plasma morphine as M6G contributes to pain relief. The highest plasma concentrations of morphine were seen at 1 and 4 h when VAS scores were high (figs 1,4). Even when patients in the i.v. group had high concentrations of morphine (mean 169 nmol litre 1 before the first PCA dose) their pain relief was inadequate (table 4). Dahlström and co-workers estimated 16 ng ml 1 ( equivalent to 56 nmol litre 1 ) as the minimum analgesic concentration. 6 This value was not confirmed in our investigation where patients had high pain scores even at morphine concentrations greater than 56 nmol litre 1 (table 4). The AUC of morphine concentrations 0 18 h did not correlate with the AUC of VAS, indicating that plasma morphine concentration was not directly influencing the extent of pain relief. Morphine concentrations at 1,4,8 and 18 h, and the AUC of morphine concentrations 0 18 h were significantly lower in the extradural group who nevertheles s had superior pain relief. This result is in accord with earlier studies showing that the pain relief from extradural morphine is mainly caused by its effect on receptors in the spinal cord while plasma morphine concentrations are more important for side effects and influence on cognitive functions. 17 Unexpectedly, no differences in cognitive functions were seen between the groups. Cognitive tests for monitoring CNS effects indicated that all patients were affected, especially at h after operation, showing that the tests were sensitive. There are few studies showing VAS scores within 1 h of awakening after abdominal surgery. In an investigation of pain relief after hysterectomy, VAS scores were 65 mm (mean) in the control group immediately after surgery. 3 There are other studies where VAS scores were mm during the first h after surgery. 4 6 We have found no study on abdominal surgery showing VAS scores less than 40 in this early period when parenteral opioids are used for pain relief. These results are similar to ours and suggest that extradural morphine or local anaesthetics have to be administered in advance and before the end of surgery. Acknowledgements Mrs Christina Alm gave considerable help and support during the planning and initial phase of the study. Mrs Birgit Silén assisted with most of the drug analysis. Anette Rickebjer provided expert secretarial assistance. Supported in part by the Swedish Medical Research Council (390) and by funds from the Karolinska Institute. References 1. Gustafsson LL, Friberg-Nielsen S, Garle M, Mohall A, Rane A, Schildt B, Symreng T. Extradural and parenteral morphine: kinetics and effects in postoperative pain. A controlled clinical study. British Journal of Anaesthesia 1984; 54: Cousins MJ, Mather LE. Intrathecal and epidural administration of opioids. Anesthesiology 1984; 61: Rawal N, Sjöstrand UH, Dahlström B, Nydahl PA, Ostelius J. Epidural morphine for postoperative pain relief: a comparative study with intramuscular narcotic and intercostal nerve block. Anesthesia and Analgesia 198; 61: Rawal N, Arnér S, Gustafsson LL, Allvin R. Present state of extradural and intrathecal opioid analgesia in Sweden: a nation-wide follow-up survey. British Journal of Anaesthesia 1987; 59: Tams en A, Hartvig P, Fagerlund C, Dahls tröm B, Bondesson U. Patient-controlled analgesia therapy: clinical experience. Acta Anaesthesiologica Scandinavica Supplement 198; 74: Dahlström B, Tamsen A, Paalzow L, Hartvig P. Patientcontrolled analgesia therapy, Part i.v.: Pharmacokinetics and analgesic plasma concentrations of morphine. Clinical Pharmacokinetics 198; 7: Weller R, Rosenblum M, Conard P, Gross JB. Comparison of epidural and patient controlled intravenous morphine following joint-replacement surgery. Canadian Journal of Anaesthesia 1991; 38: Harrison DM, Sinatra R, Morgese L, Chung JH. Epidural narcotic and patient-controlled analgesia for post-cesarean section pain relief. Anesthesiology 1988; 68: Spencer L, Carpenter RL, Neal JM. Epidural anesthesia and analgesia. Anesthesiology 1995; 8: Loper KA, Ready LB. Epidural morphine after anterior crucial ligament repair: A comparison with patient controlled intravenous morphine. Anesthesia and Analgesia 1989; 68: FAO/WHO/ UNU Expert consultation. Energy and protein requirements. Geneva: WHO Technical report series, 1985; No Baddeley A. A three-minute reasoning test based on grammatical transformations. Psychonomic Science 1968; 10: Lezak MD. Neuropsychological Assessment. New York: Oxford University Press, Svensson JO, Rane A, Säwe J, Sjöqvist F. Determination of morphine, morphine-3-glucuronide and (tentatively) morphine- 6-glucuronide in plasma and urine using ion-pair highperformance liquid chromatography. Journal of Chromatography 198; 30: Assantila R, Eklund P, Rosenberg PH. Continuous epidural

7 16 British Journal of Anaesthesia infusion of bupivacaine and morphine for postoperative analgesia after hysterectomy. Acta Anaesthesiologica Scandinavica 1991; 35: Sjöström S, Hartvig P, Tamsen A. Patient controlled analgesia with extradural morphine or pethidine. British Journal of Anaesthesia 1988; 60: Yaksh TL, Durant PAC, Gaumann DM, Stevens CW, Mjanger E. The use of receptor-selective agents as analgesics in the spinal cord: Trends and possibilities. Journal of Pain and Symptom Management 1987; 3: Arnér S. Differentiation of pain and treatment efficacy. PhD Thesis. Stockholm: Karolinska Institute, Olofsson C, Ekblom A, Ekman-Ordeberg G, Granström L, Irestedt L. Analgesic efficacy of intravenous morphine in labour pain a reappraisal. International Journal of Obstetetric Anesthesia 1996; 5: Olofsson C, Ekblom A, Ekman-Ordeberg G, Hjelm A, Irestedt L. Lack of analgesic effect of systemically administered morphine or pethidine on labour pain. British Journal of Obstetrics and Gynaecology, 1996; 103: Tamsen A, Hartvig P, Dahlström B, Lindström B, Holmdahl M. Patient-controlled analgesia therapy in the early postoperative period. Acta Anaesthesiologica Scandinavica 1979; 3: Portenoy R, Thaler HT, Inturrisi CE, Friedlander-Klar H, Foley KM. The metabolite morphine-6-glucuronide contributes to the analgesia produced by morphine infusion in patients with pain and normal renal function. Clinical Pharmacology and Therapeutics 199; 51: Sinclair R, Weslander G, Cassuto J, Hedner T. Postoperative pain relief: topical lidocaine in the surgical wound of hysterectomized patients. Acta Anaesthesiologica Scandinavica 1996; 40: Sandler AN, Baxter AD, Kaz J, Samson B, Friedlaner M, Norman P, Koren G, Roger S, Hull K, Klein J. A doubleblind placebo-controlled trial of transdermal fentanyl after abdominal hysterectoy. Anesthesiology 1994; 81: Mansfield M, James KS, Kinsella J. Influence and timing of morphine on postoperative pain and analgesic requirements. Brithish Journal of Anaesthesia 1996; 76: Eriksson-Mjöberg M, Kristiansson M, Carlström K, Ekund J, Gustafsson LL, Ölund A. Preoperative infiltration with buvpivacaine effects on pain relief and trauma response (cortisol and interleukin-6). Acta Anaestesiologica Scandinavica 1996, in press.