Immediate extubation and epidural analgesia allow safe management in a high-dependency unit after two-stage oesophagectomy. Results of eight years of experience in a specialized upper gastrointestinal unit in a district general hospital

M. V. Chandrashekar1, M. Irving2, J. Wayman1, S. A. Raimes2 and A. Linsley3

1 Royal Victoria Infirmary, Queen Victoria Road, Newcastle upon Tyne, UK. 2 Department of Surgery, Cumberland Infirmary, Newtown Road, Carlisle, Cumbria CA2 7HY, UK. 3 Department of Anaesthetics, Cumberland Infirmary, Carlisle, UK

Corresponding author. E-mail: simon.raimes@ncumbria-acute.nhs.uk

Accepted for publication: December 19, 2002


    Abstract
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
Background. The perioperative management of two-stage oesophagectomy has not been standardized and the prevailing practice regarding the timing of extubation after the procedure varies. This audit has evaluated the outcome, in particular the respiratory morbidity and mortality, after immediate extubation in patients who have had thoracic epidural analgesia.

Methods. All the patients who underwent two-stage oesophagectomy by a single specialist upper gastrointestinal surgeon were recorded both retrospectively (1993–1999) and prospectively (1999–2001). Physical characteristics, comorbid factors, anaesthetic management and postoperative events were recorded on a computer database. Analysis was undertaken to evaluate the morbidity and mortality, in particular the need for reventilation and transfer to the ITU.

Results. Seventy-six patients underwent two-stage oesophagectomy between 1993 and 2001. Seventy-three (96%) patients were extubated in theatre and transferred to a high-dependency bed. Three were ventilated electively and extubated within 36 h and made an uncomplicated recovery. Seven (10%) of the immediately extubated patients subsequently needed admission to the ICU and reventilation. Sixty-seven patients had effective epidural analgesia and nine needed i.v. morphine by patient-controlled analgesia. The 30-day or in-hospital mortality was 2.6% (2 of 76). A further two patients died within 90 days, but after discharge. Respiratory complications were responsible for half of the overall morbidity (44.7%). Respiratory failure occurred in 6.5% (5 of 76) and acute respiratory distress syndrome in 2.6% (2 of 76). Both the in-hospital deaths occurred in patients requiring reventilation and resulted from respiratory complications. The following factors were found to be significant in the reventilated patients: duration of one-lung ventilation; forced expiratory volume in the first second; and ratio of forced expiratory volume in the first second/forced vital capacity.

Conclusions. Immediate extubation after two-stage oesophagectomy in patients with thoracic epidural analgesia is safe and associated with low morbidity and mortality. Patients can be managed in a high-dependency unit, thus avoiding the need for intensive care. This has cost-saving and logistical implications.

Br J Anaesth 2003; 90: 474–9

Keywords: anaesthetic techniques, epidural; complications, extubation tracheal; surgery, gastrointestinal


    Introduction
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
Despite considerable advances in preoperative staging, operative techniques and postoperative monitoring, oesophagectomy continues to be associated with high mortality and morbidity.1 There are multiple reasons for this being a high-risk procedure. Between half and two-thirds of patients with cancer of the oesophagus have a varying degree of dysphagia, rendering them malnourished. Nearly half of the patients are over 70 yr of age and consequently tend to have significant comorbidity. The procedure is long and extensive and inflicts considerable physiological stress. The second phase of the procedure involves a right thoracotomy, necessitating one-lung ventilation for a considerable period to allow access to the oesophagus.2 One-lung ventilation is associated with microbarotrauma to the dependent lung35 and the collapsed lung is subjected to organ reperfusion injury.6

It is well recognized that mechanical ventilation, even for a brief period, is associated with barotrauma5 and therefore the policy of continued postoperative ventilation after oesophagectomy immediately following one-lung ventilation can be deleterious.

The recent National Confidential Enquiry into Perioperative Deaths (NCEPOD) has recognized respiratory complications as the main cause of high morbidity and mortality after oesophagectomy.7 Adult respiratory distress syndrome (ARDS) is a major contributor to respiratory morbidity and mortality after oesophagectomy and, once developed, is associated with a mortality of 50%.810 Several studies have compared the outcomes of early and late extubation after oesophagectomy.11 These have shown that ventilation in the postoperative period is associated with a high incidence of ARDS and prolonged ICU stay, and has major cost implications. One of the key factors that has facilitated early extubation and admission to a high-dependency unit (HDU) rather than an ICU bed has been the introduction of thoracic epidural analgesia. We present the results of an 8-yr audit of the practice of immediate extubation after two-stage oesophagectomy in a specialized upper gastrointestinal unit in a district general hospital.


    Methods
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
Preoperative assessment
All patients had a full blood count, serum biochemistry, liver function tests, ECG and chest x-ray. Pulmonary function tests were carried out to measure forced vital capacity (FVC) and forced expiratory volume in the first second (FEV1), on the assumption that patients with FEV1/FVC of less than 50% are at high risk of needing postoperative ventilation. The ability of the patient to walk up two flights of stairs was used as a simple exercise test. This is believed to be a good indicator of pulmonary function. Initially blood gas analysis was done only on patients with abnormal lung function test results, but it is now done routinely on all patients. Patients with ischaemic heart disease were assessed by a cardiologist to optimize preoperative treatment.

Staging investigations of the oesophageal carcinoma included endoscopic assessment, ultrasonography of the abdomen, computed tomography of the chest and abdomen and a staging laparoscopy with or without cytology/biopsy when the carcinoma had a significant infradiaphragmatic component. The patients were assessed by the dietician and counselled by the consultant surgeon and the oesophageal and gastric specialist nurse.

Anaesthetic details
For premedication, most elective patients received lorazepam 1 mg orally the night before operation and on the morning of operation. Normal medication was continued until surgery.

In the anaesthetic room, routine monitoring, including ECG, blood pressure and pulse oximetry, was instituted. A wide-bore i.v. catheter and an arterial line were inserted under local anaesthesia. A thoracic epidural was sited before induction of anaesthesia, using a midline approach at variable levels between the sixth thoracic and the tenth thoracic vertebra. A test dose of plain bupivacaine 4 ml 0.5% was given. The patient was anaesthetized using propofol or etomidate as the induction agent. Morphine 10 mg was given i.v. at induction. The patient was intubated with a Robertshaw red rubber double-lumen tube of appropriate size using succinylcholine if there was a danger of aspiration, or a long-acting neuromuscular blocking drug if not. The position of the tube was checked by observation and auscultation. A right-sided central venous catheter was inserted into the internal jugular vein.

In theatre, the patient was ventilated with nitrous oxide–oxygen and an inhalational agent (sevoflurane or isoflurane) using low-flow anaesthesia. The following variables were monitored and recorded intraoperatively: direct arterial pressure, central venous pressure, capnography, airway pressures, and temperature. Intraoperative heat loss was minimized with a Biar-hugger warm air blanket and all i.v. fluids were administered through a warmer. The epidural was topped up with bupivacaine 0.5% if indicated by tachycardia or hypertension, usually up to a maximum of 10 ml. A standard epidural infusion was commenced at the start of surgery. This infusion was made up of bupivacaine 0.1%, 400 ml with diamorphine 5 mg. It was run at the rate of 8–10 ml h–1 throughout the operation.

After surgery, patients were routinely extubated in theatre and transferred to an HDU bed unless there was a specific contraindication. Analgesia was continued with the epidural infusion. Boluses of plain bupivacaine 0.25% were given until the patient was comfortable and then the infusion was continued at an appropriate rate. Most patients received paracetamol and/or non-steroidal analgesia with diclofenac per rectum, if not contraindicated, from the second postoperative day.

The surgical procedure was a standard Ivor–Lewis two-stage oesophagectomy with two-field lymphadenectomy. The first stage involved a laparotomy through an upper midline incision to mobilize the stomach to allow it to be brought up into the chest, and to dissect the hiatus and relevant lymph nodes. The blood supply to the stomach was maintained by the right gastric and right gastroepiploic arteries. The second stage involved a posterolateral thoracotomy through the right fifth intercostal space. A subtotal oesophagectomy was performed and a stapled oesophagogastric anastomosis constructed high up in the chest.

Statistics
The data were collected retrospectively (August 1993 to September 1999) and prospectively (October 1999 to April 2001). The physical characteristics of the patients, comorbid factors, anaesthetic management and postoperative events were recorded on a computer database (Microsoft Access version 10). Analysis was undertaken using the SPSS (Statistical Programme for Social Sciences) version 10 (SPSS, Chicago, IL, USA). Continuous data were analysed using the Mann–Whitney U-test and significance was assumed where P<0.05.


    Results
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
Physical characteristics
Seventy-six patients underwent two-stage oesophagectomy between 1993 and 2001. Fifty of these were male patients and the male:female ratio was 2:1. Median (range) age for males was 66  (46–82) yr and for females 68 (41–76) yr. Median (range) weight for males was 75  (55–106) kg and for females 57 (40–87) kg. Two patients underwent the procedure as an emergency, one after iatrogenic perforation of the tumour and the other because of non-availability of an HDU bed at the planned elective time.

Pathology and tumour stage
The vast majority of tumours were adenocarcinoma, followed by squamous carcinoma. Four patients underwent oesophagectomy for high-grade dysplasia in Barrett’s oesophagus. Table 1 shows the types of tumour for which oesophagectomy was carried out.


View this table:
[in this window]
[in a new window]
 
Table 1 Histology of oesophageal lesions
 
Forty-two per cent of operated patients had stage III oesophageal tumours and only two of the 76 patients had stage IVa (lower oesophageal tumours with involvement of coeliac nodes). Table 2 lists the tumour stages of the oesophageal cancers.


View this table:
[in this window]
[in a new window]
 
Table 2 Tumour stage
 
Selection criteria
Approximately one-third of patients with oesophageal carcinoma presenting or referred to our unit were suitable for surgical resection. The selection criteria included resectable tumour with no metastases, ASA grade III or less, good exercise tolerance (patient able to climb two flights of stairs) and the patient’s willingness to undergo surgical resection. Two of the 76 patients who underwent oesophagectomy, having being staged as resectable with no metastasis, were found to have metastases in the coeliac nodes on the final histological evaluation.

Anaesthetists
A single dedicated upper gastrointestinal anaesthetist was involved in the treatment of 81% of the patients. For the remaining 19% (15 patients), a total of nine different anaesthetists were involved.

ASA grades
Twenty-two per cent of the patients were ASA I, 65% were ASA II and 13% were ASA III.

Epidural analgesia
Seventy-three patients (96%) had a preinduction thoracic epidural catheter placed. Three patients did not have the epidural catheter sited; this was because of technical failure in two patients, and one patient refused to have it. In sixty-seven (91.8%) patients, epidural analgesia was effective and was used for an average of 6 days. In the remaining six patients and the three patients who did not have the epidural catheter, i.v. morphine by patient-controlled analgesia was used. None of these nine patients needed reventilation. There were no epidural-related complications.

Extubation
Seventy-three patients (96%) were extubated in theatre at the end of the procedure and were successfully transferred to an HDU bed. Seven of these patients (10%) needed reventilation at some stage after the operation. Three patients were ventilated electively and were extubated within 36 h. One patient was extubated after 12 h, another at 24 h and the third patient 36 h after the operation. The reasons for the delay in extubation were that the operation was felt to be too long in one patient, the duration of the procedure being 430 (median 330) min and one-lung ventilation lasting 200 (median 130) min. In the second patient, it was difficult to maintain oxygen saturation on one-lung ventilation and intermittent two-lung ventilation was needed. In the third patient no specific reason was given. None of these three patients required subsequent reventilation.

Length of HDU stay
Sixty-six patients (86%) stayed for more than 36 h in the HDU, with a median (range) stay of 3 (2–15) days. The remaining 10 patients (14%) stayed less than 36 h, five staying for less than 24 h.

Admission to ICU and reventilation
Three patients who were not extubated in theatre were transferred to the ICU and none of these patients were reventilated. Of the 73 patients extubated in theatre and transferred to the HDU, seven (9%) were subsequently admitted to the ICU for reventilation for a median (range) of 9 (6–30) days after surgery. Table 3 shows the indication for admission to the ICU, the duration of stay and the outcome in this group.


View this table:
[in this window]
[in a new window]
 
Table 3 Indications, duration and outcome of reventilated patients admitted to the ICU. *This patient required ventilation twice
 
Factors associated with reventilation
The significance of the ventilation factors (Table 4), comorbidity and other factors (Table 5) was evaluated. The duration of one-lung ventilation time was significantly longer (P<0.005) in the group that need reventilation. FEV1 and the FEVI/FVC ratio were each significantly lower (P<0.05) when compared with the non-reventilated group. A history of smoking, chronic obstructive airway disease and neoadjuvant therapy appeared to be associated with the reventilated group, but the small number of cases precludes meaningful statistical analysis. Fifty-nine patients (76%) had a blood transfusion in the perioperative period, with a range of 1–8 units and a median of 2 units (Table 5).


View this table:
[in this window]
[in a new window]
 
Table 4 Difference in operative and preoperative ventilation variables in patients who were reventilated after surgery in comparison with those who were not
 

View this table:
[in this window]
[in a new window]
 
Table 5 Comorbidity and other factors in patients who were or were not ventilated after surgery
 
Patients who underwent oesophagectomy for squamous carcinoma were at increased risk of needing reventilation. Five out of 18 patients with squamous carcinoma compared with 2 out of 58 patients with adenocarcinoma required reventilation (Table 3). Four of the seven patients who were reventilated were anaesthetized by the dedicated upper gastrointestinal anaesthetist (4 of 61), while the remaining three patients were managed by non-regular anaesthetists (3 of 15). Nine patients (out of 76) did not receive epidural analgesia, but none of these needed reventilation, while all seven patients who were reventilated had a working epidural.

Reoperation
Seven patients (9%) underwent reoperation. Six patients underwent rethoracotomy, four for ligation of a leaking thoracic duct, one to ligate a bleeding mediastinal vessel, and one for reconstruction of the stomach conduit as the proximal part had become ischaemic. One patient had a feeding jejunostomy constructed. All the reoperated patients were again extubated in theatre at the end of the procedure except for the patient who had a re-anastomosis and returned, ventilated, to the ICU.

Morbidity and mortality
The 30-day or in-hospital mortality was 2.6% (n=2) and the 90-day mortality, including the 30-day deaths, was 5.2% (n=4). ARDS was responsible for the two in-hospital deaths. Of the two patients who died within 90 days, but after hospital discharge, one died of a pulmonary embolus and the other from a paraneoplastic neurological syndrome. It is noteworthy that all four patients who died within 90 days of surgery had squamous cancers; there were no deaths in this period in those with the more common adenocarcinoma.

Overall, 44.7% of patients suffered a postoperative complication, more than half having significant pulmonary problems. The complications are listed in Table 6.


View this table:
[in this window]
[in a new window]
 
Table 6 Complications
 

    Discussion
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
Oesophagectomy for carcinoma of the oesophagus is recognized as a high-risk procedure associated with significant morbidity and mortality.1 The majority of patients who require this type of surgery are in their seventh and eighth decades and have significant concomitant disease. The reported mortality in published series from specialized centres varies between 3 and 17%.1 The morbidity is of the order of 40–50%.3 1214 The in-hospital mortality rate in this series is 2.6%, with a morbidity of 44.7%. As found in the recent NCEPOD report,7 respiratory problems remain the major cause of both mortality and morbidity after oesophagectomy for cancer. Muller and colleagues15 noted an incidence of 27% of respiratory failure after oesophagectomy. In this series, the incidence of respiratory failure was 6.5% and the specific pulmonary complications were 25%, well below the reported incidences in published series.3

ARDS is the main contributor to the high respiratory morbidity and mortality after transthoracic oesophagectomy.7 8 The reported incidence varies from 17 to 53%.8 13 ARDS carries a very high mortality rate, with fewer than 50% surviving the complication.10 The incidence of ARDS in our series was 2.6%, which is well below the reported incidence; however, it was responsible for both of the in-hospital deaths.

During the thoracic phase of the procedure, the dependent and non-dependent lungs are both subjected to injury. The dependent lung is exposed to microbarotrauma,5 while the collapsed lung suffers ischaemia–reperfusion injury.6 A recent multicentre study on ARDS has shown that a high tidal volume is associated with increased mortality.4 During one-lung ventilation, the dependent lung receives all the preset tidal volume and most of the pulmonary circulation, and thus is subjected to a high tidal volume that could result in ventilation-induced lung injury.3 On re-expansion, the non-dependent lung is exposed to the well-recognized organ ischaemia–reperfusion injury.6 Tandon and colleagues3 noted that the longer the one-lung ventilation time and the operation time the greater the chance of acute lung injury. In our series, the duration of one-lung ventilation in the reventilated group was significantly longer. On this background of reperfusion injury to the collapsed lung and microbarotrauma to the ventilated lung, elective continued ventilation in the postoperative period after oesophagectomy could be harmful and might precipitate ARDS and respiratory failure. This could be one of the reasons for the high incidence of respiratory failure and ARDS reported in many series. We believe that avoidance of this factor, together with our low rate of postoperative septic complications, has kept our incidence of ARDS low.

Smoking and chronic obstructive airway disease (COAD) are known predisposing factors for postoperative respiratory morbidity.3 14 In our series, four of the seven patients who had to be reventilated (57%) were smokers and two of the seven (29%) had significant COAD. The other factors that appeared to be associated with reventilation were squamous carcinoma of the oesophagus and neoadjuvant chemoradiotherapy. It is noteworthy that the proportion of patients with squamous cancers who required reventilation or had postoperative respiratory problems was considerably higher than the percentage with adenocarcinoma. This subgroup of oesophageal cancer patients requires even more careful preoperative assessment and perioperative management. All the deaths within 90 days in our series occurred in those with squamous cancer of the oesophagus.

The other major problem with two-stage oesophagectomy is the degree of operative trauma inflicted on the patient. The procedure involves long abdominal and thoracic incisions covering several dermatomes (4th thoracic to 11th thoracic) and requires deliberate excision or fracturing of ribs to improve access. Therefore, optimum analgesia is paramount. It has been shown that adequate postoperative analgesia is associated with lower cardiopulmonary complications, lower mortality and reduced cost.16 17 Studies have looked at various postoperative analgesic regimens and have conclusively shown epidural analgesia to provide the most satisfactory analgesia.18 It has been demonstrated that thoracic epidural analgesia reduces the incidence of both fatal and non-fatal respiratory complications.19

It is our experience that the key factor that facilitates immediate postoperative tracheal extubation is good epidural analgesia. Terai and colleagues20 have shown that early extubation is related to the efficacy of continuous epidural administration of analgesia. Brodner and colleagues21 have shown that, by blocking the perioperative stress response, it is possible to extubate early (mean, 6.1 vs 25.1 h). However, in none of these studies has the feasibility and safety of immediate extubation been examined. The policy in our unit is to commence epidural analgesia before incision, starting with a bolus dose and giving a continuous infusion perioperatively. This facilitates the blocking of pain stimuli, thus eliminating or minimizing the stress response. This is complemented by intensive physiotherapy to the chest and early mobilization. In our series, 96% of the patients were extubated at the end of the procedure in theatre. Not only are the incidences of ARDS and respiratory failure significantly lower than in most reported series, but the incidences of admission to the ICU (10%) and reventilation (9%) are also lower.3 11 We were able to confidently admit our patients straight to an HDU bed after theatre in the knowledge that this was not associated with any increased risk. Indeed, this policy was probably a major factor in minimizing serious respiratory complications.

Another advantage of a policy of tracheal extubation at the end of the procedure is that it helps reduce pressure on overstretched ICUs, and thus there is also a significant cost-saving.

In conclusion, we have shown that in our experience immediate extubation in theatre at the end of a two-stage oesophagectomy is safe and is associated with low morbidity and mortality, provided the patient has good analgesia with a thoracic epidural. Patients can be transferred from theatre and safely managed on an HDU, avoiding the need for an ICU bed. This has implications for resources and subsequently costs. Other factors that have contributed to our results have been the team approach, in particular the availability of a dedicated upper gastrointestinal surgeon and anaesthetist, active physiotherapy input and good specialized nursing support.


    References
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
1 Begg CB, Cramer LD, Hoskins WJ, Brennan MF. Impact of hospital volume on operative mortality for major cancer surgery. JAMA 1998; 25: 1747–51[CrossRef]

2 Tugrul M, Camci E, Karadeniz H, Senturk M, Pembeci K, Akpir K. Comparison of volume controlled with pressure controlled ventilation during one-lung anaesthesia. Br J Anaesth 1997; 79: 306–10.[Abstract/Free Full Text]

3 Tandon S, Batchelor A, Bullock R, et al. Peri-operative risk factors for acute lung injury after elective oesophagectomy. Br J Anaesth 2001; 86: 633–8[Abstract/Free Full Text]

4 ARDS Network. Ventilation with lower tidal volumes as compared with tradition tidal volumes for acute lung injury and the ARDS. N Engl J Med 2000; 342: 1301–8[Abstract/Free Full Text]

5 Dos Santos CC, Slutsky AS. Invited review: mechanisms of ventilator-induced lung injury—a perspective. J Appl Physiol 2000; 89: 1645–55[Abstract/Free Full Text]

6 Jordan S, Mitchell JA, Quinlin GJ, Goldstraw P, Evans TW. The pathogenesis of the lung injury following pulmonary resection. Eur Respir J 2000; 15: 790–9[Abstract/Free Full Text]

7 Sherry K. Management of patients undergoing oesophagectomy. In: Gray AJG, Hoile RW, Ingram GS, et al., eds. Report of the National Confidential Enquiry into Peri-operative Deaths 1996/1997. London: National CEPOD, 1998: 57–61

8 Millikan KW, Silverstein J, Hart V, et al. A 15 year review of oesophagectomy for carcinoma of the oesophagus and cardia. Arch Surg 1995; 130: 617–24[Abstract]

9 Wyncol DLA, Evans TW. ARDS. Lancet 1999; 354: 497–501[CrossRef][ISI][Medline]

10 Baudoiun S. Improved survival in ARDS: chance, technology or experience? Thorax 1998; 53: 237–8[Free Full Text]

11 Caldwell MT, Murphy PG, Page R, Walsh TN, Henessey TP. Timing of extubation after oesophagectomy. Br J Surg 1993; 80: 1537–9[ISI][Medline]

12 Katsuta T, Saito T, Shigemitsu Y, Kinoshita T, Shiraishi N, Kitano S. Relation between tumour necrosis factor alpha and interleukin 1beta producing capacity of peripheral monocytes and pulmonary complications following oesophagectomy. Br J Surg 1998; 85: 548–53[CrossRef][ISI][Medline]

13 Bartels H, Stein HJ, Siewert JR. Pre-operative risk analysis and post-operative mortality of oesophagectomy for resectable oesophageal cancer. Br J Surg 1998; 85: 840–4[CrossRef][ISI][Medline]

14 Nagawa H, Kobori O, Muto T. Prediction of pulmonary complication after trans thoracic oesophagectomy. Br J Surg 1994; 81: 860–2[ISI][Medline]

15 Muller JM, Erasmi H, Stelzner M, Zieren U, Pichlmaier H. Surgical therapy of oesophageal carcinoma. Br J Surg 1990; 77: 845–57[ISI][Medline]

16 Tsui SL, Law S, Fok M, et al. Postoperative analgesia reduces mortality and morbidity after oesophagectomy. Am J Surg 1997; 173: 472–8[CrossRef][ISI][Medline]

17 Smedstad KG, Beattie WS, Blair WS, Buckley ND. Postoperative pain relief and hospital stay after oesophagectomy Clin J Pain 1992; 8: 149–52[ISI][Medline]

18 Tsui SL, Chan CS, Chan ASH, Wong SJ, Lam CS, Jones RDM. Post-operative analgesia for oesophageal surgery. A comparison of three analgesic regimens. Anaesth Intens Care 1991; 19: 329–37[ISI][Medline]

19 Watson A, Allen PR. Influence of thoracic epidural analgesia on outcome after resection for oesophageal cancer. Surgery 1994; 115: 429–32[ISI][Medline]

20 Terai T, Yukioka H, Fujimori M. Administration of epidural bupivacaine combined with epidural morphine after oesophageal surgery. Surgery 1997; 121: 359–65[CrossRef][ISI][Medline]

21 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: 228–34[Abstract]