Search

BackgroundThe proportion of avoidable hospital deaths is challenging to estimate, but has great implications for quality improvement and health policy. Many studies and monitoring tools are based on selected high-risk populations, which may overestimate the proportion. Mandatory reporting systems, however, under-report. We hypothesise that a review of an unselected sample of hospital deaths will provide an estimate of avoidability in-between the estimates from these methods.MethodsA retrospective case record review of an unselected population of 1000 consecutive non-psychiatric hospital deaths in a Norwegian hospital trust was conducted. Reviewers evaluated to what degree each death could have been avoided, and identified problems in care.ResultsWe found 42 (4.2%) of deaths to be at least probably avoidable (more than 50% chance of avoidability). Life expectancy was shortened by at least 1 year among 34 of the 42 patients with an avoidable death. Patients whose death was found to be avoidable were less functionally dependent compared with patients in the non-avoidable death group. The surgical department had the greatest proportion of such deaths. Very few of the avoidable deaths were reported to the hospital’s report system.ConclusionsAvoidable hospital deaths occur less frequently than estimated by the national monitoring tool, but much more frequently than reported through mandatory reporting systems. Regular reviews of an unselected sample of hospital deaths are likely to provide a better estimate of the proportion of avoidable deaths than the current methods.

Quality aspects of the anesthetic process are reflected in the rate of intraoperative adverse events. The purpose of this report is to illustrate how the quality of the anesthesia process can be analyzed using statistical process control methods, and exemplify how this analysis can be used for quality improvement.We prospectively recorded anesthesia-related data from all anesthetics for five years. The data included intraoperative adverse events, which were graded into four levels, according to severity. We selected four adverse events, representing important quality and safety aspects, for statistical process control analysis. These were: inadequate regional anesthesia, difficult emergence from general anesthesia, intubation difficulties and drug errors. We analyzed the underlying process using 'p-charts' for statistical process control.In 65,170 anesthetics we recorded adverse events in 18.3%; mostly of lesser severity. Control charts were used to define statistically the predictable normal variation in problem rate, and then used as a basis for analysis of the selected problems with the following results: Inadequate plexus anesthesia: stable process, but unacceptably high failure rate; Difficult emergence: unstable process, because of quality improvement efforts; Intubation difficulties: stable process, rate acceptable; Medication errors: methodology not suited because of low rate of errors.By applying statistical process control methods to the analysis of adverse events, we have exemplified how this allows us to determine if a process is stable, whether an intervention is required, and if quality improvement efforts have the desired effect.

Since the first version saw the light of day in 1991 the Guidelines have occupied a central position in the Norwegian practice of anesthesia. This document comprises part of the quality management documents held in the departments of anesthesia in Norwegian hospitals. If departments of anesthesia are unable to adhere to certain specific points in the Guidelines, it is recommended that this should be documented in writing. It has been stated by central governmental bodies and patients' insurance organizations that the Guidelines will be an important factor in medico legal cases, although it is not an obligatory legal document for hospital owners. It is our objective that the document will form the foundation for quality assurance work in the departments of anesthesia in Norway. The purpose of this document is to ensure a satisfactory standard for the practice of anesthesia in Norway. 'The Guidelines for the Practice of Anesthesia in Norway' (the Guidelines) is a series of recommended guidelines. The practice of anesthesia in this context includes general anesthesia, regional anesthesia, controlled sedation, postoperative monitoring, and other observations where anesthesia personnel are required. The Guidelines apply to all doctors, nurses, and other personnel undertaking the delegated practice of anesthesia. Deviations from the Guidelines should be explained and documented in every case. The Guidelines should be adhered to in medical emergencies as far as possible. The Guidelines must not be allowed to prevent the execution of immediate and lifesaving measures. The Guidelines should be revised at regular intervals so that it is up-to-date with current legislation and medical and technological developments and practice.

The low incidence of mortality and major morbidity in anesthesia makes it difficult to study the pattern of potential accidents and to develop preventive strategies. Anesthetic 'near-misses', however, occur more frequently. Using data from a simple routine-based system of problem reporting, we have analyzed the pattern and causes of serious non-fatal problems, in order to improve preventive strategies.We prospectively recorded anesthesia-related information from all anesthetics for five years. The data included intraoperative problems, which were graded into four levels, according to severity. We analyzed only the serious nonfatal problems, which were sorted according to clinical presentation, and also according to which factor was most important in the development of the problem. We assessed any untoward consequences for the patient, and whether the problems could have been prevented.Serious problems were recorded in 315 cases out of 83,844 (0.4%). Anesthesia was considered the major contributing factor in 111 cases. Difficult intubation, difficult emergence from general anesthesia, allergic reactions, arrhythmia and hypotension were the dominating problems. Twenty-six anesthesia related problems resulted in changes in level of postoperative care, and one patient later died in the intensive care unit after anaphylactic shock. Eighty-two problems could have been prevented by simple strategies.Analysis of serious nonfatal problems during anesthesia may contribute to improved preventive strategies. Data from a routine-based system are suitable for this type of analysis. Intubation, emergence, arrhythmia, hypotension and anaphylaxis cause most serious problems, and should be the object of preventive strategies.

To describe the frequency and pattern of drug errors in clinical anesthesia, and to evaluate whether a change to colour coded syringe labels, along with education, could reduce the problem of drug errors.We prospectively recorded anesthesia-related information from all anesthetic cases for 36 mo, totally 55,426 procedures. Intraoperative problems, including drug errors, were recorded. After eighteen months we changed to colour coded syringe labels, and the effect of this change and education on drug errors was assessed. Errors were divided into four groups: syringe swap, ampoule swap, other 'wrong drug' errors, and wrong dose errors. The problems were graded into four levels, according to severity.A drug error was recorded in 63 cases (0.11%). There were 28 syringe swaps, and muscle relaxants were erroneously given in 15. There were nine ampoule swaps. There were eight 'other wrong drug' cases, and 18 cases where a wrong dose of the correct drug was given. Three of the drug errors were classified as serious, and 27 were of moderate severity. We found no differences between the two periods except for decreased number of ampoule swaps (P = 0.04).Drug errors are uncommon, and represent a small part of anesthesia problems but still have the potential for serious morbidity. Syringe swaps occurred most often between syringes of equal size, and were not eliminated by colour coding of labels. As muscle relaxant drugs are most commonly involved, and can cause lasting morbidity, special preventive measures should be taken for this group of drugs.

The Trondheim region's (315 km2, population 154,000) emergency medical service (EMS) provides advanced cardiac life support (ACLS) with combined paramedic and physician response. This EMS system is commonly employed in Norway, yet no population based study of outcome in cardiac arrest has been published to date. This retrospective study reports incidence and outcome from every attempted out-of-hospital cardiopulmonary resuscitation (CPR) during 1990 through 1994 according to the Utstein template. Information on the patient's pre-morbid conditions and final outcome was obtained from hospital records. The incidence of cardiac arrest and CPR from all causes was 68 per 100,000 per year, with 83% primary cardiac aetiology. The median alarm to patient arrival interval for ambulance and emergency physician was 8 minutes and 11 minutes, respectively. The presenting rhythm was ventricular fibrillation or tachycardia in 51%, asystole in 34%, pulseless electrical activity in 8% and undetermined in 8%. Definite return of spontaneous circulation occurred in 211 patients (40%, 27 per 100,000 per year) and 57 patients (11%, 7.4 per 100,000 per year) survived to discharge. Most patients made a favourable cerebral outcome, although nine were severely disabled. This is the first population-based Norwegian study of outcome from out-of-hospital cardiac arrest in this combined paramedic/physician staffed EMS. Incidence, survival and neurological outcome are comparable with results obtained in other EMS systems.

The long QT syndrome (LQTS) is associated with syncopal attacks or even sudden death at a young age due to ventricular fibrillation. We report a patient with an undiagnosed LQTS who had an episode of cardiac arrest during the final part of general anesthesia, immediately after the drugs for reversal of the neuromuscular blockade were given. We suggest that the administration of glycopyrronium might have been the provoking factor in this patient.

Direct comparison of survival rates from in-hospital cardiopulmonary resuscitation (CPR) remains difficult. The objective of this study was to report outcome according to the Utstein template for in-hospital cardiac arrest and to evaluate the Utstein template itself as applied to a retrospective material.The hospital (900 beds, 37,000 annual admissions) has no established do-not-resuscitate (DNR) order policy. CPR outside the Intensive- or Coronary Care Units (ICU/CCU) is performed by an emergency medical team consisting of an anaesthesiologist, a medical resident and a nurse anaesthetist. CPR attempts during 5 years (1990-1994) were analysed retrospectively. Patient survival, cerebral and overall performance category (CPC/OPC) score of the survivors was determined. The Utstein template was evaluated in terms of clinical relevance and data availability.During 5 years, 4927 patients died as in-patients. CPR outside the CCU/ICU was attempted 244 times. CPR was primarily successful on 83 occasions (34%), and 42 patients (17%) were finally discharged with CPC 1 or 2. Survival from primary ventricular fibrillation (VF) or ventricular tachycardia was 40%, pulseless electrical activity 3%, asystole 11% and of rhythm undetermined 6%. Age or sex effects were not observed.More than 90% of in-hospital deaths in this hospital are handled without CPR being initiated. Overall survival was 17%, and almost all survivors made a favourable outcome. The Utstein template for in-hospital cardiac arrest performed acceptably as a framework for reporting outcome in this retrospective study.

Background: The aim of this study was to determine the incidence and severity of pulmonary aspiration of gastric contents during anaesthesia, to determine the short- and long-term morbidity, and to evaluate present routines for preoperative gastric emptying. During the study period, preoperative gastric emptying was done only when intestinal obstruction was suspected. Methods: We routinely record prospectively all problems during and after anaesthesia by means of a database. All data for the 5 years from 1989 to 1993, a total of 85 594 anaesthetic procedures, were analyzed. The hospital charts were also reviewed for those patients where aspiration to the lungs had occurred. Results: Pulmonary aspiration of gastric contents was detected in 25 cases; all occurred in patients receiving general anaesthesia. The incidence was 4.1 times higher in emergency procedures than in electives. There were no aspirations in 30 199 patients receiving regional anaesthesia. The complication occurred in all phases of anaesthesia, but clinical morbidity was low in most cases. Three cases showed serious morbidity immediately after the event, but recovered. Two cases showed serious long-term morbidity, but also recovered completely. No patients died. No cases, except possibly one, might have been prevented by stricter routines for preoperative gastric emptying. Conclusion: We found a low incidence of pulmonary aspiration. When it occurs, it carries a low risk for serious morbidity. Emergency cases for general anaesthesia are most at risk. Regional anaesthesia is considered safe. There is no evidence that preoperative gastric emptying should be routinely done in emergency cases, except in patients with suspected ileus/subileus.

Both the period of total circulatory arrest to the brain and postischemic-anoxic encephalopathy (cerebral postresuscitation syndrome or disease), after normothermic cardiac arrests of between 5 and 20 mins (no-flow), contribute to complex physiologic and chemical derangements. The best documented derangements include the delayed protracted inhomogeneous cerebral hypoperfusion (despite controlled normotension), excitotoxicity as an explanation for selectively vulnerable brain regions and neurons, and free radical-triggered chemical cascades to lipid peroxidation of membranes. Protracted hypoxemia without cardiac arrest (e.g., very high altitude) can cause angiogenesis; the trigger of it, which lyses basement membranes, might be a factor in post-cardiac arrest encephalopathy. Questions to be explored include: What are the changes and effects on outcome of neurotransmitters (other than glutamate), of catecholamines, of vascular changes (microinfarcts seen after asphyxia), osmotic gradients, free-radical reactions, DNA cleavage, and transient extracerebral organ malfunction? For future mechanism-oriented studies of the brain after cardiac arrest and innovative cardiopulmonary-cerebral resuscitation, increasingly reproducible outcome models of temporary global brain ischemia in rats and dogs are now available. Disagreements exist between experienced investigative groups on the most informative method for quantitative evaluation of morphologic brain damage. There is agreement on the desirability of using not only functional deficit and chemical changes, but also morphologic damage as end points.

Objectives: A possible influence of thoracic epidural analgesia on coronary hemodynamics and myocardial metabolism in coronary artery bypass grafting was investigated. Design: The study was prospective and randomized. Setting: The study was performed in a university hospital. Participants: Thirty male patients less than 65 years of age and with ejection fraction greater than 0.5 participated. They were randomized into 3 groups: the high-fentanyl (HF) group receiving high-dose fentanyl (55 μg/kg) anesthesia, the HF + thoracic epidural analgesia (TEA) group receiving the same general anesthesia plus thoracic epidural analgesia, and the low-fentanyl (LF) + TEA group receiving low-dose fentanyl (15 μg/kg) anesthesia plus thoracic epidural analgesia. Interventions: A thoracic epidural catheter, a peripheral and central venous catheter, a radial artery catheter, a thermodilution pulmonary artery catheter, and a coronary sinus reverse thermodilution catheter were inserted. Measurements and Main Results: Coronary circulatory parameters, myocardial oxygenation, and myocardial substrate utilization were investigated before bypass and for 9 hours after bypass. Before bypass, the most striking finding was a reduction in myocardial lactate extraction in all groups, but also coronary flow and myocardial oxygen consumption decreased compared with baseline. After bypass, the only significant finding was a lower coronary vascular resistance early postoperatively in the epidural groups, but coronary blood flow was adequate in all groups. Myocardial metabolism was essentially unchanged both with and without epidural analgesia after bypass. Conclusion: With regard to the coronary circulation and myocardial metabolism, no hard data supporting the use of thoracic epidural analgesia in coronary artery bypass grafting were found.

Thoracic epidural analgesia (TEA) may offer haemodynamic benefits for patients with coronary heart disease going through major surgery. This may-in part-be secondary to an effect on the endocrine and metabolic response to surgery. We therefore investigated the effect of TEA on the endocrine metabolic response to aortocoronary bypass surgery (ACBS). Thirty male patients (age65 years, ejection fraction0.5) were randomized into 3 groups; the HF group receiving a high dose fentanyl (55 micrograms.kg-1) anaesthesia, the HF+TEA group with the same fentanyl dose+TEA with 10 ml bupivacain 5 mg.ml-1, followed by 4 ml every hour, and the LF+TEA group receiving fentanyl 15 micrograms.kg-1 + TEA. Adrenalin, noradrenalin, systemic vascular resistance (SVR), glucose, cortisol, lactate and free fatty acids were followed during the operation and for 20 h postoperatively. A significant increase in adrenalin, noradrenalin and SVR was found in the HF group whereas this increase was blocked in both epidural groups. An increase in glucose and cortisol was noticed in all groups, but the increase was delayed in the epidural groups. Our results suggest that a more effective blockade of the stress response during ACBS is obtained when TEA is added to general anaesthesia than with high dose fentanyl anaesthesia alone.

There is an increasing demand for systems for measuring the quality of our medical work. In this article, we present a suggestion for how one can attempt to do this in a department of anaesthesia. It would be desirable to measure real clinical outcomes such as morbidity and mortality. However, such events are rare and not suitable for routine recording of work quality. Instead, we propose a system based on indicators of process quality and surrogate clinical outcomes. Surrogates may provide useful information if chosen carefully and checked for validity. We further suggest that such indicators be recorded routinely on every anaesthesia chart. The rate of the indicator can then be followed over time with the use of statistical process control methods. The foundation for such a system for measuring quality is the presence of a good quality culture in the department, with a good team spirit, communication and cooperation.

Anesthesiologists and nurse anaesthesists have important roles in the management of in-hospital medical emergencies, and in trauma management. We investigated the extent of these services provided by the department of anaesthesiology at St. Olav University Hospital in Trondheim in the year 2002.Missions involving emergency medical assistance were identified by a computer search in the department's database. We assessed medical aspects, place, time, and type of intervention.We identified a total of 646 missions, of which 501 occurred in the emergency department and 145 on the general wards. The majority (64%) occurred during on-call hours (4 pm-8 am). We found that as many as 4 to 7 missions occurred on any single day 40 times (i.e. days) during the year; which fits in with a Poisson statistical model. Trauma (50 %) and cardiac arrest (20%) dominated the material. Other medical emergencies included seizures, cerebrovascular events, intoxication, respiratory arrest, cardiac arrhythmias, internal bleeding, pulmonary oedema and loss of consciousness. Airway management was important, as 61 patients received bag-valve-mask ventilation and 164 were intubated and ventilated.Staffing of anaesthesiology departments must take into account the possibility of sudden medical emergencies.

Background Anaesthesia equipment problems may contribute to anaesthetic morbidity and mortality. The magnitude and pattern of these problems are not established. We wanted to analyse the frequency, type and severity of equipment-related problems in our department, and if additional efforts to improve safety were needed. Methods The study is based on a system in which anaesthesia-related data are recorded from all anaesthetic cases on a routine basis. The data include intraoperative problems and their severity. When a problem occurs, the anaesthetist responsible for the case writes a short description of the event on the anaesthetic chart. From all recorded cases of general and regional anaesthesia, we selected cases recorded with anaesthetic ‘equipment/technical problems’. These charts were retrieved from departmental archives for analysis. Results From 83 154 anaesthetics, we found the frequency of anaesthetic equipment problems to be 0.05% during regional anaesthesia, and 0.23% during general anaesthesia. One-third of problems involved the anaesthesia machine, and in a quarter, human error was involved. No patient died and none suffered any lasting morbidity. Conclusion The rate of equipment problems was low, and most often of low severity. Aside from improvements in routines for preoperative equipment checks, no specific strategies for problem reduction could be suggested. The incidence of equipment problems is not a good quality indicator because of the low rate of occurrence. However, recorded equipment problems may be useful for improving quality, by analysing causative factors, and suggesting preventative strategies.

The effect of thiopental on glutamate metabolism was studied by 13C magnetic resonance spectroscopy. Cerebral cortical astrocytes were incubated with 0.5 mM [U-13C]glutamate for 2 hr in the presence of 0.5 or 1 mM thiopental. Labeled glutamate, glutamine, aspartate, and glutathione were observed in cell extracts, and glutamine, aspartate, and lactate in the medium. Not only present in the medium was uniformly labeled glutamate, but also glutamate derived from the tricarboxylic acid (TCA) cycle, and thus glutamate release could be detected. The amounts of [U-13C]glutamate and unlabeled glucose taken up by astrocytes were unchanged in the presence of 0.5 mM thiopental and decreased to about 50% and 80%, respectively when the concentration was increased to 1 mM. The amounts of most metabolites synthesized from [U-13C]glutamate were unchanged in the presence of 0.5 mM thiopental, but decreased [U-13C]glutamine, [U-13C]aspartate, and [U-13C]lactate were observed in the 1 mM group. Surprisingly, the amounts of [1,2,3-13C]glutamate, [2,3-13C]aspartate, and [3,4-13C]aspartate (2nd turn via the TCA cycle) were unchanged. However, this was not the case for [1,2-13C]lactate and [2,3-13C]lactate. Such variations indicate cellular compartmentation, possibly caused by a heterogeneous glutamate concentration within the cells affecting TCA cycle turnover rates differently.

The topics discussed in this session include a partial review of laboratory and clinical studies examining the effects of adrenergic agonists on restoration of spontaneous circulation after cardiac arrest, the effects of varying doses of epinephrine, and the effects of novel vasopressors, buffer agents (NaHCO 3 , THAM, ‘Carbicarb’) and anti-arrhythmics (lidocaine, bretylium, amiodarone) in refractory ventricular fibrillation. Novel therapeutic approaches include titrating electric countershocks against electrocardiographic power spectra and of preceding the first countershocks with single or multiple drug treatments. These approaches need to be investigated further in controlled animal and patient studies. Epidemiologic data from randomized clinical outcome studies can give clues, but cannot document pharmacologic mechanisms in the dynamically changing events during attempts to achieve restoration of spontaneous circulation from prolonged cardiac arrest. Also, rapid drug administration by the intraosseous route was compared with intratracheal and intravenous (i.v.) drug administration. Many studies on the above treatments have yielded conflicting results because of differences between healthy hearts of animals and sick hearts of patients, differences in arrest (no-flow) times and cardiopulmonary resuscitation (CPR) (low-flow) times, different pharmacokinetics, different dose/response requirements, and different timing of drug administration during low-flow CPR versus during spontaneous circulation. The need to stabilize normotension and prevent rearrest by titrated novel drug administration, once spontaneous circulation has been restored, requires research. Most of the above topics require some re-evaluation in clinically realistic animal models and in cardiac arrest patients, especially by titration of old and new drug treatments against variables that can be monitored continuously during resuscitation.

Background: As quality assurance, risk analysis and documentation become more important, rapid access to accurate information of departmental activity is needed. In this paper we describe a system for routine data recording of anaesthesia-related information, and we give examples of how this information can be used. Methods: For each anaesthesia procedure we record 20 variables in a database. The variables are related to the patient, the procedure, and the anaesthesia. Both the type and severity of intraoperative problems are recorded. All information is routinely written on the standard anaesthetic chart, and later entered into a database. All charts are quality checked before data entry. Results: With small modifications, we have used this system for 10 years, and it is well integrated as part of our departmental routine. From 1985 to 1995 we have recorded an increased incidence of ‘minor’ intraoperative problems, while the increase in ‘severe’ problems has been small. During the period 1993–95 the total problem incidence has been stable at around 14%. We have found the incidence of problems related to technical equipment to be 0.14%, and of ‘minor’ severity, and therefore equipment failure has not warranted extensive quality assurance efforts. We also investigated the impact of our routines for gastric emptying of non-fasting patients, and found an incidence of pulmonary aspiration of 0.03%. Our results made us conclude that our recommendations concerning anaesthetic technique and gastric emtying are safe. The system has proven to be simple enough to be practicable, and yet detailed enough to be informative for educational, quality assurance, research and administrative purposes. Conclusion: A simple database for information related to anaesthetic care and problems can be a useful tool for anaesthesiology departments. We have used the system for a variety of quality assurance efforts, for educational and administrative purposes and in research. To get reliable data it is important to keep the system simple, with the information recorded kept to a minimum. The software chosen must make it possible for people within the department to write and modify reports. A constant focus on data quality is important, and motivated nurses and doctors as well as leadership enthusiasm are necessary to make the system work properly. The atmosphere of problem recording should be one of ‘openness and confidence’, rather than ‘inspection and control’. Conclusions: We conclude that development and use of a simple database related to anaesthesia care and problems are possible in a busy clinical setting. Our database is now an integrated part of our departmental routine, and plays an important part in quality assurance strategies, educational purposes, research activities and administrative matters. However, the system is not self running. It needs close management and follow up. In addition, leadership enthusiasm and allocation of resources are necessary to make the system work, particularly to ensure the quality of the data. Even in a simple system such as ours, with minimal data and great efforts being made to make data entry simple, our experience is that doctors and nurses need reminding and encouragement to make the system work properly. Most of this follow-up education is carried out during monthly problem meetings, when specific patients and group of patients are discussed. Motivation is best achieved by getting relevant information back to the suppliers, to make them feel that their efforts are worthwhile. The problem recording is a cornerstone of our system, and is a motivation factor for the whole system. This is an important aspect, as relevant and interesting feedback to the data suppliers possibly is an effective way to increase the quality of information in the whole system. The key words to make our system work as a routine in a clinical setting have been ‘simplicity’ for the data system, and ‘openness’ and ‘confidence’ for the recording of problems.

A substantial reduction in lung volumes and pulmonary function follows cardiac surgery. Pain may prevent effective breathing and coughing, and as thoracic epidural analgesia may reduce postoperative pain, we investigated the effect of epidural analgesia on pulmonary function.Fifty-four male patients, under 65 years and with an ejection fraction of more than 0.5, were randomized into two groups: a control group receiving high-dose fentanyl anaesthesia and an epidural group receiving low-dose fentanyl anaesthesia + thoracic epidural analgesia. Time to awakening and time to extubation were recorded. Further, spirometric data, arterial oxygenation, pulmonary shunt, haemodynamics, use of vasoactive drugs and fluid balance were followed for up to 6 days postoperatively.Patients with low-dose fentanyl and epidural analgesia awoke (1.6 vs 3.6 h) and were extubated (5.4 vs 10.8 h) significantly earlier than control group patients. A 50-70% reduction in forced vital capacity, forced expiratory volume in 1 s (FEV1.0) and peak expiratory flow rate (PEFR) was seen after surgery, but higher FEV1.0 and PEFR on days 2 and 3 were seen in the epidural group than in the control group. Pulmonary shunt and alveolo-arterial oxygen difference increased similarly in both groups, whereas oxygen delivery and mixed venous oxygen saturation were higher in the epidural group. Epidural analgesia gave better control of the postoperative hyperdynamic circulation.Thoracic epidural analgesia yields a slight, but significant, improvement in pulmonary function, most likely due to a more profound postoperative analgesia.

Tachycardia and hypertension may cause myocardial ischaemia in patients with coronary heart disease going through major surgery. Thoracic epidural analgesia (TEA) has been reported to be beneficial in this situation. The haemodynamic effects of TEA in aortocoronary bypass surgery were investigated in 30 male patients 0.5. They were randomized into 3 groups: the high dose fentanyl (HF) group receiving high–dose fentanyl (55 μg kg-1) anaesthesia, the HF + TEA group receiving the same fentanyl dose + TEA with 10 ml bupivacaine 5 mg ml-1 followed by 4 ml every hour, and the low dose fentanyl (LF) + TEA group receiving low–dose fentanyl (15 μg kg-1) anaesthesia + TEA. Haemodynamic parameters, the use of vasoactive and inotropic drugs and fluid balance were followed during the operation and for 20 h postoperatively. Before bypass the only significant difference between groups was a higher mean pulmonary arterial pressure in the HF + TEA group and a lower systemic vascular resistance (SVR) in the LF + TEA group, both compared to the HF group. 89% of epidural group patients needed small doses of ephedrine whereas more HF group patients were given nitroglycerine. During bypass SVR and mean arterial pressure (MAP) were significantly higher and pump flow lower in the HF group compared to the LF + TEA group. More ketanserin to HF group patients and methoxamine to epidural group patients were given. After bypass heart rate increased in all groups. Lower MAP 0.5 h after bypass and higher filling pressures in the early post bypass period in the epidural groups, most pronounced in the HF + TEA group, were noted. More propranolol, nitroglycerine and sodium nitroprusside were given to patients in the HF group. No difference in intraoperative fluid balance between groups was seen. Our results suggest more easy control of perioperative hypertension and tachycardia with TEA. This was achieved without haemodynamic side effects of clinical importance.