Impact of the Covid-19 pandemic on the performance of machine learning algorithms for predicting perioperative mortality.

BACKGROUND: Machine-learning models are susceptible to external influences which can result in performance deterioration. The aim of our study was to elucidate the impact of a sudden shift in covariates, like the one caused by the Covid-19 pandemic, on model performance. METHODS: After ethical approval and registration in Clinical Trials (NCT04092933, initial release 17/09/2019), we developed different models for the prediction of perioperative mortality based on preoperative data: one for the pre-pandemic data period until March 2020, one including data before the pandemic and from the first wave until May 2020, and one that covers the complete period before and during the pandemic until October 2021. We applied XGBoost as well as a Deep Learning neural network (DL). Performance metrics of each model during the different pandemic phases were determined, and XGBoost models were analysed for changes in feature importance. RESULTS: XGBoost and DL provided similar performance on the pre-pandemic data with respect to area under receiver operating characteristic (AUROC, 0.951 vs. 0.942) and area under precision-recall curve (AUPR, 0.144 vs. 0.187). Validation in patient cohorts of the different pandemic waves showed high fluctuations in performance from both AUROC and AUPR for DL, whereas the XGBoost models seemed more stable. Change in variable frequencies with onset of the pandemic were visible in age, ASA score, and the higher proportion of emergency operations, among others. Age consistently showed the highest information gain. Models based on pre-pandemic data performed worse during the first pandemic wave (AUROC 0.914 for XGBoost and DL) whereas models augmented with data from the first wave lacked performance after the first wave (AUROC 0.907 for XGBoost and 0.747 for DL). The deterioration was also visible in AUPR, which worsened by over 50% in both XGBoost and DL in the first phase after re-training. CONCLUSIONS: A sudden shift in data impacts model performance. Re-training the model with updated data may cause degradation in predictive accuracy if the changes are only transient. Too early re-training should therefore be avoided, and close model surveillance is necessary.

[Deep neuromuscular blockade : Benefits and risks]. / Vollrelaxation : Nutzen und Risiken.

Neuromuscular blockade (TOF count = 0) can improve tracheal intubation and microlaryngeal surgery. It is also frequently used in many surgical fields including both nonlaparoscopic and laparoscopic surgery to improve surgical conditions and to prevent sudden muscle contractions. Currently there is a controversy regarding the need and the clinical benefits of deep neuromuscular blockade for different surgical procedures. Deep neuromuscular relaxation improves laparoscopic surgical space conditions only marginally when using low intra-abdominal pressure. There is no outcome-relevant advantage of low compared to higher intra-abdominal pressures, but worsen the surgical conditions. Postoperative, residual curarisation can be avoided by algorithm-based pharmacological reversing and quantitative neuromuscular monitoring.

[Pre-operative documentation of individual in-patient therapy goals : A medical staff questionnaire]. / Patientenorientierte Therapieziele im Krankenhaus : Eine Fragebogenerhebung an medizinischem Personal.

BACKGROUND: Perioperative care demands consideration of individual treatment goals. We evaluated the attitudes of medical staff towards a short standardized advance directive (SSAD) as a means of improving patient-orientated care at the transition from operating theater to general or intensive care wards. METHOD: Multicenter anonymized standardized multiple-choice questionnaire among physicians and nurses from various operative and anesthesiology departments. Questions addressing demographic parameters and attitudes towards advance directives in acute care settings (eleven 4­stepped Likert items). Univariate analysis of group comparisons using the chi-square and Kruskal-Wallis rank-sum test. Multivariable analysis of significant differences employing ordinal logistic regression. RESULTS: The overall return rate was 28.2 % (169 questionnaires). Of these, 19.5 % said that existing advance directives were regularly reassessed preoperatively. SSAD was expected to provide improved emergency care by 82.3 and 76.6 % thought that it would help to better focus intensive care resources according to patients' needs. DISCUSSION: Our study shows the dilemma of insufficiently structured directives for changing treatment goals as well as a high number of legal procedures to obtain proxy decisions due to missing out-patient advance health planning. From a medical staff perspective there is strong support for the concept of SSAD based on medical, ethical, economic and organizational reasons.

Sugammadex and neostigmine dose-finding study for reversal of residual neuromuscular block at a train-of-four ratio of 0.2 (SUNDRO20)†,.

BACKGROUND: The aim of this dose-finding study was to evaluate the dose-response relationship of sugammadex and neostigmine to reverse a commonly observed level of incomplete recovery from rocuronium-induced neuromuscular block, that is, a train-of-four ratio (TOFR) ≥0.2. METHODS: Ninety-nine anaesthetized patients received rocuronium 0.6 mg kg(-1) i.v. for tracheal intubation and, if necessary, incremental doses of 0.1-0.2 mg kg(-1). Neuromuscular monitoring was performed by calibrated electromyography. Once the TOFR recovered to 0.2, patients were randomized to receive sugammadex (0.25, 0.5, 0.75, 1.0, or 1.25 mg kg(-1) i.v.), neostigmine (10, 25, 40, 55, or 70 µg kg(-1) i.v.), or saline (n=9 per group). Primary and secondary end points were the doses necessary to restore neuromuscular function to a TOFR≥0.9 with an upper limit of 5 and 10 min for 95% of patients, respectively. RESULTS: Neostigmine was not able to fulfil the end points. Based on the best-fitting model, the sugammadex dose estimation for recovery to a TOFR≥0.9 for 95% of patients within 5 and 10 min was 0.49 and 0.26 mg kg(-1), respectively. CONCLUSION: A residual neuromuscular block of a TOFR of 0.2 cannot be reversed reliably with neostigmine within 10 min. In the conditions studied, substantially lower doses of sugammadex than the approved dose of 2.0 mg kg(-1) may be sufficient to reverse residual rocuronium-induced neuromuscular block at a recovery of TOFR≥0.2. CLINICAL TRIAL REGISTRATION: NCT01006720.

[Preoperative patient-oriented advance planning of emergency and intensive care treatment--Necessary or imposition? : Questionnaire survey]. / Präoperative patientenorientierte Vorausplanung von Notfall- und Intensivbehandlung--Bedarf oder Zumutung? : Fragebogenerhebung.

BACKGROUND: Traditional advance directives can often not be satisfactorily implemented into patient care; therefore, patient-oriented decision-making prior to scheduled interventions and beyond the actual surgery is of particular importance. Data on inpatient advance care planning (ACP) in Germany are lacking. OBJECTIVES: This proof-of-concept study was carried out to determine the needs of inpatients undergoing surgery for advance preoperative planning of emergency care and to assess potential discomfort caused by such a program. MATERIAL AND METHODS: A voluntary and anonymous standardized questionnaire survey was carried out in scheduled surgery inpatients over 50 years old. Data collection was structured in a demographic part and statements dealing with preoperative advance planning of emergency care in hospital evaluated as Likert items. RESULTS: Out of 579 patients (mean age 66 years, 51% male) 43% indicated a basic interest in being informed about advance planning of emergency care individually during the current hospital stay. Desire for patient self-determination represented an independent factor of information needs [p = 0.036, 95% confidence interval (95% CI) 0.027-0.793]. The survey was perceived as a burden by only 7.3% of patients. This perception was independently associated with less concern about perioperative complication risks (p = 0.008, 95% CI 0.144-0.975). CONCLUSION: The results confirmed a substantial interest in patient-oriented advance planning of emergency care in a preoperative setting; however, no demographic group criteria for patients with information requirements could be defined. As the burden evoked by the topic is low, advance planning of emergency and intensive care treatment of inpatients undergoing surgery should be actively provided in the future.

Effects of single-shot and steady-state propofol anaesthesia on rocuronium dose-response relationship: a randomised trial.

BACKGROUND: Similar to volatile anaesthetics, propofol may influence neuromuscular transmission. We hypothesised that the administration of propofol influenced the potency of rocuronium depending on the duration of the administration. METHODS: After consent, patients scheduled for elective surgery randomly received rocuronium either after induction of anaesthesia with propofol (2 min of propofol, n = 36) or after 30 min of propofol infusion (30 min of propofol, n = 36). Remifentanil was given in both groups. Neuromuscular monitoring was performed by calibrated electromyography. The dose-response relationship of rocuronium was determined with a single-bolus technique (0.07, 0.1, 0.15, 0.2, 0.3 and 0.45 mg/kg rocuronium). The primary endpoints were the ED50 and ED95 of rocuronium after 2 and 30 min propofol. Data are presented as means with (95% confidence interval). The trial is registered with the Eudra-CT: 2009-012815-16. RESULTS: A total of 72 patients were included. Time to maximal neuromuscular blockade was significantly shorter in patients after 30 min of propofol [3.3 min (2.9-3.7)] compared with patients anaesthetised with 2 min of propofol [4.6 min (4.0-5.2)]. After 30 min of propofol, the slope of the dose-response curve was significantly steeper (30 min of propofol: 4.34 [3.62-5.05]; 2 min of propofol: [3.34 (2.72-3.96)], resulting in lower ED95 values of rocuronium (30 min of propofol: 0.287 mg/kg [0.221-0.368]; 2 min of propofol [0.391 mg/kg (0.296-0.520)]. The ED50 were not different between groups. CONCLUSION: The potency of rocuronium was significantly enhanced after propofol infusion for 30 min. Estimates of potency those are usually determined during steady-state anaesthesia might underestimate rocuronium requirements for endotracheal intubation at the time of induction.

Postoperative impairment of motor function at train-of-four ratio ≥0.9 cannot be improved by sugammadex (1 mg kg-1).

BACKGROUND: A train-of-four ratio (TOFR) ≥0.9 measured by quantitative neuromuscular monitoring is accepted as an indication of sufficient neuromuscular recovery for extubation, even though many postsynaptic acetylcholine receptors may still be inhibited. We investigated whether antagonism with sugammadex after spontaneous recovery to TOFR≥0.9 further improves muscle function or subjective well-being. METHODS: Following recovery to TOFR≥0.9 and emergence from anaesthesia, 300 patients randomly received either sugammadex 1.0 mg kg(-1) or placebo. Fine motor function (Purdue Pegboard Test) and maximal voluntary grip strength were measured before and after surgery (before and after test drug administration). At discharge from the postanaesthesia care unit, well-being was assessed with numerical analogue scales and the Quality-of-Recovery Score 40 (QoR-40). RESULTS: Patients' fine motor function [6 (sd 4) vs 15 (3) pegs (30 s)(-1), P<0.05] and maximal voluntary grip strength (284 (126) vs 386 (125) N, P<0.05) were significantly lower after anaesthesia compared with the pre-anaesthesia baseline. After sugammadex or placebo, motor function was significantly improved in both groups but did not reach the preoperative level. There was no difference between groups at any time. Global well-being was unaffected (QoR-40: placebo, 174 vs 185; sugammadex, 175 vs 186, P>0.05). CONCLUSIONS: Antagonizing rocuronium at TOF≥0.9 with sugammadex 1.0 mg kg(-) (1) did not improve patients' motor function or well-being when compared with placebo. Our data support the view that TOFR≥0.9 measured by electromyography signifies sufficient recovery of neuromuscular function. CLINICAL TRIAL REGISTRATION: The trial is registered at ClinicalTrials.gov (NCT01101139).

[Survey on the need for information during the preanesthesia visit]. / Fragebogenstudie zum Informationsbedarf in der Prämedikationsambulanz.

BACKGROUND: The preanesthesia informed consent document is regarded mainly as a legal prerequisite but patient autonomy in the authorization of a proposed intervention requires that the relevant information is provided in a suitable and useful way. AIM: The information needs of patients was determined in relation to demographic parameters. This study carried out to evaluate if the expected extent of information regarding anesthesia during the preanesthesia visit was dependent on group-specific variables. MATERIAL AND METHODS: A total of 699 adult patients with forthcoming elective non-cardiac surgery were anonymously interviewed concerning their expectations and informational needs during the preanesthesiavisit. The questionnaire contained 15 demographic variables, one being the question on health-related quality of life (HRQoL). The ASA classification was the only patient data assessed by the anesthesiologist after the consultation. In the second part of the questionnaire statements regarding the kind and extent of information (n = 10) as well as structural aspects of the preanesthesia visit (n = 5) could be rated using a four-step Likert scale. Point values from questions 1-10 were added to a sum score of need for information for each patient with 0 to ± 3 allotted for each question according to the direction of the question wording (i.e. more or less information desired) and the individual patient scores on the Likert scale. Variables associated with this score of need for information were assessed by regression analysis. RESULTS: Of the patients, 80.6% were classified as American Society of Anesthesiologists (ASA) physical status I and II. The HRQoL was rated fair or good by a total of 80%. On average patients were satisfied with the extent and the kind of information offered during the preanesthesia visit with a mean of the sum score of 0 (min. -10 and max. +10, SD ± 3.2). This applied to the written material to prepare for informed consent; however, the consultation was much more appreciated as a source of information. Of the patients, 278 wanted more information and 268 patients wanted less. Linear regression analysis determined education [p = 0.00018, 95% CI: 0.405 (0.194-0.615)], ASA physical status [(p = 0.047, 95% CI: - 0.558 (- 1.107 to - 0.009)] and HRQoL [(p = 0.025, 95% CI: - 0.412 (- 0.771 to - 0.053)] as being independently related to information needs, including perioperative processes as well as rare risks and complications. Interest in being educated about patient autonomy in end of life situations in the hospital was significantly correlated to the score (p < 0.001, r = 0.143). The results of this study demonstrate for the first time in a German surgical cohort a wide acceptance of preoperative healthcare planning (77.4 %). CONCLUSION: Demographic criteria can help to tailor pre-anesthetic information to individual patient needs. The explanatory power of these variables was, however, low. The relationship between self-assessed HRQoL and the demand for information underlines the necessity to adapt the amount and kind of information provided during the consultation to individual patients preferences.

Plasma oxytocin and vasopressin do not predict neuropeptide concentrations in human cerebrospinal fluid.

The involvement of the neuropeptides oxytocin (OXT) and vasopressin (AVP) in human socio-emotional behaviours is attracting increasing attention. There is ample evidence for elevated plasma levels upon a wide variety of social and emotional stimuli and scenarios, ranging from romantic love via marital distress up to psychopathology, with cause versus consequence being largely unclear. The present study examined whether plasma levels of both OXT and AVP are reflective of central neuropeptide levels, as assumed to impact upon socio-emotional behaviours. Concomitant plasma and cerebrospinal fluid (CSF) samples were taken from 41 non-neurological and nonpsychiatric patients under basal conditions. Although OXT and AVP levels in the CSF exceeded those in plasma, there was no correlation between both compartments, clearly suggesting that plasma OXT and AVP do not predict central neuropeptide concentrations. Thus, the validity of plasma OXT and AVP as potential biomarkers of human behaviour needs further clarification.

The use of sugammadex in a patient with myasthenia gravis.

Myasthenia gravis, affecting neuromuscular transmission, leads to a large variability in sensitivity to depolarising and non-depolarising neuromuscular blocking drugs. We report the successful use of the modified mu-cyclodextrin sugammadex in a myasthenic patient to reverse a rocuronium-induced deep level of neuromuscular block. After spontaneous neuromuscular recovery of T2 (second twitch of the train-of-four series), we administered 2 mg.kg(-1) of sugammadex intravenously, reversing neuromuscular blockade to a train-of-four ratio (T4/T1) > 90% within 210 s. Sugammadex, in combination with objective neuromuscular monitoring, can be used to reverse rocuronium-induced neuromuscular blockade in patients with myasthenia gravis, thereby avoiding the need for reversal with acetylcholinesterase inhibitors.

Effects of sevoflurane on cognitive deficit, motor function, and histopathology after cerebral ischemia in rats.

BACKGROUND: The volatile anesthetic sevoflurane exhibits neuroprotective properties when assessed for motor function and histopathology after cerebral ischemia in rats. Damage of hippocampal neurons after ischemia relates to a number of cognitive deficits that are not revealed by testing animals for motor function. Therefore, the present study evaluates cognitive and behavioral function as well as hippocampal damage in rats subjected to cerebral ischemia under sevoflurane compared with fentanyl/nitrous oxide (N(2)O)/O(2) anesthesia. METHODS: Thirty-four rats were trained for 10 days using a hole-board test to detect changes in cognitive and behavioral function. Rats were randomly assigned to the following groups: (A) sham/fentanyl/N(2)O/O(2) (n=7); (B) ischemia/fentanyl/N(2)O/O(2) (n=10); (C) sham/2.0 vol% sevoflurane in O(2)/air (n=7); and (D) ischemia/2.0 vol% sevoflurane in O(2)/air (n=10). Cerebral ischemia was produced by unilateral common carotid artery occlusion combined with hemorrhagic hypotension (mean arterial blood pressure 40 mmHg for 45 min). Temperature, arterial blood gases, and pH were maintained constant. Cerebral blood flow was measured using laser-Doppler flowmetry. After surgery, cognitive and behavioral function was re-evaluated for 10 days. On day 11, the brains were removed for histopathologic evaluation (hematoxylin/eosin-staining). RESULTS: Cognitive testing revealed deficits in declarative and working memory in ischemic rats anesthetized with fentanyl/N(2)O. Rats anesthetized with sevoflurane during ischemia showed a significantly better outcome. Hippocampal damage was significantly worse with fentanyl/N(2)O. CONCLUSION: The present data add to previous investigations showing that sevoflurane prevents a deficit in cognitive function and histopathological damage induced by cerebral ischemia in rats.

Muscle relaxant use for tracheal intubation in pediatric anaesthesia: a survey of clinical practice in Germany.

OBJECTIVES: Aim of this Germany-wide study was to evaluate the use and application customs of neuromuscular blocking drugs (NMBDs) for tracheal intubation in children of age 5 years or younger. MATERIAL AND METHODS: In the year 2005, a total of 3260 questionnaires were sent out to all heads of anesthesia departments at all types of hospitals as well as ambulatory anesthesia centers in Germany. The fields covered by the questionnaire were regarding institutional size, amount of general anesthesia and specifically pediatric cases, number of intubations and NMBDs used, frequency of use of the individual relaxants and techniques utilized when administering muscle relaxants. RESULTS: Of the sent-out questionnaires 66.9% could be analyzed: 82% of hospitals use 1-3 muscle relaxants in children; 91% of the ambulatory anesthesia centers use a repertoire of 1-2 neuromuscular blockers for pediatric cases. However, general anesthesia with tracheal intubation in children is often induced without using any NMBD at all. Mivacurium is the predominantly used NMBD for pediatric intubation in Germany. In contrast, the use of succinylcholine is far less in ambulatory anesthesia centers than in hospitals. Despite controversial discussion, precurarization, priming, and timing are still utilized in German anesthesia practice. CONCLUSION: In Germany, mivacurium, with its favorable pharmacologic profile for short cases, is the predominantly used NMBD for pediatric tracheal intubation. Despite the known adverse effects of intubation without muscle relaxation, this technique is also wide-spread, especially among German anesthetists in ambulatory anesthesia centers. Surveys like these are important to determine a status quo of use and application customs of NMBDs in pediatric anesthesia and provide a basis for numerous other studies.

[Application of neuromuscular monitoring in Germany]. / Einsatz des neuromuskulären Monitorings in Deutschland.

The aim of the present study was to evaluate application customs of neuromuscular monitoring in hospitals and private practice. Of the 3,260 questionnaires sent out, 2,182 could be analyzed. Of these 54% were from anaesthetists in private practice, 41% from heads of hospital anaesthesia departments and 5% from heads of level 1 hospital anaesthesia departments. In 12.1% of the hospital departments and 66.7% of private practices, no neuromuscular monitoring was available at all. In both hospital departments and private practices, clinical signs were the most often applied criteria for timing of reinjection of myorelaxants as well as for evaluation of neuromuscular recovery.

Onset properties of mivacurium measured by mechanomyography depend on the twitch height of the adductor pollicis muscle.

INTRODUCTION: The influence of the twitch height of the adductor pollicis muscle during baseline measurements on the pharmacodynamic parameters of mivacurium was prospectively evaluated. PATIENTS AND METHODS: Fifty adult patients were anaesthetized with propofol and alfentanil. Neuromuscular function was monitored mechanomyographically by measuring the force of the adductor pollicis muscle following stimulation of the ulnar nerve. Following a stabilization period of 20 min, the individual twitch height of the adductor pollicis muscle was determined before a single bolus of mivacurium (75 microg kg-1) was administered. Patients were divided into two groups. The data of patients whose thumb adduction force was below the median value of all patients were the 'low force' group (9.1 +/- 1.4 N) and the data of all other patients were the 'high force' group (13.7 +/- 1.8 N). RESULTS: In the 'high force' group, maximum neuromuscular blockade of mivacurium was deeper (0.97 +/- 0.05 vs. 0.93 +/- 0.06; P < 0.05) and onset faster (2.9 +/- 1.1 min vs. 4.0 +/- 1.2 min; P < 0.05). Neuromuscular recovery did not differ between the groups. CONCLUSION: The different onset speeds reflect either different sensitivity to neuromuscular blocking agents with respect to patients' muscle power or a problem of the mechanomyographic measuring technique.

Incidence and duration of residual paralysis at the end of surgery after multiple administrations of cisatracurium and rocuronium.

In a randomised, controlled, double-blind, multicentre trial in 338 patients, we assessed the incidence of residual paralysis following administration of cisatracurium or rocuronium. The incidence at the end of surgery was significantly lower in patients treated with rocuronium (62 of 142 patients, 44%) than in those given cisatracurium (99 of 175 patients, 57%) (p < 0.05). In contrast, with rocuronium the mean (SD) time between skin closure and extubation was 28 (28) min vs 18 (19) min for cisatracurium, and the duration 0.9 (time from administration of last top-up dose to recovery of the train-of-four ratio to 0.9) was significantly longer and more variable for rocuronium than for cisatracurium. Thus, after repeated administration, the duration and variability of duration of action are greater with rocuronium compared with cisatracurium. These pharmacodynamic differences do not necessarily translate into a higher incidence of residual paralysis, because clinicians compensate for the longer duration of action and variability of rocuronium by terminating administration of the neuromuscular blocking earlier.

[Transport of blood gas samples: is the pneumatic tube system safe?]. / Transport von Blutgasproben: Ist die Rohrpost sicher?

BACKGROUND: Transport of blood gas samples via a pneumatic tube system and subsequent analysis in the central laboratory can reduce costs and errors compared to on-site testing in the operating theatre or the intensive care unit. In this study, a modern pneumatic tube transport system was tested for its usability for this purpose. METHODS: A total of 4 consecutive blood gas samples were obtained intraoperatively from 54 different patients and sent to the central laboratory. Of these, 3 samples were transferred using the pneumatic tube system but by different methods and 1 sample was transported personally which served as a reference. The results of sample analysis concerning blood gases, electrolytes and haemoglobin were compared and examined for differences. RESULTS: No statistically significant differences could be determined between the different modes of transportation. CONCLUSION: Transport of samples for blood gas analysis via a modern pneumatic tube system is safe when samples are correctly prepared.

[Muscle relaxants in Germany 2005: a comparison of application customs in hospitals and private practices]. / Muskelrelaxanzien in Deutschland 2005: Ein Vergleich zwischen den Anwendungsgewohnheiten in Krankenhäusern und Praxen.

Aim of this study was to evaluate application customs of muscles relaxants in hospitals compared to their use in private practice. Of the 3,260 questionnaires sent-out, 66.9% could be analyzed. Of these 54% were from anesthetists in private practice, 41% from heads of hospital anesthesia departments and 5% from heads of level one hospital anesthesia departments. The first difference between private practices and hospitals was the number of available muscle relaxants: 87% of private practices use 1-3 relaxants, whereas 79% of hospitals use 3-5. Another apparent difference was the relationship between general anesthesia and the number of intubations: 60% of private practices have over 80% of general anesthesia cases, but only 50% of these patients are intubated. On the contrary, two thirds of the hospitals have 50-80% general anesthesia cases and 60-70% of patients are intubated. The main wish for an ideal muscle relaxant was independent of private practice or hospital, short onset time, followed by fast recovery. In accordance 74% of anesthetists in hospitals and 72% of anesthetists in private practice voiced the wish for a non-depolarizing succinylcholine substitute. The results of this nationwide survey suggest that time pressure in combination with an increased specialization of anesthetists in private practice are the main factors for availability and use of muscle relaxants in routine anesthesia.

Sevoflurane and propofol influence the expression of apoptosis-regulating proteins after cerebral ischaemia and reperfusion in rats.

BACKGROUND AND OBJECTIVE: Sevoflurane and propofol reduce the extent of necrosis and improve neurological outcome in rodent models of cerebral ischaemia and reperfusion. However, the effects of these anaesthetics on programmed cell death (apoptosis) are unclear. The present study investigates whether sevoflurane and propofol affect the expression of apoptosis-regulating proteins after cerebral ischaemia in rats. METHODS: Thirty-two fasted male Sprague-Dawley rats were tracheally intubated and the lungs were ventilated (isoflurane and N2O/O2 anaesthesia). After surgical preparation, the animals were randomly assigned to one of the following groups: control (n = 8): fentanyl intravenous (10 microg kg(-1) bolus and 25 microg kg(-1) h(-1) infusion) with N2O/O2; sevoflurane (n = 8): 2.0% sevoflurane (end-tidal concentration) and O2/air; propofol (n = 8): 0.8-1.0 mg kg(-1) min(-1) propofol intravenous and O2/air; sham-operated (n = 8): 25 microg kg(-1) h(-1) fentanyl intravenous and N2O/O2, no cerebral ischaemia. Ischaemia (30 min) was induced by unilateral common carotid artery occlusion plus haemorrhagic hypotension to a mean arterial pressure of 30-35 mmHg. Four hours after cerebral ischaemia the brains were removed and the expression of apoptosis-regulating proteins (Bax, Bcl-2, p53, Mdm-2) was determined using immunofluorescence and Western-blot analyses. RESULTS: The expression of the pro-apoptotic protein Bax was greater in control animals than in sevoflurane or propofol anaesthetized rats and than in sham-operated animals. The concentrations of Bcl-2, p53 and Mdm-2 were not changed 4 h after cerebral ischaemia. CONCLUSIONS: In addition to the anti-necrotic effects of sevoflurane and propofol, these anaesthetics also reduce the concentration of the apoptosis-inducing protein Bax as early as 4 h after ischaemia.