Implementation of anesthesia quality indicators in Germany : A prospective, national, multicenter quality improvement study.

BACKGROUND: In 2016 the German Society of Anesthesiology and Intensive Care Medicine (DGAI) and the Association of German Anesthetists (BDA) published 10 quality indicators (QI) to compare and improve the quality of anesthesia care in Germany. So far, there is no evidence for the feasibility of implementation of these QI in hospitals. OBJECTIVE: This study tested the hypothesis that the implementation of the 10 QI is feasible in German hospitals. MATERIAL AND METHODS: This prospective three-phase national multicenter quality improvement study was conducted in 15 German hospitals and 1 outpatient anesthesia center from March 2017 to February 2018. The trial consisted of an initial evaluation of pre-existing structures and processes by the heads of the participating anesthesia departments, followed by a 6-month implementation phase of the QI as well as a final re-evaluation phase. The implementation procedure was supported by web-based implementation aids ( www.qi-an.org ) and internal quality management programs. The primary endpoint was the difference in the number of implemented QI per center before and after implementation. Secondary endpoints were the number of newly implemented QI per center, the overall number of successful implementations of each QI, the identification of problems during the implementation as well as the kind of impediments preventing the QI implementation. RESULTS: The average number of implemented QI increased from 5.8 to 6.8 (mean of the differences 1.1 ± 1.3; P < 0.01). Most frequently the QI perioperative morbidity and mortality report (5 centers) and the QI temperature management (4 centers) could be implemented. After the implementation phase, the QI incidence management and patient blood management were implemented in all 16 centers. Implementation of other quality indicators failed mainly due to a lack of time and lack of structural resources. CONCLUSION: In this study the implementation of QI was proven to be mostly feasible in the participating German hospitals. Although several QI could be implemented with minor effort, more time, financial and structural resources would be required for some QI, such as the QI postoperative visit.

[Implementation of anesthesia quality indicators in Germany : A prospective, national, multicenter quality improvement study]. / Implementierung anästhesiologischer Qualitätsindikatoren in Deutschland : Eine prospektive, nationale, multizentrische Qualitätssteigerungsstudie German version.

BACKGROUND: In 2016 the German Society of Anesthesiology and Intensive Care Medicine (DGAI) and the Association of German Anesthetists (BDA) published 10 quality indicators (QI) to compare and improve the quality of anesthesia care in Germany. So far, there is no evidence for the feasibility of implementation of these QI in hospitals. OBJECTIVE: This study tested the hypothesis that the implementation of the 10 QI is feasible in German hospitals. MATERIAL AND METHODS: This prospective three-phase national multicenter quality improvement study was conducted in 15 German hospitals and 1 outpatient anesthesia center from March 2017 to February 2018. The trial consisted of an initial evaluation of pre-existing structures and processes by the heads of the participating anesthesia departments, followed by a 6-month implementation phase of the QI as well as a final re-evaluation phase. The implementation procedure was supported by web-based implementation aids ( www.qi-an.org ) and internal quality management programs. The primary endpoint was the difference in the number of implemented QI per center before and after implementation. Secondary endpoints were the number of newly implemented QI per center, the overall number of successful implementations of each QI, the identification of problems during the implementation as well as the kind of impediments preventing the QI implementation. RESULTS: The average number of implemented QI increased from 5.8 to 6.8 (mean of the differences 1.1 ± 1.3; P < 0.01). Most frequently the QI perioperative morbidity and mortality report (5 centers) and the QI temperature management (4 centers) could be implemented. After the implementation phase, the QI incidence management and patient blood management were implemented in all 16 centers. Implementation of other quality indicators failed mainly due to a lack of time and lack of structural resources. CONCLUSION: In this study the implementation of QI was proven to be mostly feasible in the participating German hospitals. Although several QI could be implemented with minor effort, more time, financial and structural resources would be required for some QI, such as the QI postoperative visit.

Magnetic seizure therapy in treatment-resistant depression: clinical, neuropsychological and metabolic effects.

BACKGROUND: Magnetic seizure therapy (MST), despite being in an early phase of clinical research, has been demonstrated to be associated with antidepressant efficacy. However, safety, tolerability and efficacy data in connection with functional brain activity from larger samples are lacking. The aim of this study was to determine clinical and cognitive effects of MST and the influence of MST on regional brain glucose metabolism. METHOD: Twenty-six patients suffering from treatment-resistant depression (TRD) underwent MST. Ten patients underwent a randomized trial and 16 patients an open-label study design. The primary outcome criterion was the severity of depressive symptoms assessed with the Hamilton Depression Rating Scale (HAMD). Depressive symptoms, tolerability and cognitive safety, along with social functioning and quality of life parameters, were assessed using various rating scales. A clinical follow-up visit 6 months following the completion of a course of MST and [18F]-fluorodeoxyglucose positron emission tomography (FDG-PET) scans of 12 patients were analysed. RESULTS: A significant response to MST was demonstrated by 69% of the patient sample, with 46% meeting remission criteria. Anxiety ratings were significantly reduced in responders and their quality of life was improved. Half of the responders relapsed within 6 months. No cognitive side-effects were observed. FDG-PET scans showed a metabolic increase in the frontal cortex bilaterally and a decrease in the left striatum. CONCLUSIONS: Robust antidepressant and anti-anxiety efficacy of MST was demonstrated, and found to be associated with localized metabolic changes in brain areas that are strongly implicated in depression. Thus, MST presents an effective, well-tolerated and safe treatment option for patients unable to respond to other forms of therapy for depression.

Bilateral bispectral index monitoring during and after electroconvulsive therapy compared with magnetic seizure therapy for treatment-resistant depression.

BACKGROUND: Electroconvulsive therapy (ECT) is a highly effective and established treatment for depression. Magnetic seizure therapy (MST) has recently been developed and seems equally effective while associated with fewer side-effects. Both require general anaesthesia, which could be quantified using the bispectral index (BIS). We compared ECT and MST with respect to recovery times, left-sided BIS, and left-right differences in BIS. METHODS: In this prospective, observational study, we enrolled 10 successive patients receiving ECT and 10 patients undergoing MST. Anaesthesia was performed with propofol and monitored with a bilateral BIS sensor. The seizure was elicited when the BIS was within a range from 50 to 60. The time to eye opening was measured and bilateral BIS were recorded for 10 min after seizure induction. RESULTS: A comparable anaesthetic depth was observed in the ECT and MST groups at baseline [mean (standard deviation, sd) BIS values of 94.1 (4.1) and 95.5 (3.0), respectively] and before seizure induction [mean (sd) BIS values of 52.3 (9.6) and 55.2 (10.3), respectively]. Post-ictally, MST patients opened their eyes significantly earlier than ECT patients [3.0 (1.0) vs 6.7 (1.3) min, P<0.001]. They showed a significantly higher BIS at 2 min after seizure induction [69.2 (10.1) vs 50.9 (15.9), P=0.003], and this difference was still present at 10 min after seizure induction [BIS 81.5 (6.5) vs 68.0 (16.4), P<0.001]. Significant differences between the left and right BIS were observed in neither the ECT nor the MST group. CONCLUSIONS: At a comparable anaesthetic depth, MST is superior to ECT in terms of post-ictal recovery, which is correctly reflected by higher post-ictal BIS values. Unilateral BIS monitoring is sufficient to monitor anaesthetic depth in ECT and MST patients. TRIAL REGISTRY NUMBER: NCT 01318018.

Patient state index vs bispectral index as measures of the electroencephalographic effects of propofol.

BACKGROUND: The patient state index (PSI) and the bispectral index (BIS) quantify anaesthetic depth based on the EEG using different algorithms. We compared both indices with regard to the prediction of the depth of propofol anaesthesia. METHODS: In 17 patients, propofol was infused until burst suppression occurred and stopped thereafter until BIS recovered to values above 60. This was repeated; afterwards, patients were intubated, for subsequent surgery. Without surgical stimulus, PSI and BIS were measured simultaneously and compared with the estimated effect-site concentrations of propofol. These were derived from simultaneous pharmacokinetic and -dynamic modelling in an individual two-stage and a population-based NONMEM approach. RESULTS: A close sigmoid relationship was observed between the propofol effect-site concentration and both PSI [coefficient of determination rho(2)=0.91 (sd 0.05)] and BIS [rho(2)=0.92 (0.03)], which was significantly steeper for PSI [gamma=2.2 (0.6)] than for BIS [gamma=1.8 (0.4)], and reached significantly lower values for PSI [E(max)=0.3 (1.1)] than for BIS [E(max)=5.3 (6.7)] at maximal propofol concentrations. A significantly smaller k(e0) was obtained for PSI [0.09 (0.03) min(-1)] compared with BIS [0.10 (0.02) min(-1)]. PSI and BIS correlated significantly with each other (rho(2)=0.866) and predicted propofol effect-site concentration with a comparable probability [P(K)=0.87 (0.05) and 0.86 (0.05), respectively]. NONMEM revealed E(0)=89.3 and 92.3, E(max)=1.9 and 8.6, C(e50)=1.38 and 1.92 microg ml(-1), gamma=1.6 and 1.48, and k(e0)=0.103 and 0.131 min(-1) as typical values for PSI and BIS, respectively. CONCLUSIONS: The PSI and the BIS monitors performed equally well in predicting depth of propofol anaesthesia. However, PSI was lower than BIS by approximately 10-15 points at high propofol concentrations.

The Entropy Module and Bispectral Index as guidance for propofol-remifentanil anaesthesia in combination with regional anaesthesia compared with a standard clinical practice group.

This study was designed to investigate the impact of the Entropy Module and Bispectral Index (BIS) monitoring on drug consumption and recovery times compared with standard anaesthetic practice in patients undergoing orthopaedic surgery using a combination of regional and general anaesthesia as performed by an experienced anaesthesiologist. We hypothesised that electroencephalogram monitoring would lead to a lower drug consumption as well as shorter recovery times. With institutional review board approval and written informed consent, 90 adult patients undergoing surgery to the upper or lower extremity received regional anaesthesia for post- and intraoperative pain control and were randomised to receive general anaesthesia by propofol/remifentanil infusion controlled either solely by clinical parameters or by targeting Entropy or BIS values of 50. Recovery times and drug consumption were recorded. Data from 79 patients were analysed. Compared with standard practice, patients with Entropy or BIS monitoring showed a similar propofol consumption (standard practice 101 +/- 22 microg/kg/minute, Entropy 106 +/- 24 microg/kg/minute, BIS 104 +/- 20 microg/kg/minute) and showed similar Aldrete scores (10/10) one minute after extubation: 9.1 +/- 0.3, 9.2 +/- 0.6 and 9.3 +/- 0.5, respectively. Time points of extubation were 7.3 +/- 2.9 minutes, 9.2 +/- 3.9 minutes and 6.8 +/- 2.9 minutes, respectively, demonstrating a significant difference between Entropy and BIS (P = 0.023). Compared with standard practice, targeting an Entropy or BIS value of 50 did not result in a reduction of propofol consumption during general anaesthesia combined with regional anaesthesia as performed by an experienced anaesthesiologist in orthopaedic patients.

Predictive value of initial clinical status, intracranial pressure and transcranial Doppler pulsatility after subarachnoid haemorrhage.

BACKGROUND: We examined the predictive value of initial clinical status, mean arterial blood pressure (MABP), intracranial pressure (ICP) and transcranial Doppler (TCD)-derived pulsatility and resistance indices for outcome and quality of life one year following aneurysmal subarachnoid haemorrhage (SAH). METHOD: Neuromonitoring was performed in 29 patients following clipping or coiling of an aneurysm. Mean arterial blood pressure was measured in the radial artery and intracranial pressure was assessed via a closed external ventricular drainage. Based on transcranial Doppler-recordings of the middle cerebral artery, Gosling's pulsatility (PI) and Pourcelot's resistance (RI) index were calculated. Glasgow outcome score (GOS) and short form-36 (SF-36) scores were determined one year after SAH. FINDINGS: An unfavourable outcome (GOS 1-3) was observed in 34% of patients and correlated significantly (p < 0.05) with a poor initial clinical status, as determined by Glasgow Coma Scale (r = 0.55), Hunt and Hess (r = -0.62), World Federation of Neurosurgical Societies (WFNS) (r = -0.48) and Fisher (r = -0.58) score. Poor outcome was significantly associated with high mean arterial blood pressure (r = -0.44) and intracranial pressure (r = -0.48) as well as increased pulsatility (r = -0.46) and resistance (r = -0.43) indices. Hunt and Hess grade > or = 4 (OR 12.4, 5-95% CI: 1.9-82.3), mean arterial blood pressure > 95 mmHg (19.5, 2.9-132.3), Gosling's pulsatility >0.8 (6.5, 1.6-27.1) and Pourcelot's resistance >0.57 (15.4, 2.3-103.4) were predictive for unfavourable outcome in logistic regression, however TCD-diagnosed vasospasm was not. Except for mental health, significantly reduced scores were observed in all short form-36 domains. Initial clinical status correlated significantly with the physical functioning, role physical, bodily pain, social functioning and physical component summary of short form-36. CONCLUSIONS: Mortality and morbidity following SAH remains high, especially in poor-grade patients. Outcome is mainly correlated with initial clinical status, mean arterial blood pressure, intracranial pressure, pulsatility and resistance indices. Those factors seem to be stronger than the influence of vasospasm.

Correlation of the A-Line ARX index with acoustically evoked potential amplitude.

BACKGROUND: Automated indices derived from mid-latency auditory evoked potentials (MLAEP) have been proposed for monitoring the state of anaesthesia. The A-Line ARX index (AAI) has been implemented in the A-Line monitor (Danmeter, V1.4). Several studies have reported variable and, in awake patients, sometimes surprisingly low AAI values. The purpose of this study was to reproduce these findings under steady-state conditions and to investigate their causes. METHODS: Ten awake unmedicated volunteers were studied under steady-state conditions. For each subject, the raw EEG and the AAI were recorded with an A-Line monitor (V1.4) during three separate sessions of 45.0 (1.6) min duration each. MATLAB (Mathworks) routines were used to derive MLAEP responses from EEG data and to calculate maximal MLAEP amplitudes. RESULTS: The AAI values ranged from 15 to 99, while 11.4% fell below levels which, according to the manufacturer, indicate an anaesthetic depth suitable for surgery. Inter-individual and intra-individual variation was observed despite stable recording conditions. The amplitudes of the MLAEP varied from 0.8 to 42.0 microV. The MLAEP amplitude exceeded 2 microV in 75.3% of readings. The Spearman's rank correlation coefficient between the MLAEP amplitude and the AAI value was r=0.89 (P<0.0001). CONCLUSIONS: The version of the A-Line monitor used in this study does not exclude contaminated MLAEP signals. Previous publications involving this version of the A-Line monitor (as opposed to the newer A-Line/2 monitor series) should be reassessed in the light of these findings. Before exclusively MLAEP-based monitors can be evaluated as suitable monitors of depth of anaesthesia, it is essential to ensure that inbuilt validity tests eliminate contaminated MLAEP signals.

Correlation of continuously monitored regional cerebral blood flow and brain tissue oxygen.

BACKGROUND: The purpose of this study was to investigate the relationship between continuously monitored regional cerebral blood flow (CBF) and brain tissue oxygen (PtiO2). METHODS: Continuous advanced multimodal neuromonitoring including monitoring of PtiO2 (Licox, GMS) and CBF (QFlow, Hemedex) was performed in eight patients after severe subarachnoid haemorrhage (n=5) and traumatic brain injury (n=3) for an average of 9.6 days. Parameters were measured using a flexible polarographic PtiO2-probe and a thermal diffusion CBF-microprobe. FINDINGS: Regarding the whole monitoring period in all patients, the data indicated a significant correlation between CBF and PtiO2 (r=0.36). In 72% of 400 analysed intervals of 30 minutes duration with PtiO2 changes larger than 5 mmHg, a strong correlation between CBF and PtiO2 existed (r>0.6). In 19% of intervals a still statistically significant correlation was observed (0.3

Brain tissue oxygen (PtiO2): a clinical comparison of two monitoring devices.

BACKGROUND: We investigated the difference between two commercially available sensors for continuous monitoring of brain tissue oxygen (PtiO2). One is a single parameter probe for PtiO2 monitoring (Licox), the other is a multiparamter sensor (Neurotrend) further including measurement of brain temperature, pH, and partial pressure of tissue carbon dioxide. METHODS: In seven patients after subarachnoid hemorrhage or traumatic brain injury continuous monitoring of PtiO2 was performed simultaneously using Licox and Neurotrend. FINDINGS: Mean PtiO2 was generally lower when assessed by the Neurotrend, as compared with the Licox (Licox 27.7 mmHg vs. Neurotrend 20.9 mmHg; P = 0.028). The amplitude of PtiO2 elevations during ventilation with 100% oxygen was higher with the Licox, but this did not reach statistical significance (Licox 55.2 mmHg vs. Neurotrend 50.2 mmHg, P = 0.082). Regarding clinical stability of the sensors, only one Neurotrend sensor provided valid function over the desired monitoring period. Five Neurotrend sensors dislocated or broke and one sensor did not show any function after insertion. No malfunction occurred with the Licox sensors. CONCLUSIONS: Our results suggest that PtiO2 might be lower when assessed by the Neurotrend sensor. The clinical stability of the Neurotrend sensor was of concern and allowed monitoring in one of seven patients over the desired monitoring period of several days only.

Improvement of brain tissue oxygen and intracranial pressure during and after surgical decompression for diffuse brain oedema and space occupying infarction.

BACKGROUND: We evaluated the perioperative and intraoperative changes of intracranial pressure (ICP) and partial pressure of brain tissue oxygen (PtiO2) after decompressive craniectomy in patients with diffuse brain oedema and space occupying infarction. METHODS: Ten patients suffering from medically intractable raised intracranial pressure (ICP) were included. The underlying diseases and causes for elevated ICP were diffuse brain oedema after subarachnoid haemorrhage (n = 3) and head injury (n = 3), or space occupying infarction of the middle cerebral artery territory due to vasospasm after SAH (n = 4). Continuous perioperative and intraoperative monitoring of PtiO2 and ICP was performed at the side of decompression. FINDINGS: ICP and PtiO2 improved significantly in a uniform pattern during bone flap removal and dura opening, irrespective of the underlying disease (mean ICP from 52 mmHg to 8 mmHg, mean PtiO2 from 9 mmHg to 25 mmHg). ICP, PtiO2, and cerebral perfusion pressure were further improved in the subsequent 12 hours after surgery, as compared to the preoperative 12 hours. CONCLUSIONS: Decompressive craniectomy seems to be a successful option in the treatment of intractable intracranial hypertension with associated cerebral hypoxia. These positive effects may last for several hours after the procedure irrespective of the underlying disease.

Effects of decompressive craniectomy on brain tissue oxygen in patients with intracranial hypertension.

This report examined the intraoperative course of partial pressure of brain tissue oxygen (P(ti)O(2)) and intracranial pressure (ICP) during surgical decompressive craniectomy for medically intractable intracranial hypertension due to diffuse brain swelling in three patients after severe subarachnoid haemorrhage and aneurysm coiling. The mean ICP decreased from 59 mm Hg to 10 mm Hg in a two step fashion, relating to bone flap removal and dural opening. Simultaneously, P(ti)O(2) increased rapidly from 0.8 kPa (6 mm Hg) to 3.07 kPa (23 mm Hg). P(ti)O(2) and ICP remained at non-critical ranges postoperatively. Despite these beneficial effects on ICP and P(ti)O(2), the patients' clinical status remained poor with two in a persistent vegetative state and one dead.

Continuous assessment of cerebral autoregulation: clinical and laboratory experience.

The method for the continuous assessment of cerebral autoregulation using slow waves of MCA blood flow velocity (FV) and cerebral perfusion pressure (CPP) or arterial pressure (ABP) has been introduced seven years ago. We intend to review its clinical applications in various scenarios. Moving correlation coefficient (3-6 min window), named Mx, is calculated between low-pass filtered (0.05 Hz) signals of FV and CPP or ABP (when ICP is not measured directly). Data from ventilated 243 head injuries and 15 patients after poor grade subarachnoid haemorrhage, 38 patients with Carotid Artery stenosis, 35 patients with hydrocephalus and fourteen healthy volunteers is presented. Good agreement between the leg-cuff test and Mx has been confirmed in healthy volunteers (r = 0.81). Mx also correlated significantly with the static rate of autoregulation and transient hyperaemic response test. Autoregulation was disturbed (p < 0.021) by vasospasm after SAH and worse in patients with hydrocephalus in whom CSF circulation was normal (p < 0.02). In head injury, Mx indicated disturbed autoregulation with low CPP (< 55 mmHg) and too high CPP (> 95 mmHg). Mx strongly discriminated between patients with favourable and unfavourable outcome (p < 0.00002). This method can be used in many clinical scenarios for continuous monitoring of cerebral autoregulation, predicting outcome and optimising treatment strategies.

On the number of measurement sites required to assess regional cerebral blood flow by laser-Doppler scanning during cerebral ischemia and reperfusion.

The aim of this study was to determine whether the number of measurement sites affected the precision of regional cerebral blood flow (CBF) assessment by Laser-Doppler (LD). A simulation study was applied based on data obtained by scanning the cortex in 25 rats during baseline conditions, 15 min global cerebral ischemia and reperfusion. Random samples were repeatedly collected from 1 to 100 locations and deviations from the median of the entire CBF data pool (800 locations) were determined. Single location CBF measurements missed the true median by 24.8+/-2.2 LD-units (baseline conditions, n=100 simulations, mean+/-SEM), 2.7+/-0.6 LD-units (ischemia), and 31.9+/-2.4 LD-units (30th min reperfusion), which can be reduced to 10.9+/-1.0 LD-units (baseline), 0.9+/-0.1 LD-units (ischemia), and 15.5+/-1.3 LD-units (30th min reperfusion) by scanning ten locations. Reliability is further improved by scanning 30 sites with deviations of 6.1+/-0.6 LD-units (baseline), 0.4+/-0.0 LD-units (ischemia), and 8.9+/-0.7 LD-units (30th min reperfusion). Single location CBF assessment was sufficient during global ischemia only. In order to keep the deviation from the true flow below 10 LD-units, at least 15 locations are recommended during baseline conditions and 25 during reperfusion. Laser-Doppler scanning improves the reliability and reduces the variability of CBF measurements.

Laser Doppler scanning: how many measurements are required to assess regional cerebral blood flow?

This study was initiated to determine the optimal number of measuring sites necessary to estimate regional cerebral blood flow (CBF) under pathophysiological conditions. 25 rats were exposed to 15 minutes of global cerebral ischemia. Local CBF was sequentially measured by laser Doppler (LD) at 32 locations during baseline conditions, ischemia and reperfusion using a computer-controlled scanning technique. A simulation study was performed based on 800 local measurements at each time point: random samples (size 1-100) were repeatedly drawn to estimate the variability of median flow. Accuracy was defined as probability that the simulated median differed less than +/- 5 LD-units from the true median of the 800 data. Above a single location, CBF was measured with an accuracy of 21.6 +/- 0.4% (baseline conditions, n = 100 simulations, mean +/- SEM), 85.8 +/- 0.4% (ischemia) and 11.1 +/- 0.3% (30th min. reperfusion). Accuracy increases to 75.2 +/- 0.5% (baseline conditions), 100 +/- 0% (ischemia) and 41.8 +/- 0.6% (30th min. reperfusion) if 24 locations are scanned. Scanning, therefore, improves accuracy and reduces variability of CBF measurements. With enough local CBF measurements laser Doppler assessment of regional CBF is possible. Single location CBF assessment is sufficiently accurate only during ischemia. During reperfusion, when accuracy is half reduced compared to baseline conditions, larger sample sizes are required.

Postischemic application of lipid peroxidation inhibitor U-101033E reduces neuronal damage after global cerebral ischemia in rats.

BACKGROUND AND PURPOSE: The lipid peroxidation inhibitor U-101033E was examined for effects on cerebral blood flow (CBF), cortical tissue hemoglobin oxygen saturation (HbSo2), and neuronal damage. METHODS: Fifteen minutes of global cerebral ischemia was induced by two-vessel occlusion and hypobaric hypotension. Wistar rats (n = 25) were randomized to receive vehicle (n = 9) or 40 mg/kg U-101033E (n = 9) intraperitoneally during 2 hours of reperfusion. A sham group (n = 7) had neither ischemia nor therapy. Histology was evaluated 7 days after ischemia. RESULTS: During late hyperperfusion (at 17 minutes), vehicle-treated animals had a higher (P = 0.044) cortical tissue HbSo2 (72.0 +/- 1.4%) than did U-101033E-treated animals (65.8 +/- 2.5%). Neuronal counts in the superficial cortex layer found after 7 days correlated negatively with rCBF (r = -0.76; P < 0.001) or cortical tissue HbSo2 (r = -0.56; P = 0.028) assessed during the late hyperperfusion phase. U-101033E reduced neuronal damage in hippocampal CA1 from 64.3 +/- 9.2% to 31.2 +/- 8.4% (P = 0.020), as well as in the superficial cortical layer from 53.5 +/- 14.6% to 12.8 +/- 11.7% (P = 0.046). While animals in the vehicle group had reduced counts in all four examined cortex layers (P < 0.05 versus sham group), there was significant cortical neuron loss in the U-101033E group in only one of four areas. U-101033E had no effect on resting CBF or CO2 reactivity. CONCLUSIONS: Postischemic application of U-101033E protects hippocampal CA1 and cortical neurons after 15 minutes of global cerebral ischemia. The results indicate that free radical-induced lipid peroxidation contributes to reperfusion injury, a process that can be inhibited by antioxidants such as U-101033E.