Forewarning of hypotensive events using a Bayesian artificial neural network in neurocritical care.

Traumatically brain injured (TBI) patients are at risk from secondary insults. Arterial hypotension, critically low blood pressure, is one of the most dangerous secondary insults and is related to poor outcome in patients. The overall aim of this study was to get proof of the concept that advanced statistical techniques (machine learning) are methods that are able to provide early warning of impending hypotensive events before they occur during neuro-critical care. A Bayesian artificial neural network (BANN) model predicting episodes of hypotension was developed using data from 104 patients selected from the BrainIT multi-center database. Arterial hypotension events were recorded and defined using the Edinburgh University Secondary Insult Grades (EUSIG) physiological adverse event scoring system. The BANN was trained on a random selection of 50% of the available patients (n = 52) and validated on the remaining cohort. A multi-center prospective pilot study (Phase 1, n = 30) was then conducted with the system running live in the clinical environment, followed by a second validation pilot study (Phase 2, n = 49). From these prospectively collected data, a final evaluation study was done on 69 of these patients with 10 patients excluded from the Phase 2 study because of insufficient or invalid data. Each data collection phase was a prospective non-interventional observational study conducted in a live clinical setting to test the data collection systems and the model performance. No prediction information was available to the clinical teams during a patient's stay in the ICU. The final cohort (n = 69), using a decision threshold of 0.4, and including false positive checks, gave a sensitivity of 39.3% (95% CI 32.9-46.1) and a specificity of 91.5% (95% CI 89.0-93.7). Using a decision threshold of 0.3, and false positive correction, gave a sensitivity of 46.6% (95% CI 40.1-53.2) and specificity of 85.6% (95% CI 82.3-88.8). With a decision threshold of 0.3, > 15 min warning of patient instability can be achieved. We have shown, using advanced machine learning techniques running in a live neuro-critical care environment, that it would be possible to give neurointensive teams early warning of potential hypotensive events before they emerge, allowing closer monitoring and earlier clinical assessment in an attempt to prevent the onset of hypotension. The multi-centre clinical infrastructure developed to support the clinical studies provides a solid base for further collaborative research on data quality, false positive correction and the display of early warning data in a clinical setting.

Treatment of poor-grade subarachnoid hemorrhage trial.

BACKGROUND AND PURPOSE: Management of poor-grade subarachnoid hemorrhage is based on limited evidence from small single-center retrospective observational studies. The purpose of this study was to undertake a single-center randomized controlled feasibility trial comparing a strategy of early endovascular aneurysm treatment with treatment after neurologic recovery in this group of patients. MATERIALS AND METHODS: Patients with poor-grade SAH were randomized within 24 hours of admission to early treatment or treatment after neurologic recovery. If a patient was randomized to early treatment, the aneurysm was treated endovascularly within 24 hours of randomization. Recruitment rate, safety profile, and functional outcome at the time of discharge and at 6 months were assessed. RESULTS: Fourteen of 51 patients screened were eligible for the trial. Of these 14, 8 patients were randomized (57%). All patients in the early coiling arm received treatment within 24 hours of randomization. There was no treatment-related complication. Overall, good outcome occurred in 25% of patients; the mortality rate was 75%. Patients in the early treatment arm (n = 5) had a good outcome rate of 20%, while those in treatment after neurologic recovery arm (n = 3) had a good outcome rate of 33.3%. CONCLUSIONS: This was a feasibility study that demonstrated that recruitment and randomization for comparing management strategies in poor-grade SAH are feasible. The recruitment rate among eligible patients was encouraging (57%), though a number of patients had to be excluded due to ineligibility. A multicenter study is necessary to recruit the numbers required to compare the clinical outcomes of these management strategies.

Does the cognitive measure Cog-4 show improvement among patients treated with thrombolysis after acute stroke?

BACKGROUND: Although the established measure of disability post stroke, the modified Rankin Scale emphasizes motor function and may underestimate the importance of cognitive impairment in more disabled patients. A subset of four items from the National Institutes of Health Stroke Scale has been proposed to assess cognitive function after stroke (Cog-4), and to correlate with modified Rankin Scale. Items correspond to orientation, executive function, language, and inattention. We investigated responsiveness of Cog-4 to treatment with thrombolysis and whether it offers information that supplements modified Rankin Scale. METHODS: We included 6268 patients from the Virtual International Stroke Trials Archive: 2734 received intravenous thrombolysis and 3534 were treated conservatively. We compared day 90 outcomes between treated and untreated groups, by modified Rankin Scale (illustrative) and by Cog-4 (primary measure) adjusting for age, baseline National Institutes of Health stroke scale, hemispheric lateralisation as well as baseline Cog-4 and baseline National Institutes of Health Stroke Scale excluding baseline Cog-4 separately. Analysis of Cog-4 was repeated within strata of 90 day modified Rankin Scale. Statistical analyses included proportional odds logistic regression and Cochran-Mantel-Haenszel test. RESULTS: Modified Rankin Scale showed a difference between treatment groups of expected magnitude (odds ratio 1·56; 95% confidence interval 1·43-1·72; P < 0·001). After adjustment for imbalance in baseline prognostic factors, the distribution of Cog-4 scores at 90 days was better in thrombolysed patients compared with nonthrombolysed patients (odds ratio 1·31; 95% confidence interval 1·18-1·47; P = 0·006). However, Cog-4 analysis stratified by 90-day modified Rankin Scale was neutral between treatment groups (OR 1·01; 95% CI 0·90-1·14), and Cog-4 was not responsive to treatment group even within modified Rankin Scale categories 4 and 5 despite substantial cognitive deficits in these patients. CONCLUSION: Although Cog-4 may be responsive to treatment effects, it does not provide additional information beyond modified Rankin Scale assessment.

Early warning of EUSIG-defined hypotensive events using a Bayesian Artificial Neural Network.

BACKGROUND: Hypotension is recognized as a potentially damaging secondary insult after traumatic brain injury. Systems to give clinical teams some early warning of likely hypotensive instability could be added to the range of existing techniques used in the management of this group of patients. By using the Edinburgh University Secondary Insult Grades (EUSIG) definitions for -hypotension (systolic arterial pressure <90 mmHg OR mean arterial -pressure <70 mmHg) we collected a group of ∼2,000 events by analyzing the Brain-IT database. We then constructed a Bayesian Artificial Neural Network (an advanced statistical modeling technique) that is able to provide some early warning when trained on this previously collected demographic and physiological data. MATERIALS AND METHODS: Using EUSIG defined event data from the Brain-IT database, we identified a Bayesian artificial neural network (BANN) topology and constructed a series of datasets using a group of clinically guided input variables. This allowed us to train a BANN, which was then tested on an unseen set of patients from the Brain-IT database. The initial tests used a particularly harsh assessment criterion whereby a true positive prediction was only allowed if the BANN predicted an upcoming event to the exact minute. We have now developed the system to the point where it is about to be used in a two-stage Phase II clinical trial and we are also researching a more realistic assessment technique. KEY RESULTS: We have constructed a BANN that is able to provide early warning to the clinicians based on a model that uses information from the physiological inputs; systolic and mean arterial pressure and heart rate; and demographic variables age and gender. We use 15-min SubWindows starting at 15 and 30 min before an event and process mean, slope and standard deviations. Based on 10 simulation runs, our current sensitivity is 36.25% (SE 1.31) with a specificity of 90.82% (SE 0.85). Initial results from a Phase I clinical study shows a model sensitivity of 40.95% (SE 6%) and specificity of 86.46% (SE 3%) Although this figure is low it is considered clinically useful for this dangerous condition, provided the false positive rate can be kept sufficiently low as to be practical in an intensive care environment. CONCLUSION: We have shown that using advanced statistical modeling techniques can provide clinical teams with useful information that will assist clinical care.

Trigger characteristics of EUSIG-defined hypotensive events.

BACKGROUND: Hypotension is a recognized -secondary insult after traumatic brain injury (TBI). There are many definitions of hypotension, an often cited example being the Brain Trauma Foundation's current (2007) "Guidelines for the Management of Severe Traumatic Brain Injury," which defines hypotension as systolic pressure <90 mmHg. However, this same document declares "The importance of mean arterial pressure, as opposed to systolic pressure should also be stressed, …." Our work shows that when using the Edinburgh University Secondary Insult Grades (EUSIG) definitions, which require monitoring of both systolic and mean arterial pressures, that most hypotensive events are in fact triggered by a breach of the mean arterial level of 70 mmHg. We suggest that close monitoring of mean arterial pressure would enable clinical teams to avoid more potentially damaging hypotensive events. MATERIALS AND METHODS: An analysis of 100 patients from the Brain-IT database was performed. Using the EUSIG definitions, 2,081 events can be obtained by analyzing the systolic and mean blood pressures on a minute by minute basis. A software program was written to identify and classify the trigger pattern for each event. A categorical analysis of these triggering patterns has been carried out. KEY RESULTS: Our analysis shows that most events are triggered by a drop in mean arterial pressure. In fact a large number of events (91%) occur where the mean arterial pressure is below the threshold limits whereas the systolic pressure does not cross the 90 mmHg limit at all. CONCLUSION: We suggest that more emphasis should be placed on closely monitoring mean arterial pressure as well as systolic pressure when trying to guard against hypotensive problems in traumatically brain injured patients. In future work we will study the underlying physiological mechanisms and attempt to further classify concomitant conditions that may be contributing to the onset of a hypotensive event.

The effect of the results of the STICH trial on the management of spontaneous supratentorial intracerebral haemorrhage in Newcastle.

Recently, the Surgical Trial in IntraCerebral Haemorrhage (STICH) was unable to show an overall benefit from 'early surgery' compared with a policy of 'initial conservative treatment'. Here, we evaluated the impact of the STICH results on the management of spontaneous supratentorial intracerebral haemorrhage (ICH) in the Newcastle upon Tyne Hospitals. The STICH results were released to the Neurosurgery Department at Newcastle General Hospital in November 2003; using ICD-10 data, we analysed ICH admissions before (2002) and after (2004, 2006, 2007) this. We assessed numbers of Neurosurgery and Stroke Unit admissions, numbers of clot evacuation procedures, and 30-day mortality rate (Neurosurgery vs. Stroke Unit admissions). Subarachnoid haemorrhage (SAH) admissions data were also collected to corroborate our findings. There were 478 spontaneous supratentorial ICH admissions in total; 156 in 2002, 120 in 2004, 106 in 2006 and 96 in 2007. SAH admissions remained remarkably constant over this period. Neurosurgery admissions decreased significantly across the four time periods, from 71% of total ICH admissions (n = 156) in 2002 to 55% (n = 96) in 2007, and Stroke Unit admissions increased significantly from 8% (n = 156) in 2002 to 30% (n = 96) in 2007 (chi(2) = 20.968, p < 0.001, df = 3). Clot evacuation procedures also decreased significantly from 32% (n = 111) of Neurosurgery admissions in 2002 to 17% (n = 53) in 2007 (chi(2) = 11.919, p = 0.008, df = 3). 30-day mortality increased in Neurosurgery from 14% of Neurosurgery admissions (n = 111) in 2002 to 26% (n = 53) in 2007, and decreased in the Stroke Unit, from 42% of Stroke Unit admissions (n = 12) in 2002 to 17% (n = 29) in 2007. The STICH results have significantly impacted ICH management in Newcastle, with a trend towards fewer Neurosurgery admissions and clot evacuations, and increased Stroke Unit admissions. The role of surgery for ICH remains controversial, and randomization continues in STICH II for patients with superficial lobar ICH.

The use of hyperventilation therapy after traumatic brain injury in Europe: an analysis of the BrainIT database.

OBJECTIVE: To assess the use of hyperventilation and the adherence to Brain Trauma Foundation-Guidelines (BTF-G) after traumatic brain injury (TBI). SETTING: Twenty-two European centers are participating in the BrainIT initiative. DESIGN: Retrospective analysis of monitoring data. PATIENTS AND PARTICIPANTS: One hundred and fifty-one patients with a known time of trauma and at least one recorded arterial blood-gas (ABG) analysis. MEASUREMENTS AND RESULTS: A total number of 7,703 ABGs, representing 2,269 ventilation episodes (VE) were included in the analysis. Related minute-by-minute ICP data were taken from a 30 min time window around each ABG collection. Data are given as mean with standard deviation. (1) Patients without elevated intracranial pressure (ICP) (< 20 mmHg) manifested a statistically significant higher P(a)CO(2) (36 +/- 5.7 mmHg) in comparison to patients with elevated ICP (> or = 20 mmHg; P(a)CO(2): 34 +/- 5.4 mmHg, P < 0.001). (2) Intensified forced hyperventilation (P(a)CO(2) < or = 25 mmHg) in the absence of elevated ICP was found in only 49 VE (2%). (3) Early prophylactic hyperventilation (< 24 h after TBI; P(a)CO(2) < or = 35 mmHg, ICP < 20 mmHg) was used in 1,224 VE (54%). (4) During forced hyperventilation (P(a)CO(2) < or = 30 mmHg), simultaneous monitoring of brain tissue pO(2) or S(jv)O(2) was used in only 204 VE (9%). CONCLUSION: While overall adherence to current BTF-G seems to be the rule, its recommendations on early prophylactic hyperventilation as well as the use of additional cerebral oxygenation monitoring during forced hyperventilation are not followed in this sample of European TBI centers. DESCRIPTOR: Neurotrauma.

Intraventricular hemorrhage and hydrocephalus after spontaneous intracerebral hemorrhage: results from the STICH trial.

INTRODUCTION: Intraventricular hemorrhage (IVH), either independent of or as an extension of intracranial bleed, is thought to carry a grave prognosis. Although the effect of IVH on outcome in patients with subarachnoid hemorrhage has been extensively reviewed in the literature, reports of spontaneous intracerebral hemorrhage (ICH) in similar situations have been infrequent. The association of hydrocephalus in such situations and its influence on outcome is also uncertain. PATIENTS AND METHODS: As a sub-analysis of data obtained through the international Surgical Trial in Intracerebral Hemorrhage (STICH), the impact of IVH, with or without the presence of hydrocephalus, on outcome in patients with spontaneous ICH was analyzed. CT scans of randomized patients were examined for IVH and/or hydrocephalus. Other characteristics of hematoma were evaluated to see if they influenced outcome, as defined by the STICH protocol. RESULTS: Favorable outcomes were more frequent when IVH was absent (31.4% vs. 15.1%; p < 0.00001). The presence of hydrocephalus lowered the likelihood of favorable outcome still further to 11.5% (p = 0.031). In patients with IVH, early surgical intervention had a more favorable outcome (17.8%) compared to initial conservative management (12.4%) (p = 0.141). CONCLUSION: The presence of IVH and hydrocephalus are independent predictors of poor outcome in spontaneous ICH. Early surgery is of some benefit in those with IVH.

The significance of crossovers after randomization in the STICH trial.

INTRODUCTION: Of all forms of stroke, spontaneous intracerebral haemorrhage (ICH) causes the highest morbidity and mortality. The Surgical Trial in Intracerebral Haemorrhage (STICH) found no difference in outcomes between patients randomized to surgical or conservative treatment. PATIENTS AND METHODS: Of 530 patients randomized to initial conservative treatment, 140 crossed over to surgery. This study examines the variables associated with crossover. RESULTS: Dominant features of the crossover group were: male, (p = 0.04), right-sided clot (p = 0.03), lobar clot (p = 0.003), clot volume (median 64 mL for crossovers vs. 38 mL for others, p < 0.00001), midline shift (median 6 mm for crossovers vs. 3 mm for others, p < 0.00001), superficial clot (median 1.3 mm for crossovers vs. 11.5 mm for others, p < 0.00001), and randomization within 12 hours of ictus (p < 0.0005). Thalamic location (p = 0.002) was under-represented. Intraventricular haemorrhage, hydrocephalus, and focal deficits were not associated with crossover. Craniotomy was the method of evacuation in 85% of crossover patients. CONCLUSIONS: Crossover to surgery was more likely when ICH had these features: Right side, lobar location, superficial, large volume, big shift, and early randomization. Crossovers formed a worse prognostic group compared to non-crossovers. Surgery did not affect trial results, which were analyzed by intention-to-treat.

Variability in pupil size estimation.

BACKGROUND: The clinical estimation of pupil size and reactivity is central to the neurological assessment of patients, particularly those with or at risk of neurological damage. Health care professionals who examine pupils have differing levels of skill and training, yet their recordings are passed along the patient care pathway and can influence care decisions. The aim of this study was to determine if any statistical differences existed in the estimation of pupil size by different groups of health care professionals. METHODS: A total of 102 health care professionals working in the critical care environment were asked to estimate and record the pupil size of a series of 12 artificial eyes with varying pupil diameter and iris colour. All estimations were performed indoors under ambient lighting conditions. RESULTS: Our results established a statistically significant difference between staff groups in the estimation of pupil size. CONCLUSION: The demonstrated variability in pupil size estimation may not be clinically significant. However, it remains desirable to have consistency of measurement throughout the patient care pathway.

Comparison of image quality, diagnostic confidence and interobserver variability in contrast enhanced MR angiography and 2D time of flight angiography in evaluation of carotid stenosis.

The aim of this study was to compare image quality, level of diagnostic confidence and interobserver agreement in assessment of carotid stenosis with contrast enhanced MR angiography (CE MRA) in comparison with 2D time of flight MR angiography (2D TOF MRA). 60 carotid arteries in 30 patients were examined by three observers. Image quality and diagnostic confidence were assessed on the basis of a visual analogue scale. Interobserver variability was assessed with the help of intraclass correlation coefficient. Median values on the visual analogue scale for image quality and diagnostic confidence were higher for CE MRA compared with 2D TOF MRA for all three observers. Higher intraclass correlation values were recorded for interobserver variability for CE MRA compared with 2D TOF MRA both for visual estimation of carotid stenosis as well as for measurement of carotid stenosis on the basis of North American Symptomatic Carotid Endarterectomy Trial (NASCET) and European Carotid Surgery Trial (ECST) criteria. CE MRA provides better image quality, higher level of diagnostic confidence and more interobserver agreement compared with 2D TOF MRA.

Regional differences in outcome from subarachnoid haemorrhage.

BACKGROUND: Surgeons are increasingly placed under pressure to accept publication of their results and to abide by recommendations to change practice which others derive. Considerable concern exists about misinterpretation of such data. The issue is well illustrated by this study. METHOD: Data on outcome following treatment for subarachnoid haemorrhage were prospectively collected from 1993-1998 in two centres in the British Isles: Newcastle and Nottingham. FINDINGS: Initial examination of this data suggest a substantial difference in the performance favouring Nottingham over Newcastle. The odds of a poor outcome was 1:1.86 in Newcastle compared with 1:4.26 in Nottingham giving an odds ratio of 2.3 in favour of Nottingham and this difference was highly significant with p<0.00001. On a more detailed examination taking account of confounding variables, this difference disappeared entirely. Newcastle was able to operate a less selective admissions policy than Nottingham because of the deficiency of beds at the latter unit. A summary of these results has been published elsewhere. INTERPRETATION: These results illustrate the dangers of applying statistical tools developed for simpler situations such as industrial process control to complex medical problems. We conclude that comprehensive and accurate data on all factors likely to influence the outcome for a particular treatment should be collected as an absolute prerequisite to any judgments being made on apparent statistical differences between the performances of differing units.

Effect of awake Carotid Endarterectomy under local anaesthesia on peri-operative blood pressure: blood pressure is normalised when carotid stenosis is treated under local anaesthesia.

BACKGROUND: Carotid Endarterectomy can be performed under local, regional or general anaesthesia. One of the most important effects of the type of anaesthetic used is on the systemic blood pressure. Although variations in blood pressure during and following carotid endarterectomy have been studied previously, the effects of awake carotid endarterectomy under local anaesthesia on blood pressure and its comparison with similar procedures under similar types of anaesthesia have not. METHODS: Peri-operative blood pressure measurements were collected from the records of 25 consecutive patients for each of the following five procedures; Carotid Endarterectomy under general anaesthesia (CEAGA), Anterior Cervical Discectomy and Fusion under general anaesthesia (ACDF), Cerebral Angiography under local anaesthesia (ANG), Carotid Endarterectomy patients under local anaesthesia who were symptomatic (CEALAS) and Carotid Endarterectomy patients under local anaesthesia who were asymptomatic (CEALAA). The recordings were then analysed to find out if there were any clinically significant variations in peri-operative blood pressure. FINDINGS: There is a significant and consistent difference when the pre-operative value was compared with the 4 hour and 24 hour post-operative recordings between the local and general anaesthetic groups for carotid endarterectomy. Carotid endarterectomy reduces the systolic and diastolic blood pressures post-operatively when performed under local anaesthesia and only the diastolic pressure was reduced when performed under general anaesthesia. CONCLUSION: The study provides evidence about the effect of carotid endarterectomy on the systemic blood pressure and its variations when performed under different types of anaesthesia. There is significant post-operative reduction in both the systolic and diastolic blood pressure values and the intraoperative fluctuation is minimal when local anaesthesia is used. Further studies are required to find out how this affects the long-term blood pressure and clinical outcome of the patient.

Elderly patients with severe head injury in coma from the outset--has anything changed?

Advancing age is known to be a determinant of outcome in head injury. We have sought to discover whether there has been any change in the outcome of elderly patients with severe head injury in Newcastle, where these patients have continued to be treated with maximum intervention. A review of prospectively collected data from the Newcastle Head Injury Database for the period 1990 to 2000 was carried out. All patients aged 70 years and above who had sustained a severe head injury (Glasgow Coma Score of 8 or less from the outset) were included. The Glasgow Outcome Score (GOS) was determined at 6 months. Seventy-one patients were identified. Fifty-seven (80%) died and 2 (3%) were in a vegetative state, 11 (16%) had severe disability, 1 (1%) had moderate disability and no patients made a good recovery. The natural history of this condition remains unchanged and due consideration should be given to this when evaluating interventions for elderly patients with a severe head injury.

Endovascular treatment of anterior communicating artery aneurysms: results of clinical and radiological outcome in Newcastle.

The results of endovascular management of anterior communicating artery aneurysms (ACoAA) using Guglielmi Detachable Coils (GDC) are presented. We detail the clinical and radiological features, and postoperative clinical and radiological outcome in a consecutive series of patients. We have prospectively collected comprehensive data on our patients with SAH since 1989. This study reports on patients admitted between January 1990 and December 1998, and focuses on 30 patients who had their ACoAA endovascularly treated. An independent observer (TE) carried out long-term follow-up in January 2002. Statistical analysis was performed to study the relationship between clinical factors, the radiological morphology of aneurysms and the long-term outcome. The age ranged from 25 to 74 years (median: 54) and endovascularly treated ACoAA were more common in women, 19 (63%) compared with men, 11 (37%). Seventy-seven per cent were in good grade (WFNS 1 & 2) before treatment. Three patients rebled before treatment. The aneurysms were less than 10 mm in maximal diameter in 27/30 patients. The follow-up period was from 1 to 53 months (mean 32.5, median 36.6 months). Excellent outcome was recorded for 11 patients (36.7%), good in seven patients (23.3%), fair in three patients (10%) and poor in four cases (13.3%). Five patients had died (mortality 16.7%). Favourable outcome was achieved in younger patients, women, and in those who presented in grades 1 and 2. The long-term radiological follow-up revealed residual necks in 13 patients. Only one has required retreatment and no rebleed has occurred in 3-6 years. This study reports a contemporary experience with the endovascular management of ACoAA. Long-term stability of the coil and good outcome is demonstrated. Endovascular treatment using GDC is an efficient technique for treating acutely ruptured ACoAA with little additional morbidity. The clinical and radiological results are comparable with those in the literature.

Trends over time in the management of subarachnoid haemorrhage in newcastle: review of 1609 patients.

This article reviews trends in the management of subarachnoid haemorrhage (SAH) at the Regional Neurosurgery Unit in Newcastle over 9 years. This is a comprehensive analysis of prospectively collected data on patients with SAH. We review the changes in clinical therapy and outcome with regards to conservative (non-surgical), surgical and endovascular therapy. Since 1990, the demographic and management/outcome details of patients with SAH have been recorded systematically. This study involves patients admitted over the 9 years, from January 1990 to December 1998. The data were computerized using Microsoft Access (Microsoft Inc. USA), and analysed using SPSS statistical package. A total of 1609 had aneurysmal SAH confirmed with CT, lumbar puncture and/or angiography. Sixty-seven per cent (1,073 patients) were female with a female to male ratio of 2:1. This ratio was maintained from 1990 to date. The mean age has slowly increased from 49 years in 1990 to 55 years of age in 1998, (range 18-91). Overall, 53.9% (from 66.3% in 1990 to 35.3% in 1998) were surgically treated, 8.1% had embolization (range 0.6-18.4%) and 38% (range 28.2-46.4%) were managed without surgical intervention for the aneurysm. The proportion of patients undergoing surgery has decreased since 1994 with improvements in endovascular therapy, participation in the ISAT trial and increased admission of poor grade patients (WFNS grades 4 and 5, from 17% in 1990 to 31% in 1998). The mortality rate has doubled over the years under review (18-32%). The percentage of severely disabled patients has remained constant at about 7% with none in a vegetative state. Only 54% had a favourable outcome in 1998 compared with 78% in 1990. Total morbidity and mortality has increased particularly during the last 3 years. This has been associated with double the number of admissions in grade 5. Favourable outcome occurred in 90% of good grade patients (WFNS 1 and 2) with 6.2% mortality in surgical candidates and 5.5% in patients treated endovascularly. The mortality for poor grade (WFNS 4 and 5) patients was 64%.