Teamwork in the Scottish ICU

T Reader, R Flin and B Cuthbertson School of Psychology, William Guild Building, University of Aberdeen, AB24 2UB.

Studies of patient safety have indicated the importance of having effective teamwork within the intensive care unit (ICU). In particular, lapses in team communication between ICU nurses and doctors have been found to be an important factor in the occurrence of preventable medical errors.1,2 Pronovost et al. 3 have also cited the importance of open communication between nurses and doctors, in order to create an environment where it is safe for all individuals to participate and speak up when necessary. This is consistent with many high-risk industries, where teamwork, and the processes of teamwork (e.g. communication), are recognised as being crucial.4 Due to the role of communication in medical errors, it would appear important to measure the perceptions of nurses and doctors with regards to factors influencing the quality of communication in the ICU. To data, relatively little research has been done in the UK ICU environment.

The current study reports on the perceptions of ICU nurses and doctors with respect to communication in the ICU. Employing a questionnaire tool used previously within the US, the study examines perceptions of teamwork in a number of Scottish ICUs. The questionnaire has been used across the US, with associations between interdisciplinary communication, patient length of stays, and risk-adjusted mortality rates being found.5,6 The questionnaire contains items that measure the quality of communication in the ICU between disciplines (i.e. nurses and doctors), and within disciplines (i.e. senior and junior doctors). Also measured by the questionnaire are perceptions of leadership, satisfaction with communication, understanding patient care goals, and perceived unit effectiveness. The preliminary results show that, overall, staff in ICUs have generally high perceptions of teamwork, similar to the US norm data. However, significant differences in perceptions of interdisciplinary communication openness were found between nurses and doctors. Also, there were significant associations between the leadership and communication scales, and between quality of unit communication and reported understanding of patient care goals.

References 1. Beckmann U, Baldwin I, Hart GK, Runciman WB: The Australian Incident Monitoring Study in Intensive Care: AIMS-ICU. An analysis of the first year of reporting. Anaesth Intensive Care 1996; 24:320-329 2. Donchin Y, Gopher D, Olin M, et al: A look into the nature and causes of human errors in the intensive care unit. Crit Care Med 1995; 23:294-230 3. Pronovost PJ, Wu AW, Sexton JB: Acute decompensation after removing a central line: Practical approaches to increasing safety in the intensive care unit. Ann Intern Med 2004; 140:1025-1033. 4. Cannon-Bowers J, Salas E. Teamwork competencies: The interaction of team member knowledge, skills and attitudes. In: O'Neil O, ed. Workforce readiness: Competencies and assessment. Mahwah, NJ: Erlbaum, 1997 5. Baggs JG, Schmitt MH, Mushlin AI, Mitchell PH, Eldrege DH, Oakes D. Association between nurse-physician collaboration and patient outcomes in three intensive care units. Crit Care Med 1999; 27:1991-1998. 6. Shortell SM, Zimmerman JE, Rousseau DM, et al. The Performance of Intensive Care Units: Does Good Management Make a Difference? Med Care 1994; 32:508-525.

The funding for this project comes from a PhD studentship awarded by the College of Life Sciences and Medicines, University of Aberdeen.

Recovery from the anaemia of critical illness is associated with resolution of the inflammatory state despite a depressed erythropoietin response.

A Bateman and T S Walsh. Department of Critical Care, The Royal Infirmary of Edinburgh, Little France Crescent, Edinburgh EH16 4SA.

Background: Anaemia is present in 80-90% of all patients at intensive care unit (ICU) discharge and persist for long periods in many patients.1. In chronic conditions such as renal failure, anaemia is associated with impaired quality of life and significant morbidity, which can be improved by treatment.2 We investigated the factors contributing to the persistence of anaemia after critical illness. We present data relating anaemia recovery to erythropoietin response and persistent inflammation.

Methods: Patients who received >24 hours of invasive ventilation and/or >2 organ support were screened for the presence of anaemia at discharge from the medico-surgical ICU of the Royal Infirmary of Edinburgh. Exclusions included ongoing requirement for renal replacement therapy, immunosuppresion or known chronic haematological condition. 19 of 30 enrolled patients completed followed up at 1, 3, 6, 9, 13 and 26 weeks post ICU discharge either in hospital or their homes. Blood samples were taken at each visit to measure haemoglobin concentration (Hb), reticulocyte count (retics), C-reactive protein (CRP) and serum erythropoietin (EPO) concentration. For analysis, patients were assigned to 2 groups depending upon whether or not Hb levels had normalised by 13 weeks following ICU discharge (7 'responders' and 12 'non-responders'). We compared measured parameters at 3 weeks post-ICU discharge to explore associations with poor recovery in Hb.

Results: CRP was higher and reticulocyte count lower among non-responders, but there was no difference in erythropoietin concentrations, which were inappropriately low in both groups (figure and table).

Table 1. Mann Whitney Test (paired comparison 3 weeks post-ICU discharge) Parameter P-value CRP 0.005 EPO 0.96 Reticulocytes 0.01

Conclusion: Among anaemic patients discharged from ICU erythropoietin response is inappropriately low. Slow/non-recovery of anaemia is associated with persistent inflammation in the post-ICU period and a hypo-responsive bone marrow.

References 1. Saleh, E, Walsh, T. S. Prevalence of anaemia among survivors of critical illness managed with conservative transfusion triggers. Br J Anaesth 2003;90. 2. Canadian Erythropoietin Study Group. BMJ 1990; 300

Neurones Express Macrophage Inflammatory Protein-2 Following Traumatic Brain Injury in the Rat

JKJ Rhodes, PJD Andrews and J Sharkey. Astellas CNS Research in Edinburgh, the University of Edinburgh, the Chancellor's Building, 49 Little France Crescent, Edinburgh, EH16 4SB.

The expression of inflammatory mediators and the recruitment of leucocytes into the acutely injured brain are implicated in the pathogenesis of secondary brain injury.1. However the inflammatory response to brain injury, its control and modulation remain incompletely described. The neutrophil chemotactic cytokine macrophage inflammatory protein-2 (MIP-2) is expressed invitro by glial and cerebral vascular endothelial cells. The invivo expression of MIP-2 in response to traumatic brain injury has been described previously. The production of this chemokine by glia has been implied but not demonstrated.2

We have investigated the cellular localisation of MIP-2 in the lateral fluid percussion model of focal brain injury. In accordance with the Animals (Scientific Procedures) Act 1986 and after review by the animal procedures committee, anaesthetised male Sprague Dawley rats received a moderate (1.7-2.0 atm) lateral fluid percussion injury. At 0, 4, 8, 12 & 24h after injury brains were harvested and dissected into anatomical regions. MIP-2 levels in the cortex were analysed by ELISA.

After injury MIP-2 was significantly increased in the injured cortex, peaking at 4 h after injury and declining rapidly to baseline by 12 h. Imminohistochemical staining of coronal sections from 4 h after recovery with an anti-MIP-2 antibody and the neuronal marker anti-NeuN localised MIP-2 expression to the cytoplasm of shrunken necrotic neurones.

Invitro chemokines are expressed by glial in response to pro-inflammatory cytokines. We have localised MIP-2 expression following focal brain injury to necrotic neurones. Our invivo results suggest that chemokine release may also be a fundamental primary response to tissue damage in the brain, initiating neutrophil chemotaxis.

1. Menon DK, Matta BF. Intensive Care After Acute Head Injury. Matta BF, Menon DK, Turner JM (eds). Textbook of Neuroanaesthesia and Critical Care London: Greenwhich Medical Media Ltd, 2000:35-49. 2. Otto VI, Stahel PF, Rancan M et al. Regulation of chemokines and chemokine receptors after experimental closed head injury. Neuroreport 2001;12:2059-2064.

S McLeod, C Lee, D Campbell and S Crofts Department of Anaesthesia and Intensive Care, Ninewells Hospital, Dundee DD1 9SY

The United Kingdom Safer Patient Initiative, a collaborative between the Institute for Healthcare Improvement & the Health Foundation, aims to improve patient care by ensuring the consistent delivery of best practice. It has been shown that changes in process can lead to changes in outcome. Measurement of the process rather than outcome results in rapid and meaningful data with feedback. One initiative is utilisation of Daily Goals to improve interdisciplinary communication and provide clear targets that are patient centered. When this model was implemented in one Intensive Care Unit (ICU) there was a fall in length of stay from 2.2 days to 1.1 days.1 We developed and implemented a 5 element Daily Goal chart for our unit.

Compliance in use of Goals was confirmed in a previous audit at 95.4% in our unit (unpublished data). We have now looked in more detail at the content and success of each Goal which were categorised. It was expected for all goals, with the exception of Parameters, that a time should be set to meet the goal. The attainment or not of each goal was recorded. Data collection was carried out by a member of staff not connected to the setting of the Daily Goals.

Data was collected for 133 patient days out of a potential 142 potential patient days (93.6%) over 4 weeks. 540 goals set with 496 (92%) being met. Only 48.9% of appropriate goals were time targeted. 34 goals were not met and in a further 10 the outcome was unknown. When categorised 65.6% were Tasks 95.7% with completed: 18.9% were Communications with 88.2% completed: 12.6% were Parameters with 76.4% completed: 2.2% were Miscellaneous with 100% completed: 0.7% were Hardware Removal with 75% completed.

The setting of daily goals should be multidisciplinary; this was not the case with most goals being medically orientated. The setting of more parameters would perhaps allow the multidisciplinary team to use their unique skills in patient care rather than using a medical model Communication is always an area of concern in dealing with the patient, their relatives and others, as it is usually failings in communication that lead to dissatisfaction with hospital care 18.9% of all goals focused on this with 88.2% of these goals being met. We are not yet appropriately completing goals within a time limit.

It is too soon yet to see if this intervention has led to decreased length of stay. The development process is still ongoing but it has become evident that both the methodology and the monitoring of process are powerful tools in the provision of care for an ICU patient.

References 1 Improving Communication in the ICU Using Daily Goals Pronovost P, Berenholtz S, Dorman T et al J Crit Care 2003 Jun; 18(2):71-75

Mortality in patients with alcoholic liver disease admitted to intensive care: assessment of a new scoring system.

C Goutcher, J Edwards, E Forrest and L Donaldson. Intensive Care Unit, Glasgow Royal Infirmary, 84 Castle Street, Glasgow G4 0SF.

Mortality is high in patients with alcoholic liver disease (ALD) who require intensive care (ICU) admission. Several scoring systems have previously been assessed to attempt to predict mortality in this group. Using these scoring systems in clinical practice has been difficult either because of complexity (e.g. APACHE II) or because of lack of objectivity (e.g. Child-Pugh). The Glasgow Alcoholic Hepatitis scoring system (GAHS)1 is a new, objective scoring system which has been developed and validated for predicting mortality in patients with alcoholic hepatitis. It assigns scores for age, white cell count, urea, prothrombin ratio and bilirubin to give a total score between 5 and 12. This system has not previously been assessed for predicting mortality in patients with ALD admitted to ICU.

We carried out a search of the Ward-Watcher computer database in our ICU. We identified all patients with ALD admitted to ICU from January 2000 to June 2005. Case notes / laboratory databases were checked to confirm the diagnosis of ALD and to get details of the admission. The GAHS was calculated. As in Forrest's study patients were divided into two groups (scores of 5-8 or 9-12). We also calculated a total score for the Cardiovascular (CVS) and Renal sections of the Sepsis-related Organ Failure Assessment (SOFA) score. Likelihood ratios were calculated for each group.

63 patients were identified. Overall ICU mortality was 63% (hospital mortality 74%) consistent with previous studies. ICU mortality in the GAHS 5-8 group was 53% (38/63) compared with 80% in the 9-12 group (25/63). Corresponding likelihood ratios (95% Confidence Intervals) for ICU death were 0.6 (0.4-0.9) and 2.3 (1.0-5.3). The total score for the CVS and Renal sections of the SOFA score combined with GAHS gave ICU mortalities as follows:

Patients with alcoholic liver disease who are admitted to ICU have a high mortality. Reversibility of the acute critical illness is often controversial for patients with ALD. GAHS is a simple to use, objective method for predicting those sub-groups who are more likely to survive and would benefit from aggressive ICU support, particularly when combined with the CVS and Renal section of the SOFA score.

References 1. Forrest EH, Evans CDJ, Stewart S, et al. Analysis of factors predictive of mortality in alcoholic hepatitis and derivation and validation of the Glasgow Alcoholic Hepatitis score. Gut 2005: 54: 1174-1179.

Case note retrieval funded in part by a grant from the Glasgow University Department of Anaesthesia Endowment Fund.

Management of sepsis and septic shock in critically ill patients transferred by a dedicated transport team in the West of Scotland

AJ Cadamy, I Thomson and A Binning Intensive Care Unit, Western Infirmary, Dumbarton Road, Glasgow G11 6NT

Over 500 critically ill patients are transferred by a dedicated critical care transport team each year in the West of Scotland. With the advent of international guidelines for the management of severe sepsis and septic shock there are now criteria by which the management of these conditions can be assessed.1 We undertook a prospective audit over a three month period to determine what proportion of the patients transferred have these conditions and to determine how management conforms to the guidelines.

Patients were deemed to have sepsis if they had suspicion of infection and two or more of the following: Temp. >38oC or <36 oC; WCC <4 or >12 x103 mm-3; HR >90min-1; RR >20min-1; SBP <90mmHg or MAP <65mmHg or needing a vasopressor. We adapted sepsis resuscitation bundles derived from the guidelines and devised a data collection form with relation to the following: serum lactate measurement; blood cultures prior to antibiotics; antibiotics given within 3hrs for A&E <1hr for non-A&E referrals; MAP <65mmHg and management with a minimum 20mlkg-1 fluid challenge, vasopressors, and CVP monitoring; achievement of MAP =65mmHg; measurement of central venous oxygen saturation (ScvO2)2. Data were collected for every patient transferred during June, July, and August 2005.

90 patients were transferred from a total of 19 different hospitals during the audit period. Data was available for 82 (91%). 45 patients (55% 95% CI 44-66%) met criteria for sepsis. Of these 8 patients had blood cultures prior to antibiotics (18% 95% CI 8-32%), and in 23 (51% 95% CI 35-66%) this information was not available or unclear. Similarly, 24 patients (53% 95% CI 28-68%) had antibiotics within the time window, and in 17 (38% 95% CI 24-54%) this was unclear. 28 (62% 95% CI 47-76%) patients had circulatory failure with 19 of these (68% 95% CI 48-84%) requiring more than a fluid challenge alone. MAP =65mmHg was achieved in 43 patients (96% 95% CI 85-100%). Two patients with sepsis had serum lactate measured (4% 95% CI 0.5-15%). 1 patient of the 19 who had not responded to a fluid challenge had ScvO2 measurement (5% 95% CI 0-26%).

A significant number of critically ill patients with sepsis and septic shock are transferred each year in the West of Scotland. Many have circulatory failure and this is managed consistently with fluids, vasopressor and CVP targeting. Documentation and communication of blood culture withdrawal and antibiotic therapy appears to be poor. Serum lactate and ScvO2 measurement did not appear to be part of the routine management of patients with sepsis referred for transport. The former may be due to the lack of availability of access to the assay. However, as the use of ScvO2 measurement and targeting merely requires access to a blood gas analyser, the latter finding may suggest that there is a reluctance to apply some aspects of the recommendations in this group of patients.

References: 1. Surviving Sepsis Campaign guidelines for the management of severe sepsis and septic shock. Crit Care Med 2004; 32(3): 858-73 2. Institute for Healthcare Improvement. www.ihi.org

Non-invasive assessment of central aortic haemodynamics and endothelial function in critical care

MJ Duffy, DF McAuley, P Glover and BA Mullan Regional Intensive Care Unit, Royal Victoria Hospital, Grosvenor Road, Belfast, BT12 6BA

Non-invasive pulse wave analysis (PWA) methods can generate the ascending aortic pressure wave from the radial/brachial pressure wave.1 In cardiovascular disease, measurements derived from central aortic pressure waveform analysis have been reported to be strong independent predictors of cardiovascular mortality. Moreover, PWA combined with endothelium-dependent ß2-adrenergic vasodilation has been shown to be a simple, repeatable, non-invasive means of assessing endothelial function in vivo.2 The use of PWA in critical care has not previously been reported. We present pilot data from 13 patients admitted to our intensive care unit (ICU) and 10 age-controls.

PWA measurements were performed using the SphygmoCorTM Mx system. The following central aortic haemodynamic variables were determined: the aortic augmentation index (AIx - measure of systemic arterial stiffness), the time to wave reflection (Tr - measure of aortic stiffness), and the Buckberg subendocardial viability ratio (SEVR - measure of subendocardial perfusion). Data are mean ± SD.

Compared with the control subjects, in the ICU patients there was a significant reduction in Tr (155.5 ± ?12.2ms vs 126.2 ± 13.3ms, p<0.01) and SEVR (171.2 ± 20.1 vs 114.5 ± 27.2, p<0.01), despite both groups having similar peripheral blood pressures. When the ICU patients were subdivided into septic (n=5) and non-septic (n=8), the septic patients had a greatly reduced AIx (-5.4 ± 17.0% vs 28.8 ± 11.9%, p<0.01) and a severely impaired response to endothelium-dependent ß2-adrenergic vasodilation (Figure 1).

As well as providing additional haemodynamic information useful in monitoring the critically ill, data derived from PWA may inform prognosis in the critically ill. Further research is warranted to determine the usefulness of these variables in critical illness. Of particular interest, PWA may be a useful non-invasive means by which to assess arterial endothelial dysfunction in sepsis.

References: 1) Soderstrom S, Nyberg G, O'Rourke MF, Sellgren J, Pontén J. Can a clinically useful aortic pressure wave be derived from a radial pressure wave? Br J Anaesth 2002; 88: 481-88. 2) Wilkinson IB, Hall IR, MacCallum H, et al. Pulse-wave analysis. Clinical evaluation of a noninvasive, widely applicable method for assessing endothelial function. Arterioscler Thromb Vasc Biol 2002; 22: 147-52.

A Hutchinson1, C McAllister2, G Lavery1, P Caddell1, J Adams1 and P Glover1. 1Regional Intensive Care Unit, Royal Hospitals Trust, Grosvenor Road, Belfast, BT12 6BA. 2Intensive Care Unit, Craigavon Area Hospital, 68 Lurgan Road, Portadown, BT63 5QQ.

Intensive care units are complex patient management environments in which critical incidents occur frequently.1 The design of this project is based on the previous experience of the Australian Incident Monitoring Study in Intensive Care Units, a well established system, which has proven to be successful.2 By recording and analysing incidents, we aim to identify patterns and system based failures amenable to change, thus improving patient safety.

The project was undertaken in a 17 bedded ICU / 8 bedded HDU critical care facility over 5 months (May-October 2005). A database, specifically designed for the anonymous collection of incident data, relevant to an intensive care environment, was installed in the unit. The database was designed to be user friendly, to maintain patient and staff anonymity and to prevent access to the data by anyone other than the local coordinators at a later date. An incident was defined as "any event that led to, or could have led to, patient harm if it had been allowed to proceed. It may, or may not, have been preventable and it may, or may not, have involved error." Details were requested on various aspects of the incident and contributing factors, as well as outcome. The system complemented the formal hospital incident report system.

Seventy-five incidents were reported over the 5-month project period. The majority of incidents involved airway management issues (39%), followed by procedures and lines (27%), drug errors (17%), unit management (15%) and environmental problems (3%). Medical staff precipitated 29% of incidents and nursing staff 48%. Most incidents were detected by nursing staff (80%). Incidents were detected within 1hour of their onset in 63% of cases. Reported incidents resulted in either no, or only a minor, physiological change in the patient's condition in the majority of cases (74%). During the pilot project there were 19 incidents reported through the existing hospital critical incident reporting system.

The anonymous incident reporting scheme has been well adopted by staff. As a result of this project, change has been implemented to reduce further adverse events, and a decision to continue using the database has been made. The system gives staff more direct feedback regarding incidents and staff members appear to be more comfortable using an anonymous system, reflected by the increased numbers reported. The Northern Ireland Incident Monitoring Study will prospectively audit adverse events in several intensive care units in the province.

1. Rothschild JM, Landrigan CP, Cronin JW, et al. The Critical Care Safety Study: The incidence and nature of adverse events and serious medical errors in intensive care. Crit Care Med 2005;33:1694-700. 2. Beckmann U, West LF, Groombridge GJ, et al. The Australian Incident Monitoring Study in Intensive Care: AIMS-ICU. The development and evaluation of an incident reporting system in intensive care. Anaesth Intensive Care 1996;24:314-319.

AHJ McDonald and DG Swann Critical Care, Royal Infirmary, 51 Little France Crescent, Edinburgh EH16 4SA

Severe sepsis has a mortality of 21.2-47.3%.1 Activated Protein C (APC), a mediator of the inflammatory and coagulation systems, reduced hospital mortality of severely septic patients from 34.6% to 29.4% in a randomised controlled trial.2,3 This therapeutic benefit outweighed the risk of bleeding complications in septic patients with multiple organ failure or an APACHE =25.3

This study assessed the outcomes of patients treated with APC in terms of mortality and organ failure through retrospective analysis of medical records. Expected mortality data was calculated 24 hours after admission to Intensive Care using the APACHE II score. Organ failure was assessed one day before, the four days during, and one day after treatment with APC using the Sequential Organ Failure Assessment (SOFA) score, without the neurological aspect - this was impossible to assess retrospectively.

APC was administered to 48 patients, who on the day of administration had 3-5 failing organ systems. The hospital mortality for the group was 33.3%. Predicted hospital mortality using the APACHE II score and diagnosis was 54.2% (48.2-60.2% CI). This gave an SMR of 0.61.

The figure illustrates daily SOFA scores of 40 patients with complete data. Mean SOFA scores rose in the day before treatment with APC and then fell. The fall in mean SOFA score reached statistical significance, at p<0.05, on day 4 onwards. (The paired t-test was used after confirmation of normal distribution. For non-parametric data, Wilcoxon's test was used)

By day 5, four patients had died and one transferred to another hospital. Four cases of non-fatal bleeding were reported.

Our patients with severe sepsis who received APC had a lower than expected hospital mortality using the APACHE II prediction model. Their hospital mortality was between those of the treatment and control arms of the PROWESS study. Bleeding complications were not a major problem. There was a significant improvement in organ failure after infusion of APC.

1. Padkin A, Goldfrad C, Brady AR et al. Epidemiology of severe sepsis occurring in the first 24 hrs in ICUs in England, Wales & Northern Ireland. Crit Care Med 2003; 31, 2332-2338 2. Angus DC, Laterre PF, Helterbrand J, et al: The effect of drotrecogin alfa(activated) on long-term survival after severe sepsis. Crit Care Med 2004; 32:2199-2206 3. Bernard GR, Vincent JL, Laterre PF, et al. Efficacy and safety of recombinant activated protein C for severe sepsis. NEJM 2001; 344, 699-707

Cervical spine clearance in the multiply injured, unconscious patient: current practice in the West of Scotland.

M Burwaiss and P.Edgar. Neurointensive Care Unit, Institute of Neurological Sciences, Southern General Hospital, 1345 Govan Road, Glasgow G51 4TF.

In the conscious trauma patient, clinical examination of the cervical spine (C-spine) offers the most sensitive, specific and cost effective method of excluding injury. Unfortunately in the unconscious or multiply injured population presenting to Intensive Care such clinical clearance is rarely possible. Most (90 - 95%) of these patients do not have a C-spine injury and keeping them immobilised unnecessarily for long periods is associated with a number of potentially life-threatening complications. Recently several groups have attempted to outline an optimal imaging strategy to identify those patients in whom it is safe to discontinue spinal immobilisation1, with most suggesting a combination of plain radiographs and CT of the entire C-spine within 72 hours of admission as a reasonable standard.

Our unit receives unconscious trauma victims from all over the West of Scotland and we are regularly faced with the problem of "clearing" the C-spine radiologically. There is wide variation between different referring units in the type of C-spine imaging performed. To clarify this issue further we conducted a telephone survey of eleven hospitals in the West of Scotland. We contacted the consultants on-call for both radiology and intensive care and asked them about their current practice in imaging the C-spine in unconscious trauma patients.

Seven hospitals would perform three standard plain radiographs in this situation while three would perform only a lateral C-spine view. All centres would readily progress to CT scanning: eight would scan the entire C-spine while three performed limited scans. All CT scanners were able to complete adequately detailed scans within minutes and a formal report from a consultant radiologist would be available within 24 hours of the scan taking place. If imaging was normal two units would "clear" the cervical spine while the patient remained unconscious: four hospitals awaited return of consciousness irrespective of imaging. The remaining five units would "sometimes" clear the neck in this patient group after discussion with orthopaedic surgeons and radiologists, but admitted this was "very operator dependent". Four hospitals had a protocol for this scenario, although each protocol suggested a different management strategy. One clinician felt that protocols were best avoided in this area.

Little consensus in the management of this patient group was found despite several recent high profile publications. Most hospitals performed and reported the investigations listed above within the suggested time frame, however only two were prepared to act on this and discontinue spinal immobilisation if imaging was normal. Six units did not have a protocol for this issue and felt there should be one.

1 - Morris CGT, McCoy E. Clearing the cervical spine in unconscious polytrauma victims, balancing risks and effective screening. Anaesthesia, 2004, 59. 2 - Morris C, Guha A, Farquhar I. Evaluation for spinal injuries among unconscious victims of blunt polytrauma: a management guideline for intensive care. ICS 2005

E Jack and P McConnell Glasgow Royal Infirmary, Alexandra Parade, GLASGOW G31 2ER.

The vast majority of patients in intensive care units are intubated to allow invasive positive pressure ventilation to occur (>97% in our unit). To provide a definitive airway that offers some protection against aspiration of gastric contents cuffed endotracheal tubes are used - either oral, nasal or via a tracheostomy. Most tubes are now made with a 'high volume, low pressure' profile and the recommendations are to keep the intra cuff pressure (Pcuff) between 20 & 30mmHg. This is said to reduce the chance of mucosal ischaemia and subsequent tracheal stenosis. Despite this there remains a reported incidence of 1% clinical stenosis after all intubations and a much higher prevalence in the ICU population of 11% suffering a 10 - 50% stenosis1 or up to 20.3% with a clinically determinable stenosis.2 We performed a completed audit loop within our unit which has no current guidelines on recording Pcuff.

We present the findings and discuss the interventions that we attempted: The data collected included Pcuff as measured by a manometer via a 100mm catheter taken early in the morning with the patient semi supine.The first stage survey in October revealed little knowledge of what Pcuff was desirable or the possible implications of excessive Pcuff amongst our nursing colleagues. Two doctors then undertook a simple education policy of talking on a one to one basis with all nursing staff during their shifts and introducing them to the pressure manometer. We ensured that they understood the implications of high Pcuff and how to avoid it. We supplemented this with posters around the unit during the intervening weeks and placed a reminder of what pressures were desirable at each bed space.In December we repeated the survey collecting the same data as before and the results are presented below

Our results show no significant improvement in any aspect. We have shown that simple measures to reduce morbidity fail to change practice despite intensive interventions by enthusiastic staff. We need to investigate further interventions that may improve our results before performing a third audit loop.

Reference: 1. Stauffer JL, Olson DE, Petty TL. Complications and consequences of tracheal intubation and tracheostomy. A prospective study of 150 critically ill adult patients. American Journal of Medicine 1981; 70: 65 - 76 2. Richards I, Giraud M, Perrouin-Verbe B et al. Laryngotracheal stenosis after intubation or tracheostomy in patients with neurological disease. Arch Phys Med Rehabil 1996; 77: 493-6

JG McCormack and IS Grant Intensive Care Unit, Western General Hospital, Crewe Road, Edinburgh, UK, EH4 2XU

Objective: Arterial carbon dioxide levels may be measured non-invasively using a transcutaneous monitoring system.1, 2 This study aimed to assess the value of this technique and compare accuracy with arterial blood gas sampling in adult intensive care patients.

Design and Setting: A prospective single centre observational study, in a ten-bed teaching hospital intensive care unit (ICU).

Patients and Measurements: Simultaneous arterial (PaCO2), end-tidal (EtCO2) and transcutaneous (PtcCO2) carbon dioxide levels were recorded in thirty patients admitted to the intensive care unit requiring mechanical ventilation. Details of primary pathology, ventilatory support and use of vasoactive agents were recorded. In a subgroup of patients recording continued over 16 hours to assess inaccuracy due to drift.

Results: 90 sets of complete recordings were obtained from 30 patients, with mean age 53y (29-81y) and 14 patients (47%) defined as having severe respiratory failure (PaO2/FiO2 ratio = 26.7kPa). Mean differences and standard deviations between carbon dioxide measurements are demonstrated in table 1.

Bland-Altman analysis of PtcCO2 vs PaCO2 demonstrated a mean bias of +0.45kPa (limits of agreement: -0.58 to 1.48kPa). Concurrent intermittent PaCO2 and PtcCO2 sampling for up to 16 hours demonstrated that mean PtcCO2-PaCO2 difference did trend upwards from 0.47kPa at 0-3 hours to 0.94kPa at 16 hours though this was not statistically significant (p=0.08). Of all other parameters measured, including presence of respiratory failure, level of positive end expiratory pressure (PEEP) applied, age, length of ICU stay and use of vasoactive therapies, only respiratory failure and higher PEEP levels significantly affected recordings between PaCO2 and PtcCO2.

Conclusion: The transcutaneous monitor was found to be accurate, with superior accuracy to end-tidal capnometry, giving meaningful results which may enhance the management of intensive care patients in selected clinical circumstances.

References 1Dullenkopf A.. Paediatric Anaesthesia 2003; 13: 777-784. 2McBride DS. Southern Medical Journal 2002; 95: 870-874

Statistical analysis of blood glucose and 28 day hospital mortality for Scottish intensive care patients.

KD Rooney, MJWatson, A McConnachie, FN MacKirdy and SJ Mackenzie. Department of Anaesthesia and Intensive Care, Royal Alexandra Hospital, Paisley, PA2 9PN.

The aim of this study was to investigate the association between serum glucose and mortality in general intensive care patients, after taking account of baseline mortality risk, as measured by the APACHE II risk scoring system. Data was collected prospectively between 01/01/1995 and 31/12/1997 on admissions to Scottish intensive care units (ICU). Lowest and highest blood glucose concentrations were collected as part of the minimum data set. We analysed data recorded from 12,649 patients during their first 24 hours after admission to intensive care. Logistic regression modelling was used to test the association between the highest and lowest blood glucose and 28 day mortality. The 24-hour highest blood glucose (p<0.0001) and the 24-hour lowest glucose (p<0.0001) were independently associated with in-hospital 28-day mortality. The relationship between 24 hour highest glucose and in-hospital mortality was further characterised by logistic regression modelling. The patients were subdividing into low (9.3%), medium (25.4%) and high (53.7%) mortality subgroups by APACHE II predicted in-hospital mortality. For 24-hour highest blood glucose levels between 6 and 30mmol/l all three mortality sub-groups had linear associations with the log odds of the in-hospital mortality rate.

This model showed an increased adverse effect of elevated 24-hour highest glucose in the high mortality subgroup of patients. In summary increased severity of illness was associated with an increased adverse effect of high serum glucose and potentially increased benefit from tight glycaemic control.

Acknowledgements: The audit department of the Scottish intensive care society for providing funding for this study and access to intensive care audit data.

The Robertson Biostatistics Department of Glasgow University for their invaluable assistance during the analysis of this data.

Implementaion of guideline based "sepsis bundles" in Aberdeen Royal Infirmary: a pilot audit.

I McLardy and G Houston Aberdeen Royal Infrimary, Forresterhill Road, Aberdeen AB25 2ZN

The Surviving Sepsis Campaign (SSC) was formed in 2002 and is a collaborative world wide, cross speciality group whose main goal is to reduce the incidence of sepsis mortality by 25% within 5 years.1 The group developed a core set of evidence based guidelines for the management of sepsis, which have been incorporated into "sepsis bundles" allowing for easier evaluation of treatment and clinical outcome in a format which is transferable between institutions facilitating audit.

The purpose of this audit was two fold. Firstly to see how easy these guidelines were to institute within our hospital, and secondly to get a baseline idea of how sepsis is currently managed within our institution so as to allow comparison in the future should changes in the management of sepsis be implemented.

Over a 6 week period all patients in ICU at Aberdeen Royal Infirmary were screened daily using a severe sepsis screening tool. Those meeting the criteria for severe sepsis were entered into the severe sepsis bundle pathway. Subsequently an individual chart measurement tool was completed for each patient looking at compliance with individual components of the severe sepsis bundles.

Data was collected for 23 patients over a 6 week period. At the time of writing this abstract the data has not been fully analysed, due largely to the fact that SSC database has not been made available. However we anticipate this to be available within the near future, and should have full data analysis completed by mid January.

References and acknowledgements: 1. Dellinger RP, Carlet JM, Masur H, et al: Surviving Sepsis Campaign guidelines for management of severe sepsis and septic shock. Crit Care Med 2004; 32 858-873

A Survey of Patients Requiring Ventilation on the Queen Elizabeth National Spinal Injuries Unit

CJ Rimmer and P Edgar Queen Elizabeth National Spinal Injuries Unit, Southern General Hospital, Govan Rd, Glasgow, G51 4TF

The Queen Elizabeth Spinal Injuries Unit (QENSIU) accepts patients from the whole of Scotland. The acute ward has 12 beds and is able to accept 3 ventilated patients at one time. This was a retrospective survey covering a 3-year period from 1st January 2002 to 31st December. 2004. During this time 509 patients were admitted to the spinal injuries unit as a whole, of which 33 were ventilated or required ventilation for a period during their stay. 29 case notes were available for review.

The two most common modes of injury in these patients were road traffic accidents (34%) and falls involving alcohol (34%). Falls not involving alcohol accounted for 10% and the remaining 17% were caused by various accidents, (horse riding, diving, industrial, hit by falling debris), an intra-operative event and an epidural abscess. These figures are comparable with the Spinal Injuries Association figures.1 The age range of these patients was 19y to 80y, with a mean of 48y. The level of their injuries ranged from C1 to T7 and the number of days they were ventilated ranged from 1 to 312. As with other reports the main cause of injury in the older sub-group (59<) was falls.2,3

Most of the patients were transferred to the QENSIU from other hospitals, only 5 being transferred from other departments within the same hospital. 13 of the 29 were not ventilated for transfer, including the 2 patients with C1 and C2 fractures. Most coped well with the transfer and were not ventilated until several days later, but 4 needed ventilation the same day of which 1 was described as being in extremis on arrival necessitating an arrest call. Outcomes of Patients Ventilated on the QENSIU Jan 2002 to Dec 2004

Older patients with these injuries have a greater mortality2 and patients with higher cervical lesions require a longer hospital stay.

References 1. Spinal Injuries Association Annual Report 2004 2. Irwin ZN, Arthur M, Mullins RJ, Hart RA Variations in Injury Patterns, Treatment and Outcome for Spinal Fractures and Paralysis in Adult vs. Geriatric Patients. Spine 29: 796-802, 2004 Apr 3. Brolin K, von Holst H, Cervical Injuries in Sweden, a national survey of patient data. Injury Control and Safety Promotion 9: 40-52, 2002 Mar

Intracranial pressure is not related to plasma sodium concentration in patients with fulminant hepatic failure

BJ Shippey, J Davidson and A Lee Royal Infirmary of Edinburgh, 51 Little France Crescent, Edinburgh, EH16 2SA

Murphy et al propose a linear relationship between plasma sodium ([Na]) and intracranial pressure (ICP) in patients with fulminant hepatic failure (FHF).1

We proposed such a relationship in our patients. We identified from our database patients admitted to our hospital with FHF between 2000 and 2005 in whom ICP monitoring was performed. We performed a retrospective case-note review of these patients, and recorded ICP, cerebral perfusion pressure, body temperature, partial pressure of CO2, partial pressure of O2, arterial hydrogen ion concentration, and the temporally closest [Na] at two time points - the time at which the ICP monitor was inserted, and the time at which the highest ICP was recorded. We recorded the number of ICP "treatments" given, the need for barbiturate coma and, the need for therapeutic hypothermia.

We identified 43 patients with a complete dataset. Mean ICP at time of ICP monitor insertion was 19±13.5 cm H2O, when mean [Na] was 133±6.4 mmol l-1. Mean ICP at time of highest ICP was 41±36.2 cm H2O, when mean [Na] was 133±6.1 mmol l-1. We found no linear relationship between [Na] and ICP at either point (figure), nor were the number of ICP treatments, the need for barbiturate coma or the need for therapeutic hypothermia related to [Na] at the time of ICP monitor insertion.

There is no evidence from these preliminary data for a link between plasma sodium concentration and ICP in patients with FHF. A prospective study is planned.

Reference 1. Murphy N et al The Effect of Hypertonic Sodium Chloride on Intracranial Pressure in Patients with Acute Liver Failure. Hepatology 2004; 39: 464-470

D Seath, N Sharma, SA Stott Intensive care Unit, Aberdeen Royal Infirmary, Foresterhill Road, Aberdeen, AB22 2ZN

Introduction. The National Confidential Enquiry in to Patient Outcome and Death (NCEPOD) is part of the National Patient Safety Agency in England and Wales. In 2005 NCEPOPD produced a report commenting on the care of acutely unwell medical patients in England, Wales and Northern Ireland. Although the aim of the study was to review the care of medical patients referred for Level 3 care rather than intensive care practice the report made several recommendations regarding the provision of critical care. These included pre-referral monitoring, the referral procedure and the intensive care admission process. We decided to audit our current practice against these. Methods. We audited 100 intensive care admissions to a University hospital 14 bed intensive care unit over a ten week period. Demographic data was collected using the ward watcher system. Pre-admission data was gathered from case note review and included documentation of physiological parameters, grade of person referring and evidence of senior clinician involvement. Post admission data was recorded after speaking with the receiving team and included time of referral, time between receiving referral and patient being seen and being admitted, time patient first seen by an ICU consultant and whether patient was reviewed on ward prior to admission. No patient was admitted to the unit without the duty consultants approval. Results. Of the 100 admissions 38% were referred by a medical specialty. 30% of the admissions were from theatre, 41% from a ward, 20% from A&E, and 9% from other hospitals. Average APACHE II score was 20.2. Overall ICU mortality was 24%. Referral data. 44% of the patients were referred by a consultant, 49% by an Specialist registrar, and 7% by a Senior House Officer. 82% of patients were either referred by a consultant or had documented evidence of being discussed with a consultant within 4 hours prior to admission (mortality 22%). The mortality of patients where there was no documented consultant involvement was 33%. Respiratory rate was not documented in only 12% of referrals. Three of these patients (median APACHE II 28.3) died. ICU admission. 30% of admissions were reviewed prior to ICU admission. Only 40% of non ICU survivors referred outwith theatre and not from another hospital were reviewed prior to admission by the ICU team. For those reviewed, the median time from referral to review was 30 minutes (range 0-115). Median times from referral to ICU admission were 120 (15-1050) minutes and 90 (30-375) minutes for survivors and non-survivors respectively. Median time to consultant review overall was 30 (0-795)minutes. One patient was not seen within 12 hours of admission. Median time for consultant review for admissions between 08:00-17:59 was 17.5mins (0-60)(42 admissions), between 18:00-23:59 30mins (0-795)(33 admissions) and between 00:00-07:59 240 minutes(15-555) (25 admissions). Non survivors were reviewed quicker than survivors between 0800-17:59 (12.5[0-40] and 20 [0-60]minutes respectively) and between 18:00-23:59 (5 [0-480]minutes and 30 [0-795] minutes respectively) but not between 00:00-07:59 (260 [75-240] minutes and 232 [15-555] minutes respectively). Conclusions. Within the limitations of this audit we have demonstrated that our hospital meets many of the referral, documentation and review audit criteria set down in the NCEPOD document An Acute Problem. It is pleasing to note the high incidence of recording of the respiratory rate. Although an ICU consultant is involved in every admission we do not meet the criteria of reviewing patients outside the ICU prior to admission. Reference NCEPOD. An Acute Problem? May 2005. London.

Title: Assessment of the impact of pharmacist dose adjustment of aminoglycosides and vancomycin in the intensive care unit

Authors: J West, A Boyter, A Thomson, J Kinsella Intensive Care Unit, Glasgow Royal Infirmary, 84 Castle Street, Glasgow. G4 0SF

At Glasgow Royal Infirmary a Therapeutic Drug Monitoring pharmacist liaises with the ward clinical pharmacist regarding drugs with a narrow therapeutic index. Advice is provided to medical and nursing staff on the most appropriate dosing schedule for each patient. On the Intensive Care Unit (ICU) concern was raised about excessive analysis of serum drug concentrations, multiple dosage changes before achieving target concentrations, poorly completed request forms, and difficulties with communication.

The objectives of this study were to design, introduce, and assess the impact of a pharmacist led dose adjustment protocol for aminoglycosides and vancomycin in the ICU. A retrospective longitudinal 2-phase cohort study was carried out to assess the appropriateness of drug level monitoring, before and after the introduction of a pharmacist led dose adjustment protocol. A User Satisfaction Survey was carried out to determine the satisfaction with the pharmacist led protocol. Data were collected over 2 six month periods from 34 patients (33 aminoglycoside & 16 vancomycin courses) before and 48 patients (32 aminoglycoside & 23 vancomycin courses) after the introduction of the new dose adjustment protocol. The new service resulted in a statistically significant fall (from 73 to 19) drug levels which were difficult or impossible to interpret (defined as "practically inappropriate" drug levels) and a non-significant reduction in clinically inappropriate drug levels (drug levels for which there was no clinical need). There was no difference observed in the study drug course length, time on antibiotics, renal impairment, length of ICU stay or mortality. The response to the user satisfaction survey suggested that the new service was highly valued by the ICU staff with 100% of respondents indicating a preference for pharmacist led dose adjustment protocol compared to the previous advice only service.

The pharmacist dose adjustment protocol is a popular service with the ICU staff, many of whom believe that it has optimised patient care with respect to aminoglycosides and vancomycin. No benefit was demonstrated in clinical outcome measures; however the reduction in samples without a sample time is potentially beneficial to patients as it eliminates the need for guesswork when interpreting measured drug levels. Further study is needed to assess whether there is a clinical benefit to patients.

J.P. Johnston, H.Love, P.Glover. Regional Intensive Care Unit, Royal Hospitals Trust, Grosvenor Road, Belfast, BT12 6BA.

Anaemia is a common observation in critically ill patients. The aetiology of this condition is frequently multifactorial and complex. Furthermore, the use of allogenic blood transfusions to correct anaemia, may be associated with adverse outcomes. Blood sampling is a frequently performed intervention in this group of patients, and may contribute to the development of anaemia.2 In this prospective observational study, we examined the volumes of blood taken from critically ill patients during the period of their ICU admission.

Data was collected for 49 patient days. Data collected included blood test performed, indication for sample, volume of blood taken for each day. Table 1 shows the mean (± S.E.) volumes of blood taken from patients during the first week, and remainder of their ICU stay. A two-tailed t-test was used to compare both time periods.

There was no difference in the total volumes of blood taken from patients in the first week or subsequent stay of their admission. However, more blood was taken for arterial blood gas sampling in the earlier part of their ICU stay. The volumes sampled from our patients are less than previously reported volumes.1,2,3. However, as the average length of patient stay in our ICU is 9 days, it can be projected that a patient will lose approximately 460 mls of blood during their admission.

As blood sampling may contribute to anaemia in critical illness, we recommend that clinicians examine their practice and introduce procedures that may minimise blood loss. Such procedures include modified arterial line sampling methods, minimising discarded aspirate volumes (twice the volume of dead space)4 and the use of paediatric bottles for sample collection. Local policies on the indications for blood sampling should be developed.

References. 1. Alazia M, Colavolpe JC, Botti G, et al. Blood loss from diagnostic laboratory tests. Preliminary study. Ann Fr Anesth Reanim 1996:15:1004-7. 2. Corwin H, Parsonnet K, Gettinger A, et al. Blood transfusion in the ICU. Chest 1995;108: 767-71. 3. Smoller B, Krushall M. Phlebotomy for diagnostic tests in adults. N Engl J Med 1986;314: 1233-35. 4. Rickard C, Couchman B ,Schmidt S, et al. A discard volume of twice the deadspace ensures clinically accurate arterial blood gases and electrolytes and prevents unnecessary blood loss. Crit Care Med 2003;31:1654-58.

A Baker, S Chohan Dept Anaesthetics and Intensive Care Monklands Hospital Monkscourt Avenue Airdrie ML6 OJS

Nutrition is a standard of care in intensive care units (ITU). Use of the enteral route is associated with fewer infectious complications compared with total parenteral nutrition (TPN), but not reduced mortality1,2,3,4. Achieving adequate caloric intake is helped by early enteral feeding5 and the use of protocols6,7. Prokinetic agents increase the dose of enteral feed (EF) in patients with high gastric residual volumes and aid passage of small bowel feeding tubes although the impact on patient outcome is uncertain8. An evidence based feeding protocol is in use in our ITU. We wanted to audit the success of our protocol in meeting feeding goals in critically ill adults. We looked at all patients admitted between over 3 months who were expected to stay more than 48hours (n=31). During the 1st 5days of ITU stay, we collected data on demographics, feeding routes, feed dose, reasons for not being fed, prokinetic use, and use of combined EF and TPN. Data was complete on 28 patients. Demographic data was unremarkable for this unit. 69% of patients on day1 and 95% thereafter were fed, 61% and 78% enterally. Overall TPN was used in 7% of patients. 3 patients (5 days) received combined EF and TPN. 45% of those fed enterally received target feed dose in 1st 48hours, 69% thereafter. High residual volumes resulted in not achieving target dose in 27% on day1, 64% thereafter. Overall, 64% of these patients received prokinetic agents. No patients were fed via the small bowel due to lack of equipment. In conclusion, the locally used protocol results in a majority of patients being fed, and fed early in their stay, mostly using the enteral route. We hypothesise that increased awareness, prokinetic use and use of small bowel feeding will lead to a greater use of the enteral route, and a higher delivered dose. 1.Gramlich L, Kichian K, Pinilla J, et al. Does enteral nutrition compared to parenteral nutrition result in better outcomes in critically ill adult patients? A systematic review of the literature. Nutrition 2004;20:843-848 2.Moore F, Moore E, Jones T et al TEN vs TPN after major abdominal trauma: Reduced septic morbidity. J Trauma 1989;29:916-23 3.Kudsk K, Croce M, Fabian T et al Enteral versus parenteral feeding. Ann Surg 1992;215:503-513 4.Alverdy J, Aoys E, Moss G TPN promotes bacterial translocation from the gut. Surgery 1988;104:185-190 5.Heyland DK Nutritional support in the critically ill. A critical review of the evidence. Crit Care Clin 1998;14:423-40 6.Martin CM, Doig GS, Heyland DK et al Multicentre, cluster randomised clinical trial of algorithms for critical-care enteral and parenteral therapy (ACCEPT). CMAJ Jan 2004;170:197-204 7.Spain DA, McClave SA, Sexton LK et al Infusion protocol improves delivery of enteral tube feeding in the critical care unit. JPEN 1999;23:288-92 8.Booth CM, Heyland DK, Paterson WG Gastrointestinal promotility drugs in the critical care setting: A systematic review of the evidence. Crit Care Med 2002;30:1429-35

A S Patel General ICU, New Royal Infirmary of Edinburgh, Little France, Edinburgh, EH16 4SA

Serial arterial lactate concentrations and failure to clear lactate to normal levels have both been correlated with increasing mortality in the Intensive Care Unit (ICU)1,2 and in the Emergency Department (ED)3. This study examines a range of lactate measures to ascertain which is the most accurate in predicting mortality in general ICU patients. It also examines the utility of ED lactates in screening critically ill patients. This is a prospective observational study of 137 consecutive admissions over 8 weeks to a general ICU. Arterial lactates were collected in the ED and the first 24 hours of ICU admission.

Initial ICU (5.8 and 1.9 mmol/L), post-6hr (5.3 and 1.6mmol/L), peak lactate (8.5 and 2.5mmol/L), and lactimes (24 and 1.8hrs) were significantly higher (p<0.0001) and % lactate clearance (15.7% and 41.1%) significantly lower (p=0.0212) in non-survivors than survivors. Initial ED lactate (10.1 and 3.8mmol/L) was also significantly higher (p=0.0224) in non-survivors than survivors. Areas under receiver-operating-characteristic (ROC) curves suggest that the best predictors of mortality are post-6hr (0.86), lactime (0.82) and peak lactate (0.80), whilst initial ICU (0.74), ED lactate (0.73) and % clearance (0.66) are adequate.

Lactate 6 hours post-admission is the most accurate predictor of mortality in general ICU patients, where a lactate of >2.5mmol/L predicts a mortality of 45% compared with 10% when lactate is <2.5mmol/L. Initial ED lactate >9.25mmol/L is predictive at the earliest stage of presentation, when identification of high risk patients might change subsequent resuscitative efforts and improve outcome.

References 1. Smith I, Kumar P, Molloy S, et al. Base excess and lactate as prognostic indicators for patients admitted to intensive care. Intensive Care Medicine 2001; 27: 74-83. 2. Husain FA, Martin MJ, Mullenix PS, Steele SR, Elliott DC. Serum lactate and base deficit as predictors of mortality and morbidity. Am J Surg 2003;185: 485-491. 3. Nguyen HB. Rivers EP. Knoblich BP. Jacobsen G. Muzzin A. Ressler JA. Tomlanovich MC. Early lactate clearance is associated with improved outcome in severe sepsis and septic shock. Crit Care Med 2004; 32:1637-42.

A Conway Morris, M Garrioch Intensive Therapy Unit, Southern General Hospital, Govan Road, Glasgow G51 4TF

Anaemia has been linked to poorer outcomes in patients1, and is common amongst the critically ill population. Transfusion of allogenic blood is a commonly used treatment for this problem, although raising haemoglobin levels in this fashion may not produce improved outcomes2. Attention has been turned to the causes of anaemia in critically ill patients, and it has been suggested that phlebotomy-related blood loss may be a significant contributor3,4. This study set out to determine how much blood was removed from patients and to record changes in haemoglobin and transfusion requirements in patients without active bleeding.

This study took place in a single intensive care unit. All patients admitted over a 30 day period had the amount of blood removed for tests, or as 'dead space' to clear lines, recorded prospectively. Daily haemoglobin (Hb) levels and administered blood products were also recorded for patients without active bleeding.

21 patients were included (out a total possible 24), for a mean period of 6 days. Mean age was 56 (range 22-88), , 62 % were male and median APACHE II score was 19. Mean blood volume removed was 53mls/24 hrs. Non-bleeding patients (n=16) entered intensive care with a mean Hb of 116g/L, which dropped to a mean of 93g/L by day 2 (mean DHb 7g/L/day) followed by a plateau phase. 9 non-bleeding patients required transfusion (median number of units transfused 2, range 1-4).

Several strategies exist to reduce blood removed for testing purposes. Returning 'dead space' blood to the patient and using paediatric sized tubes would reduce losses by 60% with a possible reduction in anaemia and hence transfusion.

References 1Carson JL, Duff A, Poses RM et al: Effect of anaemia and cardiovascular disease on surgical mortality and morbidity. Lancet 1996;348:1058-1060

2Hebert PC, Wells G, Blajchman MA et al. A multicentre, randomised, controlled clinical trial of transfusion requirements in critical care. NEJM 1999;340:409-17

3Greenwood M. Blood gas analysis may lead to iatrogenic anaemia in intensive care. Aust Crit Care 2000;13:30

4Pathmanathan S, Sipolo M, Seneviratne C et al . Effects of phlebotomy on blood use in the critically ill. Chest 2003;124( suppl) 127

Hypoxic response of an organ depends on its blood flow and O2 consumption rate. Brain is the organ that responds first in an alert individual. When cerebral function is impaired as in head injury or depressed as in an anaesthetised individual, cardiovascular system is the next to respond

A 47 year-old man presented with sudden collapse and a brief loss of consciousness. He was seen alert and obeying commands in our A & E. There was a gradual fall in his mental functions with confusion and lack of orientation and presented typically without any change in behaviour, distress or cardiovascular signs of hypoxia. Also we noticed a gradual fall in arterial saturation levels despite supplimental oxygen, and increasing lung crepitations. The chest X-ray was diagnostic of pulmonary oedema. Simultaneously increasing drowsiness made us to get an emergency CT scan of head that was typically massive subarachnoid haemorrhage.

It is very unusual to find a conscious and an alert individual to present without any signs of hypoxia with below normal oxygen saturation. This could be a sinister and should initiate further investigations especially when found without any signs of distress.

Hypoxia affects brain in an awake patient and the heart in an anaesthetized patient.1 Hence hypoxia in a conscious, alert and well oriented individual results in impaired judgment, drowsiness, dulled pain sensibility, excitement, disorientation, loss of time sense, and headache. The rate of ventilation increases and causes dyspnoea , cyanosis and hypertension when severe. The cardiovascular response to hypoxaemia1,2 as seen under anaesthesia is a product of reflex(neural and humoral) and direct effects. The reflex effects occur first and are excitatory and vasoconstrictive. The direct effects occur locally (inhibitory and vasodilatory) and are delayed. The net result is a balance between the two. Studies have attempted to correlate clinical signs with hypoxaemia.3

In a patient with head injury or with a neurological condition that is affecting consciousness leaves the cardiovascular response to work for hypoxia that would go unnoticed till the arterial saturation could activate the reflex mechanisms which is unpredictable in the early phase of hypoxia. This could be the probable reason for this patient presenting no signs with an arterial saturation of 85%.

1 Ronald D. Miller, Anesthesia, 5th edition, Churchill Livingstone 611-12. 2 Heistad DD, Abboud FM: Circulatory adjustments to hypoxia: Dickinson W.Richards Lecture. Circulation 61:463,1980 3 Duke T, Mgone J, Frank D. Hypoxaemia in children with severe pneumonia in Papua New Guinea. Int J Tuberc Lung Dis 2001; 5:511-19.

Effect of Left Ventricular Impairment on Mortality in Intensive Care Patients

Fitzpatrick KR, Korsah P. Intensive Care Unit, Crosshouse Hospital, Kilmarnock Rd, Kilmarnock, Ayrshire, KA2 0BE.

The aim of the study was to determine the effect of left ventricular (LV) impairment on intensive care and hospital mortality of patients admitted to Intensive Care (ICU) in 2004 that had echocardiography performed during their ICU stay.

The study was a retrospective review of patients admitted to Crosshouse ICU between 1st Jan '04 and 31st December '04. Data was obtained from ICU Ward Watcher, patient case notes, and echocardiography records held by the Cardiology department. The outcomes measured were ICU mortality and overall hospital mortality (which included patients that died in ICU and those discharged from ICU that died prior to hospital discharge).

267 patients were admitted to Crosshouse ICU in 2004 and 38 of these had echocardiograph performed in ICU. In 37 of 38 patients that had echocardiography performed satisfactory heart views were obtained to allow the technician to comment on LV function. This was satisfactory in 19 patients (51.4%) and impaired in 18 patients (49.6%). LV impairment was described as mild (3/18), mild/moderate (2/18), moderate (3/18), moderate/severe (1/18), severe (8/18), and compromised (1/18). The results of the study are shown in the table below.

The results demonstrate that patients with LV impairment on echocardiogram had both higher ICU and hospital mortality as compared to patients with no LV impairment. The number of patients studied is too small to analyse outcome correlating to the degree of LV impairment found on echocardiogram. A secondary outcome was to determine the effect of right ventricular (RV) impairment, and no difference was found in ICU or hospital mortality between patients that had RV impairment and those that did not.

In conclusion, patients that were admitted to ICU and had LV impairment demonstrated on echocardiogram had increased ICU and overall hospital mortality than patients with satisfactory LV function.

C Duncan, IS Grant Western General Hospital, Crewe Road South, Edinburgh, EH4 2XU

Patients who are admitted to the Intensive Care Unit (ICU) with pre-existing or newly acquired neuromuscular disease may fail totally to wean from ventilation (1). With death the only alternative, increasing numbers of such patients go on to long-term home ventilation programmes. We examined the notes of 34 patients referred to the Western General Hospital (WGH) Home Ventilation Service for long-term ventilation, following acute presentations in ICU between October 1992 and August 2005. Case notes of 2 further patients who fulfilled these criteria were not available to examine.

It is interesting to note that 28% of patients presented with acute respiratory failure needing ventilation as an initial presentation of their neuromuscular disease.

Total length of stay in ICU ranged from 1 - 416 days (mean 84) with approximately 50 % of that time spent after referral to the service. Additional time was spent in high dependency care in some cases.

Conlusion We have demonstrated that it is possible to stabilise most weaning failure patients on long-term ventilatory support and maintain them at home or in a nursing home, with relatively few re-admissions (19/29 - one or less) into NHS beds; lack of readmission being an index of success of home care.

(1) Ref: Bradley MD et al (2002) Outcome of ventilatory support for acute respiratory failure in motor neurone disease. J. Neurol Neurosurgery and Psychiatry 72:752-6

There are well established guidelines for the transport of critically ill patients.1,2 There is now good evidence to suggest that poorly organised transfers by inexperienced teams are associated with increased morbidity and mortality.3,4

We prospectively collected data on all patients requiring an anaesthetic transfer to another hospital over a seven month period, ending at the date when the intensive care unit expanded to provide care for all Forth Valley patients. Data was collected using a standardised transfer form, and information was checked against the ward admission book and the Ward Watcher database. Information collected included reason for transfer, accompanying staff, observations prior to and during transfer, and any adverse events. A total of 12 transfers were carried out over the time period, 2 of which were performed by the West of Scotland Shock team. Senior House Officers (SHOs) carried out 8 of the transfers, with the remaining 2 by Specialists Registrars (SpRs). All transfers teams included an Operating Department Practitioner (ODP).

The number of patients transferred is small, and the majority were transferred by the hospital's own team, apparently safely. Increases in bed capacity and renal replacement capabilities brought about by centralisation of services in Forth Valley may well reduce these figures. However this is unlikely to impact upon the number of patients being transferred for an urgent neurosurgical intervention, who make up the largest patient group. This type of information provides important quality control for our intensive care unit and may be useful when considering regional solutions to challenges in the transfer of the critically ill patient. References 1. Neuroanaesthesia society of Great Britain and Ireland and Association of anaesthetists on Great Britain and Ireland - Recommendations for the transfer of patients with acute head injury to Neurosurgical units, 1996 2. Intensive Care Society - Guidelines for the transport of critically ill adults, 2001 3. Bion JF, Wilson IH, TaylorPA. Transporting critically ill patients by ambulance:audit by sickness scoring. BMJ 1988; 296:170-1 4. Oakley RA. The need for standards for interhospital transfer. Anaesthesia 1994; 49: 565-6

Continuous Haemofiltration in Septic and Non-septic Acute Renal Failure: How continuous is it and are we achieving fluid balance?

Shakeel M, Rajendran P, Jehan S, Adey G Intensive Therapy Unit, Aberdeen Royal Infirmary, Foresterhill, Aberdeen AB25 2ZN

Introduction: Continuous Venovenous Haemofiltration (CVVH) is used for renal replacement and fluid management in critically ill patients and has become an integral part of intensive care medicine in the United Kingdom (1). We do not know how continuous CVVH actually is in our unit. There is little information available on the length of time that the patients are off-filter during CVVH (2) and the effect of this treatment free time on the fluid balance achieved.

Aims: To quantify the length of time for which the patient did not receive haemofiltration "down time" and the net fluid balance achieved compared with the target fluid balance during CVVH. The long-term aim is to ascertain if there are any correlations between "down time" and other factors such as diagnosis, anticoagulant used or type of dialysis catheter.

Design and Setting: Prospective audit of CVVH data in Intensive Therapy Unit (ITU) of tertiary care hospital. Data was collected for 86/108 (80%) patients who underwent CVVH in ITU between august 2004 to august 2005.

Patients and Method: The CVVH prescription sheet, intake-output charts, nursing and medical notes were reviewed. The following information was collected: time on and off the filter, target and achieved fluid balance, indication for and total duration of haemofiltration, number of filters used, type of cannula and anticoagulant, cause of filter failure. Data was analysed to give: average "down time" per day for septic and non-septic patients and average difference between actual fluid balance and target fluid balance for septic and non-septic patients. The data was analysed using Epi info 3.3.

Results: Out of 86 patients, 60 had sepsis and 26 had non-sepsis acute renal failure (ARF) as an indication for CVVH.Target balance was neutral in 63 (n=44 Sepsis ARF, n=19 Non-sepsis ARF), or was negative, ranging -720 to - 4800 ml per day in 23 (n=16 Sepsis, n=7 Non-sepsis ARF). The time lost per day was 1.45 hours (0-8.9) in non-sepsis and 1.88 hours (0-9) in sepsis ARF. In 72.5% of the cases, the filter failure was due to clotting. The achieved fluid balance was +1768ml per day in non-sepsis ARF and +1872ml per day in sepsis ARF. Average number of filters used in non-sepsis ARF were 2.8 (1-7) and 4.3 (1-30) in sepsis ARF. The average filter life was 20.6 hours (2-43) in non-sepsis and 16.6 hours (1-51) in sepsis ARF.The average duration of CVVH was 2.7 days (0.08 - 9.12) in non-sepsis and 3.2 days (0.04 -17.4) in sepsis ARF. The APACHE II score was 28.4 (12-55) for sepsis and 27.9 (18-43) for non-sepsis ARF. Only 21% (18/86) patients left the hospital alive.

1. 'Continuous therapy' is not truly continuous (as shown by Uchino S et al) (2) with significant down-time ranging from 0 to 9 hours/day. 2. "Down-time" is a significant factor in the failure to achieve target fluid balance. 3. This should be taken into account prescribing CVVH. 4. In patients with renal failure due to sepsis: average down time, deviation from target fluid balance, total duration of haemofiltration and average number of filters used were more than in non-septic patients. 5. In the majority of the cases filter failure was due to clotting. Its earlier occurrence in patients with sepsis probably reflects their hypercoagulable state and the hypofibrinolysis in these patients (3)

1. Wright S.E., Bodenham A., Short A.I., et al., The provision and practice of renal replacement therapy on adult intensive care units in the United Kingdom. Anaesthesia. 2003 Nov; 58(11):1063-9. 2. Uchino S., Fealy N., Baldwin I., et al., Continuous is not continuous: the incidence and impact of circuit "down-time" on uraemic control during continuous veno-venous haemofiltration. Intensive Care Med. 2003 Apr;29(4):575-8. 3. Moir E., Greaves M., Adey G.D., et al., Polymorphonuclear leukocytes from patients with severe sepsis have lost the ability to degrade fibrin via µPA. J. of Leuk. Biol. 2004, 76; 571-576.

Critically ill patients transferred by air by a dedicated transport team in the West of Scotland

Over 500 critically ill patients are transferred by a dedicated critical care transport team each year in the West of Scotland. Patients are transferred using a dedicated road ambulance adapted for the purpose of being a mobile intensive care unit. On occasion the distance involved means that transfer by road would impractical and undesirable. In these instances the team has access to the aircraft owned or chartered for use as air ambulances by the Scottish Ambulance Service. Air transfer is performed relatively infrequently by the team and, in the light of a recent fatal crash, may be perceived as presenting a greater risk to both patient and escorts 1. Patients transferred by air might be expected to be sicker due to the delays inherent in the logistics of transfer, and to have poorer outcomes due to the duration and specific physical hazards posed by air transport 2, 3. We therefore performed a retrospective audit of air transfers performed over the 6 year period January 1999 to December 2004.

We retrieved data on location, duration of transfer, and immediate transfer outcome from the electronic database maintained by the transport team. For patients transferred into intensive care units in the North Glasgow NHS Trust we obtained APACHE II and outcome data from the Scottish Intensive Care Society Audit Group database via the local servers. The primary outcome measure was mortality at hospital discharge.

During the audit period there were 3020 transfers of which 95 were conducted by air (3.1% 95% CI 2.5-3.8%). The median duration of transfer by air was approximately 6 hours (IQR 5-7 hrs) while that of road transfer was 2 hours (IQR 1.5-3 hrs). 8 patients (8% 95% CI 4-16%) were declined transfer on arrival as compared to 67 of the non-air transfers (2.2% 95% CI 1.8-2.9%). All patients who were transported survived transfer. 42 patients were transferred into North Glasgow NHS Trust and had severity of illness scoring and outcome data available. Median apache score was 18 (IQR 13-27) and hospital mortality was 38% (95% CI 24-54%).

Although we concede that there are serious limitations in our analytical method, our audit suggests that outcomes in patients transferred by air are not markedly worse than other patients admitted to intensive care. Patients who are not transferred after assessment at scene may be expected to have a high mortality and this occurs in a larger proportion of patients transferred by air. This may reflect the effects of mobilization time, the fact that these patients are largely being transferred from locations with limited critical care facilities, and the perceived additional risks to the critically ill patient posed by air transfer.

1. Durairaj L, Will JG, Torner JC, Doebbeling BN. Prognostic factors for mortality following interhospital transfers to the medical intensive care unit of a tertiary referral centre. Crit Care Med 2003; 31 (7):1981-86 2. Martin T. Handbook of Patient Transportation. Greenwich Medical Media: London 2001. 3. http://news.bbc.co.uk/1/hi/scotland/4349523.stm

MI Beatty, KL Everingham T Barber, DG Swann, Critical Care, Royal Infirmary, 51 Little France Crescent, Edinburgh, EH16 4SA

CRBSI are a significant cause of morbidity and mortality in intensive care patients, with an increasing incidence of 6000 cases per year in the UK or 8.8/1000 catheter days . CRBSI may be responsible for 17% of primary bacteraemias in ICU patients and 30-40% of nosocomial infections. As with all nosocomial infections, CRBSI may prove very difficult to treat as many are caused by multi resistant organisms .

Our critical care ward has 18 beds. Infection surveillance was carried out over 12 months from December 1st 2004 to November 30th 2005 as part of The Hospitals in Europe Links for Infection Control through Surveillance programme. The criterion for the diagnosis of a CRBSI was the identification of paired positive central venous line tip and peripheral blood cultures.

The total number of admissions during the surveillance period was 1027 with 559 patients totalling a length of stay greater than 2 days. During this time 486 patients had a central venous catheter inserted and there were 24 cases of CRBSI. This equated to a device-associated rate of 6.4/1000 catheter days. The mean age and APACHE II score of infected patients were 57.8 years and 22.4 respectively. The mean number of days the catheter had been in situ for before a CRBSI was identified was 11.The most commonly identified pathogens included Staphylococcus aureus and Klebsiella species.

This data illustrates the fact that CRBSI are a significant problem. Surveillance, education the subsequent implementation of more stringent infection control measures may help to reduce their incidence.

Trainee physicians and surgeons experiences of seeking advice, making referrals and dealing with critically ill patients - questionnaire survey

The recent NCEPOD report 'An Acute Problem'1 has highlighted that the care of patients prior to admission to intensive care (ICU) is at times sub-optimal 2. In April 2005, a questionnaire survey was sent to all medical and surgical trainees in two district general hospitals in the West of Scotland. The aim was to assess the support within their own departments, their experiences of referring to ICU and their previous training in critical care issues.

26 out of 50 questionnaires were returned giving a response rate of 52%. 19% (5) of trainees always felt well supported within their own specialty while 69% (18) usually did. 8% (2) only felt well supported on occasions. Most were always or usually happy to contact their senior for advice but were less happy out of normal working hours. Many trainees had hesitated before phoning for advice, despite knowing they were out of their depth and the patient was ill.

89% (23) of trainees would always or usually discuss a case with their senior prior to referring to ICU. This however, does not seem to happen in reality 1,3. It is already known that referrals to ICU are often late. Of the trainees surveyed, 92% said ICU had been helpful when phoned and very few said the service was unapproachable. However, they did feel that at times, informing ICU of an 'at risk' patient was seen as low priority. Of interest, 85% (22) of trainees also stated that they would not ask for help sooner if asking a critical care outreach team nurse. Only 50% (13) of trainees had received any training in critical care while at university, with 69% (18) having been trained since graduation. 50% trainees felt their teaching had been adequate but 88% (23) felt they would benefit from further teaching in specific areas.

Changes to physicians' work patterns may, in the future, result in greater supervision and support of trainees 4. However, there still seems to be a gap in their knowledge base and practical experience, despite the availability of multiple critical care courses. Perhaps trainees would benefit from teaching from intensivists about the kinds of patients who would benefit from ICU and an overview of what ICU expects of the referring trainee. It may be possible to incorporate this into the induction day programme. Hopefully the advent of the foundation years of training will also be beneficial, with more trainees having exposure to practical critical care teaching.

1. NCEPOD. An Acute Problem. May 2005 www.ncepod.org.uk/2005.htm 2. McQuillan P. Confidential enquiry into quality of care before admission to intensive care. BMJ 1998; 316: 1853-1858. 3. Allan A, McQuillan P. J, Taylor B. L, et al. Who sees the critically ill patient before ICU admission? Clin Int Care 1994; 5: 152. 4. Royal College of Physicians. Interface between acute general medicine and critical care. A report of a working party of the Royal College of Physicians 2002. www.rcplondon.ac.uk/pubs/online_home.html

P Thornton 1 , P Seigne 1, B Sweeney 2, Farrell M,3 Keohane C 2 Divisions of Anaesthesia 1 and Neuroscience 2, Cork University and Beaumont Hospitals3.

We present a very rare case of a 56 year old man with known coeliac disease who developed rapidly progressive cognitive impairment and generalised seizures which became intractable. His ICU management was complicated by highly resistant seizures, and EEG evidence of seizure activity persisted even while being treated with Thiopentone infusion. Initial CT and MRI did not show any focal lesion and CSF analysis was normal. Multiple investigations including dementia, nutritional, toxic, vasculitic, infective (including for prion disease), and metabolic screens proved non-diagnostic. Brain biopsy showed normal grey matter but there was a non-infective, non-inflammatory leucoencephalopathy. His condition progressed over 10 weeks, he developed pneumonia and died. At post mortem changes were limited to brain, largely in white matter which showed an extensive diffuse vaculoar myelinopathy. There was no evidence of an inflammatory or infective cause. Neuropathological complications of coeliac disease are well documented however the type of white matter damage in this case is not described as a known complication. There is one case report of a patient with coeliac disease with progressive Leukoencephalopathy similar to this case however that patient did not develop intractable seizures. Vacuolar Leukoencephalopathy has been reported following toxic agents such as cuprizone, tri-ethylintin hexachlorpene and heroin abuse, however none of these were a factor in this case. It is postulated that this condition may have been caused by a disorder of amino acid metabolism leading to toxic metabolites in the CSF.

Frequency and appropriateness of arterial blood gas measurement in a tertiary referral intensive care unit.

S McGuirk, GG Lavery, P. Cadell, T Coogan, B McLaughlin. Regional Intensive Care Unit, Royal Hospitals Trust, Grosvenor Rd, Belfast BT12 6BA.

Arterial blood gas (ABG) measurement is an aid to clinical decision-making and is the commonest "investigation" performed in the intensive care unit (ICU). We audited all ABGs drawn from 25 patients (consecutive admissions to ICU) during their first 10 days in ICU or until ICU discharge (if length of stay (LOS)<10 days). Our questions were (i) How often are ABGs done? (ii) What proportion of ABGs are appropriate? (iii) Does the frequency of ABG measurement differ by time of day or time since ICU admission?

The ICU had 17 level 3 beds (in 2 zones) and 2 ABG analyzers (Omni 9 Analyzer, Roche). All patients had continuous pulse oximetry and indwelling arterial lines. All ABGs were drawn by ICU nursing staff. The ICU has an automated clinical management system (ACMS) [CAREVUE 9000, Phillips Medical], which documented the timing and values of ABGs, the reason for measuring ABGs (a specific field) and the details of the patients' concurrent physiology and organ support. The appropriateness of each ABG measurement was judged by applying relevant data from ACMS to a multi-disciplinary practice guideline developed in another tertiary referral ICU and published in 1997 1.

Patient characteristics [median and (range)] were as follows; Male: Female 13:12; Age 45 yrs (14-87); Admission APACHE II score 17 (7-34); LOS in ICU 12 days (2-57). Mortality was 16% (n=4). Over the 10 day period, the 25 patients had a total of 2047 ABG measurements done - 1054 (51%) in period 0800-2000 hrs (day-shift) and 993 (49%) in the period 2000-0800 (night shift). The median number of ABG measurements performed per patient per 24 hr period was 9 and a reason was documented on ACMS in 48% (n=983). When audited against the practice guideline 1, 44% of ABG measurements (n=891) were judged to have been appropriate. The daily median number of ABGs done and the percentage judged as appropriate are shown in Table I.

Day in ICU	1	2	3	4	5	6	7	8	9	10
No of patients in audit	25	25	25	24	24	23	23	22	20	16
BGAs/24 hrs (median)	8	10	9	10	8	9	9	8.5	9	8.5
ABG appropriate %	52	48	43	44	44	41	34	39	45	38

Our investigation shows a high overall frequency of ABG measurement and no significant difference between day and night shift. LOS in ICU did not influence the number of ABGs measured per 24 hr period but the proportion of ABGs judged to be appropriate decreased over the first 7 days. We hope to use these data and a modified guideline to influence our practice prior to further audits in 4 and 8 months time. References Pilon CS, Leathley M, London R, et al (1997) Practice guideline for arterial blood gas measurement in the intensive care unit decreases numbers and increases appropriateness of tests. Crit.Care Med. 25,1308-13.

Alcohol abuse is regular excessive consumption of alcohol (>28U/wk in men, >21U/wk in women), and is particularly prevalent in the West of Scotland. It is associated with concomitant physical, psychological and social problems. Common alcohol related admissions to intensive care (ITU) include acute withdrawal syndrome, encephalopathy, respiratory complications, upper GI bleeding, pancreatitis and acute liver failure.

We aimed to determine whether alcohol consumption is recorded in case notes during the pre-ITU referral out-patient visit or ward admission, to identify any relationship between alcohol excess and reason for admission and to assess ITU mortality rates between excessive (XS) and non-excessive (non-XS) drinkers in a prospective case note review of all patients admitted to general intensive care at the Western Infirmary Glasgow, from May-July 2005. The following factors were recorded:-

· Any relationship between admission diagnosis and alcohol excess (direct, contributory or unrelated).

There were 119 admissions to intensive care from May-July 2005. Data was available for 111 subjects. Excessive alcohol consumption was identified in 27 (23.9%) patients. Prior to ITU admission, alcohol intake was recorded in 44.1% in A&E, 34.3% in referring ward and in 34.5% in ITU. 87 patients were seen as out-patients prior to admission, and intake was recorded in 62.1%. In the XS group, 24 ( %) had an admission diagnosis directly related to alcohol abuse, compared with 6 ( %) in the non-XS group (p <0.001). There was no difference in duration of ITU stay between the XS (median 3 days, IQ range 2-6 days) and the non-XS groups (4 days, 2-9 days, p=0.20). However, alcohol abusers required a significantly longer period of ventilation (3 days, 2-6 days) compared with the non-XS group (2 days, 0-3 days, p=0.02). There was no significant difference in ITU mortality rates between the groups (XS 33.3 % vs non-XS 17.7 % (p=0.089).

Alcohol excess is a major contributory factor to admission to intensive care, and is related to a longer period of ventilation in ITU. Alcohol abuse is associated with a trend towards a higher mortality rate. Improved documentation of alcohol consumption may allow early identification of higher risk patients. Further data collection from ITU patients with alcohol related problems will be helpful to assess the cost, both in health and financial terms, of excessive drinking, and to identify patients who might benefit from interventions aimed at reducing the complications of excess alcohol.

S. Black, M. Kumar Department of Anaesthesia, Aberdeen Royal Infirmary, Aberdeen AB24 2ZN

Background: Awareness is the post-operative recollection of events occurring during general anaesthesia, affecting 0.1-0.2% of all surgical patients 1. Bispectral index is a monitor of anaesthetic depth and has been shown to decrease incidence of awareness 1, 2, reduce consumption of anaesthetic agent and reduce immediate recovery 1. We conducted a survey among anaesthetists in Aberdeen Royal Infirmary (ARI), with the following aims: Establish proportion of anaesthetists in ARI who have had a patient with awareness; assess degree of concern surrounding possible awareness felt by patients and anaesthetists; ascertain whether a BIS monitor would contribute helpful information in ARI theatre suite.

Methods: A questionnaire was posted to all 91 members of the anaesthetic department in ARI. 56 completed the questionnaire (61.5% response rate), including 11 SHOs, 18 SpRs, 2 staff grades, and 25 consultants.

Findings: Twenty-seven percent of anaesthetists have had a patient with awareness. Routine visitation of patients post-operatively is mainly limited to the same day, and first post-op day. 70% anaesthetists seldom ask patients directly about possible awareness. 57% (N=32) of anaesthetists are occasionally asked by patients about awareness. 34% (N=19) of anaesthetists are only occasionally concerned about possible awareness, closely followed by 32% (N=18) who seldom are concerned. 54% (N=30) of anaesthetists believe the BIS monitor would change their practice. Only 32% believe it would be a good change. 50% (N=28) of anaesthetists believe that patients have a right to expect more advanced monitoring devices that can better safeguard them against awareness.

Conclusions: 27% of anaesthetists in ARI have had patients experience awareness, however this result may be inaccurate due to post-operative visiting practices, and patient under-reporting. If the incidence of awareness is to be established accurately, visitation should be extended beyond the first post-operative day, and direct questioning should be included to encourage patient reporting. Patients express concern about awareness only occasionally. Likewise, anaesthetists are only occasionally or seldom concerned about awareness. Further work needs to be done to assess patient's concerns regarding awareness, possibly by conducting a patient survey. Only 54% of anaesthetists believe BIS would change their practice. A locally conducted trial in ARI may provide further evidence that BIS monitoring can change practice. References 1. Myles PS, Leslie K, McNeil J, Forbes A, Chan MT. Bispectral index monitoring to prevent awareness during anaesthesia: The B-aware randomised controlled trial.[see comment]. Lancet. 2004; 363(9423):1757-1763. 2. Ekman A, Lindholm ML, Lennmarken C, Sandin R. Reduction in the incidence of awareness using BIS monitoring. Acta Anaesthesiol Scand. 2004; 48(1):20-26.

Audit of guidelines introduced for hypothermic management of patients with out of hospital cardiac arrest.

In 2003 the International Liaison Committee of Resuscitation recommended hypothermia as a supplementary management for comatose patients with an out of hospital cardiac arrest. This followed two publications in the literature that showed improved neurological outcome in this subgroup of patients if cooled to mild hypothermia1, 2.

Following an evidence based assessment of the literature it was felt appropriate to adopt the recommendations and provide hypothermic management for this subgroup of patients in our Intensive Care Unit (ICU). However we had no experience of cooling patients as part of our present practice. A further literature search to identify the practicalities and potential adverse effects was conducted 3, and guidelines were introduced after a multidisciplinary meeting.

This audit includes all 15 patients treated in the first year after guidelines were introduced. It aims to find out the following questions. How long did it take to achieve the desired temperature and what were the delays to cooling? Once the patients were cooled was it possible to keep them at the target temperature? Were our guidelines and techniques safe and were there any side effects?

Our results show that only five of the fifteen patients fulfilled our admission criteria. This was mainly due to over application of the evidence. There were delays in reaching the desired temperature consisting of both the time to transfer from Accident and Emergency (A&E) to ICU where hypothermia was initiated, and by slow induction of hypothermia in ICU due to conservative rates of cold saline infusions. It was difficult to keep the temperature at the target range of 32.5 to 33.5oC. Only 34% of temperatures measured were in this range. In addition 60% of patients had a significant temperature overshoot on re-warming that was difficult to control. However the guidelines appeared safe and did not demonstrate side effects that would prevent attempting a more rapid induction of hypothermia.

Conclusions and recommended action includes some minimal changes to broaden admission criteria, and the highlighting of the inappropriate over application of the guidelines to patient groups lacking evidence of the efficacy of hypothermia. Due to both the difficulty in rapidly achieving hypothermia and the lack of evidence of dangerous side effects the induction of hypothermia may be better commenced in A&E and continued in ITU with a faster induction by more rapid fluid administration. Due to the difficulty in maintaining steady hypothermia and the high occurrence of potentially harmful pyrexia on re-warming, consideration should be made of using an intravascular cooling device 4,5. With future adjustment of the protocol, this may also permit more rapid induction of hypothermia.

References 1. Bernard SA, Gray TW., Buist MD, et al. New Engl J Med 2002; 346: 557-563 2. The hypothermia after cardiac arrest study group. New Engl J Med 2002; 346: 549-556 3. Bernard SA, Buist M., Crit Care Med 2003; 31: 2041-2051 4. Diringer M, Crit Care Med 2004; 32: 559-564 5. Shmutzhard E, Engelhart K, Beer R et al. Crit Care Med; 30: 2481- 2488

The incidence of SIRS in patients presenting to ARI over a 6 week period. (poster only)

McLardy, I. Houston G. Aberdeen Royal Infrimary, Forresterhill Road, Aberdeen AB25 2ZN

Systemic Inflammatory Response Syndrome (SIRS) is a well recognised cause of morbidity and mortality.1 In order to better understand our patient population, and the extent of SIRS presenting to our hospital we undertook an observational study to look at the incidence and characteristics of SIRS in patients presenting as an emergency to ARI surgical wards and the A&E resuscitation room over a 6 week period. This information will be useful in the future to allow implementation of the Surviving Sepsis Campaign Guidelines for the management of Severe Sepsis and Septic Shock. All patients notes were reviewed within 24 hours of admission, and SIRS scores calculated and recorded.

Further information collected included evidence or suspicion of infection and evidence of organ dysfunction at presentation. A modified Patient at Risk Score (PARS)2 was also calculated.

Complete data was available for 380 patients, 233 from A&E and 147 from general surgical receiving. The mean SIRS score amongst all groups was 1.15, with 31.3% of people by definition having SIRS syndrome. The SIRS distribution of the 380 patients is shown below: Evidence of infection was present in 27.7% of patients with a SIRS score of 2, 40.7% in those with a SIRS score of 3 and 55.6% in those with a SIRS score of 4. Subgroup analysis was performed to look at the incidence and severity of SIRS by age and presenting speciality (A&E vs Surgical receiving). The average SIRS score for the 0-20 age group was 0.81; for the 21-40 range,1.17; for the 41-60 age group, 1.08; for the 61-80 age group, 1.18 and for the 80+ group it was 1.31. Subsequently the SIRS score at admission was compared to the PARS score and the scores were correlated. References and acknowledgements: 2. Dellinger RP, Carlet JM, Masur H, et al: Surviving Sepsis Campaign guidelines for management of severe sepsis and septic shock. Crit Care Med 2004; 32 858-873 3. Rees JE, Mann C: Use fo the patient at risk scores in the emergency department: a preliminary study. Emerg Med J 2004;21:698-699 This work was supported by a grant from Eli Lily.

Pilot study to assess the incidence and severity of postoperative hyperglycaemia in non-diabetic patients undergoing vascular surgery

JJ McNamee1, P Hodgkinson1, D Harkin2, P Glover1, BA Mullan1 Regional Intensive Care Unit1 and Vascular Surgical Unit2, Royal Victoria Hospital, Grosvenor Road, Belfast, BT12 6BA

The neuroendocrine response to surgery may result in stress-induced hyperglycaemia1. Intensive insulin therapy to maintain blood glucose between 4.4 and 6.1 mmol.l-1 can significantly improve clinical outcome during critical illness2. It remains to be determined if there is a threshold for the development of glycaemia-related complications in non-diabetic patients. Clinical information on the incidence and severity of stress-induced hyperglycaemia in non-diabetic patients undergoing surgery is currently sparse. We performed a pilot study in non-diabetic patients scheduled for elective vascular surgery (abdominal aortic aneurysm repair, carotid endarterectomy, aortofemoral or femoral-distal vessel bypass, axillary-femoral bypass, and amputations).

The study was prospective and observational. Full ethical approval was obtained and all patients gave written informed consent. The Medisense Precision QIDTM meter and Precision PlusTM electrode strips (Abbot Laboratories, Bedford, MA 01730 USA) were used to measure blood glucose. Blood glucose was measured prior to the induction of anaesthesia, immediately postoperatively and then 6, 12 and 24 hours postoperatively.

29 patients were studied (male/female 14/15; age 63+13 years). Postoperatively, blood glucose increased significantly (pre-op 5.6+1.0 mmol.l-1, immediate post-op 6.6+1.2 mmol.l-1, 6 hours post-op 8.1+1.3 mmol.l-1, 12 hours post-op 7.3+1.8 mmol.l-1, 24 hours post-op 8.3+0.7 mmol.l-1; p<0.001 repeated measures ANOVA). In 4 patients the blood glucose exceeded 10 mmol.l-1 during the 24 hour period.

Significant hyperglycaemia can occur in non-diabetic patients postoperatively. Blood glucose should probably be measured routinely in all patients after surgery. Further investigation to determine the role of intensive insulin therapy for such patients is warranted.

References: 1) Schricker T, Lattermann R, Schreiber M, et al. The hyperglycaemic response to surgery: pathophysiology, clinical implications and modification by anaesthetic technique. Clin Intensive Care 1998; 9: 118-28. 2) Van den Berghe G, Wouthers P, Weekers F, et al. Intensive insulin therapy in critically ill patients. N Engl J Med 2001; 345: 1359-67.

Pilot study to assess the incidence and severity of postoperative hyperglycaemia in non-diabetic patients undergoing vascular surgery

JJ McNamee1, P Hodgkinson1, D Harkin2, P Glover1, BA Mullan1 Regional Intensive Care Unit1 and Vascular Surgical Unit2, Royal Victoria Hospital, Grosvenor Road, Belfast, BT12 6BA

References: 3) Schricker T, Lattermann R, Schreiber M, et al. The hyperglycaemic response to surgery: pathophysiology, clinical implications and modification by anaesthetic technique. Clin Intensive Care 1998; 9: 118-28. 4) Van den Berghe G, Wouthers P, Weekers F, et al. Intensive insulin therapy in critically ill patients. N Engl J Med 2001; 345: 1359-67.

Clinical decision making for investigation and management of Intensive Care Unit (ICU) acquired pneumonia (IAP): a proposed classification system

DA Kerslake1 K Everingham1, 2, IF Laurenson2, TS Walsh1, 2 Intensive Care Unit1 and Dept Microbiology2, Royal Infirmary of Edinburgh, Little France, 51 Little France Crescent, Old Dalkeith Road, Edinburgh, EH16 4SA.

Background: ICU acquired pneumonia has significant impact on morbidity and mortality in hospitals worldwide. Controversy exists regarding optimum diagnostic and treatment strategies. Objective: To use data from a period of detailed ICU surveillance to develop a system of describing investigation and management of IAP, using independent surveillance using the HELICS system as a reference. Method: We developed a database to record clinical decision-making in relation to IAP as an addition to the HELICS database. The aim was to define fields not recorded in HELICS that relate to quality in IAP management. We present data on all IAPs during a 3-months period to illustrate the potential value of the classification as a method of tracking practice over time. Results: There were 24 IAPs during the period. Data were missing in 1 case. The classification, based on the queries we developed, is summarised below for 23 cases.

Comment: Our system is a useful method of describing clinical decision-making that could compare practice between ICUs, or within an ICU over time. We propose it as a quality tool to measure practice against evidence based guidelines.

SOFA score (CVS + Renal)	GAHS	ICU Mortality	Likelihood ratio (95% CI)
0 – 4	5 – 8	44%	0.4 (0.2 – 0.7)
0 – 4	9 – 12	80%	2 (0.7 – 6.2)
5 – 8	5 – 8	69%	1.1 (0.4 – 3.2)
5 – 8	9 – 12	100%	∞

	n	<20mmHg	20 – 30mmHg	>30mmHg	P_cuff > MAP
October	52	16 (31%)	20 (38%)	16 (31%)	2 (4%)
December	80	27 (33.75%)	34 (42.5%)	19 (23.75%)	2 (2.5%)

		Mean difference	SD			Mean difference	SD
PtcCO2-PaCO2	n=90	0.45	0.52	PaCO2-EtCO₂	n=90	1.49	1.15
t=0min	n=30	0.42	0.54	t=0min	n=30	1.55	1.13
t=30min	n=30	0.51	0.57	t=30min	n=30	1.46	1.11
t=60min	n=30	0.43	0.44	t=60min	n=30	1.46	1.22
PaCO₂<6.7kPa	n=73	0.31	0.42	PaCO₂<6.7kPa	n=73	1.11	0.66
PaCO₂>6.7kPa	n=17	1.08	0.44	PaCO₂ >6.7kPa	n=17	3.14	1.32
Severe Respiratory Failure	n=39	0.76	0.56	Severe Respiratory Failure	n=39	0.97	0.59
PEEP >12cmH2O	n=12	1.15	0.59	PEEP >12cmH2O	n=12	2.63	1.62

	Home	Nursing home	Hospital	Deceased
% patients	36%	9%	27%	21%
Mean age	38y	46y	49y	63y
Level of injury	C4/5 – T7	C45 – C56	C1/2 – C7/T1	C4/5 – T3/4

Blood and aspirate source	Duration of stay		p value
	1-7 days (n=38)	≥7 days (n=11)
ABG	24.5 +/- 1.8	17.45 +/- 2.2	0.02
Routine blood sampling	18.66 +/- 0.96	16.27 +/- 2.6	0.8
Additional blood sampling	10.8 +/-2.6	9.0 +/- 4.0	0.7
Total	53.9 +/- 3.8	42.7 +/- 7.1	0.18

Back to February Contents

Teamwork in the Scottish ICU

Recovery from the anaemia of critical illness is associated with resolution of the inflammatory state despite a depressed erythropoietin response.

SOFA score (CVS + Renal)

GAHS

ICU Mortality

Likelihood ratio (95% CI)

0 – 4

5 – 8

44%

0.4 (0.2 – 0.7)

0 – 4

9 – 12

80%

2 (0.7 – 6.2)

5 – 8

5 – 8

69%

1.1 (0.4 – 3.2)

5 – 8

9 – 12

100%

∞

Result

How many patients were diagnosed using BAL?

	Patients admitted to ICU in 2004	Patients that had echocardiogram performed	Patients with no LV impairment	Patients with LV impairment
Number of patients	267	38	19	18
Age – mean (median;25^th,75^th percentiles)	55 (57;43,70)	63 (66;56,73)	62 (65;54,73)	64 (64;56,72)
Apache II – mean (median;25^th,75^th percentiles)	20.4 (20;14,26)	23.0 (23;19,26)	21.7 (21;18,26)	23.4 (23;21,26)
ICU stay (days) – mean (median;25^th, 75^th percentiles)	5 (2;1,5)	18 (8;5,18)	27 (14;6,35)	9 (8;4,9)
ICU mortality – number (%)	55 (20.6)	13 (34.2)	4 (21.1)	8 (44.4)
Hospital mortality – number (%)	83 (31.1)	20 (52.6)	7 (36.8)	12 (66.7)