key: cord-0033768-1bwf7luk
authors: nan
title: PS 0035-0363
date: 2006-08-24
journal: Intensive Care Med
DOI: 10.1007/s00134-006-0312-5
sha: a2af1daf79afb5d9c16b18bad22e6ae8139be842
doc_id: 33768
cord_uid: 1bwf7luk

nan

CONCLUSION. "HIGH PEEP" and "LOW PEEP" ventilatory strategies provide similar attenuation of pulmonary inflammation and histological lung damage through a pro-apoptotic signal of similar magnitude. 

37 ± 19 36 ± 31 42 ± 52 0 (N = 9) 0 6 (50%) 6 (50%)

49 ± 20 33 ± 18 14 ± 14 8 9 (90%) 1 (10%)

33 ± 14 29 ± 28 29 ± 32 2 2 (29%) 3 (43%) 2 (29%)

CONCLUSION. WRC permits effective care and favorable evolution for selected patients coming from an ICU. Esteban et al. (2000) . Those with low no. of decs not reported Our results confirm those obtained in a prior study (1) . An overmortality is associated with the occurrence of an "early" septic shock". In the present study, however, severity of patients was similar at the onset of septic shock, while morbidity in the "early" group is higher than in the "late" group. We hypothesize that the patients admitted to ICU with an "early" septic shock experience a delay in their management regarding the goals recommended by the "Surviving Sepsis Campaign". REFERENCE(S). 1. Roman-Marchant O, et al. Chest 2004; 126:173-8 .

The purpose of this simulation research was to assess whether bedside nurses could better apply currently recommended therapeutic interventions for patients with sepsis by using a horizons trends clinical decision support tool, rather than just standard monitoring screen shots alone.

Simulation research participants (N=50) were first required to attend a didactic training session focusing on recognition and evidence-based treatment for critically ill patients with sepsis. Participants were then directed to apply these treatments in a simulated sepsis experience. Data were collected at 2 sites (AACN National Teaching Institute Critical Care Nursing Conference, New Orleans, May 2005 and Long Beach Memorial Medical Center's Health Skills Education Center). A METI HPS (human patient simulator) was connected to a Philips Medical Systems Intellivue MP 70 in a simulated critical care environment. Participants were given the patient history, and completed the rest of their assessment using the HPS and Intellivue patient monitoring. Data were collected to compare the use of bedside monitor displays with and without horizon screen trends in the care of patients with sepsis. Group 1 (N=26) completed the sepsis scenario using a standard screen display, and group 2 (N= 24) had the addition of horizon trends on the display.

The point that marked the onset of sepsis was when each of the physiologic parameters met the current evidence-based screening criteria being disseminated by the Surviving Sepsis Campaign. The mean time for initiation of every therapeutic decision point was shorter in Group 2 (the horizons screen group) than in Group 1 (the standard screen group). There was a 32% reduction in time to sepsis recognition, an 8% reduction in time to initiation of fluid bolus, a 22% reduction in time to initiation of vasopressor therapy, and a 6% reduction in time to get diagnostic blood cultures.

CONCLUSION. While the number of participants was too low to reach statistical significance, results of this pilot study potentially support the hypothesis that the use of horizons screen trends assisted the clinicians in making more rapid clinical decisions.

Grant acknowledgement. Partial funding was supplied by Philips Medical Systems and Medical Education Technologies.

Castro R CONCLUSION. Failure to adhere to the"sepsis bundles" was associated with an increase in mortality. METHODS. Seventy seven ICUs, representing 1251 critical care beds, with an homogeneous distribution around the country were included. During the preliminary phase of the study, we collected data from the subjective perception (SP) of ICU directors on the level of implementation of the guidelines and the general characteristics of the ICUs. During a two month period we collected prospectively data about the implementation of the SSC guidelines, before any specific educational intervention, using the "sepsis bundles" in 731 episodes of severe sepsis or septic shock. We have compared the difference between the SP of ICU directors on the current level of implementation of the guidelines and objective data.

Serum lactate was measured in 37% of episodes (SP 64%), blood cultures were collected before antibiotic administration in 54% of episodes (SP 85%), broad-spectrum antibiotic were administered within 3 hours in 66% of episodes (SP 80%), 20 ml/kg of crystalloid fluid bolus was delivered followed by vasopressors if needed to maintain MAP > 65 mmHg in 40% of episodes (SP 84%), central venous pressure > 8 mmHg was achieved in 23% of cases within 6 hours of presentation (SP 70%), central venous saturation > 70% was achieved in 6% of cases within 6 hours of presentation (SP 54%), low dose steroids was administered in accordance with standardized ICU policy within 24 hours of presentation in 43% of cases (SP 52%), Drotrecogin alfa was administered in accordance with standardized ICU policy within 24 hours of presentation in 46% of cases (SP 51%), glucose control was maintained > lower normal limit with median < 150 mg/dL 6 to 24 hours after presentation in 44% of cases (SP 63%) and median inspiratory plateau pressure was maintained < 30 cmH2O over first 24 hours after presentation in 85% of cases (77%).

The current level of implementation of the sepsis guidelines is Spain is low and worse than expected by ICU directors. Great effort should be made in order to improve the implementation of the guidelines in patients with severe sepsis and improve outcome.

Grant acknowledgement. RED GIRA (G03/063).

Ferrer R 1 , Edusepsis investigators 2 1 Critical Care Center, Sabadell Hospital. Corporacio Sanitaria Parc Tauli., 2 Critical Care Centers, Edusepsis, Sabadell, Spain CONCLUSION. In non-neutropenic patients with more than 48 h. of mechanical ventilation and severe sepsis, colonization of the upper digestive tract shows greater sensitivity and specificity in predicting multifocal colonization by Candida spp than colonization of the respiratory tract. 

De la Cruz Avila L 1 , Vazquez de Anda G 1 , Gutierrez Ruiz J 1 , Rodríguez Hernández A 1 , Zuñiga Velazquez C 1 , Arzate Villafaña J 1 1 Critical Care Unit, Medical Center ISSEMyM, Metepec, Mexico

It is known that Candida spp infection is an increasing and worrying disease in critically ill patients. It has been advocated that treatment should start as soon as possible under evidence or suspicion of Candida spp infection. The aim of this study is to present the incidence and mortality of Intravascular Candidiasis (IVC) from a general Intensive Care Unit (ICU).

December 2005 all patients admitted in the ICU were prospectively enroll in a Cohort study. Every day a check list was performed to determine de presence of Systemic Inflammatory Response (SIRS), documented or highly suspected infection, antibiotics and antifungical drugs prescribed. During admission and or at any time during the length of stay in the ICU cultures from Blood Stream (BS), tip of Central Venous Catheter (TCVC), and any other place potentially infected or colonized with Candida spp was indicated. A case of Candidiasis was defined when a culture for Candida spp was reported. A positive case for IVC was defined when a Candida ssp was isolated from BS and or a TCVC. Control was defined as a patient in whom Candida spp. was not isolated. Tables of 2X2 and Chi square or Fisher's test were performed to determine association between mortality and the presence or absence of IVC and antifungical treatment.

. 909 critically ill patients were included; Candida spp was isolated in from 59 patients (incidence of 6.4%). Intravascular Candidiasis was identified In 14 patients (incidence of 1.5%). Candida spp was isolated from BS in 7 patients (0.77%); from TCVC of 4 patients (0.44%); finally, from BS and TCVC of 3 patients (0.33%). Candida Albicans was the most common type of Candida specie isolated. The total mortality observed in the ICU during the study period was 30.3%. Associated mortality due to Candida spp was 55.9% in contrast with 25% of mortality from patients without Candida spp. The mortality rate of patients with IVC was 85.7% versus 46.66% from patients with non IVC, a relative risk for mortality of 2 (0.14-0.40). All patients with IVC, except two, received antifungical treatment. Two patients survive to the ICU. Mortality directly attributable to candidemia was observed in one patient.

CONCLUSION. In our ICU, IVC has a low incidence and lower mortality attributable to Candida spp. However, high associated mortality is observed despite of treatment. INTRODUCTION. S100B protein is a small dimeric protein which belongs to the group of calcium binding proteins. It is present in high concentrations in glial cells and Schwann cells. The level of this marker is highly specific for lessions of the central nervous system. The aim of our research was to evaluate a consequence of a dynamic monitoring of the S100B protein levels for a prognosis of pateints with a head trauma injury.

METHODS. 120 patients with a head trauma injury were prospectively monitored. All patients were admited to the Emergency Department of the University Hospital in Pilsen. The time period from injury to admission was shorter than 6 hours. All patients had a positive finding at the initial CT scan of the brain. We used scoring protocols ad admission GCS, APACHE II, ISS, TRISS, at dismissal GOS, KPS. The S100B protein level was done by LIA Essay on fully automated immunoanalyser Liaison, DiaSorin, Sweden. The S100B protein level measurement was done at admission, after 6, 12, 24 and 72 hours.

We found no correlation among S100B protein levels, gender and age. The initial S100B protein level corresponded to GCS, APACHE II and TRISS, but only partly to CT scan[ finding at admission. We found following correlations: The initial S100B protein level is very important for the prognosis especially the initial level above 1 microg/L, but it is necessary to evaluate the S100B 0h/S100B 72h ratio. Rapid decrease of the levels in the first 72 hours to normal value is associated with a good prognosis. The initial level can be relatively high, but the S100B 0h/S100B 72 h ratio must decreased minimally by 300%. Persistent high levels in the first 24 hours and decrease of the level after this time period is associated with good prognosis too, but the result is depended on the initial level more, because the S100B 0h/S100B 72h ratio is worse. Repeated increase of the S100B protein levels after previous decrease is associated with poor outcome in GOS and KPS. The high initial level and the low S100B 0h/S100B 72h ratio are typical for this group.

CONCLUSION. The S100B protein is a good prognostic marker in comparison with others scoring systems. The course of the dynamic evaluation is by our opinion useful for more accurate determination of the prognosis. 

Pino F I 1 , Navarrete P 1 , Rincón M 1 , García S 1 , Guerrero F 1 , Miñambres E 1 , Sainz-Pardo M 1 , Bermejo S 1 , Toral D 1 1 Trauma and Neuro-Critical Care, Spanish Society of Intensive, Critical Medicine and Coronary Units, Granada, Spain INTRODUCTION. ARDS is a serious complication of patient with severe trauma and it is linked to severe complications and high mortality. Our purpose was describe epidemiologic characteristics, risk factors, complications and outcomes of severe trauma ards in Spain METHODS. Multicenter, observational study achieved for 15 months. Variables collected were intra and extra-hospital. We defined severe trauma as that with revised trauma score < 11 and/or ISS > 16. Variables: demographics, severity scores, diagnosis-related groups in emergency department and according to CIE-9, therapeutics, including surgical therapeutics in the first 24 hours, ards characteristics, duration of mechanical ventilation, length of stay (LOS) in ICU, complications and mortality.

We identified 122 patients with 84.4% of males. They suffered thoracic injury (59%), severe traumatic brain injury (46.3%) and fractures of long bones (47.5%). APACHE was 17.97 ± 7.6 and ISS 26.9 ± 11.8. Onset of ards was late (first 48 H) in 68%. Incidence of pneumonia was of 58.3 and 61% of the sample showed severe sepsis. Mortality was 28.7%. Patients with multiple organ failure (MOF), acute renal failure (ARF), early ards and severe traumatic brain injury had more probability of dying with significant statistical difference: OR for severe brain trauma was of 5.98(95% confidence intervel, 1.68-21.26), OR for early ards was 6.15(95% CI, 1.69-22.33), OR for ARF was 8.49(95% CI,2.28-31.57). The catheter-related infection, bacteraemia, sepsis and pneumonia are protectors from mortality. Septic shock is not associated with more mortality. Abdominal surgery and long bone fracture are protective from mortality also. Neither prone position, permissive hypercapnia, multiple transfusion, thoracic surgery, osteosynthesis, neurosurgery, abdominal, pelvic or thoracic injury nor sex influenced in ards mortality. Exitus had worse GCS, lower revised trauma score and PaO2/FiO2, greater FiO2, less days of mechanical ventilation and ICU LOS and greater SOFA. MOF is the more frequent cause of exitus in 48% and severe hypoxaemia was cause of exitus in 18%.

Ards is a marker of seriousness in severe trauma and it is defined by anatomic and physiologic indexes and by diagnosis-related groups presents in the hospital admission. Ards is late and it is characterized by high rates of infections and metabolic complications and high rates of mortality.

GrmecŠ 1 , Cander D 1 , Klemen P 1 ,Šimat D 1 1 Emergency Medicine, Centre for Emergency Medicine Maribor, Maribor, Slovenia

The outcome of prolonged resuscitative attempts cannot be predicted. The decision about when terminate resuscitative efforts for patients with cardiopulmonary arrest is often subjective. Therefore a method of predicting the outcome of out-of-hospital cardiac arrest (OHCA) is needed. The measurement of petCO2 could be used for reliably identification of irreversible cardiac arrest.

Study was performed in 297(201 male) consecutive victims of normothermic, nontraumatic OHCA with nonshockable initial monitored rhythm (patients with post-defibrillation with similar rhythm were included). Once a patient was intubated, petCO2 and others parameters were measured every minute for first 5 minutes and than every 5 minutes for 20 minutes or until resuscitation efforts were terminated or patient was hospitalized. In this study, a hypothetical decision was made to cease resuscitative efforts based on petCO2 level of 1.33kPa or less after 20 minutes of advanced cardiac life support. The final outcome for all patients was classified as death in the field and survivor to hospital admission.

The average of intial petCO2 was significantly higher in survivors to hospital admission (2.6+/-0.9 vs. 0.9 +/-0.8 kpa; p<0.01). After 20 minutes of advanced cardiac life support average of petCO2 was significantly higher in survivors (3.6+/-1.2vs.0.8+/-0.4kPa; p<0.001). Using the petCO2 of 1.33kPa or less as a theretical threshold to predict death in OHCA successfully discriminted between the 136(45.8%) survivors to hospital admission and 161 (54.2%) prehospital deaths. In 114 of the survivors (83.8%), the first evidence ofROSC, before palpable or blood pressure occured, was elevation of petCO2 (in average for 1.6 +/-0.5kPa). Sensitivity, specificity, positive predictive value and negative predictive value were all 100%.

Measurements of petCO2 can be used to accurately predict irreversible death in patients with nonshockable initial rhythm in OHCA, and should also be considered as a useful tool to allow the discontinuation of resuscitative efforts. 

Percutaneous Tracheostomy is gaining acceptance in American hospitals. A survey of surgical residency programs was conducted to gauge the use of percutaneous tracheostomy and its influence on surgical training.

A survey was mailed to the director of every accredited general surgery residency in the United States. Recipients were asked to estimate the proportion of tracheostomies at their institutions performed by general surgeons, ENT surgeons, medical intensivists, or other specialists. They were asked to estimate the proportion done at the bedside and those done with bronchoscopy. 

Callejo-Torre F 1 , Martínez-Barrios M 1 , González -Gil J M 2 , Llata L 1 , Fernández J L 1 , García A 1 , Arroyo M 1 1 Intensive Care Unit, Hospital General Yagüe, 2 Forensic Pathology, Legal Medicine Institute, Burgos, Spain

Comparison between the clinical findings in the ICU with the cause of death established by forensic pathologists.

On January 2005, 12 workers are involved in a fire, 7 die instantly. From the 5 taken to hospital, 3(*) die in the following days. On admission, aware of combustion of polyurethane that produces hydrogen cyanide gas (CN-), treatment with hydroxicobalamin is followed, sending blood samples to the National Toxicology Center.

In spite of the critical status on admission (table), none had severe burns. CO and HbCO are normal but CN-is high in 2 of the 3 death in ICU. We think that the main cause of death was CNintoxication. The forensic pathologists confirm our assert studying the seven workers dead instantly in the fire. They all had CN-blood levels over 3mgr/L and HbCO less than 27%. Burns where similar in body surface and depth as the ones seen in the 5 patients. Neither there were severe burns in upper airways. CONCLUSION. The cause of death was both hypoxic hipoxia (O2 consumption in ignition) and histotoxic hypoxia (CN-). Cyanide inhibits mitochondrial cytochrome oxidase creating multiorganic tissular hypoxia. From 10 to 46% of death in fires are because of CN-. Hydroxicobalamin is the best terapeutical option, based on risk/benefit, if used as soon as possible. So it is necessary: first, to know that every fire victim, as CO, must be considered as a CN-intoxicated and, secondly, to have antidote treatment ready in our ICUs. 

Covaciu L 1 , Allers M 2 , Enblad P 3 , Lunderquist A 4 , Wieloch T 5 , Rubertsson S 1 1 Anaesthesiology and Intensive Care, Uppsala University, Uppsala, 2 Thoracic Surgery Lund University, Lund University, Lund, 3 Neuroscience-Neurosurgery, Uppsala University, Uppsala, 4 Radiology, 5 Laboratory for Experimental Brain Research, Lund University, Lund, Sweden

Therapeutic hypothermia after cardiopulmonary resuscitation improves neurological outcome and reduces mortality (1, 2) . The protective efficacy of brain cooling increases if hypothermia is introduced quickly after experimental brain ischemia (3). Also, selective brain hypothermia may prevent complications of systemic hypothermia especially if employed in traumatic brain injury. We therefore studied the feasibility and efficiency of a new non-invasive nasopharyngeal technique for induction of selective brain hypothermia.

Twelve anaesthetised piglets were subjected to rapid induction of selective cerebral cooling and maintained for a period of 6 hours. Brain temperature was lowered with the help of bilaterally introduced nasal balloon catheters connected to a circuit in which chilled saline by means of a heat exchanger was circulated by cardioplegia pumps. External heating was used in order to keep normal body temperature. Temperature was measured in both cerebral hemispheres, rectal, oesophageal and in right atrium. The piglets were normoventilated and haemodynamic variables were continuously measured.

Cerebral hypothermia was induced rapidly and after the first 20 minutes the cerebral temperature was lowered with a mean gradient of 2.8 • C to a mean temperature of 35.3

• C. Meanwhile the central body temperature reflected by right atrium probe dropped with 0.5 • C to a mean temperature of 37.4

• C. After six hours brain temperature reached 34.7

• C while the central temperature was 36.2 • C (p=0.0038). The animals were haemodynamically stable during the whole period.

Inducing selective cerebral hypothermia with cold saline through nasopharyngeal balloon catheters is feasible, quick and effective. Furthermore the brain temperature was maintained at target temperature for six hours while the body temperature remained within normal range. This method may warrant further consideration for induction of fast and selective therapeutic cerebral hypothermia in humans. REFERENCE(S). 1. N Engl J Med 2002 :346:549-556. 2. N Engl J Med 2002 :346:557-563. 3. Acta Neuropathol 1994 

Pichon N 1 , Amiel J 1 , François B 1 , Marsaud J 1 , Clavel M 1 , Dugard A 1 , Gastinne H 1 , Vignon P 1 1 Medical-surgical Intensive Care Unit, Dupuytren Teaching Hospital, Limoges, France INTRODUCTION. Mild induced hypothermia (MIH) improves neurological outcome of patients who sustained a cardiac arrest. Nevertheless, the optimal cooling system for performing MIH and its benefit/risk ratio remain unknown. We evaluated the efficacy and tolerance of a recently available internal cooling system (Coolgard ® ) in patients admitted after a cardiac arrest.

During a 22-month period, all patients who underwent MIH using the Coolgard ® system were studied. MIH was started as soon as possible after patient admission, target temperature was 33

• C, and temperature was subsequently increased at a rate of 0.3 • C per hour after 36 h of hypothermia. Recorded data were: SAPS 2 and neurological examination (Glasgow Coma Scale [GCS] , papillary reflex, myoclonus) on admission and Cerebral Performance of Pittburgh (CPC) category at 6 months. Core temperature was monitored and MIH was considered steady when variations of core temperature were less than 0.4 • C. Potential side effects attributable to MIH or to the cooling system were recorded.

Among 69 patients admitted after a cardiac arrest during the study period, 34 patients underwent MIH (mean age: 56±17 years; SAPS2: 58±16; duration of anoxia: 12±9 min; cardiac origin: 65%; hypoxic origin: 35%). Upon admission, GCS was 4±1, papillary reflex and myoclonus were present in 32% and 21% of the cases, respectively. Target temperature was always reached after a mean time of 187±119 min (range: 30 to 600 min) and maintained steady in all patients for 36 h. Asymptomatic hypokalemia was frequently observed and postrewarming "rebound hyperthermia" (38.9±0.9

• C) was observed in 25 patients (74%) during the first 24 h which followed MIH. Three patients exhibited a bacteremia (Staphylococcus aureus). During follow-up (11±3 months), 10 patients (29%) had a favorable neurological outcome (CPC 1 or 2). No MIH characteristics (time to initiation, time to target temperature, duration of MIH, duration of the rewarming period) apparently altered neurological outcome.

CONCLUSION. Endovascular cooling system Coolgard ® is effective in inducing and maintaining mild hypothermia (33 • C) in resuscitated patients after a cardiac arrest, without noticeable adverse effects. The potential relationship between the development of nosocomial infections and MIH remains to investigate.

Appelboam R 1 , Thomas E O 1 1 Critical Care, Derriford Hospital, Plymouth, United Kingdom INTRODUCTION. The acute and subsequent management of patients with spontaneous intracerebral haemorrhage (ICH) who require long-term oral anticoagulation (OAC) presents a therapeutic dilemma. The available evidence comes from many specialties, and is disparate. The aim of this study was to ascertain current UK practice regarding the management of this difficult group of patients.

We contacted the duty consultant of all 32 UK Neuroscience Intensive Care Units (NICU) by telephone, and asked standardised questions regarding their usual management of anticoagulated patients who suffer a spontaneous ICH. We subsequently conducted a literature search to establish an evidence base to guide future recommendations.

. 16 units were stand alone NICU (50%). 14(44%) of the consultants have a lead role in their ICU. 23(72%) use FFP for reversal of OAC and 3 use Factor VIIa -none of these units have a protocol. 12 units use Prothrombin Complex Concentrate (PCC) and 8 also use IV vitamin K. 5 units have an established protocol for OAC reversal in ICH, 4 of these use PCC and vitamin K. Over 90% would normalise INR for DVT, PE, chronic stable AF (CSAF) and paroxysmal AF (PAF) but only 56% in patients with mechanical heart valves (MHV). 66% would commence IV heparin in the first 4 days for MHV, 16% for PAF, 16% for CSAF and 9% for PE. 94% would restart OAC for MHV, 19% for PAF,16% for CSAF and 9% for PE. Following ICH 25% would recommence OAC between 2 and 7 days and 69% after 7 days.

There is considerable variation in practice amongst senior clinicians who regularly manage these patients. We found practice to be inconsistent across the units surveyed, and in many cases not evidence based. The literature would support that early aggressive reversal of the INR to within the normal range with PCC and IV vitamin K improves outcome. The risk of systemic embolisation, even with mechanical heart valves, is low at 0.016% per day and temporary cessation of all anticoagulation is safe for 8-14 days. Furthermore, the risk of recurrent haemorrhage after reintroduction of OAC is low. The literature does not support the use of IV heparin in the acute phase, or factor VIIa for reversal in patients receiving OAC. There is an urgent need for national guidelines regarding the management of these patients. 

Intracerebral hemorrhage (ICH) represents 10-30% of all strokes. Epidemiological studies have shown factors associated with its high mortality. Factors associated with morbidity are little known. The aim of our study was to find the factors (risk factors, clinical, laboratory and neuroimaging parameters) obtained on admission which may influence the mortality and functional outcome.

We made a prospective study of patients with spontaneous ICH admitted in the Intensive Care Unit (ICU) during 2003. On admission we recorded vascular risk factors (high blood pressure, diabetes mellitus), clinical parameters (arterial blood pressure, Glasgow Coma Scale score (GCS), body temperature), laboratory parameters (leucocytes, platelet counts, prothrombin time) and neuroimaging parameters (ventricular hemorrhage, subarachnoid hemorrhage, swelling, herniation brain, volume of ICH interpreted by a neuroradiologist using computed tomography (CT)). Mortality and its predictive factors were determined after 30-day and one year using univariate and multivariate statistical analysis. The functional outcome when discharged from the ICU after one year were evaluated by the Glasgow Outcome Scale (GOS), Barthel Index (BI) and Modified Rankin Scale (mRS).

. 66 patients were identified (43 males, 23 females). The 30-day mortality was 44% and at one year 57%. Most patients died within the first seven days. Age and initial GCS were the strongest independent predictive factors of 30-day mortality for all locations (p<0.001); mortality was correctly predicted (sensitivity of 76%, specificity of 73%). Age, GCS and volume of ICH on admission were the strongest independent predictive factors of mortality after one year for all locations (p<0.005); mortality was correctly predicted (sensitivity of 87%, specificity of 75%). Global percentage prognosticated was exactly 81%. At the time of the ICU discharge, only 22% had good functional condition and were independent as evaluated by the GOS and 24% by the BI and mRS. Of the patients who survived after one year, 75% were independent as evaluated by the GOS and the BI and 68% evaluated on the mRS. After one year, the global improvement was estimated at 50%; age and initial volume were the two most important functional prognosis predictors (p<0.05).

On admission at the ICU, age of the patient, in combination with the initial GCS were the two most powerful and easiest predictors of 30-day mortality and morbidity in patients with spontaneous ICH; age, GCS and volume of hemorrhage were after one year. The functional outcome was better in younger patients and smaller haematomas on CT on admission.

Pattani H 1 , Yeoman P 1 1 Intensive Care Unit, Queens Medical Centre, Nottingham, United Kingdom INTRODUCTION. The Virginia Prediction Tree (VPT) uses four variables (pupilary response, age, Glasgow Coma Scale Motor Score (GCS-MS) and the presence or absence of an intracranial mass lesion) to allocate patients into eight groups each with a specific outcome profile [1] . While this has been used to predict prognosis in individual patients, it also provides a method for comparing the outcome of two similar groups of patients in whom the above variables are known.

Data was collected prospectively on all patents admitted with severe head injuries (GCS <9, GCS-MS <6) for ten years (1993) (1994) (1995) (1996) (1997) (1998) (1999) (2000) (2001) (2002) . 759 out of 794 patients had a Glasgow Outcome Score (GOS) recorded at one year. These patients were divided into eight groups using the VPT. The Standardized Mortality Ratio (SMR) was calculated for each group using the expected and actual number of deaths in each group.

. 759 adult patients studied, 77% male, mean age 39 yrs, mean GCS total score 5.8. 

Prospective clinical audit to discover if routine near-patient estimated serum osmolality can be used interchangeably with a gold standard (laboratory measured values) to help uncover Central DI (Diabetes Insipidus) in neurointensive care patients with severe brain injury who are free from hypertonic solute infusions.

METHODS. 32 adult neurointensive care patients with severe traumatic and ischaemic brain injury were identified over period of four months. Anonymised laboratory identification was based on the need for laboratory serum electrolytes, urea and osmolality measured for suspected Central DI and in the absence of mannitol or hypertonic saline therapy over the previous 8 hours. Serum osmolality was estimated using the formula 2x Na + Blood urea + Glucose (mOsm/kg) 2 . Serum osmolality was measured using Advanced Instruments Micro Osmometer, Model 3300. We assessed the bias and limits of agreement of the two methods for determining serum osmolality using Bland and Altman's technique [3] .

We found good correlation between measured and calculated osmolalities with r = 0.907 (r 2 = 0.823). However, Bland-Altman analysis revealed a bias of -2.45 (±6.21mOsm/kg SD) and limits of agreement from -14.6 to 9.72 mOsm/kg. Excluding patients with evidence of renal dysfunction (6 patients, serum creatinine > 110µmol/l), the bias was -2.57(±3.37mOsm/kg SD) with limits of agreement from -9.2 to 4.0 mOsm/kg.

Our clinical audit suggests that estimated and measured serum osmolality can not be used interchangeably for neuroinjured patients on intensive care unless they have normal renal biochemistry in the absence of hypertonic solute therapies. Estimated serum osmolality appears to have limited validity for use in the near-patient diagnosis of Central DI in critically ill patients following severe brain injury. Recent evidence suggests that low dose corticosteroid therapy may be used in determined subsets of patients with traumatic brain injury (TBI) to avoid hypotension. We evaluated adrenal function in TBI patients undergoing barbiturate coma as treatment of their refractory intracranial hypertension.

We prospectively studied 40 patients with moderate to severe TBI. Group A (17 patients) were treated with barbiturate coma. Group B (23 patients) presented intracranial hypertension controlled with first tier measures. Adrenal function was evaluated by using the high-dose corticotropin stimulation test at 24 hours after brain injury and after barbiturate coma induction.

Both groups were comparable in baseline caractheristics. After 24 hours of TBI, adrenal function was similar in both groups. After barbiturate coma was induced, patients in the group A presented a higher incidence of adrenal insufficiency compared with group B (53% vs. 22%, p=0.03).

Patients treated with barbiturates who developed adrenal insufficiency required higher doses of norepinephrine to maintain cerebral perfussion pressure than patients treated with barbiturates without adrenal insufficiency (1.07 ± 1.04 µg/kg/min vs. 0.31 ± 0.32 µg/kg/min, p= 0.03).

Patients with TBI treated with barbiturate coma are at higher risk to develop adrenal insufficiency. This subset of patients presented higher requirements of vasoactive support to avoid hypotension. In these patients corticosteroid therapy may play a role in the treatment of barbituratesassociated hemodynamic instability. CONCLUSION. This is the first analysis of changes in rates of diagnostic errors over time in an intensive care unit. The study shows a reduction in important Goldman type I diagnostic discrepancies in intensive care patients who underwent a post mortem examination in the period 2001-2005 compared with the preceding five years. There may be several reasons for this reduction. Our unit has doubled the number of consultant intensivists from two to four. In addition, patients are now managed by multidisciplinary teams. Increased use of diagnostic radiology may have played a role. We plan to increase our use of post mortem examination to further audit diagnostic performance, and to further examine the reasons for the apparent improvement in diagnostic accuracy. [1] . The ultimate goal is to improve the quality of care in Dutch ICU's. Quality of care is defined in several domains which are all covered by our comprehensive set of indicators. The process of QI development was described before [1] . Here we report the results of a pilot study to the feasibility of registration, the reliability and validity of the QI.

Physicians and nurses were trained to collect uniformly all data for the 12 QI: availability of intensivists (hours per day), patient to nurse ratio (three times daily), strategy to prevent medication errors (10 items yes/no), measurement of patient and family satisfaction (yes/no), length of ICU stay, duration of ventilation, absolute number of interclinical transport, % of days with all ICU beds occupied, % of glucose measurements above 8 or below 2.2 mmol/l, standardised mortality (APACHE II), incidence of decubitus, number of unplanned extubations (per 100 ventilation days). Data were collected locally and transferred to a national database. Site visits, interviews and written question lists were used.

One ICU was not able to implement the data collection in daily routine. 17% of ICU's needed more than 60 min per day to collect the items, 37% 30-60 min and 46% less than 30 min per day. The pilot covered 7682 admissions and 31849 treatment days (of which 16860 ventilated). Interclinical transport was the least reliable collected item. All other indicators showed significant variability (data not shown here) to serve as a target for quality improvement programs. Multilevel and regression analyses are being developed to relate the results of different indicators and to improve insight in the validity of the set.

CONCLUSION. This set of indicators gives a quick view of the quality of care in individual ICU departments. However, computerised assistance is necessary to limit the registration workload. The set shows what the focus of future quality improvement programs should be both on local and national level. 

We started an ICU audit in September 2005 including all patients with an ICU LOS of more than 48 hrs and focussing on key indicators within two QOC components viz process and outcome. Process measures analyzed were DVT prophylaxis (DVTP), stress ulcer prophylaxis (SUP), head end elevation (HEE), EGDT in sepsis, steroid use in septic shock and glycemic control (GC). Outcome measures audited were standardized mortality ratio (SMR), unplanned extubation (UE), VAP and CRBSI rates. Statistical analysis was with chi-square test using SPSS 11.5 V.

There were 321 evaluable patients. Parameters below showed no statistically significant difference. 

Periodic audits showed a trend towards improving QOC. However we need to further intensify our efforts to acheive a significantly higher standard of care.

A questionnaire was designed based on Boynton and Heyland. All discharged patients were included in the study if they were alert, gave verbal consent and had ICU recall 3-4 days after ICU discharge during an 8 week period. Patients were surveyed by a medical student not involved with or aware of the ICU care provided.

. 42 patients were questioned with 25 excluded. Consequently 90% questioned were surgical patients. Pain was experienced by 76%, dreams/hallucinations by 55% and sleep disturbance by 72%. Analgesia was however considered sufficient by 98%. (1, 2) . Our aim was to establish a clinical -pathologic correlation of findings in necropsies of patients admitted to our ICU.

METHODS. This is a retrospective study on necroptic studies performed since 2000 to 2004, establishing relations between autopsy findings and reason for hospital and ICU admission, the quoted reasons for death (clinical documentation) and the possible changes in diagnosis or therapy strategies that could be conditions if the necroptic information was known during patients life. This possible influence was separately analyzed by two independent physicians with a similar experience.

Out of 744 deaths during the period, 61 autopsies were performed (8%) with 51 complete and available reports. Forty six of these reports (90%) showed a good correlation between clinical and necroptic quoted causes of death. Postmortem findings would represent changes in treatment in 10 cases, with possible longer life. In 27 cases, no discrepancies were found between diagnosis pre and postmortem. Both experts correlation did not show significant differences (x2) except for the importance on necroptic findings for changing the clinical diagnosis.

Necroptic studies in our series do represent a poor added value with respect to clinical orientation. In our experience there is some discrepancy when comparing the quality control value of autopsies with their teaching and clinical meaning. 

Patients undergoing PAC placement (Vigilance ® , Edwards Lifesciences (CCO) for clinical reasons were also monitored with FloTrac TM (Edwards Lifesciences) and LiDCOplus TM . PulseCO (LiDCO group) was calibrated at baseline using the Lithium dilution technique and then it was not recalibrated for the next 8 hours. Readings of FloTrac, PulseCO and CCO were taken hourly for 8 hours and compared against intermittent thermodilution the PAC performed in triplicate randomly throughout the respiratory cycle (ITD). Device measurements were compared using a Bland-Altman Plot.

. 9 patients were enrolled. 76 total pairs of data were collected for all the devices. ITD CO range was high, (3 to 10 L/min) (mean 6.6 ± 1.8 L/min). Accuracy and precision (Bias ± 2SD) when compared against ITD were -0.62 ± 2.15 L/min for FloTrac, and L/min for PulseCO and 0.3 ± 1.85 L/min for CCO. 

Patients with hemodynamic instability who needed a Pulmonary Artery Catheter (PAC) to guide therapy were also connected to the Arterial Pressure based Cardiac Output (APCO) Flo-TracTM Edwards, Lifesciences, Irvine, CA, USA. Data collected from APCO was evaluated and compared with the intermitted cardiac output measurement using the PAC. Data is analysed using the Bland-Altman method.

In this study 11 patients are analysed. Mean age is 69 years [36 -88] . Mean APACHE II score 23 [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] . Hemodynamic instability originated from severe sepsis / septic shock in 9 patients and two patients suffered from cardiogenic shock. One after cardiopulmonary resuscitation and one patient after myocardial infarction. A total of 193 cardiac output measurements have been obtained. Bland Altman plot gives a mean bias of 0.8 l/min, a standard deviation of 1.6 l / min and the limits of agreement -2.3 -3.9 l/min.

The FloTrac measures cardiac output without the need for system calibration. Flo-Trac can be used in clinical practice but more research is needed to fully understand APCO and its implications. 

Patients monitored with the LiDCO-Plus TM for therapeutic reasons were enrolled. The gold standard reference measurement for CO was the lithium dilution measurement from the LiDCO. The PulseCO was calibrated with LiDCO at baseline. Further measurements of CO were taken at 1, 2, 4, 8 and 24 hours from baseline with lithium dilution curves. PulseCO readings (not recalibrated) were recorded simultaneously at these timepoints. A percentage error of disagreement of less than 30% was taken as being clinically acceptable.

. 14 patients were enrolled. At baseline patients showed a wide range of cardiac output (mean 8.01 l/min, SD 4.27 l/min). CO measured from the two devices correlated well (r2=0.82, p <0.01) The bias and limits of agreement were clinically acceptable (less than 30%) for the first 8 hours but then widened. METHODS. This prospective study included 76 patients (38F/38M) who underwent Doppler echocardiography with measurements of TAPSE. Right and left heart function was quantified by the ratio of right to left ventricular end-diastolic area (RVEDA/LVEDA) and the left ventricular ejection fraction (LVEF). Tricuspid annular velocities were determined by Doppler tissue imaging (DTI). In 27 patients with invasive hemodynamic monitoring the examination was repeated after passive leg raising or after treatment (e.g., fluid challenge, dobutamine infusion).

At baseline, TAPSE (mean (SD) =20 (5) mm) was related to LVEF (48 (12) %, R=0.49), age (59 (17) yrs; R= -0.42), RVEDA/LVEDA (0.55 (0.16); R= -0.25) and to numerous echocardiographic variables. There was no relationship between TAPSE and gender, weight, height, arterial pressure and heart rate. The combined influences of LVEF and age explained 30% of TAPSE variability (multiple R2=0.30; p<0.01), with no additive value of the diastolic ventricular ratio or other variables. Following hemodynamic maneuvers, changes in TAPSE (from -6% to + 60%) and in LVEF (-16 to + 111%) were linearly related (R2=0.75; p<0.001). The relationship between heart function and DTI velocities was weak or not significant.

In our study population, TAPSE was more related to LVEF than to RVEDA/LVEDA (which is however a rough measure of RV function). This could be explained by the contribution of the LV to RV function via the crista supraventricularis kinetics. Further studies are needed to confirm the interest of TAPSE to monitor cardiac function in the ICU. 

Pulmonary edema (PE) is a classical cause of acute respiratory failure during weaning from mechanical ventilation. As the treatment should be different its recognition is an important issue. Only invasive method such as pulmonary artery catheterization can distinguish between PE and other causes of weaning failure demonstrating an increase in PAOP > 18 mmHg. The B-type natriuretic petide (BNP) is an established marker of wall ventricular stress. We prospectively investigated whether the plasmatic BNP level and its change during a weaning test could reliably detect a weaning-induced PE.

In 13 patients who failed at two consecutive weaning tests on a T-tube, a pulmonary artery catheter was inserted. In all patients, a subsequent weaning trial over a maximum 1-h period of SB using a T-piece was performed. Weaning-induced PE was diagnosed if intolerance to spontaneous breathing occurred while PAOP increased above 18 mmHg. Twenty-two weaning trials were analyzed and BNP was measured before and at the end of the weaning period.

Overall BNP (44 simultaneous measurements) weakly correlated with PAOP (r 2 = 0.30, p < 0.05). From the baseline to the end of the weaning period, PAOP increased from 12 ± 7 mmHg to 21 ± 10 mmHg (p<0.05). In case of weaning-induced PE (n=12), PAOP increased from 15 ± 8 mmHg to 28 ± 9 mmHg (p< 0.05). In absence of PE, PAOP increased from 10 ± 4 mmHg to 13 ± 3 mmHg (p< 0.05). Weaning-change in BNP (delta-BNP) was significantly higher in the PE group (41 ± 22% vs 12 ± 21%, p< 0.05). The best threshold value of BNP measured before the weaning trial to predict PE occurrence, was 460 pg/ml with sensitivity (Se) of 67%, specificity (Spe) of 80%, positive predictive value (PPV) of 80% and negative predictive value (NPV) of 67%. The best threshold value of BNP measured at the end of the weaning period to predict PE occurrence, was 480 pg/ml with Se of 75%, Spe of 80%, PPV of 82% and NPV of 72%. The predictive value of BNP measured after was significantly better (area under the ROC curves) than that measured before the trial.

Although BNP may provide some degree of prediction of weaning-induced PE, our preliminary results suggest that predictive values are not satisfactory enough for BNP to be used as a single marker of a weaning-induced PE. 

Castro P 1 , Delgado E 1 , Aibar J 1 , Nicolás J M 1 1 Medical Intensive Care, Hospital Clinic, Barcelona, Spain

Non-invasive assessment of continuous cardiac output (CO) may be performed by analysis of the arterial pulse contour (APCO) without in vivo calibration (Vigileo System, FloTrac TM ; Edwards Lifesciences, Irvine, CA, USA). The aim of the study was to assess the level of agreement between a continuous APCO and CO obtained with a pulmonary artery catheter in critically ill patients.

We prospectively studied all consecutive patients admitted to ICU in a two-month period in whom a pulmonary artery catheter (Edwards Lifesciences) was placed for hemodynamic monitoring, by measuring simultaneously APCO with a FlotracTM-Vigileo System. CO determinations were performed each hour since pulmonary artery catheter was placed. Thermodilution CO (TCO) measurements with pulmonary artery catheter could be either continuous or intermittent. Statistical analysis was performed using Pearson's correlation test, t test and the method described by Bland-Altman. Statistical significance was considered P < 0.05.

A total of 10 patients were included, with 303 matched CO determinations. Causes of admission were septic shock (SS) (n=5) and elective postoperative cardiac surgery (PC) (n=5). Data at admission (mean, range) were: age 63 (47-87) years (SS: 60, PC: 67), body weight 74 (52-110) kg (SS: 71.2, PC: 78.5), APACHE II 15 (7-23) (SS: 18, PC: 11) and SOFA 9 (5-13) (SS: 12, PC: 6). CO at admission was: APCO 6.0 (3.8-9.6) L/min and TCO 8.4 (4.9-16). Overall difference between APCO and TCO measurements (bias ± 2SD) was -2.1 ± 4.2 L/min. However, CO determinations disagreements in SS were greater than PC (bias ± 2SD: -2.0 ± 3.6 versus -0.15 ± 1.78 L/min, respectively, P = 0.001) for PC and -2.3 ± 3.8 L/min for SS). In fact, the bias between the two CO methods correlated to TCO (r = 0.91, p < 0.001). Finally, SS patients with TCO =< 7 L/min showed highly comparable APCO values (bias ± 2SD: -0.53 ± 2.0 L/min) whereas when TCO was >7 L/min APCO understimate TCO (bias ± 2SD: -4.3 ± 2.1 L/min).

CONCLUSION. FloTrac TM system seems to be reliable for monitoring CO in the normal range in critically ill patients, but may underestimate CO in septic hyperdymanic patients. Larger population studies as well as subgroup analysis are needed to confirm these preliminary data.

Grant acknowledgement. FIS PI020533 and PI050164.

Hamzaoui O 1 , Monnet X 1 , Richard C 1 , Teboul J L 1 1 Réanimation Médicale, CHU Bicêtre, Le Kremlin-Bicêtre, France

Using the PiCCO TM device, cardiac output can be measured either by transpulmonary thermodilution or by pulse-contour method. The algorithm for continuous cardiac output determination by pulse contour-analysis has been shown to be reliable such that recalibration every 6-8 hours is considered enough. However, in the case of acute changes in vascular tone, the reliability of this algorithm may be still questionable. The aim of our study was to examine whether cardiac index measured by pulse-contour method agrees with cardiac index measured by transpulmonary thermodilution in the case of marked changes in mean arterial pressure (MAP).

METHODS. patients equipped with a PiCCO TM device (Pulsion, Germany) for monitoring of circulatory failure. The pulse-contour system was initially calibrated by triplicate arterial thermodilution and no further recalibration was performed over the following 6 hours (deactivation of automatic calibration). Each time the attending physician decided to perform thermodilution measurements, the pulse-contour cardiac index value was also recorded.

A total of 284 pairs of measurements of cardiac index were obtained in 31 patients. The pulse-contour cardiac index correlated with the thermodilution cardiac index (r2 = 0.68). The Bland-Altman analysis showed a bias of 0.09 and a standard deviation of 0.60 L/min/m2. Ninety six times in 28 patients (mean age: 62 ± 10 yrs), changes of MAP more than 15% were observed between baseline and one of any time of measurements. At baseline, their MAP was 87 ± 19 mmHg, their cardiac index was 3.7 ± 1.1 L/min/m2. Their mean change in MAP was 28 ± 17% and in thermodilution cardiac index was 15 ± 14%. Most of these patients suffered from severe sepsis (n=19) and 16 received norepinephrine at baseline. The pulse-contour cardiac index correlated with the thermodilution cardiac index (r2 = 0.75). The Bland-Altman analysis measurements showed a bias of 0.04 L/min/m2 and a standard deviation of 0.58 L/min/m2. In 31 patients, changes of MAP lower than 15% were observed 188 times between baseline and one of any time of measurements. The pulse-contour cardiac index correlated with the thermodilution cardiac index (r2 = 0.68). The Bland-Altman analysis showed a bias of 0.15 L/min/m2 and a standard deviation of 0.61 L/min/m2.

Our study showed that: 1) over a 6-h period cardiac index measured by pulse-contour method agreed with cardiac index measured by transpulmonary thermodilution and 2) acute changes in arterial pressure did not alter this agreement. INTRODUCTION. An arterial waveform algorithm which estimates Cardiac Output (CO) from an adapted arterial pressure transducer and a waveform analysis algorithm (Vigileo/FlowTrac, Edwards Lifescience) was compared at timed intervals over 6 hours to a validated indicator dilution (ID) method for CO measurement (1) . APCO uses a pressure based algorithm and estimates stroke volume from entered patient biometrics and statistical data derived from the the arterial pressure waveform. The algorithm uses no external calibration, as the literature indicates an inbuilt compensation to allow for aortic compliance and changes in arterial tone.

Any ICU patient with a LiDCOplus in clinical use was eligible for this study. Patients were excluded if there were contraindications either to lithium dilution, or if the boundary conditions for ID were not met. APCO was set up and allowed to stabilise over 10-15 minutes. LiDCO measurements were made at the outset and at timed intervals for 6 hours with the ID and APCO results recorded simultaneously.

. 10 ICU patients with a variety of diagnoses were studied. (sepsis n=3, trauma n=3, cardiac arrest n=2, respiratory failure n=2 -5 males 5 females -age range 19-74 years). Mean APCO was 5.8 l/min (2.7 -15.8 l/min) vs. mean ID 6.5 l/min (2.9 -15.2). R 2 = 0.46 with a bias of 0.63 and precision +/-4.2 l/min. There was an 81% error of agreement between the two methods.

The data showed a poor correlation between ID and APCO. The bias changed with increasing CO and had wide limits of agreement against an ID standard. Comparison with ID indicated an accuracy outside generally accepted limits of precision (+/-30%) for interchangeability of the measurement techniques (2) . It is recognised in this comparison that the numbers of patients are small limiting the degrees of freedom for the analysis which may result in widened limits of agreement. The data is worrying however and further patients are being studied. 

Acute respiratory distress syndrome (ARDS) is associated with a diffuse pulmonary inflammatory reaction with increase of pulmonary resistances. A new Doppler echocardiographic index combining systolic and diastolic time intervals (Tei Index) seems to be useful in the evaluation of right ventricular function during normal conditions and primary pulmonary hypertension (1).

Sixteen mixed-strain newborn piglets (age 30-40 days and body weight 6-8 Kg) were anesthetized with ketamine (2.5 mg/Kg i.m) and propofol (bolus of 5-7.5 mg/Kg ev and continuous infusion of 7 mg/Kg/hr). The piglets were intubated with an endotracheal tube (4-5 mm internal diameter) and mechanically ventilated. A 4 F angiography catheter was inserted via the right jugular external vein and placed in the main pulmonary artery to measure pulmonary arterial pressure (PAP). ARDS was inducted by endotracheal administration of saline solution (100 ml). Diagnosis of ARDS was established when PaO2/FiO2 ratio was < 200. Complete two-dimensional Doppler echocardiographic examination was performed before induction of ALI (T0), and 10 (T1), 60 (T2) and 120 min after (T3). Sampled for blood gas analysis were withdrawn at same times. Data are expressed as mean±SD.

Variables were compared using unpaired Student t-test.

Study results are summarized in Table 1 . (1, 2) this demands early goal directed therapy (EGDT), including SvO2 monitoring. The objective of our study is to compare our ICU population with the population described by Rivers (1).

A total of 71 patients in 3 intensive care units were included. All patients arrived at the intensive care units either directly from the Emergency Department or after acute surgery with severe sepsis or septic shock. Other hospitalised patients and patients after elective surgery were excluded. To determine SvO2, central venous or mixed venous oxygen saturation were measured as early as possible after insertion of a central venous catheter or pulmonary artery catheter. Hematocrit and lactate were determined from the first obtained arterial bloodsample.

At admission 48 patients (68%) showed a SvO2 > 70%. Fourteen (14) patients (19%) had a SvO2 of 60-70%, 9 patients (13%) showed a SvO2 < 60% while only 2 of these were below 50%. Our population showed a significant higher SvO2 (73.1% ± 10.6 vs. 48.9% ± 12.3) and lower lactate (3.1 mmol/l ± 2.9 vs. 7.3 mmol ± 4.6) and hematocrit (29.0% ± 5.8 vs. 34.7% ± 8.5) compared to River's population.

There is a significant difference in condition of the patients at admission between our ICU's and the ED where the EGDT paper originates. Therefore it is questionable whether optimisation of SvO2 is a useful goal in our patients. 

The severe physical and psychological effects of critical illness may take years to recover. A RCT using self directed rehabilitation manuals demonstrated improvement in physical recovery (1) . The aim of this study was to evaluate the effectiveness of a therapist led, structured rehabilitation programme for patients following discharge from critical care.

We recruited 8 ICU survivors admitted for >48h and discharged home to this prospective, before and after intervention study. Physical status was assessed using the 6 minute walk test (6MWT)(2) and the incremental shuttle walk test (ISWT) (3) prior to and following a six week course of cardiopulmonary exercise. We completed an assessment of hospital anxiety and depression at the start and end of the programme. Data were analysed using Wilcoxon Signed Ranks Test.

All patients who attended the post ICU rehabilitation programme showed a significant improvement in both walking tests and Depression score (see table) . A median increase of 107.5m (31%) was seen for the six minute walk test and an increase of 80m (29.5%) for the incremental shuttle walk test. 6/8 subjects also demonstrated a decrease in both anxiety and depression scores on the HADS scale, 4 of which moved out of significant to normal levels. 

Survivors from intensive care (ICU) have a reduced quality of life (QOL) after hospital discharge. Planning optimal rehabilitation follow-up care requires a thorough evaluation of functioning after discharge from the ICU. The purpose of this study was to describe restrictions in daily functioning at discharge from the ICU and after 3 months.

In a consecutive series of ICU patients who were ventilated >48 hours, functional status was assessed within one week after ICU discharge with the Barthel Index (BI) and after 3 months with the Sickness Impact Profile 68 (SIP68) and BI. Patients were invited to a follow-up service 3 months after ICU discharge to evaluate functional status and to discuss perceived problems.

Of 56 patients assessed after ICU discharge, 4 died and 23 refused follow-up at 3 months. Consequently 29 patients were assessed after 3 months of whom 18 visited the follow-up service. Patients attending the follow-up service had better initial health status at ICU discharge (BI) than those not attending. One week after ICU discharge the majority of the patients had severe restrictions in performing basic daily activities (76% severe, 15% moderate and 9% slight dependency (BI)). After 3 months hardly any restrictions in basic daily activities were found, however in all patients, daily functioning was impaired (Median domain scores of the SIP68, range 0 to 100 with higher scores indicating poorer functioning: 'somatic autonomy' 0, 'mobility control' 42, 'social behavior' 50, 'mobility range' 30, 'emotional stability' 17, 'psychological autonomy and communication' 9). The majority of the patients reported problems related to dependency in daily activities, mobility, muscle weakness, memory and concentration, emotional instability, loss of taste and weight loss. None had returned to their job.

At discharge from the ICU, patients were dependent for basic ADL activities and three months later, daily functioning was still restricted. Considering that the follow-up service was attended only by a selection of the ICU population with a relative good health status, we conclude that it is urgent to carefully plan rehabilitation follow-up, in all patients with a prolonged ICU stay. Rehabilitation treatment should at least include support with respect to physical, psychological, and social functioning, and nutritional problems. 

The 6-min walk test (6'WT) and the treadmill ergometric test are simple and validated measurements of functional capacity that predict survival in patients with severe obesity and also cardiovascular diseases. This study aimed to compare these procedures in obese patients undergoing bariatric surgery.

Population (32 morbidly obese patients, 93.8% female, age 40.9+/-9.6 years, body mass index 51.3 -/+ 8.4 kg/m2) was analysed preoperative and 10-12 months after bariatric surgery (gastric bypasss). In the 6'WT patients were instructed to walk from end-to-end of a previously measured corridor, covering as much distance as possible in the allotted period of 6 min. They were allowed to stop and rest if desired, resuming walking as soon as they felt able to do so. In the fatigue rather than timelimited treadmill exercise (modified Bruce test) total distance, Borg scale of perceived exhaustion, and physical as well as cardiovascular variables were recorded. CONCLUSION. 1) The 6'WT reproduced the general pattern of the treadmill test with similar differences, and both were successfully completed; 2) The 6'WT was safe, inexpensive and could be conducted without specialized equipment; 3) This test is recommended for functional assessment of patients with severe obesity. 

Maximal oxygen consumption (VO2 max) is a valuable diagnostic and prognostic tool in obesity and cardiocirculatory diseases but studies with morbidly obese populations are uncommon. In a prospective protocol VO2max was investigated, aiming to define the effect of major weight loss on physiological response.

Participants were obese women (n = 65, age 40.4 ± 8.4 years; body mass index/BMI 49.4 ± 5.4 kg/m2) candidates for anti-obesity gastric bypass. VO2max was estimated by measuring the heart rate response to reference levels of submaximal work on a computerized treadmill (modified Bruce protocol). Preoperative ergometric test was compared to early (4-6 months) and late (10-12 months) post-operatory follow-up. Patients were clinically stable and free from acute disease.

Tolerance to the test was adequate and no complications occurred. BMI diminished from 49.4 ± 5.4 to 36.8 ± 3.7 and 30.4 ± 4.4 kg/m2 in the three observations, respectively (p<0.05). VO2max increased in the same proportion: 25.4 -/+ 9.3, 29.8-/+ 8.1 and 36.7-/+ 8.3 ml/ min. Progressive normalization (p<0.05) could be shown for the two postoperative periods in comparison with pre-surgical findings, which were substandard, despite the fact that no physical rehabilitation program was employed in this series.

CONCLUSION. 1) VO2max significantly benefitted from weight reduction, in parallel with greater exercise tolerance on the ergometric test; 2) Preoperative cardiorespiratory performance was inadequate and substantially recovered in the late postoperative period; 3) The ergometric test was safe and can be recommended for non-critical morbidly obese subjects; 

Interprofessional Education (IPE) aims to enhance motivation with others by securing a common knowledge base and resolve misunderstandings. IPE aims to overcome prejudices and negative stereotyping and acquires collaborative competences. IPE facilitates positive interaction that, through a chain of events, should ultimately improve patient care [1] . Adult learning theory and evidence from successful interprofessional education is used to support interventions to enhance successful working and overcome barriers. This paper aims to identify the challenges with IPE and to discuss them in relation to the implementation of the MSc in Critical Care at Cardiff University.

METHODS. An extensive review of the literature was undertaken and comparisons were made to the MSc in Critical Care at Cardiff University.

There are obstacles to IPE such as attitudes, educational, financial and organisational obstacles but, at Cardiff University, we have embraced the challenges of IPE to ensure that critically ill patients are managed through interprofessional initiatives that are based on specialist knowledge, rigorous research and advanced understanding of the concepts involved in their care. This is delivered through the MSc in Critical Care, a distance learning, international, interprofessional course which, at the time of its inception, was unique and, as recognised by Parsell et al (1998) , it encourages self directed, critical thinking, life long learners who are able to function as a team memeber [2] .

At Cardiff University we strive to address the challenges presented by interprofessional education and the students of our MSc evaluate positively the opportunity to learn alongside and from other professionals and other disciplines. We will continue to address the challenges and develop our interprofessional MSc in Critical Care to facilitate improvements patient care. 

Parenti N 1 , Manfredi R 1 , Baldisserri C 1 , Martini U 1 , Lanzoni S 1 , Lenzi T 1 1 Emergency Department, Santa Maria della Scaletta Hospital, Imola, Italy

We evaluated if a triage course improve the capability of the triage nurse to categorize correctly patients and its impact on the waiting time for physician examination.

METHODS. This is an observational study of 360 consecutive patients admitted from Jan.to Dec2005.

All medical records were examined by 4 nurses and reviewed by 2 physicians. We excluded patients with life-threatening conditions. We collected:nurse triage category, time of initial evaluation by a triage nurse and by a physician, physicians diagnosis, demographic and clinical characteristics. Quality indicators of triage efficiency were:the urgency category agreement among nurses and investigators; documentation of triage form (presenting problem, history and vital signs, initial triage category and assessment area allocated); compatibility of triage nurses urgency category and the waiting time to physician examination. We compared the previous quality indicators before and after a 2 weeks course on triage. All the triage nurses attended the course. The triage urgency category (UC) were: urgency 1=immediate response; urg. 2 assessment within 20 min; urg.3 within 60 min; urg.4 within 120 min. T or chi square test were used to compare 2 groups.

. 180 patients were included in the before-course (period 1) and 180 in the after-course (pe-riod2) group, mean age 48.7 and 56 yrs(p=0.007). The 2 groups had similar demographic and clinical characteristics (p>0.05). In the after-coruse group there were more patients in urg.cat.(UC)3(68.3% vs 61.1%; p>0.05). After the course there were more cases of full agreement among the UC assigned by nurses and investigators (81.7%vs 76.7%; p>0.05) but there were fewer triage forms with a complete documentation:98.3%vs 98.8%; p>0.05. In period 1, we found that 70% of patients in UC 2 were examined by a physician within 20 min., instead in period 2 all patients (100%) in UC2 were examined by a physician within 20 min.(p<0.01). Almost all (98%) the patients in categories 1,3,4 were examined within the expected time, in both periods. There was significant difference between the average waiting time for physician examination in UC2 before and after the course:33.2 min vs18.5min (p<0.05).

Our triage course seems to improve the capability of the triage nurse to categorize correctly emergency patients but it isn't able to improve the documentation quality indicators. Moreover the course had a strong impact on the waiting time for physician examination mainly in category 2 patients. Study limitations are:few patients and statistical power of results. METHODS. This is an observ. retrosp. study of 414 pat. admitted to our ED from Dec.2005 to Jan 2006. All records were examined by 4 nurses and reviewed by 2 physicians (ph). We excluded pat. with life-threatening conditions and collected:nurse triage category, time of initial evaluation by a triage nurse and by a ph., physician's diagnosis, demographic and clinical characteristics, the means of transport used, duty time, nurse experience in triage, ED crowding. We considered mistriage a triage with disagreement on the urgency category (UC) among examined nurses and investigators or with incomplete documentation: presenting problem, history, vital signs, initial triage category, assessment area allocated, retriage category. The triage UC were:urgency 1=immediate response; urg.2,3,4 assessment within 20,60,120 min. T or chi square test were used to compare 2 groups. We calculated the odds ratio (OR) and the logistic regression coefficient (r).

RESULTS. 307 pat. were included in the right triage group and 107 in the mistriage group, mean age 55.8 and 53.3 yrs (p=0.5). In the mistriage group 64% of pat. had an incomplete documentation and 36% had an under-triage. In the under triage group 10% of pat.in UC4 were assigned by investigators in UC3; 82% in UC3 were assigned in UC2; 8% in UC2 were assigned in UC1. In the mistriage group there were more women (57% vs 42.3%; p=0.03), more UC3 (72% vs 62%; p<0.05), more foreigners ( 

Norman S E 1 1 Anaesthetics, Cardiff University, Cardiff, United Kingdom

The United Kingdom's Department of Health (DOH) recognises that the shortage of human resources within the National Health Service (NHS) is its biggest constraint [1] . There is still little evidence to support and inform managers in the recruitment and retention of critical care nurses. The aim of this paper is to critically analyse and evaluate factors which influence the recruitment and retention of critical care nurses.

The literature was reviewed to identify current recruitment and retention strategies. This formed the foundation for the questionnaire used to perform the primary research. 176 qualified and non-qualified nurses across 2 sites were surveyed. Questionnaires were returned in sealed, unidentifiable envelopes and analysed by a computerised program. A combination of quantitative and qualitative methods of analysing the data collected was used. The recruitment and retention strategy of the directorate was identified and compared with the literature review and the evidence from the primary research.

A response rate of 69% (n=122) was achieved. The study found that factors considered to be important for recruiting staff were not necessarily the same factors needed to retain staff. For example, whilst education was identified by 66% of respondents (n=81) as an attracting factor when they joined the critical care unit, just 44% (n=54) said that it was an essential to them at the time of the study. 74% (n=90) of the nurses stated that communication and job satisfaction were essential criteria for retention. These were not identified as attracting factors. Although equitable pay was said to be essential in retention by 72 nurses (59%), only 4 nurses identified it as a top priority, and it is interesting to note that 18% of nurses (n=22) identified the enhanced overtime rates of the unit as an factor that attracted them to the unit. Promotional opportunities were only considered essential by 29 nurses, but again it is interesting to note that 46% of respondents had received a promotion during this employment. 46% of nurses had admitting to seeking alternative employment within the previous six months.

CONCLUSION. This directorate complies with the findings of the literature in the recruitment of staff by offering incentives such as educational opportunities, enhanced overtime rates and self-rostering, which are major attracting factors. However, these do not guarantee the retention of staff. The staff identified issues such as communication, support and job satisfaction to be the most important factors. These are not highlighted in the directorate's recruitment and retention strategy and is recommended. Further research in the field of rention is recommended. 

Critical Care Units prepare and administer tens of thousands of Intravenous Medications to patients every year. This process can pose serious risks to patients if the correct protocols are not observed, resulting in potential harm, and also increased organisational resource 1). Recognition of these risks has encouraged the UK Department of Health to identify I. V. Medication errors as an area that requires action to reduce patient harm (2). The Study was conducted across 8 Critical Care units in the Greater Manchester Critical Care Network.

A structured observational technique was used to assess the preparation and administration of IV drugs (3). Observers were recruited and trained to help with the study. Practice was audited against the NHS Scotland Clinical resource and audit group's 'Good practice statement (4).

A total of 1288 preparation and administration episodes of IV drugs were observed. With respect to drug preparation areas where practice fell below the audit standards included aseptic technique and checking drugs and patients with colleagues. With respect to drug administration areas where practice could be improved included problems with aseptic technique, and administering drugs that had been prepared by the previous shift.

In summary, a number of problems with the preparation and administration of drugs at the patients' bedside in critical care units have been identified. Work is now underway to disseminate information, standardise polices and provide competency based education to improve practice. METHODS. An Excel-based medication calculation test containing 27 medication calculations was developed and implemented. So far 47 nurses have been tested and more results will be reported in April 2006. The calculations contained practical clinical examples and they were systematized from easy to difficult. The time limitation (20 minutes) for the test was used to represent the normal time pressure and workload in nursing practice. The calculations of the test had also a mathematical dimension. According to that 14 questions should be able to solve within 20 minutes time limitation. Immediate feedback and test results (paper version) were printed by the Excel application after the test.

The maximum of correct answers in 20 minutes time were 26 calculations and minimum 0 (unanswered). The average sum of the correct answers was 13 calculations within the test time limitation. Most errors were made with decimal points, conversions and dosage calculations.

Competent and safe medication administration needs regular testing. The study showed that ICU nurses need more practical exercises and time to revise their calculation skills. In our unit future directions should also focus on testing the pharmacological skills of ICU nurses.

Kruger H L 1 , Spanjersberg R 1 , Ligtenberg J J M 1 , Meertens J H J 1 , Zijlstra J G 1 , Monteban W E 1 , Tulleken J E 1 1 Intensive and Respiratory Care Unit, University Medical Center Groningen, groningen, Netherlands

For each part of ICU equipment, the information provided includes an explanation of how it works, when and for how long it is generally used, and possible complications. The primary aim of our study was to evaluate the basic knowledge of frequently used medical equipment.

By means of an anonymous questionnaire, all medical workers on our 4 different ICU's in our hospital were asked to answer the following: 1. Assume an oxygen cylinder with a content of 5 litres. The pressure is 100 bar. Your patient is receiving 10 ltrs/min.oxygen. How long will this oxygen supply last? 2. Choose the syringe pumps which is used on your ICU (Asena or IVAC). They are fully charged. How long will the battery of the syringe pump last?

The medical workers were not informed in advance about this questionnaire.

. 58 medical workers completed the questionnaire. The results show that 45% answered question 1 correctly. Question 2 was answered correctly by 7% of the participants.

Critical Care nurses (25)

CONCLUSION. Education is mainly directed at complicated equipment. The adequate functioning of basic equipment is taken for granted. Although this might be justified almost always, basic knowledge of this equipment should be improved just to be certain in exceptional situations.

Akerman E 1 , Larsson M 1 , Joachim D 1 , Bergenzaun L 1 , Ersson A 1 , Chew M 1 1 Intensive Care Unit, Malmö University Hospital, Malmö, Sweden

Pulse Contour cardiac output (PiCCO) is a less invasive method of continuous haemodynamic monitoring [1] . As with any equipment, familiarity, education, perceived usefulness and safety are important issues. Since nurses are constantly at the patient bedside, we reasoned that a nurse-driven approach for calibration and monitoring may be advantageous in terms of providing better consistency and quality of measurements, and act as an early warning system for intensive care unit (ICU) physicians.

Over a 6-month period we educated ICU nurses in the use of PiCCO. We identified "core" personnel (3 nurses) who acted as reference persons with backup from 4 ICU physicians. The core personnel underwent a training program consisting of 1) a 60-minute lecture 2) computer based educational module 3) computer based competency test 4) practical simulation 5) bedside demonstration. Training was extended to all ICU nurses, consisting of a 45-minute theoretical and practical session, then supervised bedside training. Six months later, we issued a questionnaire to all nurses to evaluate the perceived value of education, ability to use PiCCO technically correctly, ability to identify haemodynamic changes and act upon them, and perceived technical difficulties.

. 41 (of 52) nurses completed the program. We have monitored over 100 patients in our ICU (10 ventilated beds, 900 admissions/year) with one adverse event (bleeding at site of insertion requiring surgery).

The nurse-driven PICCO program was implemented sucessfully. Most nurses felt that they had increased their knowledge of haemodynamics and obtained the necessary technical skills to carry out measurements correctly. Patient care was thought to have improved by earlier identification haemodynamic problems, resulting in changes in medical management. For our ICU, the program was an excellent way for increasing nurse awareness and the perceived quality/consistency of haemodynamic monitoring. 

Knowledge of basic aspects of critical care was assessed among a group of 462 nurses at the five General hospitals in Western Macedonia. A three-sections (demographic characteristics, acute care and rehabilitation) multiple choice original questionnaire was distributed to the nursing staff of each hospital to be voluntarily and anonymously filled in and returned at once. Nurses were requested to answer 8 questions of acute care related to equipment, drugs and techniques of CPR and 8 questions of rehabilitation related to tracheal suctioning, tracheostomy, lines and catheters placed in the ICU.

. 207 nurses answered (44.79% of the total), 30 males and 167 females, having 11.25±7.64 years of previous practice and 10.52±2.79 average total score (maximum total score 16). 35.7% of them were working in a surgical department and 64.3% in a medical one. Factors related to a higher score of correct answers were a three-year education, the male sex and the number of attended nursing congresses. The questions with the higher and lower percentage of correct answers concerned tracheal suctioning, the use of self-inflating bag-valve mask apparatus (>80%) and the use of defibrillator (45.7% . Within our intensive care unit weaning protocols have been used. Feedback from nursing staff highlighted a subjective lack of knowledge relating to respiratory physiology and mechanical ventilation. They also wanted to develop skills to allow confident recognition of airway and respiratory problems and interpretation of information necessary for decision making. This information formed the basis of a programme of work aimed at promoting independent practice within a group of nursing staff.

A study event was held with consultant medical staff, clinical educators and senior nurses. Eight junior nurses attended the day. Critical care nursing experience within this group ranged from 0 to 3 years. The expectations of each staff group regarding desired outcome and further support required to achieve this were identified. These expectations were evaluated after 6 months using individual interviews.

Notes from the interviews were themed with the purpose of identifying the important components of a learning culture within our critical care unit. The themes from each interview were amalgamated and the following identified by each staff group as being essential for a learning environment. 

Whilst formal and informal education is essential for enabling critical care staff to learn, the development of a learning culture requires other processes such as good relationships, mentorship and role modelling. Common to all groups of staff were the requirement for self-motivation and a culture of critical inquiry. Future education of staff to promote independent practice should incorporate these themes. 

Data concerning acid aspiration induced lung injury are mainly focused on the acute effects. Our aim was to develop a murine model of acid aspiration allowing animals' spontaneous breathing and long term survival. We also tested the efficacy of surfactant therapy in this model

During mechanical ventilation hydrochloric acid (pH 1.5; 1.5 ml/Kg) was selectively instilled into the right lung of anesthetized CD-1 mice via an intratracheal angiocatheter. Mice were divided in 3 groups studied at 12 (12H, n=6), 24 (24H, n=23) hours and 2 weeks (2W, n=22). At each time, left ventricular blood was sampled and PV curve obtained. The lungs were removed for histological procedures and Wet to Dry (W/D) ratio determination. A CT scan was obtained in four 12H mice.

In a different group of mice (12 HS, n=4) we evaluated the effect at 12 hours of early (10 minutes after injury) surfactant (2 ml/Kg) instillation.

RESULTS. 2 weeks survival was 73%. PaO2 decreased at 12H and 24H and, to a lesser extent at 2 weeks. Histological injury was limited to the right lung and was more evident at 12H and 24 H than at 2W, the latter showing a small fibrotic scar and less inflammatory cells. Lung compliance, reduced in the first 24 hours was still below control values at 2W. The W/D ratio was increased in the first 24 hours but returned to normal at 2W. CT at 12H showed areas of consolidation in the right lung, associated to hyperinflation in the left lung. Both CT and histology showed a lower injury extension in 12HS compared to 12H, associated to improved gas exchange. 

We characterized a murine model of acid aspiration, suitable for studying, both at short and long term, the evolution and/or possible therapies, like the administration of surfactant. 

We describe the whole lung lavage (WLL) technique adopted in our centre and report our series of patients with Pulmonary Alveolar Proteinosis (PAP) undergone this procedure.

Between 1990 and 2006, 32 patients (25 females, 40±13 years) underwent WLL; 8 pts needed multiple lavages: overall WLLs were 46. Under general anesthesia, pts are ventilated through a double-lumen tracheal tube; invasive pressures, temperature, weight, respiratory mechanics, gas exchanges and electrolytes are monitorized. WLL is performed in lateral position: we ventilate the dependent lung and lavage the nondependent one with tidal volumes of warmed normal saline. Electrolytes continuously lost in a dialytic-like way are replaced. To improve the removal of intralveolar material we use chest wall clapping. Procedure ends when the fluid drained becomes clear.

Mean lenght of WLL was 4.9±1.4 h, volume of lavage fluid 48±10 l. ICU stay ranged 36-72 h. At the beginning of the procedure PaO2/FiO2 (P/F) was 229±83 mmHg (PEEP 4±2 cmH2O); P/F decreased during one lung ventilation to 89 ± 63 (PEEP 4±2) and 117±95 (PEEP 8±3) when lavaging the 1 • and 2 • lung; at the end of WLL (PEEP 8±2) P/F was 248±117. During the procedure pts developed moderate hypercapnia (PaCO2 47±7 mmHg; p<0.001 vs basal value), base deficit (BE -2.4±3 mEq/l; p<0.01) and acidemia (pH 7.32±0.01; p<0.001). The day after WLL Creatine Kinase (88±76 vs 59±39 U/l; p<0.01) and Lactic Dehydrogenase (654±232 vs 593±208 U/l; p<0.005) slightly increased. Complications: 1 bronchial lesion and several not clinically significant cases of electrolytes abnormalities, fluid overload and hypothermia; no cases of massive flooding of the dependent lung.

CONCLUSION. WLL is a very complicated and potentially harmful procedure requiring careful monitoring and expert operators to minimize complications. The severe hypoxemia may make not feasible the one lung ventilation in these pts. We face the problem optimizing ventilation/perfusion ratio by lateral positioning and ventilation of the dependent lung. Lateral decubitus also increases effectiveness of chest wall clapping. In our pts the gas exchange improvement, not evident soon after WLL, became significant 15 days later, persisting in most cases at long term follow up (1). 

Tirumala S 1 , Malathi T 2 , Mantha S 1 , Yadagiri B 2 , Gopinath R 1 , Pasha S 1 , Reddy S 1 , Murthy T 1 1 Anaesthesiology and Intensive Care, 2 Biochemistry, The Nizam's Institute of Medical Sciences, Hyderabad, India

High peak inspiratory volumes and pressures and high mean airway pressures are predisposing factors for lung injury (Biotrauma). These are usually associated with volume controlled ventilation (VCV). Pressure controlled ventilation (PCV) may be beneficial due its advantages on respiratory mechanics. Studies comparing the effects of two modes in ARDS were inconclusive, could be due to confounding variables. We hypothesise release and temporal trend manifestation of the airway inflammatory mediators in patients with normal lungs reflecting biotrauma resulting from two different modes of ventilation.

After IRB approval and informed consent,15 adult surgical patients were randomised into two groups to receive PCV (G1, n=7) or VCV (G2, n=8). ARDSNET protocol followed for ventilation in either mode for 24 hours, using total intravenous anaesthesia. IL6, IL8 and TNFalpha were measured at 4 periods ie at 0, 6, 12 and 24 hours in the BAL specimens using ELISA method. Statistical analysis was performed using non-parametric repeated measures ANOVA (Friedmans test). Posthoc analysis was done for multiple pairwise comparisons. Table 1 . 

Dembinski R 1 , Hochhausen N 1 , Bensberg R 1 , Stadermann F 1 , Kopp R 1 , Kuhlen R 1 1 Intensive Care Medicine, University Hospital, Aachen, Germany

Preservation of spontaneous breathing activity and inhaled nitric oxide (iNO) have both been recommended to overcome severe hypoxemia in patients with ARDS. The aim of this study was to determine effects of iNO on ventilation-perfusion (VA/Q) distribution during different modes of partial ventilator support (PVS) in experimental acute lung injury (ALI).

METHODS. 24 pigs with ALI were randomised to four groups ventilated with assist-control (A/C), volume assured pressure support (VAPS), pressure support (PS) or biphasic positive airway pressure (BIPAP). In each group, controlled mechanical ventilation (CMV) and PVS were performed for one hour each followed by 15 minutes of 10 ppm iNO. VA/Q distributions were determined after CMV and PVS with and without iNO.

In all groups, iNO increased perfusion to lung areas with normal VA/Q distribution thereby reducing shunt during CMV and PVS (see table) . In contrast, significant improvements of VA/Q distribution due to PVS were only determined for PS and BIPAP. However, even in these groups improvements of VA/Q distributions due to combined PVS+iNO did not reach statistical significance when compared to CMV+iNO. 

ASV is a microprocessor-controlled mechanical ventilation mode that independent of the patients' activity maintains a preset minute ventilation as a function of patients' ideal body weight (IBW). Tidal volume (VT) is chosen by the ventilator. We determined which factors influence VT with ASV, such as institute, patients' height and weight, level of positive end-expiratory pressure, gender and the difference between set weight (set-W) and (IBW).

We prospectively collected data of 270 consecutive post-cardiac surgery patients in 3 Dutch IC units

Mean VT was 8.3.± 1.6 ml/kg IBW. In 43.3% the correct W was used; in 16.5%, 40.2% set-W was too low or too high respectively. A wrongly set-W was the only factor that influenced VT (P < 0.05), as shown in the graph.

Incorrectly set W in ASV results in undesirable large VT. METHODS. 185 consecutive patients of a surgical ICU were reviewed. Inclusion criteria were an assumed ICU stay of at least 6 days. Exclusion criteria were refused consent, previous pancreatic disease, operations of the pancreas, alcohol abuse, pregnancy, age of less than 16 years. Plasma levels of PaP were examined on a daily base. Murray (lung injury) and APACHE II (severity of disease classification) scores were calculated. Patients were divided according to amount and duration of ventilator support. Groups were compared according to their maximum plasmatic PaP levels and their initial PaP level (maximum within the first three days on ICU). Mean values were calculated and groups compared by t-test.

A total of 87 patients were included (34 women, 53 men, mean age 45 years, mean time on ICU 17 days, 22 died, mean APS 6.8, 48 trauma patients, 39 non-traumatic). Low maximum PaP levels were found in patients with 5 or less days of ventilation during their ICU stay (mean 468 ng/ml, n=12). Mean maximum PaP rose up to 885 ng/ml (n=29) in patients being ventilated 6 to 10 day during their ICU stay, and up to 2515 ng/ml (n=43) when being ventilated for more than 10 days. These differences for short and long ventilation periods were significant (p=0.00002 for <=5 vs >5 days of ventilation and p=0.0003 for <=10 vs >10). No statistically significant difference was found for initial PaP levels in severely ill patients in relation to total days of ventilation. Mean initial PaP levels were 116/192/450 ng/ml for total ventilator time of <=5/5-10/>10 days of ventilation.

Significantly higher plasmatic PaP levels were seen in patients with longer duration of ventilation. Our data do not support any predictive value of early plasmatic PaP measurement on the degree of respiratory failure in intensive care patients.

Ferreyra G 1 , Squadrone V 1 , Baussano I 2 , Richiardi L 2 , Marchiaro G 1 , Merletti F 2 , Ranieri V M 3 1 Department of Anesthesia, 2 Department of Epidemiology, 3 Anesthesia, University of Turin, Torino, Italy

Postoperative pulmonary complications (PPCs) are the most important determinants of acute respiratory failure (ARF) in postoperative abdominal patients with an increased risk of prolonged ICU stay due to mechanical ventilation requirements. In the present study we evaluated the efficacy of the early use of continuous positive airway pressure (CPAP) to prevent pulmonary complications, atelectasis, pneumonia, and endotracheal intubation in patients recovering from abdominal surgery.

METHODS. MEDLINE, EMBASE and COCHRANE database, were searched to identify relevant randomized controlled clinical trials examining the use of CPAP versus standard therapy in patients undergoing abdominal surgery. Outcomes were extracted from these articles, and a meta-analysis was performed.

RESULTS. 9 randomized controlled trials reached standard of quality and were included in the analysis. Overall, CPAP significantly reduced PPCs by 41%, atelectasis by 37%, and pneumonia by 67% in comparison to standard treatment (O2 treatment by face mask). Heterogeneity was negligible. CPAP showed a significant decrease of endotracheal intubation rate (RR 0.15;95% CI of 0.034 to 0.66). 

Ouragini H 1 , Baccar K 1 , Gargouri A 1 , Chaoua T 1 , Lamourou M 1 , Kaddour C 1 1 Anesthesia and intensive care, National Institute of Neurology, tunis, Tunisia

Many physiologic changes affect respiratory system and all organ during pregnancy. The respiratory system undergoes changes and is subject to additional functional and anatomic stresses. Pathophysiologic processes may result in significant ventilatory compromise, adversely affecting maternal and foetal oxygenation (1).

We conducted a prospective study over 5 years (July 2000 to June 2005) involving 116 women with pulmonary disorders in pregnancy (PDIP). These patients were selected among 389 obstetrical patients admitted in our Intensive Care Unit. The diagnosis of pulmonary disorder was made on the basis of clinical data, arterial blood gas, and/or a pulmonary artery monitoring wedge pressure of at least 18 mmHg, findings that were confirmed by chest X-ray or CT-angiography.

. PDIP incidence in our series was 29.8%. Thirty seven percent of our patient were primigestous, 41% were primiparous, and 58% were multiparous. Matrenal age average was 32 years (19 to 44). Eighty percent of women underwent caesarean section; spontaneous delivery was observed in 17% and 32% were instrumentally delivered. Pre-existing medical consists of cardiovascular diseases (n=14; 12%), thromboembolic events (n=2; 1.7%), asthma (n=3; 2.5%), hypothyroidism (n=1; 0.8%) and diabetes (n=1; 0.8%). PDIP was attributes to eclampsia (55%), to obstetrical sepsis (19%) and to haemorrhage (26%). Eighty nine percent of patients required mechanical ventilation. Global maternal mortality was 14%. Causes of death was septic shock (n=5), intracerebral haemorrhage (n=3), cardiogenic shock (n=3) and pulmonary embolism (n=1) and inhalation syndrome (n=5).

CONCLUSION. PDIP is a serious complication. Patient at risk should be delivered in appropriate conditions. Close collaboration between obstetrics and ICU department may improve prognosis of such patients. Clinicians are concerned about the optimal time period to be allowed before obtaining a blood sample following a change in oxygenation settings. The exact timing for ABGs measuring in ICU patients (pts) with oxygenation impairment has not yet been well documented. The purpose was to determine the time required for the arterial oxygen tension (PaO2) to reach equilibrium after a 0.3 change in fractional inspired oxygen (FiO2), in mechanically ventilated (MV) ICU pts.

We studied 20 adult MV pts (10 men, 10 women) admitted in a 7-bed multidisciplinary university ICU, using serial ABGs measurements. The pts were on ACMV and divided in 2 groups according to the PaO2/FiO2 ratio (1>200=10 pts, 2<200=10 pts). Their oxygenation and hemodynamic status was stable on the day of the study. The initial FiO2 was selected according to the pt needs for an adequate PaO2. Following two baseline PaO2 measurements at the initial FiO2 at 5 min intervals, to confirmed respiratory stability; the FiO2 was increased by 0.3 for 30 min and then decreased by 0.3 returning to the initial value, without any other change in respiratory or hemodynamic parameters. Sequential ABGs measurements were performed in 3, 6,9,12,15,20,25 & 30 

Sifrer F 1 , Drinovec I 1 1 Intensive Care Unit, University Clinic of Respiratory Diseases and Allergy, Golnik, Slovenia

Hyperglycaemia is associated with poor outcomes from pneumonia, myocardial infarction, stroke and acute respiratory failure in chronic obstructive pulmonary diasease (1, 2) . A study was undertaken to determine the relationship between mechanical ventilation, presence of diabetes, mortality and duration of mechanical ventilation.

Data were retrieved from medical records for patients admitted in 2005 to medical intensive care unit with acute respiratory failure and mechanically ventilated (92 patients). The patients were grouped according to presence of diabetic hyperglycaemia (25 patients, 27%), survival, duration of mechanical ventilation. Blood glucose level was determined on admission. Student t test was used to assess differences between groups.

. 67 patients (73%) were normoglycaemic nondiabetics. 44 (66%) survive mechanical ventilation. Their mean duration of mechanical ventilation was 9.9 days. Nonsurvivors (23 pts, 34%) were mechanically ventilated for 11.6 days (p NS). 25 patients (27%) were known diabetics. Survivals (16 pts, 64%) have an average blood glucose level of 9.3 mmol/l and were mechanically ventilated 5.2 days. Nonsurvivors (9 pts, 36%) have an average blood glucose level 13, 5 mmol/l (p 0.13) and were mechanically ventilated for 8.5 days (p 0.05).

The presence of diabetic hyperglycaemia is not associated with increased mortality in mechanically ventilated patients. In isolated group of diabetic patients increasing hyperglycaemia increases mortality. A prospective study is required to determine whether control of blood glucose can improve survival of diabetic patients with acute respiratory failure. 

A lateral site of insertion, in the "safe triangle", is recommended in patients lying supine to limit the risk of vessel injury and to avoid unsightly scar. Using this route, however, the chest tube might frequently enter the pleural space above the upper level of a pleural effusion and/or below the inferior limit of a pneumothorax, increasing the risk of chest tube malposition.

During an 18-month period, the charts of all patients admitted in the surgical intensive care unit with one or several percutaneously inserted chest tubes visible on thoracic CT scan (TCT) were reviewed. Among them, only patients who had had a TCT in the 24 hours preceding tube insertion were analyzed for the present study. By comparing the two TCT, it was possible to measure the distance existing between parietal and visceral pleural layers at the level of the intercostal space through which the chest tube was inserted. The local ethic committee (CCPPRB of CHU la Pitié Salpétrière) approved the protocol that did not require informed consent.

. 39 chest tubes (20 right and 19 left) inserted in 30 patients (age 44 ± 17 years, SAPS II = 38 ± 17) for pleural effusion or hemothorax (61%), pneumothorax (21%), or both (18%) were reviewed. The mean drained volume of pleural effusion was 588 mL (± 418 mL). The distance between parietal and visceral pleura at the insertion site of the chest tube was superior to 10 mm in 5 cases, less than 10 mm in 10 cases, and pleural space was virtual in 24 cases (61%). Among the 24 tubes inserted in a virtual pleural space, 7 were intraparenchymal whereas the 15 tubes inserted in a fluid-or air-filled pleural space were all correctly positioned (p = 0.03, exact Fisher test).

Physicians should keep in mind that when inserting percutaneously a chest tube by the lateral route, they may enter the thorax through a virtual pleural space, a situation that predisposes to intraparenchymal placement of the chest tube. [1] . Suctioning through a bronchoscope inserted through a tight seal connector may have similar effects. Our aim was to evaluate suction flow through a bronchoscope and monitor changes in functional residual capacity (FRC) during suctioning.

Suction flow at vacuum levels of -20 to -80 kPa was measured. In 3 ICU patients, a 16 Fr bronchoscope was inserted during volume and pressure control ventilation (VCV, PCV). FRC was monitored with electric impedance tomography calibrated by a nitrogen washout method. Tracheal pressure (Ptrach) was monitored via a 1.1 mm polyethylene catheter via the ETT lumen.

Suction flow through the bronchoscope was 17, 11, 8, and 5 l/min at vacuum levels of -80, -60, -40 and -20 kPa. During insertion FRC increased with 171-585 ml and decreased with 148-1326 ml during suctioning. In PCV Ptrach decreased to subatmospheric levels, -24 -0 cmH2O (range). After removal of the bronchoscope, FRC returned to baseline in VCV but not in PCV. See graph.

CONCLUSION. Suction flow through the bronchoscope at the vacuum levels used is above minute ventilation in most ALI patients. During suctioning, the ventilator was unable to deliver enough volume in both VCV or PCV to maintain FRC. In PCV tracheal pressure decreased below atmospheric pressure. 

Apolipoprotein A-I (Apo A-I), the major protein component of HDL, decreases during the acute phase of inflammation. In vitro, Apo A-I specifically inhibits the contact-mediated activation of monocytes by stimulated T cells, decreasing the production of inflammatory mediators. Propofol (Prop) widely used as an anesthetic agent, dissolved in a lipid emulsion, is highly lipophilic. In vitro, Prop seems to modulate the inflammation. The aim of this study was to evaluate whether Prop modulates in vivo the concentration of Apo A-I after a surgical stress. The secondary objective was to investigate whether it is Prop, or its solvent that is responsible for the action on Apo A-I.

Triple blind RCT in patients undergoing laparoscopic hernia repair comparing 3 different anesthesias: Group P: induction and maintenance with Prop, Group S: without Prop but with its solvent, Group C: without Prop or solvent. We assessed Apo A-I before (TB), after (TA) induction of anesthesia, 5 (T5) and 24 (T24) hours after the surgical stress. CONCLUSION. Prop and its solvent seem to modify the Apo A-I during the first 24 hours after the surgical stress. The mechanism by which Prop or its solvent modulate inflammation should be further investigated.

Yukioka H 1 , Kato N 2 1 Department of Anesthesia, Emergency and Intensive Care, Yukioka Hospital, 2 Department of Emergency and Critical Care Medicine, Osaka City University Medical School, Osaka, Japan

Carbon monoxide (CO) and bilirubin are produced by heme catabolism due to induction of heme oxygenase-1 by inflammation. Since increased endogenous CO production has been reported in patients with severe sepsis (1), this study was undertaken to determine the correlation between exhaled CO and plasma bilirubin concentration in patients with hyperbilirubinemia.

Sixteen adult patients with hyperbilirubinemia (total plasma bilirubin level greater than 2mg/dL) including 3 without systemic inflammatory response syndrome (SIRS), 9 with SIRS, and 4 with sepsis and eight control patients without hyperbilirubinemia without SIRS (4 patients) or with SIRS (4 patients) were studied. Exhaled CO concentrations in end-expiratory air samples collected into a plastic bag were measured using a CO analyzer (Carbolyzer mBA-2000, Taiyo, Osaka, Japan) with a sensitivity of 0.1 ppm. Measurements of exhaled CO, arterial carboxyhemoglobin (CO-Hb) using a blood gas analyzer (ABL700-series, Radiometer, Copenhagen, Denmark) and of total plasma bilirubin levels were performed several times for each patient. Data analyses were performed with Student's unpaired t-test or Spearman's rank correlation test.

In 7 healthy volunteers, exhaled CO level was 2.2±0.4 ppm, which was significantly (P<0.01) less than that in the control group (2.9±0.9 ppm). Exhaled CO and CO-Hb levels were significantly (P<0.01) greater in patients with hyperbilirubinemia than in the control group (5.1±1.9 versus 2.9±0.9 ppm, 1.7±1.1 versus 0.8±0.5%, respectively). A moderate correlation was found between exhaled CO and bilirubin concentration for the 13 patients with hyperbilirubinemia and SIRS/sepsis (rs=0.496, P<0.05), although no correlation was found between the two for the 16 patients with hyperbilirubinemia. For the 13 patients with hyperbilirubinemia and SIRS/sepsis, a good correlation was found between percent change in exhaled CO concentration and that in bilirubin level (rs=0.780, P<0.01).

Increase in exhaled CO concentration may be an important indicator of overproduction of bilirubin due to inflammation in patients with hyperbilirubinemia and SIRS/sepsis. 

The aim of our study was to compare PCT and CRP serum levels in patients post orthotopic liver transplantations (OLTx) with and without the administration of anti-thymocyte antibodies (ATG Fresenius) and without any complications.

Serum samples from patients after OLTx (21 recipients with and 7 recipients without ATG therapy) were evaluated. PCT and CRP serum concentrations were measured in OLTx recipients before induction of anesthesia, at hours 4 and 8 following graft reperfusion, and daily until postoperative day 4. PCT was also determinated in 12 patients undergoing liver resection.

PCT serum levels were slightly elevated (up to 14 ng/ml) in several patients after OLTx without ATG therapy. PCT was strongly induced in patients after OLTx with ATG administration (up to 249 ng/ml). The mean value of maximum PCT concentration was on first post-operative day 4.5±1.6 ng/ml in patients after OLTx without ATG therapy and 59.0±12.6 ng/ml in patients with ATG therapy (p<0.001). In addition, both groups are compared with 12 patients undergoing liver resection, whose mean serum PCT levels did not exceed 1.4±0.3 ng/ml. CRP serum levels in group of patients after OLTx with ATG therapy markedly increase four hours after graft reperfusion (22.6±3.2 mg/l), the highest level was post-operative day 1 (83.7±10.4 mg/l) followed by a decrease over the next days. In the group after OLTx without ATG therapy, there was a significant increase of serum CRP levels in post-operative day 1 (80.2±13.6 mg/l) followed by a decrease. In patients after liver resection the CRP level was highest on the second post-operative day (55.3± 12.0 mg/l) with decrease until post-operative day 4.

Polyclonal antibody administration to patients with OLTx is associated with a very marked increase in serum PCT levels, with peak values on post-operative day 1. However, this is without a clinical correlation in the form of a severe inflammatory response. The possible PCT release following ATG therapy should be taken into account when using this inflammatory parameter in transplant patients. The results of the study indicate that not only infectious inflammation is a stimulus for procalcitonin synthesis.

After approval by the local ethics committee and written informed consent from the individuals, 83 patients with a diagnosis of single or double vessel coronary disease undergoing coronary artery bypass grafting (CABG) were enrolled in the study (mean age 66 ). Of the patients, 47 underwent CABG with CPB (on-pump group), and the remaining 36 without CPB (off-pump group). Blood samples were drawn before surgery (T1), at the end of CPB or after finishing coronary revascularization on the beating heart (T2) and 12 hours postoperatively (T3). Serum concentration of sFas was measured by a sandwich enzyme-linked immunosorbant assay kit.

There was no statistical difference between groups with regard to age, gender and SAPS II score. The sFas levels before surgery were comparable in both groups (on-pump group vs. off-pump group, means ± standard deviation: 3.1±0.9 vs. 3.6±1.4 ng/ml). At T2 sFas levels were significantly lower in both groups as compared to T1 (on-pump: 2.7±0.9; p<0.05, off-pump: 3.1±1.2; p<0.05), without significant difference between groups. Nevertheless, the mean drop of sFas level was higher in the off-pump group. The mean values almost returned to baseline at 12 h after surgery in on-pump patients ( 

Severe hyperthermia commonly accompanies septic shock. High body temperature in absence of infection activates the inflammatory response, and is associated with a high mortality. Three years ago, our hypothesis that sustained fever is harmful in septic shock, led us to the development of a protocol aiming at decreasing hyperthermia (> 39.5 • C) by means of hemofiltration when the patients did not respond to antipyretics. We present a report of temperature and hemodynamic changes and outcome of 19 consecutive hyperthermic septic shock patients with multiorgan system failure and compare them with a historical similar group of patients in whom hyperthermia was not treated with hemofiltration.

Depending on renal function, patients were treated with continuous low-flow hemofiltration (n=8) or hemodiafiltration, (n=11). In all cases a PRISMATM (Gambro, USA) with a AN69 polyacrylonitrile filter (Hospal, France) was used. Core temperature was registered every hour. A hemodynamic index (HI) was defined (mean arterial pressure/noradrenaline dose) and used during the first 24 hours to describe the patients' hemodynamic profile by means of its percent variation starting 6 hours prior to instituting the hemofiltration.

No differences regarding age (p = 0.33), APACHE II (p = 0.73), SOFA score (p = 0.8) or HI at the moment of the diagnosis of hyperthermic septic shock (p = 0.9) were observed between both groups. All patients had multiorgan system failure involving, at least, the cardiovascular and respiratory systems. The patients' temperature decreased linearly from 39.8 ± 0.5

• C prior to hemofiltration to 37± 1.2 • C after 24h of treatment (p<0.001). The HI decreased significantly from -6 h to the onset of hemofiltration (p=0.002) and increased significantly after 24 h (p = 0.008). 28-day mortality was 32% (6/19) as compared to 100%(11/11) in the historical group (p<0.001).

Continuous low-flow hemofiltration decreased body temperature and vasopressor requirements in hyperthermic septic shock patients. The mortality was unexpectedly low. 

A new approach to SAP, which entails specialist centre referral, wide bore percutaneous drainage of infected necrosis and delayed (mean 41 days) surgery, is associated with a low mortality (5.7%) [1] . This study analyses the level of Intensive Care input required to effect this approach.

Retrospective chart and ICU database review of all patients admitted from January 1999-March 2005 with SAP.

Of the 35 patients described, 21 (60%) required ICU admission. Two other patients were admitted acutely to ICU and died < 48hrs before specialist surgical referral. All patients followed the described delayed surgical approach complemented by critical care interventions. Results are expressed as mean values and (range). The Imrie score on admission was 3 (1-5). The length of stay in ICU was 53 (3-143) days, and in hospital was 98 (24-300) days. Nineteen patients were mechanically ventilated for 24 days (1-108), five requiring percutaneous tracheostomy. Vasoactive therapy (noradrenaline+/dobutamine) was indicated in 19 patients (90.4%) and CVVHDF was instituted in 7 patients (33%). Seven (1-17) CT scans were performed per patient, with 3 (0-7) CT guided drains inserted per patient. Eight (2-21) different organisms were grown from 14 (3-38) sites per patient. Eleven patients required parenteral nutrition. Mortality was 2/21 (9.5%) in the ICU surgical group; 4/23 (17.4%) in the overall ICU group. 

Heat stroke is a life-threatening illness defined by an elevated core body temperature above 40

• C with central nervous system dysfunction which often results in neurological disabilities due to cellular destruction at extreme temperatures. We studied organ failure in relationship to duration of heat stroke [DHS] (time to hospital admission plus time to return at normal body temperature) to test the hypothesis that sustained HS is associated with increased morbid/mortality.

During August 2003 sixty-six adult patients with heat stroke in absence of other etiologies explaining the hyperthermia were admitted to 14 ICU in Europe for organ support and included in a multicentric retrospective study. For each patients DHS, severity scores, organ failure (SOFA) including cardiac failure assessed as serum troponin and endocrine failure assessed as ACTH test, Glasgow Outcome Score (GOS) at leaving from hospital and mortality were recorded. Patients were divided into two groups according to the median (12hrs) and early death:A for patients with DHS < 12 and B: patients with DHS > 12 and dead in hyperthermia.

Sixty-six patients with HS in absence of other etiologies explaining the hyperthermia took part at the study. Thirty-one died (47%) in ICU, 43 (65%) at hospital. SAPS II and SOFA score were associated with DHS. Troponin and cortisol were not different in the two groups while incidence of death (9 vs. 34, p<0.001), sepsis (4 vs.10, p=0.05) and neurologic disabilities (p<0.0001) was lower in sustained HS than in non-sustained HS group. 

Werneck P 1 , Knibel M 1 , Roderjan C 1 , Hatum R 1 , Guimarães E 1 , Vazan A 1 , Neto E 1 1 Intensive care unity, Sao Lucas Hospital, Rio de Janeiro, Brazil

Obesity has been pointed out as an independent risk factor for morbity and mortality at ICU.

Retrospective review of peritonitis cases admitted at ICU. Patients were divided into two groups regarding BMI: 1) super obese (SO) with patient on BMI above 35 kg/m2 and 2) non obese (NO) those with BMI under 35 kg/m2. 

Despite known risk associated to obesity, no differences on outcome was found in this small sample. These findings must point out caring of critical obese patient could be as successful as for a non obese one. 

Forty patients admitted in a medical ward with pulmonary disease were enrolled in this study. Blood was drawn on the day of admission. TREM-1 neutrophil and monocyte expression were measured by flow cytometry (Epics XL-HCL, Beckman Coulter) after erythrocyte lysis and labelling with PE-conjugated anti-TREM-1 monoclonal antibodies. In addition clinical and laboratory data including temperature, blood gases, full blood count, CRP, sputum gram stain and culture, blood cultures and imaging data were collected. Clinical data were evaluated by two physicians blinded regarding TREM-1 expression and bacterial infection was diagnosed in the presence of a combination of fever, productive cough, sputum neutrophils, positive sputum culture, absence of a non-infectious condition fully explaining the patient's symptoms and a favourable response to antimicrobial treatment. TREM-1 expression was compared between the groups with and without bacterial infection by the Mann-Whitney test. Receiver operator characteristic (ROC) curve was drawn. INTRODUCTION. Data about patterns of CRP in ICU patients with nosocomial BSI are sparse. We studied the dynamics of daily CRP levels to identify those patients with either good or worse outcome.

A historical cohort study of all patients with a nosocomial BSI admitted to the ICU between 1 January 2003 to 31 December 2004. Comparison of CRP levels from two days prior to microbiological documented BSI (CRP-2) till five days after onset of BSI (CRP+5).

Over the study period, 155 patients were documented with a bacteraemia during their stay in the ICU. Mean age was 52.8±17.9 yrs. Fifteen days, 30 days, ICU, and hospital mortality rates were 17.4%, 26.5%, 29.7%, and 42.6%, respectively. Mean serum CRP concentrations from day -2 till day +5 were not significantly different in survivors (n=89) compared to non-survivors (n=66) (see Fig. 1 ). Although CRP levels are already high, 2 days prior to BSI, a smooth increase is seen from day -2 till day +1 (11.5±8.5 vs 16.4±9.1, P=0.38), from which on CRP levels will decrease till day +5 after onset of BSI. This distribution is, however, identical between both groups.

CONCLUSION. CRP has no value in predicting mortality in critically ill patients with nosocomial BSI. 

Systematic colonization surveillance in the ICU permits monitoring of microbial transmission, early detection of epidemics and guidance for empiric antimicrobial treatment. We retrospectively studied the colonization dynamics of multi-drug resistant (MDR) pathogens in our ICU and analyzed the ability of colonization to predict microbial etiology of subsequent infections.

The study was performed in a new 5-bed general ICU for 28 months. Infection control policy included weekly surveillance cultures of bronchial and stool samples. Colonization dynamics were studied in patients with at least two consequent surveillance cultures (ie 2 weeks in ICU). All cases of ventilator-associated pneumonias (VAP) and bloodstream infections (BSI) were recorded and relationship between infectious etiology and most recent colonization was analyzed, based on species, antimicrobial susceptibility and molecular typing by REP-PCR of selected isolates.

A. Colonization dynamics analysis: Of 66 patients, 23 (35%) were colonized by at least one MDR pathogen on admission. For the rest, the most common colonizing microorganism was A.baumannii (78%), in an almost endemic presence. MDR P.aeruginosa (57%) and K.pneumoniae (46%) behaved epidemically, usually following admission of an already colonized patient. Colonization occured early with A.baumannii (median, 11 and 12 days, for RT and GT), later for P.aeruginosa (20 and 11.5 days) and even later for K.pneumoniae (19.5 and 24 days). B. Colonization-infection concordance analysis: Among 266 patients, 21 VAP and 74 BSI cases (41 catheter-related) were recorded. Pathogens isolated from VAP cases correlated with bronchial or stool colonizers in 83%, with prior RT colonization being most important. In both primary BSI and CR-BSI cases, Gram-negative pathogens were recent colonizers in 73% and 81%, respectively, associated with both GT and RT. REP-PCR techniques confirmed pathogen and colonizer concordance in all cases tested. Systematic colonization surveillance permitted 90% adequacy in VAP and 80% in primary bacteremia empiric antimicrobial treatment.

RT and GT colonization is strongly related to microbial etiology of subsequent infection and systematic surveillance could be helpful in implementing adequate antimicrobial therapy in the ICU. 

Elevation of serum procalcitonin (PCT) has been proposed as a marker of disease severity associated with sistemic infection. There are two methods: quantitative (PCTL) and semiquantitative (PCTQ). We intend to evaluate the correlation between the two methods.

We evaluate 222 blood samples from 83 patients with SIRS or sepsis in the ICU. PCTQ levels wee availed in four levels and compared with PCTL levels.

The analysis found a positive correlation between PCTQ and PCTL for levels > 2ng/dl and > 10 ng/dl. The sensitivity and specificity were 50% and 91% respectively.

The preliminary analysis suggests that PCTQ can be used to measure PCT level in values above 2ng/dl. Levels below that need more samples to provide more information. 

Procalcitonin (PCT) is thyroid gland prohormone, and its serum concentration is elevated in bacterial infections. The prognostic value of elevated serum levels of PCT in patients early after cardiac surgery on cardiopulmonary bypass (CPB) remains unclear. In RCT study, we investigated whether PCT is useful as prognostic marker in cardiac surgery, with respect to mortality, complications and infections, and whether procalcitonin-guidance could reduce antibiotic use in patients subjected to the open heart surgery.

METHODS. 205 patients subjected to surgical CABG, valve replacement or combined CABG + valve operations were randomly assigned for procalcitonin-guided treatment (procalcitonin group; n=102) or standard care (control group; n=103) from 02/05 till 08/05. On the basis of serum PCT concentrations, use of antibiotics was encouraged (≥ 0.5 ng/mL) or discouraged. Reevaluation was possible after 6-24-48 hours in both groups. Primary endpoint was use of antibiotics and analysis was by intention to treat. Operation data, laboratory data and clinical outcome: mortality, infections, severe complications were observed. 

Clec'h C 1 , Hamza L 1 , Jauréguy F 2 , Chouahi I 1 , Karoubi P 1 , Fosse J P 1 , Vincent F 1 , Cohen Y 1 1 ICU, 2 Microbiology laboratory, Avicenne teaching hospital, Bobigny, France

An early and accurate diagnosis of NP is of utmost importance because of its potential impact on patients' outcome. The diagnosis of ventilator-associated pneumonia (VAP) relies on protected specimen brush (PSB), bronchoalveolar lavage (BAL), and plugged telescoping catheter (PTC) with quantitative cultures. In the particular setting of NP occurring in non-mechanically ventilated patients, there is no consensus regarding the best diagnostic strategy. When mechanical ventilation (MV) becomes mandatory because of NP-related respiratory distress, tracheal aspiration (TA) performed immediately after intubation of the trachea could be thought of as a simple, fast and cheap way to identify the microorganisms involved and achieve early adequate antibiotic therapy. The aim of the present study was to compare the diagnostic accuracy of post-intubation TA (PITA) to that of PSB, BAL or PTC in patients requiring MV for suspected NP.

All consecutive patients with prior hospital stay >= 48 hours who required invasive MV for NP suspicion were prospectively enrolled. Immediately after intubation of the trachea, tracheal aspirates (PITA) were obtained by sterile suction. Within 2 hours, pulmonary samples were obtained either by PSB, BAL, or blinded PTC referred to hereafter as "reference methods" (RM). The thresholds for positive cultures were 10.3 cfu/ml for PTC and PSB, 10.4 cfu/ml for BAL, and 10.5 cfu/ml for PITA. The definite diagnosis of NP was based on a composite item of clinical, radiological and bacteriological (ie, blood or pleural fluid cultures) patterns. The agreement between PITA and RM was assessed by the kappa statistic. Sensitivity, specificity, positive likelihood ratio, and negative likelihood ratio of PITA and RM were calculated according to standard formulae, taking the definite diagnosis of NP as the reference.

Sixty-nine patients were included over a one-year period. The clinical suspicion of NP was confirmed in 44 cases (63.8%). The kappa statistic was 0.71. The sensitivity, specificity, positive likelihood ratio, and negative likelihood ratio were 77%, 84%, 4.80, and 0.27 for PITA, and 75%, 88%, 6.25, and 0.28 for RM, respectively.

CONCLUSION. PITA may be a reliable alternative to bronchoscopic samplings and blinded PTC in the particular setting of NP in newly mechanically ventilated patients. 

Ventilator-acquired pneumonia (VAP) bears a high morbi-mortality and is a complex diagnosis due to lack of accuracy of chest X-Ray and common infection markers. The clinnical pulmonary infection score (CPIS) has been proposed as an adjuntive tool for the management of VAP. The aim of the study was to assess the relation between CPIS, inflammatory response and bacterial isolates in suspected VAP.

We included 32 ICU patients ventilated for more than 48h with new pulmonary infiltrates and suspected VAP and 12 ventilated controls without infiltrates. CPIS and systemic and local inflammatory response was assessed in plasma and BAL fluid at inclusion and after 3 days (TNF-alpha and soluble receptors, IL1b, IL6, IL8, IL10, CRP and procalcitonine). A standard microbiological work-up was carried out. 

We conducted a prospective study of 44 patients receiving mechanical ventilation and without the coexistence of other infection than VAP. PCT and PCR were measured periodically in serum and bronchoalveolar lavage fluid. VAP was suspected if the patient fullfiled clinical criteria or CPIS was ≥5. VAP was diagnosed when a significant growth of quantitative cultures of bronchoalveolar lavage fluid was achieved. Systemic response was assesed by the presence of SIRS. Severity and organ failure were assesed by APACHE II, APS and SOFA scores.

. VAP was suspected in 21 patients but only confirmed in 9 cases. Serum PCT was significantly increased in the VAP group. The best cutoff value was 2.99 ng/ml; sensitivity 77.8%, specificity 97.1%. The area under the receiver-operating-characteristic curve when PCT was used to differenciate the presence of VAP was 0.87. PCT was superior to serum PCR as a diagnostic tool for VAP. Alveolar PCT and PCR were similar in all groups. Serum PCT showed a poor correlation with the presence or not of SIRS and the results of APACHE II, APS and SOFA scores.

Serum, but not alveolar, PCT seems to be a helpful parameter for the diagnosis of VAP independently of the existence of SIRS or the severity of systemic repercussion. M-BAL showed more number of positive results at controls 1 and 3, otherwise the degree of concordance was higher at baseline and control 2 (Table) . Microbiological isolations were similar with both techniques. We had a low rate of complications with 6 episodes of transient desaturation. CONCLUSION. 1) m-BAL and EA showed similar results, but m-BAL presented more positives microbiological isolations at controls 1st and 3rd (6th y 12th day of MV). We could not find any difference with regard to their value in the diagnosis of VAP, however these are preliminary results with low number of cases. 2) m-BAL is a safety procedure.

The MEDS score is validated in the emergency room (ER) pts suspect of infection. The 2001 ATS criteria define severe CAP and criteria for ICU admission. We compared their accuracy for predicting mortality among a cohort of severe CAP requiring ICU.

All pts admitted from the ER with severe CAP during a 30 months period were prospectively enrolled. Calculation of the MEDS score, the number of minor (Cminor) and major (Cmajor) 2001 ATS criteria, the SAPSII score. ROC curve analysis and Odds Ratios (OR) were used to evaluate the scores' accuracy.

For the 141pts:median and interquartile of MEDS score, Cminor and Cmajor 10 (8-13),3 (2-4) and 2 (1) (2) (3) and discrimination (AUC ROC=0.866) were good on the training set. Discrimination remained good (AUC-ROC:0.783) in the validation set. The observed incidence density by ICU was between 9.7 and 26.1 /1000 VM days but the ratio between predicted and observed incidence density was not different from 1 in all but 1 ICU.

If VAP is used as an outcome measure, the VAP risk score might be a more acurate benchmark tool. An external validation of the score is needed.

Damasceno M P C D 1 , Rodriguez F B 1 , Pillar R B 1 , Confalonieri R P 1 , Souza P S P 1 1 Intensive Care Unit, Hospital de Clínicas de Niterói, Rio de Janeiro, Brazil

The early diagnosis improves the chance of survival with the correct treatment in the adequate time. The aim of this study was to analyze the diagnostic accuracy of clinical pulmonary infection score (CPIS) in patients with VAP.

We evaluated 30 clinical suspicion of VAP during 5 months, in 25 patients admitted in a general ICU. When the senior physicians of the ICU had the VAP's suspicion, CPIS was calculated at day zero and day 3. All these suspicions had bronchoscopic bronchoalveolar lavage (BAL). We considered a CPIS ≥ 6 like a positive indicator of VAP in both evaluations in days 0 and 3, while a positive BAL was considered when we met at least one bacterial species ≥ 104 cfu/ml from BAL fluid.

In the 30 clinical suspicions, we had 25 positive BAL. In this group, we had 17 positive CPIS, while in the 5 negative BAL we had only one. The construction of a 2 x 2 table with columns to positive and negative BAL, and rows to positive CPIS present and absent, gave us a sensitivity of 0.68 and a specificity of 0.80. The positive likelihood ratio of a positive CPIS was 3.4, while the negative likelihood ratio was 0.4. 

It is hard to say that an appropriate CPR is always given for the emergency patients. When a patient passes away in Japan, members of his family are supposed to be with him, so even in the case of out-of-hospital cardiac arrest (OHCA) without witness, medical staffs are obliged to keep CPR until the family arrives, which would take more than 2 hours and causes the extra cost of medical service. The rule to discontinue ACLS is required and past OHCA cases were studied with statistical method.

ACLS were performed routinely to all 129 patients in 2004 according to the AHA guideline. In this study the case with more than two hours recovered heart beat is considered successfully resuscitated, which were 44 patients, and the other 85 were unsuccessful. To predict the possibility of CPR, the logistic regression analysis was employed for the following simple data: sex, age, the supposed interval of cardiac arrest, the supposed cause and some emergency laboratory tests (totally 34 items).

To select the variables to be included in the logistic regression model, all the variables were examined using the Chi squared and t tests. 

The accuracy of this model to predict the short-term outcome of OHCA which includes 4 variables, T, PLT, Log[NH3], and Log[PaO2] was 80% and could be clinically available. Blood ammonia level is considered a stopwatch to measure the interval of cardiac arrest because it is known to be proportional to the period of the circulatory collapse. The platelet is related to the coagulation function and could reflect the change by systemic damage. Using this model, the probability of successful CPR will be estimated and it will avoid the futile CPR effort.

Masnou N 1 , Pont T 2 , Salamero P 2 , Deulofeu R 3 , Gracia R 2 1 Transplant Coordination, Hospital Vall d\'Hebron, 2 Transplant Coordination, Hospital Vall d'Hebron, 3 OCATT, OCATT, Barcelona, Spain

In the year 2001, a follow-up protocol for neuro-intensive care patients was designed. The objective was to analyse their detection and outcome between 2001-2005, to define the donor profile, discover the time utilised in the process and suggest possible improvements.

Prospective study of all patients admitted with GSC£ 8 in any critical unit of this 1200 beds University hospital. A patient archive was designed with: demographic and medical data, and times. Definitions:"first time":admittance-detection,"second time":detection-outcome; sub-divided in:diagnosis of brain death (BD), cardiac arrest, discharge in vegetative state (PVS) or in neurological improvement. SPSS. 

Our "typical donor" is a 45-year-old male, with a HIC and initial GCS 6, detected in the first 24-36h and whose donation takes place within the following 72h. provides an assessment of cardiac output, cardiac contractile function (cardiac function index) and pulmonary oedema (extravascular lung water). We studied the frequency of these estimates using the PiCCO system in SAH patients. We also determined if elevated cardiac troponin I (cTnI) adversely affects these estimates of cardiopulmonary function.

Twenty-eight patients with aneurysmal subarachnoid haemorrhage (SAH) with a poor clinical condition or large amounts of extravasated blood. We assessed reduced mean cardiac output (CI < 3 l/min/ m2), mean cardiac contractile function (CFI< 4.0 l/min) and elevated mean extravascular lung water (EVLWI > 10 ml/kg) every eight hours in the first 5 days after SAH.

A reduced cardiac output during at least one day, was seen in 6 (21%) patients, a depressed cardiac contractile function and an elevated extravascular lung water both in 10 (36%) patients. An elevated cTnI tended to increase the risk for reduced cardiac output (OR 6; 95%CI: 0.5 -310) and was statistically significant related to decreased cardiac contractile function (OR 18; 95%CI:1.6 -850).

Cardiopulmonary complications were frequently observed in patients with severe SAH. An elevated cTnI appeared to be a good marker for the occurrence of a decreased cardiac contractile function. Transpulmonary thermodilution technique seems to be a useful tool for monitoring SAH patients who are at risk of cardiopulmonary abnormalities, especially those with elevated cTnI on admission.

Foundation (grant D98.014).

Spatenkova V 1 , Kazda A 2 , Skrabalek P 3 , Kralova D 4 , Suchomel P 5 1 Neurocentre ICU, Regional Hospital, Liberec, 2 Department of Clinical Biochemistry, Postgraduate Medical School, 1st Faculty of Medicine, Charles University, Prague, 3 Department of Clinical Biochemistry, Regional Hospital, Liberec, 4 Centre of Biostatistics and Analyses, Masaryk University, Brno, 5 Neurocentre, Department of Neurosurgery, Regional Hospital, Liberec, Czech Republic

Polyuria is very frequent in the neurointensive care unit. It can be caused by osmotic diuresis or water diuresis in central diabetes insipidus, both of which could lead to serum sodium dysbalance such as hyponatraemia or hypernatraemia. The aim of our study was to analyse diuresis over 4000 ml in patients admitted to the neurointensive care unit.

We retrospectively evaluated 789 days of polyuria in 318 patients (mean age 68, from 17 to 82 years, 179 male, 139 female) out of 1605 patients hospitalised in the neurointensive care unit over a period of three years. There were 289 patients with diseases of the central nervous system (CNS), (cerebral 266 and extracecebral 23) and 29 patients with spinal diseases. 243 patients underwent operations. 168 patients received osmotic agents, 15 patients received diuretic agents and 16 patients were given a combination of osmotic and diuretic agents.

The mean value of all 789 polyuria was 4535 ml (from 4050 to 10400 ml). They lasted from 1 to 15 days (mean 2.5 ± 2.4 days). The mean value of the highest diuresis of each patient was 5192 ml. There were no significant differences in these diuresis either in patients with CNS or with spinal diseases (p=0.229); neither were there differences between those with cerebral and extracerebral diseases (p=0.933). No effect was found between patients who had operations and those who did not (p=0.969). Significant differences were found in diuresis in patients with cerebral complications (p=0.001), who also had significantly longer periods of diuresis (p<001). There were 680 days of polyuria in 261 patients (82%) without serum sodium dysbalance and 109 days of polyuria in 57 patients (18%) with serum sodium dysbalance. No significant differences were found in the mean values of diuresis between patients with and without serum sodium dysbalance (p=0.189).

The results of our study show that polyuria in neurointensive care unit is usually unaccompanied by serum sodium dysbalance for most patients. Due to the frequency of osmotherapy, the majority of polyuria were caused by osmotic diuresis. 

We have analyzed 180 neurosurgical cases in retrospect. The history of the gastrointestinal pathology within one year prior neurosurgery was the exclusion criteria. The CNS pathology distribution: 68 pts with traumatic brain injury (TBI), 27 pts with fossa posterior tumors, 52 pts with tumors of chiasmal region, 33 pts with tumors of hemispheres and basal ganglia region.

The following main types of GI pathology were marked out. 1. Ulcerative lesions. 2. Erosions. 3. Case "1" and/or "2" accompanied by bleeding. It was verified that in case of the focal lesions of the diencephalon, brainstem and craniospinal regions the usual disposition of the erosion and ulcer is the pylorus and the bulb of the duodenum. The focal lesions of the hemispheres and basal ganglia regions commonly led to the evolution of the GI pathology in stomach curvatures. The 31% of patients with TBI had diffusive erosions of the mucous layer of the upper GI tract.

CONCLUSION. 1. The main types of the GI pathology in neurosurgical pts are the erosions and ulcers.

2. The location of the GI pathology is usually determined by the topography of the CNS lesion.

Grant acknowledgement. We wish to thank Dr. V. K. Emeliyanov for his inexhaustible support.

Ledoux D 1 , Piret S 1 , Damas P 1 , Laureys S 2 1 Intensive Care, 2 Neurology, Liège University Hospital, Liège, Belgium

Vegetative state (VS) is generally considered as a late poor neurological condition. However vegetative state is a disorder of consciousness that can be acute and reversible. The aim of the present study was to provide a description of VS at the acute stage and to examine the outcome of these patients.

We analysed data collected prospectively on all consecutive admissions over a 5 years period in a 26 beds intensive care units (ICU) at the Liege university hospital. During that period of time the best Glasgow Coma Score (GCS) was recorded daily. We defined patient with a GCS < 15 during the first 24-hour of ICU stay as having an impaired consciousness on admission. Among these patients VS was defined as: eye opening (spontaneously, to load voice or on noxious stimulation), no verbal response or incomprehensible sounds or groaning or not assessable/ventilated and no motor response or stereotyped response or normal flexion. We looked at ICU outcome and hospital survival separating traumatic from non-traumatic brain injuries. CONCLUSION. Acute VS is far from being a rare diagnosis in the ICU. The prognosis of patients who were or transit through vegetative depends greatly on the nature of the brain injury. However even in non-traumatic brain injury, an important proportion of patients experienced favourable outcome. Therefore, great caution should be taken before considering end of life limitation in these patients.

Koukoulitsios G V 1 , Stathopoulos G T 2 , Mandila C G 1 , Papakonstandinou K 2 , Karabinis A 1 1 ICU, General Hospital of Athens G Gennimatas, 2 ICU, General Hospital of Athens G Gennimatas, Athens, Greece

Hypothermia is a common occurrence in burn patients, especially when anaesthetized and mechanically ventilated. Unfortunately, the available therapeutic approaches for burnassociated hypothermia are limited, and extensive surface burns often limit the use of classic external and internal warming techniques. The use of novel intravascular cooling devices in order to warm hypothermic burn patients has not been evaluated.

We used an intravascular catheter designed to lower the body temperature (Cool Line, Alsius Corporation, Irvine, CA, USA) in order to achieve the opposite: to warm a hypothermic burn victim. We treated a 28-year-old man who suffered a full-thickness skin burn involving 45% of his body surface area. He was sedated, mechanically ventilated, and admitted to the Intensive Care Unit due to co-existing inhalation burn. During admission he had moderate hypothermia with a rectal temperature of 30.8 • C, was hypotensive, had low urine output, and exhibited excessive bradycardia alternating with flairs of atrial fibrilation.

Despite active external re-warming using air-conditioning, a heating blanket, and warm crystalloid infusions, the patient remained hypothermic and in cardiovascular instability. We then employed the endovascular cooling device to combat hypothermia and maintain normothermia. This system circulates temperature-controlled sterile saline through two small balloons mounted on the distal end of the catheter. The patient's blood is gently warmed as it is passed over the balloons. The system had been set to a target temperature of 36.5 • C. The catheter-controlled re-warmig process took six hours, after which time the patient's core body temperature reached and was maintained at 36.5

• C. With normothermia, the patient's cardiac rhythm, urine output, and arterial blood pressure returned to normal.

Burn patients are extremely prone to hypothermia associated with hemodynamic instability and impaired perfusion. Intravascular warming can be considered as a possible alternative to classic methods of external or internal re-warming.

Stensballe J 1 , Schiødt F V 2 , Lippert F K 1 , Rasmussen L S 1 , Dahl B 3 1 Anaesthesia, Centre of Head and Orthopaedics, 2 Hepatology, 3 Orthopaedic Surgery, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark

Gc-globulin is a multifunctional plasma protein involved in the extracellular actin scavenger system responsible for the removal of circulating actin released from necrotic cells. No study so far, has described the time course of actin-free Gc-globulin concentrations in the first 24 h after injury. We hypothesized that low levels of actin-free Gc-glubulin were associated with severity of the injury.

Two hundred consecutive adult trauma patients admitted a Level 1 trauma centre were included in this prospective study. All patients had plasma samples taken on admission and after 6, 12, and 24 h. The level of actin-free Gc-globulin was measured using the Gc-globulin (Actin-free) ELISA kit (AntibodyShop A/S, Gentofte, Denmark). Data were analyzed using ANOVA (mixed effect model, repeated measures) and the Bonferroni test. The local Ethics Committee approved the study.

Eighty patients (40%) had major injury (Injury Severity Score (ISS) > 15) with a significantly higher 30-day mortality than in patients with minor injury (ISS<=15) (27 (33.8%) vs. 1 (0.8%); p<0.0001, Fisher's exact test). The plasma concentration of Actin-free Gc-globulin was significantly lower in patients with major injury, table 1. This difference was present at all time points. 

Plasma concentration of actin-free Gc-globulin is reduced in trauma patients with major injury throughout the first 24 hours. The greatest difference in actin-free Gc-globulin between patients with minor and major injury is seen 6 h after admission. This finding supports the theory that measurement of actin-free Gc-globulin is a prognostic marker already on admission, and that it may be a useful marker in trauma assessment. 

Atrial fibrillation (AF) remains a common arrhythmia. Electrical cardioversion is commonly employed in its management. External cardioversion is a short but painful procedure with stimulus intensity analogous of a surgical incision. Adequate depths of sedation are important to prevent recall of an unpleasant experience and to attenuate the catecholamines surge of the stress response. We describe the use of different sedatives drugs, their side effects, and the presence of awareness and the quality of awakening of the anaesthesia.

With approval of the ethics committee and written informed consent, 50 patients ASA II-III physical status, age 37-75 who were scheduled to undergo electrical cardioversion were enrolled twenty five per group in the study. Demographic data were comparable in both groups. Cardioversion was performed during a morning session on the CCU. Group 1 received initially 2.5 mg midazolam (MID) and 50 mcg fentanyl (FED) were given intravenously, there was a two minute wait before a further 1 mg MID and 50 mcg FED bolus were given and repeated at 1-2 minute intervals till the total dose of MID reached 5 mg. Group 2 received initially as a premedication 50 mg ranitidine, 10 mg metoclopramide, 4 mg ondansetron and then 2.5 mg MID, 50 mcg FED, there was a two minute wait before a further 50 mcg FED and 50 mg of propofol were given. The patients were considered effectively sedated when they appeared sleepy and no longer continued a conversation and there was no response to soft verbal commands and mild tactile stimuli, then the shock was performed.

Fisher's exact test was used for statistical analysis. P<0.05 was considered statistically significant. No patient required intubation in both groups. Patients needed respiratory support for a little while, were in group 1 one patient and for group 2 six patients needed respiratory support. Both groups received an adequate sedation. There was no difference between the groups regarding awareness and respiratory depression on the other hand there was statistically significant difference regarding nausea (P<0.05). The p-value for any of the events between the 2 groups is 0.538. CONCLUSION. We think that the use of medication such as H2receptors, 5HT3 are useful and also the sedation with the adjuvant of a classic sedative as propofol is a safe alternative and limits the side effects (as nausea) of the opioids.

Grant acknowledgement. to the support and service of the ICU staff.

Cotogni P 1 , Antonucci A 2 , Becchio D 2 , Ranieri V M 1 1 Anesthesiology and Intensive Care, S. Giovanni Battista-Molinette Hospital, University of Turin, 2 School of Nursing, University of Turin, Turin, Italy

Victims of cardiac arrest need immediate cardiopulmonary resuscitation (CPR). The teaching of CPR is an important part of a continuing medical education program. In 1997, our 1400-bed tertiary teaching hospital introduced a 8-hr CPR plus automated external defibrillation training (BLSD course), targeted at nurses, physicians, healthcare students, and other allied healthcare professionals. A number of studies shows that learned CPR skills deteriorate over time even among nurses (N) and physicians (P). The purpose of this study was to investigate the retention of theoretical knowledge of CPR in our hospital trained N&P.

, 122 BLSD courses with 1894 participants were evaluated. During this period, 996 N&P employed in our hospital have successfully completed the course and were included in the study. A total of 460 N(76%) and P(24%) were each fully interviewed with regards to their knowledge of CPR. 536 N&P were missed to the survey due to several causes (transfer to other hospitals, retirement, illness, pregnancy). The period of time since last trained in BLSD ranged from 6-48 months. Data were collected by a structured interview during a phone call.

More than 50% N&P who had been updated in BLSD up to 18 months prior to the interview were not able to answer correctly half the questions asked. Moreover, more than 50% N&P who had been updated in BLSD up to 12 months referred to feel no longer confident with practical CPR skills. No differences in CPR knowledge retention between N&P were found. On the contrary, a significant difference (p<.001) was shown between N&P employed in critical care areas or in ambulance staff and in not monitored wards. No difference related to gender was observed. Interestingly, 44% N&P referred to be directly involved in almost one in-hospital delivery of CPR following BLSD course participation. Moreover, 61% N&P undertook all four steps of the "chain of survival" (activation of the emergency medical team, CPR plus defibrillation, and advanced life support).

CONCLUSION. Data collected clearly show a time-dependent course in resuscitation content decay. Moreover, these data confirm that in-hospital cardiac arrest can occur anywhere. The results of this study underline the need for an in-hospital planned retraining program, as well as additional studies of the optimal timeframe for BLSD refresher course to facilitate N&P retention of CPR knowledge over time. 

In a previous study, the author had evaluated the efficacy of low-dose ketamine (0.25mg/kg) in the treatment of intra-operative shivering associated with sub-arachnoid blockade. In the following study the author sought to evaluate the efficacy of low-dose (0.25mg/kg) ketamine in the treatment of post-operative (upto 72 hours) shivering in neurosurgical patients in a neuro intensive care unit. Shivering would have deliterious effects in this setting & ketamine being a NMDA antagonist, would proably have neuro-protective effects.

METHODS. 60 post-operative neurosurgical patients having undergone craniotomy with a postoperative GCS > 10 were enrolled. All patients who shivered, during a 72 hour postoperative period received either low dose ketamine or pethidine (25 mg/60 kg) in a randomised fashion. Febrile episodes were treated with tepid sponging (< 100 F)& I/M paracetamol 500mg(>100 F). The time taken to cessation of shivering was noted in each group. Also, the respiratory rate, pulse, temperature, NIBP & SpO2 ware monitored in each group. 

Bubnova I D 1 , Astakhov A 1 , Dobrynin I 1 , Busygin S 1 1 Department of anaesthesiology, reanimatology and neuroreanimatology, Urals medical academy of postgraduate education, Chelyabinsk, Russian Federation

It is known that the outcome of sepsis depends on haemodynamic (H) stability, and we always try to support the hyperdynamic type of circulation especially in acute period. This study we planed to determine the role of central regulation (CR) of H on outcome in severe sepsis.

METHODS. 42 patients with severe sepsis were studied in ICU. Most of them had different signs of encephalopathy and were sedated. In our previous works we showed the influence of different regions of brain activity on variability of EEG amplitude and main haemodynamic parameters (HP). So, all patients with sepsis were examined by complex method of synchronic registration of EEG amplitude (AEEG) and HP, such as stroke volume (SV), cardiac output (CO), cardiac index (CI), heart rate (HR), blood pressure (BP), and pulsatory amplitude of peripheral vessels (PAPV). Besides we evaluate the variability (V) of all parameters in band 0-0.5 Hz and its four ranges -UVLF (0 -0.05 Hz -metabolic influences), VLF (0, 05 -0.075 Hz -hormonal activity), LF (0.075 -0.15 Hz -baroregulation), HF (0.15 -0.5 Hz -parasympathetic or volume regulation). The rise of V in AEEG and HP we estimated as the result of adaptive brain activity, the fall of V in all ranges -as the result of impairment of CR.

All patients were divided in two groups: A (CI >2.5) and B (CI < 2.5). In group A the patients had the high SV (75±5.4), CO (7.6±0.6) and BP (130.4±6.3); In group B these parameters were much lower: SV (27.6±2.5), CO (2.7±0.2), BP (113±5.5). In group B the amplitude of EEG was also lower (4.7±1.4), then in group A (8.3±0.5). The V in group A was high in UVLF and VLF for AEEG, PAPV and in LF, HF for SV, CO, BP, HR. More often it is the result of adaptive reaction and hypothalamic activity. In group B the V was very low in all ranges of AEEG and HP.

In patients with severe sepsis the adaptive hypothalamic activity supports the needful level of volume circulation. The fall of V of AEEG and HP may be used as a marker of poor prognosis. 

A prospective 1-year study was carried out, collecting epidemiological, clinical management and results data of 370 TBI patients that were admitted to 7 Catalan ICUs.

Male patients were predominant (3:1), with a mean age of 40. The severity profile (measured by GCS score) identified 53% severe, 27% moderate and 20% mild TBI patients. According to Marshall radiological classification of intracranial injury, type II diffuse injury (39%) was more prevalent. 49% of cases presented tHSA. Utilization of neuromonitoring devices was PIC 52% (69% in severe TBI), SjO2 27% (37% in severe TBI), TCD 50% (64% in severe TBI), cerebral perfusion pressure 59% in severe TBI. During the first 15 days of stay, a mean of 2 CT scan per patient were made. Therapeutic measures were also reviewed, and manitol appeared as the main osmotic measure for intracranial hypertension control. Corticosteroids were used in 15% of severe TBI patients. More than 25% of cases received vasoactive drugs support, predominantly noradrenaline. 25% of patients received prophylactic anticonvulsants. Barbiturates were prescribed in <10% of patients. Hyperventilation was used in 80% of patients. A neurosurgical intervention was practised in 24% of patients (21% in severe, 29% in moderate and 22% in mild TBI patients). The most prevalent secondary insult during ICU length of stay was fever and hypotension, specially in severe patients. Global mortality rate was 22% (32% in severe, 13% in moderate and 7% in mild TBI patients). Independent variables associated with a greater mortality were: female, >60 years old, admission GCS, anisocoria and bilateral mydriasis, encephalic type III/IV injury, evacuated or non-evacuated mass, Fisher type-III tHSA, and mean LOS <4 days or 5-12 days.

Clinical results profile and prognostic factors obtained were in accordance with current medical literature results. Although monitorization levels were greater than levels observed in published inquiries, there is little deviation from published clinical guidelines in the management of TBI patients in catalan ICUs.

Piazza O 1 , Scarpati G 1 , Romano I 1 , Mariniello I 1 , Tufano R 1 1 Anestesiologia e Rianimazione, Università degli studi di napoli Federico II, Napoli, Italy

Stroke patients undergo alterations of the thrombotic and fibrinolytic systems with formation of thrombin antithrombin III complex (TAT) and consume of antithrombin III (ATIII). To evaluate the difference between focal and global ischemia we examined sequential changes in ATIII, TAT and coagulatory markers in ICU patients.

28 patients were included in this observational study (15 ischemic stroke-focal stroke-; 13 survivors after cardiac arrest-global stroke-). The plasma level of TAT, D-dimer, fibrin degradation products (FDP) and percent activity of ATIII were measured at the 1st and the 5th day after the stroke onset.

In all patients TAT was significantly elevated at all time points, and this elevated levels were associated with increased D-dimer and FDP levels. ATIII was not decreased in ischemic patients on the 1st and the 5th day after the stroke onset. CONCLUSION. AT III is the most important physiological inhibitor of blood coagulation as it interferes with the clotting process at various levels; it might play an active role in the pathogenesis and in the evolution of cerebral ischemia. The present study demostrated the same alterations in thrombotic and fibrinolytic markers in each subtype of stroke. 

Renal failure in the ICU is common but long term dialysis dependency is rare (1). One proposed mechanism is tubular failure through mitochondrial damage (2) . Calcineurin inhibition (CI) has been shown to ameliorate mitochondrial structural damage caused by acute sepsis in an animal model (3) but is also associated with renal failure long term. We reviewed our rates of renal replacement therapy (RRT) following ICU discharge in 3 groups (sepsis, liver transplant [Tx] and paracetamol overdose [POD] ) over a 20 month period. Liver Tx patients all received immunosuppression with CI.

To our knowledge, rates of RRT following liver Tx have not been reported.

We identified patients admitted to our ICU who survived to hospital discharge from Dec 03 to Aug 05. This was cross-correlated with the renal department database to identify those currently haemodialysed (HD) or on CAPD and the liver department database for Tx and POD patients. Previous renal impairment was defined as serum creatinine >.126mmol/L (the ULN in our hospital) within 6 months before admission.

. 295 required RRT during their stay; 130 survived to hospital discharge. We excluded patients already receiving HD. There were no results for 9 patients. The study group size was 77; 46 were admitted with sepsis, 21 received Tx and 10 were treated for POD. In the septic group only 1 patient with previously normal renal function required RRT, compared with 4 septic patients with previous renal impairment. In the Tx group, 10 patients required temporary renal support following ICU discharge. However, of the 13 with pre-existing renal impairment, only 1 patient remains on RRT. None of the 8 with normal renal function required RRT. None of the POD group required RRT. Rates of recovery were similar in all groups (p=0.38, Fisher's exact). The median time to recovery was 8 days for the septic group and 15 days for the liver Tx group (p=0.07, t-test). 

Oral N-acetylcysteine (NAC) has been shown to reduce the incidence of acute renal dysfunction by 85% after iodinated contrast injection in patients at risk for contrast media nephropathy (1) . The aim of this study was to assess the incidence of contrast media nephropathy in the ICU and evaluate the usefulness of NAC administration in its prevention.

This study was a prospective cohort comparing the incidence of contrast nephropathy [increase in serum creatinine level>=0.5 mg/dl (44 µmol/l) or the need for dialysis within 48h] in two teaching hospital medical ICUs: in one oral NAC was always used while it was never used in the other; saline hydration and low osmolality contrast media were always used. All consecutive dialysis free patients needing iodinated contrast media for computed tomography or angiography with available serum creatinine before and within 48 h after examination were included. Risk factors for contrast nephropathy, severity scores, hemodynamics, hydration, volume of contrast media, serum creatinine and urea, urine output were prospectively recorded at contrast examination. Renal function parameters were recorded within 48h after examination.

. 55 patients underwent 70 contrast examinations in the "no-NAC ICU" and 63 underwent 70 contrast examinations in the "NAC-ICU". Patients were similar in both ICUs for contrast nephropathy risk factors, severity, baseline renal function, hydration, and volume of contrast media administered. The incidence of contrast induced nephropathy was 17% (12/70) in the no-NAC ICU and 10% (7/70) in the NAC-ICU, p=0.21. The number of patient with a 25% increase in serum creatinine (18% vs 11%, p=0.23) or needing dialysis (6% vs 3%, p=0.98) within 48h were also similar.

The benefit of NAC in preventing nephropathy in the ICU, if any, was considerably smaller than previously reported. We could not demonstrate such a benefit. 

Replacement of renal function in critical patients is complex. Continuous haemodialysis seems better than intermittent haemodialysis, but there are few studies that showed significant difference in this process in overall mortality. We will present our experience with Sustained Low Efficiency Dialysis (SLED), analyse tolerance and complications of this technique.

A prospective study with collection of demographic and epidemiological data of all patients admitted to our intensive care unit (UCI) from January until December of 2005. We analysed all patients who needed renal replacement, focusing on the process of renal replacement therapy. It was used a conventional dialysis machine, using low blood and dialysate flow rate for prolonged time.

We analysed 135 SLED sessions in 17 patients (57.5±17 years; male: female, 11:6). Initial SAPS II and SOFA was 53.9±15 and 11.3±4 respectively. From all patients, 15 had medical illness, with 12 having sepsis. Six had chronic renal failure previous to admission and 2 were on regular dialysis. Thirteen patients were oliguric on admission and the mean initial creatitine was 3.1±1mg/dl and urea of 116.5±51mg/dl. The mean duration for each session was 9.9±2 hours and with a mean ultrafiltration of 2304±895ml. We used fraction LMWH in 81 sessions to prevent clotting. We had 7 interruptions, two for hemodynamic instability and five for blood clotting in the extracorporeal circuit.

In 41 sessions patients were with amine infusion, with need of readjustment in 30 (19 session was necessary increasing dose). One patient recovered complete renal function, 3 patients maintain renal failure but without need for renal replacement therapy, other 3 needed regular renal replacement and 10 died while on ICU. The mean SOFA at discharge from ICU was 8.8±6. 

In critically ill patients treated with Ampho-B by continuous infusion, mortality and need for CVVH are high. However, in two thirds CVVH is already initiated before the start of Ampho-B due to co-morbid disease. A cumulative dose of more than 500 mg by continuous infusion appears not to be associated with a higher incidence of renal insufficiency at discharge compared to the lower dose. Renal function recovered in nearly all survivors. 

De laet I 1 , De Waele J J 1 , Blot S 1 , Decruyenaere J 1 , Oeyen S 1 , Colpaert K 1 , Nollet J 1 , Roosens C 1 , Hoste E A 1 1 Intensive Care Unit, Ghent University Hospital, Gent, Belgium

is associated with mortality. Early intervention to prevent development of AKI-RIFLE may therefore improve outcome. AKI-RIFLE can only be detected after a 6-hour period of oliguria or an increase of serum creatinine with 50%. The aim of the study was to evaluate whether a shorter period of oliguria, i.e. a 2 hour period of UO<0.5 mL/kg/hr (U2), is predictive for development of AKI-RIFLE.

All patients admitted to the 28 bed adult surgical ICU of the Ghent University Hospital, during a 35 d period were prospectively evaluated for occurrence of U2 and AKI-RIFLE.

. 121 patients were included in the study. Median age was 58 yr (interquartile range (IQR) 40.70), and 63% were male. Length of stay in the ICU was 3 d (IQR 2.5), and ICU mortality was 8.3%. AKI-RIFLE occurred in 51 patients (42.1%). AKI-RIFLE was already present on admission in 8 patients (6.6%). One or more episodes of U2 occurred in 80 patients who did not have AKI-RIFLE at admission (70.8%). In 37 patients (46.3%), U2 was not followed within a 48 hour period by development of AKI-RIFLE. On the other hand, 43 patients (53.8%) with U2 did develop AKI-RIFLE within a 48 hour time period. None of the patients developed new AKI-RIFLE without U2. Maximum severity of U2 patients who developed AKI-RIFLE was Risk in 26 patients (32.5%), Injury in 16 patients (20%), and Failure in 1 patient (1.3%).

CONCLUSION. AKI defined by the RIFLE classification occurred in almost half of the patients. A 2-hour period of oliguria was a frequent finding in this cohort of surgical ICU patients, and was followed by development of AKI-RIFLE in half of the patients within a 48 hour time period. A 2-hour period of oliguria was therefore a sensitive tool for detection of development of AKI-RIFLE, and may therefore be an opportunity for early intervention therapy.

Vinclair M 1 , Broux C 1 , Chabre O 2 , Faure P 3 , Brun J 1 , Jacquot C 1 , Payen J F 1 1 Anesthesiology -intensive care unit, 2 Endocrinology, 3 Biology, University hospital, Grenoble, France

The implication of etomidate in severely-ill patients was recently pointed out as a risk factor for subsequent adrenal insufficiency (AI) in ICU patients. We investigated the prevalence and the duration of the AI following a bolus of etomidate required for endotracheal intubation by measuring serial blood concentrations of cortisol and its substrate, the 11-desoxycortisol.

With approval of the local ethical committee and informed consent, 43 critically ill patients were included (28 M/15 F, 46±19 years, SAPS II 39±14, SOFA score 6±4). The blood content of 11-desoxycortisol (nmol/l) and cortisol (µg/dl) was measured before (T0) and after (T60 min) corticotropin stimulation tests (250 µg) realized at H12, H24, H48 and H72 following the injection of etomidate. AI was defined as a cortisol response (Delta F) less than 9 µg/dl. Normal blood content of 11-desoxycortisol was determined from measurements of 15 critically ill patients having received another sedative drug for tracheal intubation. Data (mean±SD) were analysed using ANOVA.

. AI was found in 84% of the etomidate population at H12, 48% at H24, 21% at H48, 24% at H72. Evolution of Delta F is represented in the figure (* p<0.01 H12 vs H48, § p<0.01 H24 vs H48). The blood content of 11-desoxycortisol was initially high and decreased to control values at H72 (p<0.01 at H12 and H24 vs. H72).

A bolus of etomidate inhibits the 11-β-hydroxylase during 48h, resulting in a high incidence of AI. If etomidate is required, a corticosteroid supplementation could be thus indicated for 48h. 

The aim of this study was to assess whether outcome prediction can be improved in critically ill patients with hematologic malignancies by taking into account the evolution of organ failures during the first 3 days of ICU stay and the admission diagnosis.

All consecutive critically ill patients with hematologic malignancies admitted to the ICU between 2000 and 2006 were included in this study. SOFA scores on day 1-3 and a cancer specific severity of illness score (CSSIS) [1] were calculated prospectively. The admission diagnosis was assessed by an independent panel of physicians blinded to the patients outcome who categorized the patients according the presence or absence, the diagnostic certainty and the site of bacterial infection. The performance of the SOFA on d 1 and two logistic regression models consisting of the SOFA on d 1 adjusted for Delta SOFA (d 3 -d 1) and additionally adjusted for the admission diagnosis respectively, were compared with the CSSIS.

Over the study period 300 patients were admitted in the ICU and 205 were still in the ICU on day 3. ICU, hospital and 6 months mortality was 40%, 55% and 64%, respectively. The expected hospital mortality according to the CSSIS was 71 ± 26%. Mean SOFA scores on d 1 and 3, and Delta SOFA were 9.0 ± 4.0, 8.6 ± 3.9 and -0.5 ± 3.3, respectively. Patients admitted because of documented or clinically suspected bacterial infection had a higher SOFA on d 1 (9.7 ± 4.0 vs. 8.4 ± 4.0 P=0.008) but a more rapidly reversible organ failure (Delta SOFA -1.1 ± 3.1 vs. 0.0 ± 3.4, P=0.013) and a lower hospital mortality (44% vs. 64%, P=0.001) than those without. The area under the ROC ± SE and calibration statistics (chi-square, P-value) for the CSSIS, the SOFA d 1 and the Delta adjusted SOFA were 0.73 ± 0.03 and 11.1, P=0.20; 0.60 ± 0.04 and (NA); and, 0.78 ± 0.03 and 10.4, P=0.24, respectively. The performance further improved by adjusting for the admission diagnosis: 0.83 ± 0.03 and 6.6, P=0.58. Hospital mortality in patients with and without bacterial infection precipitating ICU admission was 22% vs. 49% (P=0.002) in patients with a SOFA < 8 (median) and 65% vs. 93% (P=0.002) in those with a SOFA > 8, respectively.

Outcome prediction in critically ill patients with hematologic malignancies improves by taking into account the evolution of organ failures during the first 3 days of ICU stay and the admission diagnosis. Provided that underlying hematologic illness is not rapidly fatal, it is justified to offer a 3 days therapeutic ICU trial to any of those critically ill patients to maximize chances for survival. Oncol 1998; 16:761-770 .

Claessens Y 1 , Taupin P 2 , Kierzek G 3 , Pourriat J 3 , Baud M 3 , Jais J 2 , Riou B 4 , Dhainaut J 1 , Landais P 2 1 Emergency Medicine, Cochin Hospital, 2 Biostatistics, Necker Hospital, 3 Emergency Medicine, Hôtel-Dieu, 4 Emergency Medicine, Pitié-Salpétrière, Paris, France

To develop a day-to-day composite index to detect the heat related-over mortality in Paris based on the profile of patients referred to ED.

Retrospective study from records of 99.976 adult patients (May 1st -September 30th) over the years 2001, 2002, 2003 . Demographics, social, triage and physical parameters; statistical analysis of the criteria influencing admission during the heat wave period and the development of a composite index to predict mortality in the population exposed to the heat wave.

• C, admission after ED-visits could predict the occurence of heat wave-related pre-hospital over-mortality (black line).

A composite index developed according to the profile of patients admitted in the ED, appeared suitable for alerting on the overall mortality in the Paris area submitted to heat wave.

Stachon A 1 , Holland-Letz T 2 , Kempf R 1 , Becker A 1 , Friese J 3 , Krieg M 1 1 Inst. of Clinical Chemistry, Transfusion, and Laboratory Medicine, University Hospital Bergmannsheil, 2 Department of Medical Informatics, Biometry, and Epidemiology, Ruhr-University Bochum, 3 Department of Surgery, University Hospital Bergmannsheil, Bochum, Germany INTRODUCTION. The appearance of nucleated red blood cells (NRBC) in the peripheral blood is associated with a variety of severe diseases. When NRBC are detected in the blood this is generally associated with increased mortality.

In a prospective study the detection of NRBC was analyzed with regard to the clinical circumstances and the course of laboratory parameters of organ injury. NRBC were daily measured in the peripheral blood of surgical intensive care patients (n=284).

NRBC were found at least once in 32.0% of all patients. The mortality of NRBC-positive patients was 44.0% (40/91); this was significantly higher (P<0.001) than the mortality of NRBCnegative patients (4. 2%, 8/193) . With regard to intensive care mortality, NRBC in blood showed sensitivity and specificity of 83.3% and 78.9%, respectively. The area under curve (C-statistic) was 0.86. Mortality increased with increasing NRBC concentration. All patients with more than 2000 NRBC /µl died. Moreover, mortality increased with increasing frequency of occurrence. When after first detection of NRBC in blood, during the further course of intensive care treatment the NRBC have disappeared from the circulation, the mortality again decreased to values of NRBC-negative patients. NRBC were detected for the first time, on average, 9 days (median 5 days) before death. Multiple logistic regression analysis under consideration of several other clinical and laboratory risk indicators revealed a significant association between NRBC and increased mortality, the odds ratio being 1.95 (95% confidence interval 1.35-2.82; P<0.001) for each increase in the NRBC category (0/µl; 1-40/µl; 41-80/µl; 81-240/µl, >240/µl). After the initial detection of NRBC in blood there were no significant increases regarding the creatinine concentration and the alanine aminotransferase activity, respectively. However, the appearance of NRBC coincided with increasing C-reactive protein and thrombocyte concentrations, respectively.

CONCLUSION. The detection of NRBC in blood of surgical intensive care patients is of prognostic power with regard to the patients' mortality. This prognostic significance of NRBC was independent of some clinical and other laboratory risk parameters. The appearance of NRBC in blood was not associated with a failure/lesion of the kidney and the liver, respectively. Grant acknowledgement. This study was granted by the Foundation for Pathobiochemistry and Molecular Diagnostics of the German United Society for Clinical Chemistry and Laboratory Medicine and Sysmex Europe GmbH, Germany.

Meyfroidt G 1 , Güiza F 2 , Fierens D 2 , Ramon J 2 , Van den Berghe G 1 1 Intensieve geneeskunde, UZ Gasthuisberg, 2 Computer sciences, Catholic University of Leuven, Leuven, Belgium

A randomised controlled trial on intensive insulin therapy (IIT) in a medical ICU (1) showed a mortality reduction among patients remaining in the ICU for at least a third day. Unfortunately, predicting LOS in the ICU is virtually impossible on a clinical basis. Machine learning techniques and probabilistic graphical models are statistical methods that can integrate biomedical data and clinical background knowledge (2) . The aim of this study was to evaluate them to build classification models for predicting a LOS ≥ 3 days.

We examined a database of 1548 patients from a randomised controlled trial on IIT in a surgical ICU (3). In this study, 630 patients stayed ≥3 days in ICU. A selection of clinical and lab data and the APACHE II from the 1st day in ICU, together with medical history and demographic data, were used as input variables. Data were analysed using 4 techniques: Decision Tree learning (DT), Random Forests (RF), Naïve Bayesian (NB) and Tree-Augmented Naïve Bayesian (TAN) networks. We ran performance tests using 10-fold cross-validation and averaging the measures. Discrimination was assessed by the area under the Receiver Operator Characteristic curve (aROC). Calibration (goodness-of-fit) was assessed by Hosmer-Lemeshow H statistic (p≥0.05).

Performances are summarized in table 1. NB and TAN were more discriminative than DT and RF models. TAN had a poor calibration. 

Data mining could predict a LOS ≥3 days, in a database that was not designed for clinical predictions. The learned models from this experiment may not be generalised. The cut-off discrimination (aROC) for predictions upon which a clinical decision is based (such as starting IIT) remains arbitrary. The presented data mining techniques will need prospective validation in a clinical setting. 

Severity of illness scoring systems are widely used in intensive care practice. However, their use in patients with sepsis are largely bee limited to means of stratification in clinical trials. As newer sepsis therapies become available, it may be possible to use such systems for refining their indications and monitoring their utilization. The aim is to evaluate the ability of severity of illness scoring systems (APACHE II, APACHE III, SAPS II) and organ dysfunction scores to predict hospital mortality in adult intensive care patients with sepsis.

A prospective observational cohort study was performed over 6 months in a medical ICU of a tertiary hospital. All consecutive patients with definite diagnosis of sepsis admitted to the ICU were included in the present study. Discrimination was assessed by area under ROC curves and calibration was done using Hosmer-Lemeshow goodness of fit test, calibration curves and observed / predicted mortality correlations.

A total of 88 patients were included. The observed hospital mortality was 50.5%. Results of calibration and discrimination are shown in table.

No statistically significant differences in accuracy of prognosis prediction were identified for the scales assessed. 

Dawson L 1 , Tangkau P L 1 , Salm E F 1 , Sleeswijk Visser S J 1 , Meynaar I A 1 1 ICU, Reinier de Graaf Gasthuis, Delft, Netherlands

Patients who are admitted to the ICU are mostly offered the whole range of ICU interventions. However, during the course of intensive treatment, treatment options could be limited for several reasons. This study was done to see how often treatment options are limited in patients who eventually die in the ICU.

The unit is a 10 bed closed format general ICU in a 550-bed general hospital. Two thirds of patients are surgical. All charts of all patients who died in the ICU from January 1st, 2005 until December 31st, 2005 were examined retrospectively. In our hospital patients are supposed to receive a resuscitation code upon admission from the admitting doctor. When the patient's situation or preferences change, the resuscitation code can be changed at any time. Code 1 stands for all treatment including tracheal intubation and cardiopulmonary resuscitation (CPR). Code 2A allows CPR but no mechanical ventilation outside CPR. Code 2B allows mechanical ventilation but no CPR. Code 3 allows no mechanical ventilation or CPR. There are three different codes 3: 3A allows ICU treatment, 3B allows no ICU treatment and 3C allows only palliative treatment.

In the study period, 794 patients were admitted to the ICU with a mean age of 66.7 years, range 16-100 years and a mean APACHE II score of 13.6, range 0-41. Sixty-seven patients mean age 72.6 years, range 37.5-92.9, mean APACHE II score 24.5, range 7-41, died in the ICU after a median length of stay of 1.5 days, range 0-104 days. Another 45 patients died in the hospital after ICU discharge. All 15 patients that died with code 1, recieved CPR. Ten (67% of these patients had been admitted for less than 24 hours. CONCLUSION. Most patients who die in the ICU, do so after a decision to limit treatment has been made. CPR is predominantly used in patients who have only just been admitted.

Bleeding and metabolism 0220-0231 0220

Meurant F 1 1 Intensive Care Unit, Kirchberg Hospital, Luxembourg, Luxembourg

Haemorrhagic shock (HS) is a major cause of death in European Intensive Care Units (ICU). We here report our experience with the use of Recombinant factor VIIa (rFVIIa) (Novo Seven) as rescue haemostatic strategy for the management of patients presenting with HS.

Twenty patients admitted to the ICU with HS following trauma (n=7) or surgery (n=13) were prospectively enrolled and randomized in the study during 8 months, family informed consent and institutional agreement were obtained. HS was defined as hypotension (mean arterial pressure < 60 mmHg), subcritical anaemia (Hb < 8 gr/dL), macroscopic haemorrhage and uncontrolled bleeding (transfusions of more than 3 units of blood within 2 hrs). All patients received Red Blood Cells (RBC), Fresh Frozen Plasma (FFP) and Blood Platelets (BP). Calcium, proteins, fibrinogen (>1.5 g/L) concentrations and temperature were corrected. Five out of the 20 enrolled patients received a single vial of 4.8 mg of rFVIIa, independently of the body weight, as recommended by published rFVIIa dosing policies (1) . All patients underwent a surgical exploration or radiological procedure before administration of rFVIIa. For statistical analysis a Shapiro-Wilk test and a Student T-test were used.

Transfusions requirements (RBC, FFP and BP) were compared between the two groups. The need for RBC transfusions (units, mean±SD) decreased significantly from 22.8±12.1 to 1.4±2.6 (p = 0.01) among patients treated with rFVIIa. A similar although not statistically significant trend was observed FFP transfusions (17 versus 3 units, p = 0.19) and BP (2.2 versus 0 units, p = 0.21). The mortality rate did not differ between the two groups (3/5 versus 9/15). However, haemorrhage was the cause of death in only 33% of patients who received rFVIIa and in 66% of those who did not (p<0.001).

As suggested by our study, treatment with rFVIIa of patients with uncontrolled bleeding reduces the transfusions requirements and the haemorrhage-associated mortality. Individual responses remain however unpredictable. 6.0 ± 9.7 0.196

In conjunction with surgery and conventional therapy, rFVIIa use for haemorrhage was associated with a significant reduction in blood product transfusion and normalisation of clotting parameters. Coagulation, temperature and perhaps pH should be optimised prior to rFVIIa use. rFVIIa use may be futile in profoundly acidaemic patients.

Studies have shown the importance of tight blood glucose (Glc) control to reduce mortality and morbidity amongst critically patients [1] . However, not only the optimal glucose range but also treatment strategies are still a matter of vivid discussion. We examined the effectiveness of a strictly protocol-guided insulin infusion therapy versus a liberal nursing staff dependent therapy in cardiac surgery patients.

Local ethical committee accepted, single-center, prospective study with retrospective controls. After given written consent 164 patients (Protocol Group) admitted to elective open heart surgery were treated according to an hourly-glucose-measurements-based insulin infusion protocol during the operation and within 24 hours after admission to the ICU. Blood glucose management of the control group in 108 patients (Non-Protocol Group) was liberally led by the nursing staff. The aim in both groups was to maintain blood glucose levels between 80-180mg/dl. We compared minimum, average and maximum glucose levels and the amount of the infused insulin dosages (International Units/hour, IU/h). Statistics: Data presented in mean ± standard deviation, Student's T-Test.

Preoperative base line glucose levels were not different between the Protocol-and Non-Protocol Groups (Glc mg/dl, 115±25 versus 113±32, p=0.45); postoperative Glucose levels and insulin infusion rates were as follows: Glucose mg/dl, min: 106±18 vs 118±25, average: 137±20 vs 163± 26, max: 177±38 vs 212±56, p<0.001; Insulin IU/h, min: 1.3±1.1 vs 1.2±1.2, p=0.1 average: 2.6±1.5 vs 2.1±1.5, p=0.02; max: 4.9±2.7 vs 3.4±2.9, p=0.08. We could not find a significant difference in total insulin dosage between groups.

To achieve beneficial tight glucose control clinicians may not sufficiently rely on "common practice" but use strict insulin infusion protocols. According to our data the potential of hourly adjusted insulin infusion rates in a protocol guided concept may not even lead to significant higher consumption of insulin. 

As revealed by several studies, mortality after cardiac surgery can be related to glucose blood levels (1). Epidural anaesthesia reduces the level of stress hormones that accompanies cardiac procedures (2) . The aim of the study was to evaluate if epidural anaesthesia could improve postoperative glycaemic control in cardiac surgery.

METHODS. This prospective randomized study included 50 cardiac surgical patients affected by noninsulin dependent diabetes mellitus. Twenty-five patients received combined high thoracic epidural anaesthesia with total intravenous anaesthesia (TIVA) (group A) and 25 patients received exclusively TIVA (group B). All patients received a continuous i.v. infusion of insulin until the second postoperative day. We recorded total insulin dose infused, blood glucose levels and the average time requiered to achieved normoglycaemia.

The results are shown in the following table. There were no differences between groups regarding demographic data, type of surgical procedures, duration of cardiopulmonary bypass, use of inotropes. A significant correlation between type of anaesthesia and glycaemic levels has resulted. In group A normoglycaemia was achieved in shorter periods without increasing the incidence of hypoglycaemic events. 

Hypernatremia is a common electrolyte disorder in the surgical intensive care unit (S-ICU). This study evaluates the risk of several factors of hypernatremia in the S-ICU.

A case-control retrospective study was conducted by data extracted from the electronic data base. Group 1 consisted of normonatremic patients and group 2 consisted of hypernatremic patients. The factors that evaluated were age, gender, APACHE II score, sepsis criteria, days of ICU stay, mortality, BUN levels, creatinine serum levels, and administration of furosemide, antibiotics, human albumin, and fresh frozen plasma (ffp).

The 279 patients of group 1 and the 35 patients of group 2 had no differences according to gender, age, APACHE II score, sepsis criteria, BUN and creatinine levels and furosemide administration. The mortality was higher in the second group (p<0.01). The administration of antibiotic therapy was not identified as a risk factor of hypernatremia, with the exception of cefuroxime. Cefuroxime as well as human albumin administration seems to be statistically significant prophylactic factor of hypernatremia, while ffp administration and each additional day in the S-ICU setting are statistically significant risk factors of hypernatremia. The following table shows the odds ratios of these factors. 

In our study, each additional ICU day as well as the ffp administration seems to increase approximately 3 times the risk of hypernatremia. The administration of one single dose of human albumin and cefuroxime decreases nearly 5 and 1.5 times this risk, respectively. Poster Sessions Technology assessment (I) 0232-0245 0232

Yarustovsky M B 1 , Grigorianz R G 2 , Geptner R A 2 , Abramyan M V 2 , NazarovaÅ I 2 , PlyushtchÌ G 2 1 Unit of extracorporeal support of internal organs, 2 Extracorporeal support of internal organs, Bakoulev Scientific Center for Cardiovascular Surgery, Moscow, Russian Federation

The ischemic deterioration of liver functions being part of the multiorgan failure syndrome (MOFS), after cardiosurgery is characterized by high mortality level -up to 80%. The opportunity to include the albumin dialysis method, replacing the detoxifying liver function, in the complex intensive therapy, allows to gain time while awaiting the functional recovery of the own organ.

We have an experience of 35 procedures in 16 pts. Three of them were operated for congenital heart diseases (body weight range 9 -47 kg, age range 4 months -29 years). Other pts suffered from acquired heart diseases (the average body weight 70.6±19.5 kg, the average age 50±14 years).

After the operation all pts developed MOFS, comprising heart (EFLV <40%, epinephrin >0.1 µg/kg/min.), respiratory (PaO2/FiO2<200), renal (80% of pts) and liver failure, the different stage of DIC syndrome. The failure of standart medical therapy applied to liver failure, progressing growth of bilirubin level (total bilirubin is 305.2±120.5 µmol/l), increase of liver ferments value (AST 245±168 U/l, ALT 228±187 U/l, GGT 142±68 U/l), decreasing cholinesterase level (3965±1235 U/l), breach of the hemostasis system (INR -3.1±1.3, prothrombine index -37.2±16.4%) and the increasing level of blood ammonia (over 50 µmol/l) became indications for MARS-therapy. The procedure was carried out with MARS device (Gambro, Sweden) combined with the artificial kidney Fresenius 4008B (Fresenius, Germany).

No negative MARS impact on hemodynamics was revealed. In 4 pts were revealed the 10% increase of heart index, the increase of total vascular peripheral resistance for 18% after the albumin dialysis provided the unchanging dosage of inotropic drugs. The MARS-therapy resulted in the decrease of bilirubin level at adult pts, in average by 29.0±3.9% from the preprocedural value, as for children, the decrease reached 57%. All pts demonstrated the possibility to control the level of low molecular water soluble substances (urea 50-60% from initial level). The plasma of 8 out of 10 examined pts showed the reliable growth of cholinesterase concentration 12 hours after MARS procedure. The total survival was 25% (4 out of 16 pts).

Our experience suggests that MARS therapy is a perspective method for support of the decompensated liver failure in pts with MOFS after cardiosurgical operations, and can also be safely applied in infants and patients with unstable hemodynamics. 

A pre and post analysis of expectations was carried out through a questionnaire. ICU's physicians and nurses were the population evaluated. Before full implementation of the electronic system the pre questionnaire was applied. After 3 months of complete use, the post questionnaire was undertaken. Each questionnaire, with both multiple choice and open questions, had 3 major parts: 1) demographic data 2) feasibility of input and review 3) performance and quality.

Most answers considered that recording data was moderately simple and fast (75.8% pre and 96.3% post). When we consider also prescription there were less favorable answers on the pre assessment (54.5%), but it improved after 3 month's experience (81.5%). Most users think that consulting the system is slow and difficult or just moderately simple. In the pre period 72.7% consider the system better than handwritten files, but after 3 months negative answers raised to 37%. Principal advantages found were limitation of errors and easier use of protocols. Most users won't change again to manual charting (59.3%). 

Although being a subjective evaluation this study points out interesting features in need of further investigation: 1) Low percentage of answers, eventually meaning a low level of motivation about this issue 2) Therapy prescription easier than expected initially 3) Main problems expected and felt when reviewing data, maybe the main weakness of this system 4) Overall perceived advantage of PDMES over traditional paper charting. 

The objective of surgical repair of a thoracoabdominal aortic aneurysm (TAAA) is to replace the diseased aorta without compromising the viability of the spinal cord. The most common etiology of TAAA is atheroscerosis and the typical patient is elderly with multiple coexisting diseases. Different approaches for the management of intercostal and lumbar arteries and monitoring of spinal cord integrity during repair of TAAA exist, but the ideal strategy remains to be demonstrated.

During the period 10/2002 -12/2004, 100 patients (age 63.5 ± 13.5 years) underwent elective TAAA repair with intraoperative monitoring of motor (MEP) and somatosensory evoked potentials (SSEP) in our institution. An average of 8.0 ± 2.6 segmental artery pairs were sacrificed overall, with an average of 4.5 ± 2.1 segmental pairs sacrificed between T7 and L1, where the artery of Adamkiewicz is presumed to arise. At surgical preference, an anesthetic management strategy without inhaled agents or muscle relaxants was employed that does not interfere with MEPs or SSEPs. One lung ventilation, mild hypothermia, distal perfusion, high-dose steroids, antifibrinolytics and cerebrospinal fluid (CSF) drainage were employed in all cases. CSF drainage was discontinued once lower extremity function was demonstrated to be normal.

Hospital mortality was 6%. The average CICU stay was 4.3 days and mean hospital stay 16 days. There was no reported intraoperative recall and no patient exhibited movement during their procedure. In 99 cases, MEPs and SSEPs remained unchanged during the course of serial segmental artery sacrifice, or could be returned to baseline levels by manipulation of the blood pressure. Postoperative paraplegia occurred in only two patients. After normal function of lower extremities, the drainage of CSF was terminated.

Monitoring of MEPs and SSEPs appears to add to the safety of TAAA repair. Omission of potent volatile inhaled agents and muscle relaxants from the anesthetic technique is compatible with an adequate anesthetic state and a motionless surgical field.

Jacobs 

Corstjens A M 1 , Ligtenberg J J 2 , Spanjersberg R 2 , Tulleken J E 2 , Meertens J H 2 , Zijlstra J G 2 1 Anesthesiology, 2 Medical Intensive Care, University Medical Center Groningen, Groningen, Netherlands

INTRODUCTION. An accurate and feasible bedside glucometry method is essential to obtain tight glucose regulation. We evaluated the reliability of the subcutaneous continuous glucose monitoring system (CGMS Gold, Medtronic Minimed) in critically ill patients. An advantage of the CGMS is that no blood has to be drawn from the patient. The CGMS features a subcutanuous sensor that can be used for up to 72 hours. Glucose is measured by the glucose-oxidase method; interstitial glucose measurements are sent to a monitor which records an average glucose value every 5 minutes (range of 2.2-22.2 mmol/l (40-400mg/dl)). The study was performed as a single-center, prospective, observational study in a 12-bed Medical Intensive Care Unit of a University Hospital.

The CGMS sensor was inserted in the abdominal subcutis and calibrated every 6 hr. Glucose data were downloaded after 24-72 hr. Furthermore, heparinized arterial bloodsamples were drawn from an arterial line every 4 hours and analyzed on the blood gas/blood glucose analyzer ABL715 (Radiometer Medical; previously we validated this analyzer in ICU patients). The results of the paired measurements were analyzed as a scatter plot, by the method of Bland and Altman and were expressed as a correlation coefficient.

. 60 patients were included. 786 paired readings were analyzed. Subcutaneous glucometry provided an acceptable estimate of blood glucose assessment compared to the ABL 715: the correlation coefficient was 0.87; in the Clarke error grid 100% of the paired measurements were in the clinically acceptable zones A & B.

CONCLUSION. The CGMS Gold is reliable and easy to use in critically ill patients. However, this version of the CGMS is not yet useful for tight glucose regulation, since it provides no on-line glucose data (the next version will).

Tarabrin O 1 1 Anaesthesiology, Odessa Medical University, Odessa, Ukraine

Venous thromboembolism is one of the most common complications seen in cancer patients and may be due to the hypercoagulalable state of malignancy and to it's surgical treatment.

Patients received MEDNORD (Ukraine Co analyser) analysis (HVG), a viscoelastic test, measures clot formation and includes information on the cellular, as well as the plasmatic coagulation system. A complete coaguation screen, activated clotting time (ACT), thromboelasthgraphy (TEG) and haemoviscoelastography (HVG) were performed before surgery, at the end of surgery, and on postoperative days 1, 2, 3, and 7; they were analyzed for the reaction time and the maximal amplitude (MA). We tested the hypothesis that the parallel use of standard TEG and HVG can assess postoperative hypercoagulability and can estimate the independent contribution of procoagulatory proteins and platelets.

We calculated the elastic shear modulus of standard MA (Gt) and HVG MA (GH), which reflect total clot strength and procoagulatory protein component, respectively. The difference was an estimate of the platelet component (Gp). There was a 14% perioperative increase of standard MA, corresponding to a 48% increase of Gt (P < 0. 05) and an 80%-86% contribution of the calculated Gp to Gt. We conclude that serial standard thromboelastography and HVG viscoelastic test may reveal the independent contribution of platelets and procoagulatory proteins to clot strength. Using multiple linear regression, all coagulation, TEG and HVG variabities were used to model postoperative hypercoagulation. Results showed that some components of the TEG failed to identify hypercoagulation (r < 0, 2, P > 0, 75). However, three components of the routine coagulation assay, including bleeding time, prothrombin time, and platelet count could be modeled to show prolonged postoperative hypercoagulability (P < 0, 01). We conclude that all components of the HVG test reflect postoperative coagulopaties, these results suggests that it may be usefull in determining the coagulation status of cancer patients perioperatively.

Postoperative hypercoagulability, occurring for at least 1 week after major cancer abdominal surgery, may be demonstrated HVG viscoelastotest. This hypercoagulability is not reflected completely by standard coagulation monitoring and TEG and seems to be predominantly caused by increased platelet reactivity. HVG viscoelastotest provides a fast and easy to perform bedside test to quantify in vitro hemocoagulation.

The measurement of regional lung ventilation by electrical impedance tomography (EIT) has been evaluated in many experimental studies (1, 2) . However, EIT is not routinely used in a clinical setting which is attributable to the fact that a convenient concept how to quantify the EIT data is missing. The definition of region of interest (ROI) is an essential point in the data analysis.

To date, there are only limited data available on the different approaches to ROI definition to evaluate regional lung ventilation by EIT.

For this survey we examined ten patients (mean age ± SD: 60 ± 10 years) under controlled ventilation. The EIT examinations were performed with the Goe-MF II EIT device (Viasys Healthcare, Höchberg, Germany). Sixteen self-adhesive electrodes (3M Red Dot 2239, 3M Health Care, Borken, Germany) were applied on the chest circumference in one transverse plane and used for rotating electrical current injection and voltage measurement. The EIT data were acquired at a rate of 13 scans/s during a 60-s time interval. Regional tidal volumes were quantified as pixel values of inspiratory-to-expiratory differences in relative impedance change and four types of ROIs were subsequently applied. The definition of ROI contours was based on the calculation of the pixel values of 1) standard deviation from each pixel set of impedance data and 2) regression coefficient from linear regression equations between the individual local (pixel) and average (whole scan) impedance signals. Additionally, arbitrary ROIs (four quadrants and four anterioposterior segments of equal height) were used.

In all 10 patients, good quality EIT data were acquired and the functional EIT scans clearly visualised the ventilated lung regions. No significant differences between the quantitative analyses of regional lung ventilation using the two functional approaches to ROI definition were found. Our results indicate that both approaches to ROI definition using functional statistical parameters are suitable when impedance signals with high sensitivity to ventilation-related phenomena are to be analyzed. The simple arbitrary ROIs may be combined with the functional ROIs.

CONCLUSION. The definition of the ROI contour as 20%-35% of the maximum standard deviation or regression coefficient is recommended. The simple segmental ROIs are less convenient because of the low ventilation-related signal component in the dorsal region. 

Vibration response imaging (VRI) is a novel technology that measures vibration energy generated from airflow to create a real-time structural and functional image of the respiration process. Sophisticated software and surface skin sensors are placed on the back to record, analyze and display vibrations as a non-invasive measure of airflow in the lung.

We performed VRI in two patients during tidal breathing and immediately thereafter during incentive spirometry (IS). One patient was recently extubated following hemoptysis associated acute respiratory failure ( Fig. 1 ) and the second patient was admitted with acute non-Q wave myocardial infarction (not shown). In both circumstances, the left lower lung field was poorly ventilated during tidal ventilation.

Incentive spirometry produced a striking increase in ventilation fo the left lower lung.

The current incentive for incentive spirometry is patient recognition of achieved tidal volume goal. These two cases highlight the potential utility of VRI as a non-invasive, bedside reinforcement tool in which reversal of atelectasis would be the feedback target.

Costa R 

Patient-ventilator synchrony is strongly dependent on both the ventilator settings and the interface used to apply positive pressure to the airway. The aim of this bench study is to compare three different interfaces for mechanical ventilation, the endotracheal tube (ET) the facial mask (FM) and the helmet (H) during Pressure Support Ventilation (PSV) delivered with different rates of pressurization (Timepress) and expiratory trigger (Trexp).

METHODS. PSV (inspiratory support 12cmH2O, PEEP 5cmH2O) was delivered to an head manikin, connected to a test lung (ASL 5000, IngmarMedical) with an ET, a face-mask and an helmet. We tested 3 Respiratory Rate (RR) (14, 20 and 30 breaths/min) and two ventilatory settings (Timepress 50%-Trexp 25% and Timepress 80%-Trexp 60%), applied in random order. All the data are expressed as Mean ± SD. The analysis of variance for repeated measures was performed with ANOVA test. P values lower than 0.05 were considered statistically significant.

The analysis of patient-ventilator interaction shows that the synchrony was significantly better with the ET than with the mask and the helmet, as shown by inspiratory trigger delay (Delaytrinsp) (p<0.001) and time of assistance as well (p<0.001 

and Surgical ICU (total 26 beds) and use of hemodinamic monitoring -ED (Cardio Q, Dextel, UK) or PAC (Vigilance, Edwards LifeSciences, USA). We conducted a retrospective cohort using the data bank Quati (Dixtal, Brasil). Two groups were made (ED and PAC) for comparision.

In this period 5835 patients were admitted in our ICU and 410 (7%) fulfilled the criteria. 33 patients were excluded because the usage of both methods. 125 patients used ED and 253 PAC. 

We report on the early MRI findings of HO of the knee joint in 10 patients hospitalized in the ICU for CNS trauma (6 patients), necrotizing pancreatitis (1 patient) and fat embolism (1 patient). Upon clinical and laboratory suspicion for HO, conventional radiographs (CR) and MRI of the knee was performed (19.9±7.4 days after admission).

. CR were negative, while MRI depicted joint effusion and a "lacy pattern" of the vastus lateralis and vastus medialis muscles with hyperintense septa interposed among the low signal intensity muscular fibers, on STIR images. Homogeneous high signal was observed at the innermost part of the vastus medialis. Due to enhancement of the intermuscular septa a "lacy pattern" was also observed in contrast enhanced fat suppressed T1-weighted images. On follow-up the lesion was limited at the innermost part of the vastus medialis which showed heterogeneous high signal on STIR and T1weighted images and homogeneous enhancement after contrast administration. CR depicted a calcified mass confirming the diagnosis of HO in all 10 patients.

CONCLUSION. MRI of the knee performed upon clinical suspicion, shows a distinct imaging pattern that may be useful in the diagnostic work-up of HO patients. Early diagnosis with MRI might present an important advantage, which is to facilitate effective prevention of HO. 

Lellouche F 1 , Lyazidi A 2 , Brochard L 2 1 Centre de Recherche, Hôpital Laval, Sainte-Foy, Canada, 2 Service de Réanimation Médicale, Hôpital Henri Mondor, Créteil, France

Heated Humidifiers (HH) with heated wire have been shown to deliver absolute humidity levels as low as 20mgH2O/L in situations of high ambient temperature and/or using turbine ventilators (1) , which can lead to severe adverse effects. The aim of the study was to evaluate two new devices designed to avoid such pitfalls.

We measured on bench hygrometry (psychrometric method) of new HH (HC 200, Grundler medical) and HH with new compensation algorithm (MR850, Fisher&Paykel) in different situations of room temperature and ventilator setting.

CONCLUSION. New tested HH showed better humidification performances and were not influenced by room temperature. Edwards Lifesciences, Irvine, CA, USA) only requires access to the radial or femoral artery using a standard arterial catheter and does not need an external calibration [1] . We validated this technique in critically ill patients in the intensive care unit (ICU) using COPAC as the method of reference.

After obtaining approval of the Institutional Ethics Committee, we studied 20 critically ill patients (10 male), aged 16 to 74 years (mean, 55.5±18.8 years), who required both arterial and pulmonary artery pressure monitoring. COPAC measurements were performed at least every 4 hours and calculated as the average of 3 measurements randomly distributed over the respiratory cycle, while COAP values were taken immediately at the end of bolus determinations. Accuracy of measurements was assessed by calculating the bias, precision and limits of agreement using the method described by Bland and Altman [2] .

A total of 167 coupled measurements were obtained. Absolute values of COPAC ranged from 2.80 to 10.80 L/min (mean 5.95±1.58 L/min). The bias between COPAC and COAP was -0.002 L/min±1.38 L/min with limits of agreement (2SD of the bias) of ±2.76 L/min.

The COAP algorithm shows a minimal bias with COPAC over a wide range of values in an inhomogeneous group of critically ill patients. The relative wide scattering of the data may partly be explained by the calculation mode with which the algorithm compensates for slow changes in vascular resistance using a 10 minute moving average. This method may have difficulty responding to rapid changes in vascular tone during haemodynamic instability. Improvement regarding the response time to changes in vascular tone should enhance the accuracy of COAP determination and this has been addressed in a subsequent version of the software. Our data suggest that FloTrac TM system is a promising alternative for cardiac output assessment and provides an acceptable accuracy for clinical decision making in an ICU setting.

Our objectives are to analyze if the clearance of pulmonary edema (PE) depends of the intensity of the edema and to know if the application of PEEP increases or not this clearance rate.

Experimental study on 25 pigs weighing 30±3 kg, anaesthetized and in mechanical ventilation during 4 hours with tidal volume of 10 ml/kg, respiratory rate of 20 per min and FiO2 of 0.6. Four experimental groups, combining two levels of PE with or without PEEP, were studied. Group Low PE (n=10): PE was produced with 4 ml of intratracheal saline solution (ss). Group Low PE+PEEP (n=5): the same as the previous but PEEP of 10 cm H2O were added. Group High PE (n=5): 10 ml/kg of ss were instilled intratracheally. Group High PE+PEEP (n=5): the same as the previous but PEEP of 10 cmH2O were added. Extravascular lung water (EVLW) was determined using the PiCCO system (Pulsion, Germany). EVLW, blood gas exchange and respiratory parameters were determined every 30 min during 4 hours. Mann-Whitney test was used for quantitative variables and a p<0.05 was considered statistically significant.

In absence of PEEP, the clearance rate in the 2 first hours was similar in groups of Low and High PE (22.6 ml/h and 35.3 ml/h, p=NS). During the 3 th and 4 th hours the clearance decreases in Low PE group (5 ml/h vs 32 ml/h, p<0.005). The application of PEEP in group of Low PE originates a sharp and significant increases of PE clearance (at 30 min, 11±17 ml vs 63±36 ml, p<0.05) (see Figure) . In group High PE + PEEP there are non statistically significant differences.

During the first 2 hours, the clearance of PE does not depend of the intensity of the edema. The low clearance rate during 3 th and 4 th hours in the Low PE group indicates that the PE is practically resolved. The application of PEEP increases the PE clearance early, in particular in low levels of PE. 

Acute pulmonary edema is one of the most common problems in critically ill patients and has a great effect on the outcome. Although the pathological change of pulmonary edema is the accumulation of fluid in the lungs, it has never been clearly defined quantitatively. It is known that extravascular lung water (EVLW) estimated by the PiCCO system (Pulsion Medical Systems) correlates closely with gravimetric measurements of lungs in experimental animal models [1] . It is also demonstrated that EVLW correlated well with oxygenation in ALI/ARDS patients in septic shock and prognosis in critically ill patients [2, 3] . Thus, the aim of the present study was to define acute pulmonary edema quantitatively by EVLW using the PiCCO system.

All 75 patients who had a central venous catheter and a thermistor-tipped arterial thermodilution catheter (PiCCO system) for hemodynamic management treated in three hospitals between July 2004 and January 2006 were included. The correlation between PaO2/FiO2 ratio and EVLW was evaluated to investigate the relationship between EVLW and oxygenation in patients with pulmonary edema. We also evaluated the relationship between EVLW and postmortem lung weight in eight patients in whom autopsies were carried out within 48 hours after thermodilutional EVLW measurement.

To define the pulmonary edema quantitatively, a PaO2/FiO2 ratio of 200 and postmortem lung weight of 27mg/kg was considered as the cutoff value of pulmonary edema according to the literature.

Measurement of EVLW using the PiCCO system was very closely correlated with PaO2/FiO2 ratio (R=-0.57 P<0.001) in patients with pulmonary edema and with gravimetric measurement of lung weight (R=0.93 P=0.002) in humans. From the correlation between EVLW and PaO2/FiO2 ratio, 14ml/kg was derived as the cutoff value for pulmonary edema, and the same value was derived from the correlation between EVLW and lung weight.

CONCLUSION. EVLW of 14ml/kg measured by the PiCCO system may be the quantitative definition of acute pulmonary edema. 

Nitti C 1 , Gentili T 1 , Rapino K 1 , Salvi A 1 1 Emergency Medicine, General Hospital, Ancona, Italy

Pulmonary embolism has multifarious clinical presentations and has been divided into two groups: massive PE, often associated to shocks, and non-massive PE. In between the two extremes there lies a continuum range having different degrees of seriousness, with a rate of death ranging from over 30% in the shock-associated forms, that reduces to 1% in the asymptomatic or paucisymptomatic forms. Right ventricle disfunction is considered a very useful indicator of a serious case, but its assessment is not univocally standardized and not easy to gain in emergency.

In a perspective study we examined 125 patients with a pulmonary embolism diagnosis and with a minimum six months follow-up. We assessed survival in relation to age, sex, troponine I, RVD, PE massive/non-massive on a clinical evaluation, COPD, cardiopathy, trombolysis (yes/no), neoplasia (yes/no), shock index (HF/SBP). 

Our data highlight a high early death (within one month) and this seems to show that a more aggressive and timely therapeutic approach may improve survival rates. A very useful element to predict the PE disease was the Shock Index, which proved particularly useful in emergency, as it can be easily gained. In our experience RVD assessed by echocardiography was not particularly meaningful, probably because of the inaccuracy in standardization of parameters. 

A low maximal EVLW value correlated well with survival. There seems to be a correlation between positive fluid balance, a high capillary leak index, subsequent increased EVLW and poor prognosis. In our preliminary results, a high EVLW however was not an independent predictor of prognosis. The study will be continued until 150 patients. .. Consequently, physicians are often reluctant to setup a PEEP in the peri-intubation period, fearing a deleterious hemodynamic impact which would exceed the potential favorable effect on gas exchanges. Hypothesis: setting a 5cm H20 (PEEP) vs zero PEEP (ZEEP) in the peri-ntubation period does not significantly affect the hemodynamic status of MICU patients in a peri-intubation period of 90 min and does not influence the outcome.

Methods: a prospective one-center interventional blinded randomized pilot trial, including consecutive patients with clinical indication of intubation. The primary objective was a variation of mean arterial pressure (MAP) from baseline up to 90 min post intubation. Secondary issues addressed mean duration of intubation, level of MAP support after intubation, and 28-days mortality. Neither anesthestic procedures nor levels and types of intervention for maintaining MAP after intubation were codified, and all were left to the judgment of the on-duty physician. Levels of MAP support after intubation (0 to 4 scale): 0 (2 vs 3); 1 (2 vs 6); 2 (8 vs 7); 3 (17 vs 9); 4 (6 vs 4)(p>0.05).

In this clinical pilot trial, there is no evidence that setting a prophylactic PEEP of 5 cm H2O neither adversely affect the short-term hemodynamic status nor the outcome of MICU patients in a peri-intubation period.

Roch A 1 , Blayac D 1 , Michelet P 1 , Ramiara P 2 , Forel J 1 , Lambert D 3 , Carpentier J 2 1 Medical intensive care unit, Hopital sainte marguerite, 2 Intensive care unit, HIA Laveran, 3 Intensive care unit, Hopital sainte marguerite, marseille, France

The present study compared the effects of isotonic versus hypertonichyperoncotic resuscitation on pulmonary edema and histological injury induced by a severe hemorrhage in a pig model.

19 pigs (43 ± 4 kg) were anesthetized, ventilated and randomized in 3 groups: without hemorrhage (n=5), hemorrhage group resuscitated with 0.9% NaCl (NS group, n=7) or hemorrhage group resuscitated with a solution of 7.2% NaCl / 6% hydroxyethylstarch 200/0.5 (Hyperhes, Fresenius)(HSHES group, n = 7). The protocol consisted into five periods: hemorrhagic shock during 2 hours (MAP<40 mmHg and cardiac index < 60% of baseline and SVO2<30%), resuscitation during 2 hours (2 ml/kg/min of NS or 4 ml/kg of HSHES in 10 minutes followed by an infusion of 0.2 ml/kg/min of HSHES in order to obtain a CI of 90% of baseline and a SVO2 of 50%), 1h of transfusion (Hb> 80% of baseline) and 1 hour of observation. After euthanasia, extravascular lung water (EVLW) was measured by gravimetry and histological score was determined on lung biopsies.

The volume of removed blood was 37±6 ml/kg in the NS group and 39±3 ml/kg in the HSHES group. Resuscitation endpoints were achieved in similar times with 90±17 ml/kg of NS (34±4g of Na) and 6.8±1.9 ml/kg of HSHES (20±5 g Na). CI, SVO2, oxygen delivery, and blood lactate were not different between loading strategies during the whole protocol. Hemorrhage induced an increase in EVLW which was independent of the resuscitation strategy (9.7±1.8ml/Kg in NS group vs 9.2±1.4 ml/kg in HSHES group and 6.4±1 ml/kg in control group, p<0.05 vs control). The degree of alveolar membrane focal thickening and the degree of interstitial neutrophil infiltration were significantly more severe in the hemorrhage groups without effect of the type of fluid loading. Hemorrhage did not induce hypoxemia or alveolar injury but increased indexed pulmonary resistances (488±143 vs 233±32 dyne.sec.cm-5.m2 at baseline). A pulmonary hypertension occurred after resuscitation in the NS group but not in the HSHES group (MPAP= 23±4 mmHg in NS group vs 15±5 mmHg in HSHES group (p<0.05) and 12±1 mmHg in control group (p<0.05)).

Despite a very negative fluid balance, small-volume resuscitation did not attenuate early hemorrhage-induced pulmonary edema and interstitial histological injuries. These results do not promote the use of a small-volume resuscitation as a strategy aiming at limiting early hemorrhageinduced pulmonary edema. 

In this short-term, fluid resuscitated, severe model of endotoxic sepsis, different patterns were observed in tissue PO2in the four organs studied, with rises in bladder and falls in muscle and (particularly) liver. Changes were not related (where measured) to blood flow. These data suggest that early responses to sepsis are organ-specific and may relate to local changes in oxygen supply, demand and utilisation.

A mitochondrial dysfunction has been described for human peripheral blood mononuclear cells (PBMC) during septic shock. Little is known about consequences of these metabolic alterations on PBMC functions. Purposes of this present study are: 1) to assess global oxygen consumption rate (VO2) of PBMC and fractions for NADPHoxidase and 4 metabolic routes which consume also ATP, 2) to study modifications of these patterns by septic plasma incubation.

METHODS. PBMC have been isolated by gradient centrifugation from the whole blood of 8 healthy volunteers. Cells were incubated for 3 hours either in their own plasma (control) or in pooled plasma obtained from 6 septic shock patients. VO2 has been measure amperometrically for quiescent and stimulated cells (PMA and ionomycin). To quantify the fractions of VO2 for the main ATP consuming pathways, inhibitors of protein synthesis (cycloheximide), RNA/DNA synthesis (actinomycin D), Na+, K+ ATPase (ouabain) and Ca2+ ATPase (lanthanum chloride) were applied. Oxygen consumption rates of NADPH oxidase and of respiratory chain were also quantified by inhibition with DPI (for NADPHoxidase) and antimycin A (for complex III). Satistics: Wilcoxon rank tests, results: mean ± SD.

RESULTS. VO2 of quiescent PBMC decreased when PBMC are incubated in septic plasma (5.7±1.43 and 7.22±1.74 ng.atom O2/min/107 cells (U) for septic and control plasma respectively, p=0.011). A trend to a mitochondrial respiration rate decreased was observed. Only VO2 linked to Na+, K+ AT-Pase activity was significantly reduced in septic plasma (p=0.021). The fraction of total VO2 related to NADPHoxidase in quiescent cells was similar in the two conditions. Stimulation increased NAPH oxidase VO2 in both experimental conditions but increased more in septic plasma (+3.86±3.08 and +0.6±1.34 U for septic and control condition, respectively, p=0.036).

Septic shock plasma modifies bioenergetics of healthy PBMC, especially by reducing cells global oxygen consumption. Stimulation with PMA and ionomycine increased NADPHoxidase oxygen consumption rate, but more in incubation in septic plasma. Septic plasma increases reactive oxygen species production, and also oxidative stress under PBMC stimulation. [1, 2] . It is conceivable that the experimental sepsis model has an impact on mitochondrial performance.

METHODS. 33 anesthetized pigs were randomized to either saline (C, n=12), continuous endotoxin infusion (E, n=9) or fecal peritonitis (P, n=12) for 24 hours. Cardiac index (thermodilution, ml/kg/min) and systemic mean arterial pressure (MAP, mmHg) were recorded and glutamate-dependent State 3/4 (isolated muscle mitochondria, nanoatom O2/min/mg protein), respiratory control ratio (RCR) and tissue ATP (mumol/g) measured at baseline, after 6 hours and at the end of the experiment.

RESULTS. a p< 0.05 vs. baseline; b p< 0.05 vs. control. 

Groesdonk H V 1 , Siebach A 2 , Hoffarth B 1 , Heydrich S 1 , Senftleben U 1 1 Department of Anesthesiology, 2 Department of Cardiac Surgery, University of Ulm, Ulm, Germany

Sepsis-induced apoptosis of lymphocytes is associated with immune dysfunction and fatal outcome. Interestingly, prevention of lymphocyte apoptosis improves survival of septic mice. However, the molecular mechanisms leading to lymphocyte death during sepsis are not clear. Regular activation of NF-κB is essential to prevent T cells from TNFα-induced apoptosis. As sepsis is characterized by increased levels of TNFα, the NF-κB pathway might play an important role in this scenario.

Here we address the question whether thymocyte apoptosis in murine sepsis is associated with reduced NF-κB activity. Moreover, via enhancement of NF-κB activation in lymphocytes we tried to prevent lymphocyte death and to improve survival rates of septic mice.

Severe sepsis was induced in female C57BL/6-mice by cecal ligation and puncture [20G] (CLP). Thymus and spleen were harvested up to 24h from CLP and sham-operated mice. NF-κB activity was determined by Electrophoretic Mobility Shift Assay (EMSA) using cellular protein extracts. Tissue sections were H&E and trichrome (Masson-Goldner) stained. Apoptosis was analyzed via TUNEL-Assay. Mortality was determined up to 48h. IκBα-deficiency is known to cause increased NF-κB activation. To analyze the effects of enhanced NF-κB activation, we adoptively transferred Iκbα-/-or wild-type (wt) fetal liver stem cells into sublethally irradiated lymphopenic Rag1-/--mice (RC) and CLP was performed.

. NF-κB DNA binding activity in thymus and spleen of Sham-mice showed a distinct increase within 6h and a second peak after 24h. In contrast, NF-κB activity in both organs of CLPoperated animals was clearly diminished after 6h and showed no further increase. Apoptosis was dramatically increased after 12 and 24h in thymus as well as in spleen in CLP-mice compared to Shammice. H&E and trichrome staining showed increased organ destruction. Interestingly, CLP-operated Ikbα-/-RCs showed clearly reduced apoptotic rates and less organ destruction compared to sham mice. Moreover, CLP-treated B6-mice and RAG-wt-RCs were severely sick during the whole observation period and died between 32 and 48h. Instead, Iκbα-/-RCs showed increased survival.

During severe sepsis increased apoptotic rates in spleen and thymus are associated with reduced NF-κB activity in thymocytes. Enhancement of NF-κB activity prevents lymphocyte death and improves survival. Our data suggest that NF-κB signaling in lymphocytes is essential for adaptive immune responses during sepsis.

Matsuda N 1 , Yamazaki M 1 , Hattori Y 2 1 Anesthesiology, 2 Pharmacology, Toyama University Graduate School of Medicine, Toyama, Japan

Dysfunction of vascular endothelium has been implicated in the development of the sepsis-related complications, including disseminated intravascular coagulation and multiple organ failure. We investigated the effects of short double-stranded RNA fragments termed small interfering RNAs (siRNAs) targeting caspase-3 and -8, a family of cysteine proteases that act as pro-apoptotic regulators, on endothelial cell injury in a septic mouse model.

In BALB/c male mice (8-12 week-old), sepsis was induced by cecal ligation and puncture, and aorta was harvested 10 and 24 h after the onset of sepsis. Sham-operated animals underwent the same procedure except for ligation and puncture of the cecum. SiRNAs targeting caspase-3 and -8 were introduced to the mouse 10 h after the onset of sepsis via intravenous injection by means of the liposome method.

Western blot analysis showed that phosphatidylinositol 3'-kinase (PI3K) protein expression relative to that of sham-operated control was decreased by 60% in aortic membranes from mice 24 h after the onset of sepsis. Furthermore, a 60% reduction in the phosphorylated level of Akt was found in septic aorta without any significant change in total Akt expression. While the anti-apoptotic marker, phosphorylated Bad, was significantly down-regulated by sepsis induction, both caspase-3 and -8 were time-dependently up-regulated in septic mouse aorta (4∼5-fold at 24 h). Transfection with caspase-3 and -8 siRNAs resulted in complete restoration of the septic changes in membranous PI3K, Akt phosphorylation, Bad phosphorylation and caspase-3 and -8. Transmission electron microscopy of septic mouse aorta revealed detachment of endothelial cells, endothelial denudation, and microthrombus in the endothelial surface area, all of which were significantly improved by silencing of caspase-3 and -8 using siRNAs.

These results suggest that derangement of pro-apoptotic regulator expression plays a critical role in the development of endothelial histological injury after induction of sepsis. Since the PI3K/Akt pathway is considered to represent an effecter mechanism that promotes cell survival in the setting of sepsis, interruption of PI3K/Akt pathway activation could also contribute to sepsis-induced morphological injury in endothelial cells. 

Mitochondrial oxygen consumption (VO2) depends on the flow of electrons through the electron transport chain (ETC) and is normally coupled to energy (ATP) production. Electrons enter the ETC through Complex I (CI) or II (CII). Inhibition of CI, but not CII, has been described during sepsis (1) . The aim of this study was to investigate if succinate, a CII-specific electron donor, can improve mitochondrial VO2 during sepsis.

Skeletal muscle fibres were isolated from naïve (n=15), sham-operated (n=15), mildly (n=10), moderately and severely (n=10) septic rats, 48 hours after induction of faecal peritonitis and fluid resuscitation. Mitochondrial (cyanide-sensitive) VO2 was studied in the presence of glutamate plus malate (CI-specific electron donors) and after the addition of succinate (CII-specific electron donor) using a Clark oxygen electrode (Rank Brothers, UK).

Mitochondrial VO2 rates (nmol O2/min/mg of tissue) in the presence of glutamate plus malate and after the addition of succinate are reported (as means±SE) in Table 1 . Relative changes recorded after the addition of succinate (Delta VO2) are also presented. During moderate and severe sepsis, skeletal muscle mitochondrial VO2 in the presence of glutamate plus malate was significantly lower than in naïve and sham-operated animals; after the addition of succinate, it increased to the same level as that of the controls. 

Liver dysfunction is common in sepsis [2] . This may be related to regional blood flow limitation or redistribution [1] . We hypothesized that infection which is primarily located in the abdominal region (peritonitis) interferes to a greater extent with liver perfusion and function than systemic inflammation induced by endotoxin.

METHODS. 27 anesthetized pigs were randomized to either saline (C, n=11), continuous endotoxin infusion (E, n=7), or fecal peritonitis (P, n=9) for 24 hours. Systemic (thermodilution) and regional hepatic (ultrasound Doppler) and microcirculatory blood flow (laserDoppler, blood perfusion unit, BPU) were measured and oxygen transport and hepatic lactate exchange (mumol/min/kg) calculated. At the end of the experiment, glutamete-dependent liver mitochondrial respiratory ratio (RCR) was assayed polarographically.

RESULTS. a p<0.05 Friedman test, b p<0.01 Friedman test, c p<0.05 vs C. anesthetized, mechanically±Twenty-one female (body weight: 27.5 ventilated, hemodynamically monitored sheep received 1.5 g/kg body weight feces intraperitoneally to induce sepsis. Ringer's lactate and 6% hydroxyethyl starch solutions were infused throughout the experiment to prevent hypovolemia. No antibiotics or vasoactive agents were used. Two hours after injection of feces, animals were randomized to one of three equal groups: hypercapnia group -exogenous CO2 was given to maintain PaCO2 55∼65 mmHg; dobutamine group -g/kg/min dobutamine continuous intravaenous infusion; and control group -noµ7 treatment. All animals were studied until spontaneous death.

All animals developed a hyperdynamic phase characterized by hypotension, increased cardiac index and decreased systemic vascular resistance. All had metabolic acidosis with hyperlactatemia. PaCO2 was significantly higher in the hypercapnia group than the other two groups. The hypercapnia and dobutamine groups showed improved mean arterial pressure, higher stroke volume, higher heart rate, higher oxygen delivery and lower lactate concentrations compared with controls (p<0.05). there were no significant differences in any of these variables between the hypercapnic and dobutamine groups. The hypercapnic group had a higher PaO2/FiO2 ratio, lower mean airway pressure and lower wet/dry ratio compared with the other two groups. There was no difference in survival time among the three groups (p = 0.65).

CONCLUSION. Hypercapnia improved gas exchange, decreased pulmonary edema formation, had similar hemodynamic effects as dobutamine infusion in our clinical relevant septic shock model in sheep.

Hypercapnia improved gas exchange, decreased pulmonary edema formation, and had similar hemodynamic effects to dobutamine infusion in our clinically relevant septic shock model in sheep.

Nielsen J S 1 , Rix T 1 , Nybro Sørensen R 1 , Gjedsted J 1 , Krog J 1 , Larrson A 1 , Tønnesen E 1 1 Dept. of Anaesthesiology and Intensive Care Medicine, Aarhus University Hospital, Aarhus C, Denmark

Beneficial effects of Activated Protein C (rhAPC) on mortality and morbidity have been demonstrated in patients with severe sepsis and septic shock. rhAPC has antithrombotic, profibrinolytic, anti-apoptotic and anti-inflammatory properties. The anti-inflammatory properties remains relatively unsettled. Aim: By analysing plasma cytokines we tried to asses the anti-inflammatory effects of rhAPC in a porcine model of acute endotoxemia.

Eighteen female landrace pigs were subjected to anaesthesia and endotracheal intubation. Group I were subjected to general anaesthesia and lipopolysaccharide (LPS) infusion for 5 h. Group II were exposed to LPS and rhAPC for 5 h. After a stabilization period LPS infusion was started at a rate of 2.5 µg · kg-1 · h-1 and increased stepwise to 15 µg · kg-1 · min-1 during the following 30 min's. For the remaining trial period the infusion continued at a rate of 2.5 µg · kg-1 · h-1. The rhAPC infusion was commenced at 100 µg.kg-1.h-1 15 min before the stabilizing period had ended. Blood for cytokine analysis (IL-1, IL-6, IL-8, IL-10, TNF-α) was collected at -15 min, 60, 120, 180, 240, and 300 min. Measurements of APC, thrombin anti-thrombin(TAT) complex, and PAI-1 were obtained at -15 min, 180 min, and 300 min.

Both groups elicited a marked pro-(TNF-α, IL-1, IL-6 and IL-8) and anti-inflammatory (IL-10) cytokine response, but without significant differences between the groups (TNF-α(p=0.49), IL-6(p=0.60), IL-8(p=0.67), IL-10(p=1.00). PAI-1 levels were significantly lower in the rhAPC treated animals (p<0.01). There was a tendency towards decreased TAT levels in the rhAPC treated group, but this was not statistical significant (p=0.058).

No significant effect of rhAPC were seen in plasma cytokine levels. Though PROWESS showed reduced IL-6 plasma levels in septic patients, other human LPS-trials have also showed no effect of rhAPC on plasma cytokines. We showed the pro-fibrinolytic effects of rhAPC as decreased plasma PAI-1 levels. Insignificant TAT-differences could indicate that our study is underpowered or differences in porcine and human coagulationsystems. Conclusion: We did not show any modifying effects of rhAPC on pro-or anti-inflammatory cytokines in this porcine model of acute endotoxemia. Our results suggest that if APC has anti-inflammatory effects these are not elicited through plasma cytokines. 

At least two studies using Prowess data demonstrated in the sepsis patients treated with activated protein C (APC) an early improvement in cardiovascular failure as assessed in this slide Thus, it is possible that a part of APC efficiency is due to specific cardiovascular properties METHODS. We used sedated and ventilated Wistar rats. Monitoring includes the continuous measurements of MAP, abdominal and mesenteric aortic blood flow, muscular laser Doppler and muscle PO2. The measurements also included blood lactate, nitrite and nitrate and TNF. Endotoxin was used at 10 mg/kg and rats were continuously resuscitated with 18 ml/kg/h leading to an hypokinetic model. Three groups were investigated, the first one was treated by 100 µg/kg of APC bolus in the meantime of endotoxin infusion, the second one was treated by 24 µg/kg/h continuously infused and the third by saline.

Continuously infused APC was associated to an improvement in MAP, cardiac output, heart rate, muscular blood flow, muscular PO2 and lactate. Cardiac effects of APC were confirmed using dose response curves with epinephrine and not with phenylephrine. In the APC groups the decrease in CO due to the increase in myocardial afterload was attenuated by APC suggesting that APC improves the answer to beta 1 stimulation as also suggested by a higher heart rate in the APC group.

CONCLUSION. In this experimental model, the association of vasoactive and myocardial effect make APC a unique therapeutic agent. 

Toft P 1 , Schmidt R 1 , Broechner A 1 , Larsen N 1 , Lillevang S 2 , Bollen P 3 , Olsen K 4 1 Intensive care, 2 Immunology, 3 Biomedical lab, 4 Pathology, Odense University Hospital, Odense, Denmark

Plasmapheresis has been used to treat sepsis. Two randomised trials showed no benefit in adult patients with septic shock whereas one randomised trial showed improved survival with plasmapheresis. The effect of plasmapheresis on the cell mediated immunity during sepsis has not been investigated. The aim on the present study was therefore to investigate the long term effect of plasmapheresis on the cell mediated immunity in pigs made septic by endotoxin infusion.

METHODS. 20 pigs were divided into 2 groups. All the pigs were anaesthetized and mechanically ventilated. 10 pigs received 30 microg/kg of Escherichia Coli endotoxin within 20 minutes. 40 minutes later these pigs were treated with plasmapheresis. During plasmapheresis which lasted 4 hours 40 ml/kg bodyweight of the pigs plasma were exchanged with 5% human albumin. 10 pigs received the same bolus of endotoxin and served as a control group. Supportive treatment was continued for 24 hours.

The adhesion molecules CD18, CD44 and CD62L were measured using monoclonal antibodies. The ability to respond with an oxidative burst was measured by means of flow cytometry using 123-dihydrorhodamine. The number of neutrophils was counted in peripheral blood and in lunge tissue. The lymfoproliferative respons and inflammatory cytokines were measured.

The infusion of endotoxin was followed by initial granulocytopenia and later on granylocytosis, activation of CD18, CD62L and increased oxidative burst. The level of cytokines were increased and granulocytes accumulated in lung tissue. Plasmapheresis had no effect at the adhesion molocules (CD18, CD62L), did not influenze on the number of granulocytes and did only slightly reduce the level of cytokines. There was a tendency towards reduced accumulation of granulocytes in the lunges following plasmapherfesis.

Though plasmapheresis was initiated within one hour after the induction of endotoxin induced sepsis it did not significantly attenuate the activated cell mediated immunity in sepsis. 

Perner A 1 , Perner T 1 1 Dept. of Intensive Care, Rigshospitalet, Copenhagen, Denmark

Inhibition of nitric oxide synthases (NOS) may be a therapeutic target in sepsis, but increased mortality was the result of a large clinical trial of a non-specific inhibitor of NOS in patients with septic shock (1) . The aim of the present study was to identify models of sepsis that may predict the outcome of NOS inhibition in patients.

We did a systematic review and meta-analysis of published studies on the effect of pharmacological inhibition of NOS on mortality in animal sepsis. PubMed and Embase were searched and studies of non-specific inhibitors of NOS in sepsis induced by live bacteria were included if mortality data were available. Included studies were graded according to scientific quality.

Eleven studies of a total of 611 animals were included of which 3 were done in large animals and the rest in rodents. Overall the survival rate of sepsis was unaltered by non-specific inhibition of NOS (odds ratio, 95% confidence intervals: 1.1, 0.4 -2.9), but the design of the studies varied considerably and half of the studies had low quality. There was a tendency of increased survival with NOS inhibition in studies of high quality (3.1, 0.8 -11.2) and large animals (6.0, 0.7 -55.9) and decreased survival in studies of rodents (0.7, 0.3 -1.8).

Studies of NOS inhibitory compounds in animal models of sepsis vary considerably in design and quality. This meta-analysis shows that safe conclusions about the predictiveness of a single model for effects in patients cannot be made. Thus extreme caution should be taken if drugs that reduce NO bioavailability are to be tested in septic patients. 

Endothelin-1 (ET-1) plasma levels are increased in sepsis and correlates negatively to cardiac function and survival. The main receptors of the ET system, ETA/ETB, may have differing roles in the myocardial response to sepsis and may mediate certain aspects of myocardial depression. We hypothesized that dual ET receptor antagonism could produce mixed myocardial effects during endotoxemia.

METHODS. 14 anesthetized, mechanically ventilated pigs were subjected to endotoxin-infusion (0.25 mcg/kg/h). Contractile and diastolic function was assessed by left ventricular (LV) pressure-volume analysis (conductance volumetry). After 3 hours of endotoxin, 7 pigs were subjected to tezosentan (TEZO, a dual ET receptor antagonist) at 1 mg/kg/h for 2 hours.

RESULTS. TEZO during endotoxemia caused an increase in cardiac index, decrease in mean pulmonary (MPAP) as well as mean arterial pressure. Left ventricular stiffness (LV end-diastolic pressure/volume) was improved as was isovolumic relaxation (tau). In contrast, systolic function was impaired by TEZO (maximal power/end-diastolic volume, maximal dP/dT/end-diastolic volume, tendency for decrease in preload recruitable stroke work p=0.057)). CONCLUSION. This is the first ex-vivo demonstration of concurrent VP hypersensitivity and NE hyporeactivity in blood vessels taken from a clinically realistic septic animal model. We hypothesize that VP and NE utilise different Ca 2+ mobilisation pathways, and that modulation of these in sepsis is responsible for the observed results. We will therefore use this approach to further examine VP and NE signalling at both tissue and cellular levels. 

Paraforou T 1 , Nikaina I 1 , Paraforos G 1 , Paterakis K 2 , Chovas A 1 , Mpakopoulou M 2 , Tasopoulos K 1 , Komnos A 1 1 Intensive Care Unit, General Hospital of Larissa, 2 Neurosurgery Department, University Hospital of Larissa, Larissa, Greece

To investigate the association between the microdialysis parameters and the outcome in patients with spontaneous intracerebral hematomas METHODS. 30 patients with a mean age of 59 ± 10 years and GCS<=12 were included in this prospective study. Their outcome was evaluated with the GOS in 6 months. The patients were under multimodal brain function monitoring (ICP-CPP and ptiO2 measurements and a microdialysis catheter). The microdialysis samples were collected every two hours and were automatically analyzed.

From the patients included, 22 (73.33%) had favorable outcome (GOS=4,5) and the remaining 8 (26.67%) unfavorable outcome (GOS=1,2,3). The two groups were similar, in terms of patient's age and gender, initial GCS score and therapeutic interventions. The group of patients with the favorable outcome had lower glucose, lactate, lactate/pyruvate ratio and glycerol values, and higher pyruvate values comparing to the group with the unfavorable outcome. None of these differences reached statistical significance, with the exception of the differences in mean lactate and maximum lactate values, which were indicative (table 1) . 

Although microdialysis is a useful tool for the prognosis in patients with traumatic brain damage, its value seems limited when used in spontaneous intracerebral hematomas.

Dias C 1 , Canitrot C 2 , Rey B 2 , Fonseca S 2 , Pimenta C 2 , Leao A 2 , Duraes G 2 , Moutinho R 3 1 Intensive Care, 2 Anesthesia, Hospital S. Joao, 3 Anesthesia, Hospital Gaia, Oporto, Portugal

Recent literature suggests a marked decrease in plasma concentrations of valproic acid (VPA) associated with concomitant administration of carbapenems. This interaction may increase the risk of epileptic seizures in neurocritical care patients.

we retrospectively analyzed the interaction between meropenem and VPA in patients admitted in our NCCU between 2000-2005. During this period we had twelve patients that received concomitante treatment with VPA and meropenem. Seven female and five male with mean age of 59 years admitted with neurocritical diseases needed antiepileptic treatment either because of previous history of epilepsy or recent seizures. During their stay in the NCCU developed Gram negative nosocomial infections with indication for treatment with meropenem.

In all cases serum VPA levels fell imediately after meropenem therapy was started, in spite of increasing VPA daily intake until maximum dose. Patients were monitored with EEG and none of them developed acute seizures during this period. Meropenem seems to inhibit the hydrolytic enzyme involved in the hydrolysis of VPA-glucoronide to VPA, resulting in a decrease in plasma concentration of the active drug.

In concluison it is advisable that high risk patients for developing seizures should be closely monitored with VPA serum levels determination and EEG evaluation. 

Patients in ICU with renal failure requiring treatment with renal replacement therapy (RRT) have poorer survival at hospital discharge and at 1 year, and this effect persists after adjustment for potential confounding factors (1) . However, studies may not have adjusted for all important factors. Furthermore, little is known about the survival of these patients beyond one year. This study aims to evaluate the independent effect of renal failure in ICU patients requiring RRT on short and long-term survival.

A retrospective cohort study of adult patients admitted to a 22-bed general ICU between 1987 and 2002. An ICU clinical database and two administrative databases (Hospital Morbidity Data System and Death datasets) were linked using probabilistic matching. Follow-up time from the index ICU admission was possible for at least 1 year for all patients, and up to 16-17 years (7% patients).

Survival was evaluated at: (1) hospital discharge (logistic regression); (2) 1 year for patients who survived to hospital discharge (Cox regression); (3) subsequent survival, up to 17 years, for all patients alive at 1 year (Cox regression).

The study cohort of 21583 patients had a median follow-up of 6.6 years. Patients (n=915) having RRT in ICU (duration of therapy 1-40 days) had poorer unadjusted survival (54% vs 93% at hospital discharge, 46% vs 86% at 1 year and 33% vs 77% at 5 years). However, patients receiving RRT were sicker with higher severity of illness and comorbidity. Adjustment was made for covariates including age, gender, comorbidity, acute physiology score on admission, number of organ failures, elective surgery, ICU diagnosis, length of stay in ICU, mechanical ventilation, and year group of admission. Administration of RRT was not associated with worse survival to hospital discharge (OR=0.80, 95% CI 0.66,0.99); survival at 1 year, for patients who survived to hospital discharge, (HR=1.05, 95% CI 0.79,1.39); and on subsequent survival for patients who were alive at 1 year (HR=0.92, 95% CI 0.76,1.10).

Hospital and long-term survival is worse in patients who have RRT for renal failure in ICU. However this is a consequence of age, comorbidity, diagnosis and severity of illness, and not an independent effect of renal failure. INTRODUCTION. BNP has been studied as a promising follow-up marker and risk predictor for cardiac surgery (CS) patients (pts); however, its behavior in the postoperative (PO) period of CS is still controversial. The objective this study is to correlate preoperative and PO BNP levels of pts undergoing elective CS with Multiple Organic Dysfunction Score (MODS) in the PO period.

Prospective study with a cohort of 83 CS pts selected between August 2003 and September 2005. Their mean age was 67.0±8.55 years, 23 (27.3%) were females, and the mean Euroscore was 4.0±2.60. BNP was quantitatively measured by use of immunofluorescence (Biosite Triage BNP Test) in the preoperative period (BNPPre), and in the first (BNP1) and sixth (BNP6) PO hours. MODS of the first (MODSD1) and third (MODSD3) PO days was used, the cut-off point being ≥ 3. The statistical analysis comprised Spearman rank correlation, Kruskal-Wallis test and ROC curve.

The mean BNP levels were as follows: BNPPre = 181.0 ± 368.95 pg/mL (MED = 58.7); BNP1 = 156.0 ± 276.9 (MED = 67.2); and BNP6 = 280.3 ± 567.4 (MED = 146.0). The mean MODSD1 was 2.7 ± 1.8 (MED = 2) and the mean MODSD3 was 1.7 ± 2.0 (MED = 1). The Kruskal-Wallis test showed no significance of BNPPre, BNP1 and BNP6 to predict MODSD1 ≥ 3 (0.99; 0.73; 0.19) and MODSD3 ≥ 3 (0.30; 0.21; 0.55). By using the Spearman test, a correlation was observed between BNPPre and MODSD3 (rho = 0.256; p = 0.02). The area under the ROC curve of BNPPre to predict MODSD3 ≥ 3 was 0.57.

In the population studied, BNP was not a good predictor of organic dysfunction in the PO period of CS; a small correlation with MODSD3 was observed. Further studies with more critically ill patients are required. 

Do not attempt resuscitation (DNAR) directives should be an integral part of patient management. There is no published data about DNAR decision-making on high dependency units (HDU). Our objective was to investigate prevalence and factors associated with an explicit DNAR directive in HDU.

Patients admitted to the HDU over 6 months were enrolled. We recorded demographics, severity scores, admission diagnosis, and resuscitation directives. Where no explicit directive was made, an implicit directive for resuscitation was assumed.

. 205 consecutive patients were included; age 62±18 yrs, mean APACHE II 16±6.2 and mean SOFA 4.9±2.8. 67 patients (33%) had directives established, of which 5 (7.4%) were established prior to HDU admission. Directives were established within 24 h of HDU admission (95% CI 0.4-1.6 days). No directives were established by family physicians. Of these 67 directives, 9% were for resuscitation. Independent predictors of an explicit directive are shown in table 1. The strongest predictors were prior functional impairment, previous stroke, and the inability to make an informed decision. 

Our study suggests that DNAR decisions relate most strongly to prior functional state, and thus could be predicted and discussed prior to admission to HDU. After admission, patients for whom these decisions are made, are unlikely to be able to participate, through limited capacity.

Edwards FloTrac TM sensor and Vigileo TM monitor: easy, accurate, reliable cardiac output assessment using the arterial pulse wave

Measuring agreement in method comparison studies

Mechanism of the drug interaction between valproic acid and carbapenem antibiotics in monkeys and rats

Meropenem decreses serum level of valproic acid

3 Department of Clinical Research, Instituto Nacional de Câncer, 4 Department of Hematology

The SOFA (Sepsis-related organ failure assessment) score to describe organ dysfunction/failure

Effect of cardiopulmonary bypass on serum procalcitonin and C-reactive protein concentrations

Perioperative determinants of morbidity and mortality in enderly patients undergoing cardiac surgery

Challenges in end-of-life care in the ICU

Communication of end-of-life decisions in European intensive care units

Sokolowski J 1 , Niewinska K 1 , Jakubaszko J 1 1 Chair of Emergency Medicine, Wroclaw Medical Univercity, Wroclaw, Poland INTRODUCTION. Carbon monoxide (CO) poisoning is a significant problem in Poland. Nonspecific spectrum of symptoms cause that many cases of poisoning are not recognized. CO poisoning is not common in pregnancy and can be easy unrecognized. It may have many adverse effects both to the mother and fetus. CO is fetotoxin and teratogen. Safety and efficacy of HBO2 in the treatment of pregnant women is not well established. The aim of this article is assess safety of HBO2 in the treatment of pregnant women with CO poisoning and the influence of HBO2 on time of delivery and children delivery in 1 year observation period.

Record of patient treated with HBO2 for acute CO poisoning in Hyperbaric Oxygen Therapy Unit, Wroclaw Medical University, from January 2004 to January 2006 were collected and analysed. Symptoms of poisoning, approximate time of exposure, and time from the end of exposure to start HBO2 were analysed. Data of the time of delivery, Apgar scale were collected. Children delivery were assessed using standard protocols for 1 year old children in Poland.

From January 2004 to January 2006 57 patient were treated with HBO2 for CO-poisoning. There were 5 pregnant women in this group. The diagnosis was made on the basis of medical history and symptoms and eleveted COHb level. The mean age of pregnants was 29 years (23-33), mean gestational age -32.25 hbd (22-38hbd) . Symptoms: loss of consciousness in 4 cases, which lasted 3 minutes, 5 minutes, 15 minutes, 1 hour, respiratory insufficiency which demand intubation and mechanical ventilation for over half hour in one case, headache, and retrograde amnesia in all cases. Time from estimated exposure to HBO2 varied from 45 minutes to 4 hours (mean 2.44 hours). The reason of this was: 3 women were transported to our unit from others hospitals. The fetal CTG showed unprovoked decelerations to 100 bpm in 4 cases, and marked bradycardia in 1 case. Breathing with 100% oxygen was administered on the field, and patient were transported to our unit by ambulances. Patients were treated with HBO according to our protocol (1x2.8 ATA-90 minutes, 2x2.5 ATA-90 minutes in the first 24 hours. Then patients were observed in the Department of Toxicology, and consulted by obstetricians every 24 hours (CTG) for 7 days. All patients delivered in term at: 2x39, 2x40, 41 weeks gestation. The Apgar scores were 9 points in 1st minute and 10 points in 10th minutes after delivery in all cases. In one year observation all children have developed promptly.CONCLUSION. 1. HBO2 is the treatment of choice for CO poisoning and all pregnant women suffering CO intoxication should be referred for HBO2 therapy (irrespective to maternal symptoms and maternal COHb level). 2. HBO2 is safe and well tolerated by fetus.

Paramythiotou E 1 , Aggelopoulou H 1 , Kalogeropoulou O 2 , Karakitsos D 1 , Ntagiopoulos P 1 , Karabinis A 1 1 ICU, 2 Pharmacy department, George Genimatas general hospital, Athens, Greece INTRODUCTION. Antibiotics represent one of the most commonly prescribed drug classes in the Intensive Care Unit setting (1, 2) . Several side effects are therefore noticed. Our main purpose was to record the prevalence of side effects of commonly used antimicrobials in our unit.

During a twelve -month period (1/1/2005 -31/12/2005) , all antibiotics side effects which were noticed in patients hospitalized in our 10 bed multivalent ICU were prospectively recorded.

During this year one hundred seventy two patients were admitted in our Unit. There were 88 men and 84 women. Their mean age was 51 ± 20.2 y. Adverse side effects were noticed in 12 patients (7%). Seventy four patients were treated with piperacillin / tazobactam. Four between them (5.5%) presented an eruption. None of them had a history of allergic reactions to any antimicrobial. After the drug was interrupted and antihistamines were given the eruption disappeared. Among sixty eight patients who received imipenem / cilastatin three presented side effects (4.4%). The first patient presented seizures, another one presented neutropenia and the third an allergic skin eruption. In the first case, seizures were attributed to a long duration of therapy while neutropenia in the second case to an overdose of the drug. Finally among 45 patients treated with linezolid, five (11%) presented severe thrombocytopenia (mean PLT number 22.000 / dl). Number of platelets continued to diminish for some time after the drug interruption. Thrombocytopenia appeared after a mean period of 8.8 ± 4.6 days. Platelets returned to normal slowly after an average of 12 days.CONCLUSION. ICU physicians must remain aware of the fact that adverse side effects of antimicrobial drugs are not very rare. Attention must be paid so that the responsible drug is interrupted on time and more severe consequences are avoided. 

Connelly K A 1 , Gonzalez I M 1 1 Intensive Care, The James Cook University Hospital, Middlesbrough, United Kingdom

Recent publications have suggested sub-optimal ward management of patients referred to critical care services (1, 2) which is associated with an increased mortality (1) . Our audit looked at the adequacy of resuscitation and antibiotic therapy prior to admission to ICU. We felt that patients were often being admitted late to ICU and had been sub-optimally managed on the wards. Our aim was to quantify the extent of this problem.

This was a prospective audit on the ICU in a large teaching hospital. The notes of every emergency admission directly from the ward over a six months were reviewed. The ward care prior to involvement of the critical care team was assessed by a questionnaire focusing on the recognition and management of clinical deterioration. We aimed to objectively assess resuscitation in terms of airway, breathing and circulation. Medical management was considered optimal if basic therapies appropriate to their underlying illness had been instituted. Where appropriate antibiotic therapy should have been commenced.

The notes of 56 patients were audited. A summary of the adequacy of resuscitation is presented in Table 1 . Basic medical management was considered adequate in only 68% of patients. Antibiotics were considered appropriate in 73%. Only 90% of these had antibiotics prescribed and only 22% had been written up for a stat dose. There was a discrepancy of between 1 to 4 hours between the prescribed dose and administered times. 

Our findings confirmed that resuscitation, management and antibiotic therapy before ITU admission are often sub-optimal. Reasons for this are lack of recognition of clinical deterioration, failure to start appropriate therapy and lack of senior medical input. Although there is an element of subjectivity in such assessment our results support recent findings (1, 2) . More education of ward staff in the recognition and management of acute illness is required. 

Rundgren M K 1 , Karlsson T 2 , Nielsen N 3 , Johnsson P 4 , Fribeg H 1 1 Anaesthesia and Intensive Care, Lund University Hospital, Lund, 2 Anaesthesia and Intensive Care, Malmoe University Hospital, Malmoe, 3 Anaesthesia and Intensive Care, Helsingborg Hospital, Helsingborg, 4 Cardiothoracic surgery, Lund University Hospital, Lund, Sweden

Early prediction of neurological outcome after cardiac arrest is essential to avoid futile intensive care. Induced hypothermia as a post cardiac arrest neuroprotective strategy demands sedation, making a clinical neurological exam unreliable. We conducted a prospective trial of the neurobiochemical markers S-100b and NSE as tools for predicting neurological outcome in hypothermia treated cardiac arrest patients.METHODS. NSE and S-100b were measured in serum at 2+/-1h, 24+/-4h, 48+/-4h and 72+/-4h after cardiac arrest in patients with mixed initial rhythms and different locations of cardiac arrest. All patients were treated with induced hypothermia (33 • C for 24h and a slow rewarming 0.5 • C/h). After sedation was withdrawn the patients were evaluated neurologically. In patients who remained deeply unconscious (GCS 3-4) 72h after the time point when normothermia was reestablished, active treatment was withdrawn. All patients were evaluated according to the Cerebral Performance Categories (CPC) scale when leaving the ICU, before leaving hospital and 6 months after cardiac arrest. A CPC of 1-2 at any time was considered a good outcome.

. 104 patients, 86 in whom NSE were evaluated, were included from August 2002 until August 2005. The patients had a mean age of 63 years, 70% were male. 84% had out-of-hospital cardiac arrests; the initial rhythm was VF/VT in 62%, Asystole/PEA 35% and unknown in 3% of the patients. 47% of the patients were classified as having good outcomes. With an S-100b cut off of 0.55 at 24h we reached a specificity of 93% with a sensitivity of 53% for bad outcome. NSE values of 24 (24h) and 27 (48h) resulted in a specificity of 100% and a sensitivity of 49% and 70% respectively. The early values of S-100b showed a large range independent of good (0.11-3.1) or bad (0.38-11) outcomes.CONCLUSION. The best markers for neurologic outcome in this mixed group of hypothermia treated cardiac arrest patients were an S-100b level of 0.55 at 24h or a NSE of 27 at 48 h. Low levels of S-100 at any time should be interpreted cautiously due to low sensitivity, and initial S-100 levels were unreliable from a prognostic point of view.

Coskun R 1 , Guven M 1 , Sungur M 1 1 Internal Medicine and Intensive Care Unit, Erciyes University Medical Faculty, Kayseri, Turkey INTRODUCTION. Organophosphate (OP) insecticides inhibit both acetylcholinesterase and pseudocholinesterase activities. The clinical course of OP poisoning may be quite severe and may need intensive care management. We report our experience with the intensive care management of serious OP insecticide poisonings.

A retrospective study was performed on the patients with OP poisoning followed at our medical intensive care unit. Sixty-one patients were included. Diagnosis was performed from the history taken either from the patient or from the patient's relatives regarding the agent involved in the exposure. Intravenous atropine and pralidoxime was administered as soon as possible. Pralidoxime could not be given to 18 patients: 2 patients did not receive pralidoxime because they were late admissions, and 16 did not receive pralidoxime because the Ministry of Health office was out of stock. Data are presented as mean ± standard deviation.RESULTS. There were 32 female and 29 male patients. Fourty (65.57%) were suicide attempts and 21 (34.43%) were accidental exposure patients. The mortality rates for the patients who did and did not receive pralidoxime were 30.2 and 16.7%, respectively and were not statistically different. Sixteen patients (26.23%) required mechanical ventilation. The mortality rate for the patients who required mechanical ventilation was 50%, but the rate was 17.8% for those patients who were not mechanically ventilated. Intermediate syndrome was been observed in 11 (18.03%) patients. Complications were observed in 40 (74.4%) patients. The duration of the intensive care stay was 5.80 ± 3.37 days.CONCLUSION. OP insecticide poisoning is a serious condition that requires rapid diagnosis and treatment. Because respiratory failure is the major reason for mortality, careful monitoring, appropriate management and early recognition of this complication may decrease the mortality rate among these patients. 

There has been an increasing concern among intensive care physicians with timely diagnosis and intervention in critical illness. The interface Emergency-Intensive Care Unit is in focus as an area for possible improvement. Hospital of S. João is a tertiary referral academic center with more than 160000 admissions a year to the Emergency Department. Of these, a small proportion is critically ill or trauma patients and enters in a 5-bed space called "Emergency Room" (ER) where are also admitted hospital inpatients with clinical deterioration. An Intensivist from the Intensive Care Service works in the Emergency Department in a weekly-based Rota with responsibility for directing the overall plan of care for these patients, for triage decisions and intensive care allocation. The purpose of this study was to assess whether care provided by an intensivist to critically ill patients in the ED has a positive impact in the process of care in terms of time-effectiveness and diagnostic and treatment expertise.

Retrospective study of admissions to the ER from January to December 2005 as registered in the written log. Data collected included: origin of the patient, length of stay, and destination in terms of admission to Intensive Care Units, special care units, operating room, general ward, death or discharge.

In 2005 there were 2789 patients entering ER (of those, 513 where major trauma patients and were assisted by the Trauma Team). The other 2276 patients included 1437 (63, 1%) coming from the community; 324 (14, 2%) were inter-hospital transfers; 359 (15, 7%) came from in-hospital wards and 156 (6.8%) from Intermediate Care Units. Patients were assessed regarding indication for ICU admission. When life support treatment was withheld, strategies for palliation of symptoms were implemented on site. The mean length of stay was 5.8 hours. It's possible that this space functions as a buffer to lack of beds in either intensive or intermediate care units expending resources and leading to overcrowding. Only 11% of the patients (384) were admitted to Intensive Care Units. Opportune and efficient treatment may have averted many ICU admissions. Patients admitted to Intermediate Care Units -844 (37%); to general wards -336 (14.7%); to operating theatre -135 (5.9%). Two hundred ninety eight patients (13%) died.

Having an intensivist in the front line of admission of critically ill patients is highly cost-effective. It extends the ICU model of care to the ED with focus in the level of care the patient needs regardless the location he is.

Escudero D 1 , Otero J 1 , Muñiz G 1 , Tenza E 1 , Rodríguez P 1 , Forcelledo L 1 , Quindós B 1 , Taboada F 1 1 ICU, Hospital Universitario Central de Asturias, Oviedo, Spain

Objetivo: To evaluate the BIS (Bispectral Index Scale) monitor as a method of brain death (BD) detection.

Observational prospective study. Intensive care unit (ICU) of a university hospital. Patients hospitalized in a non consecutive way in the intensive care unit, with a serious neurological pathology and evolution towards a brain death. A BIS monitor, XP model and the sensor "BIS Quatro" were used. The BIS values were continuously recorded: suppression ratio (SR), quality of the signal index (QSI) and electromyographic (EMG) activity.

. 33 patients (19 male). Ages ranged from 34 to 76 years (mean 62, 6). Aetiology of BD: 7 traumatic brain injury, 9 intracerebral hemorrhage, 13 subarachnoid hemorrhage and 4 stroke. E. Glasgow mean on admission in ICU was 4.9 (3-13). The 4 patients with stroke NIHSS Scale of 30, 21 and 19. The BD diagnosis was made through neurologic clinical exploration and electroencephalogram (EEG) in all the cases. Additionally, the transcranial Doppler (TD) was used in 23 patients (70%). Coinciding with the clinical worsening, it was observed that there was a gradual decrease of the BIS value, together with a raise in the SR. In all the patients in which the BD diagnosis was confirmed, the BIS showed values of 0 and suppression rates of 100. Only one patient showed interferences, due to electromyographic activity, being the same problem detected when doing a conventional EEG. After using a neuromuscular blocker, the values of BIS and SR were of 0 and 100 respectively. 1 patient, with BIS of 0, and atrial fibrilattion, during apnea test, presented a cardiac frequency of 170 bpm and increasing of BIS until values of 20. In 3 cases, with BIS of 0 the cough reflex persisted between 30 minutes and 5 hours.

The BIS is a non invasive method, simple and easy to interpret. All the patients with DB diagnosis, except for one, had a BIS value of 0 and TS of 100, showing a perfect correlation with the other methods of diagnosis used. False negatives transient can be observed by electromiographic interference and electrocardiogram artifact. The BIS cannot be used on its own for the confirmation of the BD, but it is a very useful tool in order to detect the beginning of brain herniation.

Azim A 1 , Kumar A 1 , Baronia A K 1 , Gurjar M 1 1 Critical care Medicine, Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow, India INTRODUCTION. Dyselectrolytemia is common in hospitalized patients and hyponatremia (Serum Na <135 meq/l) is the commonest electrolyte disturbance encountered in critically ill patients. An association between osmotic demyelination syndrome (ODS) and the rapid correction of sodium in hyponatremic patients was established almost three decades back. We present a retrospective analysis of ten patients with the diagnosis of central pontine myelinolysis (CPM) and extra-pontine myelinolysis (EPM).

A retrospective analysis of medical records of ten patients diagnosed with ODS was done during their stay in our intensive care unit (ICU) over a period of three years. We reviewed the clinical features, etiology and clinical outcome of these patients RESULTS. Study included ten patients ranging in age from 12-71yrs (mean age-32yrs) with male female ratio of 4:6. Worsening of sensorium was the most common presenting neurological symptom (40%) at the time of admission to our ICU followed by new onset of alteration in sensorium (20%), seizures (20%) and generalized motor weakness (20%). Mean admission APACHE II was 12. Malnutrition (30%) and prolonged diuretic therapy (30%) were the most common predisposing factors. All patients were of chronic hyponatremia (>48hrs) with serum sodium less than 120 meq/l. In all patients the rate of correction of sodium was more than 10 meq/l/24 hours. Hypokalemia was the most common associated biochemical abnormality (60%) followed by hyperglycemia (20%), hypophosphatemia (20%) and elevated serum creatinine (20%). MRI was done in all patients after two weeks of first detection of hyponatremia. Imaging revealed CPM in 5 patients, EPM in 3 patients and both CPM and EPM in 2 patients. Only two patients had complete neurological recovery at ICU discharge (22%). Two patients were discharged in vegetative state, 2 patients died during the ICU stay due to intercurrent complications and 4 patients were discharged with moderate neurological recovery.

Osmotic demyelination syndrome is a complication of treatment of patients with life threatening hyponatremia. Outcome remains poor as majority (75%) of survivors in our series remained dependent for self care even after six months. Length of ICU stay and hospital stay were prolonged Recognizing the patient at risk (malnutrition, associated illnesses, patients requiring aggressive fluid therapy); preventing rapid correction of hyponatremia (>8 meq/l/24 hours); optimal correction of associated biochemical abnormalities especially concurrent hypokalemia and good holistic intensive care can reduce the incidence and improve the outcome of this syndrome INTRODUCTION. Neuromuscular abnormalities are common in critically ill patients with systemic inflammation and organ failures. For the diagnosis of critical illness polyneuropathy (CIP), electrophysiological tests are recommended but frequently not available. We aimed to assess the incidence of clinically diagnosed CIP and its potential impact on length of hospital stay and mortality.METHODS. 30 consecutive critically ill patients on mechanical ventilation for 48 hours and with the presence of greater than or equal to 2 SIRS criteria were prospectively studied. Clinical neurological examinations were performed daily during sedation stop. Clinical diagnosis of CIP was defined as symmetric limb muscle weakness with no explanation other than CIP (e.g. myasthenia gravis, Guillain-Barré syndrome, prolonged effect of sedation) in patients with normal neurology at ICU admission.

. CIP was diagnosed clinically in 11 patients (37%). 10 patients (33%) had sepsis and 5 of them also had CIP (50%). shown as the elective one, but the safety profile and the timing to perform it have still to be defined. This group is very sensitive to changes in the intracranial pressure (ICP), cerebral perfusion pressure (CPP) and ventilation. This study tries to define these variations as well as the role of the association of fever (F), use of vasoactive drugs (UVD), tracheal cuff puncture (TCP), endoscopy (E) and ventilatory mode (VM) during the procedure.

We studied prospectively 64 neurocritical patients under monitorization of the ICP, invasive mean blood pressure (MBP), arterial oxygen saturation (SaO2) and end-tidal pCO2 (pCO2et). The method of Ciaglia with the succesives technological improvements of the percutaneous dilational tracheostomy (PDT) was chosen, associating endoscopic view occasionally (just to verify the appropriate position of the wire) in the patients with difficult anatomic references. The patients were under general anaesthesia and MV with FiO2:1 and PEEP<10 cm of water. The moment to perform it was designed by consensus once the patient was hemodinamically and respiratory stable and without intracranial hypertension crisis. The variables are expressed as mean and standard deviation. 

CCTs underwent internationally-recognised training in the insertion of ICP monitors. Following this, CCTs undertook local, supervised training and assessment. Time from request to ICP monitor insertion was recorded, along with number of attempts. Notes were reviewed for 7 days post insertion for any complications. Supervision of junior doctors was performed by CCTs using a standardised training programme and competency-based assessment tool.

. 165 consecutive ICP monitor insertions were audited over a 5 month period. 145 of these were considered urgent. No statistical differences were noted in the time taken to insertion between operators. No episodes of local infection or intracranial bleeding were recorded. 

Cerebral autoregulation (CA) monitoring might enable us to optimise the therapy of severe traumatic brain injury. The non-invasive cross-correlation method uses slow spontaneous oscillations of arterial blood pressure (ABP) and middle cerebral artery flow velocity (FV) for continuous measurement of CA. The aim of this study was to evaluate the cross-correlation method in a pig model of acute subdural hematoma coupled with autoregulatory disturbance. The static rate of regulation (sROR, [1] ) was used to confirm the status of CA.

After approval by the local ethics committee 12 male pigs (30 kg) were anesthetized, intubated, mechanically normoventilated and craniectomized. We continuously measured ABP, intracranial pressure, and bilateral FV by Doppler sonography. At baseline (intact CA), the time delay at the positive maximum of the cross-correlation function between low-pass filtered (0.1 Hz) ABP-and FVoscillations was calculated and averaged over 10 minutes. For the calculation of sROR [1] , ABP was elevated about 20% of baseline values by continuous infusion of arterenol. After recovery of ABP, a unilateral subdural hematoma was induced by autologous blood injection. To reach autoregulatory failure, the injection was continued until cerebral perfusion pressure persistently remained below 50mmHg. Calculation of the time delay and sROR was repeated. Wilcoxon tests were used for statistics.

The trauma led to significant disturbance of CA (mean sROR before trauma 1.0 ± 0.4, afterwards 0.2 ± 0.4 (p<0.01)). Cross-correlation analysis detected failure of CA (mean time delay before trauma -2.9 ± 1.4s, afterwards +0.5 ± 0.7s) and significantly correlated with the sROR (Spearmans rho -0.58, r 2 = 0.34, p<0.01).CONCLUSION. The applied model is useful to study CA measurement methods. Cross-correlation analysis can reliably distinguish between an intact and impaired CA. As the method is non-invasive, it is suitable for clinical monitoring. 

Pulse pressure variation has been proposed as a means of assessing fluid responsiveness in critically ill patients. However, this index may be sensitive to ventilatory settings. We hypothesized that pulse pressure variation may decrease at high respiratory rates.

We investigated six patients ventilated in volume control mode with a tidal volume higher than 8 ml/kg at a respiratory rate (RR) lower than 15/min and with a pulse pressure variation higher than 10%. Cardiac index was measured in all patients, either with a pulmonary artery catheter (n=5) or with a PiCCO system (n=1). Cardiac index, mean arterial pressure (MAP), pulse pressure variation and the ratio between heart rate and RR (HR/RR) were measured at baseline (RR of 14/min), after increasing RR to 30/min and then to 40/min, and again at 14/min after volume expansion with 1000 ml of crystalloids. The inspiratory to expiratory ratio was kept constant during the entire procedure. Data are presented as median [percentiles 25-75]. We applied a Friedman test followed by a Wilcoxon rank test with Bonferroni adjustment for multiple comparisons.

Plateau pressure, total positive end-expiratory pressure (PEEP) level and transpulmonary pressure remained constant throughout the experiment. The evolution of the principal variables is shown in the table. 

We included spontaneously breathing patients with acute circulatory failure for whom the attending physician decided to perform a fluid challenge. At baseline, during PLR, before and after fluid challenge (500 mL NaCl 9% in 15 minutes), the stroke volume was calculated as the product of aortic valve area by the velocity time integral of aortic blood flow (VTIAo). At baseline, we also obtained two static markers of preload: the left ventricular end-diastolic area (LVEDA) and the ratio of mitral E wave velocity to the early diastolic velocity of the mitral annulus (E/Ea), which was measured by tissue Doppler imaging.

Twenty patients (65 ± 16 years old) were included. The main origin of circulatory failure was septic in 15 cases and non septic in 5 cases. At baseline, mean arterial pressure was 64 ± 11 mmHg, indexed stroke volume (SVi) was 39 ± 13 ml/m 2 , and heart rate was 98 ± 22 beats/min. The increase in SVi after fluid challenge correlated with the increase of SVi during PLR (r 2 =0.49, p=0.0006) but not with baseline LVEDA (p=0.8) and E/Ea (p=0.9). In the 15 patients who increased their SVi by more than 10% after fluid challenge (responders), SVi increased by 21 ± 11% (p<0.0001) after PLR and by 25 ± 9% (p<0.0001) after fluid therapy. In non-responders (n=5), SVi did not change after PLR. For predicting fluid responsiveness, an increase in SVi after PLR above 8% had a sensitivity of 100% and specificity of 100%.

The PLR-induced change in stroke volume measured by transthoracic echocardiography appeared as a good indicator for predicting fluid responsiveness in patients with spontaneous breathing.

Ridel C 1 , Lamia B 1 , Monnet X 1 , Anguel N 1 , Osman D 1 , Hamzaoui O 1 , Richard C 1 , Teboul J L 1 1 Réanimation médicale, EA4046, CHU Bicêtre, Université Paris 11, Le Kremlin-Bicêtre, France

The aim of this study in patients with spontaneous breathing or arrhythmias was to investigate whether the response of pulse contour cardiac cardiac output to passive leg raising (PLR) could distinguish responders to volume expansion from nonresponders.

Fifty two patients with circulatory failure were included. Arterial pressure variation could not be used for predicting fluid responsiveness because of spontaneous breathing (n=44) and/or arrhythmias (n=10). The pulse contour cardiac index (CI) was monitored by a PiCCO device (Pulsion, Germany). At baseline, during PLR and after fluid infusion (500 mL saline in 15 minutes), CI and arterial pressure were measured. The global end-diastolic volume (GEDV), a static marker of preload, was also obtained at baseline.

The PLR-induced increase in CI correlated with the fluid-induced increase in CI (r=0.62, p<0.0001) but not with baseline GEDV (p=0.2). There was a significant correlation between the rate of increase of pulse pressure during PLR and that after fluid challenge (r=0.40, p=0.003). The increase of CI after fluid challenge above 15% defined responders. At baseline, cardiac index was 3.1±1.0 L/min/m 2 in responders and 3.0±0.9 L/min/m 2 in nonresponders (p=0.8), GEDV was 720±206 mL/m 2 in responders and 776±273 mL/m 2 in nonresponders (p=0.3). In responders (n=27), CI increased by 25±20% (p<0.001) during PLR and by 34±20% (p<0.001) after fluid challenge while the pulse pressure increased by 16±17% (p<0.001) during PLR and by 24±28% (p<0.001) after fluid challenge. In nonresponders (n=25), neither CI nor pulse pressure changed during PLR. An increase in CI during PLR > 12% predicted fluid responsiveness with sensitivity (Se) of 70% and specificity (Sp) of 92%. The predictive value of PLR-induced increase in pulse contour CI was significantly better (area under the ROC curve = 0.80±0.06) than that of the PLR-induced increase in pulse pressure (threshold: 11%, Se: 59% and Sp: 80%; area under the ROC curve = 0.68±0.07).

The short-term changes in pulse contour CI induced by PLR provide a good prediction of volume responsiveness in patients with spontaneous breathing and/or arrhythmias.

Dubin A 1 , Ferrara G 1 , Pozo M O 1 , Murias G 1 , Martins E 1 , Canullán C 1 , Canales H S 1 , Ince C 2 1 Cátedra de Farmacología, Facultad de Ciencias Médicas, Universidad Nacional de La Plata, La Plata, Argentina, 2 Clinical Physiology, Academic Medical Center, University of Amsterdam, Amsterdam, Netherlands

Our hypothesis was that changes in microcirculation are earlier indicators of hypovolemia than commonly used variables.METHODS. 8 anesthetized sheep were stepwise bled (5, 10 and 15 ml/kg) at 30' intervals. Ileal intramucosal-arterial PCO2 (∆PCO2), mean arterial blood pressure (MAP), cardiac output (CO) and intestinal blood flow (Qgut) were measured. Sublingual and ileal mucosal microcirculation were evaluated by sidestream darkfield (SDF) imaging. Microvascular flow index (MFI) was determined (1) .

The change in ∆PCO2 was correlated with mucosal MFI (r2 = 0.45, p < 0.0001). 1.9 ± 0.6* 82 ± 10 1 ± 3 2 . 5± 0.6* 5 ± 4 2 . 7± 0.3* 2.3 ± 0.3* 10 ml/kg 1.6 ± 0.3* 71 ± 17 -1 ± 5 4 . 0± 1.5* 8 ± 4* 2.6 ± 0.2* 2.0 ± 0.4* 15 ml/kg 1.4 ± 0.3* 60 ± 17* -4 ± 5* 5.8 ± 2.2* 17 ± 15* 2.3 ± 0.2* 1.7 ± 0.4* *p < 0.05 vs. basal.

Before changes in MAP, BE or ∆PCO2 could be detected, there were significant alterations in sublingual and intestinal microcirculation and in lactate levels. SDF imaging might contribute to early detection of perfusion deficits. REFERENCE(S). 1. Boerma E C, Mathura K R, van der Voort P H, Spronk P E, Ince C. Quantifying bedside-derived imaging of microcirculatory abnormalities in septic patients: a prospective validation study. Crit Care 2005;9:R601-6.

Haider C 1 , Kainz C 1 , Schnoell F 2 , Bolzano K 3 , Pichler M 4 1 Neurorehabilitation, 2 Cardiology, Rehabilitation Center Großgmain, Großgmain, 3 Hemodialysis, 4 Cardiology, SALK, Salzburg, Austria

Echocardiographic assessment of diastolic function respectively dysfunction remains difficult under variable loading conditions, changes of preload during intensive care are frequently documented, presence of diastolic dysfunction however contributes to congestive heart failure and has an impact on choice of drugs.METHODS. 16 patients on hemodialysis (HD), 10 male, median age 54 years SD13, exhibiting one or more Doppler echocardiographic signs of diastolic dysfunction: isovolumic relaxation time (IVRT)<100ms, peak velocity E< 100cm/s, Deceleration time>240ms or<150ms, E/A ration<1.1. Patients were investigated immediately before/after hemodialysis as a model for rapid preload reduction without apparent influence on myocardial contractility by pulsed wave and continous wave transmitral velocity profile.

Hemodialysis with a median weight reduction of 2.5kg (range 0.5-3.2) in all patients, left ventricular ejection fraction 55% (28-62), left ventricular muscle mass 141(89-213) g/m2. Heart rate before HD 74/min SD13, after HD 77/min SD13, systolic blood pressure before HD 151mmHg SD23, after HD 138mmHg SD25, diastolic blood pressure before HD 80mmHg SD6, after HD 77mmHg SD9. Doppler echocardiographic parameters before/after hemodialysis: peak velocities of E and A wave, deceleration time and isovolumetric relaxation time were significantly altered by different loading conditions. IVRT before HD 100ms SD15, after HD 120ms, p=0.0002. Peak velocity E wave 97cm/s before 78cm/s after HD, p=0.0006. Deceleration time before 178ms after HD 244ms, p=0.0001. Acceleration time before 58ms after HD 56ms, p=0.6, non significant.CONCLUSION. Patholologic reduced acceleration time as a parameter of left ventricular diastolic dysfunction seems to be a Doppler echocardiographic parameter stable against short term changes in loading conditions. 

Yepes D 1 , Bejarano J 2 , Panesso R 3 1 Intensive Care Unite, Clinica Medellin, 2 Intensive Care Unit, Clinica Bolivariana, 3 Intensive Care Unit, Clinica Las Americas, Medellin, Colombia INTRODUCTION. For many years, pulmonary artery catheterization has been a method of choice for assessing preload at the bedside in critical care units. Despite many recent concerns about its use, and the lack of evidence for benefits in medical trials, it is estimated that nearly a million pulmonary artery catheters are used per annum in the USA. Hemodynamic changes associated with mechanical ventilation have been proposed as predictors of preload in critical care patients. The Valsalva maneuver has been described in previous trials as a useful tool for assessing preload; it appears to be a safe procedure, but it has not been tested in seriously ill patients.

We enrolled mechanically ventilated critical care patients who had pulmonary artery catheters and arterial lines for various clinical conditions, hospitalized in Colombian critical care units between August 2004 and June 2005. Local staff-trained personnel measured pulmonary artery occlusion pressure (PAOP) through the pulmonary artery catheter, and we compared the results with the PAOP estimated by the Valsalva maneuver. For statistical analysis, we used SPSS Version 13.

We included 120 patients in the trial, comprising 3 categories: (1) 72 patients with shock, (2) 21 surgical patients, (3) 27 patients with miscellaneous conditions, described as 'others'. PAOP was measured directly through the pulmonary artery catheter and estimated by the Valsalva maneuver. There was a statistically significant correlation between the two measurements for the surgical patients (r2 =0.518, p<0.05) and the 'others' (r2=0.621, p<0.001). We found no statistically significant correlation for the shock group or for the overall patient population.

Pulmonary Artery Oclussion Pressure (PAOP) estimated by the Valsalva maneuver is a safe procedure for use in mechanically ventilated critical care patients. In surgical patients and patients who are not in shock, there is a good correlation between the PAOP measurements taken from a pulmonary artery catheter and estimated by the Valsalva maneuver. 

Lambert P 1 , Sloth E 2 , Smith B 3 , Kjaersgaard Hansen L 1 , Koefoed Nielsen J 1 , Larsson A L S 1 1 Anesthesiology and Intensive Care, Aalborg and Aarhus University Hospitals, Aalborg, 2 Clin Inst, Aarhus Univ Hosp,Århus, 3 MMDS, Aalborg University, Aalborg, Denmark

It has been proposed that the respiratory variations in the diameter of the inferior vena cava (IVC) indicate fluid responsiveness (1) . We hypothesized that application of positive endexpiratory pressure (PEEP) by hampering venous return and increasing the volume in the capacitance vessels would decrease the variations even if fluid responsiveness exists.METHODS. 8, 20-22 kg, anesthetized, muscle-relaxed and mechanically ventilated pigs were subjected to 0.10, and 20 cmH2O of PEEP at 10% (of estimated blood volume) hypovolemia, normovolemia, and 10% hypervolemia. Hypovolemia was achieved by venesection and hypervolemia by infusion of a starch-solution. Subcostal echocardiography was performed and a cross section image of the IVC, covering one fully respiratory cycle, was recorded and stored in digital cine-loop format. The area together with the anterior-posterior (APD) and left-right (LRD) diameter during end-expiration and inspiration were measured. Maximum and minimum DIVC values over a single respiratory cycle were collected and the DIVC variation (DDIVC) calculated as the difference between the maximum and the minimum DIVC value, divided by the minimum value and expressed as a percentage (1) . Statistics: Kruskal Wallis and Wilcoxons tests, p<0.025 was considered significant. Median and 25.75% are shown.

All measures, i.e., variations in APD, RLD and areas followed the same pattern. The APD values are shown. At hypovolemia and normovolemia, DDICV% decreased significantly by application of 10 and 20 cmH2O PEEP (from 44 (15.56)% to 7 (2.10)% and 4(2.10)% at hypovolemia, and from 9 (6.25)% to 5 (4.15)% and 2 (0.5)% at normovolemia) while at hypervolemia was no significant changes occurred (from 12 (5.13)% to 12 (10.15)% and 6 (4.9)%).

The respiratory variations in the diameter of the inferior vena cava were significantly reduced by application of PEEP suggesting that this method might not be useful for assessment of fluid responsiveness during mechanical ventilation with moderate or high PEEP. 

Chevreuil C 1 , François B 1 , Pichon N 1 , Clavel M 1 , Dugard A 1 , Ait Hssain A 1 , Frat J 2 , Gastinne H 1 , Vignon P 1 1 Medical-surgical Intensive Care Unit, Dupuytren Teaching Hospital, Limoges, 2 Medical Intensive Care Unit, Jean Bernard Teaching Hospital, Poitiers, France

The hemodynamic criterion of ARDS relies on the measurement of a pulmonary artery occlusion pressure (PAOP) less than or equal to 18 mmHg, which is usually obtained by pulmonary artery catheterization (PAC). We evaluated the ability of transesophageal echocardiography (TEE) to predict PAOP in ventilated patients.METHODS. 82 patients were studied (mean age: 63±14 years; SAPS 2: 46±12). All patients were in sinus rhythm and underwent both a TEE and CAP. Invasive PAOP measured with CAP was used as reference. In each patient, the following conventional Doppler parameters were measured (mean of 3 end-expiratory measurements): mitral Doppler E/A ratio, pulmonary vein Doppler S/D ratio and systolic fraction (SF = VTI S/VTI S+D). In the last 26 patients, new Doppler parameters were also measured: early diastolic blood flow propagation velocity with color M-mode (Vp), and early tissue Doppler imaging (TDI) velocities of the lateral aspect of the mitral ring (Ea), with E/Vp and E/Ea ratios. Inter-and intra-observer variability in measured parameters was 1 to 13% and 2 to 7%, respectively. ROC curves were determined for each Doppler parameter to predict a PAOP less than or equal to 18 mmHg, and areas under curve (AUC) were compared.

All Doppler parameters were significantly correlated with invasive PAOP. Although AUC were similar between Doppler indices to predict PAOP, new Doppler parameters yielded highest values (table) .

Compelling experimental evidences have suggested that females may have a more active immune system offering them a better protection against severe infection. This would be related to sex hormones differences. However clinical studies are far more controversial. We hypotheses that these discrepancies may be related on the population studied: severe sepsis, shock septic, and hormonal status (pre or postmenopausal).

We studied 1692 severe sepsis patients from the 4860 patients included in the Outcomerea database ® over an 8-year period. We first determined variables independently related to hospital mortality using a logistic regression method. Based on these variables, we developed a propensity score for mortality and used it to carefully match women with men. We evaluated the sex gender's influence on the outcome on severe sepsis, septic shock through the use of a conditional logistic regression. Subgroups analyses were performed on the age before <50 year (pre-menopausal) or after 50 year (post-menopausal).

We matched 1000 men with 608 women with severe sepsis; the hospital mortality was lower in women (OR: 0.75 [0.657-0.97], after adjusting for confounding variables, p=0.02). This effect was even more significant in older women (>50 year old) vs. men (OR: 0.69 [0.52-0.93, p=0.014]). Interestingly, we did not observe any difference in premenopausal (<50 year old) versus men (OR: 1.01 [0.52-1.97, p=0.98]). No sex gender differences were unmasked when considering only patients with septic shock (n= 310 vs. n=491, respectively): there was no difference in outcome in this subgroup of patients (OR: 0.92 [0.65-1.3]).

Women with severe sepsis have a lower risk of hospital mortality than men. This better chance of survival was more pronounced after the age of 50 (postmenopausal). Furthermore, there was also no difference in hospital mortality in the subgroup of patients with septic shock, irrespective of the hormonal status.

Outcomerea is supported by nonexclusive grants from Aventis Pharma, Wyeth, and the Centre National de la Recherche Scientifique (CNRS). 

The objective was to study frequency, prognostic factors and outcome of critically ill patients with infection and to evalaute sepsis criteria in this cohort of critically ill patients.

Substudy of a prospective multicentre, multinational cohort study. A total of 16,784 patients consecutively admitted to 303 intensive care units (ICUs) from 14 October to 15 December 2002 (SAPS 3 database) were studied. Details about the cohort and methods have been reported elsewhere [1, 2] .

Out of the whole cohort, 7,566 patients (45%) presented with SIRS. 3,505 patients (21%) presented with an infection already at ICU admission. Generally, infected patients were older (65 vs. 63 [49-74] years, p < 0.001), exhibited more often female gender (42 vs 39%, p<0.001) and were significantly more often admitted unplanned (86 vs. 59%, p<0.0001). With respect to the surgical status, infected patients were more often admitted after acute surgery. Severity of illness as measured by the SAPS 3 Admission Score was significantly higher in infected patients (62 [52-72] vs. 44 ), as were hospital mortality rates (41.8 vs 18.7%). ICU length of stay was significantly increased in infected patients (5 [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] vs. 2 [1] [2] [3] [4] [5] days). Out of the infected patients, 704 presented with infection, 190 (5%) with sepsis, 1768 (50%) with severe sepsis and 843 (24%) with septic shock. Hospital mortality rates increased from 32% (infection) to 58% (septic shock).

Infection was present in 21% of admitted patients and was associated with increased ICU length of stay and increased ICU and hospital mortality rates. 

Croitoru M 1 , Croitoru S 2 , Krimerman S 1 , Barmeir E 3 , Dor Y 4 , Altman E 5 1 ICU, 2 Imaging, 3 MAR Institute, 4 General Surgery, Bnai Zion Medical Center, Haifa, 5 Thoracic Surgery, Western Galilee Hospital, Nahariya, Israel INTRODUCTION. Tracheostomy (surgical and percutaneous) is a well-established procedure for prolonged mechanical ventilation in the critically ill adult patient. Fiberoptic laryngotracheoscopy (FLT) is used in the evaluation of the trachea following decannulation but it involves discomfort. Today it is possible to perform virtual endoscopy by multislice CT within seconds even in debilitated patients. Our aim was to confirm the value of virtual laryngo-tracheoscopy (VLT) in the diagnosis of cervical tracheal granulations and stenosis of various grades in clinic and in an animal model experiment.

We examined 28 patients (19 males and 9 females) who were recruited from ICU after tracheostomy and decannulation. Examinations were performed between a few days and 12 months after decannulation on a multidetector CT scanner. Navigation through the laryngo-tracheal lumen as well as reformatted coronal and sagittal images were performed on GE AW 4 workstation. The animal studies were done on 10 adult dead pigs in three series. In the first series, normal parameters of the larynx and cervical trachea were examined with adult bronchoscope Olympus (FLT) and after that by VLT. In the next two series, cervical tracheal granulations and stenosis were simulated and examined by both methods. The parameters of the larynx and cervical trachea were accurately measured RESULTS. The pathological changes in the clinical study found with the help of VLT were: mural granulations, polypoid mass in the tracheal wall, tracheal wall flap and flap with a persistent tract. When the FLT was performed it showed the same findings. The animal study confirmed our supposition that endoscopic pictures and virtual images provided similar macroscopic appearance. In most lesions, the measurements and localization of the abnormalities with VLT were more precise and easy to perform compared with FLT.

Comprehensive evaluation of the trachea can be achieved by combining these two techniques. We suggest that in the evaluation of critically ill patients following decannulation VLT examination should be done first. In cases where endoscopic biopsy or treatment is required, FLT should be then performed. Prospective studies will help to find out patients with complicated healing after tracheotomy and give opportunity to treat them properly.

Welch C A 1 , Harrison D A 1 , Rowan K 1 , Short A 2 1 ICNARC, Intensive Care National Audit and Research Centre, London, 2 Intensive Care Services, Broomfield Hospital, Chelmsford, United Kingdom

Alcohol consumption in the United Kingdom (UK) has steadily increased since the 1950s, with annual per capita consumption of 100% alcohol doubling from 4 litres to over 8 litres in this time [1] . We investigated the effect of increased alcohol consumption on the number of admissions to intensive care units (ICUs) in the UK each year.

Using data from 385,429 admissions to 166 adult ICUs in England and Wales from 1996 to 2005, we investigated how the number of admissions with alcoholic liver disease, their case mix, outcome and length of stay has changed during this time. Admissions were identified from the Case Mix Programme Database if they had 'alcoholic cirrhosis' or 'alcoholic hepatitis' specified as a reason for admission or other relevant condition. These figures were extrapolated to estimate the total number of admissions with alcoholic liver disease each year to all 229 adult ICUs in England and Wales.

We identified 4,219 admissions with alcoholic liver disease. The percentage of admissions with alcoholic liver disease increased from 0.7% in 1996 to 1.4% in 2005, but case mix, outcome and length of stay remained similar. The percentage of ICU bed-days occupied by these admissions has also increased. The projected total number of admissions to all ICUs in England and Wales increased from 500 in 1996 to 1500 in 2005 (figure), and the projected total number of bed-days occupied increased from 3,000 to 10,000. 

Gómez E 1 , Parejo M 2 , González C 3 , López V 1 , Calvo R 1 , Broch M 1 , García M 1 , Lacueva V 1 1 Intensive Care Unit, Hospital Sagunt, Sagunt, 2 Intensive Care Unit, Hospital de la Ribera, Alzira, 3 Preventive Medicine Department, Hospital Sagunt, Sagunt, Spain

Despite technological advances in coronary care, mortality related to Acute Coronary Syndrome (ACS) is entirely proportional to the delay to contact health system and receive an appropriate treatment. Our aim was to evaluate the impact of an educational campaign about ACS on the delay of demanding assistance and providing adequate care.

Quasi-experimental design with retrospective (pre-test) and prospective (post-test) data and non-equivalent control group. The target population included adults older than 30 from the catchment area of our hospital. The educational campaign was designed to provide comprehensive information about ACS and improve coordination among all the health system links involved in coronary care. Data from 242 patients pre-intervention and 241 from the intervention with diagnosis of acute myocardial infarction (AMI) admitted to ICU were collected. As a control group, we used data from a different catchment area similar in population during both periods including, respectively, 182 and 180 patients with AMI admitted to another ICU.

Comparing before and throughout the intervention, the intervals Symptoms-First contact and Symptoms-Fibrinolysis were shortened, but only the interval Hospital-Fibrinolysis was significantly reduced (52'(39.85) vs 39'(29.70); p= 0.002) (expressed as minutes (percentile 25.75). In the control group, none of these intervals showed a significant reduction in time during the study period. The access to the hospital was mainly by private transport (62.4%) in pre-intervention group, whereas the use of public health transport was increased after the campaign (from 37.6% to 57.4%; p<0.001). In patients who arrived at the hospital using private transport, a significant reduction in the delay between the onset of symptoms and fibrinolytic therapy (180' vs 124'; p=0.009) before and after the intervention was observed.CONCLUSION. 1. The educational campaign and a better coordination among health system links reduced the delay in receiving fibrinolytic therapy in patients with AMI. The interval from arrival at the hospital to fibrinolysis was significantly shortened, compared to a control group.2. Despite the increase in the use of public health transport to arrive at the hospital, patients using private transport received fibrinolytic therapy earlier during the intervention period.

Grenouillet-Delacre M 1 , Bégaud B 2 , Guisset O 1 , Gruson D 1 , Hilbert G 3 , Haramburu F 2 , Moore N 2 , Gabinski C 3 1 Medical critical care, University Hospital, 2 Pharmacology, U657 INSERM, 3 Medical critical care, Victor Segalen Bordeaux 2 University, Bordeaux, France INTRODUCTION. More than 10000 of deaths were attributed to heat wave during August 2003 in France. Classic heat stroke was uncommon in this country. Underidentification of drug related heat stroke could worsen organ failures and outcomes in intensive care unit (ICU). Few studies have assessed drug related morbidity in heat waves (1, 2) . The aim of this study was to assess frequency and characteristics of drug related heat stroke at admission to ICU during the august 2003 heat wave and type of drugs involved in heat stroke.

Patients admitted during the heat wave were included in a prospective study to a university hospital medical ICU. All drugs taken the month before admission to ICU were collected and a standardized questionnaire filled. For each patient, an independent expert panel determined relashionship between drugs and heat stroke, after follow-up and by using imputation method. Follow-up focused on morbidity, mortality, workload for personnel, and length of stay. Heat stroke patients were defined by an elevated core body temperature above 40• C with central nervous system dysfunction, in the absence of other etiologies explaining hyperthermia (1, 2) .

Between the 2 and 14th of August 2003, 43 were admitted in ICU. Eight patients (18.6%) were admitted for heat stroke. On admission, 6 of 8 heat stroke patients had 3 drugs or more known to interact with thermoregulation process. Lithium, neuroleptics, antiparkinsonian, and antihypertensive drugs were involved. In 4 cases, patients have association of neuroleptics therefore, a malignant syndrom was first suspected in the first hours in 2 cases. Despite intensive cooling, and artificial support (hemodialysis in 3, blood transfusions, vasopressive drugs and mechanical ventilation in 4), 3 of them died. Workload for personnel, length of stay were the same between patients with or without drug related heat stroke.

Heat stroke were frequently worsened by psychiatric, antiparkinsonian and antihypertensive drugs (1) . Drugs, delayed diagnosis and subsequent treatment may increase heat stroke related mortality. 

Drew R 1 , Cunney R 1 1 Dept. of Microbiology, Children's University Hospital, Dublin, Ireland

Many clinicians in Ireland have felt that there has been an increase in the severity of bronchiolitis over the last decade. During an infant's first episode of RSV, 0.5%-2.5% require hospitalisation (Centers for Disease Control and Prevention). The aim of this study was to examine if there has been an actual rise in the number of cases of bronchiolitis and also if there has been a rise in the percentage of cases requiring admission to ICU/HDU. The cases were also reviewed to determine if there was any correlation between admission to ICU/HDU and a shift in age profile of the infants.

All cases of bronchiolitis admitted to Temple St Hospital were reviewed between September 1996 and February 2006. Data was obtained from the infection control records of the Microbiology Department. Data regarding admission to ICU/HDU was available from September 1998. All cases were RSV positive by nasopharyngeal aspirate and aged less then 365 days on the day of virus isolation. Each bronchiolitis season was taken from September to February.

In total there were 667 cases over the ten years. There has been a steady increase in the percentage of children requiring admission to ICU/HDU from 5% in 1998-9 to 17% in 2005-6. The average age has not changed significantly, with a range of 104-137 days. The percentage of children in the group aged less then 49 days has varied greatly from 0% to 31% of the whole group. The increase in the admission rates of all infants is strongly correlated to a rise in the admission rate amongst infants aged less then 49 days (r=0.776). The total number of children in the group less then 49 days also strongly correlates with the admission rate (r=0.593) while there is no significant correlation of children aged 49-90 days (r=-0.004) and over 90 days (r=-0.004) with admission rates.CONCLUSION. The recent rise in the admission rates to ICU/HDU is strongly correlated to the rise in percentage of the infants aged less then 49 days in the group as a whole, and also the rise in the percentage of children less then 49 days requiring admission to ICU/HDU. Maternal protective antibodies are postulated to play a significant role in the protection of these infants aged less then 49 days. An antigenic shift in the RSV may be occurring resulting in fewer infants receiving transplacental immunity to the virus. To further investigate this, a prospective national audit is being carried out this year to determine this, and other possible causative factors.

Bakker J 1 , Mulder P 2 1 Dept. of Intensive Care, 2 Epidemiology and Biostatistics, Erasmus University Medical Centre, Rotterdam, Netherlands

Reduction of duration of mechanical ventilation (MV) in ICU patients is related to decreased morbidity and mortality and may increase ICU bed capacity. We studied the effect of remifentanil-based analgo-sedation (RAS) on the duration of MV and length of ICU stay (LOS) compared to conventional analgesia and sedation (CAS).

16 ICU's participated in an open label, centre-randomized, centre-crossover study. Patients with an expected MV-time of 2-3 days requiring analgesia and sedation were included. Study medication was given for a maximum of 10 days; either: CAS (morphine or fentanyl combined with propofol, midazolam or lorazepam according to Dutch IC Society guidelines) or: RAS (remifentanil at a starting dose of 6-9 mcg/kg/h, titrated to 0-45 mcg/kg/h and combined with propofol when required). When required, patients received CAS after day 10. Duration of MV (start study medicationextubation) and LOS (start study medication -ICU discharge) were analysed using a Cox Proportional Hazards model, stratified by study centre. Subjects not reaching the endpoints, including death or still ventilated after day 10, were censored.

. 205 ICU patients were included (109 CAS, 96 RAS). Age, SAPS, type of admission (post-surgical or medical) and diagnosis were well balanced. CAS contained morphine (58%), fentanyl (38%), propofol (46%), midazolam (81%) and/or lorazepam (7%). In RAS, the median weighted remifentanil dose was 8 mcg/kg/h and propofol was added in 65% of cases. Median duration of MV is 5.0 days (95% CI 3.4-6.6) for CAS and 3.9 days (95% CI 2.6-5.2) for RAS (crude Kaplan-Meier). Treatment effect on duration of MV was time dependent. For day 1-3 the extubation rate in RAS was 1.85 times higher than in CAS (p=0.019). Based on this ratio of 1.85, remifentanil increases the cumulative percentage of extubated patients as compared to CAS from 8% to 14% on day 1, from 23% to 38% on day 2 and from 34% to 54% on day 3. Between day 4 and 10, there was no significant difference (p=0.85). At day 10, 23 CAS-patients vs. 8 RAS patients were still mechanically ventilated (p=0.012). The ICU discharge rate during day 1-3 was 1.9 times higher in RAS than in CAS (p=0.044). The 85th percentile LOS values were 2.9 days for CAS and 1.8 days for RAS.

In patients with an expected duration of MV up to 3 days inclusive, remifentanil based analgo-sedation significantly decreases the duration of MV and length of ICU stay.

On behalf of the UltiSAFE investigators. Sponsored by GlaxoSmithKline BV.

Christensen S 1 , Larsen K M 2 , Jensen R 2 , Pedersen L 1 , Larsson A 3 , Tønnesen E 2 , Sørensen H 1 1 Department of Clinical Epidemiology, 2 Department of Anaesthesiology and Intensive Care, Aarhus University Hospital, Aarhus, 3 Department of Anaesthesiology and Intensive Care, Aarhus University Hospital, Aalborg, Denmark

A high prevalence of ICU admissions is related to chronic alcohol abuse. Limited data exist on the outcome of critically ill alcohol abusers requiring ICU admission. To clarify this issue we conducted a population-based cohort study to examine the 30-day mortality of critically ill alcohol abusers requiring ICU admission.

Through Aarhus University Intensive Care Study Cohort we identified all first-time ICU admissions to three multidisciplinary ICU's within Aarhus University Hospital from 1999 through 2004. Alcohol abusers ("previously hospitalized with an alcohol-related disease" or "previously prescribed antabus (disulfiram)") and comorbidity were identified through linkage to hospital discharge registries and prescription databases. Complete follow-up for mortality were obtained through the Danish Civil Registration System. We constructed Kaplan-Meier survival curves, based on the date of ICU admission, for the main study variables (alcohol abuse, age, gender, comorbidity, renal replacement therapy (RRT), mechanical ventilation (MV)) and computed 30-day mortality. We used Cox's regression analysis to estimate 30-day mortality rate ratios (MRR) for alcohol abusers compared with non-abusers adjusted for potential covariates.

Of 9597 patients with a first-time ICU admission, 583 (6%) where alcohol abusers. Alcohol abusers were generally younger (82% vs. 47% were younger than 60 years of age) and were more likely to be male (68% vs. 41%) and had similar levels of comorbidity (42% vs. 44%). The prevalence of alcohol abusers treated with MV (34% vs. 43%) and RRT (5% vs.7%) was lower than of non-abusers. The 30-day mortality among alcohol abusers was 19% compared with 21% among non-abusers corresponding to an adjusted 30-day MRR of 1.34 (95% CI: 1.10-1.63). Among patients referred from departments of medicine we found a 30-day mortality of 18% among alcohol abusers compared with 28% among non-abusers corresponding to an adjusted 30-day MRR of 0.84 (95%CI: 0.64-1.11). Among patients referred from surgical departments we found a 30-day mortality of 22% among alcohol abusers compared with 17% among non-alcohol abusers corresponding to an adjusted 30-day MRR of 1.80 (95% CI: 1.37-2.37).

We found a 30% increase in 30-day mortality of critically ill alcohol abusers requiring ICU admission compared with non-abusers; however, the increased mortality was a superimposing of two findings. Among patients referred from departments of medicine the 30-day mortality of alcohol abusers was reduced by 15% whereas patients referred form surgical departments had a nearly two fold increased 30-day mortality.

Rizk A F 1 , El Naggar A I 1 , Saad I M 2 1 Critical care Dept, Cairo University Hospitals, 2 Vascular Surgery, Mansoura Yniversity, Cairo, Egypt

Risk profiling in deep vein thrombosis (DVT) has been classically concerned with traditional factors of obesity, postoperative status, prolonged recumbency, long standing varicosity, etc with subsequent stagnation of blood & damage to vascular endothelium. Only recently, there has been increa sing concern with procoagulant factors as protein C, protein S, antithrombin III deficiencies as well as elevated F VIII, hyperhomocysteinemia, dysfibrinogenemia, etc.all of heredo familial nature.The present study is intended to assess the prevalence of two genetic disorders promoting coagulation namely, the mutant form of factor V (Leiden) and prothrombin gene in Egyptian patients with acute DVT.

We studied 30 pts admitted with acute DVT (16M,14F, mean age 44±14y) & 30 control subjects (19M,11F, mean age (37±10y). Excluded from the study were pts known to have bleeding dia thesis, those with acute or chronic liver disease and those on oral or parenteral anticoagulation. Following clinical evaluation including twelve leads ECG and routine labora tory tests, all pts were subjected to venous duplex and gene identification. The latter comprised DNA extraction, PCR amplification, and gene mutation detection using the THROMBO TYPE reagent kit.

Compared to control subjects, pts with acute DVT had signif icantly higher prevalence of factor V Leiden Gene mutation (66.7% vs. 23.3%, p 0.003). Compared to non carriers of this mutant form, carriers exhibited significant more frequent familial incidence (55% vs. 15%, p 0.035), younger age of pre sentation (40y vs. 51y, p 0.048) & more frequent complications (55% vs10%, p 0.049). Prothrombin gene mutation was exhibited by 3 out of 30 pts with acute DVT (10%) & was associated with factor V Leiden in two of them None of control subjects ex hibited this mutant form of prothrombin gene 20210.

Acute DVT among young pts and particularly with recurrent DVT should urge the cardiologist to search for factors pro moting coagulation. Our data show abnormally high prevalence of mutant form of F V Leiden (associated with prothrombin gene mutation in a minority). Besides the diagnostic value, gene mutation detection has therapeutic and prognostic imp lications through the need to adjust the dose & the durati on of oral anticoagulation.

Serclova Z 1 , Antos F 1 , Marvan J 1 1 Surgical department, University Hospital Bulovka, Prague 8, Czech Republic

Abdominal sepsis, tertiary peritonitis and re-operative abdominal surgeries lasting for many hours are inflicted with number of postoperative complications including operative wound complications. Necrotizing fascitis, dehiscence of laparotomy or small dehiscence of anastomosis with entero-cutaneous fistulas prolong the hospitalization of a patient at the ICU. Deep purulent complications in the operative wound demand expensive, painful and many weeks lasting dressing changes.

We have been using Vacuum assisted closure (V.A.C.) system in the treatment of those complications since June 2004. It is a sandwich method where we insert semi-permeable folia or mesh non-permeable folia into open laparotomy. Polyurethane foam and adhesive non-permeable folia is placed on top of that and into which centre target with vacuum catheter is placed. In case of fascitis the foam is placed directly onto the wound and its edges are sealed up with non-permeable folia with a target. The wound is then continuously or intermittently evacuated and dressing is changed once every 4-5 days.

. We used this system in case of 11 patients in the period between 06/2004 -03/2006. The system must have been changed 4 times in average, then the subcutaneous layer and skin was sawn up or a stomy device was possible to apply in case of reduced wound around the entero-cutaneous fistula. The wound healed within 22 days in average.

VAC system is a promising possibility in the management of complicated dehiscences in place of laparotomy. Active vacuum prevents pus retention, effectively lowers pus secretion, decreases edematous inflammatory reaction and supports granulation production. Main advantages are seen in faster healing and lower number of painful dressing changes. This system is almost irreplaceable only possibility of treatment of large dehiscent wound and enterocutaneous fistula. The next advantage, especially for ICU, is closeness of this system and therefore lower risk of spreading the hospital related infection. 

Aprotinin (A) is a serine protease inhibitor that is being used widely in cardiac surgery in order to reduce blood loss. In a recent study, use of A was associated with a significant increase in the risk of myocardial infarction or heart failure probably because of intravascular thrombosis (1) . The effect of A during OFF pump coronary surgery (OPCAB) is still not widely described. OPCAB is associated with postoperative hypercoagulability which may compromise the patency of bypass grafts (2) . Then, the aim of this study was to evaluate the risk of myocardial infarction in patients undergoing OPCAB with or without A.METHODS. 78 consecutive patients undergoing OPCAB with preoperative use of A (GrA) were retrospectively compared to 160 patients operated without A (GrT). Patients in group A received a bolus of 2 x 10(6) KIU during 30 minutes, followed by a continuous infusion of 0.5 x 10(6) KIU per hour until the end of surgery. Troponin I and CKMB levels were assessed at arrival in ICU (H0), 6 & 18 hours later (H6, H18). Statistical analysis was performed with t test, chi-2 test and anova.RESULTS. The 2 groups were similar for age, weight, ejection fraction and EuroSCORE. Significant inter-group differences were found in troponin I levels (p=0.023, fig), in CKMB values and cumulative blood loss within 24h. 

We retrospectively reviewed records of patients who underwent cardiac surgery in our 16-bed ICU with microbiological or histopathological evidence of Aspergillus during SICU stay.

Surgical procedures included: ascending aorta replacement (n=1), valve replacement (n=3), CABG and AVR (n=1). Engineering records were reviewed and no hospital construction or renovation projects took place during this period. All cases clustered in two operation theaters during a 5-month period. The exclusive occurrence of Aspergillus infection among surgical patients led to a closer examination of air quality in this area. A Biotest Air Sampler RCS (D-6000) Hycon used to measure the number of Aspergillus colonies/m 3 (colonies/m 3 =colonies on agar strip/sampling time). Sampling from the two operating theaters (no3, 4) showed a high Aspergillus colonization far exceeding the allowed norms (no3: 125cfu/m 3 and no4: 113cfu/m 3 ). Our 4 operating rooms are supplied with filtered air from two independent ventilation systems. Theaters no3 and no4 where the Aspergillus was identified are supplied from the same unit. Despite that the examination system indicated that no filters needed replacement; they were all immediately replaced and the air tunnels meticulously cleaned. After that no additional invasive Aspergillus infections have been occurred up to the end of March 2006.

Nosocomial transmission of Aspergillus infection occurs most often during or after hospital or surroundind buildings construction or renovation and usually observed in immunocompromised patients. However, the development of postoperative invasive Aspergillus infection is unusual. The investigation of invasive Aspergillus infections among post-surgical patients identified contamination in an operating theater air-handling system. Although contamination of air-handling system is uncommon, this outbreak suggests that some modification in current air quality guidelines is necessary in ICU and operating theaters.

Tessier Y 1 , Berthet M C 1 , Casez M 1 , Robin S 1 , Bach V 2 , Durand M 1 , Girardet P 1 1 Anesthesia, 2 Cardiac Surgery, CHU de Grenoble, Grenoble, France

Preoperative statin (S) therapy may reduce the risk of mortality after CABG with CPB (1. S induces biologic effects independent of lipid lowering which could explain this improvement, including anti-inflammatory effects (2) . The aim of this study was then to test the hypothesis that pretreatment with S before cardiac surgery without bypass (OPCAB) reduces postoperative inflammatory reaction.

Data were collected retrospectively for 277 patients undergoing non emergent OPCAB including 178 patients treated with S (GrS) and 99 non-S-pretreated (GrC. Primary outcomes was Creactive protein levels preoperatively (D-1) and from day of surgery (D0) to day 10 (D10). Statistical analysis was performed with t or chi-2 test and anova.

Preoperative use of S significantly reduced mortality (table) and CRP levels (Fig,  p<0 .0001).

Gordon S E 1 , Griffiths M J D 1 1 Unit of Critical Care, Royal Brompton Hospital, London, United Kingdom

The vascular endothelium exerts control over local thrombosis, inflammation and blood flow in addition to its barrier function. Traditionally endothelial function has been assessed using venous forearm plethysmography and flow mediated vasodilatation. These techniques are unsuitable for assessment of patients in the peri-operative period. Pulse wave analysis (PWA) is a novel technique that may have potential for endothelial function assessment in these patients. We have investigated the use of PWA and the response to salbutamol whose effects are partially mediated by the endothelium.

Patients with triple vessel coronary artery disease were enrolled (n=10). Studies were undertaken following abstinence of caffeine and recent fatty meals with patients lying supine in a quiet temperature controlled room. Recordings were taken from the Pulse Trace apparatus that analyses the digital pulse waveform. The response to incremental doses of inhaled salbutamol was recorded. Blood pressure and heart rate were monitored throughout the study.

A small dose-dependant reduction in the reflection index (RI): 3.96 +/-1.86%, (p = 0.38, one-way ANOVA) was observed only in response to relatively high doses of inhaled salbutamol 400mcg. Values were averaged between 5 and 15 minutes after inhalation.CONCLUSION. The reduction in RI following inhaled salbutamol in patients with severe coronary artery disease is very small. This may indicate that this technique is unsuitable for assessment of endothelial function in this population. 

The manifestations of CMV infection in immunosuppressed patients rang from asymptomatic virus shedding to severe organ disease. Patients who are not treated with immunosuppressive drugs and who are HIV-negative are usually not considered immunocompromised, and are therefore not at high risk for CMV infection. However, some reactivations have been observed in other populations, for example, after trauma, as well as in patients with cirrhosis, renal failure and those receiving dialysis. The objective of this study was to determine the prevalence, associated findings, and consequences of CMV viremia on surgical intensive care unit (SICU) patients.

Between a 12-month period we retrospectively reviewed records of patients who underwent cardiac surgery in our 16-bed SICU, and had positive CMV screening test by a molecular test assay. Patients with HIV infection and transplant recipients were excluded. Patients were tested for CMV based on clinical judgment of the attending physician.

. From a total of 1450 patients (p), 6 p (0.4137%) were found positive for CMV by the method of PCR (mean age: 70±7.51years). 2 p underwent valve replacement, 3 p CABG and 1 p Bentall and CABG. CMV viremia was diagnosed within 25.83±9.62 days after ICU admission. All patients fulfilled the sepsis criteria, were on mechanical ventilation, experienced renal failure and 4 p needed dialysis. All were on enteral nutrition. All p had been transfused but blood transfusion is considered a contamination risk when blood products were not leukocyte depleted. In our practice, all blood products used were leukocyte depleted. But this process usually does not eliminate all leukocytes, and few leukocytes are required for a CMV viremia to be positive. 1p received a 14-days antiviral therapy.CONCLUSION. CMV viremia is not an uncommon diagnosis in critically ill CSICU, regardless of their immune system status. The clinical significance of CMV is unknown and the differentiation between CMV detection and CMV disease represent a difficult diagnosis dilemma. CMV has known immunosuppressive effects, which may predispose chronic SICU patients to subsequent bacterial and fungal infection, and subsequent organ failure and death. Several factors suggest pathogenicity, but further study is needed to define causalty. 

The aim of this study is to compare evolution of procalcitonin (PCT), interleukin 6 (IL6), IL8, CRP and white blood cells (WBC) in the case of cardiac surgery with cardiopulmonary bypass (CPB).

Prospective observational study, including 100 adult patients who underwent a cardiac surgery with CPB. Patients with documented infection were included in group GI, the remaining patients in group GNI. PCT plasma level was determined before surgery, at hour 1 post operative (po), then daily until day 7 po. The threshold was considered as significant to indicate the occurrence of infection according to ROC curves.

Documented infection took place in 17 patients (bacteraemia, pneumonia, mediastinitis or parietal infection). Patients demographics as well as per operative clinical parameters were comparable in both groups. ROC curves analysis found that PCT plasma level in day 3 po (2ng/ml) had the highest predictive value of infection with a specificity of 90% and a sensibility of 77%. Air under the curve was 0.87. For CRP, WBC, IL6 and IL8 it was respectively 0.73, 0.75, 0.65 and 0.66. Compared with other markers of inflammation, PCT had the best positive and negative predictive values of post CPB infection.

In cardiac surgery, a procalcitonin plasma level > 2ng/ml after CPB, on the 3rd day post operative or just later would signal the occurrence of an eventual infection.

Fernández-Calderón C 1 , Suárez L 1 , Moreno I 1 , García de Lorenzo A 1 , Palomero M A 1 , Criado A 1 1 Anestesiologia y Reanimación, Hospital Universitario La Paz, Madrid, Spain

The objetive of this work was to analyzed the prognostic value of CRP during the cardiac surgery.

CRP of 34 patients for cardiac surgery with extracorporeal circulation were analized before surgery and 0,5,12 and 18 hours after the end of the procedure. We compared the CRP levels of patients with complications (divided in global, cardiac, respiratory and other complications) versus those without these events. Results are expressed as mean (SD) and statistical differences like p<0.05.

There were no statistical differences in CRP levels in patients with complications and non complications divided in global, cardiac, respiratory and other complications (table1).INTRODUCTION. The autonomic nervous system, endocrine, immune system form the stressresponse of the organism.

1. Heart rate variability (HRV) is analyzed a computerized beat-to-beat heart rate. Thus are investigated SDNN, RMSSD. 2. Plasma concentration of cortisol is analyzed using radioimmuno-assay technique. 3. Cytokine production by means of study granulocyte-macrophage colony-stimulating factor (GM-CSF) of blood were defined as markers of a condition of immune system. 4. For an estimation of microcirculation were defined free myoglobin. All given parameters were studied initially, at the moment of application 30 and 60 minutes and 30 minutes after the termination of an intravenous infusion of a medicinal preparation. During increase adrenergic responses the patients received a single i/v of clonidine (0.018-0.027µ g/kg). During hypervagal responses the patients received dobutamine (4.6-6.9 µg/kg per min.).

Dobutamin decreased total variability of R-R intervals: SDNN from 89+/-18 to 27+/-3.5msec(**-p<0.01); rMSSD from 115+/-26 to 30+/-7.6msec(**-p<0.01). Dobutamin increased of level cortisol in plasma from 399+/-100.0 nmol/l to 466+/-109nmol/l. Dobutamin decreased of level myoglobin from 277+/-53 to 145+/-76 ng/ml(*-p<0.05) and decreased of GM-CSF from 100% (initially) to 74%(*-p<0.05). Clonidin increased total variability of R-R intervals: SDNN from 9.8+/-1.5 to 27+/-5.63msec(*-p<0.05); rMSSD from 3.9+/-0.5 to 25+/-7.2msec (**-p<0.01). Clonidin decreased of level cortisol in plasma from 538+/-105 nmol/l to 444 +/-99nmol/l. Clonidin decreased of level myoglobin from 267+/-40 to 98+/-19 ng/ml(*-p<0.05) and decreased of GM-CSF from 100%(intially) to 57%(*-p<0.05).

Monitoring of neuroendocrine and immune parameters during of an intravenous infusion of a medical preparation has the potential to detect physiologic response to therapy in intensive care unit. 

Consecutive adult patients who underwent elective cardiac surgery with cardiopulmonary bypass (CPB) in our hospital were prospectively included. Exclusion criteria: heart trasplant recipients, previous adrenal insufficiency, HIV or treatment with steroids in the 14 previous days or during surgery. Epidemiologic, surgical, hemodynamic and analytic variables (baseline cortisol concentration and after 60 minutes of an injection of 0.25 mg of ACTH) were analysed. The diagnosis of adrenal insufficiency was made if the baseline cortisol concentration was < 15 mcg/dl or an increase < 9 mcg/dl after ACTH administration. The results were analysed with the SPSS (Statistical Package for Social Sciences) software for Windows (version 11.0).

A total of 96 patients were included (men 68%), with a mean age of 63 ± 2 years old, and a mean Euroscore of 5 ± 1. Most of the patients underwent valvular (46%) or coronary (32%) surgery, with a mean CPB time of 120 ± 12 minutes. Mean intensive care stay of 2 days. The mortality rate was 2%. In 29 patients (31%) a distinct inductor of etomidate was used. We found adrenal insufficiency in 65 (70%) of all patients. 

Infective endocarditis (IE) has a high mortality rate, and the option of surgical treatment must be individually evaluated. In this study we describe the outcomes in that patients who finally underwent cardiac surgery in our hospital.

Patients who underwent cardiac surgery between january 2000 and march 2006 in our tertiary care centre, because of IE, were prospectively analized. Of each patient, we collected age, gender, site of endocarditis, native or prosthetic, results of microbiological tests, indication for surgery, complications in the intensive care unit (ICU), ICU and hospital stay, and in-hospital mortality. The statistical analysis was made with the SPSS (Statistical Package for Social Sciences) software for Windows (version 11.0). Predictors of prolonged ICU stay were obtained with linear regression.

A total of 68 patients (men 79%, mean age 61 +/-4 years old) were included (2% of heart interventions). Mean Euroscore 8 +/-2. Native valve IE: aortic in 19 (28%), mitral in 11 (16%), and aortic + mitral in 8 patients (12%). Prosthetic valve IE: 30 patients (44%). S. Aureus and S. Epidermidis represents the 25 and 22% respectively. The main indication for cardiac surgery was heart failure (45%). Eight patients (12%) were in shock before surgery. Twelve patients (18%) underwent surgery because of uncontrolled infection. Mean time of cardiopulmonary bypass : 124 +/-14 min. The mean ICU stay was 9 +/-4 days, and the mean hospital stay was 42 +/-6 days. Complications during ICU stay: acute renal failure in 34 patients (51%), but only in 8 (12%) a continuous renal replacement (CCRT) was used; low cardiac output or prolonged inotropic support in 39 patients (43%); an arritmic event or need prolonged epicardic estimulation in 25 (37%); respiratory complications in 27 (40%), with a median extubation time of 19 +/-8 hours; neurologic complication in 17 (25%) and multiorgan disfunction (MOD) in 14 patients (21%). Eleven patients (16%) need reintervention because of IE relapse. Infection, MOD and CCRT were related with prolonged ICU's stay. The in-hospital survival was 74% (with a 76% of ICU survivors).

Most of the patients had prolonged in-ICU and in-hospital stay, with high morbidity and mortality rates. Infection, MOD and need for CCRT were predictors of prolonged ICU stay. Surgical treatment is nowadays an useful option in selected patients.

Kiessling A H 1 , Isgro F 1 , Skuras J A 1 , Kammerer I 1 , Lehmann A 1 , Saggau W 1 1 Klinikum Ludwigshafen, Cardiac Surgery-Anaesthesiology, Ludwigshafen, Germany

Selenium is an essential part of the intracellular antioxidant system as a component of the glutathione peroxidase enzymes. Selenium seems to play a role in the regulation of inflammatory processes. The aim of the study was the evaluation of clinical outcome in patients with postoperative extended ICU stay and selenium therapy.

In this prospective, randomized, open controlled, not blinded study, we compared the effects of intravenous selenium application. The clinical endpoints were defined as mortality, length of stay (LOS) and time on respirator (TOR). Between 01/2005 to 09/2005, a total number of 74 patients underwent cardiac surgical procedures and had a prolonged postoperative stay on the ICU (> 5 days). The selenium group was treated with an initial dose of 2000mg IV and a preservation dose of 500mg for 10 days.

The primary reasons for a prolonged ICU stay were caused in a cardiac low output syndromes or respiratory failure. The treatment with selenium was initiated after the 5 day on ICU. We could not detect any beneficial aspect of the therapy. Selenium does not influence the mortality rate, LOS, TOR or the incidence of a septic shock.

Long term treatment of critically ill patients admittedly is loaded by a high inhospital-and follow up mortality. Selenium therapy had not the potential to improve the clinical outcome. A randomized, prospective, double blinded multicenter phase III study including the measurement of selenium serum levels and markers of the inflammatory response, is necessary to describe possible benefits. (1), and alterations in cerebral histopathological structures have been described after bacteremia (2, 3) . The aim of this study was to assess the relationship between cerebral perfusion and histopathology during early endotoxemia.

12 Animals (weight: 42kg ±4; mean±SD) were exposed to E.coli lipopolysaccharide or saline infusion (n=6, each) for 12 hours. Cardiac output (thermodilution; ml/kg/min), systemic (MAP) and pulmonary artery blood pressure (PAP, mm Hg), and cerebral blood flow (laserDoppler, arbitrary units) were continuously measured. At the end of the experiment, formalin fixed brains were cut in coronal sections and embedded in Paraffin. Afterwards, the sections were cut at 5 microns and stained with HE.RESULTS. 4 of 6 endotoxemic animals but none of the control group had cerebral tissue lesions (areas of encephalomalacia with spongy degeneration of the white matter and axonal swelling and ischemic neuronal thalamic necrosis). Out of the 4 animals, 3 had also significant vascular changes at the level of venules and small veins predominantly in the brainstem. -group-interaction, p<0.05, c:control, e:endotoxemia CONCLUSION. Endotoxemia induced histopathological brain damage which was not related to systemic or cerebral blood flow and pressure. 

Lyazidi A 1 , Galia F 1 , Thille A 1 , Di Marco F 1 , Rodriguez P 1 , Brochard L 1 1 Réanimation médicale, Henri Mondor Hospital, Créteil, France INTRODUCTION. The aim of this study was to evaluate the performance during PSV mode among 13 ICU ventilators (figure).

To stimulate spontaneous ventilation, we used a two-chamber lung test (Michigan Instruments). To test inspiratory trigger we used 2 levels of P0.1 (2 & 4 cmH2O) and two levels of PEEP (0 & 5 cmH2O); to assess the pressurization we used 3 levels of respiratory drive (low, moderate and high) and 4 levels of pressure support (5, 10, 15 and 20 cmH2O). We evaluated: 1) time delay (TD, between the onset of the effort and the rise of the Paw) 2) the maximal decrease in airways pressure (DPaw) and 3) the pressure-time product (PTP) at 0.3 sec and 0.5 sec, reflecting the efficiency of pressurization. 

Button D 1 , Fodor P 1 , Alpiger R 1 , Zollinger A 1 , Hofer C K 1 1 Institute of Anaesthesiology and Intensive Care Medicine, Triemli City Hospital Zurich, Zurich, Switzerland INTRODUCTION. The aim of this study was to compare CO determined by two pulse contour analysis devices (FloTrac, Edwards Lifesciences = FCO and PiCCO plus, Pulsion Medical Systems = PCO 1 ) with intermittent thermodilution (ICO) in patients after cardiac surgery.METHODS. 20 patients were included, for one set of data (A = hemodynamic stable) CO was assessed after hemodynamic stabilization. Triplicate FCO and PCO values were recorded before ICO was determined by 3 injections at 4 time points with 15 min intervals. For the second set of data (B = hemodynamic changes) triplicate FCO, PCO and ICO measurements were recorded 15 min after inducing different body positions (supine, 30• head-up, 30 • head-down, supine). Statistical analysis was done using ANOVA (Bonferroni) and Bland-Altman analysis for absolute values and % changes (δ).

CO during A ranged from 3.0 to 7.1 l min -1 without significant CO changes between measurement points (δFCO = 0.9±13.0%, δPCO = 1.0±11.3%, δICO = 0.6±11.2%). Mean bias±2SD (limits of agreement) was 0.04±1.31 l min -1 for FCO-ICO and -0.01±1.09 l min -1 for PCO-ICO. Differences of δCO were comparable (mean bias±2SD = 0.3±24.1% for δFCO-δICO and 0.4±20.2% for δPCO-δICO). A range of CO from 2.85 to 8.60 l min -1 were obtained during B with significant changes of FCO, PCO and ICO between the measurement points ( Table 1) . Mean bias±2SD was -0.10±1.80 l min -1 for FCO-ICO and -0.15± 1.21 l min -1 for PCO-ICO. For δFCO-δICO mean bias±2SD was -2.2±46.7% and for δPCO-δICO -4.9±21.2%. 

Schulz-Stübner S 1 , Harris B 1 1 Anesthesia, University of Iowa Hospitals and Clinics, Iowa City, United States

Power injections of contrast material have become a routine part of many CT protocols for the brain (e.g. CT-Angio), chest (e.g. Pulmonary artery protocols), abdomen and pelvis. Data regarding the feasibility and safety of power injections through central venous catheters is limited. Unfortunately, in many critically ill patients no peripheral access can be obtained. We therefore designed an in-vitro experiment to test the use of the MAC Multiaccesscatheter (Arrow international, Reading USA) for power injections in a range from 3 ml/sec to 10 ml/sec.

The MAC consists of three lumina (Lumen A: 12 Gauge, Lumen B: Distal 9 F with an 8 F catheter) and a one way valve secured introducer port for PA-catheters or additional multilumencatheters). MAC's from different lots were immersed in a water bath and connected to a Power Injector (Stellant Dual Injector by Medrad). Contrast medium (Omnipaque 300) was injected with flow rates of 3 ml/sec, 5 ml/sec, 8 ml/sec and 10 ml/sec. Contrast flow was recorded on video and catheter movement was graded as none, minimal, mild or severe by an observer blind to the pressure injected. Pressure curves were obtained from the Injector. Lumen A and B were injected with and without the presence of a PA-catheter through the introducer port and with or without a simulated catheter obstruction (hemostat on the distal third of the catheter). A conventional triple lumen catheter was used as control at the mentioned flow rates.

Power injections up to 8 ml/sec were possible through Lumen A and B, with or without a PA-catheter in place, without any signs of catheter damage. At a flow rate of 10 ml/sec the injection was stopped by the Injector because the pressure limit of 325 psi was reached. With complete obstruction an injection through Lumen B caused the one-way membrane of the introducer port to rupture and release contrast medium at a continuous flow rate with 134 PSI. While the conventional triple lumen catheter demonstrated a jet phenomenon with a severe whiplash movement the stiffer MAC did not show movement at flow rates up to 3 ml/sec and minimal movement at the higher flow rates.

Power injections up to 8.0 ml/sec can be safely performed through a MAC with or without a PA catheter in place through the introductory port. Complete catheter obstruction can result in destruction of the one-way valve of the introductory port. The reduced mobility of the MAC makes him favorable compared to conventional triple lumen catheters for power injections.Grant acknowledgement. The catheters for this study were provided free of charge by Arrow international, Reading USA. 

Bispectral index (BIS) monitoring of depth of anesthesia has pioneered the field for more recent monitoring devices like the A-line ARX Index (AAI) or the state (SE) and response entropy (RE) monitoring devices. Following an observational design the present study aimed to compare simultaneously in the same, normothermic patient recorded BIS, AAI and entropy (e.g. RE, SE) values.

Data from patients (n = 33) undergoing minor gynecological operations were analyzed. Boli of fentanyl and of propofol were given to induce anesthesia. To maintain anesthesia propofol and remifentanil were infused. At the anesthetist's preference, either an endotracheal tube or a laryngeal mask was used for airway management. Before induction of anesthesia commercially available AEP electrodes, BIS and entropy sensors were simultaneously placed on the forehead. BIS, AAI and entropy value recordings were started at 3 minutes before induction of anesthesia and continued until transfer of the patient to the postanesthesia care unit. Markers were set at defined landmarks. The anesthetist responsible for the patient was deliberately unaware of AAI, BIS and entropy values during the operation.

The prediction probability (Pk) of BIS (0.99+/-0.0004) and AAI (0.97+/-0.001) was comparable and better than that of SE (0.88+/-0.0075) or RE (0.86+/-0.0072).

During uneventful anesthesia, e.g. without patient movement provoked by surgical stimulation, BIS and AAI values showed better correlation than did AAI and entropy or BIS and entropy values. Agreement, however, between BIS, AAI and entropy on patient state (e.g. awake, inadequate anesthesia, optimal or deep anesthesia) did not exceed 20%.

Benoit D D 1 , Vandijck D M 1 , Depuydt P O 1 , Offner F C 2 , Peleman R A 3 , Roosens C D 1 , Noens L A 2 , Decruyenaere J M 1 1 Intensive Care Medicine, 2 Hematology, 3 Infectious Diseases, Ghent University Hospital, Gent, Belgium

Developing septic shock after having received chemotherapy is often considered as the worst case scenario in hematology. The aim of this study was to compare the characteristics and outcome in septic shock patients with hematologic malignancies who received vs. who did not received i.v chemotherapy within 3 weeks before ICU admission.

Prospective observational study of all consecutive patients with hematologic malignancies admitted in the ICU with septic shock between 2000 and 2006. The diagnostic certainty (documented vs. clinically suspected bacterial infection and presence or absence of non-bacterial infection) and the site of infection (pulmonary vs. non-pulmonary) were assessed by an independent panel of physicians blinded to the patients' outcome. Severity of illness on day 1 was assessed by the SOFA and a cancer specific severity of illness score (CSSIS) [1] . Multivariate logistic regression was used to identify early (<24 hrs) predictors of hospital mortality.

Over the study period, 100 septic shock patients with hematologic malignancies were admitted. Forty-five percent received previous i.v chemotherapy. The ICU, hospital and 6 months mortality in patients with vs. without previous chemotherapy was 29% vs. 53% (P=0.025), 48% vs. 64% (P=0.15) and 54% vs. 71%(P=0.08), respectively. Patients who received chemotherapy were younger (52 ± 19 vs. 64 ± 12 years, P<0.001), more often had a high grade malignancy (80% vs. 36%, P<0.001), were more often neutropenic (62% vs. 20%, P<0.001), less often had a pulmonary site of infection (42% vs. 69%, P=0.009) and were less often ventilated (62% vs. 84%, P=0.02) and dialysed (13% vs. 35%, P=0.02) during ICU stay compared to those without. However, we found no difference in SOFA (11. CONCLUSION. Septic shock complicating active chemotherapeutic treatment has a better prognosis than commonly perceived. This might be explained by the lower incidence of pulmonary infection in this group. 

Vibration response imaging (VRI) measures vibration energy generated from airflow to create a dynamic image of respiration process and the software in the VRI device provides a waveform displaying the total vibration intensity from the lungs over time. We investigated the effect of end inspiratory flow on configuration of VRI waveform in mechanically ventilated patients.

We performed serial VRI recordings during assisted volume control ventilation with the square (VC-sq) and decelerating (VC-dec) flow patterns and pressure support (PS) ventilation in 5 mechanically ventilated patients. Time between peak inspiratory and peak expiratory vibration was measured and t-test was performed.

There is a difference in the separation of peak inspiratory and expiratory vibration between VC-sq, PS and VC-dec with p values of 0.0178 for VC-sq versus PS, 0.01 for VC-sq versus VC-dec and 0.048 for PS versus VC-dec (Fig 1) .

End inspiratory flow determines the distance between peak inspiratory and expiratory vibration in VRI respiratory cycle waveforms. In mechanical ventilation, the flow at the end of inspiration in VC-sq is maximal, less in PS and almost zero in VC-dec. Lower flow at the end of inspiration causes greater separation in VRI waveforms. When inspiratory flow persists at the time of ventilator cycling, this energy must be dissipated prior to reversal of flow with expiration. The lack of separation between peaks reflects the persistence of this energy. VRI waveform analysis may provide clinically useful physiologic information for adjustment of mechanical ventilation. decline in the serial CD14 and IL-I levels from day1 to day 5 occurred in IVIG-treated entire group of pts as well as severe sepsis and septic shock subgroups of pts. Mortality rate was significantly lower in IVIG-treated pts with severe sepsis compared to conventionally treated pts (40% vs 70%). In IVIGtreated pts with septic shock, the mortality rate was 100% just like that of the conventionally treated pts with septic shock.

Used early enough, IVIG therapy induces substantial improvement in morbidity and mortality in patients with severe sepsis, but couldn't demonstrate any beneficial effect on morbidity or mortality rate in patients with septic shock. The improvement in the clinical course and ultimate outcome of severe sepsis is paralleled by a consistent decline in serum levels of IL-6, PCT and CD-14. Key words: Sepsis, IVIG, Cytokines, PCT, APACHE II score and SOFA score.

Van Nuffelen M 1 , Piagnerelli M 1 , Blecic S 2 , Robin V 3 , Vincent J 1 1 ICU, 2 Stroke unite, 3 Heamatology, Erasme, Bruxelles, Belgium

In-hospital cardiac arrests, intensive care unit (ICU) admissions and unexpected deaths are commonly preceded by warning signs 24 hours before the acute event. Appropriate alarm system detected deterioration are needed to allow an early intervention, this is the focus of our study.

The Bio Sign algorithm (Oxford signal, oxford, UK) uses 5 physiological parameters (heart rate [HR], respiratory rate [RR], blood pressure [BP], arterial oxygen saturation [SpO2] and temperature [T]) to generate a single measure. This so-called Bio Sign Index. represents the probability that a patient's physiological data is different from the normal data for a representative group of patients on whom the Bio Sign algorithm was trained. Whenever the Bio Sign Index exceeds a preset threshold of 3.0 for four minutes out of a five-minute period, the monitor is deemed to have detected an event. We monitored 16 patients on three different wards (a stroke unit, a haematology unit and an intensive care unit) including 8 patients at risk of development of complications and 8 patients at the end of their life (with "do not resuscitated" order) RESULTS. The recording were analyzed by two independent observers (Table 1) . Positive predictive value of the Bio Sign signal was 86%. Most of the artefactual Bio Sign Alarms were due to the temperature probe losing contact (15 out of 38 artefactual Alerts -39%). Despite frequent artefactual decreases in the SpO2 readings, only 4 of these caused an artefactual Bio Sign Alarm. In the 8 patients who died without intervention (end-of-life decision) the mean time between the Bio Sign alarm and the cardiac arrest was 2h 46 min [ range : 20 min -7h ]. 

Chiarandini P 1 , Mozzoni L 1 , Bertossi G 1 , Covello D 1 , Tripi G 1 , Cristofaro A 1 , Costa M 1 , Della Rocca G 1 1 Anesthesia and Intensive Care, Az. Policlinico Universitario, Udine, Italy

Indocyanine green plasma disappearance rate (ICG-PDR) obtained non invasively in the early postoperative period after orthotopic liver transplantation, has been shown to be a valuable parameter to predict graft function and clinical outcome [1] . However few data exist on the reliability of this system using low dosages of ICG in patients with an hyperdynamic circulation and poor liver function [2] .

We studied 29 patients after OLTx, who received a 4F aortic catheter with an integrated fiber-optic device and a thermistor (Pulsiocath 4F PV2024L, Pulsion Medical Systems, Munich, Germany) via the femoral artery sheath connected to a computer system (COLD-Z021, Pulsion Medical Systems), and an ICG finger clip connected to a liver function monitor (LiMon, Pulsion Medical Systems). Through a venous central access 0.3 mg/Kg of ICG (Pulsion Medical Systems) was injected, and data were simultaneously obtained. Most of the patients had more than one ICGPDR evaluation in different days. For statistical analysis data were compared with linear regression and according to Bland-Altman analysis. 

Fast-paced intensive care testing demands that laboratories deliver blood test results as rapidly and accurately as possible. The goal of this study is to define limits of agreement between sodium, potassium, chloride, glucose, lactate and haemoglobin results from the delocalized Rapidlab 1265 blood gas analyzer and the central laboratory in our critically ill patients. In a previous study, 89 minutes were so gained (1).

The 3-month prospective observational study was approved by the ethics committee. For each studied patient, a single arterial blood sample was taken. A first part was drawn by a 2 ml syringe to be processed by the blood gas analyzer Rapidlab1265 located inside the intensive care unit, the second part was collected by three appropriate tubes and sent by pneumatic conveyor system to the central laboratory. Na, K, Cl, glucose, lactate were measured on Modular System and haemoglobin on a Celldyn 4000 analyzer. The Bland and Altman's method determined the limits of agreement between the results.

. Table 1 describes for each parameter the number of comparisons, the observed range values in the central laboratory (minimum -maximum), the skew and the 95% limits of agreement (lower and upper). The skew represents laboratory value minus Rapidlab value. 

Aloizos S P 1 , Liapis G 1 , Maragiannis D 1 1 ICU, 401 General Army Hospital Athens, Ag. Paraskevi, Greece

Assessing the intravascular volume status of critically ill patients can be exceedingly difficult. Due to concerns about the efficacy and safety of using invasive haemodynamic monitoring, non-invasive diagnostic testing has gain increasing importance. We try to compare the reliability of VPW as an indicator of overload, in patients of a general ICU, with a method of invasive haemodynamic monitoring that has proved in literature and in every day practice his efficacy.

Vascular Pedicle Width (VPW), which represents the mediastinal silhouette of the great vessels, was compared to the haemodynamic measurements, which were obtained with the method of transpulmonary thermodilution (PiCCO Plus, Pulsion Munich). We measured the VPW in anteroposterior chest x-rays in supine position, with standard parameters, in 50 patients without prior cardiac surgery, prior mediastinal irradiation, obesity, severe ARDS and PEEP. In every patient we perform invasive haemodynamic monitoring with PiCCO Plus. ITBI > 1000 ml/m2 and GEDI > 800 ml/m2, ELWI > 7.0 ml/kg were considered as the markers of significant volume overload.

The mean VPW in overloaded patients was 79.85 cm compared to a mean of 63.71 cm for the rest. The results were subsequently analyzed using Spearman's non parametric test and we found good correlation (0.802, 0.788, 0.510) between VPW and GEDI, ITBI, ELWI, respectively. The results were considered statistically significant (p < 0.000, p < 0.000, p < 0.005) respectively.CONCLUSION. VPW when appropriately assessed at bedside using portable chest x-rays might give very useful information regarding volume status of the patients, results that are comparable in their efficacy to those obtained with invasive and more expensive methods. (21 vs 12), diagnosis more frequently are heart diseases (58% vs 30% in survivors)(ischemic heart diseases 54%), and surgical diseases (22% vs 12%). Different intensive procedures in died patients (85% vs 64%), the more frequent is a mechanical ventilation (75% in died vs 10% in survivors), and the less frequent is the pace maker (20% in died vs 35% in survivors).

The very oldest patient admitted in ICU is a short percentaje, origining a very short lenght of stay. Mortality and the severity of illness is higher than the rest od the population described in the bibliography. Mortality in very old patient admitted to ICU are associated to severity of illness, heart diseases, emergency surgical treatment and the needed of mechanical ventilation too. The age was not explained the practice of providing less aggressive care to elderly patients, because higher than 80% of the patients survive, thus it is not relevant criterion for ICU admission. Through those facts the number of old patients admitted to our ICUs is increasing. Aim of following paper was to objective the treatment and outcome of very old patients (older than 85 years) at a medical ICU over a two years period (01.01.1999 -31.12.2000) .

1098 patients had been admitted to our ICU during the study period, 60 (5.5%) older than 85 years. At admission the APACHE II-score ranked between 16 and 36. Indications had been mainly cardial (36), metabolic/intoxications (9), gastrointestinal (6), outside CPR (6) and acute respiratory failure (3).

It was necessary to ventilate 12 patients (20%) for 1-8 days (mean 3 d), 7 patients received a cardiac pacemakersystem, 5 underwent endoscopical interventions, 4 thrombolysis (due to acute myocardial infarction 100mg Alteplase "front loaded"), 4 patients PCI (in two patients an IABP was inserted) and one female patients ACBG. Duration of stay had been 3.8 d (overall 5.6d), mortality 26.7% (overall 14.75%).

Comorbidity and mortality had been naturally higher in patients older than 85 years compared to all patients admitted to our ICU during the study period. 6 month after the ICU stay 24 patients (54.5%) had been still alive with reported good quality of life. We belive that despite higher mortality at the ICU even very old patients benefit from ICU stay and critical care interventions. 

Cesta M A 1 , Wakefield C 1 , Price K 1 , Nates J L 1 1 Critical Care, MD Anderson Cancer Center, Houston, United States

The decisions regarding the processes at the end of life of are complex and influenced by many factors including religious, education level, and cultural. We investigated the common characteristics of the individuals who died in our ICU with focus on the frequencies of withdrawal of supportive measures.

We performed a retrospective chart review of all patients who died within our 53 bed combined oncological medical and surgical ICU. All patients admitted between 9/01/04 to 8/31/05 were included in the study. Data collected included, but were not limited to, demographics, cancer type, the initiation of the terminal wean, and utilization of Comfort Measures Order Sets (CMO). An extensive chart review including nurses, pharmacy and respiratory therapy records to determine data accuracy was performed. Race description and religious affiliation was patient reported. Financial class is reported by this institution in relation to methods of payment. CONCLUSION. Whites had a significantly higher frequency of self determined discontinuation of life supportive measures than nonwhites. Since we found no differences in the types of cancers, religious affiliations, or financial class; we suspect the differences are probably related to cultural viewpoints within these groups. However, this cannot entirely be determined from this study. Further, there is a small difference in the CMO utilization rate between the two groups studied, but, these numbers are not statistically significant and thereby do not allow for any conclusions. METHODS. Data collected prospectively in 37 European ICU's in 17 countries during 18 months (m) and divided in 3 consecutive groups of 6 m, including EOLP category [1] (CPR, WH and WD therapy, shortening of dying process -SDP), information about pt wishes, discussions with pt and families, and time from ICU admission to first limitation of therapy and from first limitation of therapy to death.

Of 31417 pts, 4248 were included. Changes in EOLP [n (%)] over time are seen on table 1. Median time from ICU admission to 1st limitation of therapy (days) was 3.5 (0-6 m), 2.5 (7-12 m) and 3.0 (13-18 m), p=0.001. Median time from 1st limitation of therapy to death (hours) was 15.3 (0-6 m), 17.6 (7-12 m) and 11.3 (13-18 m), p=0.0001. There is increasing public interest and expectation that there is more family involvement in these complex decision making processes. EOL care decisions, processes, and discussion with family are often poorly documented and not subjected to regular evaluation or audit [1] . We undertook a retrospective study of patients who had been admitted to a university teaching hospital ICU during 2005 where there had been a decision to withdrawal treatment.

Case notes of 100 consecutive patients where an EOL decision was made were reviewed. Data collected included age, sex, APACHE II score on admission, timing of EOL decision, mode of withdrawal, time to death, and communication with next of kin (NOK). CONCLUSION. Communication NOK was significantly higher than previously documented [2] . In our series median time to death from EOL decision was less than 5 hours and extubation did not appear to influence time to death. 

Chronic ventilation is an available treatment option for patients with various endstage pulmonary, cardiac and neurological diseases. In Israel, legal, social and religious issues prohibit disconnection from the ventilator in dependant patients. Therefore, the population of chronically ventilated patients is rising, requiring adequate chronic ventilation and weaning facilities. This study aims to evaluate attitudes of first degree relatives of patients chronically ventilated in one of these facilities in Israel, regarding end-of-life decisions for their relatives in comparison with their attitudes towards themselves and patients in general.

We collected patient demographics, relevant medical information and cognitive function. First degree relatives were interviewed using a structured questionnaire. They were required to address interventions for their relatives such as dialysis, antibiotics and transfusions. They were presented with various hypothetical scenarios such as chronic ventilation, resuscitation and disconnection from the ventilator, in patients with preserved or reduced consciousness. For each scenario they were asked about their wishes for themselves and for patients in general. 

Lugarinho M 1 , Souza P 1 , Castro P 1 , Silva L 1 , Silva S 1 1 ICU, Hospital de Clínicas Mário Lioni, Rio de Janeiro, Brazil

Complex issues, as interruption of treatment in terminal patients are present in every day activity in an ICU, as so as in media. To know the point of view of the staff is essential to understand attitudes and decisions. This study tried to analyze quantitatively medical prescriptions of patients who died and staff's attitude about therapy discontinuation.

Study, retrospective was done in a private general ICU with 23 beds. It was analyzed prescriptions of patients who died after more than 24 hours after admission in the ICU. Period of study was the year of 2005. It was defined as Therapeutic Intensity (TI) the number of item of medical prescription, except the drugs on demand (SOS). The TI was registered in the first 24 hours (TI-1) and in the last 24 hours (TI-2) of ICU admission. The TI was calculated from the ratio TI-2 / TI-1.A research was done with staff about therapeutic withhold of mechanical ventilation, nutrition, fluid management, antibiotics, vasoatives drugs, sedation and analgesia in patients which death is imminent and irreversible.

. From 990 admissions, 93 patients were studied. Mean age was 62.5 (SD 15.5), with mean time in ICU of 14.5 days (SD 12.6). TI-1 found was 10.07 (SD 2.07) and the TI-2 was 13.76 (SD 3.89). The analysis of sub-groups with time in ICU less than 7 days did not show difference compared with sub-group with time greater than 7 days. CONCLUSION. There is diversity in definitions. The administration of morphine is sometimes determined as PS, or, when doses of morphine are increased with the intention to shorten life, as DTL. ICU nurses and physicians judge APS, PS and DTL at the request of the patient as acceptable.

Hughes M 1 , Labram A 2 , Prior L 2 1 Intensive Care Unit, Royal Infirmary, 2 Intensive Care Unit, Western Infirmary, Glasgow, United Kingdom

Withdrawal of care is the most frequent cause of death in intensive care units. Information on the decision making process and the methods by which care is withdrawn is sparse. We wished to assess staff attitudes in Intensive Care Units (ICUs) in Scotland.

A questionnaire was distributed to all ICUs in Scotland. 18 detailed questions were asked.

. 571 responses were received (56%). In dying pateints, 66% alway diable alarms and 25% do so mostly. 46% felt alarms should always be disabled, 35% thought they should be disabled most of the time. Table 1 : If the patient is alert and orientated, is and should their poor prognosis and plan of care discussed with them? Table 2 . Do and should the family help to decide on level of monitoring in a dying patient? CONCLUSION. These results give an insight into staff attitudes to death and dying in ICU, as well as current practice. Some of the answers are unexpected. Not only is a poor prognosis and subsequent plan of care not always discussed with an alert and orientated patient, but also staff feel these discussions should be less frequent. In an era of patient autonomy, paternalism of this sort is becoming outmoded. We are unclear why alarms should so frequently be enabled in dying patients. It may be that staff feel insecure without the backup of alarms. The level of family involvement in decisions about monitoring is more contentious and the spread of answers understandable. We hope that these results will eventually allow a consensus to emerge in Scotland about some aspects of palliative care for dying patients in ICU.

Andruszkiewicz P 1 , Kanski A 1 , Sicinski M 1 , Gelo R A 1 , Konopka P 1 1 Department of Anaesthesia and Intensive Care, Warsaw Medical University, Warsaw, Poland

The opinions concerning indications for cardiopulmonary resuscitation (CPR) have been modified for the last 25 years. CPR is known to have a low success rate especially when attempted for patients in poor general condition or terminally ill. Several factors such as ethnic origin, religion, sex have been specified to have impact on physician's acceptance of "do not attempt resuscitation" (DNAR) order. In western countries where individual's autonomy is well recognized patients are deeply involved in decision making process. Although this issue has been in the center of debate all over the world it has not been properly explored in East Europe so far. The objective of the survey was to examine an actual clinical practice and current opinions of the physicians about several aspects of CPR with special emphasis on process of making decisions.

Two questionnaires were specially created for the study purposes. The first one examining current practice was filled in by physicians who diagnosed cardiac arrest. The questionnaire exploring opinions was filled in by a group of 168 physicians out of 200 drawn among 500 employed in Warsaw University Hospital.

Physicians on-call were main decision makers (45%) with no input from the patient when the decisions DNAR were made. Their knowledge about patients' medical condition was often scarce. DNAR decisions were usually informal and communicated to medical team only orally (98%). "Slow codes" are common in current clinical practice in Poland. Majority of examined physicians (63%) are not familiar with DNAR term. 20% of doctors declare that patients should be involved in the decision making process concerning CPR. More then 30% responders indicate the need for collegial elaboration of this extremely important decision. In opinion of 80% of doctors once such a decision is made it should be formally recorded. Sex, age and professional experience modify physicians' attitudes towards different aspects of resuscitation.

Results of the study confirm that attitudes of polish doctors concerning several aspects of resuscitation differ meaningfully from a model accepted in majority western European countries. Paternalistic and informal model of decision making still predominates, but many responders indicate that this approach should be changed. Current opinion of physicians differs strikingly from clinical practice in almost every respect.

Boles J M 1 , Boumediene A 1 , Boles C 2 , Tonnelier J M 1 , Prat G 1 , L'Her E 1 , Renault A 1 1 Service de Réanimation médicale et Urgences médicales, Hôpital de la Cavale Blanche -CHU, 2 Association ASP Iroise, Association ASP Iroise, Brest, France

We introduced non professional hospice volunteers (NPHV) in our 15-bed medical tertiary care ICU in september 2003 with the aim to help conscious patients and family members to relieve the burden of an ICU stay. French laws and regulations recommend the presence of NPHV in palliative care settings. We performed a study to collect patients and family members' appreciation.

We signed an official agreement with the association "ASP Iroise", a member of a national network of NPHV associations, defining NPHV' presence and role. Four NPHV, women 40 to 65 yrs, took alternate turns in the ICU on Tuesday afternoons one week out of two from 09/2003 to 03/2004 and each week since 04/2004. NPHV were free to visit any conscious patient or family who wished so and the ICU staff asked the NPHV to visit people who seemed particularly distressed. NPHV wrote a brief commentary about their visits in a special logbook which can be read by the ICU staff. All living patients and family members who met NPHVs were sent an anonymous questionnaire in 01/2005. 

Little is known on the information delivered in ICUs by booklets. We surveyed the content of ICU information booklets in French ICUs.

A sample of 105 ICUs, was enrolled. A questionnaire focusing on the modalities of patient family information, including the request to rate the Drs opinion on informing relatives on specific important issues was sent. Doctors (Drs) of ICUs were also asked to send their ICU information handout. This material was analyzed and we compared the percentage of information issues (items) found in these booklets and recommended by the two French ICU societies (1, 2) and an American. (3) We quoted how often Drs deemed important some specific items should be explained to families.

. 59 ICUs answered (56%). Numbers represent median percentage of response or median percentage of presence in booklets for the item considered. We found a significant difference (p<0.01) in booklet contents by type of ICU thus showing some "cultural" differences in information strategies. The amount of items recommended by reference documents (1-3) and found in booklets was low (41%), even if it was higher than the amount of items not recommended by reference texts present in booklets (6%)(p<0.01). Considering the items we believed as important and which were submitted to the judgment of ICU Drs, we found that they were scarcely present in booklets (14%) despite they were rated as important and deserving to be incorporated in booklets (68%)(p<0.01) Of note, some items were not mentioned in the three reference ICU societies but were present in a large percentage of booklets, e.g. modalities of children visits (63%), practical organisation of visits (23-44%), information to close relatives only (21%), role of, and response to the alarms (21%), possible need to physically restrain patients (13%), data describing the ICU architectural organisation (13-50%). The following items were marginally present in booklets despite often mentioned as needed by Drs: iatrogenic events (39% vs 2%), organ donation (51% vs 2%), inclusion in research studies (51% vs 6%), patient privacy (73% vs 21%), commitment to treat pain (76% vs 15%), how families manage stress (75% vs 13%).

There is a discrepancy between information recommended, observed and whished by Drs. INTRODUCTION. to analyse changes in the attitude of relatives to tissue donation so as to modify interview strategies.

A comparative, prospective and descriptive study of the interviews carried out betwen: 2001-2003 and 2004-2005 with all families of potential tissue donors (Ptd). An specific protocol was used:epidemiological data, interview data, perception of family attitude.

First period: 3400, evaluated 519 as Ptd with 42% refusals. Second period: 3826 deceased, we evaluated as Ptd 947 with 33% refusals (p=0.017). We carried out 784 interviews and 1012. 6% knew something of Spanish transplant law and 2% carried a donor card or had made a living will. Percentages were similar in both periods. The interviewed relatives who refused donation were respectively (2001-03/ 2004-05): parents 8-4%, partner 37-37, offspring 40-49, siblings 10-7 and other relatives 6-2(ns). Relating the cultural level of the relative with the refusals, we found that this was: low 20-26%, medium 62-46 and high 16-28(ns). The reasons for family refusals, in both periods, were as follows: deceased refusal whilst alive 33-22%, family opposition 23-25, problems with the health system 7-4(p=0.029), problems with the corpse image 3-3, deceased's wishes unknown 12-12, religious problems 1-2, flat refusals 12-26. Others 19-6. Reasons justifying acceptance:wish to donate whilst alive 15-10%, refusal turn-around 11-2, post-information acceptance 70-60(ns)

There is still an important lack of awareness with regards the needs for tissue. Further work on diffusion and informing about donation and transplant is required. In the latter period, the high level of studies is not directly related to a higher level of donation. There were not significant changes in the attitude to tissue donation, deceased's will and simply familiar negatives are the most argued reasons. We felt people less prone to change their opinion because the transplant coordinator explanations.