key: cord-014670-e31g8lns
authors: nan
title: Poster Sessions 313-503
date: 2004-10-05
journal: Intensive Care Med
DOI: 10.1007/s00134-004-2406-2
sha: 
doc_id: 14670
cord_uid: e31g8lns

nan

The improvement in P/F ratio in the HFOV group compared to CMV was statistically significant [at 8 hours (p=0.05) and 24 hours (p=0.001)] and this trend continued through the study period. Though the FiO2 was higher in the HFOV group at baseline, by 4,8 and 24 hours this was significantly lower as compared to CMV group(P<0.01and P<0.05 and P<0.01 respectively)and again this trend continued over the study period. 

We studied 18 patients with heatstroke admitted in ICU during the August 2003 heat wave in France. Plasma samples were available at the admission for all 18 patients and during the course of the disease in 13 patients. To assess the extent of the inflammatory response in the patients, plasma concentration of cytokines was studied by ELISA. Leucocyte activation was evaluated by the expression of eta2 integrins and L-selectin by flow cytometry using specific MoAbs; reactive oxygen production (ROS) by chemoluminescence and metalloproteases MMP9 and MMP2 by gelatin zymography. As markers of cell activation and/or apoptosis, microparticles (MP) isolated from plasma were double-stained with annexinV (AV) and cell specific MoAbs against platelets (CD41) or granulocytes (CD15) and analyzed by flow cytometry. Microparticles procoagulant phospholipids were measured by prothrombinase assay. Whole blood and microparticles TF was determined by a specific clotting assay.

Increased levels of IL6, IL8 and IL1-RA were observed whereas IL18, IL1-eta and TNFlpha were normal or undetectable (table) . Blood leucocyte activation was demonstrated by:-an up-regulation of eta2-integrin expression, -a down-regulation of L-selectin expression, -an increased ROS production. Moreover, pro-MMP2 and 9, possibly released by activated granulocytes were increased in the 5 tested patients with presence of active MMP9 in three. Markers of DIC (thrombocytopenia, decreased FVII and FV levels, presence of soluble fibrin and increased levels of TAT) were observed in 17/18 patients. Whole blood TF was increased in all patients (235±199 pg/ml, mean ±SD) vs controls (< 40 pg/ml). Compared to healthy controls, the number of AV positive MPs was not increased, but the cellular origin was different, with a significant decrease in platelet MPs (p<0.05) and a significant increase in granulocyte MPs (p<0.05). Furthermore, MPs TF was increased and contributed for a large part to the high procoagulant state. High levels of inflammatory cytokines play a crucial role in leucocyte activation leading to down regulation of L-selectin, ROS production and active MMP9. Our results suggest a major role of MPs of granulocyte origin in the TF-dependent procoagulant state that correlates with the severity of the disease.

Kalenka A 1 , Münch E 1 , Fiedler F 1 1 Departement of Anesthesiology and Critical Care Medicine, Faculty of Clinical Medicine Mannheim, University of Heidelberg, Mannheim, Germany INTRODUCTION: Recombinant factor VIIa (rFVIIa; NovoSeven, Novo Nordisk, Agsvaerd, Denmark) has been approved for prevention and treatment of bleeding in patients suffering from hemophilia with inhibitors. Numerous case reports and retrospective studies submitted to a webbased drug surveillance have been published, that record the sucessfull use of rFVIIa to treat lifethreatening haemorrhage in patients without pre-existing coagulation disorders. However, it is unknown whether rFVIIa induces hemostasis in septic patients with disseminated intravascular coagulation. Therefore, the objective of this study was to investigate clinical efficacy of rFVIIa in septic and non-septic patients with bleeding complications METHODS: Between 09/2001 and 03/2004, 25 adult patients with life-threatening haemorrhage without pre-existing coagulopathy were entered into the study retrospectively. Out of 6 patients with severe sepsis and DIC (known infection and at least 1 organ dysfunction) and 17 non-septic patients with severe bleeding complications due to different reasons, we reviewed coagulation parameters and the amount of transfused blood products prior to and 48 hours after application of rFVIIa. The patients´s underlying diseases and dosages of administered rFVIIa are illustrated in Table 1 . Relevant thromboembolic complications were recorded RESULTS: Prior to administration of rFVIIa, 17.4 ± 3.3 (mean ± standard error of the mean) units of Red Blood Cells (RBC), 9.7 ± 1.6 units of Fresh Frozen Plasma (FFP) and 3.2 ± 0.6 units pooled platelets (Plt) were substituted to the non-septic patients. After administration of rFVIIa, significantly less RBC 4.4 ± 1.2 (p<0.01), FFP 3.3 ± 1.6 (p<0.01) and Plt 1.3 ± 0.8 (p<0.05) were transfused. Coagulation analysis demonstrated normalisation of international normal ratio (1.3 ± 0.01 versus 1.1 ± 0.01; p<0.01) and Partial Thromboplastin Time (58 ± 11.0 s vs. 34 ± 2.6 s; p<0.05) after administration of rFVIIa. No differences were detected concerning the platelet count. In the septic-group administration of rFVIIa neither resulted in a reduction of transfused blood products nor in a reversed coagulopathy. Three thromboembolic complications were observed in the non-septic patients. Mortality rate was 29% in the non-septic group vs. 83% in the septic group (p=0.02) 

Marsilia P F 1 , Imperatore F 1 , Munciello F 1 , Scarpelli M 1 , Teodori R 1 , De Cristofaro M 1 , Occhiochiuso L 1 1 Unit of Anaesthesia and Intensive Care, Department of Emergency, A. Cardarelli Hospital, Naples, Italy INTRODUCTION: Acute septic descending mediastinitis (ASDM) has been defined as the mediastinum infection which results from the spreading of oropharyngeal infections, or deep neck structures (1). Drotrecogin-alpha (activated) or recombinant human activated protein C (rhAPC) is the only biological agent approved for use in severe sepsis syndrome that has demonstrated efficacy in reducing 28-day all-cause mortality and new data suggests a trend towards longer term survival(2). Very few data in literature report the use of rhAPC in the treatment of ASDM. Aim of this study is to report our experience of 7 cases of ASDM complicated by a severe sepsis-induced multiple organ failure syndrome (MOFS) and successfully treated with the infusion of rhAPC.

Four male and three female patients were admitted to our Intensive Care Unit for ASDM due to oropharyngeal infection and complicated by a severe sepsis-induced MOFS. All patients were undertaken to combined cervicotomy and thoracotomy surgical operation with toilette and drainage of infection. Moreover the infusion of rhAPC at 24 gamma/kg/min for 96 hours was given together with all other certain established procedure. Respiratory failure was managed with sedation, orotracheal intubation and invasive mechanical ventilation. Cardiac failure was managed with invasive arterial and central venous pressure monitoring together with the infusion of catecholamines up to standard doses; while renal failure was managed with continuos venous-venous haemodiafiltration. Antibiotics were given first empirically and then according to lab test results.

In cardiac surgery, more than in any other type of surgery, in early postoperative period can occur low cardiac output syndromes due to hypovolemia or to myocardial failure. Our objective is to evaluate the accuracy of the systolic pressure variations (SPV) and of its negative component ∆ down under mechanical ventilation in predicting the response to volume loading and to diagnose hypovolemia.

In a group of 50 patients who underwent CABG surgery, in the early postoperative period we monitored: CO/CI, CVP, PCWP, BP (S/D/M), VPS and ∆ down. The including criteria were: sinus rhythm, CI ≤ 2.5 L x min-1 x m-?, PCP < 18 mm Hg. All the patients underwent a fluid challenge (500 ml of colloids in 10 min). According to the CI variation the patients were then divided in two groups: group A (28 pts) with a raise of CI > 15%, and group B (22 pts) with a CI variation < 15%. We analyzed the variations of the parameters mentioned above due to fluid loading and the differences between the two groups.

In the following table are the data obtained. All parameters are measured in mm Hg and are expressed as the average value ± standard deviation (*p<0.05).Statistical analysis shows significant differences between the two groups regarding only the initial value of SPV (14.07 ± 1.82 mm Hg in group A, 6.45 ± 2.32 mm Hg in group B, p<0.01) and ∆ down (9.78 ± 2.14 mm Hg in group A, 1.5 ± 1,65 mm Hg in group B, p<0.01). There also significant differences of the values before and after the fluid challenge only in group A and for the same parameters: SPV (from 14.07 ± 1.82 mm Hg to 9.75 ± 1.26, p < 0.001) and ∆ down (from 9.78 ± 2.14 mm Hg to 5.42 ± 1.23 mm Hg, p<0.001). In predicting a significant raise of CI after volume loading a SPV > 12 mm Hg 

The new parameters of preload dependency tend to replace the classic pressure parameters in hemodynamic assessment, being more accurate as predictor of CI response to volume loading and as detector of hypovolemia (1). The low costs and accuracy of the SPV method advocate for using it in critical care settings, even in cardiac surgery. 

Marangoni E 1 , Volta C A 1 , Alvisi V 1 , Bertacchini S 1 , Ragazzi R 1 , Orlando A 1 , Alvisi R 1 1 Anesthesia and Intensive care, University of Ferrara, Ferrara, Italy

Several studies provide compelling evidence on the clinical role played by fluid optimization. Till now, the assessment of the intravascular volume has been based on data derived by pulmonary-artery catheter. However, some studies suggest that the use pulmonaryartery catheter is associated with an increased mortality and hence the central venous pressure (CVP) remains the only parameter to be used.

Nine patients undergoing mechanical ventilation were enrolled. CVP was determined while patients were breathing spontaneously (SB) and during assisted control ventilation (ACV) at different PEEPe levels (0-4-8-12 cmH2O). Patients clinical characteristics are (mean±SD): age (yr) 59±23; weight (kg): 63±7; tidal volume (ml, ACV): 405±6; respiratory rate (b.min-1, ACV): 19±5; static compliance of the respiratory system (ml.cmH2O-1, ACV): 35±7. The following parameters were determined: mean airway pressure (ACV), mean systemic blood pressure (MAP), CVP.

The most dramatic increase of CVP was registered when the patients were ventilated in ACV compared to spontaneous ventilation ( fig. 1) . Surprisingly, CVP values were much less influenced by progressive rise of mean airway pressure obtained by different PEEPe level. Moreover, MAP variations were closely linked to those of CVP.

Our data shows that CVP monitoring is useful for assessing the intravascular volume in patients requiring high PEEPe level. However, the CVP values obtained during ACV are very different from those calculated during SB and hence, when possible, it is advisable to determine CVP during SB.

Pottecher J 1 , Bouyges S 1 , Caron S 1 , Moreau X 1 , Beydon L 1 1 Dept of anesthesia, Hôpital Larrey, ANGERS, France

Routine use of pulmonary artery catheters (PAC) measuring continuous cardiac output but not SvO2 is controversial in cardiac surgery. Oxygenation derived variables like arteriovenous oxygen difference (AVDO2) could better reflect supply-demand balance than cardiac index (CI). Moreover, bedside, respiratory changes in arterial pulse pressure (DeltaPP) are a more reliable indicator of fluid responsiveness than pressures obtained from PAC (CVP and PCWP). The goal of our study was to compare two hemodynamic assessment methods: one based on PAC derived variables, the other taking AVDO2 and DPP into account.

Fifty consecutive mechanically ventilated patients emerging from cardiac surgery in a university hospital ICU were included without informed consent since PAC insertion is systematic in our institution.An independent observer simltaneously recorded PAC variables (CVP, PCPW, CI) (first set, S1), AVDO2 and DeltaDPP (second set, S2). Initially, the caring physiscian could get the first set of value, was asked about volemia, inotropism and vasomotor tone and gave a therapeutic option. Then he had access to the second set of value, answered the same questions and was able to maintain or to change his opinion. Patients were divided in four categories, according to DeltaPP (<13 or >=13) and AVDO2 (<5 or >=5) and discrepancies between the two sets of answers were analysed.

Based on S1 data, 41 new drugs or therapeutic were introduced (82% of patients): 4 inotropic drug, 27 fluid loading, 5 diuretic, 4 vasodilatator, 1 vasopressor. After knowing S2, in 25 patients (50%) we challenged the initial decision. DeltaPP did not correlate with CVP nor with PCPW. Similarly, CI did not correlate with AVDO2. According to table: case 1: 11 therapies decided knowing S1 challenged by S2: all should not have been fluid loaded; case 2:4 therapies decided knowing S1 challenged by S2: they should have received a fluid load; case 3: 6 therapies decided knowing S1 challenged by S2: 2 should not have been fluid loaded and 4 should have received inotropes; case 4: 4 therapies decided knowing S1 challenged by S2: 4 should have been fluid loaded and 2 should not have received inotropes.

DeltaPP >= 13 (n) DeltaPP >= 13 (n) AVDO2 <5

19 (11) case 1 9 (4) case 2 AVDO2 >=5 12 (6) case 3 10 (4) case 4 CONCLUSION: Conventional PAC measurements do not allow optimal therapeutic guidance, postoperatively in cardiac surgery. Indeed, CI does not reflect metabolic requirements at best measured by AVDO2. Also, CVP and PCPW do not reflect hypovolemia contrarily to DeltaPP. The association all data provided by S1

Albers J 1 , Heggemann F 1 , Kayhan N 1 , Bahner M 2 , Vahl C F 1 1 Abt. Herzchirurgie, ICU, 2 Abt. Radiodiagnostik, Chirurg. Universit, Heidelberg, Germany

Noninvasive imaging of coronary artery disease (CAD) using multidetector computer tomography (MDCT) provides theoretically additional information to the classical 2D coronary angiography. The objective was to determine: (1) How accurate is the 3D imaging compared to the 2D method? (2) Is it feasible to profit from the additional 3D information in the clinical setting?

METHODS: Study population consisted of consecutive patients with diagnosis of 3-vessel-CAD (n = 30).Every patient underwent both, 2D coronary angiography and MDCT scanning (Siemens Somatom Plus 4 VZ, slice thickness 1.5 or 3.0 mm, pitch 1.5, contrast medium 70 ml). Retrospective gating was used. 3D visualization was performed using raytracing. Comparison of 2D and 3D imaging was performed in a blinded manner, 3 blinded investigators scored applicability of the coronary segments (CS) for aortocoronary bypass grafting (ACB) (necessary/not necessary) and stenosis (stenosis <25%/26-50%/51-75%/76-100%).

(1) Agreement in applicability for ACB was (CS number/% agreement) 1/73. 3, 2/76.7, 3/53.3, 5/70.0, 6/93.3, 7/70.0, 8/63.3, 9/46.7, 11/80.0, 13/30.0 . CS rarely being object to ACB, showed poor agreement: 4/16.7, 10/10. 0, 12/16.7, 14/16.7, 15/40.0 . Agreement in quantification of stenosis was: 21.7% for the right coronary artery, 40.4% for the left anterior descendent and 26.1% for the circumflex coronary artery.

(2) 3D volume data were acquired in a single breathhold. Temporal resolution was 170 ms (reconstruction time 1 min/image) enabling calculation of stroke volume. 3D visualization showed distribution of coronary calcifications together with non-calcified lesions.

Severe CAD was identified noninvasively in all cases studied.

(1) Accuracy of the 3D method was sufficient for bypass planning purposes. However, quantification of stenosis was not acceptable.

(2) Data acquisition was quick, safe and provided additional data superior to conventional 2D (calcification, soft plaques, 3D quantification of stroke volume). In conclusion, patients with severe CAD can be diagnosed with high accuracy using noninvasive imaging.

Boulo M 1 , Fleyfel M 1 , Robin E 1 , Lebuffe G 1 , Lecoutre H 1 , Onimus J 1 , Tavernier B 1 , Vallet B 1 1 Anesthesiology and Intensive Care Medicine, University Hospital, Lille, France

Aortic surgery can be taken as a model of fluid and blood losses leading to volume status variation and hemodynamic impairment. These variations together with aortic clamping may compromise tissue perfusion. In intubated and ventilated operated patients, respiratory pulse pressure variation (∆PP) reflects ventricular preload dependency (1). ∆PP is a good predicting marker of increase in stroke volume index (SVI) after a fluid challenge (FC) (2). The aim of this study was to evaluate whether preload dependency as assessed by ∆PP measurement was associated with impaired tissue perfusion.

After approval from the local Ethics Committee, 15 patients undergoing aortic surgery were prospectively enrolled. Intraoperative hypovolemia was suspected when heart rate increased and/or systolic blood pressure dropped more than 20% from baseline. A 250 mL colloidal FC was then systematically performed. Automated gastric tonometry (Tonocap, Datex-Ohmeda, Finland) was used to assess PgCO 2 -PetCO 2 before and after FC (CO 2 gap-1,CO 2 gap-2) . An increased CO 2 gap larger than 20 mmHg can be taken as a threshold value of decreased tissue perfusion. An increased SVI larger than 15% was identified as responder (R) and FC was repeated until SVI did not increase more than 15% again. An increase in SVI of less than 15% was identified as non-responder (NR CONCLUSION: Increased production of gut wall lactate was previously shown to be associated with increased leakage of macromolecules across the gut wall due to gut barrier dysfunction (1). In this study we were able to show that even short and uncomplicated CPB leads to increased gut wall lactate detected by gut luminal microdialysis, indicating gut barrier dysfunction. Simultaneous tonometry proved to be insensitive to these changes. We propose that gut luminal microdialysis in the rectum may be a good method to estimate markers of metabolism and gut barrier dysfunction during surgery and in the critically ill patient. 

Matamis D 1 , Tsagourias M 1 , Vakalos A 1 , Synefaki E 1 , Kareklas M 1 1 ICU, Papageorgiou General Hospital, Thessaloniki, Greece INTRODUCTION: COPD patients are often aged, smokers and may suffer from right ventricular (RV) failure. Moreover, they may have smoking and age -related diseases of the left heart, as ischemic heart disease or valvulopathies. Left ventricular (LV) failure may induce hypercapnic respiratory failure (HRF) and mimic the clinical picture of COPD, especially in patients with heavy smoking history. The aim of our study was to identify, in patients with HRF leading to mechanical ventilation (MV), the prevalence of RV, LV, or biventricular failure, the presence of severe valvulopathies and the impact of the targeted cardiac treatment (according to ECHO findings) on the MV days and mortality.

Over a period of seven years 82 patients (58M and 24 F) with a mean age of 68±8 years and a mean APACHE score of 18±6 included in the study. Heart function was assessed with ECHO within the first 24 hours from intubation. Patients were divided in three groups (RV failure, LV failure and Normal Heart group) according to ECHO criteria. Cardiac treatment was given according to ECHO findings. Respiratory treatment and weaning process was identical in all patients. ANOVA and chi square test were used for statistical analysis. 

The objective of this study was to measure the cost-effectiveness of an IgM enriched immunoglobulin preparation in adult patients treated for severe sepsis and septic shock. We performed a meta-analysis followed by an economic analysis conducted from the hospital perspective in Germany.

Effectiveness data from a meta-analysis of eight randomised trials (N=383) was used to assign probabilities in a decision model to estimate cost-effectiveness of IgM enriched immunoglobulin preparation and its comparator standard therapy. Analysis of effectiveness data used all cause hospital mortality as the primary outcome and intensive care (ICU) length of stay (LOS) as a secondary outcome. Benefit was expressed as lives saved (LS). Published ICU treatment cost data was applied to assess differences in treatment costs. Cost-effectiveness was calculated as the incremental cost per LS. 

We develop a systematic data collection of all the admitted patients in our ICU through a home-made software and database leading to a broad description of the population and activities of the ICU during the last five years; this was correlated with the classical scoring systems of ICU patients. The system was utilisator friendly made by automatically generating hospitalisation ICU reports which guaranteed the use of the database and its completion. At the end of each year and after a data validation period standard reports were generated with classical parameters: mean age of the population, mean length of stay, mortality rate, readmission rate, daily repartition of all the ICU admissions ... and correlated with the severity of the ICU population.

The following observations were noted during this five years period: non significant elevation of the mean age and of the mean ICU length of stay, but significant reduction of the global mortality while the mean ICU scores remain stable. In the same period the number of ICU technical procedures (right catheterisation days, artificial ventilatory days, CVVHDF days...) was also significantly higher but without significant influence on nosocomial infection rate. Interesting management data were also available: more than thirty procent of the ICU patients were admitted during the night shifts (between 08 PM and 08 AM) which can be an important data to discuss the staffing problems (during the night shifts for example). The use of antibiotics was also significantly reduced during the same period.

A five years data collection period of standard ICU indicators correlated with the use of severity scores can be an interesting ICU management tool to promote quality of care (and communication procedures inside and outside the ICU) and to reduce mortality and morbidity in ICU populations. Further studies are necessary to confirm these interesting observations.

The National Intensive Care Evaluation (NICE) registry (www.stichtingnice.nl) aims to analyse and improve the quality of Dutch intensive care. The NICE registry contains 112 data items for each patient admitted to one of the 29 participating ICUs. To support the individual ICUs in comparing their population and performance to several standards NICE has introduced an internet application, NICE Online.

Users of NICE online compose their own data analyses by selecting a)functions, b)split-elements, c)comparisons, d)subpopulations. Figure 1 (left side) shows an example of a request for a graph which presents the mean length of stay (function) for survivors and nonsurvivors (split-element) of the user's own ICU and of all participating ICUs together (comparison). Readmissions to the ICU are excluded (subpopulation). Results of analyses, presented in graphs or tables, can easily be copied, e.g. to management reports. Privacy of patients and of ICUs is ensured by 1)login and password, 2)encryption of transferred data, 3)using a copy of the original NICE database without patient-or ICU-identifying information, 4)disabling combinations of functions and comparisons which may lead to identifiable information.

In Figure 1 

Transport of the critically ill patient between hospitals or intrahospital remains an hazardous road trip where the patient is exposed to less controlled circumstances outside the ICU. To gain more insight in the incidence of complications, circulatory-and respiratory instability (CI/RI) related to intrahospital transport, we investigated transports from our surgical ICU.

All data concerning transports from the surgical ICU to the department of Radiology from 2000-2003 were retrospectively reviewed. Clinical relevant circulatory instability (CI) was defined as the necessity to start vasoactive medication(VAM)or to change the existing dosage of VAM during or directly after transport to maintain a MAP ≥ 65. Clinically relevant respiratory instability (RI) was defined as the need to change the settings of the mechanical ventilator (FiO2, Peep, Minute volume) during or directly after transport. We also evaluated the administration Additional Opiates and Sedatives (AOS) prior to or during transport and the possible impact of i.v. administration of contrast fluid jopromide (Ultravist ®) on renal function. 

Sedation and analgesia are essential components of patient care in the Intensive Care Unit (ICU). "Bottom up" costing of intensive care is more accurate but more labour intensive and difficult to perform compared to "top down" costing (1). "Bottom up" cost of sedative, analgesic and neuromuscular blockade drugs have not been reported. We therefore performed an audit of the cost of these drugs in our ICU using the "bottom up" costing approach.

Over a 3 month period, we prospectively recorded the daily amount of sedative, analgesic and neuromuscular blockade drugs administered to patients in a 12-bedded ICU and multiplied the amounts by the cost of drug per milligram using pharmacy costing figures. Patients were divided into 4 groups that corresponded, roughly, to the length of stay quartile marks.

Out of 178 patients admitted during the study period, data were collected for 155 (92%). We also collected data for 94% (990 days) of ICU patient days. Table 1 shows cost of sedation per group, patient and ICU day. Around 94% of the cost was on drugs administered to the 50% of patients who stayed in ICU for more than 48 hours. Propofol and alfentanil were the commonest drugs used (administered to 88% and 68% of patients respectively) and the most expensive ( 5,577and 11,502 respectively). Total cost was 21,342 which was 81% of the pharmacy ("top down") cost.

Cost of sedation per group. 

Several medical specialities have conducted surveys among their residents during their training, in order obtain a feed back and assess possible improvement issues. At our knowledge, not such a survey has been conducted in Europe. In France, ICU specialisation can be obtained by two separate trainings: either specifically in medical ICUs or jointly to the anaesthesia training which is followed by the majority of future ICU practitioners. A shortage in ICU doctors is expected, urging raised efforts to make this training as attractive as possible. This inquiry was intended as a first step in this purpose.

A 22 question questionnaire was mailed to the first year residents (Y1) whereas a 40 question one was sent to the forth year residents (Y4) registered to the joint ICU-anaesthesia training program (lasting 4 years) in France. Anonymity of answers was insured.

A total of 96 questionnaires were received (44%) for Y1 and 123 (77%) for Y4. Main Y1 answers were: age: 25 ± 1 year, (39% female). They chose ICU for good job opportunities (29%), a clinical (24%) and dynamic (15%) speciality. They had discovered ICU-anesthesia during medical studies (10%). Night shifts were: not disturbing (21%), are shared by many other specialities (23%), and are more interesting that non-specialised night wards (23%). A 53% of them hesitated before starting this specialisation. Half of them changed geographical region for specialisation. Medical English was spoken and red: 66%. A computer was owned personally: 56%. Main Y4 answers were: 29 ± 2 year, (40% female), their 4-year training was judged for theory: excellent (7%), good (66%), fair (32%); for practical teaching: excellent (19%), good (70%), fair (12%). Medical English was spoken and red: 82%. Only 30% had a position at the end of their training. They wished to join public institutions: 66%, private: 29%. Research was a professional issue for 5%. Future practice restricted to ICU was aimed by 28%, whereas 57% preferred anaesthesia and 15% emergency medicine and pain clinics. They had published in French at least once during training: 38%, in English: 15%. Additional degrees (mainly in infectiology and as sub-specialisation in some ICU techniques) were obtained by 48% during specialisation. Periods spent in non-academic hospitals (one year) were rated as excellent: 47%, good: 44%, fair: 8%; and medical supervision was judged excellent-good in 87%. They spent an average 726 euros/year for medical furniture (books, computer). A computer was owned personally: 56%. Finally, 96% did not regret choosing ICU (and anaesthesia) as a speciality.

Training in ICU (and anaesthesia) was judged as good especially for practical aspects by most of residents. About a third aimed at working exclusively in ICU. More efforts should be performed to improve formal job offer at the end of the training. CONCLUSION: Diagnostic and imaging testing represents a significant amount of hospitalization costs in our ICU due to the high number of tests performed per patient per day and especially for ABG analysis. Thus, a better control of lab tests ordering will result in costs reduction and a cost containment policy is becoming mandatory.

Kanevetci B N a c i 1 , Dosemeci L 1 , Y?lmaz M 1 , Cengiz M 1 , Ramazanoglu A 1 1 Anesthesiology and Reanimation, Akdeniz University, Antalya, Turkey

The common problem arising in the ICU's is the use of the beds for the patients who are expected not to benefit from ICU treatment. In this prospective study, we aimed to determine the proportion, costs, length of ICU stay and prognosis of those patients who were expected to die according to our clinical experiences.

One hundred and forty five patients over 18 year of age admitted to our 24-bed ICU between March 2003 and February 2004 and expected not to benefit from ICU treatment according to the experiences of the physicians working in ICU. The APACHE II, SAPS II and GCS scores were noted and according to those scores, the estimated mortality rates were determined by the formulas. Also ICU stay, ICU beds occupied by those patients, ICU and hospital discharge mortality and morbidity and costs were determined. We didn't change the treatment strategies of those patients. The patients who had high risk of mortality but could possibly recover completely after given therapy were not included in the study.

The mean age was 55.9 ± 18.9. The most common underlying diseases were nontravmatic intracerebral hemorrhage ( % 26.2 ), cerebrovascular accident ( % 15.9 ), head injury ( % 14.5 ), metastatic tumors and vascular diseases of gastro-intestinal tract ( % 10.3 ), cardiac arrest ( % 9.7 ), lung cancer and end-stage lung diseases ( % 9.7 ) and the others ( % 13.7 ). The mean GCS, APACHE II, SAPS II scores were 6.7 ± 3.6, 24.4 ± 6.1, 54. This represents a 9% increase in activity with no change in the number of beds available or alteration to admission criteria. In both time periods bed occupancy was >100%.Length of ICU stay decreased from a mean of 9.72 days to 7.13 days (p<0.01)and the length of ventilation decreased from 7.5 days to 5.3 days (p<0.01). This reduction was observed across all specialites excluding neurosurgery.

The use of information technology to provide iterative feed back has reinforced the adoption of the ventilator and euglycaemia carebundles making the desired elements of the care bundle the default mode within the intensive care unit. The culture of the ICU has changed enabling more reflective practice ready for the adoption of the sepsis care bundle and other packages of evidence based treatment.

Papadopoulos A C 1 , Karakoulas K K 1 , Vassilakos D 1 , Filippidou M 1 , Skourtis C T h 1 , Giala M M 1 1 Anaesthesiology and Intensive Care, AHEPA University Hospital, Aristotle University of Thessaloniki, Thessaloniki, Greece

Proper heparinization is perhaps the most important aspect of sampling technique for arterial blood gas and pH analysis.

(1) The aim of our study was to evaluate two sampling techniques for arterial blood gases and pH measurements and their possible cost implications.

We obtained 103 paired samples from 20 postoperative spontaneously breathing patients in ICU having an indwelling arterial line. Body temperature and blood Hb of patients were within normal ranges. Commercially available preheparinized (QUIK A.B.G.TM, Marquest Medical Products, CO, USA) and self-prepared with liquid sodium heparin syringes were used.

(2) The PaO2, PaCO2 and pH values were obtained from the same analyzer within 5 minutes of sampling. The cost of each sampling technique was also estimated. Data were analyzed by Bland and Altman analysis.

The mean differences (+/-SD) between the results of the sampling techniques were -2.56 (+/-12.4) mmHg for PaO2, 0.44 (+/-1.87) mmHg for PaCO2, and -0.003 (+/-0.012) for pH. The 95% confidence interval was 0.97-0.99 for PaO2, 0.78-0.89 for PaCO2, and 0.81-0.90 for pH. The correlation coeffiecient (r) between measurements from the two syringes was r=0.98 for PaO2, r=0.85 for PaCO2, and r=0.86 for pH. The cost of each sampling technique was 0.9 Euros for preheparinized and 0.104 Euros for self-prepared syringes, per sample.

Our data demonstrated a relationship between the results of the two sampling techniques, close enough, to justify the use of self-prepared heparinized syringes for arterial blood gas and pH measurements. A significant 9-fold cost reduction would result by the routine use of this latter technique. 

We checked the performance degree of the following measures in two stages, pre and post intervention: prophylaxis of pulmonary embolism (PE), elevation of the head of the bed to >30 degrees, intensive insulin therapy, lower tidal volumes in acute lung injury, daily trial of spontaneous breathing, daily withdrawal of sedation, stress ulcer prophylaxis and use of nimodipine in subarachnoid hemorrhage.The intervention consisted in the handing over of written information and talk on the therapeutics measures to the medical staff and to the nursing. CONCLUSION: With only one diagnose, the system is unable to identifie the complex cases. The stay and most of the ICU procedures, do not interfere nor modify the weight of the DRG.

Dias F S 1 , Nagel F 1 , Wawrzeniak I 1 , Fonseca C 1 , Guerreiro M 1 , Froemming J 1 , Canabarro M 1 1 General ICU, Hospital São Lucas da PUCRS, Porto Alegre, Brazil

There is no information regarding the impact of the resident in critical care medicine on the outcome of critically ill patients in our environment.

We performed a comparative study between two periods in an ICU, the first without CCMSP (PI) and the second with CCMSP (PII). We collected prospectively the following data: gender, age, APACHE II and MODS in days 1 (D1), 3 (D3) and 7 (D7), prevalence of sepsis/septic shock, duration of mechanical ventilation (MV), use or renal replacement therapy and ICU survival.

Between January 1998 and June 2003, there were 2432 ICU admissions, of which 932 in the PI period (1998) (1999) and 1500 in the PII (2000) (2001) (2002) (2003) . 

After the implementation of a CCMSP, despite an increase in organ dysfunction in D1 and in prevalence of sepsis/septic shock, there was a significant reduction in the utilization of renal replacement therapy and mortality. These findings suggest that, the participation of a CCMSP medical resident was an important factor in the support of septic patients, reduction in renal replacement therapy and mortality.

Pachl J 1 , Haninec P 2 , Tencer T 1 , Tomas R 2 , Mizner P 1 1 Anaesthesiology and CCM, 2 Dept. of Neurosurgery, Charles University, 3rd School of Medicine, Prague, Czech Republic INTRODUCTION: Delayed cerebral ischemia due to vasospasm is a major cause of death and disability in patients after subarachnoidal hemorrhage (SAH). The outcomes of several experimental studies designed to investigate an effect of nitric oxide donors on the treatment and prevention of this life-threatening condition appear controversial [1, 2] . The purpose of our study was: 1) to specify the influence of prophylactic subarachnoidal administration of sodium nitropruside (SNP) on the incidence of vasospasm 2) to determine the role of brain tissue monitoring-PbtiO2, PbtiCO2 and pHbti, measured in the area of high risk of vasospasm, for management of SNP administration.

Prospective observational study on patients with non-traumatic SAH (Hunt-Hess grade I-IV) with secured ruptured aneurysma. In postoperative period all patients underwent triple-H protocol with calcium channel blocker. Subarachnoidal preventive SNP was administred in initial dose of 1mg by catheter which was inserted to basal cisterns during neurosurgical procedure. The timing of following dosage (period of 6 or 24 hrs) was directed by the changes of PbtiO2, PbtiCO2 and pHbti after SNP administration. SNP administration did not exceed a period of 12 postoperative days. The brain tissue respiratory values were estimated by Codman Neurotrend Multiparameter Sensor®. The blood flow velocity was simultaneously measured on circuit of Willis by transcranial Doppler sonography (TCD). In case of detected signs of vasospasm the dosage of SNP was increased and maintained by monitoring modalities (TCD and values of PbtiO2, PbtiCO2, pHbti).

: 16 patients were enrolled. No brain infarction was developed in the studied group. All patients survived. The vasospasm was identified in two patients by TCD and simultaneously by changes of tissue respiratory values.These patients arrived at hospital with delay of several days from the beginning of symptoms. The overall outcome was good in 11 out of 16 patients including patients with vasospasm.

Preventive subarachnoidal administration of SNP controled by TCD and brain tissue multiparameter sensor might increase the effect of triple-H protocol with calcium channel blocker. Multimodal brain tissue monitoring could be the way to maintain titratable prophylactic SNP administration. The therapeutical intervention requires considerable raise in doses frequency and its effect cannot be evaluated until now. 

Basílio C 1 , Rio E 2 , Barbosa S 1 , Paiva J 1 , Mota A 1 1 Department of Anesthesiology and Critical Care, 2 Department of Neurology, Hospital S. João, Porto, Portugal

The search for drugs to minimize neuronal lesions after prolonged seizures has been the goal of treatment of patients with status epilepticus. Topiramate is a new anticonvulsant with multiple mechanisms of action: potenciation of GABA, blockade of glutamate receptors (AMPA), inhibition of sodium and calcium channels. Recent studies state that topiramate is effective in treating refractory status epilepticus and may reduce post-status epilepticus neuronal lesions.

Description of two clinical cases of patients, admitted in the ICU, with status epilepticus refractory to conventional therapy were treated with topiramate.

Case 1: Woman, 77 years old, with a history of alcohol abuse and psychiatric disease was admitted with encephalitis and tonic-clonic seizures. The patient was treated during 16 days with hidantin, valproate sodium and thiopental. Despite this, the EEG showed a periodic epileptiform activity. Topiramate (100 mg daily) was added to valproate with clinical improvement and an absence of ictal discharges on EEG after 4 days. Case 2: Male, 77 years old, admitted with hematemesis, shock and eventual cardiopulmonary arrest. Admitted in the ICU for postoperative care of duodenal ulcer surgery. Two days later, the patient developed partial status epilepticus. He was treated with midazolam, clonazepam and phenytoin. The EEG showed bilateral periodic epileptiform discharges with no recent cerebral lesions in the cerebral CT scan. Pentobarbital coma was induced for seven days. Status epilepticus persisted despite appropriate measures and on the 13th day topiramate (100 mg daily) was added to clonazepam. After 2 days, there was clinical improvement and EEG showed periodic generalized slow wave activity with motor response to painful stimulus wich was a prediction for a better outcome.

In both cases, topiramate was able to induce clinical improvement and disappearance of ictal discharges in the EEG in case 1. However, in case 2 the pattern of EEG persisted with signs of better prognostic.

It needs further investigation with larger prospective series to better confirm the results. 

Van Tulder L 1 , Chioléro R 1 , Regli L 2 , Revelly J 1 , Berger M 1 1 Surgical intensive care unit, 2 Neurosurgical, University -CHUV, Lausanne, Switzerland

Hypovolemia is deleterious in patients developing a vasospasm after subarachnoid haemorrhage (SAH). Fluid resuscitation to induce hypervolemia is considered by many as the cornerstone of management. The efficacy of this approach is however not established.The aim was to assess the effect of fluid resuscitation on blood volume and fluid balance during the initial phase of ICU management for cerebral vasospasm after SAH

Retrospective analysis of the database of a clinical information system (Metavision, iMD Soft). Patients with the diagnosis of vasospasm after SAH (angiography) were studied. Cardiac index (CI) and intra-thoracic blood volume (ITBV) measured with transthoracic thermodilution (PiCCO, Pulsion) were determined as part of the clinical management. Fluid supply consisted of isotonic saline. The value of mean arterial pressure (MAP), CI, ITBV, as well as fluid supply and fluid balance at time 0, 6, 12, and 24 hours were analysed (Presented as mean±SD). Comparisons between these time-points were performed with one-way analysis of variance for repeated measurements. P<0.05 was considered significant.

RESULTS: Ten patients were studied. Fluid supply amounted to 1.0±0.5 l at 6H, 2.7±1.0 at 12H, and 4.8±1.5 at 24H (p<0.05). Initial MAP was 110±12 mmHg, CI 3.5±0.7 l/min/m^2, and ITBV 863±295 ml/m^2. There was no significant change over time for these variables, although ITBV increased to 1102±290 at 6H, and decreased to 924±212 ml/m^2 at 24H. Cumulative fluid balance amounted to 0.3±0.5 l at 6H, 0.3±0.4 at 12H, and minus 0.3±0.7 at 24H (not significant), due to a diuresis.

Despite aggressive volume loading with normal saline, the fluid balance of patients with vasospasm was not significantly altered over the first 24 hours of ICU admission. These preliminary data suggest that, these patients may become rapidly resistant to fluid loading due to induced natriuresis. This escape phenomenon, may contribute to the absence of documented benefit of fluid expansion in vasospasm 

This was a prospective study design. All cases of SAH(n=15)of ages between 18 to 72years which were admitted over a period of thirteen monthswere included in this study. Aneurysm detection criterion was Digital Substraction Angiography or CT Angiography. Mean Hunt-Hess Grading in coiling group was 3, whereas in clipping group it was 4.16.Fisher Grade was >2 in all patients clipped and between 1-5 in patients coiled. Exclusion criteria included cases withA-V malformation, preexisting neurological deficit or where no intervention was done. Outcome analysis was done using Modified Rankins Scale (MRS) and World Federation of NeurologicalSurgeons GradingWFNS).Stastisticallyanalysed using Chi-square and Standard Error of difference between two Means-tests RESULTS: 60% aneurysms (n=9) were coiled and 40%(n=6) were clipped. There were no significant differences in age race, gender, but there was a significant difference in the Hunt-Hess Grading ( 

Ajderian S Stepan 1 , Petrov N 1 1 Department of Anesthesiology and ICU, Military Medical Academy, Sofia, Bulgaria

Background: The modulation of the intracranial pressure in patients with head injuries is important precondition in the optimizing of the therapeutic management. The goal of the authors is to study the influence of the hypocapnea over intracranial pressure as a part of complex treatment of the head injuries METHODS: : There are encompassed 12 patients with head injury CT-scan data for brain contusion and perifocal edema who assessed according Glasgow Comma Scale under 8 pct. All patients had ventricle drainage placed in one of the lateral brain ventricles in order to measure intracranial pressure and received standard therapy. The patients divided in two groups: I -in 6 patients we applied controlled hypocapnea with ????2 value between 30 -35 mm Hg and II group -6 patients with ????2 values between 25-30 mm Hg. Statistical program was used -SD, tcriteria and p-value.

We received significantly dropping of the intracranial pressure in patients set at artificial ventilation with moderate hypocapnea. In first group the mean value of intracranial pressure was 22.16 mm Hg. In this group 3 (50 %) patients died. 

Botsis P 1 , Litis D 1 , Nikolopoulou I 1 , Chatzivasiliadis H 1 , Ioannidou H 1 1 ICU, KAT Hospital Athens Greece, Athens, Greece

Acute pancreatitis and acute acalculus cholecystitis are frequent and serious complications in severely head traumatized patients. The aim of this study is to find if the early administration of somatostatin plays a role in the prevention of these complications.

METHODS: 50 brain injured ICU patients ( 35 men and 15 women), with no abdominal or thoracic trauma, were entered the study. Age 40 +/-15, apache II score ≤ 10, GC5 8 +/-2 on admission, with no signs of preexisting gallstones in the u/s of the upper abdomen. Surgical on conservative therapy for the brain injury, with drugs known not to interfere with the pancreatic and biliary system and early E.N. via nasogastric tube, was performed. After the admission in ICU they were divided into two groups: A) At Group A (25 patients) somatostatin (250 ?g/h) was added in the standard therapy for the next 10 days. B) At Group B (25 patients), the standard therapy was continued, as planned. Daily laboratory tests for: temperature, blood type, bilirubin, blood sugar, Ca, alkaline phosphatase, serum-urine amylase, serum-lipase and daily clinical examination as well for: pain and tenderness of the abdomen, vomiting, distention, decreased bowel sounds. Every second day an u/s of the upper abdomen was performed.

At 4 patients of Group B (8%) and at 1 patient of Group A (2%) acute pancreatitis was developed the 5th day-of-stay in ICU. At 8 patients of Group B (16%) and at 6 patients of Group A (12%) acute acalculus cholecystitis was developed as well at the same day. Diagnosis of both complications was based on clinical and laboratory findings.

Early administration of somatostatin in brain injured ICU patients diminishes the possibility of development of acute pancreatitis, but does not influence the development of acute acalculus cholecystitis in these patients.

Rijnsburger E R 1 , Girbes A R 1 , Spijkstra J J 1 , Peerdeman S M 2 , Polderman K H 1 1 Department of Intensive Care, 2 Department of Neurosurgery, VU university medical center, Amsterdam, Netherlands

Hypothermia is widely used to improve neurological outcome in various types of neurological injury; however, this has not yet been well studied in patients with subarachnoid haemorrhage (SAH), where cooling has been used mainly to prevent or treat vasospasms [1] . Hypothermia has been used to treat refractory intracranial hypertension in patients with TBI and severe stroke; however, its potential to treat cerebral oedema in patients with SAH has not been well studied [1] . Only one small feasibility study dealing with this issue has been published, but here various interventions (such as induction of barbiturate coma and mild hypothermia) were applied simultaneously [2] . Thus it remains to be determined whether induction of hypothermia per se can decrease ICP in patients with SAH.

METHODS: 36 Patients admitted with SAH and refractory intracranial hypertension (ICP>25 mmHg lasting longer than 5 minutes despite prevention of hypovolemia or induction of hypervolemia, induction of hypertension, and treatment with nimodipine, mannitol and hypertonic saline, and following coiling or clipping of the cerebral aneurysm) were treated with induced hypothermia (32-35oC) according to a protocol guided by ICP.

Hypothermia was induced using cooling blankets and infusion of refrigerated fluids. Target temperatures were achieved within 82 (range 32-152) minutes. ICP decreased from 36.2±15.4 to 16.8±10.8 (normal value: <15mmHg). ICP<25mmHg was achieved in 34/36 patients; in 2/36 patients ICP decreased but remained at levels between 25-30. Hypothermia was maintained until normal ICP had been observed for 12 hours, after which patients were slowly rewarmed (again guided by ICP). Hypothermia was maintained for an average of 74±41 hours. No patients died during treatment with hypothermia. Two patients (5.6%) died in the ICU after hypothermia was discontinued; 7 (19.4%) died in the subsequent 3 months. Good functional outcome at 3 months (Glasgow Outcome Score 4-5) was achieved in 14 patients (38.8%). Previous studies had reported a high incidence of side effects such as pneumonia in patients treated with hypothermia. We observed no increase in infectious problems, perhaps because our patients were treated with SDD.

CONCLUSION: Induced hypothermia can be safely and effectively used to treat refractory intracranial hypertension in patients with SAH. Vasospasms and intracranial hypertension are thought to be the two key factors in the development of additional brain injury in SAH; however, it remains to be determined whether this treatment also improves neurological outcome and survival in these patients. 

To evaluate the usefulness of magnetic resonance images (MRI) in patients suffering from severe brain injury, unfavourable clinical progress and Marshall brain scans types I and II.

Fifteen patients with severe brain injury were retrospectively studied, considering their age, gender, initial GCS initial head scans and MRI upon admission to Intensive Cares Unit (ICU), their outcome (GOS) at discharge form ICU. MRI level I was defined as being when the subcortical white matter was affected, MRI level II being level I plus affectation of the corpus callosum, and MRI level III was defined as being MRI-II as well as damage to the brain stem and spinal cord. GOS was also evaluated, defining GOS I-II as positive and GOS III, IV and V as negative.

The average age of the patients studied was 24.8 years old. Nine (60%) were males and six (40%) were females. The average GCS on admission was 5.53.

Five patients (33%) had and initial Marshall scan type I and ten (66%) had Marshall scan type II. Two patients (13.3%) showed MRI-I, five (33%) MRI-II and eight (53.3%) MRI-III. In the two patients with MRI-I their initial GCS was 6.5 and both progressed favourably and were discharged from intensive care unit. The five patients with MRI-II had an initial GCS of 6. Three of these (60%) did not progress well. In the eight patients with MRI-III, the initial GCS was 5.2. Seven of these ( 82.5%) progressed unfavourably when discharged from intensive care.

Magnetic resonance images are related to the severity of head damage and have a high diagnostical and prognostical value for use with patients suffering from diffuse axonal lesions.

Pakulski C 1 , Badowicz B 1 , Bak P 2 , Kwiecieñ K 3 , Mikulski K 3 , Surudo T 3 1 Department of Emergency Medicine, Pomeranian Medical University, Szczecin, Poland, 2 Intensive Care Unit, Regional Hospital, Pasewalk, Germany, 3 TraumaCentre, Pomeranian Medical University, Szczecin, Poland

The management of patients with severe head injury should include monitoring of mean arterial pressure (MAP) and intracranial pressure (ICP), cerebral perfusion pressure (CPP) and levels of jugular bulb oxygen saturation (SjO2). The aim of the study is to present the outcome in patients with severe central nervous system injuries treated in the TraumaCentre, Pomeranian Medical University.

This retrospective study evaluates the methods of treatment in 103 patients with severe brain injury treated between July 1st, 2001 and December 31st, 2003 in the our TraumaCentre. These patients were admitted to our institution directly from the accident sites or from the referring hospitals during the first post-injury day. Glasgow Coma Score of 8 or less was the inclusion criterion. In all patients MAP and ICP values were monitored, CPP values were calculated, and additionally in 51 patients SjO2 values were measured. The initial treatment protocol was always the same: analgosedation (fentanyl, midazolam), normoventilation, osmotic diuretics (mannitol 0,5-1,0 g/kg/day in 6 doses and furosemide 0,05-0,1 mg/kg/dose in 6 doses), supine position. The protocol was modified with regard to MAP, ICP and SjO2 values (brain ischemia or brain hyperemia). The patients with increased ICP values resistant to osmotic diuresis were scheduled for unilateral or bilateral decompressive craniectomy. The results of treatment were evaluated with Glasgow Outcome Classification after 12 months following the injury.

The mortality in our ample was 34,95% -36 deaths out of 103 treated patients. Isolated brain injury was the cause of death in 12 patients, and in 24 patients-multi-organ injury. Sixty seven (65,05%) patients were transferred from the Trauma ICU for further treatment to other wards. Out of the patients discharged from the Trauma ICU 10 patients died -GOC 1 (14,92%), none of the patients were in the neurovegetative state -GOC 2 (0%), 5 patients with persistent aphasia or hemiparesis were classified as GOC 3 (7,46%), 19 patients with mild neurological deficits that didn't impair their social life were classified as GOC 4 (28,37%), and finally 33 patients without any neurological sequelae were classified as GOC 5 (49,25%). Out of 51 patients with treatment modified according to SjO2 values 30 patients survived. Brain hyperemia was found in 14 non-survivors and severe brain ischemia was found in 7 non-survivors.

The outcomes in our patients treated with the protocol based on monitoring of CPP and SjO2 are encouraging. Monitoring of SjO2 is a significant element of the modern treatment protocol for patients with brain injury and the best method of diagnosing both hyperemic and ischemic episodes.

Anastasiou E 1 , Tsaousi G 1 , Giannakou M 1 , Efthimiou K 1 , Geka E 1 , 

Albanèse J 1 , Boyadjiev I 1 , Chaabane W 1 , Antonini F 1 , Leone M 1 , Martin C 1 1 Intensive Care and Trauma Center, Hopital NORD, Marseille, France

In severely head-injured patients, it is often needed to add vasopressive amines to maintain adequate cerebral perfusion pressure (CPP). Norepinephrine (N) and Dopamine ((D) are proposed, but their vasoconstrictive effects may be deleterious for regional circulations.Objective is to compare the effects of D and N on cerebral, splanchnic, and renal circulations when used to raise CPP after severe head injury.

Prospective, randomized, cross-over study including 15 patients with head trauma, requiring intracranial pressure (ICP) monitoring and vasopressor therapy. After 30 and 120 min of administration of D or N, were studied : systemic hemodynamics (mean arterial pressure (MAP), cardiac index (CI), central venous oxygen saturation (SvO2), cerebral circulation (ICP, CPP, transcranial Doppler : mean velocity in the middle cerebral arterey (Vmca)), splanchnic circulation (gastric intramucosal pH (phi)), renal circulation (urin flow (UF), creatinine clearance (Clcreat) and metabolic data (energy expenditure) (ES), oxygen consumption (VO2), and lactate (Lac)). The Wilcoxon Signe Test was used with p< 0.05 considered significant.

They are presented in table 1. No significant differences were observed in systemic hemodynamics when the two drugs were compared. None of the studied local circulation were altered with any of the studied drugs. 

Vakalos A 1 , Doukelis P 1 , Kareklas M 1 , Setzis D 1 , Matamis D 1 1 I.C.U, Papageorgiou General Hospital, Thessaloniki, Greece

In patients with severe head injury the main complication is cerebral edema and intracranial hypertension that may cause cerebral ischemia, disability and in certain cases brain death. Transcranial Doppler (TCD) is a non-invasive, bedside technique which detects the blood flow velocities in the great intracranial arteries. The aim of our study was to investigate if there is a relationship between TCD findings and the outcome of patients with severe head injury.

METHODS: 59 patients with severe head injury (GCS<8) were included in our study. From these 59 patients 49 were males and 10 females. Their mean age was 36.8 years, with a range from 12 to 71 years. Among these 59 patients 20 died in the ICU (33.89%). In each TCD examination we measured the maximum, mean and the end diastolic velocity (Vmax, Vmean and Vmin respectively), and we calculated the pulsatility index (PI). The patient's outcome was recorded at the disharge from the ICU according the Glasgow Outcome Scale as following: Good recovery (GR), Moderate disability (MD), severe disability (SD), persistent vegetative state (PVS) and death. The patient's outcome was compared with the cerebral flow velocities and the PI index.

There was a statistically significant difference in the mean values of all velocities between the outcome categories of the patients. We found the stronger difference in the mean values of Vmax and Vmin between the death and the categories PVS and SD, and in the mean values of Vmean between the death and the categories PVS, SD and MD. 

Inter-hospital transfers of the critically ill patient raises important medical and ethical dilemmas 1 . In 1999 a transfer questionnaire assessed the views of intensivists in Scotland, regarding the problem, that when no intensive care bed is available, is it ever acceptable to transfer an existing patient to another facility to create a bed for a new referral 2 .7% of Scottish consultants would not transfer a stable patients to create space for a new patient under any circumstances. Reasons given included no intrinsic benefit to the current patient and that there was a designated transfer team who were experienced in transferring critically ill patients and providing critical care without walls.

This time the questionnaire was repeated among intensivists in South Thames to determine any regional variation. The questionnaire was sent to 65 consultants in 15 intensive care units (ICUs). Consultants were asked if they would ever consider transfer of an existing patient to another hospital in order to admit a new referral and if so what would they consider the most compelling reasons for doing so.We also asked whether or not formal consent was sought prior to undertaking a transfer and what risks, if any, were explained to the patient and their family. Prior knowledge of their wishes regarding this form of therapy is essential in order to preserve their autonomy. Objective: to examine the knowledge of COPD patients related to the illness, about mechanical ventilation as a potential treatment and their wish to participate in the health care decision-making process and advanced care planning.

METHODS: a qualitative research of an intentional sampling of homogeneous subgroups with COPD outpatients (II and III by GOLD score) was performed between November 2003 and March 2004 in an ambulatory setting by means of semistructurated interviews and later content analysis with a sample size defined by saturation criteria.

: 40 male outpatients with COPD were interviewed (age range from 48 to 85 years). They feel to be correctly informed and trust their respiratory physician or family doctor, but in most cases there have not been prior discussions with the health care team concerning MV as a potential treatment of their disease. They consider themselves to have a good quality of life although their health is not good. They are interested in participating in the health care decisionmaking process. In case of treatments and cares as MV or admittance to an Intensive Care Unit they accept any option that keeps their usual quality of life at the same level. This sample of COPD patients are in favour of advanced care planning and show the aim to shape their own specific advanced directives.

CONCLUSION: patients with COPD do not have enough information to take autonomous decisions. Although MV is a potential treatment for COPD patients with acute exacerbations, most of them were unaware of MV as a possible treatment option for them because discussions about this topic occur infrequently between physicians and patients. They are in favour of participating in health care decision-making with physicians and accepting any therapy that makes them able to keep their quality of life. The patients consider advanced directives as an opportunity to express their preferences in order to be considered when they are not able to communicate with the health care team. CONCLUSION: WH/WD in Ireland is common (69%) and similar to European practice(72%). There was no SDP although the prevalence was 2% in Europe. The increased use of sedation (9%) in association with WD suggests an awareness of patient comfort. Despite only one advance directive, patient wishes were known in 28%. ICU physicians were the primary initiators of EOL discussion (59%) suggesting their important role in ICM practice. The majority (55% v 45%) of EOL decisions were taken during 'office hours'. This is finding warrants further study; it may represent the inexperience of on call personnel or may reflect the complexity of the decision. Despite popular opinion extubation is infrequent and comfort measures tend to be continued. 

An anonymous questionnaire including 16 questions was sent out to all staff mebers including medical staff, nurses and physiotherapists. Questions included potential benefits to staff and patients and concerns regarding the use of restraint.

Of the 135 questionnaires sent out, 63 were returned completed. Most respondents (60, 96%) felt that sedative drugs are used as a form of restraint in care of the patients. Most (59, 95%) thought there is a place for physical restraint in critical care and that it might benefit patients, 54 (87%) believed it might benefit staff but 36 (56%) of responders had concerns regarding the use of physical restraint. A majority (70%) stated that concerns would be answered if undertaken as part of a clear unit policy. 58 responders (93%) would be happy to use physical restraint with sedation to ensure patient's safety, but only 51 (82%) would agree to its use if they were a patient.

The majority of responders (93%) felt that some form of physical restraint may be appropriate. Practice should be re-examined and consideration given to the use of some physical restraint in addition to sedative drugs. In some patients shock is so severe that extremely high doses are needed to elevate their blood pressures. Studies show that 90% of ICU physicians withhold or withdraw vasopressor administration because patients did not respond to "maximal" therapy. However the "maximal" dosage of vasopressors is not defined so each physician has his/her own limit as to the highest dose of adrenaline or noradrenalin that he/she will administer to a patient. Many ICUs physicians order doses of up to 50mg of adrenaline or noreadrenalin per hour (1 mg/kg/hour) we hypothesize that this dose is futile.

Following Helsinki approval, all intensive care charts from 2001 were reviewed (689 patient charts 3444 patient days) 72patients were found to have received a vasopressor. Demographic data as well as APACHE II scores, ICU days and total hospital days, biochemistry, liver functions, blood gases, diagnoses and secondary complications were recorded. Vasopressors, adrenaline and noradrenalin, maximal doses , initial dose, number of days/hours patient received all subsequent doses, and mortality were recorded.

The data showed that all patients who received more than 2 microgram per kilogram per minute of adrenaline or noradrenalin died. (p value < 0.0001). The length of time that the patients received vasopressors had no influence on survival. These data showed a direct correlation between the number of days a patient received vasopressors and the length of hospitalization. The length of time a patient received low dose of vasopressors had no significance on mortality, but the vasopressor dose had an indirect association with survival. There was no significance difference in age between survivors and non-survivors. The elderly (over 75 years of age) and the young had the same of survival rates when receiving vasopressors.

Patients who received more than 2 microgram per kilogram per minute of noreadrenaline or adrenaline died. It thus appears that therapy with such high doses is futile.

17th Annual Congress -Berlin, Germany -10-13 October 2004 S103

Ehrmann S 1 , Mercier E 1 , Bertrand P 2 , Dequin P 1 1 Medical Intensive Care Unit, 2 Department of Biostatistics, Bretonneau University Hospital, Tours, France

To carry on indefinite invasive treatments in the intensive care unit for patients with a high probability of death in the short term, is ethically objectionable. On the individual level it extends the agony and suffering of the patient and on the community level it consumes precious limited resources. As there is no tool that can objectively and reliably help the physician to make ethical decisions of therapeutic limitations, we examined whether the absolute change of the LOD score between the day of admission to the ICU and the third day of unlimited treatment could be predictive of death in the ICU.

METHODS: 154 consecutive patients admitted to the ICU were prospectively included during a three-month period. For all of these patients the Simplified Acute Physiologic Score II (SAPS II) and the LOD score (LOD1) were calculated on the day of admission. The LOD score was calculated again between the 72nd and the 96th hour in the ICU (LOD2) for the 93 remaining patients (52 patients left the ICU before the 72nd hour, 7 of whom died) without exclusion criterions (9 therapeutic limitations before the 72nd hour). The endpoint was death in the ICU. The performance of the DeltaLOD = LOD 2 -LOD1 index to predict death was examined through univariated and multivariated analysis and through calculation of the positive predictive value of death (PPV) for different cut-offs. 

After 72 hours of unlimited treatment in the ICU, DeltaLOD appears to be a good predictor of death in the ICU, independent of the initial severity of disease. The PPV is not high enough even for high cut-offs to assist with making individual therapeutic limitation decisions. Accordingly to the Bayes theorem, the performance of DeltaLOD deserves to be evaluated in a population of patients exhibiting greater severity of disease. developed as an answer to the stressful everyday practice. Indeed, in ICU, the caregivers are under a great pressure induced by several factors: the explicit urgency to act, the implicit burden to face patients and families in critical situations and the requirement of a high level of technicity. Such a creative workshop was set up for physicians and nurses in our unit, in order to give the opportunity to express the "unspoken", to share the experiences, explore the patient's perspectives.

METHODS: 2 sessions of 2 hours were planned. At the first session, the participants were proposed to imagine a situation starting from one of 2 pictures of ICU patient. They were invited to write as "I" or "you" in order to take the patient's or the relative's place through their imagination. The written texts were distributed to all participants and discussed at the second session. The workshop was organized and moderated by a senior specialist in intensive care medicine and the person in charge of the Medical Humanities teaching program. The participation to the workshop was optional.

: 9 physicians (4M, 5F) and 11 nurses (all F) and 3 (1F,1M) medical students participated to 4 workshops. One month later, they answered a questionnaire. All participants gave a positive global appreciation and underlined the importance of the discussion which allowed the sharing of their experiences. 4 felt encouraged to adopt the patient's perspective, 3 were reassured about their feelings of their practice. The relevant themes of the texts were the behaviour of the caregivers, the abrupt change of worlds, a strange perception of time, the importance of noises in the ICU environment. Despite the similarity of the themes, the way the narrators shaped the story was very different. The author's selection of words, details, and literary devices confers the personal touch of his/her experience.

None of them considered the writing as an obstacle to their expression, even if they first felt difficult to step in the writing process. All of them were satisfied with the workshop and with the sharing of their experiences. The most important point reported was the awareness that the others, either physician or nurses, had the same preoccupations and feelings. The fact that this workshop did not give any concrete recipe for the resolution of problems induced some frustration among ICU caregivers. 

There have been few studies to investigate how well the results obtained by CO2 rebreeding to assay cardiac output (CO) (1-3). A reliable non-invasive CO monitor could enhance patient safety and reduce risk. This study evaluates a NICO measurement and calculated derived parameters from CO. CO was based on differential from of the CO2 Fick equation.

Twenty three CO measurements with derived parameters were obtained from 3 male patients admitted to a medical intensive care unit, St Vincent Hospital, Medellín, Colombia, 2003. The NICO monitor (Novametrix Medical Systems Inc) was connected between the ventilator circuit and tracheotomy tube. Previously, multilumen Swan-Ganz Thermodilution catheters, Edwards Labs were placed into the external jugular vein via an introducer sheath. CO was calculated from pulmonary blood flow by correcting for shunt. The difference between consecutive thermodilution and NICO measurements was calculated. Also, calculated derived parameters from S-G catheter such as CI, SVRI, LVSWI, RVSWI, and SVI were compared to measurements derived from NICO. On NICO measurements, central venous pressure replaces the value of PCWP. Correlation between the two methods was determined by Pearson´s correlation. A Bland-Altman analysis was used to compare the bias and precision of the two methods, and a difference > 30% was considered as a limit of accuracy. Significance was assessed at the 95% confidence interval.

Twenty-three matched pairs of consecutive changes in CO and calculated derived parameters measurements were recorded in three critically ill patients. With a mean (±SD) age of 30,6 y. Relationship between changes in thermodilution and NICO CO measurements was significant (r = 0.60, p = 0.002). None of calculated derived parameters (CI, SVRI, LVSWI, RVSWI, and SVI) were considered significant. Only CO and RVSWI showed difference between means to compare the degree of agreement measurements (CO 12.46% and RVSWI 19.8%% respectively).

The results of the current study agree with those from previous studies where is suggested that NICO monitor would provide a good alternative to invasive CO measurements on critically ill patients. However, NICO cannot be a substitute to get calculated derived parameters when pulmonary artery occlusion pressure is a necessary value. (STD) has been demonstrated to be sufficiently accurate for estimation of intrathoracic blood volume (ITBV) and EVLW when compared with the clinical standard, i.e., transpulmonary thermo-dye dilution (TDD) [2] . In this study, we examined the reliability of STD for estimation of ITBV and EVLW with respect to several influencing factors.

We retrospectively analyzed data of 174 critically ill patients patients (114 male, 60 female; age 10 -88, mean 52 ± 20 years) who underwent extended hemodynamic monitoring by the transpulmonary thermo-dye dilution technique. The agreement between ITBVSTD/ ITBVTDD and EVLWSTD / EVLWTDD was determined as mean bias and standard deviation (SD) within different categories (level of PEEP, PaO2/FiO2 ratio and EVLW). Linear regression analysis was applied to compare overall bias between EVLWSTD and EVLWTDD with the different factors.

: Mean bias ± SD within the different categories are shown in 

The clinical judgement of an adequate volume status in critically ill patients remains a challenge. Current clinical parameters to assess the adequacy of resuscitation often do not adequately reflect the volume status of the patient. Therefore additional information about the adequacy of circulating blood volume in critically ill patients could be of great value.

On 53 occasions in 28 critically ill patients on a surgical intensive care unit the adequacy of circulating blood volume (BV) was clinically judged by the parameters central venous pressure, mean arterial blood pressure, heart rate, and urine production. Clinically estimated blood volume was compared with measured blood volumes using pulse dye densitometry with Indocyanine Green (DDG-2001A/K, Nihon Kohden, Japan). Obtained BV measurements were categorized in low blood volume (LBV), normal blood volume (NBV), and high blood volume (HBV) using reference values for men and women 60-75 ml/kg and 55-65 ml/kg respectively(1).

Clinical judgements led to 9 hypovolemic (HV) versus 46 not hypovolemic (NHV) cases. There was no statically significant relation between the clinical judgement of volume status and measured BV. In HV patients no LBV was measured and in clinically NHV patients LBV as well as HBV were measured. No significant correlation between measured BV and calculated fluid balances was found.

CONCLUSION: There seems to be a discrepancy between the clinical judgement of circulating blood volume and the measured circulating blood volume in critically ill ICU patients. These results emphasize the difficulty of judging the volume status by current clinical parameters in critically ill patients. 

Schütz N 1 , Romand J A 1 , Stotz M 1 , Gerard I 1 , Bendjelid K 1 1 APSIC, Geneva University Hospitals, Geneva-14, Switzerland

We recently demonstrated the accuracy of a new miniaturized transcutaneous sensor (Tosca Monitor, Switzerland) to monitor non invasively PaCO 2 (TcPCO 2 ) in white skinned patients [1] . The objective of the present study is to analyse the same accuracy in a subgroup of dark skinned patients.

Eight post operative patients (mean 37 ± 22) were included. TcPCO 2 sensor was applied at the ear lobe. The simultaneously obtained TcPCO 2 and PaCO 2 values (measured using a blood gas analyser) were compared by linear regression analysis. The difference between PaCO 2 and TcPCO 2 values were compared using the method of Bland and Altman.

: 64 paired measurements were correlated. TcPCO 2 correlated with PaCO 2 (r 2 =0.56, p<0.0001) in the PaCO 2 range 3.2 to 6.6 kPa. The mean bias between the two methods was 0.13 ± 0.38%.

Our results demonstrate that skin pigmentation affects slightly the accuracy of the sensor. (1). In critically ill patients it is difficult to gain knowledge of the intravascular volume using the conventional clinical parameters such as mean arterial blood pressure, central venous pressure (CVP), heart rate and urine production. New insights in the assessment of hemodynamics such as central venous saturation (SvO2) and blood volume (BV) monitoring may give additional information of a patient's intravascular volume status.

METHODS: Blood volume measurements were performed in 28 critically ill ICU patients on 53 occasions using pulse dye densitometry with Indocyanine Green (ICG)(DDG-2001 A/K, Nihon Kohden, Japan). Blood volume measurements were compared with the parameters SvO2 and CVP in assessing the patient's intravascular volume status. Also the relation between BV and Albumine and Colloid Oncotic Pressure (COP) was investigated. 17th Annual Congress -Berlin, Germany -10-13 October 2004 S105

Hofmann D 1 , Sakka S G 1 1 Anesthesiology and Intensive Care Medicine, Friedrich Schiller University, Jena, Germany INTRODUCTION: Patient management guided by extravascular lung water (EVLW) is associated with reduced mortality of patients with pulmonary edema [1] . Recently, single transpulmonary thermodilution (STD) has been demonstrated to be sufficiently accurate for estimation of intrathoracic blood volume (ITBV) and EVLW when compared with the clinical standard, i.e., transpulmonary thermo-dye dilution (TDD) [2] . In this study, we examined the reliability of STD for estimation of ITBV and EVLW with respect to several factors of pulmonary function.

We retrospectively analyzed data of 174 critically ill patients patients (114 male, 60 female; age 10 -88, mean 52 ± 20 years) who underwent extended hemodynamic monitoring by the transpulmonary thermo-dye dilution technique. The agreement between ITBV STD / ITBV TDD and EVLW STD / EVLW TDD was determined as mean bias and standard deviation (SD) within different categories (level of PEEP, PaO 2 /FiO 2 ratio and EVLW). Linear regression analysis was applied to compare overall bias between EVLW STD and EVLW TDD with the different factors.

Mean bias ± SD within the different categories are shown in 

acute-on-chronic (AoC) liver failure but data on the use of this procedure is yet scarce. We communicate our experience after centralizing all the procedures of our centre in the ICU

Data of all treatments performed in our centre (liver transplant program; ICU with continuous renal replacement therapies -CRRT-experience; a unique protocol for surgery, hepatology and ICU patients and a prospective registry). ALF patients stay in the ICU but AoC patients are admitted every other day for the procedure. MARS is performed with a Prisma monitor and sessions are aimed for a length of al least 15 hours if feasible (in ALF continuously with 24h changes). We analysed clearance, metabolic control, tolerability and technical and clinical complications. We used Pearson correlation coefficient and linear regression analyses to detect relation between hours of treatment and clearance capabilities 

Electrical impedance tomography (EIT) is a promising technique to assess continuously respiratory function with high temporal resolution1. Changes in thoracic gas volume lead to corresponding changes in thoracic impedance. The aim of this study was to evaluate air distribution during volume controlled mechanical ventilation.

Five adult patients undergoing elective thoracic surgery with single lung ventilation were included. EIT data were collected during ventilation of both lungs (tidal volume (tv): 800 ml), left lung (tv: 400 ml), and right lung (tv: 400 ml), respectively. EIT was performed using 16 electrodes placed around the thorax. Data are presented as percent of impedance change of both lungs (normalized electrical impedance (nei)).

During one lung ventilation nei was reduced to 47.7 ± 18.5 in right lung and 49.6 ± 13.4 in left lung compared with both lungs with a clear separation between ventilated and nonventilated lung. In addition we found an imbalance of distribution of ventilation along the vertical axis in favor of the ventral part of the lungs (Fig.1) .

CONCLUSION: EIT seems to be a sensitive non-invasive method for monitoring distribution of ventilation. 

The use of PET in adults with TMA has dramatically improved outcome. Resistance to PET, which is observed in 1/3 of such patients and may affect mortality, remains however incompletely understood. We retrospectively studied 25 adults with TMA treated by PET in our unit to evaluate the short and long term outcome and to identify predictive factors of mortality and of resistance to PET.

All records of adults with TMA treated by PET between 1999 and 2003 were reviewed. TMA associated with bone marrow transplantation were excluded from the study. Age, sex, cause of TMA were collected. Glasgow and SOFA scores were estimated at the admission. Clinical data including: neuroligical or pulmonary disorders with mechanical ventilation (MV), renal failure, and therapeutic delay (TD) to PET; biological data including: hemoglobinemia, platelet count, and LDH; plasmatic volume exchange per procedure and number of plasmapheresis sessions were also collected. Mortality was assessed at one month and at one year follow-up. All data were analyzed and compared between survived/deceaded and between responders/nonresponders (R/nR) patients.

: 19 females and 6 males were included. Mean age: 46.8±16.3 yo, mean Glasgow coma score: 11±3 , mean SOFA score: 5.8±2.8. Etiologies of TMA: post-immunologic 4, post-infectious 4, post-neoplastic 4, drugs associated 4, idiopathic 9. Two patients were in MV, 3 underwent hemodialysis and 10 had at least two organ dysfunction. The mean TD for PET was 2.7±5.2 days and the mean plasmatic volume exchange per procedure was 35.9±8.6 ml/kg. 19 patients (76%) partially or fully responded to PET. 20 patients (80%) survived after one month and 19 (76%) after one year follw-up. The comparison between survived and deceaded patients showed that response to PET (18/20 vs 1/5 responders respectively) was the only significant determinant parameter. The comparison between R and nR showed that a longest TD (2.5±1.8 vs 3.3±2.0) and neoplastic cause of TMA (3/19 vs 3/6) were significantly discriminant for a non-response to PET. Almost all of the R patients (95%) exhibited a positive response to PET before the tenth plasmapheresis session. In a median follow-up period of 23,6 (1-48) months, relapses episodes occured in 7 patients (35 %).

Adults with TMA, characterized by a mild to important severity, treated by PET have a relatively good outcome since survival reached 80% at 1 month and was maintained at 76% after 1 year follow-up. Among the parameters studied, lack of response to PET was the only predictive factor of mortality. Two factors were predictive of resistance to PET: neoplastic etiology of TMA and a longer TD to PET. Adult with TMA non-responding to PET after the tenth plasmapheresis session could be considered as totally non-responder and should benefit shortly from another therapy. 2) . Additional in-formation can be gained by visual conditioning of large data sets. We report a novel approach on data inter-pretation by visualizing the gastric tonometry values in a case of necrotizing pancreatitis and septic shock and the correlating clinical events ( (38y., male, APACHE II Score 24, 8308 measurements, 58 days running).

PgCO2 was measured every 10 min. with a gastric tube (TRIP, NGS catheter) and an automatic gas analyzer (Tonocap, Finland). We recorded these values continuously and calculated the frequency distribution in an 8h interval and in an area between 20 and 130 mmHg (increment 5). This histogram was displayed as a contour-plot. In that kind of visualization the frequency is displayed as colour in the area and not as third axis in a graph.

Our form of data processing provides additional information on pathological patterns at an early stage. In context with other parameters, this can be helpful in guiding treatment, e.g. volume substitution, catecholamines or blood transfusion. In this case we can see episodes of normal tonometric values as well as periods of pathological patterns like periods of septic shock (T1, T2), daily abdominal lavage in a period of severe sepsis (T3), major abdominal surgery (T4), unsuccessful trials of enteral feeding (T5), successful enteral feeding (T6), weaning period and extubation (T7).

The graphical presentation of the frequency distribution of a large number of data easily allows to conceive the information of the data. The aim of future activities has to be the development of a real-time bedside display of the progress of changes in the measurement of PgCO2-values. 

Our study population consisted of patients undergoing elective cardiac surgery (n=38). Blood samples for TNF alpha mRNA were taken preoperatively (baseline), 1hr and 6hrs postoperatively. Total RNA was extracted from purified peripheral blood mononuclear cells (RNeasy, Qiagen). We utilised real time RT-PCR to quantify TNF alpha gene expression after cardiac surgery using ABI PRISM 7000 Sequence Detection System and normalised against an endogenous reference GAPDH. The patients were divided into two groups: Group A: Eighteen patients who developed complications post surgery as defined by i) hypotension requiring inotropes (n=15) ± intra-aortic balloon pump counterpulsation (n=1) and /or ii) lactate > 4mmol/l (n=12). Group B: Control group of patients with an uneventful postoperative course (n=20). Statistical analysis was performed using the Kruskal-Wallis test. 

Ischaemia of the colon is a recognised but infrequent complication following cardiac surgery. Colonic ischaemia is thought to lead to a disruption in the intestinal barrier and this has been implicated in the progression to the systemic inflammatory response syndrome (SIRS) with some patients going on to develop multi-organ dysfunction syndrome (MODS). Little is known of the early pathophysiological processes occurring in the colon during cardiac surgery. Thus, the aim of this study was to investigate the early histological changes within colonic mucosa and cytokine release during CABG surgery.

METHODS: 20 patients undergoing coronary artery bypass surgery (10 on-pump, 10 off-pump) were prospectively recruited. Mucosal biopsies of the sigmoid colon were obtained after induction of anaesthesia and immediately at the end of the procedure. Microscopic examination was performed using Haematoxylin and Eosin staining. Peripheral blood was assayed intraoperatively for cytokines IL-6 and IL-10 and for up to 48 hours post-operatively.

On-pump surgery produced a 2.5-fold increase in columnar epithelium apoptosis. No other histological changes occurred. There was a 15-fold rise in IL-6 in both two groups intraoperatively. Post-operatively, IL-6 continued to rise to 30-times baseline levels in contrast with the off-pump group which remained at intra-operative levels. IL-10 did not change significantly in the off-pump group. In the on-pump group there was a 22 fold increase in IL-10 associated with initiation of cardiopulmonary bypass (p<0.05 Student's t-test). Post-operatively, IL-10 levels returned to baseline levels.

Apoptosis of colonic mucosa occurs during on-pump CABG but not in offpump CABG. This precedes the inflammatory process. Thus, we identify apoptosis, rather than necrosis, as the principal mode of cell death following on pump CABG surgery. Further elucidation of this process may identify targets for pharmaceutical prevention colonic mucosal apoptosis.

CABELLO B 1 , Rubio O 1 , Delgado M 1 , Vera P 1 , Mancebo J 1 1 intensive Care, Hospital Sant Pau, Barcelona, Spain

Liberation from mechanical ventilation can be interfered by the development of congestive heart failure (CHF). This issue has been poorly studied as a cause of weaning failure. We designed a clinical-physiologic study to analize the mangnitude of the problem and its physiological characteristics.

During a two months period we daily screened all the intubated mechanically ventilated patients in our 16-bed ICU looking for those who meet usual weaning criteria. These patients went on a t-piece trial (SB) during 30 minutes. Patients who presented respiratory distress were studied with esophageal-gastric balloon and a Swan-Ganz catheter. Hemodynamic and respiratory measurements were collected in Assist Control Ventilation (ACV), Pressure support (PS) of 7 cm H20 with PEEP of 5 and again SB. We defined CHF when the pulmonary wedge pressure (PCWP)was normal during ACV and above 18 mm Hg during SB. 

Unsuccessful extubation in COPD patients is associated with increased morbidity and hospital mortality, and accurate prediction of post-extubation acute respiratory failure (ARF) is potentially important. Our hypothesis was that two parameters i.e., 1 the airway occlusion pressure at 0.1s (P0.1) and 2 the expiratory flow limitation (EFL) determined by applying a negative expiratory pressure (NEP) during tidal breathing, both recorded repeatedly after extubation, could be good indicators of postextubation ARF in COPD patients.

COPD patients were included prospectively after extubation. A specially devised system (Micro 5000; Medisoft , Dinan, Belgium) was used to measure EFL and P0.1. Each patient was placed in half sitting position and breathed spontaneously. After stabilization of the patient, a NEP of -5 cm H2O was applied at the beginning of expiration and maintained throughout the ensuing expiration. The test breath was the breath during which the NEP was applied, and the preceding expiration served as control. Five test breath separated by periods of quiet breathing were recorded. The expiratory flow-volume loops generated with NEP were compared by superimposition with those obtained during the immediately preceding breaths. The portion of the tidal expiration over which there was no appreciable change in flow with NEP was considered as flow-limited and was expressed as a percentage of the expired control tidal volume (%VT). The module of NEP was replaced by that allowing to measure the P0.1. Five measurements of P0.1 were made, spaced by at least 15 seconds. Gas exchange, P0.1 and EFL under NEP were measured at the 1st , 6th, 24th and 48th hour following extubation. If a limitation of flow was evidenced at a given time, the subsequent measurements were not carried out. Post-extubation ARF was defined by a respiratory rate of more than 25 per min, a respiratory acidosis with a PaCO2 > 45.2 mmHg and a pH lower than 7.35 without metabolic acidosis.

To date, 25 patients have been included. Heighteen of them (72%) presented a ELF at 8 ± 10 hours following extubation. Nine patients presented an ARF at 54 ± 17 hours in post extubation. These 9 patients had a ELF and a P0.1 significantly higher than those without postextubation ARF (respectively 80.2 ± 8.4 % vs. 49.7 ± 12.1 %; and 5.2 ± 0.6 cm H2O vs. 2.0 ± 0.6 cm H2O; p < 0,05). Seven patients (28%) did not have ELF and did not present ARF in post extubation.

CONCLUSION: ELF by NEP and P0.1 are easily measured in the period following extubation in COPD patients. This preliminary report seems to demonstrate that P0.1 and EFL, measured precociously then repeatedly after extubation, could be good indicators of postextubation ARF in COPD patients.

Serebriysky I I 1 , Galstian G M 1 , Gorodetsky V M 1 . 1 Intensive Care Unit, National Centre of Hematology, Moscow, Russian Federation

Acute respiratory failure (ARF) is the most frequent and serious complication in patients (pts) with hematological malignancies. Respiratory insufficiency in this group of patients can be caused by a combination of increased vascular permeability, heart failure and liquid overload. The aim of this study was to analyze the effects of colloid replacement therapy in pts with hematological malignancies and ARF.

We examined 9 pts with acute leukemia, complicated by sepsis and ARF (bilateral radiographic infiltrates, PaO 2 /FiO 2 = 200 ± 60). Three of them had PAWP>18 mmHg, respectively 19, 22 and 33 mmHg. All the patients received infusions of 20% albumin (A.). The first Infusion of 150ml was carried out during 45 minutes, later the speed was 50 ml/30 min. We measured extravascular lung water index (ELWI), pulmonary vascular permeability index (PVPI) by PICCO-plus (Pulsion, Germany), central hemodynamics parameters by Swan-Ganz catheter.

In the first group that received A. in doses of 2.5 ml/kg (6 pts) there were no significant changes in CI, PAWP, ELWI, PVPI, apart from 1 pt (1st pt, see 

Should critical illness polyneuropathy (CIP) itself prolongs mechanical ventilation or whether this prolongation is the effect of concurrent risk factors for weaning failure is a matter of debate. Our primary objective was to evaluate the impact of CIP on the length of mechanical ventilation after controlling for coexisting risk factors for weaning failure. We also set out to assess the impact of CIP on the length of the stay as well as to determine the costs associated with this neurological complication.

A prospective cohort study. Setting: ICU of a tertiary hospital. Patients: All patients with severe sepsis or septic shock that required mechanical ventilation for at least 7 days who were considered ready to discontinue mechanical ventilation. Patients underwent a neurophysiologic evaluation at onset of weaning from mechanical ventilation. 

Gianesello L 1 , Pavoni V 1 , Paparella L 1 , Gritti G 1 1 Dept. of Critical-Medical Surgical Area, Section of Anaesthesia and Intensive Care, Florence, Italy

Extubation failure (EF) has an important effect on length of ICU and hospital stay,ICU and hospital mortality(1).EF can occur secondary to upper airway obstruction or to an inability to manage respiratory secretions a cause of laryngeal dysfunction (LD) and ineffective cough.LD can result from depressed mental status or local trauma after intubation.Pre-admission functional status can also delay post-estubation swallowing impairment in critically ill elderly patients(2).

Over a 12-month period patients who needed reintubation after successful trial of weaning and planned extubation, in a polyvalent intensive care unit (ICU) were identified.Data including clinical features (age, sex, SAPS II on admission, Glasgow Coma Score (GCS) on day of extubation, type of patient, length of intubation and mechanical ventilation (MV) before extubation, length of ICU stay (LOS), ICU and hospital mortality) were collected.Moreover we considered two parameters that asses airway patency and protection like predictors of EF:cough strength and suctioning frequency after extubation.Cough strength on command was measured with a semiobjective scale of 0 to 5 (0= weak cough, 5= strong cough). (3/23)(13.2%), pulmonary embolism(1/23)(4.3%)and severe sepsis(1/23)(4.3%).Seven of patients who received reintubation a cause of defective airway manage needed at least one suctioning every two hours; moreover the same patients and other three with alteration in neurological function had weak cough (grade 0 to 2).The LOS of EF patients was 23±24.3 days, their ICU and hospital mortality were 39.1% and 47.8%, respectively, both higher when compared with not reintubated patients.Results of logistic regression showed that SAPS II is the only independent risk-factor of reintubation (odds ratio 1.056, sig. 0.004),while age, type of admission,length of intubation and GCS seem to do not influence EF.Data were analysed using the SPSS 11.0 for Windows. CONCLUSION: EF can depend from defective airway protective mechanisms due to alteration in consciousness or glottic incompetence.This event influences negatively LOS and outcome.Severity of illness is the only independent risk-factor.

Over a 5 year-period, 16 patients with DILD (12 males; age: 62±18 years; SAPS II: 41±7) were retrospectively studied. Among them, 9 patients were immunocompromised and 4 patients sustained complications attributable to the procedure: airleak (n=3), pneumothorax (n=1). The median duration of the chest tube drainage was 7±6 days (range: 1 to 25 days). Both the PaO2/FiO2 ratio and mean level of PEEP were comparable before and 24 hours after the OLB (Table 1) . No patient died in the perioperative period. OLB allowed to establish a diagnosis in 14 patients (87%). In 11 patients (70%), the DILD was idiopathic (Table 2) . In 3 patients, histologic diagnoses obtained from OLB were not suspected clinically or by radiological investigations: invasive pulmonary aspergillosis, diffuse amyloidosis, methotrexate lung toxicity.

178 ± 94 173 ± 102 PEEP level (cm H2O) 10 ± 4 9 ± 5

Idiopathic interstitial pneumonia Diagnosis UIP AIP* BOOP N. of patients 3 3 5 *: Acute interstitial pneumonia (ARDS) CONCLUSION: In ventilated patients, OLB can be performed with acceptable morbidity at bedside in the ICU. In this study, OLB established a definite diagnosis in 87% of patients and corrected the clinical and radiological diagnosis in 44% of the cases.

Fonsato V 1 , Mariano F 2 , Triolo G 2 , Camussi G 1 , Nederlof B 3 , Tetta C 3 1 Laboratory of Renal Immunopathology, University of Turin, 2 Unit of Nephrology and Dialysis, CTO Hospital, Turin, Italy, 3 Department of Research, Fresenius Medical Care, Bad Homburg, Germany INTRODUCTION: High porosity membranes may enhance cytokine elimination by convection and also diffusion. However, there is need to balance the high permeability between cytokine removal and a clinically acceptable loss of plasma proteins. Here,we studied the sieving coefficients (SC)and clearances of different cytokines (TNFa; IL-1b;, IL-6, IL-8, IL-1ra) and protein permeability profile (albumin, cystatin C, IgG)in an ex vivo hemofiltration (HF), hemodiafiltration (HDF)and hemodialysis (HD)circuit of nanostructured high porosity polysulfone membranes with different albumin permeabilities of 5%(Type A)and 13%(Type B).

Three hundred ml of fresh normal human blood was incubated with endotoxin (1 mg, E. Coli, Sigma, 37°C, 4 hr and overnight at room temperature). We set up the three circuits under the following conditions: I) post dilutional HF, at three different blood flow rates (100, 150 or 200 ml/min) and with a fixed (20%) ultrafiltration rate (UFR: 1.2, 1.8 and 2.4 L/hour, respectively). The circuit operated at zero balance. Samples for SCs and Clearances were obtained conventionally at 10, 30, 60, 120, 240, 360 and 480 min; II) HD, at a dialysate flow rate of 3 L/h and 5 L/h; III) HDF, dialysate flow rate of 3 L/h and 5 L/h included 0.5 L/h of ultrafiltrate. Both HD and HDF were conducted always at blood flow rate of 150 ml/min. Cytokines were determined by commercially available kits, albumin, Cystacin C and IgG by nephelometry (Beckman).

Median SC was nearly up to 1 for IL-1b and IL-1ra, at about 0.6 for IL-6, 0.4 for IL-8 and 0.7 for TNFa (Type A vs B, p >0.05). Despite similar high cytokine clearance (15 and 30 mil/min), permeability profile showed a higher SC for albumin, cystacin C and IgG for Type B than for Type A (p<0.001). SC for all cytokines was significantly reduced in HD (at both 3L/hr and 5 L/hr) as compared with HF and HDF. It was of interest that in HDF SC of IL-6 and IL-8 at 3 L/h were overlapping those obtained in HF. However, SC of IL-1b, IL1-ra and TNFa in HDF were about half of those obtained with HF. In addition, increasing dialysate flow (from 3 L/h up to 5 L/h) in HD and HDF at a constant blood flow of 150 ml/min led to decrease SC of IL-6, IL-8, TNFa and albumin. Albumin clearance was 1.22±0.43 and 0.57±0.02 ml/min in HDF and HD, respectively.

Our data show that high cut-off polysulfone membrane are associated with high clearances of cytokines independently from blood flow rate and UFR. Tailoring membrane porosity on the basis not only of cytokine clearances but also on ex vivo plasma protein permeability was instructive to formulate their clinical application in mixed convective-diffusive treatments rather than in pure convective or diffusive modes.

Nunomiya S 1 , Momose K 1 , Ohtake K 1 1 Division of Intensive Care, Dept of Anesthesiology and Intensive Care Med, Jichi Medical School, Minamikawachi, Japan INTRODUCTION: Several clinical and experimental studies have reported recently that direct hemoperfusion using a polymyxin B immobilized fiber column (PMX-DPH) is effective for septic ARDS and improves pulmonary oxygenation. Unfortunately, however, little is known about the exact mechanism in such effects. Therefore, we studied the role of circulating leukocytes activities in endotoxemic pigs undergoing PMX-DPH.

Eleven anesthetized pigs were received endotoxin infusion (ETX) to develop ARDS state and submitted to either PMX-DPH group or CTRL group. ARDS state was defined when PaO2/FIO2 ratio decreased to the level less than 70% compared to the point before ETX. Extracorporeal circulations (ECC) were done for 2 hours in both groups. Blood samples were obtained at 4 points; the time before ETX (T-1), ARDS state (T0), 1hour (T1) and 2 hours (T2) after the start of ECC. Leukocyte activities were measured as the abilities of oxygen radical productions from leukocytes using chemiluminescence assay.

One was dead within 1 hour and another was dead within 2 hours after the start of ECC in CTRL group, whereas no animals were dead in PMX-DPH group during the study period. Time courses of PaO2/FIO2 ratio and leukocyte activities in both groups are shown in Tables.

Changes in P/F ratio compared to T-1 

Cotogni P 1 , Muzio G 2 , Trombetta A 2 , Canuto R 2 , Trompeo A 1 , Viale A 1 , Ranieri M 1 1 Anestesia e Rianimazione, 2 Patologia Generale, University of Turin, Turin, Italy

In the early phase of ARDS, intense inflammatory reactions occur in the alveolar space. In this setting, the balance between pro-and anti-inflammatory cytokines may be a critical component for prognosis. Evidence is accumulating that n-3/n-6 polyunsaturated fatty acids (PUFA) ratio may influence inflammation, since the eicosanoids formed from n-3 PUFA and those developed from n-6 PUFA have opposite effects upon inflammatory mediators production. In standard artificial nutrition -both in parenteral and in enteral formulas -n-3/n-6 PUFA ratio is quite low (between 1:5 and 1:7), since most nutrients are richer in n-6 than in n-3 PUFA. Though, the most favourable n-3/n-6 PUFA ratio is not yet defined. Our study tested the hypothesis that n-3/n-6 PUFA ratio may modulate inflammatory cytokines production in a cell culture of human pneumocytes exposed to lipopolysaccharide (LPS).

A549 cells, a human pulmonary cell line with type II pneumocyte properties, were cultured (30000/cm2) in HAM F-12K medium. In all cultures but in controls, LPS was added 24 hours after seeding, to obtain a final concentration of 400 mug/ml. Three hours after LPS, PUFA were added as docosahexaenoic acid (DHA) (n-3) and arachidonic acid (AA) (n-6) in different n-3/n-6 ratios. Four hours later, all culture supernatants were collected to determine the release of TNFalpha, IL-6, IL-8, and IL-10 (ELISA).

Pro-inflammatory cytokines production was significantly reduced by a 1:1 ratio of n-3/n-6 PUFA, but increased by a 1:4 ratio. A higher ratio (3:1) was not associated with further cytokines reduction (Table 1) . *n-3/n-6 CONCLUSION: In a human pulmonary cell culture stimulated with LPS, inflammation can be modulated by PUFA, through appropriate changes of n-3/n-6 ratio. 

High doses of selenium could be a promising way for septic shock treatment. However, selenium (Se) toxicity is supposed to be related to oxidative stress through a reaction with thiols. In the situation of an oxidative stress such as severe sepsis, it is to be feared that selenium toxicity could be increase, despite the fact that preliminary results are in favor of two different pathways for lipopolysaccharide (LPS) and Se toxicity.

After approval by the CRSSA ethical committee, 110 wistar male rats were studied. Rats were quarantined for 8 days. Then, Lipopolysaccharide (LPS) followed one hour later by selenium, as sodium selenite (LPS-Se group) or Se alone, as sodium selenite, (Se group) were administered intraperitoneally. In ten rat LPS-Se groups, LPS were administered at the dose of 26 mg/kg followed by Se at increasing doses from 0.014 to 3 mg/kg. In ten rat Se groups, Se was administered with increasing doses from 0.35 to 4.5 mg/kg. Mortality rate was observed at 48 hours. Surviving animals were sacrificed under anesthesia by halothane. Blood samples were taken on two surviving rats of each group. Plasma selenium concentration was measured using Electrothermal Atomic Absorption.

Mortality related to Se appears for lower doses in LPS-Se groups than in rats receiving Se alone. Mortality rate of rats receiving 26 mg/Kg LPS alone was 65% (13/20). For doses of more than 0.35, septic rats died in respiratory distress in less than one day. For LPS alone or followed by Se at the dose of 0.014 mg/Kg, rats were rapidly sick. They rolled up into a ball. Their fur was dull, and stood on end. They were asthenic and had diarrhea. Se rats developed an encephalopathy the first day and later recovered, except 4 rats with extremely high doses of Se, according to the literature on selenium acute toxicity.

Mortality related to Se (mg/Kg) mg/Kg 0.014 0.35 0.8 1 3 4.5 LPS-Se 7/10 8/10 10/10 10/10 10/10 Not Do Se 0/10 0/10 0/10 0/10 0/10 4/10 CONCLUSION: In a 70% mortality non reanimated LPS rat model, mortality related to selenium administration appears at lower doses those administred in healthy rats. Mortality was related to respiratory distress in LPS followed by Se rats. Doses of 0.014 mg/Kg, presently considered as the maximum selenium level administration, seems not to modify the spontaneous evolution of sepsis in this model.

Zimmermann T 1 1 Department of Visceral-, Thoracic-and Vascular Surgery, Technical University Dresden, Dresden, Germany

Selenium plays a dual role in the regulation of the inflammatory response in mononuclear blood cells. First, selenium enzymes (Gpx 4, TRR) are essentiel for the physiological regulation of the redoxsensitive transcription factor NF-kB (key role in inflammation). Second, selenium is capable to inhibit the activity of NF kB. Another transcription factor (AP-1) is being specific activated via the subunits (c-jun,c-fos) by means of selenium. The authors investigated 28 patients with severe sepsis within the SIC-Study (Selenium in Intensive Care).

Mononuclear blood cells: NF-kB-and AP-1 binding activity, p50/p65 (NF-kB)-protein concentration in the nucleus and cytoplasm. mRNA-expression of IkB, TNF, tissue factor, MIF, Gpx-4 and TRR (selenoenzymes), intracellular synthesis of MIF and IkB. ROS in whole blood. Blood was taken on the 1.,3., 7., 14., 21., 28. day of sepsis.

Septic patients with supplementation of selenium showed a increase of the NF-kBand a strong increase of the AP-1 binding activity during the course of the sepsis. In the same time a rigorous reduction of the mRNA-expression of IkB (inactivator of NF-kB) and MIF could be observed. The mRNA-expression of the tissue factor and TNF was not influenced. Supplementation of selenium lead to a amplified translocation of p50/p65 (NF-kB) within the nucleus, whereas in the placebo group this effect was not shown. In contrary to septic patients, only the NF-kB bindung activity was strongly suppressed in healthy controls.

Selenium seems to possess a regulatory role in der inflammatory response of mononuclear blood cells. The positive effect of selenium in septic patients could be dependent on the time point of the supplementation, within the inflammatory (anti-or hyperinflammatory) response. This could be one explanation of "non-responders" of selenium supplementation. 

High dose of selenium (Se) could be a promising way for septic shock treatment. However, selenium toxicity is supposed to be related to oxidative stress through a reaction with thiols. In the situation of an oxidative stress such as severe sepsis, it is to be feared that selenium toxicity could be increased. Presently human administration of sodium selenite of more than 700 µg per dose must be avoided, outside carefully conducted study. However preliminary results are in favor of two different pathways for lipopolysaccharide (LPS) and Se toxicity, which leads to think that selenium, especially as sodium selenite, could be a new way of treatment.

After approval by the CRSSA ethical committee, 64 Wistar rats were studied. Rats were quarantined for 8 days. Sixty four rats received 26 mg/Kg of LPS intraperitoneally, followed one hour later by 3 milliliters of saline water (placebo) (n =20), or 0.014 mg/Kg selenium as sodium selenite (n=10) corresponding to around 1mg for a 80 Kg man, or 0.25 mg/Kg selenium as sodium selenite (n=34). Mortality rate was observed at 24 hours. Videos were performed during the 24-hour course. Surviving animals were sacrificed under anesthesia by halothane. Blood samples were taken on two surviving rats of each group.

There is a tendency of the mortality decrease in this post-treatment septic rat model. Moreover, rats receiving LPS alone or supplemented by 0.014 mg/Kg sodium selenite were rapidly sick. They rolled up into a ball. Their fur was dull, and stood on end. They were asthenic and had diarrhea. LPS non-surviving rats died in an asthenic syndrome, and surviving LPS alone rats remain very asthenic at 24 hours. On opposite, surviving LPS followed by Se rats were much more dynamic, even quite normal. 17th Annual Congress -Berlin, Germany -10-13 October 2004 S111

Kepa L 1 , Oczko-Grzesik B 1 1 Department of Infectious Diseases, Silesian University Medical School, Bytom, Poland

Cytokines and neutrophiles play an important role in pathogenesis of bacterial sepsis with purulent meningoencephalitis (bs-pme). Experimental studies in animals revealed that pentoxyfiline (PF) exerted inhibitory influence of cytokines on these cells with beneficial outcome of the disease. The aim of the presented study was the estimation of PF influence on clinical course and outcome of bs-pme in adults.

Between 2000-2003 bs-pme was recognized in 18 patients treated in our centre. Neisseria meningitidis and Streptococcus pneumoniae were etiological agents in subsequently 39% and 11% of cases. In the remaining 50% of subjects the etiology was not elucidated. All patients were divided at random way into two groups: I -8 patients (mean age 39yrs.) treated with antibiotics, symptomatic drugs and PF (3mg/kg/day) beginning from the first day of treatment, II -10 patients (mean age 41 yrs.) treated only with antibiotics and symptomatic drugs. Cerebrospinal fluid (CSF) samples were taken on the 1st, 4th and 14th day of therapy with estimation of pleocytosis and protein, glucose, lactic acid, TNF-alpha, IL-1beta and CRP concentrations.

Mean periods of consciousness impairment, fever persisting as well as hospitalization were comparable in both groups of patients. Faster normalization of CSF protein, glucose, lactic acid and CRP concentrations were recorded in patients of group I, who survived, compared to subjects of group II, but the differences were not statistically significant. CSF parameters remained abnormal in fatal cases. Most frequent sequeles of bs-pme were: partial deafness, deafness, paresis and paralysis. Side efects of PF were not observed.

Death Sequels Group I 4 (50%) 2 (25%) 2 (25%) Group II 5 (50%) 3 (30%) 2 (20%)

CONCLUSION: Pentoxyfiline used as adjunctive therapy in adult patients with bacterial sepsis and purulent meningoencephalitis did not reveal evident beneficial influence on clinical course and outcome of the disease.

Zahorec R 1 , Setvak D 1 , Cintula D 1 , Blaskova A 1 , Belovicova C 1 1 Of Anesthesia and ICU, St. Elizabeths Cancer Institute, Bratislava, Slovakia

Sepsis is a common cause of acute renal failure (ARF). ARF in early phase of severe sepsis occurred in 36-60% septic patients and is associated with significant influence on sepsis mortality. The aim of this observational study was to measure the incidence of ARF syndrome and to evaluate the efficacy of noradrenaline and furosemide infusion (Martin et al.1990 (Martin et al. , 2000 for the treatment ARF in early phase of severe sepsis.

An observational study of 17 consecutive critically ill cancer patients with severe sepsis (8) and septic shock(9). Acute renal injury/ARF syndrome was detected according Bellomo et al (2001) criteria. The surrogate markers of renal dysfunction involve serum urea, serum creatinine and urine output (diuresis per hour). We measured creatinine clearance and excretion fraction of sodium from collected urine.The severity of severe sepsis was measured by APACHE II and SOFA score during the first 24-48 hrs of ICU stay. We monitored in all pts invasive sAP,MAP,CVP,temperature, pulse oximetry, urine flow per hour and per day.Blood sampling were done every 12 hrs for WBC counts, platelets count, procalcitonin, CRP, urea, creatinine and lactate.

: 17 severe septic patients with MODS (initial SOFA score were 9,1 and 8,7 p., and APACHE II score 20,4 and 17,2)received full intensive therapy. Severe sepsis was documented by proven infection and high serum levels of procalcitonin (mean 69,8 ng/ml) and CRP(mean 210 mg/L). Acute renal injury (6 pts) and ARF(8 pts) syndrome was detected in 14 patients (82%) out of 17 septic cancer pts. We used the combination of noradrenaline infusion (0,1-0,12 mcg/kg/min) and furosemide infusion(10-30 mg/hr) for hemodynamic and renal support. We induced polyuria and reverse ARI/ARF to nonoliguric ARF in 11 pts (79%) from 14 severe septic pts. We used no renal replacement therapy.We recorded 35 % hospital mortality.

Acute renal injury and acute renal failure syndrome occurred in 82% of severe septic patients. Criteria for ARI/ARF syndrome diagnosis are very simple and useful in early detection of renal dysfunction. Renal rescue protocol (combination of noradrenaline and furosemide infusion) seems to be very effective modality in the treatment for ARI/ARF syndrome in early phase of severe sepsis, when it is instituted very early with low/moderate dosage of noradrenaline and furosemide. 

The purpose of the present study was to evaluate the effects of intravenous lornoxicam on hemodynamic and biochemical parameters, serum cytokine levels, patients' outcome in humans suffering from severe sepsis METHODS: 40 patients were included to the study. After applying, lornoxicam 8 mg was administered intravenously every 12 hrs for six doses vs placebo. Hemodynamic parameters (heart rate,mean arterial pressure), nasopharyngeal body temperature, arterial blood gas changes (pH, PO2, PCO2 ), plasma cytokin levels (interleukin 1-b, interleukin 2-R, interleukin 6, interleukin 8, tumor necrosis factor-a), biochemical parameters (lactat, leucocyt, trombocyt, creatinin, total billirubin, serum glutamat oxalat transaminase), staying time in the intensive care unit, time of mechanical ventilation support, mortality, with the control group were recorded. All measurements were obtained at baseline (before start of the study) and were repeated immediately at 24th , 48th and 72 th h. after lornoxicam.

No differences were found differences in major cytokines, duration of ventilation and ICU stay, and Fi02/ Pa02 intravenous lornoxicam vs placebo (p>0.05).

We found that the effect of intravenously lornoxicam did not effect hemodynamic and biochemical parameters, or cytokine levels or in patients' outcome in severe sepsis in humans. Because of the limited number of patients in our study and the short period of observation, our findings need to be confirmed by larger clinical trials of intravenously lornoxicam in a dose-titrated manner

Bernard GR, Wheeler AP, Russell JA, et al: N Engl J Med 1997, 336: 912-918. 

Bubenek-Turconi S S t 1 , Sefu F 1 , Stelian E 1 , Boros C 1 , Miclea I 1 , Timofiev L 1 , Moldovan H 2 , Iliescu V 2 1 1st Cardiovascular Anaesthesia and Intensive Care Dept., 2 1st Cardiac Surgery Dept., Institute of Cardiovascular Diseases C. C. Iliescu, Bucharest, Romania

Severe systemic inflammation with a vasodilatory syndrome occurs in about one third of all patients after cardiac surgery with cardio-pulmonary bypass (CPB). We studied the effects of early continuous veno-venous hemofiltration (CVVH) on the course and outcome of the patients with severe systemic inflammatory response syndrome (SIRS) after cardiac surgery.

A group of 40 patients with severe SIRS [fulfilling the criteria of ACCP/SCCM Consensus Committee (1)] in early postoperative period after cardiac surgery with CPB was divided in two subgroups: A -21 patients receiving conventional therapy and B -19 patients who received CVVH for a period of 24 h. Criteria for receiving CVVH was a severe cardiovascular dysfunction (catcholamine support required in large amounts, norepinephfrine or epinephrine > 0.1 ?g x kg-1 x min-1, for maintaining a MAP > 70 mm Hg or a SVR > 800 dyne x sec x cm-5).

Of those 19 patients 6 had also a severe respiratory dysfunction with PaO2/FiO2 < 200.

There were no significant diferences regarding demographic data and type of surgery between the two groups. The patients from group B had a dramatic improvment of the cardiovascular function, the catecholamine support being tapered off faster than in group A even the initial dose was very much higher in group B. Also the patients with respiratory dysfunction from group B were extubated earlier than those from group A, with the same amendament regarding the severity of the dysfunction. The result are sumarized in the following 

Ricci Z 1 , Salvatori G 1 , Bordoni V 1 , Bonello M 1 , Ratanarat R 1 , D'Intini V 1 , Ronco C 1 1 Nephrology Dialysis and Transplantation, Ospedale Civile S.Bortolo, Vicenza, Italy

Sepsis and MODS are associated with a disruption of normal homeostasis and alteration of biological systems. The accumulation of pro-apoptotic factors in plasma may contribute to organ dysfunction. Removal of such factors by extracorporeal blood purification techniques may help to re-establish homeostasis and cell function. We investigated the effect of treatment dose comparing standard and high volume hemofiltration.

In a prospective, randomised, cross over study two hemofiltration regimes in two consecutive days were administered to 6 anuric septic patients: we studied 5 hours high volume hemofiltration (HVHF:4L/h) followed by 5 hours standard hemofiltration (CVVH:2L/h) and viceversa. Replacement solution was administered pre filter and performed by 2 m 2 polysulfone membranes. Blood flow rate was 250 ml/min. Routine laboratory and clinical data were collected including illness severity scores. Prefilter plasma and ultrafiltrate were collected at treatment start, at 1 hour, at 5 hour, for each hemofiltration regimen. Plasma samples and ultrafiltrate were frozen at -80°C. Samples were close labelled. Samples from normal human blood were used as control. Samples were studied for apoptosis using a U937 monocyte cell line. A quantitative analysis of the apoptotic U937 cells in culture was carried out by fluorescence microscopy at 96hours. U937 cells were also assayed for caspase 3, 8 activation.

During the sequence HVHF/CVVH cell apoptosis significantly decreased after 1 hour of 4L/h treatment start (p<0,01); after 5 hours of 4L/h treatment apoptosis rate continued to decrease significantly (p<0,05). After passing to 2L/h regime the percentage of apoptosis remained constant. The fold-increase of caspase-3 measured at 96 hr correlated with the above findings (r=0,92). Similarly when the inverse sequence (CVVH/HVHF) was studied cell apoptosis did not show a decrement in the first 5 hours, while after switching hemofiltration dose to 4L/h apoptosis was significantly decreased either at the first and at the fifth hour (p<0,05; correlation between apoptosis and caspase-3 fold increase: r=0,98). The results where independent from the administration sequence.

CONCLUSION: High hemofiltration rates seem to correlate with a decrease in plasma apoptotic pattern during CRRT in anuric septic patients. The clinical relevance of such findings may contribute to explore new therapeutic options in septic patients. Epinephrine (E). We conducted a casenote review of patients with septic shock.

The unit ICNARC/MIDAS database was searched for all patients admitted over a two year period with septic shock. The ICU notes and charts were then retrieved and data found on physiology, choice and dose of catecholamine given.

: 109 patients were identified, of whom 92 were treated with NE and 28 with E. There were notable differences in outcome between the two vasopressors commonly used. Patients receiving either drug were seen to have an increased mortality in association with higher doses used. No patient survived to hospital discharge who was treated with a dose of E above 4.6 micrograms/ kg/ min or NE above 5.8 micrograms / kg/ min. CONCLUSION: There is an increased mortality seen in patients with septic shock receiving E. Whilst they may be older, with worse APACHE scores and calculated risk of death; this doesn't explain the degree of the problem. Some of the answer may lie in their worse glucose metabolism. There was also an increasing mortality seen with increasing dose of vasopressor given. This was independant of APACHE 2 score and as such may repesent a drug effect rather than a marker of illness severity. (VILI) .But positive end expiratory pressure with or without low tidal volume is protective against lung injury. In this study we investigated the effect of different inspiratory times on VILI.

METHODS: 42 Sprague Dawley rats were used. All were started to ventilate on pressure controlled ventilation mode, after anesthetized and tracheostomized, with the parameters of 14 cmH2O peak inspiratory pressure (PIP), 0 cmH2O PEEP, FiO2:1.0, 30 breaths/min and I/E:1/2. After 15 minutes stabilization period baseline blood samples were taken for blood gas and cytokine analysis, then the rats were randomized into 7 groups due to their peak inspiratory pressure, PEEP and inspiratory/expiratory ratios as follows: Other ventilator settings were kept as baseline values. The rats were ventilated with these parameters for two hours. At the end of experiment before sacrification of rats, blood samples were obtained for blood gas and cytokine analysis. Then the lungs were taken out and the left lung was used for measurement of wet weight/dry weight ratio (WW/DW).

There were no differences in baseline pH, PaO2, PaCO2, MAP values among groups. As compared to baseline values PaO2 decreased in LP1/2, HP1/2, 2/1, 3/1 groups and HPP 2/1,3/1 groups but significant differences was found only in HP1/2 group(p=0.001). At the end of experiment MAP decreased in all HP groups and HPP3/1 group. WW/DW ratio was found lower in HPP groups when compared to HP groups (p<0.001). IL-6 level was found higher in HP groups than LP and HPP groups at end of experiment.

High PIP caused lung injury with deterioration of oxygenation and increase in WW/DW ratio. While application of PEEP was protecting lungs from VILI changing inspiration expiration ratio did not.

Dragazis I 1 , Mariatou V 1 , Kopteridis P 1 , Kapetanakis T h 1 , Karidis N 1 , Balanika M 1 , Michalia M 1 , Armaganidis A 1 1 2nd Critical Care Department, Athens University Medical School, Athens, Greece

Our purpose was to investigate whether temperature modulates ventilatorinduced lung injury (VILI).

We perfused (constant flow 300 ml/min) 48 isolated sets of normal rabbit lungs and ventilated them using 3 different perfusate temperatures and two different ventilatory settings (6 groups). After initial stabilization all preparations were ventilated for 20 min using pressure controlled ventilation [PCV] with PEEP 3 cm H2O and PCV 12 cm H2O above PEEP. Following the results of randomisation the necessary adjustments were made during this period to obtain in the perfusate: 1) a pH 7.40 with a partial pressure of CO2 40 mm Hg and 2) a perfusate temperature of 33 oC, 37 oC or 40,5oC. Two groups of preparations were tested at each temperature level: a control or Low Pressure (LP) group ventilated with PEEP 3 cm H2O and PCV 12 cm H2O above PEEP for 60 min and a High Pressure (HP) group, in which a PCV = 22 cm H2O above PEEP (=3 cm H2O) was applied for 60 min. The weight gain (deltaWG in g/min) observed in each group during this period, as well as changes in ultrafiltration coefficient (Kf in gr/min/ cm H2O/100g) were used to assess VILI (indexes of pulmonary edema and of vascular permeability respectively).

Our results are summarized in Table 1 . DeltaWG in hyperthermic isolated, perfused lungs was significantly higher than deltaWG in any other group. Significant Kf changes were observed only in HP groups, with a significantly higher deltaKf in the HP_40.5 group (p=0.035). There were no important differences between normothermic and hypothermic preparations. METHODS: 15 Sprague-Dawley rats were anaesthetized, paralyzed and mechanically ventilated. Rats were ventilated similarly (Vt=10 ml/Kg, RR=20 bpm, FiO2 0.5), but were randomized to PEEP 0, 3 or 8 cmH2O (n=5 per group). The abdomen was then inflated stepwise with helium up to 10 mmHg of abdominal pressure (IAP, intra-peritoneal direct measurement). Airway pressure (Paw), esophageal (Pes) and gastric (Pga) pressure were also measured, together with invasive blood pressure. Data were simultaneously recorded and digitally stored for subsequent analysis. This allowed to consider end-expiratory (Pes exp), mean (Pes m) values of Pes and the difference between Pes at end-inspiration and end-expiration (DPes). Data are presented as mean±SD. 

We conclude that in our study study the closed tracheal suctioning system did not decrease the incidence of ventilator-associated pneumonia, not even the exogenous pneumonias. We believe that the respiratory secretions suction may be done with guarantee with an open tracheal suctioning system if it is performed with suitable asepsis measures. And we think also that it is not necessary the high cost that the routine use of a closed tracheal suctioning system represents. However the closed tracheal suctioning system may be recommended in patients with severe impairment of gaseous exchange. 

In order to avoid decrease in bacterial count due to empiric AB before sampling, we evaluated the feasability of delaying the cultures of Broncho-Alveolar Lavage (BAL) frozen at -20°C et -80°C for 24 hours. The results from these 2 delayed processing were compared with those from immediate ones. A total of 115 BALs were performed on 90 ICU patients suspected of nosocomial or community-acquired bacterial pneumonia. Each sample was divided in three, one for immediate culture (H0), the 2nd and 3rd for a delayed processing after storage at -20°C and -80°C for 24 hours (H24) respectively.

All negative H0 samples (n=38) were also negative at H24 except for one sample that yielded 10 and 40 cfu/ml of Streptococcus sp on -20°and -80°H24 culturing respectively. Seventy seven BALs yielded one or more microorganisms, with a total of 160 microorganisms in one or both samples. H0 and H24 (-20°& -80°) 

Hájek R 1 , Nìmec P 1 , Zezula R 1 , Fluger I 1 , Rù?ièková J 1 1 Cardiac Surgery, University Hospital, Olomouc, Czech Republic INTRODUCTION: Thrombelastography (TEG)is a method frequently used in perioperative assessment of haemostasis in cardiac surgery. This bedside examination can reveal some specific disorders of haemostasis especially hypercoagulation and fibrinolysis.

One hundred fourteen consecutive patients with acquired heart desease were assessed. All the patients were operated electively and the cardiopulmonary bypass was used. Standard laboratory perioperative assessment of coagulation was performed. These results were compared with TEG performed afer indtroduction of anaesthesia, after 30 minutes of CPB and immediately after admission on ICU after operation. Preoperative anticoagulation therapy, blood loss and the necessity of transfusion were evaluated.

Only 29 patients nad no anticoagulation medication preoperatively. In laboratory assessment all the patients had normal results preoperatively, 12 patients had coagulation disorder and 14 patients thrombocytopenia postoperatively.

TEG examination revealed hypercoagulation status in 70 patients and hypocoagulation in 13 patients preoperatively. During operation increased fibrinolysis was found out in 28,9% patients (11,4% during operation, 13,1% after operation and 4,4% both during and after operation), only in 25% of them aprotinin was used because of increased bleeding. Thrombocytopatia was revealed in 16,7% patient and only in 1 of them thrombocyte infusion was required. In 5 patients the residual high level of heparin was confirmed. The average blood loss during operation was 350 ml and during first 24 hours was 693 ml. No patient was reoperated because of bleeding. Correction of hypocoagulation was made with FFP in average dose 2 TU (4 case with normal postoperative TEG tracings versus 9 cases with pathological TEG).

CONCLUSION: TEG revealed hypercoagulation status in many patient preoperatively, which was not confirmed by standard laboratory tests. During operation mainly fibrinolysis and thrombocytes dysfunction was present but any specific therapy was usually not necessary. The use of blood products depends more on clinical status of the patient than on the TEG results. Clinicians are facing the challenge to differentiate between postoperative inflammation a condition considered to be benign and early signs of infection. The aim of our study was to define the timecourse of SIRS and severe SIRS after cardiac and thoracic surgery.

We utilised a structured data mining process to the prospectively collected data within the Patient-Data-Management-System (PICIS Caresuite V. 6.3) from the Cardiothoracic ICU of a University Hospital between January 1999 and May 2003. Data from all monitoring device are collected in intervals of 10 minutes, laboratory data and blood gas analysis was done according to institutional standards. In this data mining process we determined in a first step the fulfillment of each individual item of the SIRS criteria (ACCP/SCCM Consensus Conference) during a minimum of one hour. In the second step we identified the first occurrence of simultaneous fullfillmment of at least 2 criteria as the starting point for SIRS. Severe SIRS was defined as SIRS with at least two criteria for organ dysfunction as defined in the SOFA score. We used three categories SIRS, SIRS with low blood pressure (SIRS low BP) and severe SIRS with additinal organ dysfunction (SIRS severe).

A total of 1629 patients were admitted during the observation period. SIRS was present in 1001 (61.4%), SIRS with hypotension in 877 (53.8%) and SIRS with additional signs of organ dysfunction in 461(28%). The timepoints of first fullfillment are given in the table. The timeprofile with very early fullfillemtn was not changed by censoring the first hours after admission since the identified state persistent for a prolonged period. Timepoint of first SIRS fullfillment: 

In this large cohort of patients after cardaic and thoracic surgery we found dystinct profiles for SIRS with additional signs of organ dysfunction. In the majority of the patients the three different SIRS categories occurred within the first 6 hours. Further research is necessary to determine whether any of these categories are indicative of a changed outcome depending on the starting point.

17th Annual Congress -Berlin, Germany -10-13 October 2004 S121

Samalavicius R 1 , Misiurine I 1 , Norkiene I 1 , Juozaitis M 1 , Urbonas K 1 , Bubulis R 1 , Baublys A 1 1 Anaesthesiology and intensive care, Vilnius University Hospital Santariskiu Clinics, Vilnius, Lithuania INTRODUCTION: Preoperative risk stratification for predicting mortality and morbidity is widely used in cardiac surgery. The goal of this study was to assess the value of serum lactate level in predicting mortality and morbidity following coronary artery bypass grafting (CABG) procedures.

METHODS: 600 consecutive CABG patients, operated on from 2003.01.01 to 2003.12.30, were included in this prospective observational study. All patients were operated using cardiopulmonary bypass. Serume lactate levels were measured before cardiopulmonary bypass, before declamping of the aorta, after heparin neutralisation and at the ICU admission.

: Lactate level greater than 5 mmol/l was found in 2.5% of patients during cardiopulmonary bypass, in 6.7% of patients shortly after weaning from CPB and in 10.8% of patients at ICU admission. Mortality rate of patients, with hyperlactemia at ICU admission was 19.3% and morbidity 59.6%. Mortality rate of patients without hyperlactemia was 3.2% and mobidity -10.3 %. Lactate levels on ICU admission were raised in non survivors (median 4.9 +/-3.45, range 0.9-12.0 mmol/l)compared with survivors (median 2.5 +/-1.76, range 0.3-13.5).

CONCLUSION: Increased serum lactate levels following coronary artery bypass grafting allows to identify a group of patients with increased risk of postoperative mortality and morbidity. 

Since the year 2000, we have been studying prognosis in cardiac surgery (CS) and noticed the lack of models with similar populations in the literature. The objective this study is create a predictive score (Rio Score-Pre) of in-hospital mortality in patients (pts) undergoing CS based on preoperative variables.

Classical cohort with data of 1458 pts, 437 of whom undergoing valvular surgery (VS) and admitted to 2 intensive care units (ICU), public and private, consecutively selected between June 2000 and February 2003. All 19 variables were previously defined. The data underwent univariate analysis with the chi-square, Student t, Mann-Whitney, and Pearson tests, followed by logistic regression, and stepwise (likelihood ratio), with the chi-square linear tendency test and a classification table.

The score created (appendix) allows the following prediction: from 0 to 4 -low risk; from 5 to 9 -medium risk; and from 10 to 15 -high risk. 

Pérez-Vela J 1 , Renes E 1 , Escribá A 1 , Alonso M 1 , Corres M 1 , García A 1 , Perales N 1 1 Intensive Care Unit, Hospital 12 de Octubre, Spain, Spain

Vital parameters monitorization is an usual practice in the management of critically ill patients. Cardiac index (CI) is one of more important perfusion parameters used. PiCCO system is a device that offer the quantification of intermittent CI by transpulmonary thermodilution (CITP) and in a continuous manner by arterial pulse contour analysis. Objective: To compare the agreement between the standard thermodilution monitorization system (CITD) with the transpulmonary system. Also, we analysed the complications secondary to the PiCCO system.

Prospective study, in patients in the immediate postoperative period after cardiac surgery with cardiopulmonary bypass. CI by standard themodilution was measured with pulmonary artery catheter Abott OptiQ SvO2/CCOâ. We made a transpulmonary thermodilution with 15 ml physiologic fluid injection with a temperature less than 15 celsius degrees, through a central venous line, and we analysed the thermodilution in the femoral artery catheter thermistor (a 4-Fr gauge, 20 cm long arterial with a thermistor embedded in its wall: pulsiocath PV2014L) using the PiCCO system from Pulsion Medical System (Munich; Germany). We calculated CI (both methods) after inserting the PiCCO system, one hour later and then, every two hours. Also we measured parameters when staff considered appropriate to value the results of a therapeutic attitude. Results between techniques were compared by lineal regression analysis and the Bland-Altman method.

We analysed a total of 126 pair of data obtained in 18 patients, 7 male and 11 female, in the immediate postoperative period of 9 valvular replacements (5 mitral, 3 aortic and one both), 2 aortic grafts, 5 myocardial revascularizations, 1 mixoma and 1 pericardiectomy. Mean age: 67±8.4 years, mean CITD 2.75±0.67 and CITP 2.75±0.69 l/m. The range of measured CI: 1.3 a 4.2 l/m/m2. In comparison we obtained a r= 0.88 and a bias of -0.027±0.662. In tables we have the realised statistic analysis. We did not have complications attributed to the system.

Both CI measurement methods are comparable, showing a good agreement between systems, indicating that CITP is as reliable and precise as standard thermodilution. This suggests that PiCCO is a monitorization system applicable to clinical routine in critically ill patients. We did not observe complications attributed to the system.

Gomes R V 1 , Rouge A 1 , Nogueira P M M 1 , Fernandes M A O 1 , Olival S A 2 , Campos L A A 1 , Dohmann H F R 1 , Santos M 1 1 Surgical Intensive Care Unit, Hospital Pró-Cardíaco -PROCEP, 2 Surgical Intensive Care Unit, Instituto Nacional Cardiologia Laranjeiras, Rio de Janeiro, Brazil

The Left Ventricular Ejection Fraction (LVEF) has been extensively studied as a prognostic marker in cardiac surgery (CS); our group, however, has found a correlation between Left Atrial Diameter (LAD) and several outcomes in CS. The objectives this study is show the importance of LAD as a prognostic marker by assessing the following outcomes: In-Hospital Mortality (HM), Surgical Intensive Care Unit Length of Stay (SICULOS), Pneumonia (PN), and Need for Hemodialysis (NHD).

Compilation of data collected in the databank of several cohorts with 2211 patients (pts) of 2 SICU from June/00 to February/04. The 46 variables studied underwent uni-and multivariate statistical analysis.

The Few reports exist about LAD on echocardiography as a risk marker for CS; in our studies, however, LAD has reached greater significance than the subjective analysis of LV function. The great prevalence of valvular surgery (VS -30%) might be another possibility. These findings should be validated in a cohort with other centers. CONCLUSION: Amylase level > 500un/ml and bilirubin concentration of 5 mg/dL in duodenal aspirated fluid has a high positive predictive value. Position of the feeding tube within the gastrointestinal tract can be determined objectively by using simple pH and bilirubin reagent strips.

Webb I 1 , Gibbs T 1 , Beale R 1 , Jones A 1 . 1 Department of Intensive Care Medicine, Guy's and St Thomas' Hospital, London, United Kingdom

Septic shock may be accompanied by dysfunction of the hypothalamic-pituitaryadrenal axis (HPA). In a recent multi-center randomized controlled clinical trial, treatment with hydrocortisone and fludrocortisone significantly reduced the risk of death in patients with septic shock and relative adrenal insufficiency (as determined by ACTH stimulation) but not in patients with an adequate adrenal response (1). Subsequently it has become accepted practice to start corticosteroid replacement in patients with septic shock following an ACTH stimulation test to determine the presence (non-responder-NR) or absence (responder-R) of "adrenal insufficiency". In those patients with "adequate" adrenal function, corticosteroids are withheld or withdrawn. It is known that in patients who recover, this "adrenal insufficiency" is temporary. However, less is known about the temporal changes in HPA function within the period of critical illness.

METHODS: Our clinical information system (CareVue, Philips Medical Systems, UK) was interrogated to find all patients with septic shock who underwent repeated ACTH stimulation (1\mug) testing in an 12 month period. Baseline cortisol, NR/R status (\deltacortisol <9 \mug/dl), vasopressor requirements and use of hydrocortisone were identified.

We identified 8 patients who underwent repeated ACTH stimulation testing within a single episode of septic shock, who received no or limited steroid replacement therapy. In 3 subjects (6,7,8 

The recognition that HPA abnormalities exist in sepsis and that exogenous steroids are beneficial in some individuals has changed practice over recent years. However, the best indicator of which patients would benefit from corticosteroid replacement remains unclear. In addition, this preliminary data suggests that an individual patient's response to ACTH stimulation may change during an episode of septic shock. Of particular concern are patients who are initially "responders" who would not recieve beneficial therapy if only single estimates of adrenal dysfunction are used. 

Twenty four neonates with TOF were divided into two groups after getting parents consent and local ethical commitee approval into two groups ( 12 each); Group I : after general anaesthesia they had continous thoracic paravertebral block with a catheter placed at right fifth space with continous infusion of 0.75 ml/kg of 0.375% ropivacaine every 90 minutes to be maintained post operatively in NICU this. Group II :had balanced general anesthesia. Measurements: 1-Number of neonates required post operative ventillatory support in both groups. 2-Mean total dose of opioids required for analgesia in both groups. 3-Days of stay in NICU in both groups. 4-Mortality in both groups RESULTS: 1-There was statistically significant less need for ventilatory support in group I (16%) in comparison to group II (41 %). 2-Mean total dose of opioid analgesia was higher in group II 3-More days of stay in group II. 4-Three cases of mortality in group II ,while one case of mortality in group I. CONCLUSION: Picu mortality was relative high, partly due to high PRISM scores and the high proportion of MV pts. Mortality continues to increase up to 1y and stayed the same thereafter. The majority of our pts reached their preadmission cognitive status (PCPC) at two y. On the other hand they didn't reach their overall functional status (POPC) even after two y mainly due to the high proportion of pts with mild disability, POPC 2, which is however compatible with near normal and independent life.

S75-S82. 3.Jalan R, Williams R. Blood Pruif

Predictors of transfusion requirements for cardiac surgical procedures at a blood conservation center

REFERENCES: 1 Maillet JM

Grant acknowledgement: Chris Stoutenbeek Foundation REFERENCES: 1.Debra Henry Fiser

 22 (5,4) 15 (11,7) 5 (5,9) 2 (1) E coli 22 (5,4) 7 (5,4) 6 (7,1) 9 (4,6) S pneum+ 15 (3,7) 10 (7,8) 4 (4,7) 1 (0.5) Others**+ 137 (33,9) 35 (30,9) 28 (33,3) 24 (15,2) *p=0.02 G1/ G2; **p<0.05 G2/G3,+All groups CONCLUSION: In our country, Late-onset VAP showed important variations in aetiology considering 5th and 7th days and that should influence antimicrobial prescribing practices. 

Nosocomial pneumonia represents a significant cause of morbidity and mortality in intensive care units (ICU). The high incidence of nosocomial pneumonia among ICU patients can be attributable to dysregulation of lung immune responses elicited by systemic inflammation. In a previous study with endotoxin-challenged mice, defects of lung adaptive immunity were heralded by reduced numbers of lung-resident CD4+ T-lymphocytes. The depletion of CD4+ T-lymphocytes was associated with a higher susceptibility to lung infection caused by Staphylococcus aureus in some clinical and experimental studies. The aim of our study was to evaluate the mechanism by which endotoxemia reduces the number of lung-resident CD4+ T-lymphocytes and increases susceptibility to S. aureus in the lungs.

Experimental systemic inflammation was initiated in BALB/c mice (n=6) with 10 \mug of endotoxin (LPS) given intraperitoneally; 24 hrs after this challenge, the animals were anesthetized and 5x10 6 CFU of S. aureus (S.a.) were administered into the trachea. For the control group, mice were challenged only with S. aureus (n=6). Mice were sacrificed 72 hrs after the challenge with S. aureus. Lung-resident lymphocyte subsets were obtained by enzymatic digestion of lung tissue. Lung-derived and circulating total T-(CD3+) and B-(CD19+) lymphocytes, CD4+ and CD8+ T-lymphocytes as well as NK cells were enumerated with monoclonal antibodies, single platform method and cytometric analysis. Colony forming units (CFU) of S. aureus were obtained from lung tissue homogenates using a plate dilution method. The differences between groups of animals were evaluated by one-way ANOVA with a level of significance P<0.05. Data are presented as mean standard ± error.

Results are shown in the table 1 (number of cells is expressed as cellsx10 3 /ml for the blood and cellsx10 3 lobe for the lung). CONCLUSION: Our results demonstrate that mice challenged with endotoxin and S. aureus have reduced recruitment of CD4+ T-lymphocytes to the lungs when compared to animals infected only with S. aureus. Despite this finding, the susceptibility to secondary lung infection due to S. aureus was significantly decreased after endotoxin challenge indicating its protective effect against staphylococcal infection.

Aranha F G 1 , Rouge A 1 , Gomes R V 1 , Dessen M 1 , Nogueira P M M 1 , Fernandes M A O 1 , Campos L A A 1 , Dohmann H F R 1 1 Surgical Intensive Care Unit, Hospital Pró-Cardíaco -PROCEP, Rio de Janeiro, Brazil

The need for dialytic support (HD) in the PO period of CS relates to a significant increase in costs and length of hospitalization, in addition to high rates of morbidity and mortality.The objective this study is assess the incidence of HD and its association with mortality in adult patients (pts) undergoing CS.

Historical cohort with data of 770 pts undergoing CS collected from June/00 to January/04. The pts were divided into 2 groups as follows: 1) group I, 716 pts who did not undergo HD; and group II, 54 pts who required HD, accounting for 7.01% of the sample. Their mean age was 65.27 years, 68.3% were males, 24.6% were diabetic, and 55% of the CS were elective. The mean AHA mortality score was 4.8±3.5, and the mean Euroscore was 5.27±3.82. Analysis of frequency and the chi-square test were used for comparing mortality.

Fifty-six (7.27%) pts died in the hospital, 29 (4.05%) in group I, and 27 (50%) in group II. Thirty-nine (5.06%) pts died within 30 days, of whom 16 (29.6%) were in group II. An important statistical significance (P<0.00001) was observed between both groups. The PO intensive care unit length of stay was significantly longer in group II, in which 75% of the pts remained hospitalized for more than 7 days (P<0.00001).

The PO intensive care unit length of stay and mortality were significantly greater in the group of pts undergoing HD in the PO period of CS. A high percentage of patients underwent HD (7.01% of the sample), which may be explained by the profile of the population studied. In group II, in-hospital mortality was 50%, and mortality in 30 days was 29.6%. In the entire sample, these indices were 7.27% and 5.05%, respectively, and the Euroscore predicted a mortality rate greater than 11.5%. 

The conventional choice for type-B aortic dissection has been medical treatment. Surgical repair has been kept for cases presenting complications. Both treatments are associated with high mortality rates. Endovascular stent-graft placement opens up new perspectives in the controversial treatment of thoracic aorta dissections. The objective of this paper is to describe our experience in the post-operative handling of type B aortic dissections treated with endovascular stent grafting.

Twelve patients admitted to an all-purpose ICU from January 2001 to March 2004 treated with endovascular stent grafting. Ten patients with acute type B dissection and one patient with traumatic rupture of thoracic aorta. The pre-operative study included transesophagic ecography and CT to evaluate the extent of the dissection, the relation with the left subclavian exit, true and false lumen size, and vascular complications.

Placement of the endovascular stent-graft (TALENT type) was successful in all cases. Three patients died within the first 30 days, two of them in the ICU, with a mortality rate of 25%. Complications: one patient had retroperitoneal hematoma, and another presented perioperative AMI. Two cases were observed of paraplegia, and one case of perioperative acute cerebellar ischemia in relation with type A retrograde dissection of the thoracic aorta. The mean stay in the ICU was 8.5 days. Mean mechanical ventilation time was 2.95 days. Five patients (45%) presented nosocomial infection: four infections by catheter (66%), one episode of urine infection (16%) and one episode of pneumonia associated with mechanical ventilation (16%). Four patients presented acute kidney collapse (36%), without the need for hemodialisis in any case.CONCLUSION: Endovascular stent-graft placement can be an alternative to open surgery in the treatment of type B aortic dissection. Preliminary results on post-operative morbimortality are promising. Randomized and controlled studies are needed to assess the therapeutic potential. 

Durand M 1 , Gardelin M 1 , Bertet M 1 , Tessier Gonthier-Maurin Y 1 , Bouzat P 1 , Girardet P 1 1 Anaesthesia, CHU de Grenoble, grenoble, France

Global tissue hypoxia is associated with a poor outcome after cardiac surgery [1] . The best predictor of anaerobic metabolism in septic patient seemed to be the ratio of venoarterial CO2 difference (dPCO2)/arteriovenous O2 (Ca-vO2) content [2] . The aim of the present study was to verify if this ratio had the same predictive value after cardiac surgery.

We performed a retrospective analysis of 95 patients with simultaneous 271 measurements of arterial and venous blood gases and arterial lactate levels during the first 24 hours after surgery. We tested the predictive value of heart rate (HR), cardiac index (IC), mixed venous oxygen saturation (SvO2), dPCO2, Ca-vO2, dPCO2/Ca-vO2 and oxygen consumption (VO2) to predict anaerobic metabolism (lactate > 2 mmol/L). The area under ROC curves was calculated for the main parameters. Results are expressed as mean +/-SD.

: 172 results of lactate were below 2 mmo/L (Gr 1), 99 were above (Gr 2). dPCO2 (kPa) was significantly higher in Gr2 than in Gr 1 ( 

Numerous prospective, randomized studies in critically ill patients indicated that enteral feeding is superior to parenteral feeding and that early enteral feeding, compared with delayed enteral feeding, improves patient outcome as measured by length of stay or complication rates. Ideally, tube insertion would be inexpensive and would require minimal time and technical expertise. We inspected a simple bedside technique for positioning the feeding tube.

All 40 included patients received a polyurethane feeding tube with a flexible wire stylet (20-9431 silk enteral feeding tube, Corpack, Wheeling, IL).One size 103 cm -10Fr of feeding tubes was used in this study. Feeding tube position was confirmed by an abdominal radiograph. Each radiograph was reviewed by a radiologist. Equipment to measure Ph and Bilirubin consisted of color -coded paper (Multistic 10sc Bauer Corp.USA), Amylase and Bilirubin (second test) were measured in the central clinical laboratory.

Successful aspiration of duodenal fluid was performed in 36(91%) patients. Median time for perform bilirubin and Ph by color -coded paper (Multistic 10sc Bauer Corp.USA) -up to 20 seconds. Median time for perform bilirubin and amylase analysis in duodenal fluid in central laboratory was 3.2 ( + 1.5) hours.

Koulenti D 1 , Mis M 1 , Myrianthefs P 1 , Tsigou E 1 , Ioannidis C 1 , Gavala A 1 , Grigoriou P 1 , Baltopoulos G 1 1 ICU, KAT Hospital, Athens, Greece INTRODUCTION: Liver dysfunction is very common in critically ill patients due to a variety of reasons including trauma, sepsis, congestive heart failure, gall stones, hemorrhagic shock, transfusions, and drug hepatotoxicity. The purpose of the study was to investigate the characteristics of liver dysfunction in ICU patients.

We prospectively collected data concerning demographic characteristics and liver biochemistry in critically ill patients for a total period of 6 months. Liver dysfunction was defined as an increase in liver enzymes by twofold times including SGOT, SGPT, Alkaline phosphatase, gamma-GT, and bilirubin.

During the study period, 88 patients were admitted in our ICU. Mean age was 54.9±2.6, SAPS II was 45.3±2.1, APACHE II was 17.8±0.9 and MODS was 5.5±0.4. ICU LOS was 13.6±2.3. Forty-five patients (51.1%) developed liver dysfunction. Five of them (11.1%) had more than one episodes of liver dysfunction. Mean peak values of liver enzymes in patients developing liver dysfunction were SGOT 480.8±205.0, SGPT 305.6±83.7, ALP 133.4±15.3, gamma-GT 283.6±38.6, total bilirubin 2.5±0.6, and direct bilirubin 1.2±0.3. Mean duration of liver dysfunction was 6.1±0.7 days. Mean day of liver dysfunction developed was on 5.5±0.8 day. Confirmed aetiology of liver dysfunction included sepsis (10 pts), trauma-rhabdomyolysis (10 pts), cholestasis (3 pts) and drugs (3 pts). We found statistically significant differences (p<0.005) between the patients developing liver dysfunction and those who did not concerning LOS (17.9±3.5 vs.9.0±2.8 days), SAPS II score (51.4±2.4 vs.39.1±2.9), APACHE II 21.16±1.1 vs. 14.36±1.3 and MODS score 6.8±0.5 vs.4.1±0.5. Mortality was also significantly higher in patients developing liver dysfunction (35.6 vs. 16.3%).

Half of the critically ill patients may develop liver dysfunction during ICU hospitalization due to a variety of reasons which may be related to increased LOS, increased illness severity and other organs dysfunction and worst outcomes.

Hoeksema M 1 , Wester JP 1 , Bosman RJ 1 , Oudemans-van Straten HM 1 , Van der Spoel JI 1 , Haak EAF 2 , Leyte A 3 , Zandstra DF 1 1 Intensive Care Unit, 2 Clinical Pharmacy, 3 Clinical Chemistry, OLVG, Amsterdam, Netherlands

In critically ill patients with multiple organ dysfunction (MODS), thrombocytopenia is frequently observed. Heparin-induced thrombocytopenia (HIT) accounts for 6-11% of all causes of thrombocytopenia. As HIT may be complicated by arterial and venous thrombosis (HITT), alternative anticoagulation is indicated. Fondaparinux sodium (Arixtra®) is a newly developed synthetic pentasaccharide and acts by selective antithrombin-mediated indirect factor Xa inhibition resulting in subsequent thrombin inhibition. Fondaparinux sodium has no cross-reactivity to heparin and has not induced an immune-mediated thrombocytopenia in non-ICU patients. The elimination is almost exclusively renal. Its major drawback is the increased risk of bleeding, to which patients with MODS are prone. Data on treatment schedules in critically ill patients are non-existent. We describe our experience with fondaparinux anticoagulation in the treatment of HIT.

We have treated 3 patients with MODS and laboratory-proven HIT with fondaparinux sodium between December 2003 and February 2004. Treatment with unfractionated heparin or nadroparin calcium was stopped and laboratory tests for HIT were performed with the HIT-antibody ELISA test. Awaiting the test results, fondaparinux sodium (Arixtra®, Sanofi-Synthelabo, The Netherlands) was administered as a once daily subcutaneously injection or a continuous infusion of 1.25-2.5 mg/day without loading dose. Study endpoints were increase of platelet counts, thrombo-embolic and bleeding complications, and need of transfusion.

One female and two male patients, aged between 69 and 82 years, with APACHE II scores between 26 and 33, were diagnosed of HIT due to concomitant nadroparin calcium anticoagulation. Minimum platelet counts varied from 45 to 92 G/l. HIT-antibodies were present in all patients. All patients suffered acute renal failure and were treated with continuous venovenous hemofiltration. Treatment with fondaparinux sodium varied from 5 to 28 days. Platelet counts improved during fondaparinux sodium. One patient died and autopsy revealed a new myocardial infarction. In another patient recurrent major bleeding resulting in acute tamponade and hematothorax occurred under treatment of both unfractionated heparin and nadroparin as well as under fondaparinux. The third patient suffered a minor bleeding complication. Totally, 16 units of erythrocyte concentrates, 3 units of plasma, and 3 units of platelet concentrates were transfused during 45 treatment days.CONCLUSION: Treatment with low-dose fondaparinux sodium in patients with MODS and HIT may be an alternative to treatment with direct thrombin inhibitors. The efficacy and safety need to be determined.

Caballero Zirena A 1 , Cortés Díaz S 2 , Álvarez Terrero A 2 1 Intensive Care Unit, Hospital Virgen de la Concha, 2 Intensive Care Unit, "Virgen de la Concha" Hospital., Zamora, Spain INTRODUCTION: Acute pancreatitis is an "acute inflamatory process of the pancreas with variable involvement of other regional tissues or remote organ systems". The definitions of severe pancreatitis accepted generally are: Acute Physiology and Chronic Health Evaluation (APACHE II) score greater than 8, three or more Ranson´s criteria and CT Grading System of Balthazar. Predicting severity of pancreatitis early in the course of disease is very important to prevent and minimize organ dysfunction and complications.

From 1999 to 2003 a total of 521 patients were hospitalized with the diagnosis of acute pancreatitis. Of these, 27 patients (5%) were admitted to the Intensive Care Unit. The aim of this study was to compare APACHE II score, Ranson´s criteria and CT Grading system of Balthazar for predicting severity and fatal outcome in severe pancreatitis.

: 27 patients were identified. There were 17 men and 10 women. The mean age was 63 years (range 32-83). The most common cause of severe acute pancreatitis were gallstones (66%) and alcoholism (18%). The mean of APACHE II score at the admission was 18,8 (range 9-33). Most of the patients had higher CT score. All of them had more than three Ranson´s criteria. The overall mortality was 48% (13 patients). The intensive care unit length of stay ranged from 1 to 57 days ( mean 7 days).

High APACHE II or Ranson´score at admission significantly determined survival. Ranson criterium has the disadvantage of delay. APACHE II score is useful in organ failure prediction. Balthazar score is superior in predicting pancreatitic necrosis. None of the parameters tested achieved sufficient predictability when used alone. 

Claessens Y E 1 , Marque S 2 , Chiche J D 2 , Mira J P 2 , Dhainaut J F 3 , Cariou A 2 1 Emergency Medicine, 2 ICU, 3 ICU and Emergency Medicine, Cochin Hospital, Paris, France INTRODUCTION: Saving red pack cell (RPC) transfusion is an important goal in critical care management. The need for RPC transfusion after ICU discharge has never been evaluated.

Prospective monocentric study in critically ill patients admitted between July and December 2003 in the medical ICU of a teaching hospital. Data collected: demographics; SAPS2 and LOD (d0 & discharge) ; comorbidity; diagnosis, treatments, ICU and hospital lenght of stay ; Hb level at ICU admission and discharge ; RPC transfusion in ICU and during the 8 days following ICU discharge with Hb threshold and active haemorrhage. Information letter was given to patients and families RESULTS: Population: 550 consecutive pts (52 (19) yrs, SAPS2 40(21), (med(SD)). ICU mortality 22%. Hb at admission 11.4(2.5) g/dl. 17.6% needed RPC transfusion, threshold 7.13 (1.15) g/dl) ; 42% mortality among RPC transfused pts.

, however, with unknown impact on the pituitary-glucocorticoid axis (key mediators: ACTH / cortisol, affected key metabolite: serum glucose). Both, the ACTH / cortisol system [2] and blood glucose levels [3] are increasingly regarded important in intensive care medicine. Since the effects of LEVO on this system may depend on the state of consciousness, we studied respective endocrine effect of LEVO both in the awake and anesthetized state. We compared respective effects of LEVO with those of established inotropes, MILrinone and DOButamine.

Awake and anesthetized (1.5 MAC sevoflurane, ventilated) dogs (total: 36 experiments) randomly received LEVO (10µg/kg plus 3 steps:0.125-0.5µg/kg/min), MIL (5.0µg/kg plus 1.25-5µg/kg/min) or DOB (2.5-10µg/kg/min). Under steady state conditions (each dose: 30 min) we measured arterial ACTH-, cortisol-and glucose-levels. Statistics: Data presented as mean±sem, Wilcoxon test, p<0.05, alpha-adjusted for multiple testing.

: LEVO preserved the levels of ACTH both in the awake state (2.9±0.7 vs. 2.9±0.6pg/ml, baseline and highest drug dose) and during anesthesia (4.3±0.6 to 6.5±2.2pg/ml). LEVO dose-dependently -but insignificantly-increased cortisol under both conditions (awake state: 6±3, 7±4, 15±5 and 19±6ng/ml; anesthesia: 14±5, 22±8, 20±9 and 31±11ng/ml). LEVO preserved (as did MIL and DOB) arterial glucose at ~100-120mg/dl under all conditions. MIL maintained ACTH in the awake state (2.2±0.4 to 2.1±0.5pg/ml) and during anesthesia (3.2±0.4 to 2.5±1.0pg/ml), also cortisol (5±4 to 6±5; 22±7 to 29±6ng/ml). DOB maintained ACTH in the awake (2.5±1 to 2.3±1pg/ml) and anesthetized state (2.9±0.7 to 4.3±1.1pg/ml), and caused insignificant increases in cortisol (10±8 to 19±16; 18±5 to 36±11ng/ml). HFOV is an ideal method of ventilation to minimize VILI. However, there is limited data regarding outcome in children treated with HFOV. We therefore report our experience with HFOV at our PICU.

We retrospectively analysed the chartrecords of all children treated with HFOV after failure on CMV between 2001-2004. The following were recorded: demografic variables, admission diagnosis, PIM II scores, and OI and AaDO2 at several timepoints before and after transition to HFOV. End points included survival at 28 days post-admission to PICU and total number of ventilation days (CMV and HFOV).

Twenty-four children aged 1 day to 6.5 years were treated with HFO. Seven died and seventeen children survived. Non-survivors had a significant higher PIM score (25.8 vs 6.6), shorter duration of pre-CMV (22 vs 108 h). The OI and AaDO2 between non-survivors and survivors were 15.5 vs 21.1 and 355 vs 480, respectively. Both OI and AaDO2 did not decrease over time in the non-survivors. Total ventilation days were lower in the non-survivors (89 vs 156 h).CONCLUSION: HFOV was associated with a high survival percentage (71%)in a selected group of children were CMV failed.

Olsen P 1 , Rasmussen M 1 , Tønnesen E 1 , Zhu W 2 , Stefano G 2 1 Deptartment of Anaesthesia, Århus University Hospital, Århus, Denmark, 2 Neuroscience Research Institute, State University of New York, New York, United States

Exogenously administered morphine has immune modulating effects. The discovery of endogenously synthesised morphine and increased synthesis in response to surgical stress (1,2) and endotoxin infusion imparts a role to endogenous morphine in the immune response. Morphine may also affect cancer progression. However, in vitro and xenograft experimental studies illuminating morphine's role in carcinogenesis show conflicting results. The aim of the present study was to analyse human gliomas for the content of endogenous morphine.

The study was approved by the Regional Ethical Committee on Human Research.Twelve gliomas were extracted during craniotomy and frozen instantaneously in liquid hydrogen. Patients did not receive morphine intra-or postoperatively. Pathological analyses confirmed the diagnoses glioma. Upon preparation samples were analysed for morphine content with radioimmunoassay (RIA) and specificity was confirmed with mass spectrometry.

All tumours contained endogenous morphine with concentrations ranging from 2,01 ng/g -169,92 ng/g. The identity of morphine was subsequently confirmed by mass spectrometry.

The demonstration of endogenous morphine in gliomas suggests its potentially role in carcinogenesis either as an inherent protective measure or as a result of neoplastic transformation. However, it remains to be clarified where the endogenous morphine production takes place. It is also unknown whether the presence of morphine is a pan-cerebral phenomenon or specific to cancerous tissue. The present study revealed a high content of endogenous morphine in human gliomas, providing further support to the idea of potential influence of endogenous morphine in cancer growth. . The majority of infections were exogenous, i.e., the bacterium was introduced into a normally sterile organ, directly from the PICU environment. One quarter of the infections were primary endogenous, i.e., the child developed an infection due to a micro-organism present in the admission flora. The death of one child was unrelated to infection.

This study shows an infection and mortality rate of 27% and 4%, respectively. Low level pathogens caused practically all infections which were mainly exogenous following breaches of hygiene. SDD was effective as endogenous infection due to AGNB was controlled. scale (GCS),hypotension (systolic blood pressure < 90 mmHg) and hypoxia (cyanosis or pulse oximetry <90%)on admission, other traumatisms, head computed tomography (CT) based on Marshall's classification (TCDB), intracranial pressure (ICP) monitoring,jugular bulb oxygen saturation (SjO2), transcranial doppler (TCD), intracranial hypertension (HTIC) defined as ICP>20 mmHg.Brain edema treatment,length of stay in critical care unit (ICU) and hospital,GCS at ICU and hospital discharge, and mortality. We also studied hemodinamic and respiratory (PaO2/FiO2<200) complications, fever (axilar temperature >38,5ºC) and electrolytic disorders (sodium <130 mmol/L or >150 mmol/L). 

Lewejohann J C 1 , Hansen M 1 , Zimmermann C 1 , Muhl E 1 , Bruch H P 1 1 Surgery-ICU, Universitätsklinikum Schleswig-Holstein-Campus Lübeck, Lübeck, Germany

Propofol infusion syndrome (PRIS) is a very rare and often fatal syndrome in critically ill patients undergoing long-term propofol infusion at high doses. Until today 14 cases of PRIS in adults have been described in the literature and 12 of them died. The aim of our representation is to demonstrate the clinical course of a patient with severe rhabdomyolysis subsequent to a multiple trauma and sedation with propofol and to make obvious the importance of this life-threatening syndrome.

A 27 year old multiple trauma patient of about 90kg bodyweight was admitted to our surgical-ICU at a university hospital. He had a severe head trauma, a fracture of the cervical vertebra, an ARDS, multiple rip fractures, severe lower leg fractures with severe vascular damage and the nead of amputation 24h after admission, fractures of the femora, pericardial effusion and hematoma of the spleen. He received from the beginning on high doses of catecholamnies (norepinephrine, epinephrine), hemofiltration because of renal failure. After resection of his right lower leg one day after admission he received propofol 2% in a dose range between 10 to 25 ml/h over a time period of 7 days. An initial Myoglobin level of 6937µg/l as a result of the multiple trauma on admission decreased to 3865µg/l when the propofol infusion was started with 14ml/h at first. Myoglobin level decreased to 398µg/l after 73h. Propofol infusion then was increased to 20ml/h and after 105h to 25ml/h. In the following 70h we saw a dramatic increase of the myoglobin level to a peak level of 17414µg/l. The propofol infusion was stopped then because of the severe rhabdomyolysis and because we thougt about the recently publihed review about the propofol infusion syndrome. Soon after removal of propofol myoglobin level decreased rapidly and the patient survived later on.

The propofol infusion syndrome is a very rare complication subsequent to propofol use. Our patient was severe head injured and received high doses of catecholamines as triggering factors like the patients described in literature. Rhabdomyolysis decreased rapidly after stopping the propofol infusion.CONCLUSION: Think about the propofol infusion syndrome in patients with severe rhabdomyolysis receiving high dose propofol long-term sedation and consider alternative sedative agents. 

In Northern Ireland the process of co-ordinating appropriate and timely therapeutic intervention for severe traumatic brain injury (STBI) is somewhat fragmented. The objectives of the audit were to obtain baseline epidemiological data for STBI in Northern Ireland and to review current regional critical care management.

This was a 6 month prospective audit. STBI patients were identified for inclusion through referrals made to the Regional Neurosurgical Unit (RNSU).

During the audit period 195 adult patients were referred to the RNSU. 53 patients were admitted to the Regional ICU (RICU). 79% of these patients were male and mostly in the 20-39 year age group. 42% of the injuries were due to falls, 33.5% road traffic accidents and 24.5% assaults. Alcohol was detected in 43% of the patients. ICP monitoring was utilised in 85% of cases, and on day 1 of admission intracranial hypertension (ICP >20mmHg) was diagnosed in 45%. This figure fell to 8% by day 7. Muscle relaxants were used for ICP control in 70% of patients on day 1 and in 15% on day 7. Over 50% of individuals developed a ventilator-associated pneumonia (VAP) during their RICU stay. This significantly increased the length of stay, but did not increase individual mortality. 45% of patients required a tracheostomy prior to discharge.

There was a high incidence of VAP in STBI patients in Northern Ireland. This may be related to the increased frequency of alcohol intoxication in these patients. Heavy reliance on muscle relaxants for ICP control may be a further contributing factor. In light of these findings new critical care management guidelines for STBI are being considered.Pradl R 1 , Chytra I 1 , Kasal E 1 , Bosman R 1 , ?idková A 1 , ?tepán M 1 1 Dept. of Anaesthesia and Intensive Care Medicine, Charles University Hospital, Plzen, Czech Republic

Transesophageal Doppler was confirmed as useful non-invasive tool for hemodynamic optimisation in group of elective surgery patients. The aim of prospective randomized study was to evaluate the efficacy of early hemodynamic optimisation in multiple trauma patients using transesophageal Doppler in comparison with traditionally used basic hemodynamic monitoring (arterial blood pressure, heart rate, central venous pressure).

Patients with multiple trauma and expected blood loss more than 2000 ml admitted and mechanically ventilated on interdisciplinary ICU of University hospital in 2003 were randomized in protocol group (Doppler) and control group (Control). Hemodynamics of Doppler group patients were immediately after admission to ICU managed according to the protocol based on data obtained by transesophageal Doppler. Hemodynamics of control group patients was aimed at generally used resuscitation endpoints -mean arterial pressure (MAP), central venous pressure, heart rate (HR), urine output and skin perfusion. The age, the APACHE II score and Injury Severity Score (ISS) were assessed. MAP, HR and blood lactate level (Lact) were evaluated at the time of ICU admission (MAP-1, HR-1, Lact-1) and after 24 hours of ICU stay (MAP-2, HR-2, Lact-2). Mann-Whitney, Wilcoxon, unpaired and paired t-test were used accordingly; p<0,05 was considered statistically significant.

A total of 35 patients (28 men and 7 women) were enrolled and randomized in Doppler (n=16) and Control (n=19) group. No differences between Doppler and Control group in age (40,6 ± 20,2 vs 28,1 ± 15,7), APACHE II score (26,0 ± 4,9 vs 21,9 ± 5,8) and ISS (27,6 ± 7,7 vs 27,0 ± 13,7) were found. No differences between both groups in MAP-1, MAP-2, HR-1, HR-2, Lact-1 and Lact-2 were detected, however significant differences between MAP and blood lactate level at the admission to ICU and after 24 hours of ICU stay were observed in Doppler group (see table) .MAP-1 mm Hg MAP-2 mm Hg p Lact-1 mmol/l Lact-2 mmol/l p Doppler 69,8 ± 8,4 87,5 ± 6,2 * 3,6 ± 1,3 1,9 ± 0,7 * Control 85,1 ± 8,2 86,1 ± 8,1 n.s. 3,7 ± 1,9 3,3 ± 2,3 n.s. n.s. -non-significant, * -p < 0,05 CONCLUSION: We conclude that early hemodynamic optimisation by transesophageal Doppler in multiple trauma patients can contribute to better tissue perfusion and elimination of oxygen debt.

The study is supported by a research grant IGA MZ CR ND/7712-3

Fotouhi Ghiam A 1 , Abootalebi S 1 , Tavana R 2   1 Neurology Unit, Internal Medicine Department, Al-Zahra Hospital, Bushehr University of Medical Sciences, Bushehr, 2 Shiraz University of Medical Sciences, Shiraz ,Iran INTRODUCTION: Awareness of the relative prevalence of diseases causing loss of consciousness (LOC) in a particular geographic locality could greatly facilitate the approach to patient management. So this study has established to determine the etiologies responsible for nontraumatic LOC and hospital outcome in an emergency ward (EW).METHODS: 483 patients older than twelve years old who present with LOC were enrolled in this cross sectional study during the 12-month period in the EW of the Al-Zahra teaching hospital. LOC was defined as a clinical state manifested by any decrease in level of consciousness ranging from confusion to deep coma.

These numbers of patients (accounting for 7% of the EW patient volume) were identified with a mean age of 49.9 years (54.8% men). Etiology was metabolic in 42.9% , structural in 40.1% and infective in 6% of patients. It remained unknown in 11% despite extensive investigation. The most prevalent causes in subgroups were cerebrovascular accidents (30.6%) , drug intoxication (22.35%), and hypoxic-anoxic conditions (11.7%) respectively. The history taking and physical examination were most useful in diagnosis. Computed Tomography (CT) scan plays an important role in diagnosis of structural causes. Lateralizing signs (25%) and Nausea/Vomiting (16.6%) were particularly evident in the presenting symptoms. Prognosis is highly dependent on etiology. The admission Glasgow Coma Scale significantly correlated with outcome (P < 0.001). Overall series hospital mortality was 24.7%. Most of the patients have been referred to center in less than 6 hours after LOC onset.

Metabolic causes were the commonest overall etiology. The number of undiagnosed cases are significantly higher than other similar domestic and foreign (USA , Europe , Asia , Africa) studies, so emphasis on educating the medical staff to approach to LOC and establish CPR committee in EW should be considered. Poor outcome was associated with low GCS score. 

Endobronchial blockade represents an alternative to a double-lumen tube (DLT)(1). The wire-guided Arndt Endobronchial Blocker (WEB, Cook Inc) can be coupled to a fiberscope and directed as a unit through an endotracheal tube into the area to be blocked. This is of particular interest in patients with a difficult airway in whom intubation with a DLT is contraindicated(2). In contrast to a DLT, that results in complete blockade of either the left or right lung, the WEB can be positioned in almost any portion of the airway, thereby allowing to isolate a single lobe. We report on the use of the WEB in a patient with bronchopleural fistula and pulmonary hemorrhage.

A 17yr-old male was admitted after being hit by a truck. Orotracheal intubation had been performed at the scene. CT scan revealed fractured ribs, severe bilateral lung contusion, bilateral pneumothorax, pneumomediastinum and -percardium. Chest tubes placed in the right thoracic cavity were suggestive of bronchopleural fistula. Bronchoscopy revealed a tear in the right lower lobe bronchus and significant bleeding into the airway. Due to massive leakage, selective ventilation of the left lung was decided. Because of severe mediastinal emphysema, the risk of airway loss during tube exchange seemed high. We decided to perform selective blockade using the WEB. The WEB was inserted through the endotracheal tube together with the fiberscope and endoscopically directed into the right lower lobe bronchus with its cuff proximal to the bronchial tear. Once the cuff was inflated, the bronchopleural fistula closed, and ventilation improved to normal within minutes. The WEB was left in place for 24hrs, and the fistula did not reccur thereafter. The patient's trachea was extubated on day 10, and he was transferred to a peripheral ward on day 12 in good condition.

The WEB for use in single-lung ventilation with single-lumen intubation proved to be an appropriate tool in an emergency situation caused by severe bronchopleural fistula. Intubation with a DLT was considered a high-risk maneuver because of severe mediastinal emphysema and difficult airway. With the WEB inserted through the endotracheal tube it was possible to isolate the injured right lower lobe from ventilation, to prevent spread of hemorrhage, and to avoid the risk of airway loss during tube exchange. Because the WEB is fixed to the fiberscope with a wire loop, both fiberscope and blocker can be navigated through the tracheobronchial tree as one unit, the WEB released as soon as in place.

Our expercience with the WEB prompts us to recommend this device as a highly practicable alternative to a DLT whenever one-lung ventilation or lobe isolation is required.

(1) Campos JH, Kernstine KH. Anesth Analg 2003;96:283-6.(2) Arndt GA, et al., Acta Anaesthesiol Scand 1999; 43:356-8. 

Chaparro M 1 , Prieto M 1 , Aragonés R 1 , Muñoz J 1 , Curiel E 1 , Arias D 1 , Delgado M 1 , Ruíz M 1 . 1 Intensive Care Unit, Hospital Materno-Infantil, Málaga, Spain

Postpartum haemorrhage is one of the most common causes of maternal morbidity and the primary cause of maternal mortality. Only a few case reports have shown that recombinant activated FVII (rFVIIa) successfully controlled intractable obstetric bleeding (1). Three obstetric patients with massive bleeding and clinical and analytical repercussion, without previous coagulopathy are presented.

The use of rFVIIa in three consecutive obstetric patients with unresponsive lifethreatening haemorrhage admitted to our intensive care unit within the last six months is reported. Demographic data, rFVIIa doses, timing of treatment and diagnosis, among other variables are presented in the 

Floros J 1 , Maratheftis N 2 , KolliasS 1 , Vletsas C 1 , Roussos C 1 1 ICU, 2 Neurosurgery, Evangelismos, Athens, Greece

Cerebral microdialysis is a relatively new technique for measuring the levels of brain extracellular chemicals, which to date has predominantly been used as a research tool. There are many reports which emphasize the importance microdialysis to monitor patients with head injury.

We describe a significant relation lactate/pyruvate ratio and ICP in ten severely head injured patients admitted in the ICU in the perioperative period. Microdialysis catheters inserted via a bolt fixation device together with the ICP catheter. The catheters implanted into the brain to reflect changes in the penumbra of a lesion under computed tomographic control. We used the standardized equipment (CMA Microdialysis OMA600).

The lactate/pyruvate ratio is a better marker of ischemia in these patients. There is a strong difference between the values (repeated measured ANOVA) L/P (p<0.001) and ICP (p<0.5) in tracking secondary ischaemic and edema events. The lactate/pyruvate ratio was increased in all ten patients 12-18 hours before any change in the CT scan. The lactate/pyruvate ratio is also a better marker of ischemia (p<0.005) than lactate alone (p<0.1).CONCLUSION: 1.Microdialysis is an effective tool for studying extracellular chemistry and, thus, has great potential for exploring the pathophysiology of secondary brain damage. 2. The sensitivity and specificity of microdialysis for ischemia and secondary damage are better than ICP.3. There are data to confirm that microdialysis can be used to direct therapy and influence outcome.

Moriwaki Y 1 , Sugiyama M 1 , Toyoda H 1 , Fujita S 1 , Yamagishi S 1 , Kanaya K 1 , Hasegawa S 1 , Kosuge T 1 1 Critical Care and Emergency Center, Yokohama City University Medical Center, Yokohama, Japan

Recently, most of trauma patients can be non-operatively treated. One of the most important issues is a few chances of experience of surgery for trauma patients (on-jobtraining). Many training courses (off-job-training) for initial trauma care are held frequently, which training course obtain good results in many country. However, it is obvious that on-jobtraining is more effective training method. The objective of this study is to clarify the hourly incidence of trauma patients and surgery for them in one typical urban emergency center and how surgeons are effectively trained for initial care of trauma patients in this center in the education of surgical specialty.

Our city Yokohama is one of the biggest city in Japan and has 4 of third level emergency center, including our center, for 3400,000 people. We examined 1,384 of trauma patients treated mainly by surgeons (neck-chest-abdominal trauma cases and polytrauma cases) in our emergency department (ED) including 270 cardiopulmonary arrest (CPA) patients, and 495 of them including 270 CPA who underwent emergency operation. The planning of training in the education of surgical specialty was discussed from a viewpoint of an hourly incidence of trauma patients and surgery for them.

Trauma patients were mainly transferred during the night shift: 14.9 of non-CPA trauma patients (26.7%) were transferred during the day shift and 40.8 during the night shift per 3 months. Surgeries for them were also performed mainly during the night shift: 3.0 of non-CPA patients (26.2%) and 7.0 of all trauma patients (including CPA) underwent surgery during the day shift and 8.3 and 17.8 patients, respectively, during the night shift per 3 months.

We conclude that trainee for surgeon in Japan can have adequate opportunity of the initial care and surgery for trauma patients if they belong to the emergency center as an exclusive staff and are on frequent night duty.