Effects of PEEP above the L-Pflex on gas exchange, hemodynamic and gastric tonometry in ARDS patients

OBJECTIVE
To find out if the status of a patient on arrival in the intensive care unit had any bearing on the immediate ICU outcome and if there are any correctable factors.


DESIGN
A retrospective survey.


SETTING
Paediatric intensive care unit (PICU) at Parirenyatwa Hospital in Harare, Zimbabwe between January 1997 and December 1998.


SUBJECTS
147 emergency surgical admissions.


RESULTS
There were 147 surgical patients admitted to the PICU during the period of whom 77 were male and 43 were emergency surgical procedures. All patients were coming from the operating theatres (OT). Factors associated with a poor ICU outcome were emergency surgical admission (p<0.001), hypothermia on arrival in the PICU (p<0.001) and requirement for ventilation (p<0.001). On multivariate analysis only the first two were associated with a poor ICU outcome.


CONCLUSION
Some simple measures could be undertaken to improve immediate ICU outcome in surgical patients in a resource limited environment, such as improving the ambient temperature in OT during surgery and standardising transportation to PICU.

The awareness of the diagnostic difficulty and the documented high mortality risk of perioperative myocardial infarction (PMI) has led to the wide use of work up to rule out PMI after major noncardiac operations. This has caused stable postoperative patients to be kept in monitored hospital beds for extended periods of time and to be subjected to additional tests. We hypothesized that the mortality of PMI is high and, therefore, the wide use of postoperative work up to identify these patients is justifiable. We performed the following study to prove our hypothesis. All patients in the recovery room after major noncardiac operations who underwent work up to rule out PMI were identified and followed. The PMI work up included care in an electronically monitored unit, physical assessment, continuous ECG monitoring, and three 12lead electrocardiograms and cardiac enzymes obtained at six to eight hour intervals. Data collection included patient demographics; preoperative cardiac risk factors; incidence of intraoperative hypotension, hemorrhage and ECG changes; type of anesthesia and operative procedures and their durations; postoperative ECG and cardiac enzyme results; the incidence of PMI and patient outcome.
Two hundred patients were studied; 85 males and 115 females. Their mean age was 62.9 years. Preexisting conditions included hypertension in 162 patients, peripheral arterial disease in 102, diabetes mellitus in 97, angina in 30, previous myocardial infarction in 41, and smoking in 107. Of 200 patients, 164 had an abnormal preoperative ECG. Vascular operations were performed in 104 patients, nonvascular abdominal operations in 48, and other operations in the remaining 48. Intraoperatively, hypotension occurred in 29 patients, blood loss of >500 ml in 25 and ECG changes in 10. There were no deaths. PMI occurred in 5/200 (2.5%) patients. Four had undergone vascular operations and one had had an abdominal operation. The mean age of the patients with PMI was 64.2 years. The duration of operation and blood loss were similar to those of patients without PMI. None of these patients developed cardiac failure or cardiogenic shock and none of them died.
It is common to evaluate patients with acute myocardial infarction (AMI) according to the Forrester classification. A high PCWP value is a good predictor of pulmonary congestion although there is no documented correlation between PCWP and the degree of pulmonary congestion in patients with normal PCWP. In this study, we sought to investigate the correlation between (1) PCWP and thoracic fluid contents (TFC) as an indicator of pulmonary congestion, (2) PCWP and cardiac index (CI) and (3) CI and TFC in patients with AMI with PCWP values <18mmHg. The thoracic electrical bioimpedance (TEB) method was used to measure TFC and CI. CI was also measured by the thermodilution (TD) method.
Thirty-six patients with a diagnosis of AMI were included. Patients were monitored with the TEB device (BioZ System, CardioDynamics International Co, USA) and the pulmonary artery (PA) catheter, in addition to conventional monitoring. We measured CI by means of TEB and thermodilution (TD) methods (CI-TEB and CI-TD, respectively) and compared values obtained from each patient with those two methods. We also measured PCWP with the PA catheter, and TFC values as an indicator of pulmonary congestion with the TEB device. The correlation between PCWP and CI, PCWP and TFC, CI and TFC were also examined.
In this study, we confirmed that TEB method might be clinically useful for CI measurement in patients with AMI. PCWP was negatively correlated with CI-TD indicating that decreased CI in our patient group was not due to systemic dehydration but to decreased cardiac contractility. The weak correlation between PCWP and TFC means that the tendency of pulmonary water retention exists with an increasing PCWP even in patients with normal PCWP. The clinical importance of this increase in TFC is not clear. TFC values had no significant correlation with CI values. This means that the degree of pulmonary congestion is unpredictable from CI provided that PCWP remains normal. This is consistent with the Forrester classification.

Evaluation of thoracic fluid contents in patients with acute myocardial infarction H Okawa, A Suzuki, I Sakai, H Tsubo, H Ishihara and A Matsuki
University of Hirosaki School of Medicine, Department of Anesthesiology, 5  Introduction: The aim of this study was to quantify the volume of pleural effusions (PEs) in the critically ill using ultrasound. PE was suggested on the daily postero-anterior chest radiography [1] in the semirecumbent position. All patients with suspected PE were investigated with ultrasound. We hypothesize that there is a strong correlation between the maximal width of the fluid lamella along the lateral chest wall (seen on sonography) and the volume of pleural fluid punctured.

Materials and methods:
The study was approved by the hospital Ethics Committee. Eighty-seven consecutive critically ill patients underwent a pleural puncture when ultrasound analysis revealed a lamella of more than 2 cm [2]. A total of 138 pleural punctures was performed in 87 individuals. Ultrasound was performed using the ACUSON, sequoia 512, with the patient in the semirecumbent position. The deepest possible puncture side in this position was marked for pleural puncture. The PE was gradually drained, 200 ml at a time, until the fluid was completely evacuated. The width of the pre-puncture effusion lamella as measured with ultrasound was compared case by case with the actual punctured volume. Statistical analysis was performed using linear regression analysis and Spearman (rank) correlation coefficient.
Results: Due to technical difficulties or missing data, 7 punctures had to be excluded. 131 punctures (67 on the left hemithorax and 64 on the right) remained for analysis. No complications were encountered as a consequence of the pleural puncture.
The sonographic measurements correlated very well with actual effusion volume on the left (r s =0.83) and the right side (r s =0.77). For both sides the level of statistical significance was taken as P<0.001.
The thickness of the fluid lamella was taken as the independent variable, while the actual effusion volume was taken as the dependent variable. The linear sonographic method was represented by the equation y=208.77×-317.12 (left), and y=178.38×-159.7 (right). Y is the predicted effusion volume in milliliters and×is the sonographically measured thickness of the effusion lamella in centimeters. The mean predicted error was 199.7 ml for the left side and 285.3 ml for the right side.
Discussion: There is a strong correlation between the sonographic measurements and the actual effusion volume which was punctured. Unlike the findings of other authors, we were not able to predict the punctured volume, based on the width of the lamella measured by sonography probably due to the wide spread of obtained data (scatter plot).  Introduction: Aiming for the improvement of 'spaghetti syndrome' caused by many leadwires for the measurements of biological signals such as electrocardiogram (ECG), BP, SpO 2 , EtCO 2 and lines for intravenous infusion, we developed the new small wireless ECG electrode (size: 3cm×10cm) consisting of a battery-operated telemeter equipped with two electrodes. The purpose of this study was to evaluate the ECG with the wireless ECG electrode placed on critically ill patients in the emergency room.
Methods: ECGs were recorded with both a wireless ECG electrode placed on the right subclavian area and a conventional three-lead ECG (II) in 30 critically-ill patients whose ECGs were not only of normal rhythm but also arrhythmic. The 12-lead ECG (II) was also recorded. The wireless ECG electrode was also evaluated during cardiopulmonary resuscitation.

Results:
The wireless ECG electrode was quickly and easily placed on the patients compared with the conventional ECG electrode. ECGs were recorded very well in all patients utilizing the wireless ECG electrode. It was also possible to detect various arrhythmias such as VPC, SVPC, Af, VF, VT with the wireless ECG electrode as clearly as with the conventional ECG electrode. The influence of body movement was less than for the conventional ECG electrode. The wireless ECG electrode was operative even during DC shock.
Results: Based on statistical analysis, PDs predicted compliance and failure with a sensitivity of 94% and a specificity of 84%. PDs also predicted improvement and deterioration at P<0.01. As illustrated in the figures, traditional hemodynamics diagnoses a patient as stable as late as 10 min before the occurrence of flash edema. The PD diagnoses a failing and deteriorating patient as early as 2.5 h prior to occurrence of flash edema and diagnoses the patient as critically ill 25 min prior to the occurrence of the flash edema.

Discussion:
This study suggests utility of PDs in accurately diagnosing compliance or failure which would allow early intervention and monitoring of the effects of intervention in real time as compared to traditional hemodynamic evaluation.

Department of Anaesthesia and Intensive Care, Faculty Hospital, Ostrava, Czech Republic
Introduction: High cardiac index (CI) and low systemic vascular resistance (SVRI) are frequently observed in septic patients and are essential in consideration of further therapeutic interventions. As these findings could be observed even in healthy individuals (i.e. during hard exercise) we decided to analyse the relationship between CI and SVRI in patients without apparent clinical signs of the inflammatory response and in patients with the different phase of the inflammatory response.
Methods: Thirty-one critically ill patients with pulmonary artery catheter inserted were included in this prospective study (average age 48±16 years, average sum of SOFA score 2.4±1.36, etiology: 60% traumatic patients), all of them with apparent clinical signs of the inflammatory response (heart rate >90 bpm, WBC >12000 or <4000, BT >38ºC or <36ºC, CRP >50mg/l) with microbiologically confirmed infectious etiology. In accordance with our findings (published in Int Care Med 1997, 23(Suppl.1.):S72) we identified 16 periods of generalization of the inflammatory response (fall in platelet count, antithrombine-III activity and serum albumin, positive fluid balance, hemodynamic instability). Within these periods, 101 hemodynamic measurements were performed, measured parameters are shown in Fig. 1. In periods without symptoms of generalized inflammatory response we performed 115 hemodynamic measurements, results of which are shown in Fig. 2. 130 hemodynamic measurements in patients without apparent clinical signs of the inflammatory response (25 patients, age 51±17 years, pulmonary artery catheter inserted in setting of preoperative hemodynamics optimization) we used as the control group. Results are shown in Fig. 3.

Results: See Figures
Conclusion: The relationship between CI and SVRI shows a different course in the phase of the generalized inflammatory response compared to the phase of the localized inflammatory response. The course of the relationship between CI and SVRI in patients without inflammatory response is nearly identical to that in patients with symptoms of the localized inflammatory response. On the basis of these findings, we presume that assessment of the relationship between CI and SVRI (but not isolated CI and/or SVRI) is needed for accurate therapeutic decision making.

Figure 3
Organ dysfunction development. Objectives: We investigated effects of 4% and 8% gelatinbased solutions, hydroxyethyl starch (HES) and ringer's solution on albumin escape rate (AER) and maintenance of plasma volume (PV) in a porcine fecal peritonitis model.

Design:
This was a prospective randomized, controlled animal laboratory study.
Measurement and main results: Fasted, anaesthetized, mechanically ventilated and multi catheterized pigs (20.8±1.8 kg) received 1g of feces per kg of body weight into the abdominal cavity to induce sepsis and were observed over eight h. At induction, animals were randomized to resuscitation with gelatin 8% (n=5), gelatin 4% (n=5), 6% HES 200/0.5 (n=5), ringer's solution (n=5) and compared to a nonseptic control group (n=5) receiving 6% HES 200/0.5. The infusion rate was set to maintain a central venous pressure of 12 mmHg. Red blood cell volume was analyzed using chromium-51-tagged erythrocytes (RBC: 51 Cr). PV was determined from standard formulae. AER was calculated using 99m-labeled technetium serum albumin. Systemic hemodynamics and oxygenation were obtained before induction of sepsis, and 4 and 8 h afterwards.
Conclusion: PV was increased by gelatin 8% only. This suggests a better remaining volume in circulation during sepsis as there were no differences between infusion volume in all colloid groups. On the other hand, systemic oxygenation was significantly better in animals treated with HES than with gelatin 8% or 4%. In this respect, it is interesting to note that histological investigation of lung and kidney revealed a bluish blubbery fluid in small blood vessels, glomerula and tubules of the animals treated with gelatin 8% or gelatin 4%. Introduction: Bacterial lipopolysaccharide (LPS) is believed to be one of the major pathogenic factors of Gram-negative sepsis. Sepsis may be accompanied with thrombocytopenia, platelet activation and elevated platelet-derived microvesicles. Higher formation of microvesicles has been associated with higher mortality in endotoxemic pigs.
Design: Prospective, randomized animal laboratory study.

Conclusion:
In this porcine sepsis model, volume replacement with colloid solutions seemed to reduce the detectable number of platelet derived microvesicles, which could be seen after the use of ringers solution. This may be a sign of less activation with colloids.

University Leipzig, Department of Internal Medicine I, Philipp-Rosenthal-Strasse 27, 04103 Leipzig, Germany
Background: In previous studies platelet dysfunction was described as a part of disseminated intravascular coagulation (DIC). These findings are also important in septic patients. In this present study, the association of the platelet function with the systemic inflammation and the development in different parts of the septic process in patients of an internal intensive care unit were investigated.
Methods: Twenty-five patients fulfilling clinical, hemodynamic and blood chemistry results of a systemic inflammation were included. The investigations were done in a two-day period. The characterization of the immune-state with IL-6, TNF-α and Procalcitonin used standard methods. The platelet activation marker P-selectin (GMP-140) was analyzed by flow cytometric detection ex vivo and after stimulation using 5 µM ADP and 10 µM TRAP-6. DIC was also characterized by standard laboratory results (platelet count, aPTT, AT III, fibrinogen, TAT, D-Dimer). The APACHE II-score evaluated the clinical situation.

Results:
The activation status of platelets was significantly associated with the process of inflammation. Pre-activated platelets (ex vivo) were seen in all patients with systemic inflammation (PCT P<0.03). During the measurement over a two-day period (five results for each patient) the pre-activation and the reagibility of TRAP-6-stimulation were significantly correlated to plasma levels of IL-6 (P<0.01) and TNF-α (P<0.04). Furthermore the detection of changes in platelet function started earlier then the measured results of common laboratory tests in DIC.

Conclusion:
Platelet function was tightly associated with the process of systemic inflammation. The dysfunction of the cellular part of coagulation could be an important marker of changes in the hemostatic system and the development of the disseminated intravascular coagulation. Introduction: Thrombocytopenia is a major concern in intensive care medicine. The incidence is greatly varying depending on the clinical setting.
Objective: To determine the incidence, severity, prognosis and therapeutic implications of thrombocytopenia in our 12-bed medical (noncoronary) ICU.

Methods:
We evaluated, in a prospective observational study over 13 months (1.11.1997-30.11.1998), all patients who stayed longer than 48 h in the ICU. Thrombocytopenia was defined as a platelet count below 150 000/µl.
Conclusion: 45.2% of patients became thrombocytopenic during their ICU stay. The degree of newly developed thrombocytopenia was highly correlated with mortality, length of stay, initial APACHE II score and the consumption of blood products.
Hemodynamic treatment scheme with ringer´s solution was based on ensuring baseline intrathoracic blood volume (ITBV). Hemodynamics and COP were measured before sepsis induction and 4 h and 8 h afterwards. Cumulative fluid balance was calculated after 4 and 8 h.
Abdominal and pleural effusion volumes were obtained after sacrificing the animals.
Results: See Table. before

Medical Clinic I, Department of Cardiology-Angiology, Justus-Liebig-University, Klinikstrasse 36, 35385 Giessen, Germany
Background: Impairment of the microcirculation is one of the main causes for multiple organ failure (MOF) in critical care patients. Trials with AT substitution in patients with sepsis showed a positive effect on the development and course of MOF. Recent experimental studies indicate that there might be specific AT-effects independent of the coagulation cascade, which can modulate endothelialleukocyte interaction and vascular permeability.

Methods:
The effect of AT substitution on LPS-induced microvascular leakage (ML) and leukocyte adhesion (LA) in the rat mesentery was investigated by means of intravital microscopy. Male CD rats were infused with 0.5 mg/kg LPS (E. coli) over 80 min. Vascular leakage was detected with FITC-marked rat serum albumin by fluorescence microscopy and evaluated by grey-value analysis with image processing software. Light microscopy was used to evaluate leukocytes adherence to the vessel wall. Two treated groups received 500 U/kg AT either 20 min prior to or 20 min after the begin-ning of LPS-infusion. Animals not infused with LPS, either untreated or treated with placebo (albumin), served as controls. One pre-treated group additionally received heparin at a clinically used dosage. Furthermore, interleukin-2 (IL-2) was used to induce ML and treated with AT prior to infusion.
Results: LPS-infusion led to a significant increase of ML and LA compared to controls. Both effects were reduced to the level of controls by the substitution of AT. No significant differences were found between the pre-treated or the post-treated group. IL-2 induced ML was also reduced by treatment with AT.
Conclusion: Substitution of AT, even when given after LPS, ameliorates vascular leakage and leukocyte adhesion to the vessel wall. Together, these effects improve flow conditions in the microcirculation. Hence IL-2-induced leakage could also be modulated by AT and a direct effect by AT on leukocyte adhesion seems to be likely.

Conclusion:
In the AT III treated group both abdominal and pleural effusion volumes were lower than in the controls. COP values 4 h after sepsis induction were higher and the cumulative fluid balance lower. This may indicate that AT III has a positive effect on capillary leakage in septic shock.
Methods: PMN of healthy donors were isolated by lymphoprep ® followed by hypotonic lysis of contaminating erythrocytes. Chemotactic activity was determined in modified Boyden chambers by the micropore-filter leading front assay in a 48-well system. Respiratory burst activity of PMN was measured fluorimetrically.
Results: Pre-incubation in vitro of human PMN with Kybernin ® P, immune-adsorbed ATIII or both isoforms significantly inhibited the migration toward fMet-Leu-Phe (fMLP), interleukin-8 (IL-8), and ATIII itself, in a concentration-dependent manner following a bell-shaped curve. Maximum ATIII effects were seen at a concentration of 1 U/ml, which is the level of ATIII found in normal human plasma. Purified ATIII did not deactivate PMN chemotaxis toward GRO-α, whereas fMLP, IL-8, platelet factor-4 (PF4) and GRO-α itself did. Checkerboard analyses indicate that the inhibition of migration was not due to a general impairment of the chemotactic activity of the cells but most probably involved receptor regulatory processes. Deactivation of chemotaxis toward IL-8 but not GRO-α suggests differential interaction at the CXC receptors 1 and 2. At comparable concentrations purified ATIII and Kybernin ® P did not affect resting or stimulated respiratory burst activity of PMN.

Conclusion:
Our results suggest that ATIII may be a physiologic regulator of the acute inflammatory response by protecting cells from premature activation.

NIBSC, Blanche Lane, South Mimms, Potters Bar, Herts EN6 3QG, UK; *Centeon Pharma GmbH, 35002 Marburg, Germany
Sepsis-induced disseminated intravascular coagulation is usually caused by endotoxin evoked production of cellular procoagulant tissue factor (TF) and pro-inflammatory cytokines, which further perpetuate the generation of TF. Large doses of antithrombin (AT) reduce mortality and morbidity in septic patients and there is increasing evidence to suggest that AT has anti-inflammatory properties in addition to its anticoagulant properties. In the present study, we investigated the effect of AT in three in vitro cellular systems: citrated whole blood, human umbilical vein endothelial cells (HUVECs) and mononuclear cells (MNCs). The cells were stimulated with lipopolysaccharide (LPS) for 4-6 h in the presence and absence of AT. Procoagulant tissue factor (TF) activity was estimated by a TFdependent clotting or chromogenic assay and interleukin-6 (IL-6) was measured by ELISA. In all three systems, 5-40 IU/ml AT was found to inhibit TF and IL-6 production in a dose-dependent manner. This inhibitory effect was not attributable to excipients or co-purified components of AT. Experiments with chemically modified AT and a low heparin binding fraction of AT indicated that binding to heparin and/or cell surface glycosaminoglycans is important for the inhibitory activity. Up to 40 µM of a specific thrombin inhibitor, r-hirudin, did not inhibit the production of TF or IL-in a 48-well system, after pretreatment of PMN with antibodies, pertussis toxin or enzymes blockers.
Results: Preincubation of PMN with effective concentrations of pertussis toxin, staurosporine or 3-isobutyl-1methylxantine completely blocked ATIII-induced migration whereas treatment with bisindolylmaleimide (GFX), a selective protein kinase C blocker, wortmannin or tyrphostin-23, had no effect. In assays of IL-8-induced migration, PMN responses to the enzyme blockers were comparable, whereas PF4-induced responses differed as they were also sensitive to GFX. Migration of PMN toward ATIII was significantly reduced by pre-treatment of cells with CXCR1 but not CXCR2 mAb, whereas migration toward IL-8 was antagonised by both CXCR1 and CXCR2 mAb. The mAbs had no effect on fMPL-induced PMN migration. A mAb to IL-1R, which was used as a control, was inactive as well.
Conclusion: Effects of ATIII on PMN migration appear to involve specific signalling pathways including CXCR1, Gproteins and phosphodiesterase and staurosporine-sensitive enzymes other than protein kinase C (i.e. protein kinase A or phospholipase D). Objective: Several investigations and our recent experience indicate that the intravascular inflammatory response and the clotting alterations registered during severe preeclampsia and HELLP syndrome are not an epiphenomenon but the cause of the clinical syndrome. The aims of this study were to investigate the effects of the ATIII substitutive treatment on cytokine plasma concentrations, assumed as a marker of endothelial damage and of the associated systemic inflammatory response. A secondary objective was to correlate the ATIII treatment with the evolution of both single or multi-organ dysfunction syndrome (MODS).

Materials and methods:
The study involved four patients with severe pre-eclampsia and three HELLP syndrome patients. Diagnostic criteria for severe pre-eclampsia were those published by the American College of Obstetricians and Gynecologists (diastolic blood pressure >110 mmHg and proteinuria ≥0.5g/l). Diagnostic criteria for HELLP syndrome required, in association with hypertension and proteinuria, thrombocytopenia (<150 000 cells/µl), evidence of hepatic dysfunction (aspartate aminotransferase [AST] and alanine aminotranserase [ALT] levels >40 IU/L with lactate dehydrogenase [LDH] level of >600 IU/l) and evidence of hemolysis (increased LDH and anemia). Plasma levels of tumor necrosis factor (TNFα and interleukins [IL1β and IL6]) were measured by enzyme-linked immunoadsorbent assay (ELISA). Plasma concentrations of Antithrombin III (ATIII) and Protein C (P-C) were measured by a chromogenic assay. ATIII was administered after the first sampling (admission) by a loading dose of 3.000 IU in bolus infusion and a maintenance dose of 1500 U/12 h over 4 days. Results were compared using the Mann-Whitney test. Maternal parameters and clinical data were compared using unpaired Student t-test.
Results and discussion: Clinical data are listed in Table 1 and results are shown in Table 2. These results demonstrate significant changes in fibronectin and ATIII concentrations between levels at admission compared with levels after ATIII treatment. It is very interesting to register a decrease in blood levels of cytokines and a reduced proinflammatory activity. The pro-inflammatory activity was increased in women with severe pre-eclampsia and HELLP syndrome patients and this result could also be explained by specific effects which are independent from the coagulation cascade (anti-inflammatory actions).

Conclusions:
The preliminary results of this study confirm our hypothesis as well as the utility of marker monitoring and substitutive treatment. We believe that it is necessary to develop a large study (as a phase III trial) to confirm our hypothesis and to achieve other significant results. Data were presented as mean(±sd). 6 in either of three cellular systems, suggesting that inhibition of thrombin might not be the main mechanism by which AT prevents the production of TF and IL-6. The results of this study have shown that, apart from the inhibi-tion of thrombin and other activated coagulation factors, AT may also downregulate the cellular expression of proinflammatory responses and therefore may have an added value in the treatment of sepsis-induced DIC. The acute septicemic form of Burkholderia pseudomallei infection or Melioidosis is associated with substantial release of endotoxin, TNF-α and IL-1. This inflammatory response leads to endothelial injury, activation of the extrinsic coagulation cascade, depletion of naturally occurring anticoagulants, microvascular thrombosis, organ failure and death. Plasma samples drawn at baseline and at time points during the illness from 30 patients with Melioidosis were assayed for D-dimer levels, Protein C and Protein S antigen levels and Antithrombin III functional activities. Results of samples drawn during the illness were averaged for each patient. Baseline and continued deficiencies of Protein C, Protein S and Antithrombin III were predictive of poor outcome in a statistically significant fashion by logistic regression.

P30 Decreased Protein C, Protein S and Antithrombin III Levels are predictive of poor outcome in
Endothelial injury as a result of inflammatory response to Burkholderia pseudomallei infection leads to coagulopathy and depletion of the natural anticoagulants Protein C, Protein S and Antithrombin III. Early and continued deficiency of these factors is predictive of poor outcome. Replacement therapy of depleted factors to achieve normal levels may be a worthwhile strategy for patients with Gram-negative sepsis. Background and purpose: Hypercoagulability and endothelial cell activation and/or injury are mutually related and often found in acutely ill septic patients, and recently they have been reported to be related to multiple organ dysfunction syndromes (MODS).It is not clear, however, which parameters indicating coagulopathy are most closely related to MODS. In this report, we analyze correlations among the severity of the illness including MODS and parameters related to coagulopathy including TFPI in acutely ill septic patients, in order not only to ascertain their close relationships but also to find sensitive and predictive markers of the severity of septic patients.

Materials and methods:
Five acutely ill septic patients with glucose intolerance were analyzed. Their blood glucose levels were strictly controlled by means of a bedside-type artificial pancreas (AP), STG-22, manufactured by NIKKISOH corporation in Japan. We selected septic patients in whom blood glucose levels were strictly controlled with STG-22, because some parameters related to coagulopathy, including Plasminogen Activator Inhibitor-1 (PAI-1), are influenced by blood glucose levels.

Interpretation and conclusions:
The severity of disease in acutely ill septic patients, indicated by the progression of MODS and endothelial cell injury, was closely related to coagulopathy characterized by hypercoagulability with decreased fibrinolysis. Treatment for hypercoagulability seemed to be justified in severely septic patients. Although the mechanism was unclear, TFPI, which seemed to be a parameter of hypercoagulability and endothelial cell activation, was thought to be a sensitive and possibly predictive marker of the disease severity in septic patients. Introduction: The problem of treating the coagulate disorders in cases of gestosis remains of topical interest in obstetrics; heparin is dangerous because of the possible development of thrombocytopenia, heparin-induced thrombosis, osteoporosis, and increased consumption of antithrombin III.

Methods:
The investigation was carried out in 2 groups of pregnant women with gestosis. To the first group (consisting of 34 patients aged 18-32), the injections of fraxiparin (2850 ME) were given once a day. To the second group (consisting of 32 patients aged 19-30), the injections of heparin (5000) were given 4 times a day. The effectiveness of the drug was checked clinically and according to the investigations of coagulate and thrombocyte gemosta-sis the day before the operation and on the 1st, 4th and 7th days after the operation.

Results:
In the group of patients with gestosis who received heparin, reduction of the number of thrombocytes to 21.4% (P<0.01) and growth of their aggregation activity to 18.3% (P<0.05) were observed on the 7th day of the postoperative period in comparison with the period prior to operation. In the group of pregnant women who received fraxiparin, a change in the number of thrombocytes and an increase in their aggregation activity were not observed.

Conclusion:
The use of fraxiparin is more effective than treating with heparin and, as a result, it can be used for prophylaxis and treatment of thrombosis in pregnant women.

P35
The effect of using a heparin-free flush system for central venous and pulmonary artery catheters on a general medical and surgical intensive care unit There was no statistical difference in the thrombus-associated complication rates between the two groups. http://ccforum.com/supplements/4/S1 tion disorders. However, a more valid monitoring for rhirudin based anticoagulation in continuous renal replacement therapy is required.   Blood samples were drawn at 0, 10, 60 and 240 min after starting the procedure from before and after the membrane. Serum and plasma samples were drawn at 0, 10, 60 and 240 min after the onset of CVVHF or IHD. After 30 min at 4°C, the blood samples were centrifuged at 3000 × g.

Discussion
Malonic dialdehyde (MDA) determination was done by HPLC according to the method of Wong et al Neopterin (NPT), Elastase, Procalcitonin (PCT) and antibodies to oxidised LDL (oLAb) were determined by commercially available ELISA methods. Determination of peroxides was performed by an enzymatic method based on the peroxidase reaction with tetramethylbenzidine as a chromogenic substrate.

Statistical analysis:
In the case of a Gaussian distribution the statistical analysis was done by the t-test. Other distributions were assayed by the χ 2 test.
Results: PCT and oLAb remained more or less constant in all observations. No statistically significant differences were observed before and after membrane passage or between the CVVHF and the IHD group. PMN-elastase concentrations increased constantly from start to the end of CVVHF and IHD. MDA levels reached the maximum 10 min after the onset of CVVHF and IHD. Again lower concentrations were observed in the IHD group compared with CVVHF. Peroxides were not detectable in most of the samples of the CVVHF group. In the IHD group, there were statistically significant differences in peroxide concentrations before and after the membrane passage. Neopterin concentrations decreased significantly after membrane passage in both CVVHF and IHD group. The decrease was more pro-nounced in the IHD group. Details in data for all parameters in both groups are presented in Table 1.

Discussion:
For inflammation parameters (oLAb, NPT, PCT) no significant increases in concentrations could be observed. This indicates that no pronounced inflammatory reactions arise due to the dialysis procedure. The decrease in Neopterin concentrations after membrane passage shows efficient clearance of low molecular weight substances during CVVHF and IHD. In contrast, PMN-elastase concentrations increased from start to the end of CVVHF and IHD thus indicating PMN activation rather than activation of macrophages. In case of lipid peroxidation parameters, we found a peak in MDA concentrations 10 min after onset of CVVHF and IHD, which might be a consequence of PMN activation after contact with the membranes. In conclusion, our data suggest that even beneficial procedures like CVVHF or IHD may activate inflammatory and lipid peroxidation processes. ence between CUPRO and PMMA was detected when age and APACHE II score were entered as possible confounders in a logistic regression model. There was also no difference between the two study groups regarding time on dialysis, number of dialysis sessions required, need for mechanical ventilation, or total parenteral nutrition.

Conclusion:
In summary, this controlled, prospective randomized trial did not reveal any differences in the outcome of dialysis-dependent ARF patients treated with CUPRO vs PMMA dialyser membranes.
Objective: In a previous experiment with crystalloid solution we have shown that during CVVH(D) CO 2 losses are proportional to the fractional ultrafiltration (UF) rate of circulating solution, with only a marginal impact of dialysis and temperature [1]. We report here results of a similar experiment with packed RBC.
Methods: Packed RBC were heparinized, diluted with crystalloid solution with bicarbonate to hematocrit of 0.35 and continuously saturated in a special mixing chamber to targeted pCO 2 value of 5 kPa. After equilibration, the mixing chamber was connected to a PRISMA monitor (Prisma, Hospal, France) at a blood rate (Qb) of 150 ml hr -1 and different CVVH(D) settings were tested for CO 2 elimination at two temperature levels (37 and 40°C). Blood samples were drawn in triplicate at each setting before and after the filter for blood gases analysis (ABL 520, Radiometer, Denmark). CO 2 content was calculated using standard formula [2] and CO 2 removal was expressed as pre-postfilter CO 2 content difference.  Background: Tissue factor (TF) is the most important initiator of intravascular coagulation. In monocytes the expression of pro-inflammatory cytokines and TF is con-trolled by the same transcription factors. Catecholamines, frequently used in sepsis therapy of critically ill patients, have been shown to inhibit endotoxin-induced expression of monocyte cytokines, such as TNF-α or IL-6 [1]. The aim of our study was to prove whether epinephrine (EPI) and norepinephrine (NOREPI) may also affect TF expression.

Methods:
To induce TF and TNF-α expression in monocytes we incubated citrated human whole blood (WB) for four hours with LPS (E. coli 055:B5; 50 ng/ml). TF expression on monocytes was determined by flow cytometry, TNF-α secretion was measured by ELISA.
Results: Incubating WB samples with 50 ng/ml LPS we observed an increase in the number of TF-positive monocytes from 3.9±0.7% to 42.9±2.8%. 5.5 nM EPI inhibited LPS-induced TF expression by about 16% (P=0.03), but the inhibitory effect was attenuated at higher EPI concentrations (>550 nM). LPS-induced TNF-α secretion (4.2±0.5 ng/ml) was inhibited by about 60% in presence of 5.5 nM EPI, and maximum inhibition of 75% was reached at 55 nM. Compared to EPI, higher concentrations of NOREPI were needed to get a significant inhibition of monocyte activity. Inhibition of TF-expression by 13% was observed at 550 nM and the same inhibition was observed at 5,500 nM. TNF-α secretion was inhibited in a clear dose-dependent manner and amounted at 5.5 nM NOREPI to about 35% and at 5500 nM to about 75%.
Conclusion: EPI is more effective than NOREPI to inhibit LPS-induced monocyte TF expression and TNF-α secretion. At higher concentrations EPI also seemed to have a stimulatory effect on TF expression, which could be due to an interaction of EPI-activated platelets [2] with monocytes.
Introduction: Splanchnic mucosal perfusion abnormalities have been implicated in the development of sepsis and multiple organ failure. In clinical and experimental settings, administration of dobutamine can increase hepatosplanchnic perfusion [1,2]. The administration of enoximone during and after cardiopulmonary bypass diminished endotoxin levels in liver venous blood indicating a beneficial effect on tissue perfusion and barrier function of the gut [3]. The aim of this study was to investigate the impact of dobutamine and enoximone upon liver perfusion and function in fluid optimized septic patients.

Methods:
After approval by the local ethical committee and obtained written informed consent by next of kin, 48 septic patients were included in this study. After fluid resuscitation according to an optimal left atrial filling pressure established by plotting left ventricular stroke work index against pulmonary artery occlusion pressure, patients were randomly treated with dobutamine (initially 5 µg/kg/min, increasing dose up to a maximum of 20 µg/kg/min) or enoximone (initially 2.5 µg/kg/min, increasing dose up to a maximum of 10 µg/kg/min) for 10 h up to a dosage where no further increase in left ventricular stroke work index was achieved. Hemodynamics, liver blood flow (LBF), hepatic tumor necrosis factor (TNF) and monoethylglycinexylidide (MEGX) kinetics to assess hepatic function were performed within the first 12 h of sepsis, before administration of inotropic support as well as 12 h and 48 h after treatment. Statistical analysis was performed using Wilcoxon signed rank sum and Friedman test.
Results: Basic patient characteristics (age, sex, APACHE III) did not differ between groups. Oxygen delivery and oxygen consumption increased in both groups without significant difference between groups.
Conclusion: Administration of enoximone in fluid optimized septic patients may be favorable in comparison to dobutamine. Enhanced DO 2 and VO 2 in both regimens resulted in an increased regional perfusion and improved splanchnic function only in the enoximone group indicated Baseline regional distribution of cardiac output (CO) depends on the physiological determinants of the endorgans. In congestive heart failure (CHF), this distribution is modified in favour of the coronary, brain and kidney circulations, depending on the main regulatory mechanisms of the vasomotor system, including the sympathetic nervous system, the vascular endothelium and the local non-endothelial mechanisms.
In this study, we measured the renal blood flow (RBF) and CO in 20 critically ill cardiac patients (17 males, mean age 49±10 years), utilizing both techniques of color coded duplex and echocardiography. All patients had CHF due either to dilated cardiomyopathy (12 patients) or ischemic heart disease (8 patients). A control group of 14 males (mean age 41±8.6 years) was also included. Following clinical evaluation, including the history and 12-lead ECG, all patients and control group were subjected to duplex assessment of RBF (ml/min/m 2 ), in post-absorptive state by measuring the internal diameters of both renal arteries at their aortic origins, and then estimation of the renal artery pulsed flow wave to measure the time average velocity, (the average of the velocity spectrum in one second) and then multiplied by 60 to express the flow in one min. The CO was measured echocardiographically through the pulsed wave Doppler trans-aortic flow. The whole procedure was repeated, first under low dose (5 µg/kg/min), and then under high dose (20 µg/kg/min) dobutamine, with 24 h apart. The regional renal percentage (RRP) was calculated and expressed as percentage of the cardiac index. Background: Sympathomimetics are frequently used to treat hypotensive newborns with one of the goals being the preservation of cerebral blood flow. No animal model has substantiated the efficacy of this practice. Although inotropes have been extensively studied in healthy animals, little is available concerning their efficacy in "sick" hemodynamically impaired newborn animals.
Design/methods: A laparotomy was performed in anesthetized piglets (10±1 days old, n=40) to clamp a major branch of the superior mesenteric artery for 30 min. One hour after, a persistent state of impaired cardiac function was produced. Cardiac output remained 24±2% below initial baseline. There was a parallel decrease in carotid blood flow (21±5%)(CBF) while the decrement in systemic mean blood pressure (BP) was small (73±2 to 67±2 mmHg; P<0.01). Those parameters remained the same for the next two hours when treatment consisted only of intravenous normal saline at a rate of 35 ml/kg/h. One hour after the start of mesenteric reperfusion, animals were randomized to additionally receive: Dobutamine (DOB), Dopamine (DP), Epinephrine (EPI) or no inotropes (CONT). The sympathomimetics (SYMP) were given at 20 min intervals, using a randomized latin squares design, at a rate of 5, 10, 20, 40 (DOB and DP) or 0.5, 1, 2 and 4 µg/kg/min (EPI). The animals were instrumented to measure: aortic, pulmonary artery and superior sagittal sinus blood gases and lactate, cerebral oxygen extraction (CBF ext); mean systemic and pulmonary artery pressures(PAP); cardiac output (thermodilution),carotid blood flow (ultrasonic flow transducers).
Results: See Table. Mean ±s.e.m of the % change from levels immediately prior to treatment.

Conclusions:
Depressed cardiac output is associated with a significant decrease in cerebral blood flow even though BP is minimally reduced. EPI may be more efficient than the other inotropes in supporting neonatal CBF when CO is impaired. Septic shock is characterized by hypotension, persisting after fluid replacement and requiring vasopressors. Mean arterial pressure (MAP) as endpoint of treatment providing adequate regional perfusion is unclear. In an acute endotoxic shock model, norepinephrine was used to reverse hypotension in seven fluid-resuscitated pigs, anesthetized with α-chloralose and equipped with flow probes around the portal vein and renal artery, renal and jejunal mucosal laser Doppler flowmetry and jejunal tonometry. MAP was increased by 10 and 20 mmHg above the shock level with norepinephrine. Seven shocked, fluid-resuscitated only animals served as control. Measurements were performed before 2 h-endo-toxin infusion and at the end of each increased MAP level. Raising MAP with norepinephrine by 10 mmHg increased significantly cardiac output, systemic oxygen extraction, portal vein blood flow, improved metabolic balance and tended to restore renal and jejunal mucosal flows to preshock levels. Increasing MAP by 20 mmHg further increased cardiac output and oxygen delivery but reduced portal vein blood flow and tended to decrease renal and jejunal mucosal flows. In conclusion, using norepinephrine to increase MAP by 10 mmHg in volumeresuscitated, acutely septic animals improved systemic and regional perfusion. Higher MAP increase did not add any benefit despite an enhanced cardiac output. Method: MOF was induced by two intraperitoneal (IP) injections of TAA (500 mg/kg) eight hours apart. Three groups were studied, Group1 receiving TAA only. Group 2 and 3 followed thr protocol for Group 1, however, Group 2 was pre-treated with the NO precursor L-Arginine (300 mg/kg IP) once daily and Group 3 was pre-treated with NO synthase inhibitor aminoguanidine (100 mg/kg SC) for three days.

Histology:
The histological sections show markedly less organ damage in the aminoguanidine group (Group 3).
Conclusion: Inhibition of iNOS using aminoguanidine significantly improves the incidence of MOF and mortality in the TAA model of MOF. Introduction: Excess NO production due to iNOS activation and cellular toxicity resulting from peroxynitrite (ONOO -) may contribute to organ dysfunction in septic shock. Therefore, we studied the effect of the combined ONOOscavenger and selective iNOS-inhibitor mercaptoethylguanidine (MEG) [1] on hepato-splanchnic hemodynamics and energy metabolism during long-term hyperdynamic porcine endotoxemia [2].
Results: See Table. Conclusion: MEG allowed for hemodynamic stabilization due to blunting of the progressive endotoxin-induced fall in MAP while maintaining CO but did not influence the parameters of hepato-splanchnic energy metabolism. Ongoing oxidative stress resulting from inadequate dosage of the compound may account for this result [3]. Introduction: In isolated-perfused lungs of endotoxin-challenged rats, inhibition of nitric oxide synthase 2 (NOS2) by aminoguanidine (AG) improved responsiveness to inhaled nitric oxide (NO) [1]. In dogs treated with lipopolysaccharide (LPS), S-methylisothiourea (SMT) and AG prevented worsening of hemodynamics and gas exchange [2]. In this study, inhibition of NOS2 by SMT and AG modulates responsiveness to inhaled NO in lungs of endotoxin (LPS)challenged rats in a dose-dependent manner. Results of both groups were expressed in fluorescence intensity (FI) and summarized in the Table. The functional assessment of Mos and PMNS in our shock model revealed an enhancement of phagocytic activity in endotoxin-tolerant animals both in Mos and PMNS with the highest increase in the latter. Introduction: Independently from leukocyte adherence, endothelial factors and mast cell activation promote microvascular permeability [1]. The platelet activating factor (PAF) has been shown to play a significant role in endotoxin-induced leukocyte adherence [2]. The aim of our study was to investigate if there is also a role of PAF in mediating leukocyte-independent microvascular permeability changes and activation of mast cells during endotoxemia. Therefore, during endotoxemia, microvascular permeability and mast cell activation were determined after inhibition of the L-selectin mediated leukocyte-adherence by fucoidin and after inhibition of PAF effects by the the PAF receptor antagonist BN52021.

Methods:
In male Wistar rats, microvascular permeability (MP), leukocyte adherence (LA) and mast cell activation (MCA) were determined in mesenteric postcapillary venules using intravital microscopy at baseline and at 60 and 120 min after the start of a continuous infusion of endotoxin (ETX; 2 mg/kg/hr, E. coli O26:B6) (group A, n=10). Leukocyte-endothelial interaction was blocked using fucoidin (25 mg/kg b.w.). In addition to the proce-dure in group A, group B (n=10) received BN52021 (5 mg/kg b.w.) after baseline measurements. Group C (n=10; control group) only received equivalent volumes of NaCl 0.9%. Statistical analysis was performed using student's t-test. A P-value <0.05 was considered significant.

Results:
In groups A and B, fucoidin prevented LA, and so there were no significant differences in LA between these groups. In group A, MP and MCA significantly increased starting at 60 min (P<0.05 vs baseline). There was no significant increase in MP and MCA in groups B and C. Differences between groups A and B in MP and MCA were significant at 120 min.

Conclusions:
The results of this study demonstrate that PAF plays a significant role in the initiation of endotoxininduced leukocyte-independent plasma extravasation and mast cell activation. Introduction: TNF is a potent inflammatory promoting agent that can potentiate organ and tissue injury. A possible role of TNF in causing local tissue damage following snakebite was recently demonstrated in an intact rat model. We compared the systemic effects of TNF on the hearts and lungs of rats following an intramuscular injection of a sub-lethal dose of Vipera asis venom (500 µg/kg, experimental groups) to equivalent injections of saline (control group).
Results: Systemic TNF activity, heart rate and blood pressure as well as lung permeability and neutrophil sequestration were then evaluated in both groups. The venom caused a significant reduction in heart rate and arterial blood pressure, and the serum TNF levels peaked after two hours. These values remained unchanged in the control group. In contrast, lung microvasculature permeability and neutrophil sequestration were not significantly different between the experimental and control groups. Orpegen, Heidelberg, Germany). The normal range of phagocytosis (expressed as mean intensity, MnI) was defined by a control group of 11 healthy volunteers.
Results: Phagocytosis of healthy volunteers was in the range of 34 MnI and 149 MnI. All survivors and 7 of 10 nonsurvivors (i.e. 84% of all patients) had a lower phagocytotic activity on more than 50% of days in septic shock compared to that of healthy volunteers. No patient had a phagocytosis greater than 149 MnI, the upper value of the normal range.
Conclusions: Phagocytosis is decreased in most patients during septic shock compared to healthy volunteers. Thus diminished phagocytotic activity may contribute to an impaired mechanism of bacterial elimination or a reduced resolution of infection in these patients. Moreover, phagocytotic activity of granulocytes does not discriminate survivors and nonsurvivors of septic shock. We reported that T 3 eliminated acidosis and improved survival rate in a rat model of sepsis. The aim of this study is to clarify the effect of T 3 on cytokine levels.
Male Sprague-Dawley rats, weighing 350-420 g, were ligated in the cecum with puncture (CLP method) under anesthesia with pentobarbital (40 mg/kg). After CLP, 50 ml/kg of saline was injected subcutaneously for fluid resuscitation. Rats were assigned two groups; no treatment group (control group, n=15) and a T 3 treated group (n=10). The T 3 treated group was given 3 ng/hr of T 3 using a osmotic pump embedded subcutaneously. Survival rate, levels of IL-1β, TNF-α, IL-6, IL-8, and IL-10 were studied 24 h after surgery.
Three animals died in the control group, whereas no animals died in T 3 treated group. There was no significant difference on mean value of IL1-β and TNF-α between two groups. Mean levels of IL-6 and IL-8 in the control group were 237.4 pg/ml and 5342.7 pg/ml, respectively, however, those in the T 3 group were suppressed to the level of 183.5 pg/ml and 55.2 pg/ml, respectively (P<0.05). Mean IL-10 level (6.0 pg/ml) in the T 3 treated group was lower than that (72.4 pg/ml) in the control group (P<0.05).
It is concluded that T 3 prevents the cytokine storm and improves the survival rate in this rat model of sepsis.
Introduction. IL-17 is a newly discovered cytokine implicated in the regulation of hematopoiesis and inflammation. Since IL-17 production is restricted to activated T lymphocytes the effects exerted by IL-17 may help to understand the contribution of T cells to the inflammatory response. We investigated the role of IL-17 in leukocyte recruitment into the peritoneal cavity. Objectives: To discriminate patients with strong or weak ex vivo TNF-α synthesis in response to whole blood stimulation with endotoxin (LPS) and to relate this ex vivo response to cytokine balance during and after cardiac surgery with cardiopulmonary bypass (CPB).

Patients and methods:
19 patients scheduled for coronary artery bypass graft (CABG) with CPB and 29 adult volunteers were entered into the study. Whole blood was stimulated with LPS, incubated for 16 h and TNF-α concentrations determined in the culture supernatant. Results obtained in both patients and volunteers were pooled. Individuals showing TNF-α concentrations lower or higher than the median value of 1052 pg/ml were defined as low and high responders, respectively. In the 19 patients, cytokine balance during and after CPB was evaluated by the plasma concentrations of TNF-α, interleukin (IL)-8 and IL-10.

Results:
In patients undergoing cardiac surgery, perioperative cytokine plasma levels were not significantly different in both low and high responders. In these two groups ex vivo TNF-α production did not correlate with either TNF-α or IL-8 plasma levels measured during and after CPB. In contrast, ex vivo TNF-α production inversely correlated with IL-10 plasma concentrations observed four hours after the end of CPB in the low responder group (Spearman, -0.76, P=0.01) and tended to do so in the high responder group (Spearman, -0.62, P=0.1).

Conclusion:
In our series, the ex vivo production of TNFα does not allow us to predict the magnitude of the systemic inflammatory response related to cardiac surgery, as evaluated by the plasma levels of TNF-α and IL8 during and after CPB. The negative correlation observed between the ex vivo TNF-α production and the production of IL-10 after CPB suggests that patients with a low ex vivo capability to synthesize the pro-inflammatory TNF-α have a high potential to synthesize the anti-inflammatory cytokine IL-10 in vivo. Whether this profile can be considered a risk factor for morbidity after cardiac surgery, as it has been shown for infectious diseases remains to be assessed. Methods: 16 young pigs were assigned to a temperature (T°) regimen during standardized CPB: normothermia (T°3 7°C; n=8) and moderate hypothermia (T° 28°C; n=8). TNFα and IL10 were determined by a pig specific ELISA. Six hours post-operative, tissue probes of the heart, liver, lung, and kidney were taken for standard-and immuno-histological examinations. Apoptotic cells were detected by an in situ apoptosis detection kit (TUNEL). Necrotic cells were counted by light microscopy.
Results: Pigs operated on in normothermia showed significantly higher TNFα production and lower IL10 production during and after CPB, and higher degree of cellular necrosis but lower degree of cellular apoptosis than pigs operated on in moderate hypothermia. In all animals, postoperative IL10 levels were negatively correlated with the production of myocardial apoptosis (P=0.005) and TNFα levels correlated significantly with necrosis in the heart and liver (P<0.05).
Conclusion: Cardiac operations are related to cell death via both the necrotic and the apoptotic pathways. In this setting, IL10 could be an inductor of apoptosis. By increasing the production of IL10 and the ratio apoptosis/necrosis, moderate hypothermia during CPB might therefore provide organ protection. Introduction: Sepsis, as well as coronary artery disease (CAD), is associated with high levels of circulating adhesion molecules [1,2]. Elevated plasma levels of these molecules may have clinical significance as markers of endothelial injury [2]. The purpose of the study was to investigate whether CAD influences the liberation of plasma adhesion molecules, cytokines and markers of the NO-system in septic patients and if these parameters may serve for identification of patients at greater risk for cardiovascular complications in sepsis.

Methods:
After approval by the local ethical committee and obtaining informed consent from relatives, 44 septic patients were included in this study, of whom 24 patients met the criteria of CAD [3]. Measurements were performed within 12 h (early sepsis) of onset of sepsis and after 72 h (late sepsis). Soluble adhesion molecules sEselectin, and soluble intercellular adhesion molecule (sICAM)-1 were analyzed by commercially available ELISA. NO-markers were measured by the Griess reaction and cyclic guanosine monophosphate (cGMP) by competitive radioimmunoassay.
Results: Basic patient characteristics did not differ between groups. CAD patients showed significantly increased sICAM-1 levels in early and late sepsis (Table 1) but this failed to be predictive for cardiovascular complications. ICAM-1 and cGMP correlated positively in both groups with the intravenously administered norepinephrine dosage in early and late sepsis. Cardiac index, oxygen delivery and mixed venous oxygen saturation were significantly decreased in CAD patients in early and late sepsis (P<0.05), whereas VO 2 did not differ significantly between both groups. Mortality was increased in tendency in CAD patients (75% vs 50% in non-CAD).
Conclusion: Plasma levels of sICAM-1, sE-selectin and cGMP were elevated in CAD but may not serve as markers for cardiovascular complications in sepsis. The tendency in increased mortality rate could be due to an impaired endothelial reserve in CAD patients. Introduction: Scoring systems such as APACHE-III and SAPS-II often fail to reliably predict the individual mortality risk particularly in patients with severe sepsis. Cytokines and procalcitonin (PCT) have been shown to play a crucial role both in the pathogenesis of sepsis and as diagnostic tools for the inflammatory process. The objective of this study was to investigate the relationship between plasma cytokine and PCT concentrations and the actual and predicted mortality using APACHE-III and SAPS-II in patients with severe sepsis.

Material and methods:
The SCCM/ACCP selection criteria were used to identify the patients with severe sepsis. After severe sepsis was diagnosed SAPS-II and APACHE-III scores were calculated. In addition, blood samples were taken daily for the analysis of IL-6, IL-10 and PCT plasma concentrations for three consecutive days. IL-6 and IL-10 were analysed using a sandwich type of a immunoenzymatic assay (Immunotech, Marseille, France) PCT plasma concentrations were measured using a specific, ultra sensitive immunoluminometric assay (LUMItest PCT assay, BRAHMS Diagnostica, Berlin, Germany).

Conclusion:
Our results demonstrate that the APACHE-III and the SAPS-II scoring systems evaluate a similar mortality rate in severe sepsis. However, in terms of individual mortality risk prediction, the reliability and precision of both scores is limited and does not allow a precise statement about the mortality risk. In contrast, the plasma concentrations of PCT and IL-10 were significantly higher in the non-survivors and showed a good correlation with the actual mortality. IL-6 was not useful in this context. Thus, daily analyses of plasma PCT and IL-10 concentrations might be a helpful tool to improve the individual mortality risk prediction in patients with severe sepsis, possibly in combination with scores assessing the severity of illness. Objects: The present study was undertaken to examine the relationship between the severity of burns and TNFRI and TNFRII levels in the acute post-burn period and the entire follow-up period.

Materials and methods:
Tumor necrosis factor α (TNF-α) and TNF receptor I and II (TNFRI and TNFRII) were studied in 24 burn patients who had a total burn surface area (TBSA) of 50.2±20.4%. TNFRI and TNFRII were quantified by enzyme-linked immunosorbent assay (ELISA) using a kit (Amersham, Buckinghamshire, UK). TNF-α was quantified by ELISA (Medogenix, Fleurus, Belgium). Introduction: Sepsis is a clinical syndrome resulting from an inflammatory response of the host to infection, mediated by the production of cytokines in circulation. The need to understand better the mechanisms that lead to multiple organ failure, the authors (AA) began a study of cytokines in these patients.

Material and methods:
In 1998 the AA began a prospective study of soluble cytokines, consisting at the moment of 54 patients, 36 with criteria of Sepsis/MODS (Multiple organ dysfunction syndrome) and 18 with SIRS (Systemic inflammatory response syndrome). The soluble cytokines (TNFα, IL-6, IL-8 and IL-10) were measured in serum samples by the quantitative sandwich enzyme immunoassay (solid phase) -RD systems, Inc. Minneapolis (CITOMED, PORTUGAL) -with a previously established protocol.

Results:
In the sepsis/MODS group, 7 patients with IL-6 <300 pg/ml survived; among the ones with IL-6 >300 pg/ml, only the five patients submitted to pulses of methylprednisolone and/or emergent surgery survived. The patients with SIRS always presented low concentrations of cytokines, and 14 of them survived.

Conclusion:
In the patients with Sepsis/MODS, the clinical severity was always related to high serum levels of IL-6, and/or its increase. Serum values of IL-6 >300 pg/ml were predictive of death. Although TNFα and IL-8 values were high in some patients, these showed a non-uniform behaviour, and the correlation of these values with clinical severity or outcome was not statistically significant as was IL-6. In those patients who survived, the clinical improvement was always followed by a remarkable decline in the serum values of IL-6. IL-6 is an important severity serum marker in equal or superior value to others described and accepted world-wide for patients with sepsis.

Department of Paediatrics, Sophia Children's Hospital, Dr Molewaterplein 60, 3015 GJ Rotterdam, The Netherlands
Introduction: Meningococcal septic shock is a life-threatening clinical syndrome that may lead to pronounced neuro-endocrine and metabolic alterations. In critically ill adults a GH resistant state has been well documented. Less information is available in critically ill paediatric patients. In the present study we evaluated the time course and pulsatile pattern of the GH secretion and the levels of serum IGF-I, IGFBP-1 and IGFBP-3.

Methods:
Children with a septic shock and petechiae/purpura requiring intensive care treatment were enrolled in this study. The paediatric risk of mortality (PRISM) score was used to monitor the severity of disease. A GH profile was done in the first 6 h after admission and at day 3. The serum IGF-I, IGFBP-1 and IGFBP-3 levels were measured on admission (T=0), and 24 and 48 h after admission. The serum levels of IGF-I and IGFBP-3 were corrected for sex and age, expressed as SDS-levels.
Results: Twelve patients fulfilled the inclusion criteria, 7 boys and 5 girls, with a median age of 22 months. The median PRISM score was 23. Three of the 12 patients died. The serum levels of IGF-I and IGFBP-3 were decreased in all patients, the median SDS being respec-tively -2.6 and -5.7. There was a significant difference between survivors and non-survivors regarding median GH levels during the first 6 h (7 mU/l vs 131 mU/l, P=0.01), serum IGF-I levels at T=0 (43 ng/ml vs 7 ng/ml, P=0.03), IGFBP-3 SDS levels at T=0 (-5.6 vs -7.7, P=0.05), and PRISM score (18 vs 34, P=0.01). The levels of serum IGFBP-1 at T=0 were increased in the non-survivors in comparison with the survivors (median being respectively 1315 and 262 ng/ml), but this difference didn't reach significance (P=0.15). Non-survivors showed very high GH levels without GH peaks during the first 6 h after admission. In survivors, the GH secretion pattern at day 1 showed a larger number of GH peaks and larger GH peak amplitudes than at day 1. A significant increase in serum IGF-I levels was found in all patients between T=0 and T=24 (P=0.04), followed by a significant decrease in serum IGF-I levels between T=24 and T=48 (P=0.01).

Conclusion:
Mortality in children with a meningococcal septic shock is associated with extremely increased serum GH levels, no oscillatory GH secretion, high levels of IGFBP-1 and a high PRISM score. In those who survived there are significant changes during the first 48 h in the GH/IGF-I axis, suggesting a decrease in the GH resistance.

Service de Réanimation Adultes, Hôpital Purpan, F31059 Toulouse, France
Introduction: Serum sPLA2 and CRP levels increase in patients with systemic inflammatory response syndrome (SIRS). High serum levels of PCT have been detected in patients with inflammatory conditions from invasive bacterial and fungal infections. The aim of this study was to determine the diagnostic value of sPLA2, CRP and PCT in septic shock.

Patients and methods:
Consecutive patients admitted to the ICU with shock were included. Patients were screened in two groups: septic shock (American college of chest physicians criteria) and non-septic shock. Upon admission, serum sPLA2, CRP and PCT were simultaneously analysed. Data, expressed as means ±SD, were analysed by an independent investigator not involved in ICU. The catalytic activity of sPLA2 was detected by fluorimetric assays (normal 10 mU/ml). PCT was analysed by commercially available Lumitest ® kit (BRAHMS, Berlin).
sPLA2, CRP and PCT values were significantly higher in patients with septic shock. The areas under the curve (ROC) of sPLA2, CRP and PCT were respectively 0.896, 0.792 and 0.765. The area under the curve of sPLA2 was significantly higher than PCT area (P<0.05).
Conclusion: PCT does not appear to be a better marker to discriminate septic shock and non-septic shock than sPLA2 and CRP.  The aim of my study was to evaluate how major abdominal surgery induces interleukin 6 and procalcitonin by itself and to determine the usefulness of these markers in diagnosis of infectious complications.

Study design: Prospective clinical study approved by local Ethics Committee.
Patients: Thirty patients, ASA II-IV, undergoing elective surgery of the gastrointestinal system were studied. Surgical procedures included: gastrectomy or oesophagus resection (N=10), Whipple's operation (N=10), bowel resection (N=5), partial liver resection (N=2), others (N=3). In all patients, antibiotic prophylaxis (24-48h) was applied. Eleven patients developed local infection (wound N=6, lung infection N=3) or general infection (N=2). None of the analysed patients died; the total period of hospital treatment varied from 15 to 110 days.
Methods: Blood samples were taken during induction of anaesthesia and on days 1, 3, 5 and 7 after surgery. IL-6 measurements were performed by immunoradiometric assay (IL-6-IRMA, BIOSOURCE) and procalcitonin was measured by illuminometric method (PCT LUMItest, Brahms). Number of SIRS criteria and SOFA score were assessed at the same time points.

Results:
We observed low plasma levels of both IL-6 and PCT prior to surgery and a slight increase in PCT at the first postoperative day in patients with signs of infection accompanied by increases in SIRS and SOFA assessment.
In two patients with general infection, the highest levels of PCT were 8.41 and 7.95 with IL-6 levels of 3500 and 3700 respectively. Introduction: Of the few biochemical parameters available to monitor critically ill patients and to control the course of therapy in severe inflammation, procalcitonin (PCT) and polymorphonuclear (PMN) elastase enzyme stand as innovative diagnostic parameters with features different from other presently routine indicators of the inflammatory response.

Methods:
Our work aims at assessing the potential role of PCT and PMN elastase enzyme in early diagnosis and early prediction of prognosis in patients (pts) with sepsis and septic shock. To achieve this goal we studied 20 pts with septic shock (16 male and 4 female, mean age 50.15 years), together with a second group comprising 10 pts (9 male, mean age 49.2 years) with systemic sepsis without shock. A third group including 20 healthy volunteers matching with age and sex, served as controls. Serum PCT and PMN elastase enzyme levels were estimated on admission for both pts and control groups with other laboratory investigations and clinical parameters. A multivariate, discriminate analysis was performed using the following variables: PCT, PMN elastase enzyme, albumin, α-1-antitrypsin, α-2-macroglobulin, and C-reactive protein (CRP), as independent parameters. With further patient subdivision to 6 survivors and 24 nonsurvivors only PCT could be identified as independent predictor for short term prognosis in patients with sepsis with an overall predictive accuracy 80% and with cut off value of ≥78ng/ml. While PMN elastase enzyme could be used as a parameter inside a model including the whole previous parameters with overall predictive accuracy 76.7%, and with cut off value of ≥142 µg/l.

Conclusion:
Serum PCT and PMN elastase enzyme are independent useful diagnostic markers for early detection of systemic inflammatory response syndrome with or without shock. However PCT has the advantage over the above mentioned parameters in being significantly predictive of short-term prognosis, with overall predictive accuracy of 80%. Procalcitonin, PMN elastase enzyme, α-1-antitrypsin, α-2-macroglobulin, CRP, and albumin, is a model which could be used for early prediction of complications of sepsis patients with overall predictive accuracy of 76.7%.

Department of Anesthesiology and Intensive Care, University of Istanbul, Medical Faculty, Istanbul, Turkey
Although the exact origin of procalcitonin (ProCT) in infection remains unclear, its association with the presence and severity of infection has been demonstrated. This prospective study aims to evaluate ProCT as a diagnostic marker of infection in critically ill patients with SIRS, and to investigate the relationship of ProCT levels with sepsis severity and outcome. Plasma ProCT values were higher in nonsurvivors in septic patients, and area under the ROC curve of ProCT in prognosticating outcome was higher than for both CRP and WBC. The present data demonstrated ProCT as a better diagnostic marker of infection in patients with inflammatory response. This laboratory procedure also seemed to be closely correlated with sepsis severity and outcome.

Departments of *Anesthesiology and † Clinical Chemistry, University Hospital, D-89070 Ulm, Germany
Objectives: To investigate whether procalcitonin (PCT) serum concentrations differentiate severity of disease in postoperative/post-traumatic patients with septic shock, in which severity of shock is defined by different ranges of dosages of norepinephrine, necessary for cardiovascular stabilization.
Methods: Over a six month period, 192 patients admitted to the intensive care unit (ICU) were studied. Out of them, 18 patients were in a first phase of septic shock lasting at least four days. PCT serum concentrations in these patients were measured on days 1, 2, 4, 6, 8, 10, 12 and 14 after onset of septic shock until death or discharge from ICU. PCT was measured using a commercial immunoluminometric assay (BRAHMS Diagnostica, Berlin).
Conclusions: PCT serum concentrations are not a marker of severity of disease in patients with septic shock, if severity of shock is defined by certain ranges of dosages of norepinephrine. Methods: Primary human cells (peripheral blood monocytes, umbilical vein endothelial cells) and cell lines (liver, renal parenchymal and lung fibroblastic lines) were cultivated under standard conditions. Basal and stimulated mRNA expression of PCT was investigated using a semi-quantitative reverse transcriptase polymerase chain reaction (RT-PCR). Intracellular PCT protein expression was verified by Western blotting and surface-enhanced laser desorption/ionization (SELDI). Experiments elucidating the intracellular location of PCT were performed after protein fragmentation in different fractions by secondary immunofluorescence and laser scan confocal microscopy.
Results: (1) A basal and inducible mRNA expression of PCT was found only in human peripheral blood monocytes.
(2) In these cells, a distinct influence of various proinflammatory mediators was observed. (3) Western blotting of monocyte lysates using various primary antibodies directed against PCT showed a strong intracellular protein expression. (4) Experiments with SELDI revealed a molecular weight for PCT in monocytes of 12.1 kDa. (5) Human monocytes express PCT protein in association with cytoskeleton. No PCT was found in cytoplasmic fractions.

Conclusions:
Since human peripheral blood monocytes produce PCT and its expression depends strongly from sepsis-related mediators, we conclude, that this cell population is one important source of elevated PCT serum levels during sepsis. Further experiments analyzing the role of Kupffer cells and liver parenchymal cells are in progress.

Hôpital R. Debré, 51092 Reims cedex, France
Introduction: Differentiation between acute bacterial infection from other types of inflammation is often difficult in ICU. Procalcitonin (PCT) was reported to be a new potential specific marker for infection. The aim of this study was to assess, at the admission, a PCT cutoff of infection in ICU patients.

Methods:
In a prospective study, we studied all patients admitted to our ICU between January 1999 and July 1999. PCT was measured at admission (J0) and on second day (J2). Patients were grouped according to Bones classification (SIRS, sepsis, severe sepsis, septic shock, SDMV). Statistical analysis was performed using SYSTAT (GENDEL). Discussion: At admission to ICU, PCT seems to be an interesting marker of early diagnosis of infection when the level is higher than 1.5 ng/ml. On the second day, a PCT level of 3 ng/ml seems to be more effective for diagnosis of infection than 1.5 ng/ml. Incidence of injuries and other health problems is known during prolonged running and recent studies suggest that exercise-induced damage may trigger production of proinflammatory molecules.

Results
The aim of the study: The present study investigates to what extent prolonged strenuous exercise influences the plasma concentrations of procalcitonin as well as cytokines inhibitors and anti-inflammatory cytokines.
Methodology: After informed consent, 4 marathon runners and 10 half-marathon runners were recruited and investigated 12 h before running, immediately after the end of the run and 24 h later. Blood samples were collected at each time and PCT concentrations in plasma were measured with a luminometric method (Brahms). The plasma concentrations of TNF α, IL-1 β, IL-6, IL-1 ra, sTNF-r1, sTNF-r2 were measured by ELISA. Statistical analysis used Systat (Jandel).

Results:
The mean duration for marathon running was 235±18 min and 105±10 min for half-marathon running. The highest concentration of IL-6 and TNF α was found immediately after the run. Time course (1 =12 h before; 2 =immediately after and 3 =24 h later) of blood PCT, ng/ml is presented in the figure (P<0.001).

Discussion:
This study suggests that long-distance running may trigger production of TNF α and pro-inflammatory cytokines, which stimulate the production of PCT.
As it has been demonstrated in sepsis, PCT may be an early marker of elevated cytokines and muscle and/or splanchnic damage. Introduction: The role of radionuclide imaging to diagnose sources of infection is rather confusing [1].The aim of this study was to evaluate the diagnostic value of gallium-67-scintigraphy in critically ill febrile patients and its impact on treatment and outcome.
Methods: Forty whole-body Ga-67 scans performed in 37 patients in a five year period (1994-1999) were retrospectively analyzed.
Result: 34 patients were surgical and 3 medical with a median age 69 years and median APACHE II 23. All patients were febrile with systemic inflammatory response syndrome (SIRS) and all, except two, were ventilated for more than three days. Ga-67 scan was performed either within the first two weeks (median 7 day) because of no improvement of sepsis (group 1) or later than two weeks (median 22 day) because of the resurgence of SIRS (group 2). Group 1 consists of 24 patients with APACHE II 21, multisystemic organ failure (MSOF) in 6 patients and 5 exitus. Four from the 24 Ga-67 scans showed positive findings (two soft tissue abscesses, one intra-abdominal collection and one osteomyelitis) leading to changes in treatment and improvement of outcome. Group 2 consists of 13 patients with an APACHE II 24, MSOF and exitus in 7 patients. On the 16 Ga-67 scans performed in this group, only one showed retrosternal Ga fixation and had no impact on treatment and outcome.
False negative results were observed in 9 of 15 patients with known lung and thoracic pathology proved by other investigations. In spite of rather high sensitivity in patients with abdominal pathology, the results were non-specific.

Conclusion:
Ga-67 scan in the critically ill may be helpful in localizing a pyrogenic focus in severe unimproving septic patients in the early phase after their admission. However, in patients with resurgence or sepsis in the late phase of hospitalization, Ga-67 scan appeared to be of very little diagnostic help and its value remains questionable.

Department of Intensive Care Unit, "IASO" Maternity Hospital, Kifissias 37-39, Marousi, Athens, Greece
Background: Postpartum endometritis is the most common infectious complication of pregnancy, which has a serious impact on women's health and can lead to severe sepsis.

Material and methods:
We encountered 150 women with clinical evidence of early (1-48 h) post-partum endometritis and SIRS (fever, tachycardia, tachypnea and increased white blood cell count and C-reactive protein) after either vaginal delivery or caesarian section. We obtained blood cultures (aerobic and anaerobic) at 38.5°C and before administration of antibiotics.
Conclusion: Bacteremia in early post-partum endometritis was confirmed in 10% of cases. Isolated bacteria were similar to those referred to in the literature (80% Gram negative bacteria, 20% Gram positive and anaerobic bacteria). Early recognition and appropriate management ensures successful outcome and prevention of complications such as severe sepsis.

Conclusions:
Overall, the distribution of micro-organisms in BSIs in Belgian ICUs has remained quite stable since 1992. The earlier reported increasing trend in Enterobacter aerogenes BSIs was again confirmed.

Department of Surgery, Nutrition Unit, "St. Andrew" General Hospital, Patras, Greece
Catheter guidewire exchange has been suggested as an effective method of treatment of catheter-related infection (CRI) in nontunneled central venous catheters. However, the old line is usually intraluminally colonised, so contamination of the new line through the guidewire occurs early.
To minimise this event, we evaluated a method of treatment consisting of catheter guidewire exchange followed by intraluminal antibiotic administration.
Methods: Eight consecutive patients with CRI were studied. They received TPN through a subclavian vein catheter for a period of 8-19 days prior to CRI. In the suspicion of CRI, the catheter was removed and exchanged over a guidewire with a new catheter, followed by catheter tip culture and peripheral blood culture. TPN administration was stopped for 2-3 days until isolation and identification of the responsible bacteria. After identification of the bacteria, TPN administration was started again and the appropriate antibiotic, as indicated by microbiological sensitivity, was administered through the new central venous catheter for a period of 10 days. All new catheters were removed after 10 days and cultured.

Results:
In all the patients was noted defervescence of the clinical signs of infection following catheter removal. The responsible pathogens were Staphylococcus (S. epidermidis in 4, S. aureus in 3, all sensitive to vancomycin), and Enterococcus in one (sensitive to ampicillin). There were no further problems with the new catheters in 7 patients, and the new cultures were sterile after 10 days of intraluminal antibiotic treatment; however, one patient developed a new episode of CRI in the 9th day of treatment, and the catheter tip culture showed colonisation by Proteus. In this patient, CRI was finally treated by removal of the catheter and replacement at a different site.

Conclusion:
Our results suggest that, in the majority of cases, CRI can be successfully treated by exchange of the catheter over a guidewire followed by administration of the appropriate antibiotic through the new catheter for at least 10 days.

UZ gasthuisberg, Inwendige ziekten, Herestraat 49, B-3000 Leuven, Belgium
Objectives: To test the hypothesis that urgent or semiurgent orotracheal intubation can induce short-lived bacteremia with oral flora. To find predictive factors for post-intubation bacteremia.

Methods:
Prospective study in a 17 bed medical ICU. Patients in need of orotracheal intubation (OI) could be included if no cardiopulmonary resuscitation was performed. One aerobe and anaerobe blood culture (BC) was taken immediately before OI, as soon as possible (preferable less than 10 min) after and 60 min after OI. If it was impossible to take a BC before intubation, patients could be included if both BCs after OI were taken. The indication for OI, ease of OI, experience of the doctor (<10,<100 or >100 OI previously) and the antibiotics used before OI were registered. Introduction: Tetanus is a toxic disease generally prevented by immunization, but it is still present in the Third World and it shows high mortality rate.

Results
Objective: To evaluate morbidity and mortality of tetanic patients in ICU during 18 years of follow up.

Methods:
This trial is an historical and a prospective cohort that studied 285 patients from October 1981 to October 1999. They were classified in two groups according to modified Ablett's scale: not severe (mild and moderate) tetanus and severe tetanus. The following variables were compared: incubation period (IP), onset period (OP), symptomatic period (SP), age, period of autonomic instability (AI), period of administration of benzodiazepine (Pbenzo), curare (Pcur), mechanical ventilation (PMV), clinical and infectious events and mortality during ICU follow up. In statistical analyses, the continuous variables were presented as the mean and standard deviation. Student t test was used to compare the two groups. The level of significance was P<0.05.
Results: This trial evaluated 71 patients with not severe tetanus and 214 patients with severe tetanus. The group with severe tetanus had a smaller IP and OP (P=0.002 and P=0.003, respectively), but a longer Pbenzo, Pcur and PMV (P=0.001) than the not severe group. There was no difference concerning the mean age between the two groups (P=0.26). The most common clinical events in tetanic patients were cardiorespiratory arrest (25.6%), pneumothorax (10.17%) and acute renal failure (10.17%). The incidence of respiratory, urinary and catheter infections were 76.14%, 38.24% and 7.71%, respectively. The mortality rates were 4.28% in not severe tetanus and 28.9% in severe tetanus (P<0.001).

Conclusions:
The severe tetanus group presents a high mortality rate, probably as a result of autonomic instability, despite intensive care. Infections related to prolonged mechanical ventilation and invasive procedures were the most frequent events noticed in this study.

ICU at Clinicas Hospital of Porto Alegre and ICU at Nossa Senhora da Conceição Hospital, Porto Alegre, RS, Brasil
Introduction: Leptospirosis is, in general, a self-limited disease but it can be associated with important complications such as multiple organic dysfunction and high mortality [1].
Objective: The goal of this paper is to evaluate the clinical characteristics and the morbimortality of severe leptospirosis in general Intensive Care Units from two general hospitals.
Methods: All cases with the diagnosis of leptospirosis confirmed by blood macroagglutination test and admitted from 1990 to 1999 were studied. We have analyzed their clinical and laboratory characteristics, the occurrence of multiple organ dysfunction and their mortality rate. We have also compared survivors with non-survivors. The quantitative variables have been compared by unpaired t-tests and the qualitative variables by a Chi squared test.

Results:
We described 33 adult patients, aged 40±16 years, of which 27 were men and 6 women. The most frequent clinical manifestations were fever (n=30), myalgias (n=29), jaundice (n=27) and dyspnea (n=27). All patients showed some level of organic dysfunction such as respiratory (n=29), renal (n=24), hepatic (n=24) and cardiovascular (n=20). The mortality rate was 52% (n=17). The comparison of non-survivors with survivors showed they have higher incidences of respiratory, renal, cardiovascular and neurological failures as well as higher levels of acidosis (P<0.05).

Conclusions:
In endemic regions leptospirosis has to be considered as a cause of multiple organic dysfunction with a high mortality rate mainly when respiratory, renal, cardiovascular or neurological failures are present. Results: See Table. Conclusions: 1) Over the period of years, early use of Quinine has given good and consistent results (10 mg/kg, up to 600 mg thrice daily for 5-10 days) and can cure all stages of Falciparum and other Plasmodia, save the hyp-nozoites of Vivax. 2) Parasites have developed resistance to Chloroquine, Sulfadoxine-Pyremethamine. Quinine is poorly tolerated when given after Mefloquine due to the addition of adverse effects especially those of GIT. 3) If patients who have been treated with chloroquine or other drugs, including if radical cure has been attempted, continue to spike, it is best to restart with quinine even if a species other than Falciparum is the culprit. 4) Hypoglycemia, otological complications were not really severe enough or irreversible to warrant stopping Quinine. Although ECG(QTc) or other cardiac disturbances were not seen in those group,a daily ECG is a must. 5) In severely toxic patients, a simultaneous intramuscular dose of 60 mg of Artesunate on each buttock followed by a daily dose of 60 mg for the next 5 days reduces the para- Ominous signs of Falciparum 1) Peripheral smear: due to invasion of young RBCs and sequestration in tissues, a large number of parasites may be invisible on the smear, thereby giving false assurance of a "low parasite index".
2) Premunition: in holo and hyperendemic areas, due to some degree of acquired immunity, patients are largely asymptomatic early on. It is almost always late when symptoms manifest overtly.
3) Cerebral Malaria is a "diffuse symmetrical encephalopathy" and therefore sparing mild neck rigidity, signs of meningeal irritation are largely absent. Mild delerium, obtundation, convulsions etc. are to be taken seriously.
4) Pregnant women are spared gross parasitemia for a long time due to heavy "parasitisation of the placenta".
site load as it is relatively new and thereby less resistant and is an effective schizonticide. It is, however, devoid of any action on the gametocytes. This was the only other group of drugs which promised future potential for concomitant use with Quinine. 6) It is best to wait for a minimum 72 h before discontinuing quinine or switching to an alternate regime. 7) Intravenous Quinine may be used only in the most toxic and orally intolerant population. 8) Primaquine must be given in Chloroquine resistant Vivax even after usage of Quinine for radical cure. Objectives: To assess the demographics, the morbidity and mortality of patients infected with Candida species.

Materials and methods:
The intensive care unit records of 26 patients diagnosed with fungaemia during 1998 were reviewed (eight medical, twelve surgical, six paediatrics).
Results: See Table. Conclusion: Despite treatment, fungaemia carries a high mortality.The actual mortality of 69% is higher then predicted by the APACHE II/PRISM scores. Although positive blood culture for fungus is well recognized as a hallmark of systemic therapy, sensitivity of blood culture is low. Detection of fungal DNA in the blood is expected to serve as a new clue to initiate antifungal therapy.
Purpose: To assess the incidence of positive fungal DNA in the blood of critically ill patients.
Methods: 34 blood samples were obtained from 24 patients. Blood samples were processed for polymerase chain reaction (PCR) to detect fungal DNA. DNA extraction was performed with the modified methods of Buchman. PCR primers to amplify conserved DNA sequences of 18S rRNA genes shared by most clinically important fungi were used. The amplification was performed in a thermal cycler through the cycles as follows.

Infectious Diseases Division University Hospital of Geneva, Geneva, Switzerland
Introduction: The incidence of fungal infections has been increasing for the last three decadess, especially in neutropenic, cancer and critically ill patients. It is associated with high mortality rates.

Methods:
We retrospectively reviewed medical charts of adult patients with fungemia from 1989 to 1998 at the University Hospital of Geneva. For patients with candidemia, demographic information, risk factors (antibiotic treatment, preceding surgery, cancer, intravascular devices, ICU stay), outcome, and the incidence of fungal species was analyzed as well as the overall consumption of fluconazole.
Results: Of 291 patients with fungemia, we reviewed 278 (96%) medical records, 184 (66%) were from male patients. The mean age was 60 years and the overall mortality 45.5%. Candida albicans (61%) was the most commonly identified species followed by C. glabrata (15%), and C. parapsilosis (6%). The incidence of candidemia ranged from 0.2 to 0.56 per 10 000 patient-days with the highest incidence in 1993 and the lowest in 1997. Of the risk factors analyzed only ICU stay was statistically significant (P<0.0001) (OR 3.42; CI 1.98-5.92). When candidemia was revealed, 38% of the patients were hospitalized in an ICU. Most candidemias occurred several days after admission, that is 52% after two weeks and 70% after three weeks.

Conclusion:
The regular use of fluconazole since 1990 may explain a decreasing trend of candidemias due to C. albicans, whereas those due to other Candida spp remained stable. The predisposing risk factors analyzed associated with a high mortality were similar to those described in other series. For the last two years, fungemia remained low and stable in our institution. No significant changes in the incidence of fungal species during the 10 year study period were observed. . In group B the daily dose of V was calculated on the basis of the NM even though the SCM was also performed but its result was not known. We fixed the therapeutic range of V between 5 and 10 mg/l for C min [1].

Results:
In group A, C min of vancomycin was inside the therapeutic range in all patients: In group B it was above 10 mg/l in 4 patients (14.3±2.7) and below 5 mg/l in 3 patients (3.9±0.3) ( Table 1). C max was 20.4±3 mg/l in group A and 20.1±6 mg/l in group B.

Discussion:
The results obtained show that the SCM is a useful tool to maintain the serum concentration (SC) of vancomycin inside the therapeutic range, while the NM is not so effective. The possibility of maintaining the right SC of V is essential to prevent both the side effects from overdosing (oto, nephro-toxicity) and underdosing the drug (uneffective antibiotic activity, bacterial resistance), but it can be problematic in the critically ill patient whose physiopathological characteristics may interfere with the pharmacokinetics of the drug. In the 4 patients who showed a C min of vancomycin above 10 mg/l we did not find any correlation between the elevated level of V and the clinical conditions; in the 3 cases with C min below 5 mg/l we observed an increase in the volumes of distribution because of peritonitis and fluid overload.
We conclude sustaining the usefulness of the serum concentration monitoring to establish the right dose of vancomycin in the critically ill patient, and underlining that the clinical conditions characterized by the increase in the volumes of distribution represent a risk factor of underdosing the drug. Table 1 Number of patients with the minimal serum concentration (C min) of V below, between and above the limits of the therapeutic range. Purpose: The purpose of the study was to assess the peak and trough levels of 2 g of vancomycin (VAN), given once daily (od), in critically ill patients, irrespective of renal function, instead of lg VAN twice daily.
Methods: Prospective, observational, open study. 2 g of VAN, diluted in 500 mls of 0.9% normal saline, infused over four hours. VAN blood-level assays were done at 1, 2 and 12 h after completion of the infusion, and 12 hourly there-after or until the next dose of VAN. (Random VAN levels are done routinely on the ICU at 08h00 daily.) Another dose of VAN was given if the random level was below 10 mg/l.

Conclusion:
Our peak levels did not exceed 80 mg/l, a level which has been associated with ototoxicity. The incidence of nephrotoxicity with VAN alone is about 5% and is unrelated to serum VAN concentrations. We did not assess the effect on renal function, since most of our patients requiring VAN were already on continuous renal replacement therapy.
Concerns have been voiced regarding the short postantibiotic effect (PAE) of VAN (<2 h). In aminoglycides the larger, od dose resulted in a prolonged PAE. The PAE of once daily VAN is unknown and at present we are maintaining trough levels above 10mg/l.
Our results show that od VAN 2 g should not cause an increase in ototoxicity as peak VAN levels are well below 80 mg/l and if trough levels are maintained above the minimum inhibitory concentration of the organism, then concerns regarding the short PAE are eliminated. The results concerning the univariate analysis and the discriminant analysis (Table 1), the respiratory tract infection rates and the nosocomial infection rates (Table 2) are given.

Conclusions:
The colonization pressure, vancomycin use and length of ICU stay ≥29 days were the main risk factors associated with VRE colonization.
Measures against VRE colonization resulted in a significant decrease in the respiratory and nosocomial infection rates. This decrease was probably due to reduction of other pathogens colonization pressure.  From these data, we concluded that MPO and elastase were released in the alveoli by activated neutrophils and that NTP were formed in situ by the oxidant activity of stimulated neutrophils (in situ production of peroxynitrite and/or activity of MPO on nitrite or peroxynitrite). This intra-alveolar oxidant activity led to the production of BAL fluids which were cytotoxic on alveolar cells and which enable the activation of the signal transduction pathway. However, the exact consequences of this NFκB activation and the particular compounds of BAL responsible for cytotoxicity remain to determine.
The paired t-test method was used to compare the results at the four different time frames. Patients with necessary manipulation in the infusion rate of fluids-vasopressors and/or the parameters of mechanical ventilation during the 6 h period were excluded from data analysis.
Conclusions: BAL in our patient population was a safe procedure. It did not affect oxygenation and pulmonary shunt. Compliance and P plat showed after a nonsignificant initial deterioration gradual improvement. The delayed fall of CI did not persist at the end of the study. The increase in PVRI needs further evaluation.

Department of Anaesthesiology and Intensive Care, University Hospital, Alej svobody 80, 30460 Plzen, Czech Republic
Introduction: Infections acquired in the intensive care unit (ICU) commonly complicate the course of critical illness. The administration of empiric antibiotics was identified as an independent risk factor for the development of subsequent nosocomial infection [1]. The aim of our study was to evaluate the influence of empiric antimicrobial therapy on the acquired pulmonary infection in patients with a chest injury on interdisciplinary ICU.

Methods:
We analyzed a prospectively-collected database of 64 patients with a chest injury admitted for at least four days to the interdisciplinary ICU of university hospital in 1999. Patients were retrospectively divided into a group receiving, from admission to the ICU, antimicrobial therapy (ATB group, n=41) and a group without antimicrobial therapy during first 48 h after admission (No ATB group, n=23). The age, APACHE II, ISS, TRISS, length of ICU stay, duration of mechanical ventilation, rate of isolated chest injury and associated diagnoses (multiple injury, abdominal injury, head injury), ICU mortality, and rate of acquired pulmonary infection were evaluated. χ 2 test, Mann-Whitney and unpaired t-test were used accordingly; P<0.05 was considered statistically significant.

Results:
No differences in most of the parameters were found between groups, except for the length of ICU stay, duration of mechanical ventilation and the incidence of acquired pulmonary infection (Table 1).
Conclusions: Empiric antimicrobial therapy in patients with the chest injury on ICU may increase the rate of pulmonary infections, prolong duration of mechanical ventilation and extend the time of ICU stay. There is a need for controlled studies to define the role of antibiotics in empiric therapy in trauma patients.

Department of Cardiac Anaesthesia and Intensive Care, Poliambulanza Hospital, Via Bissolati 57, 25124, Brescia, Italy
Objective: To evaluate the efficacy of NINMV in cardiac surgical patients. In order to verify if, beside already established advantages in critical care patients [1] (less intubation related complications, minor discomfort for patients and less need for sedation), NINMV can be useful in managing the weaning process in the postoperative cardiac surgical course.
Design: Retrospective survey. We used NINMV with two indications: (1) patients with unexpected postoperative respiratory failure requiring reintubation; (2) patients extubated early and with incomplete postoperative recovery of cardiorespiratory stability.
Setting: Five bed intensive care unit at a cardiothoracic surgery centre.
Patients: Twenty NINMV patients representing 4.1% of the 484 patients operated on from October 1998 to September 1999 in our centre. Urgent surgery occurred in 12% of cases, emergency surgery in 4.75% of cases. Overall in-hospital mortality was 14/484 (2.2%).
Measurements and main results: Twenty patients. were treated by NINMV (two with indication 1 and 18 patients. with indication 2). The two patients with indication 1 were successfully weaned and discharged. None of the other 484 patients in this series needed reintubation. Only 3 patients had to be reintubated (n°3 for unexpected haematological complication, n°8 for psychotic disturbances, n°18 failed full haemodynamic recovery). Aim: To study the association between time to extubation and outcome after esophagectomy.

Methods:
We retrospectively reviewed the records of all esophagectomies between 1990 and 1998. Early extubation was defined as extubation within 12 h. The primary outcomes evaluated were respiratory complications, hospital mortality and length of ICU stay. Outcomes between the two time periods 1990-94 (first period) and 1995-98 (second period) were compared. In the second time period, the relationship between time to extubation and outcomes was also evaluated.
Results: There were 78 esophagectomies in the first period (1990-94) and 84 in the second period (1995-98). In the second period, more patients received epidural analgesia (64% vs 34%, P=0.001) and more patients were extubated early (41% vs 6%, P<0.001) compared to the first period. There were no significant differences in the rates of respiratory complications (27% vs 34%) and hospital mortality (6% vs 10%) when comparing the second with the first time period.
In the second time period (1995-98), 31 of the 84 patients were extubated early (median time to early extubation was 6.8 h postop). More patients in the early extubation group received epidural analgesia (90% vs 50%, P=0.001). There were again no significant differences in the rates of respiratory complications (29% vs 30%) and hospital mortality (6.7% vs 7%), but the late extubation group stayed longer in the ICU (median stay 2 days vs 1 day, P=0.006).

Conclusions:
Our experience suggests that early extubation after esophagectomy in stable patients is safe, and may decrease the length of ICU stay with cost-saving implications. We believe that thoracic epidural analgesia may facilitate early extubation by enabling patients to cough and cooperate with physiotherapy in the early postoperative period.

Conclusions:
The OCC-C1-SAC is significantly less affected by balloon laryngoscopy.

Department of Anaesthesia, Frenchay Hospital, Bristol BS16 1LE, United Kingdom
Introduction: The Combitube airway (Kendall UK Ltd., Basingstoke, England) is included in the European Resuscitation guidelines for the management of the emergency airway [1]. In the trauma patient tracheal soiling is usually from the upper airway [2]. Although the proximal, large cuff of the Combitube may protect the airway, this has not been formally assessed.

Method:
After local ethics committee approval, 10 ASA 1 and 2, starved patients undergoing routine elective general anaesthesia involving non-depolarising neuromuscular blockade and mechanical ventilation of the lungs gave informed consent. Only patients with grade 1 direct laryngoscopic views were included to ensure adequate views of the vocal cords. The small adult sized Combitube was placed into the oesophagus, all patients being of the appropriate height, and the cuffs inflated. 10 ml of 0.1% methylene blue dye was instilled into the mouth. Full monitoring was used, and adequate ventilation ensured throughout. At the end of surgery all dye was suctioned away and the oral cavity and proximal Combitube cuff manually cleaned with absorbent gauze. The Combitube was removed and the airway examined laryngoscopically.

Results:
One patient had an initial airway leak that settled with minor Combitube realignment prior to dye instillation.
No laryngeal dye was detected in 9 patients. One patient had significant blue staining of the vocal cords.

Conclusion:
The Combitube therefore protects the majority of patients airways from aspiration of dye from the oral cavity. In the trauma setting blood and oral debris may therefore be prevented from entering the trachea. The one failure was probably a result of patient movement by surgeons as this same patient developed a ventilatory leak mid-operation. However, patient movement would be expected in the trauma population. Introduction: Percutaneous tracheostomy (PT) is recognised as a safe and effective method for prolonged ventilatory support, avoiding the complications of lengthy endotracheal intubation. Bleeding has been reported as the most common perioperative complication in most series [1]. Patients with severe liver disease frequently have a coagulopathy and thrombocytopaenia. We wished to assess the safety of PT in this group of patients, who are at high risk of haemorrhagic complications.

Methods:
A retrospective study of PT's performed in patients admitted to a dedicated liver intensive care unit in a supraregional liver centre over two years. All procedures followed the Ciaglia technique [2] and were performed by, or under the supervision of, experienced operators. All patients had liver disease, which included decompensated chronic disease, hyperacute and acute liver failure and liver transplantation with severe postoperative complications.
Patients considered to be at particularly high risk for bleeding, based on platelet count (Plts) and international normalised ratio (INR) received transfusions of platelets and/or fresh frozen plasma prior to the procedure. The model was applied on ventilated adult patients while inpiring different levels of PCO 2 . Estimates of PaCO 2 and alveolar deadspace were compared to the results of invasive determination.

Results:
Preliminary results indicate a good correlation between invasively and non-invasively determined PaCO 2 and alveolar deadspace (Table 1).
Conclusions: Using gas void of CO 2 as one of the two inspiratory gases further simplifies the formula. Our preliminary findings show that a difference in the levels of inspired PCO 2 of at least 3 kPa gives an estimate of PaCO 2 with an accuracy of 10% or better. We provide a non-invasive method for the accurate estimation of PaCO 2 and alveolar deadspace ventilation. We suggest its implementation in ventilators for the close monitoring of pulmonary treatment response. Proportional Pressure Support (PPS ™ ) is a partial ventilatory support that overcomes patient respiratory system elastance (Ers) and resistance (Rrs) by applying pressure 'proportional' to volume (volume assist, VA) and flow (flow assist, FA) respectively. The aim of this study was to evaluate its clinical use at the bedside in critically ill intubated patients affected by acute lung injury (ALI).

Methods:
Thirty previously healthy patients affected by post-traumatic (emopneumothorax and lung contusion, N=21), infective (pneumonia, N=8) and inflammatory (pancreatitis, N=1) ALI (American-European Consensus Conference on ARDS) were studied. After several days (range 2-7, mean 4) of assist/controlled (A/C) ventilation (BIPAP ™ ), patients were switched to PPS once they were stable. Ers and Rrs were determined during a short period of volume control ventilation (square flow wave, tidal volume 10 ml/kg) using the inspiratory hold technique. On the ventilator (EVITA 4, Draeger) FA and VA were set to 80% value of patient Ers and Rrs respectively. The Automatic Tube Compensation (ATC ™ ) was also used at 100% compensation. O 2 inspiratory fraction (FiO 2 ) and positive end-respiratory pressure (PEEP) were the same as in A/C mode. Progressive reduction of FA, VA, PEEP and FiO 2 was used as the weaning technique, during the healing phase of the disease. The following parameters were ascertained daily: arterial blood gas analysis, respiratory rate (RR), tidal volume (Vt), minute ventilation (Ve), patients' comfort, number of respiratory distress that needed changes in ventilatory mode and failure to wean. The FA and VA values at which the patients were able to sustain spontaneous breathing and extubation were registered.
Results: PaO 2 /FiO 2 ratio remained unchanged or slightly better compared to A/C mode, and progressively returned to normal during recovery. PaCO 2 ranged between normal values and pH ranged between 7.45 and 7.40 for each patient during the study period. The respiratory pattern changed greatly in terms of RR and Vt from one patient to another and in the same patient during the days, while Ve remained more stable. All patients, except two, were always comfortable and they were successfully weaned and extubated at mean values of FA=7±2 cmH 2 O/l/sec and VA=8±3 cmH 2 O/l. The two patients that experienced respiratory distress were affected by pneumonia and needed a change of ventilatory mode.
Conclusion: PPS ™ with ATC ™ may be used to assist the ventilatory needs of patients with post-traumatic, infective or inflammatory ALI. The clinician must not be frightened by an unusual respiratory pattern in the presence of a comfortable patient, because with PPS ™ patients are able to breathe as they like to maintain a normal pH, most likely because the neuroventilatory coupling is improved by this mode, as described by Younes [1]. Extensive studies are needed to evaluate the role of PPS ™ with ATC ™ among the ventilatory modes now available to support the patient with ALI. Introduction: During Acute Respiratory Distress Syndrome (ARDS), the lung density increases along a gravity gradient, causing a hydrostatic pressure gradient, even if the edema is homogeneously distributed. In an oleic acid induced ARDS experimental model, we studied by CT scan the influence of lung hydrostatic pressure on regional pleural pressure.

P112 Selected biochemical values and organ dysfunction assessment in prediction of difficult to wean patients
Methods: ARDS was induced in eight sedated, paralyzed and mechanically-ventilated dogs, by an oleic acid dose (0.075 mg/kg) injected in the pulmonary circulation. Before and after ARDS induction, a CT scan basal section was taken, and pleural pressure was directly measured throughout two flat pressure sensors (wafers) positioned into the pleural space, in the most upper part and in the most dependent part of the lung. We calculated the superimposed pressure (SP) by mean density (ρ) and height (h) of the CT scan section, according to the following formula: SP=ρ×h [1]. SP represents an estimation of hydrostatic pressure that weighs at the lowest level of the CT section.

Conclusion:
Pleural pressure seems to change as a function of superimposed pressure, and both these pressures change as a function of the sterno-vertebral level. However, the slope between ∆Ppl and ∆SP is lower than 1 because also other variables (thorax shape, regional compliance, etc.) probably influence the SP effect on the pleural pressure changes. PEEP titrated by the L-Pflex of static PxV curve of respiratory system associated with recruitment maneuvers improved oxygenation and mortality in ARDS patients. In order to study the effects of PEEP 2 cmH 2 0 below and above the L-Pflex without and with a recruitment maneuver on gas exchange and hemodynamic parameters, we analyzed foyr children (7 to 14 months) with ARDS criteria: acute bilateral RX infiltrates, PaO 2 /FIO 2 <200, echocardiography without left cardiac dysfunction, (<5 days of installation). All the children were sedated and paralyzed. After a ventilatory control period (TV=10 ml/kg, PEEP=5 cmH 2 0), a constant flow (0.2 l/min) PxV curve was done for each child and the L-Pflex was calculated. Then the children were ventilated in VCV (6 ml/kg, RR=20/min-Servo-Siemens-300) for 30 min with PEEP 2 cmH 2 0 below the L-Pflex without recruitment maneuver, then 30 min with the same PEEP level associated with a recruitment maneuver (PCV=15 cmH 2 0, PEEP between 30 and 40 cmH 2 0, according to an immediately previous thoracic CT-scan for each child that showed less than 5% of the lung area between -100 and +100 HU). Then the children were ventilated for 30 more min with PEEP 2 cmH 2 0 above the L-Pflex without a recruitment maneuver, and finally 30 min with the same PEEP level associated with the recruitment maneuver.

P116 Effects of a lung recruitment maneuver keeping PEEP before and after L-Pflex on gas exchange in child ARDS patients
Results: See Table. Conclusion: A recruitment maneuver according to the findings of the thoracic CT-scan improved oxygenation mainly after PEEP set 2 cmH 2 0 above the L-Pflex of the PxV curve, improved CO 2 exchange without hemodynamic impairment, in child ARDS patients. Introduction: Static inspiratory pressure-volume (P-V) curves are typically S-shaped with a lower (LIP) and an upper inflection point (UIP), probably because of successive recruitment of lung units. It is often said that PEEP should be set above LIP to avoid derecruitment during expiration. However, as shown in a rabbit model [1], the effect of PEEP on lung volume is determined by the expiratory part of the P-V loop, which has a markedly different shape.
Methods: Eleven lambs with gestational age 119-141 days (term =145days) delivered by cesarean section were given surfactant and subsequently mechanically ventilated.
Using an interrupter technique, static P-V loops were recorded between an end-expiratory pressure of a) 0 cmH 2 O (ZEEP) or b) 5cmH 2 O and an inspiratory pressure of 35 cmH 2 O.
Results: Loops obtained during insufflation from ZEEP had a large hysteresis; LIP was 16-21 cmH 2 O, while UIP was above 30 cmH 2 O, if seen at all (Fig. 1 loop A) and the expiratory limbs showed a high deflation stability down to a pressure of 6-11 cmH 2 O with a steep segment at lower pressures. Loops from 5 cmH 2 O had a nearly linear inspiratory limb, and much less hysteresis ( Fig. 1 loop B).

Conclusion:
In preterm lambs with fully-recruited lung volume, derecruitment can be avoided by a much lower PEEP than indicated by LIP. During mechanical ventilation of immature subjects, the upper inflection point of the expiratory limb (UIP exp ) is probably a better indicator of optimal PEEP than LIP of the inspiratory curve.

Figure 1
Loops at 15 min of age in a 132 day GA lamb. A: loop from ZEEP; B: loop from positive end-expiratory pressure (5 cmH 2 O).

Dipartimento di Bioingegneria, Politecnico di Milano, Italy; *Istituto di Anestesia e Rianimazione, Ospedale Policlinico, IRCCS, Universita' di Milano, Italy
The effect of PEEP on improving oxygenation and FRC is well known [1], not so the effects of PEEP and tidal volumes on gas exchange and FRC. Objectives: To study the cardiovascular and respiratory effects of a recruitment maneuver using PEEP in postoperative cardiac surgical patients.

Methods:
A prospective, randomized, double-blind study was performed. All patients were stabilized on a standard ventilatory protocol (10 ml/kg at 10/min; with FIO 2 of 0.8 and PEEP=5 cmH 2 O). PEEP was either left at 5 cmH 2 O (control) or increased (1 cmH 2 0/min) until 15 cmH 2 0 (recruitment), when cardiovascular endpoints were reassessed. PEEP was then returned to baseline at the same rate. Primary endpoints were the proportion of patients achieving predetermined oxygenation criteria for extubation (PaO 2 ≥70 mmHg on ≤50% FiO 2 ) at 1 h, and time to extubation.
Results: 100 patients (mean age 63 years; 79% male) were randomized to either recruitment (n=55) or control (n=45). In the recruitment group, at peak PEEP there was a small but statistically significant increase in PAWP and CVP, with a small decrease in mean BP and cardiac index.
Time to extubation was similar in each group. Oxygenation criteria for extubation at 1 h were achieved by 93% of the total recruitment group versus 82% of controls (P=0.13).
In the subgroup with an initial PaO 2 /FIO 2 ratio <300, 94% achieved oxygenation criteria for extubation at 1 h versus 61% of controls (P<0.01). Chest radiographs taken after completion of the protocol showed significantly less basal pulmonary collapse in the recruitment group (20% versus 53% in controls; P<0.01).

Conclusion:
The use of this incremental PEEP maneuver after cardiac surgery improves pulmonary gas exchange, probably via recruitment of otherwise collapsed pulmonary segments. These benefits are more pronounced in those patients with initially poor gas exchange. We found that both PEEP and TV, but not their interaction, have an effect on volume distribution in the UT and AB (P<0.05). At increasing PEEP and TV, the increased abdominal and decreased upper thoracic contribution to inspired volume were associated to a decreased compliance of the UT. Finally, we found significant (P<0.001) exponential relationships between UT and AB contributions to TV and E UT .

P121 Effects of PEEP and tidal volume on elastances and distribution of volume changes of the different chest wall compartments A Aliverti, R Dellacà, A Lo Mauro, E Carlesso, W Del Frate, P Pelosi*, D Chiumello* and A Pedotti
We conclude that at high PEEP and TV, the elastance of the upper thorax decreases and it causes a more 'abdominal' distribution of TV.

ICU, Hospital Israelita Albert Einstein, São Paulo,SP-Brazil, Av. Albert Einstein 627/701-5 andar-CTI-adultos-CEP-105651-901, São Paulo, SP-Brazil
PEEP titrated by the L-Pflex of static PxV curve of respiratory system improved oxygenation and mortality in ARDS patients. In order to study the acute effects of PEEP on oxygenation, CO 2 exchange, hemodynamic parameters and gastric tonometry we analyzed six ARDS patients (<5 days of installation) after 30 min on PEEP of 5 cmH 2 0, then 30 min on PEEP 2 cmH 2 0 above the L-Pflex and then 30 min after PEEP of 5 cmH 2 0 again. They were all sedated and paralyzed. A Swan-Ganz catheter with a continuos cardiac output (Baxter ® ) and a continuos gastric tonometer (Tonocap ® ) were inserted in each patient. The tidal volume and respiratory rate were kept constant (8 ml/kg and 20/min-VCV-Ventilation-Servo-Siemens-300). L-Pflex was titrated by the static PxV curve of the respiratory system (random volumes).
Results: See Table. Conclusion: PEEP titrated by the L-Pflex of the static PxV curve of the respiratory system improved oxygenation without impairing global as well as regional hemodynamic parameters in ARDS patients. Introduction: In the Acute Respiratory Distress Syndrome (ARDS), although the edema is homogeneously distributed, a hydrostatic pressure gradient squeezes the gas out from the alveoli, causing a lung density increase along a gravity gradient. We studied by CT scan, in an oleic acid induced ARDS experimental model, the influence of different V T and PEEP levels on the regional nonaerated tissue during volume-controlled ventilation.
Methods: ARDS was induced in six sedated, paralyzed and mechanically-ventilated dogs by an oleic acid dose (0.075 mg/kg) injected into the pulmonary circulation. After ARDS induction, the volume-controlled ventilation was maintained, changing three tidal volumes (V T ) (12, 24 and 36 ml/kg) and two PEEP levels (5 and 15 cmH 2 O). During each of the six randomized steps a CT scan section has been taken, at end-expiration and end-inspiration, 5 cm above the diaphragm. For data analysis, we divided the CT scan section into an upper and a lower part. The nonaerated tissue mass of the upper or lower region was calculated from the densities (ρ) (-200 H <CT <+100 H) and the volume (V), according to the formula m=ρ×V [1]; then it was standardized for the area of each lung region. To evaluate the effect of V T , PEEP level and lung region (upper/lower) on the amount of nonaerated tissue, we performed analysis of variance, in condition of end-expiration and end-inspiration, respectively.

Results:
The behavior of nonaerated tissue (g/cm 2 ) is shown in the table (data expressed as mean ±SD).
The amount of nonaerated tissue is greater in the lower region than in the upper one, in both end-expiration and end-inspiration (P<0.01). In both regions PEEP and V T independently affect the amount of nonaerated tissue (P<0.01). There is interaction between PEEP and V T only at end-inspiration (end-expiration: P=n.s.; end-inspiration: P<0.01).

Conclusion:
The effects of PEEP level and V T on lung recruitment are greater in dependent lung regions and at end-inspiration.

State Scientific and Clinical Center of Coal Miners' Health Protection, Leninsk-Kuznetsky, Russia, 652509
Purpose: Early diagnosis of acute injuries of lungs in patients with severe brain injury (SBI) during acute period providing, with the comparison of results in dynamic control, reduction of radial load and of examination time.

Materials and methods:
We examined 25 patients with SBI (in admission the severity rate was less than 8 points according to the Glasgow scale) at 1, 2, 3, 5, 7 and 14 days. We developed the algorhythm of patients examination: high resolution computer tomography (HRCT) of four levels with assessment of densitometric index in four points of each level.
Results: Roentgenologic signs of 96% patients during 1 day; from these patients 62.5% had first rate with the increase of density index by 13.7%, 37.5% patients had second rate with the increase of density index by 25.2% in comparison with density index. Main localization of pathologic changes registered in posterior basal parts of the lungs. Pleural exudate was diagnosed in 50%, hydrothorax in 4%, aspiration pneumonia in 8% patients.

Conclusion:
The proposed method, in contrast to standard CT, allows the appreciation of the revealed changes, to diagnose the early symptoms of ARDS on the first day, to analyze the development of pathologic in dynamics and to conduct the treatment correction in time.
Objectives: To investigate the effects of two different gas ventilation strategies during partial liquid ventilation (PLV) on gas exchange and lung mechanics in acute lung injury.
Interventions: Acute lung injury was induced by the infusion of oleic acid and repeated lung lavages with 0.9% NaCl (20 ml/kg). After a stabilization period of 30 min the animals were randomized to two groups with different PEEP levels and tidal volumes (Group A: Vt >12.5 ml/kg, PEEP 6 cmH 2 O; Group B: Vt 9 ml/kg, PEEP 12 cmH 2 O). The pre-oxygenated and warmed perfluorocarbon liquid (30 ml/kg) was instilled into the trachea over 15 min without changing the ventilator settings.

Measurements and main results:
Airway pressures, tidal volumes, static respiratory compliance, inspiratory and expiratory airway resistances and arterial blood gases were measured. Data were obtained at baseline, after lung injury and at 60 and 120 min of PLV. See Table 1.

Data analysis:
Values are given as mean ±SEM. Comparisons were made by ANOVA for repeated measures. A Pvalue of <0.05 was considered significant. The infusion of oleic acid (0.07 ml/kg) combined with five lung lavages induced a significant reduction of PaO 2 /FiO 2 from 486±24 torr to 60±3.8 torr (P<0.01).

Conclusions:
During PLV both gas ventilation strategies significantly improved oxygenation in an animal model of acute lung injury. Static compliance of the respiratory system was not different between both gas ventilation strategies. Inspiratory airway resistance was significantly higher during the low PEEP-high Vt gas ventilation strategy, whereas the high PEEP-low VT strategy resulted in significantly higher PaCO 2 values.  Results: Murray, APACHE II and MOF scores were similar in both groups during admission and there was no difference in oxygenation response to prone positioning between two groups (Table 1). However, the MOF score was higher in extrapulmonary ARDS on the day of prone positioning. Mortality rate was higher in the extrapulmonary group and on discharge this group had higher Murray, APACHE II and MOF scores ( Table 2).

Conclusion:
Our retrospective data suggest that oxygenation response to prone positioning was similar in both groups, but mortality rates were higher in extrapulmonary ARDS patients, which correlates with MOF scores.   The available study is designed by physicians and nurses. Besides the question of patients' benefit concerning gas exchange, handling, the acceptance of the nurses, and economic consequences are proved.
Method/material: Patients with ARDS, or those patients identified as requiring to be nursed in the prone position with a Horowitz-Quotient (PaO 2 /FiO 2 ) <250, were turned over into the prone position. In a randomised procedure the patients were placed into a 180° prone position (face down), a 135° prone position (near side position) or they were treated in a Rotation bed (RotorestR). Gas analyses defined the clinical effect of the position on gas exchange. Also changes in skin integrity, skin status and the clinical outcome of proning were documented.
Additionally, the number of nurses/physicians being involved in positioning the patient and the time taken was documented.
The prone position interval is fixed at 4-6 h depending on clinical data and the personnel situation.
End of the positioning-treatment is defined by clinical data and a Horowitz quotient >300. Also the patients treated in rotation bed shows a similar clinical outcome concerning the gas exchange.
But here also the handling and the economic resources necessary are incomparably high and at every time you need a special introduced nurse.

Conclusion:
This study is being continued to get evident data for a clear point of view. The target is to develop evident criteria for the kind of prone position, not only in dependence of the clinical outcome but also concerning the economic and personnel possibilities in an ICU. A clinical treatment is not only orientated in doing the best for the patient but it is also limited by personnel acceptance and economic resources. The kind of treatment is established when the benefit for the patient is recognisable, when the handling is simple, does not need complicated equipment and a lot of personal resources. It must be safe and show a small quantity of complications.
These criteria, as shown by our study, are applicable for the near side position. Our goal is to develop criteria for the different kind of positioning possibilities concerning the clinical problems in gas exchange during the disease process.

Department of Pediatrics, University of Graz, Austria
Objective: To evaluate the effects of inhaled nitric oxide on gas exchange and outcome in pediatric patients with acute respiratory distress syndrome (ARDS).
Design: Case series report.
Setting: Pediatric intensive care unit of a tertiary care children's hospital.
Patients: Seventeen pediatric patients with ARDS requiring mechanical ventilation with an FiO 2 >0.5 at a positive end-expiratory pressure >6 cmH 2 O, and whose PaO 2 /FiO 2 ratio was <100 torr were enrolled in this study.
Interventions: Initially inhaled (NO) was applied at 10 ppm using a microprocessor based system. A positive response after 15 min of NO inhalation was defined as an increase in arterial oxygen saturation >5%.

Measurements and main results:
At the start of NO inhalation the oxygenation (OI=Paw*FiO 2 *100/PaO 2 ) and ventilation (VI=PaCO 2 *PIP*RR/1000) indices were 29±2.8 and 43±4.1, respectively, the PaO 2 /FiO 2 ratio was 69±7 torr, and the static compliance of the respiratory system 0.51±0.04 ml/cmH 2 O/kg. An initial positive response to inhaled NO was observed in 59% of patients. Many years mortality in ARDS, often complicating CCT, is still very high and reaches 70-80%. In ARDS patients, hypoxia is a result of a 'vicious circle' when acute respiratory insufficiency intensifies acute cardiac insufficiency which, in turn, deepens respiratory failure.

Patients and methods:
The study is based on the data obtained from 45 patients with moderate-severe CCT. During conventional treatment in ICU, after 48-72 h, in 21 patients (48%) an ARDS developed, and three of these patients were treated with HBOT in a mono-placed barochamber. All treatments were conducted under 1.6-2.0 ata for 40-60 min each every day during 4-15 days, depending on progress of the recovery. Monitoring of cardiac function was carried out using non-invasive impedance cardiography (IC). Stroke Volume Index (SVI) and Cardiac Output Index (COI) were calculated. Respira-tory functions were checked using a gas-analyzer. All ARDS patients were divided into three groups: group A-4 surviving patients after conventional therapy; group B-14 deceased patients after conventional therapy, and group C-3 patients surviving after addition of HBOT.

Results:
Obtained data was exposed to statistical analysis using Student's unpaired t-test and the results are presented in Table 1.
Conclusions: This study clearly shows that: 1) in ARDS patients after CCT the state of cardiac function is the factor determining development of respiratory hypoxia; 2) HBOT is a decisive treatment improving cardio-respiratory function which leads to the favorable outcome; 3) patients with CCT must be treated with HBOT just after the trauma, before ARDS has developed.  Phases  I  II  III  I  II  III  I  II  III  I  II  III The differences between the means was considered significant if P<0.05 (* or # ), P<0.01 (** or ## ). * -vs phase I; ** -vs phases I and II; # or ##vs group B.

John Farman Intensive Care Unit, Box 93, Addenbrooke's NHS Trust, Hill's Road, Cambridge, CB2 2QQ, UK
Background: There is a wide variation in the reported incidence of acute respiratory distress syndrome (ARDS) due to the use of different diagnostic criteria. The publication by the American European Consensus Conference in 1994 of diagnostic criteria for ARDS has made comparisons of the incidence of ARDS more reliable. Only one study has looked at the incidence of ARDS in the UK. This was a retrospective survey in 1988 [1] that reported an incidence of 4.5/100 000 population/year.

Method:
We report a prospective observational study of the incidence of ARDS in 10 ICUs in 9 hospitals covering a population of 1.89 million people over the age of 15 years. The consensus conference criteria were used to diagnose ARDS. Acute onset was defined as occurring within five days of onset of illness. A study co-ordinator in each participating ICU identified patients and there was regular feedback from the study organisers to ensure all cases were identified.
Results and discussion: Over the first 10 months of data collection, 68 patients met the criteria for the diagnosis of ARDS, giving an incidence of ARDS of 4.3/100 000 population/year. The mean age was 50.9 years. The mean length of stay on ICU of survivors was 16.9 days and 7.8 days for non-survivors, the intensive care mortality was 38%. These interim results give an incidence of ARDS that is similar to that reported by the previous UK study and recent studies from other countries [2]. This figure would now appear to be a more accurate estimate of the incidence of ARDS in an unselected population. The study is ongoing and it is planned to follow up the survivors on a long-term basis.

International Mechanical Ventilation Study Group
We determined if outcome differed in patients who developed ARDS before and after the institution of mechanical ventilation (MV). A prospective study of 5156 patients who underwent MV >12 h was done in 361 ICUs from 20 countries. Among those patients, 261 (5%) had ARDS at the onset of MV (prior-group), and 313 (6%) developed ARDS 48 h or more after the onset of MV (after-group). The mean time from onset of MV to development of ARDS was 4±2 days. In the after-group, MV was initiated for the following conditions: sepsis (19%), aspiration (16%), pneumonia (15%), and trauma (15%). MV parameters were similar in both groups. The patient characteristics and outcomes are shown in Table 1 (mean±SD).
Patients who developed ARDS after onset of MV were more likely to have ventilator associated pneumonia (38% vs 26%, P<0.05) and coagulopathy (31% vs 23%, P=0.03). In summary, patients who developed ARDS after the onset of mechanical ventilation were more likely to develop complications, required longer duration of ventilator support, and had higher ICU and hospital mortalities.

Faculty of Applied Physics and Engineering, University of Applied Sciences, Hüfferstr. 27, D-48153 Münster; * Department of Anaesthesiology and Intensive Care Medicine, University of Münster, Münster, Germany
Introduction: By 13 C isotope spectroscopy, a noninvasive diagnostic tool for specific functions of different organs is available [1]. A 13 C labelled tracer is administered and is metabolised by a specific pathway to 13 CO 2 . The increase of exhaled 13 CO 2 is analysed and gives an indication of a specific function of the organ under test. The availability of sensitive (resolution 0.3‰) and stable isotope-selective infra-red (IR) spectrometers has replaced expensive massspectrometers. The present IR analysers rely on two conditions: 1) on co-operating patients who collect breath samples into containers and; 2) on equilibrated test gas samples.
Methods: Online monitoring of ventilated ICU patients over a long period requires an automated collection of the exhaled gases and an equilibration of the CO 2 content without compromising the lung ventilator set-up. The sample should only contain exhaled gases, a dilution would increase the necessary sensitivity of the analyser. A collection and mixing device has been developed that collects the exhaled gases and does not interfere with the safety and monitoring concept of the lung ventilator. The unit collects the sample gas from an Evita ventilator (Dräger) and supplies it to an Infra-Red Isotope-selective Spectrometer (IRIS, Wagner) which has been modified to analyse a continuous gas stream. The expiratory concentrations of 13 CO 2 and 12 CO 2 , the ratio of 13 CO 2 / 12 CO 2 as well as the cumulated hourly percentage of the recovery of the administered 13 C label are determined by evaluation IRIS software (V2.3 1998, Wagner) and stored on a computer.

Results:
The risk analysis and technical set-up have been reviewed and certified by the TÜV-Rheinland. The collection unit connecting the lung ventilator to the analyser has proven not to interfere with the functionality of the therapeutic device. The set-up was validated in volunteers and in ICU patients. Results of a first study determining the gastric emptying by isotope techniques will be discussed in parallel on this congress.

Conclusion:
The stable IR isotope spectroscopy can be applied to ventilated patients. This allows a non-invasive focus e.g. on liver enzyme activities and on other organ specific problems.

Departments of Anesthesiology and Surgery, Lund University Hospital, S-221 85 Lund, Sweden
Introduction: Septic states might induce regional variations in metabolism. The aim of the present study was to evaluate if regional metabolic differences could be identi-fied in a model of severe acute pancreatitis in the rat, with the use of a microdialysis technique.
Methods: Under full anesthesia, microdialysis probes were inserted in the parenchyma of the pancreas and liver and under the serosa of the small intestine. Microdialysate was collected every 10 min. After a baseline period of 60 min, acute pancreatitis was induced by intraductal injection of 0.20-0.29 ml 5% sodium taurodeoxycholate. The animals were studied for three h after induction of pancreatitis after which they were killed. Arterial blood samples were taken every 60 min. The microdialysis fluid and blood were analyzed for glucose and lactate. Six groups were studied with six animals in each group; Sham, Pancreatitis and four groups with treatment given 15 min after induction of acute pancreatitis (early treatment). Treatment administered included N-acetylcystein (NAC; 200 mg/kg i.v), a platelet-activating factor (PAF) antagonist (lexipafant, 5 mg/kg) and monoclonal antibodies against the adhesion molecules ICAM-1(0.2 mg) and PECAM-1 (0.2 mg).

Results:
The levels of lactate and glucose in pancreas and lactate in blood were higher (P<0.05) in the pancreatitis group compared with the sham group. The increases of glucose and lactate in the pancreas were higher than in the blood, demonstrating an earlier and stronger change in metabolism in the pancreas as compared with the rest of the body. This finding was higher (P<0.05) in the pancreatitis group than in the sham group. There were no clear differences between the various pancreatitis groups, with or without treatment. and no differences in between the treatment groups. Within the sham group, glucose showed no regional difference, but there was a significant increase in the pancreas in pancreatitis animals as compared to the liver and intestine. Regarding lactate, there were differences between the pancreas, liver and intestine in both groups with higher levels seen in the pancreas.
Conclusion: An increase in pancreatic lactate concentrations in pancreatitis animals was seen and this increase seemed to have its origin in the pancreas and not in the whole body. Similar changes were also found for glucose. Early treatment, as described earlier, had no obvious effect on these parameters. The results imply that microdialysis could be of potential future value in monitoring metabolic locoregional differences in critical illness.

National Hospital Tokyo Disaster Medical Center, Tokyo, Japan
Acute necrotizing pancreatitis is histologically studied focusing on the degree and extent of necrosis and apoptosis.

Materials and methods:
Evaluation was done by lightmicroscopic examination (Hematoxylin-Eosin stain) and electron-microscopic examination. In order to judge the existence of apoptosis, the TUNEL method was applied.
Only exocrine tissue was surveyed. The number of specimens is three, which were obtained by distal pancreatectomy in the early phase of acute necrotizing pancreatitis because they resisted intensive care. The normal control was one specimen in which 1) negative control and; 2) positive control (TACS-Nuclease-treated control) were used.

Results:
Pancreatic tissue was divided into three areas by light-and electron-microscopy before TUNEL: 1) micro-scopically normal tissue; 2) microscopically apparently necrotized area; 3) minimal or almost normal tissue by the light microscope, although electron-microscopically mild or moderate destructive change of the cytoplasm and nuclear structure was apparent. The TUNEL method revealed the existence of apoptosis in both the the necrotized area (2) and in the minimally destroyed area lightmicroscopically (3).

Discussion and conclusions:
From our study, the existence of apoptosis in acute necrotizing pancreatitis is thought to exist. Several important themes seem to be pointed out: 1) new classification of acute pancreatitis, including apoptosis may be necessary; 2) apoptosis will influence the prognosis or mortality of necrotizing pancreatitis, hitherto believed to depend on the ratio of necrosis; 3) the criteria of severity should be revised.

Servicio Medicina Intensiva, *Servicio de Bioquímica, H.U. Virgen Victoria, 29010-Málaga, Spain
Objective: The object of the study was to relate, in the first 72 h following hospital admission, biochemical markers of inflammatory response and prognostic scores to mortality in acute pancreatitis (AP).

Results:
Results are reported in the table. Four patients had a rapid weaning, were extubated, discharged from ICU and are still enjoying good health. Two patients, despite a good liver function recovery, died with MOF.

Conclusion:
The evaluation of cardiac preload is a useful tool to guide fluid replacement and vasoactive drug admin-

Conclusion:
The use of somatostatin 14 AA shortens the time of closure of the enterocutaneous and pancreatic fistulas and allows an important reduction of the hospital costs. The evaluation of the therapeutic response 48 h after the beginning of the treatment is a fundamental criterion for optimising the use of this drug.

Unidad de Reanimacion, Departamento de Anestesiologia, Reanimación y Terapia del Dolor, Hospital Clinico Universitario de Granada, Spain
Introduction: Six retrospective case-control studies showed that the only difference between those patients who underwent elective infrarenal aortic aneurysm repair (IAAR) developing mesenteric traction syndrome (MTS) in the operation room (OR) and multiorgan dysfunction syndrome (MODS) during their ICU stay, was the decrease in urine output before aortic cross-clamping. Two of them presented the socalled 'hepatohemorrhagic syndrome' (HHS) [1] characterised by the association of disseminated intravascular coagulation (DIC) and acute ischemic hepatitis (AIH).

Methods:
Data from six patients with MTS was obtained from the anaesthesia records (1995)(1996)(1997)(1998) and clinical evolution during their ICU stay.

Results:
After the placement of the intestinal traction device, all patients developed MTS (facial flushing, reduced mean arterial pressure (MAP) and systemic vascular resistance (SVR) with increased heart rate (HR) and Cardiac Index (CI). They were treated with volume replacement until they reached the previous MAP before cross-clamping. Diuresis showed an increase that remained elevated during the intervention in all the patients but two. These two developed DIC intra-operatively, with bleeding before unclamping and posterior MODS. We believe that the initial release of PGI2, as the main product triggering the MTS, is also responsible for the increase in diuresis, but dysregulation of the renal function by an unknown mechanism [2] in the context of the prostanoid disbalance, would explain the two cases of DIC and posterior MODS in the two patients which, after a little increase in urine output, suddenly without any explanation turned to fall [3].

Conclusions:
First, although it has been postulated that HHS can not be explained by the operative course before the release of the aortic cross-clamp, we found a close cause-effect relationship between the presence, or not, of MTS during the anaesthesia period, and the posterior course in which urine output fell. Second, we also think that this syndrome (at least the two HHS we have seen) would be included in the context of MODS. Third, in this setting, the kidneys are the first organs in dysfunction by unknown mechanism, and this is due to the impaired renal response to the disbalance of prostanoids.

Department of Anesthesiology Intensive care and Medicine, Shaare Zedek Medical Center, Jerusalem, Israel
Introduction: The concept of bacterial/toxin translocation from the gut as the source of sepsis is well recognized. Continuous monitoring of gut permeability in correlation with clinical signs of sepsis is poorly documented. We measured daily intestinal permeability of patients admitted to the ICU and correlated it with signs of sepsis. Conclusions: Increased intestinal permeability is associated with clinical signs of sepsis, thus substantiating the role of gut permeability as a major defect in the evolvement of sepsis and its sequelae.

SGPGIMS, Lucknow, 226 014, India
Purpose: Encephalopathy with raised intracranial pressure is the major cause of death in hepatic failure [1]. We investigated role of arterial hypocapnia, achieved by hyperventilation, in the management of hepatic encephalopathy.

Methods:
In a prospective study of twenty-two patients, with fulminant hepatic failure with grade IV encephalopathy receiving mechanical ventilation, two groups of twelve (group I) and ten (group II) patients had PaCO 2 in the range of 25-30 mmHg (group I) and 35-40 mmHg (group II). Tidal volume was kept in the range of 8-10 ml per kg and respiratory rate in the range of 12-20 per min.

Discussion and conclusions:
The reduction in procedure and ICU invasiveness throughout the years allowed for a reduction in morbidity and may confirm that appropriate medical care can overcome the adverse influences of many of the negative predictor factors suggested to influence OLT recipients outcome.
Design: Prospective epidemiological study.
Site: Multidisciplinary pediatric intensive care unit (PICU) in a tertiary care teaching hospital.
Methods: UGIB was considered to be present if hematemesis occurred or if blood was present in the gastric tube. UGIB was qualified as clinically significant if 2 out of 3 experts independently concluded that at least 1 of 6 complications (transfusion, decreased hemoglobin concentration, hypotension, surgery, multiple organ system failure or death) was attributable to it. The mean time from PICU admission to onset of UGIB or CS-UGIB was obtained by the summation of the number of events multiplied by duration of PICU stay for each event, then divided by the total number of events. The mean incidence density was calculated by dividing the total number of patients with UGIB or CS-UGIB by the cumulative number of patient-days.
Conclusion: UGIB and CS-UGIB occur soon after admission to PICU. Prophylaxis to prevent CS-UGIB should be given early to patients at risk of developing this complication.

Ankara University Medical Faculty, Anesthesiology and Reanimation Department, Sihhiye, Ankara, Turkey
One of the mechanisms causing BT is impaired host defense. Metoclopramide (M) restores the depressed immune function after hemorrhage [1]. To our knowledge, there is no study investigating the effect of M on BT.

Procedures:
We investigated the administration of M on BT in a dog model of ischemia-reperfusion (IR) injury induced by thoracic aortic cross-clamping and declamping. Twenty-two mongrel dogs were randomized into three groups: sham-operated group (n=7) (without cross-clamping), M group (n=8) and placebo group (n=7). Placebo and M group received placebo and M (0.15 mg/kg iv), respectively, before cross-clamping. During 45 min of ischemia and 30 min reperfusion arterial blood gases and hemodynamic data were continuously recorded. 72 h later dogs were re-operated on, and peritoneal swabs, blood samples and specimens from duodenum, jejunum, ileum, colon, mesenteric lymph node (MNL), heart, spleen, kidney, liver and lung were obtained for bacteriological analysis. Blood samples were also obtained before the first operation.

P145 Is redistribution of microcirculatory blood flow within the small intestinal wall the cause of prolonged paralytic ileus in the critically ill? L Hiltebrand, V Krejci, D Erni*, A Banic* and GH Sigurdsson
Autoregulation of microcirculatory blood flow (MBF) and redistribution of flow within the intestinal wall (from the muscularis to the mucosa) are known mechanisms to maintain O 2 delivery in the gut during hypovolemia. During sepsis, however, autoregulation of blood flow is impaired [1]. The aim of this study was to monitor the distribution of MBF within the intestinal wall in different parts of the gastrointestinal tract during the development of septic shock.

Materials and methods:
We measured systemic (CI), regional (mesenteric artery; SMA) and microcirculatory blood flow (MBF) in 11 sedated and ventilated pigs. MBF was measured with multichannel laser Doppler flowmetry in the gastric, jejunal, and colonic mucosa and the corresponding muscularis layers. Septic shock was induced by faecal peritonitis. After 240 min, i.v. fluids were administered to alter hypodynamic shock to hyperdynamic septic shock.
Results and discussion: During the first 240 min (hypodynamic shock) CI, SMA and MBF in the stomach mucosa decreased by 50% (Figs 1 and 2), while MBF in the jejunal and colonic mucosa remained virtually unchanged (Figs 3 and 4). In the muscularis of the jejunum and colon, on the other hand, MBF decreased significantly more than CI and SMA (Figs 2 and 4). Administration of i.v. fluids at 240 min resulted in a significant increase (above baseline) in CI and SMA (Fig. 1) as well as in MBF in the mucosa of the stomach, jejunum and colon. The 'overshooting' increase in MBF in the jejunal and colonic mucosa, indicates that there was some hypoperfusion present in the mucosa during the hypodynamic phase despite maintained blood flow. This is supported by a significant decrease in jejunal pHi (Fig. 5) during the same time period. There was virtually no increase in MBF in the jejunal and colonic muscularis after fluid administration indicating a prolonged, perhaps endothelin induced, hypoperfusion.

Conclusion:
It was concluded that MBF in the mucosa of the jejunum and colon remained unchanged, despite a 50% decrease in systemic and regional flows, suggesting a largely intact autoregulation during septic shock. MBF in the jejunal and colonic muscularis decreased significantly more than systemic and regional flows, suggesting that there is an active mechanism to maintain adequate blood flow to the mucosa during septic shock. Redistribution of blood flow from the muscularis to the mucosa causes severe hypoperfusion of the muscularis which contributes to intestinal atony frequently observed in critically ill patients. The tonometric determination of the arterialmucosal PCO 2 -gap (∆a-rPCO 2 ) is used to monitor adequacy of gastrointestinal perfusion. As PrCO 2 also integrates other phenomena [1], we analyzed the influence of the intestinal villus microcirculation on increased ∆a-rPCO 2 during long-term hyperdynamic porcine endotoxemia.

Material and methods:
Anesthetized and ventilated pigs received continuous i.v. endotoxin (ETX, n=12) for 24 h or placebo (Sham, n=6). Hydroxyethylstarch was infused to maintain MAP >65 mmHg together with a sustained increase in cardiac output [2]. Before the start of ETX (0 h), as well as 12 and 24 h afterwards, portal venous blood flow (Qpv, ultrasound flow probes) and lactate/pyruvateratios (L/P pv), ileal mucosal ∆a-rPCO 2 (fiberoptic sensor) and bowel wall capillary Hb-O 2 -saturation (Hb-O 2 cap, remission spectrophotometry) were assessed together with intravital video records of the ileal mucosal microcir- Objectives: Evaluation of the feasibility of a measurement device using an optical fiber sensor for continuous measurement of gastric PCO 2 .
Methods: The COMOCADOF system consists of a fiberoptical catheter and the measurement unit. A CO 2sensitive polymer layer is inserted into a probe head, which is fixed at the end of the optical fiber. The in vitro tests were performed in a glass test-tube containing 20 ml of sample solution, which was maintained at a temperature of 37°C. The sample solution (either physiological saline or a physiologically adjusted hydrochloric acid of pH 1.0) was flushed with 6 gas mixtures containing 0.0 to 15.0% CO 2 for defined time intervals and at a gas flow rate of 50 ml/min. The sampling rate was typically 0.8 s. The tests were performed over a period of up to 48 h. In addition, the influence of gastric juice, colored or turbid samples (such as e.g. enteral nutrition) on the stability of the optochemical sensor was tested.

Results:
The table shows the original minimum requirements (target at the beginning of the project) and the achievements obtained up to now with the optical sensor for the detection of gastric carbon dioxide.
No significant influence of gastric juice, physiologically adjusted hydrochloric acid of pH 1.0, colored or turbid samples (such as e.g. enteral nutrition) was seen.

Conclusion
The COMOCADOF System allows direct continuous measurement of CO 2 in gastric juice. The main targets of the project (response time, long-term stability and accuracy) were achieved. The system seems to perform better than the existing air tonometry method.
Comparative studies with an air tonometry system both in vitro and in vivo are in progress. Introduction: pHi is important in the evaluation of critically ill patients. When pHi data is needed quickly, a handy-type blood gas analyzer may be useful but it must be accurate.

Methods:
We infused normal saline into the balloon of a gastrointestinal tonometer (Instrumentarium Corp, Finland) located in the patient's stomach. After a 60 min equilibration time, the saline was withdrawn and the CO 2 (PrCO 2 ) measured by the four gas analyzers (the analysis of each sample of normal saline by the four analyzers were completed within 5 min). At the same time, we withdrew and analyzed arterial blood. Then, pHi was calculated from the Henderson-Hasselbalch equation, and the CO 2 gap (PrCO 2 -PaCO 2 ) was also calculated.

Discussion:
We can now use blood gas analyzers at the bedside and get results quickly. However, the value of pHi obtained differs according to the kind of analyzer used, a point that needs to be kept in mind. Our important finding is that the data obtained with the handy analyzer tested here differed quite considerably from those obtained using the other three. We conclude that although the new handy analyzer is convenient, it does not allow accurate calculations of pHi and CO 2 gap.
Introduction: It has been under discussion for many years, whether acid-base-equilibrium during hypothermia should be managed using the alpha-stat or the pH-stat concept. Several studies in cardiac patients undergoing hypothermic cardiopulmonary bypass (CPB) led to controversial results. Until now there are only few data available, on whether to adjust PiCO 2 and CO 2 -gap to temperature or not. Thus we calculated PiCO 2 and CO 2 -gap in cardiac surgery patients with and without temperature correction looking for differences in the prediction of postoperative complications.

Methods:
After IRB approval, we studied 69 patients undergoing elective aortocoronary bypass surgery, ASA class II-III. In addition to standard monitoring each patient received a pulmonary artery catheter and an air-filled nasogastric tonometry catheter connected to a Tonocap™. Documentation of gastric-intramucosal PCO 2 (PiCO 2 ) followed at 15-min-intervals, documentation of hemodynamics, arterial and mixed-venous blood gas analyses and lactate followed seven times until extubation (T1-T7). We used mild hypothermic CPB and the alpha-stat concept. PaCO 2 and PiCO 2 were adjusted to temperature by the following algo-rithm: PCO 2 t =PCO 2 ×10 0.0185×(t-37) [mmHg] (t=body temperature; PCO 2 t =temperature adjusted PCO 2 ).
Statistical analyses were done using the paired t-test, P<0.05 was regarded as significant. Introduction: Many studies reported that measurements of gastric PCO 2 with the air tonometer method were more precise than those obtained with saline tonometer. Gastric tonometers, however, were usually validated in vitro in a solution bubbled with CO 2 . In this study, we evaluated the accuracy and equilibration characteristics of PCO 2 in the balloon of a gastric air tonometer located in saline solution either with or without bubbling of CO 2 .

Methods:
A gastric tonometer was immersed in a 0.9% saline solution maintained at 37ºC, in which certified calibration gases at three different CO 2 concentrations (5%, 9%, and 12%) were bubbled. When solution was saturated with CO 2 gas, PCO 2 in the balloon was measured every 10 min for 30 min by an infra-red analyzer with or without CO 2 bubbling. Bias and precision were calculated from the measured and expected PCO 2 values.

Conclusion:
It has been demonstrated that quick equilibration of PCO 2 between inside and outside of the tonometer balloon needs the mixing of sample solution. This result strongly suggests that measurements of PCO 2 with the air tonometer would underestimate gastric PCO 2 when it is located in the gastric juice with little mixing.

Universitätsklinik f. Anästhesiologie, Universität, D-89070 Ulm, Germany
Introduction: In patients with acute lung injury (ALI), the prone position is a well-established method to improve gas exchange by reopening atelectasis and secret drainage. This approach might lead, however, to increased intra-abdominal pressure and thereby impaired gastric mucosal perfusion. Therefore, we studied the effect of the prone position on both intragastric pressure and gastric mucosal-arterial PCO 2 gradients in patients with ALI.
Patients/methods: So far, 10 patients with ALI (PaO 2 /FiO 2 <250 mmHg) and clinical indication for mechanical ventilation in the prone position were studied. In addition to ventilator settings, systemic hemodynamics and gas exchange, we semi-continuously measured the gastric mucosal PCO 2 (PCO 2 gm) (Tonocap) via a nasogastric tube as well as the intragastric pressure continuously via one lumen of the nasogastric tube using a common pressure transducer. After 60 min of stable conditions baseline measurements were made. Then the patient was turned to prone position and further measurements were obtained 60 and 120 min later. Table. Friedmann test/Student-Newman-Keuls: # P<0.05 vs baseline.

Conclusion:
Despite major individual variability, we could not find a significant overall change in gastric mucosal-arterial PCO 2 gradient after turning to prone position. Increased intragastric pressure (>5 mmHg), however, was associated with an increased gastric mucosal-arterial PCO 2 gradient despite stable systemic hemodynamics. Positioning induced variations of intraabdominal pressure, hence, may impair gastric mucosal energy balance possibly due to compromised regional blood flow. Background: The effects of intravenous live bacteria infusion on the time course of splanchnic oxygenation variables have not been adequately defined, as well as the effects of large volume crystalloid infusion on these variables.

Methods:
Twenty-seven anesthetized mongrel dogs (17.1±1.77 kg) were challenged by a 15 min intravenous infusion of live E. coli (6×10 9 CFU/kg) and followed for 90 min. The animals were then randomized in two groups over 60 min: CT (controls, n=13), no fluid infusion or LR (lactated Ringer's 32 ml/kg/h, n=14). Cardiac index (CI in l/min/m 2 ), mean arterial pressure (MAP in mmHg), mesenteric blood flow (MBF in ml/min, ultrasonic flowprobe), oxygen-derived variables, lactate levels (in mmol/l) and gastric PCO 2 (gastric tonometry in mmHg) were assessed throughout the 165 min experimental protocol. PCO 2 -gap was defined as the difference between gastric and arterial PCO 2 and lactate flux was defined by the standard formula: ([portal lactate levels-arterial lactate levels]×MBF).
Results: Data are presented as mean ±SE. E.coli infusion significantly reduced MAP, CI, MBF, and increased PCO 2gap in both groups. Fluid infusion increased CI, systemic DO 2 and stabilized PCO 2 -gap. Although MBF was similar in both groups, only controls presented gut lactate pro- duction. Table 1 shows the time course of MAP, CI, MBF, lactate flux and PCO 2 -gap of the two groups.

Conclusion:
Although large-volume crystalloid infusion fails to restore mesenteric and gastric mucosal perfusion, it seems to prevent gut lactate production in a canine model of experimental septic shock.
P154 Routine blood gas analysis does not provide information on regional metabolism of the stomach obtained by gastric tonometry in patients with left ventricular failure J Graf, E Karassimos and U Janssens

Medical Clinic I, RWTH Aachen, Pauwelsstr. 30, 52057 Aachen, Germany
Purpose of the study: To define the association between variables reflecting regional perfusion and metabolism of the stomach obtained by gastric air tonometry with routine blood gas analysis and global hemodynamics in patients with cardiac failure.

Methods:
In 50 mechanically-ventilated patients with severe left ventricular failure gastric intramucosal PCO 2 (PgCO 2 ) was assessed utilising the Tonocap. Simultaneously hemodynamics were measured and routine blood gas analysis was carried out. Independent variables were correlated and the discriminative power to predict a low pHi <7.32 was calculated using Receiver Operating Characteristics (ROC). Conclusion: Both PgCO 2 and CO 2 Gap are reliable predictors of low pHi, thought to reflect gastric mucosal hypoperfusion. None of the cardiorespiratory parameters were capable of predicting a pHi <7.32. Thus, in patients with severe left ventricular failure assessment of regional perfusion indices using gastric tonometry cannot be replaced by one of the tested routinely applied monitoring modalities. Introduction: The gut hypoperfusion could contribute to development of multiple organ dysfunction in septic patients. However, there is no definitive study correlating the time course of gastric mucosal PCO 2 and organ dysfunction.

Methods:
We have studied prospectively 27 adult severe septic patients consecutively admitted in two large mixed ICUs. Each patient had a nasogastric tonometer and a pulmonary artery catheter. Every 8 h, systemic hemodynamic and oxygen variables, plasma lactate levels and gastric mucosal pCO 2 (PgCO 2 ) were measured. All these variables were measured on inclusion day (Day 0) and on the 1st, 2nd and 3rd days. pCO 2 -gap was calculated as the difference between PgCO 2 and PaCO 2 . Daily, we measured the SOFA score to characterize organ dysfunction evolution. We used the median pCO 2 -gap and lactate values of each day and time course of these variables to correlate to organ dysfunction and outcome (mortality on day 28). Changes over time were analyzed using a Kruskal-Wallis test and the relative risk (RR) was calculated.

Results:
The median age was 55 years and median APACHE II score was 18.5. The overall mortality rate was 52%. Table 1 shows the RR and CI 95% of each variables on day 0, 1, 2 and 3. The best cutoff values of pCO 2 -gap, lactate values and SOFA score were 15 mmHg, 2 mEq/l and 11, respectively. On the 0, 1st, 2nd and 3rd days, the patients that had pCO 2 -gap values of more than 14 mmHg, either the SOFA score remained high or the patients died on the 10th day (P<0.05). In contrast, on the same days, lactate levels did not discriminate either organ dysfunction development or death on the 10th day.

Conclusions:
There was a good correlation between pCO 2gap and multiple organ dysfunction assessed by SOFA score. Lactate levels were not able to predict outcome.

P156
The relationship between data of gastric tonometry by determination of polymorphonuclear leukocytes (PMNs, i.e., myeloperoxidase activity) and gut mucosal hypoperfusion in elective cardiac surgical patients A Pietsch, T Uhlig, H Vogelsang and P Schmucker

Department of Anesthesiology, Medical University of Luebeck, Ratzeburger Allee 160, 23538 Luebeck, Germany
Introduction: The expression of polymorphonuclear leukocytes (PMNs) as a result of local and systemic inflammation in patients undergoing cardiac surgery by using the CPB is considered to be an important cofactor in tissue hypoxia and the pathogenesis of multiple organ failure [1]. Therefore, an early detection of tissue hypoxia is very useful for the prevention of MOFS. It has been shown that gastric tonometry improves clinical applicability of gut mucosal oxygenation monitoring [2]. In a previous paper the correlation between gastric tonometry and PMNs was warranted [3]. This was done in the following study.
Methods: Following institutional approval, 66 adult male patients undergoing elective cardiac surgery using cardiopulmonary bypass were studied. They were monitored with the Tonocap™ in addition to standard monitoring (ECG, CVP, arterial pressure) and the pulmonary artery catheter. The PiCO 2 was measured every 15 min after admission to the ICU. In addition, documentation of CO 2gap, arterial and pulmonary arterial hemodynamics, arterial and mixed-venous blood gas analyses and lactate followed four times until extubation. The patients were shifted into three groups by using the terciles of preopera-tive expression of the PMNs (group I <1451 µg/l; group II ≤2487 µg/l; group III >2487 µg/l).

Results:
The course of CO 2 -gap was parallel to the PMNs. Both parameters increased after admission to the ICU. Using the Wilcoxon-Wilcox test the increase in CO 2gap and PMNs was significant in all groups. These differences did not occur in the parameters obtained from blood gas analyses (BE, lactate, mixed venous saturation) and in hemodynamic parameters. Background: Gastric tonometry has been used to detect intramucosal acidosis. However, there remains controversies about the methodology and interpretation of the results. In recent years, sublingual PCO 2 (PSLCO 2 ) has been introduced as a noninvasive method to detect perfusion failure. Yet, there has not been enough evidence to prove that the PSLCO 2 is clinically of use.

Purposes:
To define the efficacy of PSLCO 2 .

Methods:
We investigated 6 patients who underwent open-heart surgery. PSLCO 2 were recorded continuously with a PCO 2 sensor based on an ion-sensitive field effect transistor (ISFET, NIHON KOHDEN). Arterial lactate was also measured.

Result:
A substantial increase of PSLCO 2 came into our notice from 377 mmHg, at the time of induction to 5712 mmHg, during the cardiopulmonary bypass (CPB). Gradual decrease of PSLCO 2 after the aorta declamping was observed down to 485 mmHg, 6 h after the operation. Significant increase in PSLCO 2 indicates that oxidization is not fairly sustained to the tissue during CPB despite enough pump flow. The increase in blood lactate concentration was followed parallel to PSLCO 2 while CPB was performed.
Conclusion: PSLCO 2 is an attractive parameter to be served for the swift detection of perfusion failure during open-heart surgery.

Intensive Care Unit and Department of Radiology, Hellenic Red Cross Hospital, 1 Erythrou Stavrou str. Athens 11526, Greece
Introduction: The aim of our study is to investigate the change of blood and muscle Carnitine levels in patients of the Intensive Care Unit and investigate how the change in Carnitine levels affects muscular mass.

Method:
In our double blind clinical study, forty-six ICU patients were randomly divided into two groups. Patients with either renal or liver insufficiency were excluded. Both groups of patients were on enteral or parenteral nutrition with 30 kcal/kg/day intake with 1 g/kg/day protein intake. The mean duration of hospitalization was 23.6±8 days for group A and 26.9±9 days for group B. All patients were under physical therapy for 60 min/day. Only group B patients received intravenous supplements containing 100 mg/kg/day L-Carnitine. Total and free Carnitine were measured from the blood of each patient at the time of admission and at the time of exit from the ICU. Muscle Carnitine levels were also studied with a triceps muscle biopsy at the same time. Muscle mass thickness was estimated by the same utrasonographer in all cases. A 10 MHz frequency, linear high-resolution transducer was held longitudinal to the biceps muscle in the plane of maximum thickness.

Conclusions:
In group A patients, blood and muscle Carnitine are decreased. Muscle mass is concomitantly decreased. The opposite results are obtained from group B (L-Carnitine supplementation). We conclude that L-Carnitine supplementation prevents muscle atrophy in critically ill ICU patients.

Surgical Clinic and Department of Anaesthesiology, Clinical Hospital Centre Rebro, Kispaticeva 12, 10000 Zagreb, Croatia
Objectives: Pseudocholinesterase (PChE) is an enzyme (EC 3.1.1.8) and like albumin is synthesised in the liver. Although the two proteins are synthesised in the liver they are not interdependent and both have been used independently as an index of liver function. Because of the contrary literature data, the aim of the present study was to measure PChE activity and albumin concentration in healthy infants and compare them with adult values [1,2]. We also compared PChE and albumin values between healthy infants and infants with biliary atresia and cirrhosis. PChE activity and albumin concentration were measured in 25 infants with biliary atresia and accompanying liver cirrhosis before Kasai operation, in 10 healthy infants, as well as in 10 healthy adults undergoing elective surgery.
Methods: PChE activity was determined by the spectrophotometric method of Ellman, using butyrylthiocholine as substrate. Plasma albumin concentration was measured by electrophoresis.

Results:
In infants with liver cirrhosis, PChE values were about 50% lower then in healthy infants of the same age i.e. 2-6 months (P<0.001). In healthy infants PChE activity was lower by 20% than in adults (P<0.001). Significantly lower albumin concentration in all infants than in adults has also been established (P<0.05).

Conclusion:
Our results showed diminished synthesis of PChE and albumin in liver cells of infants (2-6 months). These observations would suggest that infants might be expected to differ significantly from adults in the extent to which drugs might be protein bound and in the rates of hydrolysis of ester-type drugs. It is well known that PChE is involved in the hydrolysis of muscle relaxants succinyl-choline and mivacurium. According to our results, 50% lower PChE activity in infants with biliary cirrhosis may cause prolonged duration of mentioned drugs.

Department of Anaesthesia and Intensive Care, Prince of Wales Hospital, Shatin, HKSAR, China
Introduction: It had been shown that serum albumin measured within 48 h of ICU admission is as accurate as the APACHE II score in predicting hospital mortality. If this result could be confirmed, then albumin would be a cheaper and more convenient predictor of mortality than APACHE II.
Aim: To evaluate serum albumin as a predictor of hospital mortality in critically ill patients.

Methods:
All consecutive adult ICU admissions over an 18 month period (April 1997 to September 1998) were reviewed. The data were retrieved from a computerised database. Serum albumin concentrations between survivors and nonsurvivors were compared. Accuracy of outcome prediction using albumin was assessed by area under the curve (AUC) of the receiver operator characteristics (ROC) curve.

Results
: 1003 records were reviewed. There were 647 (64.5%) survivors and 356 (35.5%) nonsurvivors. The sur-vivors had a significantly higher serum albumin both on ICU admission and in the first 72 h after admission (P<0.001) (see Fig. 1). In both groups, the serum albumin fell markedly after ICU admission. The ROC curves for hospital mortality are shown in Fig. 2. The AUC of APACHE II (0.79) is significantly higher than that of admission albumin (0.63) and albumin at 24 h (0.64) (P<0.001).
We used regression analysis to combine APACHE II with admission albumin (AUC 0.80) and APACHE II with albumin at 24 h (AUC 0.79) to predict hospital mortality. The addition of albumin did not significantly improve the accuracy of outcome prediction over that of APACHE II alone.

Conclusion:
There is an association between serum albumin concentration and hospital mortality. However, as a single variable, serum albumin is a poor predictor of hospital outcome. Combining APACHE II with admission albumin or albumin at 24 h did not improve the accuracy of outcome prediction over that of APACHE II score alone.  Serum albumin concentration is frequently used as an indicator of nutritional response in patients receiving PN. The acute-phase response following injury and sepsis is characterised by increased hepatic synthesis of specific secreted proteins while production of albumin is decreased. We have investigated the effect of acute administration of PN on absolute and relative rates of hepatic albumin synthesis during sepsis in vivo. Two groups of male Wistar rats (mean weight 239 g) underwent caecal ligation and puncture (CLP), with a third group as unoperated controls allowed free access to chow and water (ad lib). Between 18 and 24 h later CLP survivors were infused by tail vein with 0.9% NaCl or a PN solution delivering 33% of calculated daily energy and protein requirements. Total hepatic protein synthesis rate (TPS), albumin synthesis rate (TAS) and relative albumin synthesis rate (RAS) were determined with a flooding dose of 3 H-phenylalanine and the use of anti-rat albumin antibody to isolate the protein for radioactive counting.
Results: (mean ±SEM) are displayed in the table, with statistical analysis by ANOVA.
We have demonstrated that whereas TPS is increased in sepsis, TAS falls. Provision of PN has no effect on TPS, but significantly further depresses RAS. Reduced hepatic albumin synthesis in sepsis is not reversed by substrate provision, hence changes in serum albumin concentration are unlikely to be a useful monitor of efficacy of PN. Introduction: In clinical practice, gastric emptying and the success of gastric feeding is assessed by measurement of gastric residuals. A new setting easily allows bedside measurements in ventilated ICU-patients by 13 C-technology to evaluate gastric transport.
Methods: 13 C-acetate added to tube feeding will be absorbed into the blood stream from the duodenum and oxidised to 13 CO 2 in the liver. The ratio of 13 CO 2 / 12 CO 2 in the expired air can be determined by infrared isotopeselective spectrometry (IRIS). Increasing expiratory concentrations of 13 CO 2 indicate transport of the applied tracer from the stomach to the duodenum. In 11 healthy volunteers breathing spontaneously via a tight CPAP mask on a Dräger Evita respirator system, 13 CO 2 enrichment in the expired air was measured for two hours after ingestion of 150 mg 13 C-acetate in 50 ml tube feeding. In nine ventilated ICU patients measurements were performed to determine expiratory 13 CO 2 / 12 CO 2 ratios after administration of the same amount of tracer and tube feeding via gastric tube. Gastric reflux and the total amount of tube feeding were recorded before measuring. http://ccforum.com/supplements/4/S1

Results:
The new setting was validated in volunteers. About 20 to 40 min after ingestion of the tracer, peak values (% of 13 C dose/h) were recorded. Within two hours the cumulative recovery was 20%. In the ICU patients only a small quantity of the applied tracer (<5%) was recovered independent of reflux amounts.

Conclusion:
We could not quantify gastric emptying by this new setting in ICU patients. Failure might be due to changes in transport mechanisms, resorption in the duodenum, metabolism and pool exchange compared to healthy volunteers. Methods: Ileal segments of adult guinea-pigs were mounted in silanized organ baths that contained oxygenated Tyrode solution (30 ml, 37°C). Prewarmed Tyrode solution was infused into the intestinal lumen, the infusion rate being 0.5 ml min -1 . The fluid passing the gut lumen was directed into a vertical outlet tubing which ended 4 cm (reflecting a pressure of 400 Pa) above the fluid level of the organ bath. This arrangement caused gradual filling of the intestine. When the intraluminal pressure reached a threshold an aborally moving wave of circular muscle contraction, measured as a spike-like increase in intraluminal pressure, propelled the intraluminal fluid to leave the system and thus caused emptying of the segment. The pressure threshold for eliciting peristaltic waves (peristaltic pressure threshold -PPT) was used to quantify the effects of drugs on peristaltic activity. Inhibition of peristalsis was reflected by an increase of PPT. After registration of normal peristalsis the segments were exposed to propofol (0.1-100 µM), which was administered cumulatively into the bath, i.e., to the serosal surface of at least 6 intestinal segments from 6 different guineapigs either alone, after application of vehicle (DMSO diluted with Tyrode solution), the GABA A receptor antagonist bicuculline (BIC, 10 µM), or the opioid receptor antagonist naloxone (NAL, 0.5 µM).
Results: Propofol (0.1-100 µM) concentration-dependently increased the PPT, however, never caused complete inhibition of intestinal propulsion in the doses tested. While BIC (10 µM) failed to prevent the inhibitory effect of propofol, the increase in PPT was absent after pretreatment with NAL (0.5 µM).
Conclusions: Propofol inhibits intestinal peristalsis by increasing the PPT, but is less potent and effective compared to barbiturates or clonidine since it did not abolish peristaltic activity. The inhibitory effect of propofol seems not to be mediated through binding to GABA A receptors, but seems to involve enteric opioidergic pathways.

Department. of Anaesthesiology, University of Wuerzburg, Josef-Schneider-Strasse 2, D-97080 Wuerzburg, Germany, and * Department of Exp. and Clin. Pharmacology, University of Graz, Austria
Introduction: The α-adrenoceptor agonist clonidine provides a good alternative for sedation in patients in intensive care units (ICUs). Inhibition of intestinal transport, the development of ileus, and subsequently other complications are major side effects of sedatives used in ICU patients. This study examines whether clonidine exerts an inhibitory effect on intestinal peristalsis.
Methods: Ileal segments of adult guinea-pigs were mounted in silanized organ baths that contained oxygenated Tyrode solution (30 ml, 37°C). Prewarmed Tyrode solution was infused into the intestinal lumen, the infusion rate being 0.5 ml min -1 . The fluid passing the gut lumen was directed into a vertical outlet tubing which ended 4 cm (reflecting a pressure of 400 Pa) above the fluid level of the organ bath. This arrangement caused gradual filling of the intestine. When the intraluminal pressure reached a threshold an aborally moving wave of circular muscle contraction, measured as a spike-like increase in intralumal pressure, propelled the intraluminal fluid to leave the system and thus caused emptying of the segment. The pressure threshold for eliciting peristaltic waves (peristaltic pressure threshold -PPT) was used to quantify the effects of drugs on peristaltic activity. Inhibition of peristalsis was reflected by an increase of PPT. After registration of normal peristalsis the segments were exposed to clonidine (1 nM-100 µM), which was administered cumulatively into the bath, i.e., to the serosal surface of 6 intestinal segments from 6 different guinea-pigs either alone, after application of Tyrode solution (vehicle), or of the adrenoceptor antagonist yohimbine (1 µM).
Results: Clonidine concentration-dependently increased the PPT. While Tyrode solution (vehicle) and 0.1-1 nM clonidine were without any effect on the PPT, 10 nM-10 µM clonidine caused an increase of PPT and complete abolition of peristalsis occurred after 100 µM clonidine in 4 of 4 segments tested. Pretreatment with yohimbine (1 µM) prevented clonidine (0.1-100 nM) from having any inhibitory effect on PPT.
Conclusions: Clonidine concentration-dependently inhibits the ileal peristaltic reflex in vitro through adrenoceptors located in the intestine. It is assumed that clonidine also affects propulsive peristalsis in ICU patients and thus might contribute to other complications such as ileus and multiple organ failure in those patients. Previous research has demonstrated a decrease in complication rates, morbidity, and mortality. Little data exist, so far, about the use in pancreas-kidney-transplant recipients with enteral drainage. Because of the enteral anastomosis, most surgeons restrain from feeding the patients enterally, and postpone the start of enteral nutrition until at least five days after the procedure. No data is published on early enteral immunonutrition in this type of patient. The aim of this preliminary study, is to analyze the feasibility and complication rates of early enteral immunonutrition in pancreas-kidney-transplant recipients with enteral drainage.

Patients and Methods:
Prospective analyzes of five patients after PKTx was performed. Nasogastric feeding with Impact ® (Novartis, Switzerland), 25 cc/h was started immediately after surgery. This was continued for the first five days after PKTx. Additionally, patients recieve a parenteral standard nutrition program and substitution of fluids as needed to maintain diuresis. Further medication included propulsive agents (metoclopramide 20 mgs i.v. every 6 h, cisapride 10 cc p.o. every 6 h) and immunosuppression (tacrolimus 0.05 mgs/kg or cyclosporine 4 mgs/kg, ATG 5 mgs/kg, prednisolone 20 mgs/day, mycophenolatmofetile 2 g/day). During the study for each patient basic demo-graphic and clinical characteristics including age, sex, SAPS II score, ICU and hospital mortality were recorded. The patients were screened for the occurrence of rejection, opportunistic infections and surgical complications.
Results: Early enteral immunonutrition was tolerated by all the patients. No surgical complications, especially no leakage at the enteric anastomosis, were observed. There was also no occurrence of transplant pancreatitis and a reduced incidence of bowel distention and vomiting. Except for one CMV-infection, no opportunistic infections or other infectious complications were observed. All patients developed rejection; rejection was mild and steroid-sensitive in four patients, one patient required OKT 3 treatment and tacrolimus rescue therapy.

Discussion and conclusion:
The number of five patients is to small to make a valid conclusion and the results are rather descriptive. They show that early enteral feeding in this group of patients is possible without an increase in postoperative complications. Furthermore, early regeneration of mucosal gut barrier function may decrease infectious complications and improve patient outcome. In order to confirm this hypothesis further research should be performed to investigate the impact of this strategy. Variables like infection rates, surgical complications, and rejection, as well as length of ICU stay, morbidity and mortality will be analyzed.

Medical Dpt 1, *Medical Dpt 2, Charles Uni Hospital, 304 60 Pilsen, Czech Republic
Objectives: When added into all-in-one (AIO) bags, total needs of exogenous insulin decrease compared to amounts given by perfusor. We studied the influence of carrier solution on insulin availability.

Materials and methods:
A 20 ml polyvinylchloride (PVC) syringe was filled either with saline or AIO solution aspirated from original 3 l bags (Nutrimix, Brown, Melsungen, Germany). Then 8 IU of insulin (Actrapid HM, Novo-Nordisk, DK) was added into syringes (calculated concentration 400 IU×l -1 ) and a PVC perfusion line (5 ml volume) was filled. Baseline samples (0.5ml) for immunoreactive insulin assay (IRI) were taken. Perfusor rate was set at 2 ml×hr -1 and samples taken directly from the hose at 5, 10, 30, 60, 90 and 210 min, placed on ice and stored at 4°C before analysis. Fifteen sets of measurements were done both for saline and AIO solution. Values are presented as means ±SD. MANOVA, ANOVA for repeated measures and paired T-test with Bonferroni correction were used when appropriate; P<0.05 was considered significant.
Results: Figure 1 shows IRI concentrations in both saline and AIO solutions during the experiment.
IRI differed significantly giving higher IRI yield in AIO solution compared to saline (MANOVA group by time effect P<0.001). Changes in IRI concentration depending on time was seen in saline only (ANOVA time effect P<0.001 for saline; P=0.26 for AIO). When separate time points for saline were analyzed major changes were seen at the very beginning of the experiment (a decrease from 206±34 at baseline to 152±33 IU*l -1 at 5 min; P<0.001). The mean yield of IRI expressed as ratio of measured to calculated concentration at baseline was 0.52 for saline and 0.70 for AIO.
Conclusion: Insulin bioavailability in AIO solution is better than in saline with respect to the yield and stability. Rationale: AIO has become an easy and practical way to deliver total parenteral nutrition (TPN) in patients on a regular ward. However no data are available yet concerning the use of AIO in critical ill patients. Here the SBS is still the gold standard.
Method: A prospective randomized trial was conducted in surgical patients who were expected to need TPN for more than five days after major surgery or trauma. Patients were randomized to receive for at least five days an isocaloric and isonitrogenic TPN delivered as AIO or as SBS. The TPN met the caloric requirements, which were determined by indirect calorimetry. Substrates delivered, numerous laboratory parameters, metabolic disturbances as a result of the nutrition, as well as expenses for equipment and manpower were documented daily. Forty patients are required by the protocol to complete the study.
Results: 37 patients have entered the study since 2/98. 3/37 refused to continue after the first day (drop out). 16 patients received AIO (86 days recorded) and 18 pat received SBS (101 days recorded). There was no difference in age, sex, diagnosis, risk factors between the treatment groups. Three patients died during the hospital course (18, 14 and 123 days after completing the study), all in the SBS group. Cause of death was in no case related to the TPN. The nutrition could be delivered without problems in both groups. Reduction of the TPN for hyperglycaemia was necessary only in one patient in each treatment group. Reduction of the TPN for hypertriglyceridaemia was necessary in one patient in each group. The SBS required more equipment e.g. infusion pumps and disposable infusion systems as well as more manpower e.g. to fix technical alarms.
Conclusion: AIB seems to be an useful alternative to SBS in critically ill and metabolically stable surgical patients.

Department of Anesthesiology and Intensive Care Medicine, Charité -Campus Virchow Clinic, Humboldt-University, 13344 Berlin, Germany
Background: Previous clinical studies led to the hypothesis that glutamine-enriched total parenteral nutrition (TPN) has beneficial effects on clinical course and on changes in the immune system of patients in postoperative catabolic state.
Patients and methods: 20 patients with esophageal carcinoma were enrolled in a controlled, prospective, randomized and double-blinded study after esophageal resection: 10 patients received glycyl glutamine-enriched (0.15 g glutamine/kg/d) TPN (Glamin ® , Baxter Inc.) from day 2 postoperatively, 10 patients served as control (TPN without glycyl-glutamine: Vamin ® , Fresenius Inc). Clinical scores and immunological parameters were observed before and during the first 8 days after surgery. Lymphocyte subpopulations were measured using flow cytometry analyses (FACS-Flow ® , Becton Dickinson), plasma cytokine levels were determined using standard ELISA technique.
Results: No significant differences were found, neither in clinical nor in immunological parameters: SAPS II scores as well as GORIS scores showed similar courses in both groups and ICU-days did not differ. Postoperative alterations of the immune system, measured by plasma IL-6 and procalcitonin levels as well as changes in lymphocyte counts and T4-T8-ratio, were nearly similar.

Conclusion:
A beneficial effect of glutamine-enriched TPN could not be found in our study population. Further investigations are necessary to find out if higher doses and/or other patient groups are able to demonstrate similar effects as shown in former studies.

Intensive Care Unit and Department of Radiology, Hellenic Red Cross Hospital, 1 Erythrou Stavrou str. Athens 11526, Greece
Introduction. In the present study we profit from the high U/S sensitivity and specificity in order to study the muscular mass (MM) and subcutaneous fat (SBF) changes during Intensive Care Unit hospitalization. In addition, we investigate the influence of the administration of muscle relaxants and corticosteroids.

Materials and methods:
Thirty patients (19 male aged 51.1±14.15 years hospitalized for 18.84±3.5 days and 11 female aged 64.36±12.4 years hospitalized for 19.18±4.6 days) entered the study. Clinical examination and computerized tomographic scan proved intracerebral hemorrhage in 14 of them and brain infarction in 16 of them. The patient's nutrition schedule included 30 Kcal/kg/day with protein uptake of 1 g/kg/day. Physiotherapy (passive movements of the limbs) was applied for 1 h/day in all patients. Muscle relaxation was applied to 9 patients with atracurium (0.6-0.7 mg/kg). The mean duration was 2.77±0.44. Dexamethasone (24 mg/d iv) was applied to 14 patients for a mean duration of 4 days. Two ultrasonographic measurements were done for each patient; one on the first and another on the last hospitalization day in the ICU. Upper arm biceps MM and SBF thickness were assessed by M-mode ultrasonography using a 10 MHz linear-array high resolution transducer. All scans were performed with the same ultrasonographer. MM and SBF thickness are expressed in cm. MM, and SBF changes were calculated as differences of the 2nd mea-surement minus the 1st and they were expressed as ∆MM[cm] and ∆SBF[cm], respectively.
Results: A significant loss of MM and SBF was observed in all patients included in the study. SBF loss (mean ∆SBF%=41.1±16.70) was significantly higher than corresponding MM reduction (∆MM%=21.4±4.90). A significant correlation was found between ∆MM% and the duration of hospitalization (19±4 days), (r=0.62, P<0.05).
On the contrary ∆SBF% was poorly correlated with the duration of the hospitalization (r=0.10, P>0.05). Neither ∆MM% nor ∆SBF were correlated with patients' age.
Patients who took corticosteroids (n=5) showed a nonsignificantly higher percentage of MM loss in comparison to those who did not take corticosteroid (n=15; corresponding ∆MM%: 22.1±5.22 versus 20.0±4.83%). Nevertheless, ∆MM% for the patients, who did not take muscle relaxation was slightly different (20.4±4.53) from ∆MM% of patients who did not take corticosteroids or corticosteroids and muscle relaxants. A significantly higher reduction of MM was observed in patients who took corticosteroids and muscle relaxants in comparison to those who did not take either corticosteroids or muscle relaxants (corresponding ∆MM%: 23.8±4.90 versus 19.8±4.47). SBF did not differ significantly between patients who took and did not take corticosteroids (corresponding ∆SBF%: 47.5±26.47 versus 40.6±18.91). Not significant was also the SBF difference between those who took corticos-teroids and muscle relaxants (∆SBF%: 35.9±17.39) in comparison to those who did not take either corticosteroids or muscle relaxants (∆SBF%: 42.0±12.69). ∆SBF% for patients who did not take muscle relaxation (43.2±15.95) did not differ significantly from corresponding ∆SBF values of the other formatted groups.

Conclusions:
Our study shows that MM loss is correlated with the duration of hospitalization as well as with the synchronous corticosteroids and muscle relaxants administration. On the contrary SBF, although decreased during the period of our study, is not statistically correlated with any of the studied factors. Cortisol deficiency due to hypothalamic or pituitary injury has been uncommonly reported after traumatic brain injury (TBI), and the incidence is unknown. Diabetes insipidus after TBI, which usually reflects posterior pituitary injury, is however frequently recognized.
Blood cortisol concentrations are usually high in ICU patients -often 2-3 times greater than the normal morning peak values of 500 nmol/l. Cortisol has a relatively short half-life and a permissive action on catecholamine efficacy, so cortisol deficiency could manifest rapidly, and cause hypotension.
In 80 patients with severe TBI (GCS <9), and in 12 controls (severe multi-trauma without TBI), we measured blood cortisol and ACTH concentrations on days 1,3 and 7 of ICU admission.  [1,2] are only 21-30% effective at correcting serum concentrations in severely hypophosphataemic critically ill patients, with 60-85% of patients redeveloping hypophosphataemia. We evaluated a two day regimen employing a rapid infusion rate.  Sodium phosphate (0.1 mmol/ml) was infused at 10 mmol/h. Serum phosphate and ionised calcium concentrations and calcium phosphate product were determined pre-infusion, immediately post-infusion and 6 h post-infusion for each dose and after 24 h. Twenty-two patients were receiving CRRT.

Method
Results: See Table 1.
Only seven calcium phosphate product values were slightly raised (range: 4.9-5.7 mmol 2 /l 2 ). Twenty-two patients had ionised calcium concentrations of ≤1.05 mmol/l on occasions during the regimen but the extent of the hypocalcaemia was mild (mean [SD] ionised Ca: 0.99 (0.07) range: 0.68-1.05 mmol/l [n=50]) and usually transient. Furthermore the degree of change in ionised calcium concentrations from pre-treatment values for all patients was not significant (see Table 1). No clinical evidence of hypocalcaemia was noted.

Conclusion:
The two day intravenous phosphate regimen described is safe and highly effective at correcting hypophosphataemia in critically ill patients.

EMS-PHU ZD Maribor, Ul. Talcev 9, 2000 Maribor, Slovenia
Introduction: The classic presentation of hypoglycemia includes tachycardia, anxiety, extreme hunger, tremor, palpitations, pallor, weakness and an altered level of consciousness. The hypoglycemic patient's presentation may also lead to suspect another condition such as cerebrovascular accident, seizure, head injury or acute psychosis.
A clinical study was done to determine whether there is a correlation between the manifestations of hypoglycemia and the patient's age.

Methods:
We studied 191 prehospital patients in the period from February 1998 to September 1999. For each patient we studied: the level of blood glucose, age and clinical manifestations of hypoglycemia.
We found out that there existed a statistically important difference in patient's age between clinical manifestations of hypoglycemia (Table).

Conclusion:
In our study we found out that psychosis as a manifestation of hypoglycemia presents in significantly younger patients than other manifestations (Student t-test <0.05). In elderly patients hypoglycemia presents more often with focal neurological deficit, seizure, confusion and coma. Our cases illustrate the importance of considering hypoglycemia in all patients who present with alterations in mental status even when the clinical findings seem to be explained initially by other etiologies. Emergency physicians must be aware of such a presentation of hypoglycemia and the need for rapid testing, since testing is easily performed and therapy is most often curative with a good outcome. Introduction: There has been a resurgence of interest in the interpretation of acid base abnormalities using Stewart's Strong Ion Theory [1], central to which is the solving of a complex quadratic equation that accurately quantifies the contribution of both organic and inorganic anions to pH. In accordance with this theory, we hypothesised that an increase in organic anions (lactate and unmeasured anions) during metabolic acidosis would cause a compensatory fall in inorganic chloride thus preserving electroneutrality. The aim of this study was to determine if this compensatory hypochloraemia, expressed as a fall in the Cl:Na ratio, could be a useful method to detect these elevated organic anions in patients with metabolic acidosis.

Patients and methods:
Paired blood samples were taken on admission and at 24 h from 341 patients (median weight 8kg, IQ range 3-15) admitted to a regional PICU, from which serum electrolytes, lactate and blood gases were measured and the strong ion gap calculated to quantify unmeasured anions (UMA). A metabolic acidosis was defined as being present if the base deficit (BD) was ≥5mmol/l and was chosen rather than pH since it better defines the metabolic component of an acidosis [2].Using the above criteria for acidosis, 136 samples were selected and further divided into two subgroups using  Background: The kidney-derived peptide urodilatin has been suggested to be important in the regulation of natriuresis and diuresis. This hormone has been used clinically as a drug for the treatment of incipient renal failure in cardiac surgery patients and is known to be increased in these patients if renal function is preserved [1]. However, the mechanisms involved in the regulation of the production of urodilatin are far from being clear. This study was designed to determine, if Angiotensin II may influence the renal excretion of urodilatin (V URO U).  (Table 1). RVR: renal vascular resistance; VU: urine flow; V Na U: urinary excretion of sodium; GFR: glomerular filtration rate; V URO U urinary excretion of urodilatin.

Materials and methods:
Conclusions: These observations suggest that V URO U is affected -directly or indirectly -by vasoconstrictory concentrations of Angiotensin II and may help to explain recent findings on the regulation of the renal excretion of urodilatin; especially the positive relationships of V URO U with plasma sodium [2] and renal perfusion pressure [3].

Slovak Institute of Cardiovascular Diseases, Pod Krasnou Horkou 1, Bratislava, Slovak Republic
There is evidence that a transplanted heart does not provide the recipient with functionally normal cardiac function and many differences in resting haemodynamics and alterations in electrophysiology and neurohumoral disbalance can be seen. Therefore, the plasma levels of functionally important metabolites -endothelin-1 (ET-1) as well as A-and B-type natriuretic peptides (ANP and BNP) -are released into the blood bed in response to fluid retention and vasoconstriction in heart disorders. Besides these haemodynamically-interesting parameters, in various intervals after the heart transplantation (HTx), the actual endogenous antioxidant capacity of patients was studied. Before HTx and successively three times during the first week and then weekly during the first month after HTx, seven transplanted patients were carefully clinically and biochemically investigated. In their blood plasma (or erythrocytes) the total antioxidant status (TAS), activities of superoxide dismutase (SOD) and glutathione peroxidase (GPx) as well as plasma levels of thiobarbituric acid reactive substances (TBARS), ET-1, ANP and BNP were estimated.
During the whole investigation period the gradual decrease of TAS was noted. Compared with the pretransplant levels, a gradual decrease of plasma levels of ET-1 and successive rise in ANP and BNP levels were observed, especially during the first week after the HTx. In conclusion, we have shown that HTx-patients demonstrated decreased total antioxidant capacity and characteristic movements of plasma levels of analysed peptides reflects the response of the cardiovascular system to the haemodynamically different situation.

P176 Management and outcome of symptomatic hyponatremia in emergency department J Kato and Y Tokuda
Okinawa Chubu Hospital, 208-3 Miyazato, Gushikawa city 902, Okinawa, Japan Introduction: Severe symptomatic hyponatremia was thought to be extremely dangerous. It is recommended that acute hyponatremia should be treated without delay and rapidly at a rate of at least 1 mmol/l/hour. Symptomatic hyponatremia was said to lead to death or permanent brain damage. However, some authors did not support quick correction and suggested that most deaths were caused by underlying diseases. We evaluated the clinical outcome of cases with symptomatic hyponatremia in terms of different management strategies in the emergency department.

Method:
In the emergency outpatient department of a community hospital, Okinawa, Japan, we retrospectively collected adult cases of symptomatic hyponatremia (serum sodium <130mmol/l) from April 1995 to October 1999. Etiology, treatment and clinical outcomes were evaluated. Neurological sequelae were assessed in all cases.
Result: Eighty-two cases of hyponatremia were determined (mean age, 54 years old, 48 males and 34 females). The mean sodium level was 118 mmol/l (the lowest 104 mmol/l). There were 45 cases with consciousness disturbance and 33 cases with seizure. Fifty-three cases were acute and 29 chronic. Underlying etiology included water intoxication, SIADH and hypoadrenalism. Death occurred in four cases and all were caused by underlying diseases. Thirty-five cases were treated with isotonic saline, 15 cases with hypertonic saline, and 32 cases with no active fluid treatment. Although correction rate for hyponatremia was not rapid (<0.6 mmol/l/h) for all cases, there were no cases of mortality and neurological sequelae from hyponatremia.
Conclusions: Slow correction of serum sodium was not complicated with permanent brain damage from acute symptomatic hyponatremia. Rapid correction is not neces-sary for treatment of symptomatic hyponatremia just because the serum concentration is extremely low.

References:
1 Intensive care unit, CH Saint-Philibert, LOMME 59462, France Severe hyponatremia induces neurological symptoms in connection with its rate of appearance. The aim of this study was to measure the variations of EAEP during hyponatremia and to determine its prognostic interest.
Patients and method: 65 patients (44 women and 21 men, mean age 67.3±15.7 years) with hyponatremia <125 mmol/l at their admission to the emergency room or intensive care unit, were included. The EAEP (monaural stimulation by alternative clicks of 0.1 ms at 90 Db) were recorded daily until hyponatremia correction was achieved (250 records). EAEP parameters (I, III and V peaks latency, I-III, III-V and I-V intervals) were studied according to the level and the mechanism of hyponatremia, and according to the neurological symptomatology and survival of the patients.

Results:
The mean value of the initial natremia was 118±7 mmol/l, by dilution in 78% of cases, with neurological signs in 69%. Correction speed was 4.8±2.2 days. Fifteen patients died but no cases of centropontine myelinolysis occurred. There was a linear correlation between natremia decrease and brainstem conduction lengthening (III and V peaks, intervals, P<10 -4 ). These variations were corrected by the normalization of natremia (P<10 -5 ).
Hyponatremia was associated with a lengthening of the I-V interval in 74.5% of patients. The mechanism of the hyponatremia, the presence of neurological signs or the patient outcome had no effect on EAEP parameters.

Conclusion:
Severe hyponatremia must be considered a cause of EAEP lengthening. EAEP evaluation does not constitute a prognostic factor of hyponatremia.

Department of Paediatric Medicine and Department of Neurosurgery, KK Women's and Children's Hospital, 100 Bukit Timah Rd, Singapore 229899
Introduction: Sodium disturbances and polyuria in children after craniotomy for intracranial lesions are not uncommon. Diabetes insipidus (DI) of central origin is often cited as a cause of high serum sodium [1] while cerebral salt wasting (CSW) is a recognised cause of hyponatraemia in these patients [2]. DI and CSW can both result in polyuria; patients with central DI suffer from free water loss while patients with CSW have severe salt loss. Both syndromes may occur at various times in the postoperative period. A combination of the above mentioned diseases could occur in the clinical setting and this can lead to difficulty in diagnosis and cause problems in maintaining normal fluid and sodium status. Proper differentiation is essential, as the treatment for each entity is different. CSW and DI in the same clinical setting have not been previously described in the paediatric age group. We report two cases of combined DI and CSW in the immediate postoperative period.
Case reports: Case 1. CXH is a 10 year old girl who presented with 1 month's history of visual and gait disturbances.
MRI of the brain showed a large suprasellar tumour with hydrocephalus. A ventriculo-peritoneal shunt was inserted immediately. She needed two operations to completely excise the tumour. Histology showed pilocytic astrocytoma. Postoperatively, she developed evidence of central diabetes insipidus and required intravenous pitressin before urine output was successfully controlled. She remained stable till the fifth postoperative day (POD) when she developed hyponatraemia that was very resistant to treatment despite hypertonic saline replacements. She was also clinically dehydrated. Urine sodium was 296 mEql/l/l and FeNa was 10.8. With aggressive saline replacements, hyponatraemia was corrected and kept within normal limits on a regimen of enteral feeds and intravenous saline replacements according to urine output and sodium measurements. Fludrocortisone was started. After tenth POD, urinary sodium began to decline consistently below 150 mEql/l by the fifteenth POD and less than 70 mEql/l by a month POD.
Case 2. NWS presented at 1.5 months of age with rhinorrhea and nasal mass. MRI showed a nasal mass with intracranial extension. Consent for surgery was not given until a year and a half later. By that time, a repeat MRI showed the tumour had extended into the left orbit and ethmoid sinus. A lateral rhinotomy, craniotomy with tumour excision and craniofacial reconstruction was performed. Histology revealed a low-grade nasal glioma. Postoperatively, he developed central diabetes insipidus requiring intravenous pitressin before urine output was controlled. However, serum sodium continued to drop even though urine output did not change significantly. On the second POD, the child developed a generalised tonic clonic seizure that was aborted with intravenous Valium and dilantin. Computer tomography scan of the head showed residual tumour in the left orbit and suprasellar region, evidence of CSF leak but no haemorrhage or cerebral oedema. Central venous pressure had dropped to +3cmH 2 O. Serum and urine sodium was 117 mEql/l and 176 mEql/l. FeNa was 9.6. Hyponatraemia was corrected gradually with normal saline and hypertonic saline replacements. However, urinary sodium levels continued to rise over the ensuing days, reaching a peak of 295 mEql/l on the sixth POD and there was massive solute diuresis. He required aggressive replacements with hypertonic saline to keep serum sodium within limits. However, urinary sodium persisted between 120-135 mEql/l thereafter and on the tenth POD, fludrocortisone was started at 10 µg/kg/day. Subsequently, urinary sodium dropped to 56 mEql/l two weeks after surgery and remained below 30 mEql/l after three weeks post surgery. Intravenous pitressin was successfully weaned off and the patient started on intranasal DDAVP.
Discussion: CSW syndrome is characterised by hyponatraemia (<130 mEq/l), dehydration, and inappropriate urinary sodium loss that responds to fluid and saline replacements. Osmotic diuresis often accompanies this syndrome. The main feature of central DI is massive diuresis of dilute urine with low sodium content thus resulting in hypernatraemia and dehydration. Both clinical states are characterised by diuresis but in CSW, hyponatraemia is present while in DI, the patient experience hypernatraemia.
In the neurosurgical intensive care setting, ensuring sodium homeostasis is important to maintain an environment without major flux in osmolality. CSW may develop in patients with an established diagnosis of DI. Early recognition can be achieved by measuring osmolality and sodium in the plasma and urine as well as plasma ANP levels. The cornerstone of CSW management remains replacing sodium and water loss with normal or hypertonic saline. In our experience, mineralcorticoid therapy appears to be an effective adjunct [3]. The aim of this study was to determine a role of CNP and NO for cerebral vasospasm following subarachnoid hemorrhage (SAH). Both of them are endothelium-derived relaxing factor (EDRF), and CNP exists in the brain as a cerebral vasodilator. We conducted a one week monitoring of CNP and NOx levels in the plasma and cerebrospinal fluid (CSF) in 26 patients who received clipping operation within 24 h after the occurrences of SAH, and classified to the Group A (angiographical spasm +), and Group B (angiographical spasm -). And we examined CNP and NO‚ levels in the CSF of the patients who received spinal anesthesia for a small operation as a reference patients.
Results: CNP levels in the CSF on day 1 of hospitalization was higher than that of the reference patients and the level decreased significantly on day 7, but we did not observe any significant difference between the groups. CNP levels in the plasma did not change.
NOx levels in the CSF in the week following SAH were higher than the levels of the reference patients, and the level in Group A was higher than in Group B on day 1. Only in Group A did NOx levels decrease significantly from day 3. NOx levels in the plasma did not change significantly.
Conclusions: CNP levels in the CSF were high in the acute phase following SAH, while plasma CNP levels were maintained at a constant. But any specific role of CNP for cerebral vasospasm was not indicated from our findings.
We considered that NOx in the CSF has a role as an inhibitor for cerebral vasospasm following SAH, because NOx levels decreased significantly only in the vasospasm group, while plasma NOx levels did not change significantly. Histopathology: all skeletal muscle biopsies showed pathologic alterations. Fiber diameters varied from atrophy to normal. In muscle fibers, internal nuclei, fiber splitting and focal regressive changes with increased activity of lysosomal acid phosphatase were found. In 50% of cases, regenerating fibers were present. Endomysial fibrosis was common. All the changes were classified as myogenic, no convincing signs of neurogenic atrophy were noticed. One patient revealed disuse atrophy.

Conclusions:
We found electrophysiologic signs of axonal sensory-motor polyneuropathy in all our patients. In spite of that, no neurogenic changes in the muscle were found in the biopsies, the changes in skeletal muscles had a myopathic pattern. These data suggest that in acute quadruplegia of the critically ill, both the nerve and the muscle are affected simultaneously.

Unidad de Medicina Critica, Departamento de medicina, Hospital Medico Quirurgico (ISSS), Alameda Juan Pablo II, San Salvador, El Salvador
Introduction: Critical illness polyneuropathy (CIP) has been associated with sepsis and multiorgan dysfunction syndrome (MODS). Studies have reported an incidence of CIP that oscillated among 21-81.8% [1,2,3].We conducted this study to assess the incidence of CIP in our intensive care unit.
Method: During the period of 05/01/1999 to 10/30/1999 359 patients entered our unit, from this group 154 needed mechanical ventilation for more then five days. The APACHE II score and GORIS score of multiorgan failure was evaluated. When they were transferred from the unit EMG study was realized according to the protocol published before [1].
Patients with a history of polyneuropathy, chronic renal failure, diabetes mellitus, chronic alcoholism and thus patients with polytraumatism of lower extremity, in which it was not possible to realize the study, were excluded.
Results: See Table. Our results show a smaller incidence (6.6%) of CIP than other studies [1,2].These studies had a major proportion of patients with neurotrauma and polytrauma and our sample was only 13%. Nevertheless, we can observe a score of FMO larger than or equal to 5, and a more extended mechanical ventilation tim,e which can also explain the difference that was found. Methods: Over a period of 3 weeks, all neurosurgical patients (elective and emergency) admitted to a regional ICU had blood taken on the 1st and 3rd morning after admission. Blood was also taken from controls. Neutrophil (PMN) respiratory burst activity was assayed over 17.3 min using a BioOrbit 1251 Luminometer to detect diluted (1:20) whole blood chemiluminescence (CL) in the presence of luminol. Circulating and maximal CL were measured by stimulation with phorbol 1,2-myristate 1,3-acetate in the absence or presence of tumour necrosis factor alpha (TNF) respectively (Fig. 1). The peak signal for each was obtained and the maximal value dived by the circulating value. This ratio represents the capacity to increase circulating respiratory burst activity. Soluble p55 TNF receptor (anti-inflammatory marker) and interleukin 6 (IL6, pro-inflammatory marker) were measured. Clinical parameters were recorded.    Methods: Twenty-three patients (age 36.09±17.4 years) with severe traumatic brain injury (TBI) (n=12) or vasospasm after subarachnoid haemorrhage (SAH) (n=11) were included prospectively. Patients were considered for lumbar CSF drainage if they suffered persistent intracranial hypertension (>25 mmHg) for a period of more than 15 min and failed to respond to high intensity treatment. Lumbar CSF drainage was not performed in patients with tight basal cisterns. After institution of the lumbar drain, 5-20 ml CSF were initially aspirated and then continuous CSF drainage was maintained. ICP and CPP before and after bolus-aspiration were documented. The neurological outcome of the patients was scored according to the Glasgow Outcome Scale (GOS) 6 months after injury.

Conclusions:
Controlled lumbar CSF drainage reduces therapy-resistant intracranial hypertension significantly. The hazard of transtentorial or tonsillar herniation might be limited by considering lumbar drainage only in the presence of discernible basilar cisterns.

P186 Central neuroaxial blockade improves case-mix adjusted mortality of the critically ill surgical patient AJ Prabhu, KR Burchett and MC Blunt
Queen Elizabeth Hospital, Gayton Road, King's Lynn, Norfolk, PE30 4ET, UK Introduction: Central neuroaxial blockade is known to provide good postoperative analgesia, hastens the return of gut function and may attenuate the 'stress response' to surgery [1]. However, there may be haemodynamic instability in critically ill patients. Despite improvements in many physiological parameters, there is little published evidence of improvement in outcome following surgery with such a block. Comparison of mortality outcome in intensive care units must take into account differences in the case-mix of patients in order to be meaningful. We investigated the effect of neuroaxial blockade on mortality standardised for case-mix using the APACHE 2 prognostic indicator (SMR) [2].
Methods: 205 adult patients admitted to the ICU for >8 h following major abdominal surgery were categorised according to whether they had received neuroaxial blockade (block: 91; no-block 113) in a retrospective contemporaneous cohort-controlled design. APACHE 2 scores, calculated risk of death, types of surgery, ICU lengths of stay and hospital length of stay were collected from the ICU database, and SMRs were calculated for each group. Demographic data was assessed using ANOVA and SMR by Poisson distribution.
Results: There was a significant improvement in SMR in the block group (P<0.01). The APACHE 2 scores and calculated risks of death were significantly lower in this group. Both hospital and ICU length of stay were significantly shorter in the block group. There were significant differences in the type of surgery performed between the two groups (Table).

Conclusions:
Within the methodological restrictions of the non-randomised design this study demonstrated an improvement in outcome in critically ill patients who receive neuroaxial blockade. The use of SMR as the primary end-point ensures a meaningful comparison of the groups despite the obvious difference in severity of illness between the groups as it allows outcome comparison for critically ill patients against a previously defined standard. Background: Measurements and assessment of current perception threshold with Neurometer in patients with chronic pain is well established. However, pathogenesis of acute pain in severely injured patients is not fully understood.
Purposes: To measure current perception threshold in severely injured patients with and without intravenous fentanyl and to determine if pain threshold is increased with sufficient analgesia.

Patients and methods:
Twenty severely injured patients with normal mental status were studied. Patients were given bolus injection of fentanyl (1 g/kg) upon arrival at the emergency room. After resuscitation and surgery, patients were transferred to ICU and given continuous intravenous fentanyl started at 50 g/h and increased up to 150 g/h on patient's demand. Pain threshold (PT) was measured with a Neurometer. VAS (visual analogue scale) and VRS (verbal rating scale) were also measured before and after bolus injection of fentanyl and during ICU care (every hospital day until seventh).
Findings: Before bolus injection of fentanyl, PT of the patients was significantly lower than that of control (106 mA vs 221 mA; P<0.01). After continuous intravenous fentanyl in ICU, VAS decreased significantly. However, no change was observed in PT of the patients during continuous analgesia with fentanyl.

Conclusions:
Severely injured patients were more sensitive to nociceptive stimuli than normal healthy controls. Fentanyl is an effective analgesia in trauma patients and decreased VAS significantly in the patients. But pain threshold itself did not show any change even with sufficient analgesia.

Ambroise Pare University Hospital, Mons, Belgium
Objectives: Pain is often the main complaint of patients coming into casualty wards but its management is rarely initiated early. Pain management must be considered as an aspect of the general management of the patient coming into the emergency ward [1,3]. It is consequently essential to relieve the pain quickly while continuing the diagnosis [2]. The treatment of pain is a progressive process and there should be no hesitation in combining different drugs with a maximum analgesic benefit while controlling the analgesia in order to avoid secondary effects. The following protocol will allow rapid and reassuring analgesia.
Method: Upon arrival in the ward, the adult patient is dealt with by a nurse who evaluates the analgesic need using the Visual Analogue Scale (VAS). After the consent of the doctor by signing a standard protocol, proparacetamol (2 g IV) and diclofenac (75 mg IV) are administered every 6 and 12 h respectively, irrespective of the VAS level. If the VAS is greater than 3 in the half hour following the initial administration of proparacetamol and diclofenac, the analgesia is completed by subcutaneous morphine (the dose is defined on the written protocol in relation to the weight and age of the patient). This administration must be repeated with an interval of 4 h if the VAS is greater than 3. The contraindications for the different drugs must be respected: chronic renal insufficiency, arterial hypertension with heart failure and gastric ulcers for diclofenac; hepatic insufficiency, atopy, nasal polyps, asthma and eczema for proparacetamol; chronic respiratory insufficiency and drug addiction for morphine.

Results:
The cases collected (n=200) show a demographic equivalence (45% women, 55% men) and an average age of 49 years. The distribution of the cases was mainly orthopedic (72%), with renal colic representing 9.6%. Other less common indications made up the remaining 18.4% (neuralgia, arthritis, knife wounds, gout, colitis). The Table shows the VAS over time, from admission up to 28 h (for 63 patients). ). But these scores can be very subjective, dependent on a observer and can show quite great interindividual differences. The Bispectral index (BIS) is a processed EEG parameter and shows continuously the level of inhibition of brain function during application of drugs for anesthesia and sedation.
We have tried to find if there is any relation between the value of the BIS index and the Ramsay or Cook score. In the case of a positive result, there can be a chance of exact titration of drugs and possible economical profit.

Methods:
The group of 20 patients hospitalized in our ICU (October 1998-September 1999). Including criteria-GCS 15 and the need of continual sedation or analgosedation.
Monitoring of all patients by the monitor 'Aspect A-1000 TM'. The values of the BIS index were recorded to protocol at any time of change in the level of sedation. At the same time this level of sedation was evaluated by two physicians (to prevent interindividual differences) by the use of Ramsay and Cook scores. All 75 examinations were recorded to study protocol. Results were assessed by help of means (median) + SD (range) and graphs.

Results: See Figs 1 and 2.
Conclusion: Figures 1 and 2 demonstrate that a narrow span of BIS index values includes a very wide scale of a depth of sedation and analgosedation assessed by Ramsay and Cook scores. These results and especially our experiences with continual monitoring show that to rely only on BIS index or to titrate drugs according to these values is impossible. We admit that bigger number of patients is needed for better evaluation.   Introduction: Bispectral Index (BIS) is an EEG parameter which gives a continuous reading between 0-100 depending on depth of anaesthesia. It has been used in ICU as a monitor of sedation [1]. Clinical assessment of sedation requires the patient to be stimulated, which may cause elevation of the BIS. We proposed that this change in BIS in response to stimulation is important when assessing depth of sedation. The aim of this study was to assess the feasibility of BIS as a monitor of sedation in ICU and to quantify within different sedation groups the changes in BIS score before and after stimulation.
Method: ICU patients being ventilated and sedated were studied. BIS was recorded on an Aspect A2000 monitor. A nurse blinded to the BIS reading assessed sedation using the observer's assessment of alertness/sedation score (OAAS 1-5: 1=no response, 5=fully awake). Sedation was assessed hourly, average unstimulated BIS (BISu) measured before clinical assessment of sedation was compared to average stimulated BIS (BISs). Spearman's correlation coefficient was used to assess association between BIS and OAAS, Mann-Whitney U was used to compare differences between BIS in OAAS groups 4-5 (lightly sedated), OAAS groups 1-3 (deeply sedated)   (2), pelvic fracture (2), lung contusion (1), haemodynamic instability (1), sepsis (1), in addition to crush injuries in the lower extremities (14) and upper extremities (6). Decompressive fasciatomies had been performed for all patients with crush injuries before acceptance to our ICU. Patients were treated for acute renal failure (14); DIC (13); sepsis (13) [originating from wounds (11), urinary system (9) and respiratory system (7)], respiratory failure (10), hepatic dysfunction (7) and cardiac dysfunction (3)  Eventually two patients with MOF died and 16 patients were discharged to other clinics with some sequels like neuropathies (10) and cardiomyopathy (1). One patient was discharged for hyperbaric oxygen therapy because of deep soft tissue infection unresponsive to medical therapy.

National Hospital Tokyo Disaster Medical Center, Tokyo, Japan
A criticality incident occurred September 30th, in Tokaimura in Japan. Neutron radiation was emitted in a widespread area. In addition to three patients, who were exposed to high dose neutron radiation, people who were living around the incident plant (within 350 m) were urged to take shelter according to an evacuation order, and people who were living within 10 km‚ were also recommended to remain within their houses for three or four days.
We had an opportunity to examine and give advice to the people who feared the radiation exposure. The experience is reported.

Results:
The criticality incident continued for nearly one day. From the next day after the end of criticality was declared, we started health examination. During the consecutive three days, a total of 1852 people were examined. The complete blood count, including the hemogram or lymphocyte counts, and blood and urine chemistry were evaluated in all the patients. The survey of radiation contamination was done when patients requested. The history of the health condition and estimation of the radiated amount was evaluated by our members (physicians) in all the patients. Although many people complained of the fear of the after effects of radiation, the degree of fear was extremely strong in children and pregnant women. Regarding the results of blood and urine examination, there were no patients who had abnormalities that were thought to be caused by high amounts of radiation exposure, although there were several patients who had decreased lymphocyte counts.

Discussion and conclusions:
Consultation only seems to be justified because, so far as we are informed officially (from the Japanese Agency of Science and Technology), the amount of radiation exposure was within the safe range. However, several patients who had decreased lymphocyte counts are recommended to receive repeated examination of the blood counts. Mental care and long-term follow-up may be necessary for people who are fearing the exposure and for children/pregnant woman.

P197 Establishment of emergency medical services in Addis Ababa, Ethiopia
O Benin-Goren*, P Halpern*, S Amir † and MB Tesfay ‡

*ED Tel Aviv Sourasky Medical Center Israel; † Sapir Medical Center Israel; ‡ Armed Forces General Hospital, Addis Ababa Ethiopia
In February 1998, an Israeli Medical Delegation that included two ED Directors and one Deputy Head Nurse of the ED, was founded in order to introduce the emergency medicine (EM) profession and to help in establishing an EM system in Addis Ababa. The program was produced in coordination with an Ethiopian physician who visited the ED in his fellowship and was very impressed with the Israeli system.
With the approval of the International Relation of the Foreign Bureau of the State of Israel, the team traveled to Ethiopia to consider the possibility that the program could be changed according to the needs in the field.
The program consisted of several issues in EM: administrative concepts of ED building; organization of the ED; the role of Emergency Nurses in the team; trauma and the treatment of critical care; study and practices of Basic Life Support; advanced Cardiac Life Support; and Trauma Life Support.
The participants on the program included twenty physicians and ten nurses who work in emergency services. The first meeting was dedicated to study the existing system in Addis Ababa. With poor conditions, lack of necessary equipment, medication and resources, and with lack of basic knowledge, the health system in Addis Ababa cannot start re-organization without help.
Yet there is a new movement of physicians who are ready to take responsibility and be part of the change. The Israeli team gave the recommendation to the authorities in Israel to continue education among the Ethiopian staff-nurses, physicians, technicians, and Paramedics.
This paper examines EM in Addis Ababa, the course, and recommendation for the future. Furthermore, in addition to the program that has been developed until now, the Israeli team with the cooperation of some American physicians plans to continue the program in December 1999 that will include studies and meetings with policy makers in Addis Ababa. Methods: Data of 31 children with severe traumatic brain injury (STBI), admitted in our ICU, were reviewed retrospectively. Cases were analysed according to the following criteria: 1) age in years, 2) GCS (upon admission), 3) CT-scan grade, 4) injury severity score (ISS), 5) paediatric trauma score (PTS), 6) paediatric risk of mortality (PRISM III-24h), 7) length of ICU stay in days and 8) outcome at ICU discharge (Glasgow outcome score -GOS). The demographic data of the studied group are shown in the Introduction: The Kingdom of Saudi Arabia has a national organ procurement organization (SCOT) with rigid criteria for the definition of brain death and a standard method of approaching the families of potential organ donors. As one of the major ICUs in Riyadh, we have reviewed our experience with organ donation.

Methods:
Retrospective review of all confirmed brain deaths in the 14 bed Adult ICU during the two-year period from January 1997 to December 1998.

Results:
Out of 210 deaths in the ICU, there were 40 (19%) confirmed brain deaths using the SCOT criteria. Mean age was 21.7±12.1 years with male:female ratio of 9:1 reflecting the fact that 24 (60%) of the cases resulted from road traffic accidents. Other trauma accounted for 4 (10%), cerebro-vascular accidents for 7 (18%) and brain anoxia for 5 (20%). Thirty-one (78%) patients were Saudi nationals and 9 (22%) were non-Saudi. SCOT approached the relatives of all patients that were confirmed as brain dead and was successful in obtaining consent for organ donation from 14 (35%) of these cases. Eight of these were Saudi and 6 were non-Saudi. Twentysix kidneys, 10 livers and 22 corneas were successfully transplanted and 13 hearts were used for valves.
Conclusion: 1) The organ retrieval rate of 35% compares favorably with other national and international data. This rate is probably related to the active involvement and coordination by SCOT. 2) Our incidence of brain death (19%) is above average and possibly related to the high incidence of road traffic accidents due to the unique traffic conditions with extreme heat and lack of seat belt use. 3) Family refusal for organ donation remains the main reason for failure to donate and greater effort is still required to increase public awareness in the Kingdom about the concept of brain death and the importance of organ donation. in the evening hours and early morning (Fig. 1). TACHY were treated electrically (n=19), pharmacologically (n=139) or combined (n=77). The antiarrhythmic drugs most frequently used were amiodarone (n=129), diltiazem (n=47), ibutilide (n=21), lidocaine (n=24) and digitalis (n=18). Proarrhythmia occurred due to haloperidol (n=4), cisapride (n=2), ibutilide (n=1), and amiodarone (n=4). Sedoanalge-   [1,2]. The objective of this prospective study was to evaluate the clinical efficacy of biphasic waveforms for cardioversion of atrial fibrillation.

Methods:
The pulse is an asymmetric quasi-sinusoidal biphasic waveform. The peak current of the second phase was approximately half that of the first phase. Transthoracic cardioversion (29 emergent, 71 urgent and 41 elective) were performed in 141 patients who were receiving antiarrhythmic drugs (e.g., amiodarone). Ischemic heart disease was the most common (about 90%) etiology. Shocks were delivered through 11.5 cm paddles in the antero-apical position. The maximum delivered energy was 195 J.
Results: See Table. Conclusions: Our clinical results demonstrate that the biphasic waveform with a delivered energy of ≤195 J was highly effective in cardioverting atrial fibrillation. Introduction: Patients are frequently admitted to ICU following elective surgery either because of surgery-specific or patient-specific risk. However, ICU beds are expensive and not always available. We assessed whether there were significant differences in outcome between patients admitted to an ICU and a similar group of patients managed in general surgical wards.
Methods: All patients undergoing elective surgery over a 17-month period and for whom a place in ICU was requested were included in this prospective study. The anesthetic and surgical departments were not made aware of the study. Following surgery, patients were either admitted to ICU if a bed was available (group 1, n=97) or transferred to the general surgical wards after a mean stay of 1.9+3.7 h in the recovery room (group 2, n=47). We recorded preoperative factors (demographics, reason for Results: See Table. Conclusion: We found no difference in three outcome variables in high-risk patients who were either admitted to the ICU or treated in a general surgical ward. Criteria for ICU admission need to be reassessed in these patients. Introduction: Abdominal sepsis after emergency laparotomy has a high mortality rate because is associated with a great number of serious complications. The aim of this study was to assess abdominal sepsis associated mortality and to determine which variables could predict the possibility that abdominal sepsis occur.

Materials and methods:
We prospectively studied 128 patients, 65 onwards, who underwent emergency laparotomy (abdominal resection, perforation or ischaemia). In every patient we recorded several variables of the pre, intra and postoperative period. By patient's history, radiographic and scan studies, laboratory tests specific for the type of surgery performed, wound inspection and surgical reexploration, we determine the presence of abdominal infection. We also studied mortality associated with this type of infection and variables that could predict its appearing (multiple logistic regression analysis, using a model in which probability an event P will occur is 1/(1 + e z ), where z is the linear combination, z=B 0 + B 1 X 1 + B 2 X 2 +.....B P X P; B0, B1, B 2 ,...B p are coefficients estimated from the data,×is the independent variable and e is the base of the natural logarithms, approximately 2.718). Statistical analysis was made with SPSS for Windows 5.01 ® (multiple logistic regression analysis and Fisher exact test as required) and P<0.05 was regarded as significant. Variables that can predict its appearing are: decreasing O 2 tissue delivery factors, events that increase the possibility of nosocomial infection or those circumstances favouring multisystemic organ failure (lung or kidney).

P220 Baseline audit of manipulation and management of intravenous therapy delivery systems C Martinsen, A Hughes and M Smithies
Critical Care Services, University Hospital of Wales, Cardiff CF14 4XW, UK Background: We are developing local evidence-based guidelines on the management of intravenous delivery systems in a 14-bedded Teaching Hospital General ICU1. A baseline audit was carried out to assess current practice prior to the publication of our proposed guidelines, and re-audit.

Methods:
We performed an observational audit of the setup or change of an intravenous infusion and the management of intravenous delivery systems. ICU staff were aware that an observational audit was in progress but blind to what was being observed. The observations were carried out over a three-week period.
Results: See Table. Conclusion: Practice is short of the stringent care necessary to avoid infection related to the delivery of intravenous therapy. The lack of consistency is also an issue. This audit highlighted the need for clinical guidelines and also the value of clinical audit of practice. Guidelines will now be introduced and developed in the light of appraisal of evidence, further audit and user feedback. Is a clean plastic apron worn? 8 5 Are clinically clean gloves worn? 8 5 Was alcohol swab used prior to disconnection? 0 13 Was connection allowed to dry before disconnection? 0 13 Was set disposed of as unit policy? 7 5 Was alcohol swab used on valve prior to connection? 0 11 Was valve allowed to dry prior to connection? 0 11 Set up or change of an intravenous infusion (n=10) Objective: To evaluate the effect of organizational and managerial intervention, aimed at increasing the professional collaboration between medical and nursing staff, upon the clinical outcomes in the ICU.
Design: Prospective, randomized, multi-center and multinational study.
Methods: All consecutive admissions were enrolled, during two periods of two months each.
Patient data: Admission data; first day SAPS II score; hourly registration of Critical Events (CrEv) defined as the duration (in hours) of out-of-range measurements of four parameters (heart rate, blood pressure, urine output and oxygen saturation); Sequential Organ Failure Score (SOFA) at admission and then every 24 h; ICU outcome.

Measurements:
Final outcome: ICU mortality. Intermediate outcomes: SOFA, CrEv (expressed as the % of time spent in CrEv; single or combined), and length of stay (LOS).
Intervention: 1) training of nursing and medical staff of 25 ICU's in inter-professional collaboration; 2) the use of a specific manual of instructions, supported by the daily use of two protocols covering awareness of processes of care, and professional dialogue (6 months).

Analysis:
The effect of the intervention upon final and intermediate outcomes was evaluated by comparing the values of these variables in the experimental and the control groups of ICU's, before and after the intervention (Table  below). SOFA computations included total daily score and Delta(∆)-SOFA [1]. A ∆-CrEv variable was also constructed. The analyses were made on Day 1, 3 and 5, in the group of patients staying ≥3 and ≤10 days in the ICU. χ²test was used for comparing means; significance if P<0.05 Results: Data on 1,633 patients were obtained; there were no significant differences in the 4 'study cells' for Age (median=69), SAPS II (median=35) and LOS (median=4 days). A significant decrease in ICU mortality was observed for the intervention group (16% control group; 9% intervention group).

Conclusion:
Collaborative practice has a significant and beneficial effect upon clinical outcomes in the ICU. Introduction: Guidelines for intensive care unit (ICU) admission recommend taking the premorbid functional status and comorbidity of patients into consideration when selecting those who are likely to benefit from ICU care [1]. We analyzed the prognostic value of pre-admission functional status, comorbidity and daily APACHE II scores in medical ICU patients.

Patients and methods:
All patients ≥18 years of age who stayed >24 h in our non-coronary medical ICU were eligible for inclusion. Functional status referred to the two months prior to admission and was assessed using the domains of basic physiologic activities and activities of daily life of the PAEEC quality of life questionnaire [2]. Comorbidity was categorized using the Charlson Index (CI) [3]. ICU, hospital and six-month follow-up mortality rates were documented. Statistical analysis used the cox model for censored data.

Conclusion:
In our population of critically ill medical patients with a high rate of preexisting chronic diseases, pre-admission functional status and comorbidity are independent variables with a highly significant association with on-unit and hospital mortality.  Anxiety and depression can be evaluated through different methods. We studied the current associations between the commonest evaluation methods that are present in scientific literature.

Materials and methods:
We analyzed a population of 1392 nurses working in intensive care units (ICUs) and in general medicine units (GMUs), distributed in 101 Italian hospitals (with a mean age of 32±6.9 years, 73.8 % female). We considered the following evaluation scales: the Hospital Anxiety and Depression scale, divided in anxiety (HAD A) and depression (HAD D) status; the STAI scale, divided in acute anxiety (Y-1) and chronic anxiety (Y-2) status; the Maslach Burnout Inventory-Human Services Survey (MBI.), divided in Emotional Exhaustion (EE), Depersonalization (DP) and Personal Accomplishment (PA). Assuming the HAD as a reference scale, we evaluated the influence of the others scales to determine HAD. The population was divided, distinctly for anxiety and depression, into three groups, according to standardized parameters of HAD -'non-cases' (HAD 0-7), 'doubtful cases' (HAD 8-10), and 'cases' (HAD 11-21). We used multiple linear regression models; statistical significance was accepted as P<0.05.

Results:
The regression coefficients of the multiple linear regression models are expressed in the table, with the variables that result in statistical significance. For depression, we considered doubtful cases and cases together (last being only 39).
Conclusions: 1) Besides HAD A, STAI Y-2 also seems to be useful for evaluating anxiety, whatever the degree of it.
2) Besides HAD D, STAI Y-1 and MBI EE also seem to be useful for evaluating depression, whatever the degree of it.  It's a current belief that stress is an outstanding feature of intensive care units, in particular within nursing staff. The aim of this study was to compare some variables belong-ing to stress (i.e. anxiety, depression and 'Burnout' syndrome) between nurses working in intensive care units (ICUs) and general medicine units (GMUs).

Materials and methods:
We studied a population of 883 nurses working in ICUs, distributed in 79 Italian hospitals (70.1 % female) and 509 nurses working in GMUs, distributed in 35 Italian hospitals (80.2 % female). We asked them to fill in a form including: 1) general data and his/her work environment; 2) different evaluation standardized scales -the Hospital Anxiety and Depression Scale, divided into anxiety (HAD A) and depression (HAD D) status 0-7 'non cases', 8-10 'doubtful cases', 11-21 'cases'; the S.T.A.I. scale, divided into acute anxiety (Y-1) and chronic anxiety (Y-2) status; the Maslach Burnout Inventory-Human Services Survey (MBI.) divided into Emotional Exhaustion (EE), ≤18 'low', 19-26 'average', ≥27 'high', Depersonalization (DP) and Personal Accom-plishment (PA). We also evaluated the different reasons of anxiety through individual questions (higher value, more anxiety): A1, a critically ill patient; A2, a young patient; A3, an old patient; A4, a suicidal patient; A5, a terminal patient; A6, presence of mechanical supports; A7, relationship with patients' relatives. The comparison between the two groups was performed by the Mann-Whitney Rank Sum test and z-test; statistical significance was accepted as P<0.05.

Results:
The results, expressed as median value, with 25th and 75th percentile in brackets, are shown in Tables  1 and 2. Table 1 also shows the proportions of nurses that had a highest value of HAD A and M.B.I. EE.

Conclusions:
Pathologic anxiety and emotional exhaustion are more prevalent in nurses working in GMUs. Thus, contrary to a common belief, 'stress' is a more distinctive peculiarity of general medicine units than intensive care units.  Background: Anxiety and depression have a major impact on the ability to make decisions. Characterization of symptoms reflecting anxiety and depression in family members visiting ICU patients may be of major relevance to the ethics of involving family members in decision-making, particularly about end-of-life issues.
Methods: Prospective multi-center study in 43 French ICUs (37 adult and 6 pediatric). Each unit included 15 patients admitted for longer than two days. ICU characteristics and data on the patient and family members were collected. Family members completed the Hospital Anxiety and Depression Scale (HADS) to allow evaluation of the prevalence and potential predictors of anxiety and depression.
Findings: 637 patients were included in the study and 920 family members completed the HADS. All items were completed in 836 HADS questionnaires, which formed the basis for this study. The prevalences of anxiety and depression in family members were 69.1% and 35.4%, respectively. Anxiety or depression were present in 72.7% of family members and 84% of spouses. Factors predictive of anxiety in a multivariate model included patientrelated factors (absence of chronic disease), family-related factors (spouse, female gender, desire for professional psychological help, help being received by usual doctor) and caregiver-related factors (absence of physician-nurse meetings on a regular basis, absence of a room used only for meetings with family members). The multivariate model also identified three groups of factors predicting depression: patient-related (age), family-related (spouse, female gender, not of French descent), and caregiver-related (no waiting room, perceived contradictions in the information provided by caregivers). Background: ICU caregivers should specifically seek to develop a collaborative relationship with ICU patient family members, based on a rich and free exchange of information and aimed both at helping family members to cope with their distress and at providing them with the opportunity to speak for the patient if necessary.

Methods:
We conducted a prospective multicentre study of family member satisfaction as evaluated using the Critical Care Family Needs Inventory (CCFNI) developed by the Society for Critical Care Medicine. Forty-three French ICUs (37 adult and 6 paediatric) participated in the study; each was required to include 15 consecutive patients hospitalised for longer than two days. ICU characteristics, demographic data on patients and family members and questions about satisfaction were collected. The additive score of all CCFNI items was the dependent variable. Factors associated with dissatisfaction were identified using multivariate analysis. Spearman correlation coefficients were also computed between the different scales that were found to be highly correlated (level of sig. >0.01). Stepwise multiple linear regression of CES-D upon various independent variables (i.e. age of patient, ICU diagnosis, APACHE II Score, age and gender of relative, degree of kinship, relative's socioeconomic and marital status, and the score on STAI-Trait inventory) showed that the severity of the depressive symptomatology depends significantly solely on the relatives STAI-Trait score.

Conclusions:
The main findings of the present study are: a) treatment in the ICU acts as a high impact stressor for the patient's relatives, equivalent to other traumatic events of an exceptionally threatening nature; b) females are significantly more prone than males to develop a severe anxiety-depressive reaction; they also experience more intense intrusive symptoms; and c) individuals with higher pre-exposure anxiety levels, particularly females, are more vulnerable to the development of an acute stress reactiontype symptomatology. Background: The use of general outcome prediction models in the ICU remains controversial: 1) none of the existing systems is able to precisely predict individual outcome; 2) based on data collected within the first 24 h after admission, their prediction power is insensitive to alternative courses of care (and of ways-of-working) in the ICU.
Objectives: To evaluate the importance of intermediate outcomes of care upon the final outcome of patient care in the ICU; to evaluate whether the consideration of intermediate outcomes of care do increase the predictive power of SAPS II score.
Design: Prospective, multi-center and multinational study.
Methods: All consecutive admissions were enrolled during a four-month period.
Patient data: admission data; first day SAPS II score; hourly registration of Critical Events (CrEv) defined as the duration (in hours) of out-of-range measurements of four parameters (heart rate, blood pressure, urine output and oxygen saturation); Sequential Organ Failure Score (SOFA) at admission and then every 24 h; ICU outcome.  When comparing a characteristic (e.g. outcome) between two groups, tests of continuous (as opposed to categorical) data that assume parametric (as opposed to nonparametric) distributions are the most powerful. Currently, we measure outcome differences in sepsis trials in two ways. Typically, we compare mortality rates at a given time-point using categorical, parametric tests (e.g. χ 2 or Fisher's Exact test of differences in mortality at day 28) or we compare survival times using categorical, non-parametric tests (e.g. the Log-rank test to compare Kaplan-Meier curves). But survival after sepsis decreases exponentially [1]. Thus, survival could be described by exponential curves, which can be compared using continuous, parametric tests, such as the Cox's F-test. We therefore used this approach in a cohort of septic patients to determine sample size requirements in comparison to traditional approaches.

Methods:
Patients: 1102 patients with severe sepsis enrolled in a US multi-center trial.

Sub-groups:
We divided patients into those with and without septic shock to select two groups with a difference in survival (10-15%) typical for many power calculations in sepsis trials.
Statistical procedures: For each sub-group, we plotted the survival to day 28 and fit distributions with exponential curves using the maximum likelihood procedure. Curves were then compared using the Cox's F-test. We also compared differences in outcome using the Fisher's Exact test (for day-28 mortality) and the log-rank test. Statistical significance was assumed at P<0.05.
Sampling procedure: After comparing tests on the entire sample, we then drew progressively smaller random samples of the cohort and repeated the test comparisons to determine the point at which statistical significance was lost for each test.

Results:
Patients with shock had a higher mortality than those without shock (see Table). This difference was statistically significant by Fisher's Exact and Log-rank tests until sample size fell below 500. Survival in all sub-groups was modeled by exponential curves with excellent fit (R 2 >0.98). Comparing these curves by Cox's F-test, statistical significance was maintained with a much smaller sample size (see Fig.).

Conclusion:
Taking advantage of the parametric distribution that characterizes survival after sepsis, we can apply a test that finds statistical differences in survival with smaller sample sizes than traditional approaches. These data suggest that the application of exponentially-modeled sur- Objective: To evaluate the performance of the Logistic Organ Dysfunction (LOD) system for the assessment of morbidity and mortality in multiple organ dysfunction/ failure (MOD/F) in an independent database and to evaluate the use of sequential LOD measurements for the prediction of outcome.
Setting: Thirteen adult medical, surgical, and mixed intensive care units (ICUs) in Austria.
Patients: A total of 2893 consecutive admissions to 13 adult medical/surgical intensive care units (ICUs) in Austria.
Outcome measure: Patient vital status at ICU-and hospital discharge.
Results: Univariate analysis showed that the LOD was able to distinguish between survivors and nonsurvivors (2 vs 6 median score, P<0.05). Within organ systems, higher levels of the severity of organ dysfunction were consistently associated with higher mortality. For the prediction of hospital mortality, the original prognostic LOD model did not perform well in our patients as indicated by the goodness-of-fit statistic (C=37.2, 9 df, P<0.0001). Using multiple logistic regression, we developed a prognostic model (using the LOD of the first ICU day) with a satisfactory fit in our patients. The integration of further measurements during the ICU stay could not increase the accuracy of the prediction.

Conclusions:
The LOD system can be used to quantify the baseline severity of organ dysfunction. Moreover, after customization of the predictive equation, the LOD was able to predict hospital mortality in our patients with high precision. It thus provides a combined measure of morbidity and mortality for critically ill patients with MOD/F. Design: Prospective cohort study.
Setting: General medical and surgical ICU in a tertiary teaching hospital in city of São Paulo, Brazil.
Patients: 318 consecutive, unselected patients over the period from February to June of 1999. Cardiac surgical and burns patients were excluded.
Outcome measure: Patients vital status at ICU discharge.

Interventions: None.
Measurements and statistical analysis: Data required to calculate the patient's predicted mortality by the considered scoring systems were collected. Area under the receiver operating characteristic curve (ROC curve) was calculated to evaluate discrimination of each scoring system. The goodness-of-fit statistic was used to assess calibration of the models and the strength of the association between the predicted and observed outcome.
Main results: The mean age was 55.2±19 years. The length of ICU stay was 10.64±13.09 days, and the mortality rate was 26.1 %. Calibration and discrimination results and the predicted risk of death are shown in the Table. Conclusions: The results showed good discrimination for all the models.
LODS, MPM 0 and SAPS II showed good calibration, i.e, there were no significant discrepancy between predicted and observed mortality. MPM 24 showed a poor fit.
According to this, MPM0 can be used as a predictor of death at patients admission, SAPS II as a predictor of death in the first 24 h of admission, and LODS as end point in studies of organ dysfunction in our ICU.
Scoring systems need to be evaluated in terms of predictive accuracy within a single institution before applying them to make quality of care assessments. Results: APS and SAPS II were not significantly different between survivors and nonsurvivors on day 1 and 2 but were different on day 3 (see Table). APS and SAPS II decreased from day 1 to 2 in both survivors and nonsur-vivors (-29% vs -31%, P=NS for APS, -24% vs-21%, P=NS for SAPS II). However, while APS and SAPS II continued to decrease from day 2 to 3 in survivors it increased in nonsurvivors (-15% vs +8%, P=0.03 for APS, -7% vs +5%, P=0.26).

Conclusions:
In the postoperative liver transplant patients, the following are concluded: 1) in contrast to the progressive decrease in APS and SAPS II in survivors, these scores increased in nonsurvivors between day 2 and 3 after an initial drop; 2) day 3 APS and SAPS II scores better differentiated survivors from nonsurvivors than day 1 and 2 scores; 3) a mortality prediction system based on day 3 score or on serial scores is likely to be more accurate than the traditional systems using the first 24 h data. Introduction: Severity stratification in acute pancreatitis has long been a subject of debate. The availability of instruments specific for this pathologic condition lead some intensivists to argue for their use in this condition. However, to the best of our knowledge, no published study competed all these scores with general severity scores and organ failure scores on the same cohort. The objective of this work is to compare six disease-specific scores with two general severity scores (APACHE II and SAPS II) and one organ failure score (Sequential Organ Failure Assessment [SOFA] score) in patients admitted with acute pancreatitis to a mixed medical-surgical ICU.

Material and methods:
We analysed all the patients discharged from the UCI from July 1 1991 to November 30 1999 with a diagnosis of acute pancreatitis. Basic demographic and clinical data were registered, as were outcome at ICU and hospital discharge as well as APACHE II, SAPS II, SOFA score (at admission, 24 h, 48 h and maximum during ICU stay), admission Ramson score, Ramson score at 48 h, Imrie score, Osborn score, Blamey score, Balthasar score, collected according to the original descriptions. Raw data necessary for the computation of the scores has been registered prospectively, using a proprietary computerised system.
The discriminative power of the scores was evaluated through the use of the area under the Receiver Operating Characteristics (ROC) curve. Two-sample student T-test was used for the comparison of survivors and nonsurvivors. The outcome measure used was vital status at hospital discharge.
Results: During the study period, 49 patients were discharged with a diagnosis of acute pancreatitis. Biliar tract disease (n=26) and alcoholism (n=8) where the most In conclusion, in this study we could demonstrate an activity based costing methodology which measures the patient-related costs of care in both, ICU and IMCU. The TISS is a valuable tool when evaluating costs in the ICU and in the IMCU. Charging patients to diagnosis related groups allows calculation of costs and resource allocation to different specialties. Comparing the ICU and the IMCU there was a daily cost difference of 171 DM (87.43 Euro) per patient due to a lower workload and lower costs of the nursing staff in the IMCU. These findings support the potential cost saving of managing patients in an IMCU. To reduce costs without decreasing quality is a challenge.
Objectives: To demonstrate a cost management program in a 12-bed ICU of a general hospital with 72 beds. We've chosen three drugs of great impact in the bill and we've established a program that rationalized or replaced these drugs and a way to control if there was any negative impact on quality.

Methods:
The period of the study was from January 1998 to July 1999. The three drugs chosen were human albumin, omeprazole and midazolan. To decide when and how to use albumin and omeprazole, we employed evidence-based medicine concepts; with midazolan, we changed to diazepam, in equivalent doses, and controlled the "ideal doses" by the Ramsay scale.

Results:
We achieved a considerable reduction in global costs, specifically US$ 20 000 for albumin, US$ 3200 for omeprazole and US$ 2000 for midazolan, each month. The global economy was US$ 277 000 a year, without major mortality (compared to predicted by the APACHE II score), increased length of stay (LOS), discomfort or difficulties inside the medical group. The benefits are maintained until now (19 months later).

Conclusion:
Control costs without loss of quality is a challenge and a duty. We proved that this is possible without any problem. It's necessary to have clear targets and scientific basis, as evidence-based medicine, in conducting the protocols. Other areas are being included in this approach in our group.