key: cord-0033869-y6y5fcq0
authors: nan
title: Poster Sessions 292 - 465
date: 2003-09-23
journal: Intensive Care Med
DOI: 10.1007/s00134-003-1973-y
sha: c6c38b432c6caac88df6fe86ca3b68ab68d2d374
doc_id: 33869
cord_uid: y6y5fcq0

nan

Patients with severe hypoxemia and bilateral infiltrates on the chest X ray had no reduction in Cst,rs and showed only consolidation in the dependent zones on CT scan. Ventilatory and hemodynamic strategies were not different between the 2 groups. This suggest that severe hypoxemia in brain injured patients recognizes pathophysiological features different from those observed in non-brain injured patients with ALI. Grant acknowledgement: On behalf of Brain IT group supported by the EEC grant QLCT-2002-001160

Villagrá A 1 , López-Aguilar J 1 , Bernabé F 1 , Camponovo E 1 , Piacentini E 1 , Peña R 1 , Artigas A 1 , Blanch L 1 1 Critical Care Centre, Hospital de Sabadell. Institut Universitari Parc Taulí. UAB, Sabadell, Spain

The relative contribution of inspiratory flow rate on the development of ventilator induced lung injury (VILI) is unknown. Our hypothesis was that decelerated flow (DF) with a higher inspiratory flow rate contributes to VILI.

Twelve rabbits were ventilated in volume controlled mode (tidal volume (VT) 10 ml/kg, positive end-expiratory pressure 0 cmH 2 O, respiratory rate to obtain normocapnia and inspirated oxygen fraction (FiO2) 0.4) for 30 minutes. After 60 minutes ventilated with FiO2 of 1 to induce reabsorption atelectasis, animals were randomized to constant (CF) or DF group (pressure controlled mode with peak pressure to maintain previous VT) for 3 hours. Gas exchange and lung mechanics were assessed at baseline, after reabsorption atelectasis induction, and 1 and 3 hours later. Groups were compared with ANOVA for repeated measures.

RESULTS. Data are expressed as mean ± SD. 39.2 ± 2.9 40.9 ± 4.3 42.6 ± 6.9 37.9 ± 2.2 43.0 ± 5.5 51.3 ± 6.4 VT*,ml/kg 11.7 ± 0.8 11.9 ± 0.8 11.8 ± 0.8 10.9 ± 0.5 9.5 ± 0.8 8.4 ± 1.0 Crs*,ml/cmH 2 

There was no tidal FL in any animal at any time during the investigation. The increase in resistance was mostly due to DeltaRL (table 1) . The Est,w/Est,L ratio was less than 1 at T0 and further decreased after injury, specially in the SL model. 

In 6 female pigs (35-38 kgs) anesthetised, tracheostomised and mechanically ventilated, acute bronchoconstriction was induced by subcontinuous nebulisation of MCH (FDC 88, Mediprom, Paris, France). The objective was maximal inspiratory pressure (Pmax) of 40 cm H20. Once this achieved, the nebulisation was stopped and measurements of tidal FL using the NEP test (1) and respiratory mechanics using rapid airway occlusion technique were performed including: total positive end-expiratory pressure (PEEPt), interrupter (Rint) and tissue (DeltaR) resistance and static elastance (Est). The measurements were done before MCH (T0) and at time of target-pressure (T1). Values are mean±SEM.

Before MCH, tidal FL was not detected in all the animals. At T1, FL was present in 4 pigs ( 

Cunha Goncalves D 1 , Larsson A 2 , Bjorklund L 1 , Curstedt T 3 , Ingimarsson J 4 , Johansson J 5 , Perez de Sa V 1 , Robertsson B 6 , Tonnesen E 7 , Werner O 1 1 Anest, Univ Hosp, Lund, Sweden, 2 Anest, Gentoft, CPH, Denmark, 3 Clin Chem, KS, Stockholm,

In total 100 paired IAP and Cdyn measurements were performed in 4 body positions (supine, anti-trendelenburg, trendelenburg and upright) in 22 ventilated ICU patients. The Cdyn (ml/cmH20)= tidal volume/(plateau pressure -PEEP). The M/F ratio was 1/1, BMI 25.8±5.3, age 66.1±14.5, APACHE-II 27.7±9.8, SAPS-II 60.2±13.3, MODS 7.4±3.2, and SOFA 10.3±3.5.

The IAP was significantly higher in the anti-trendelenburg and upright position versus the supine, and significantly lower in the trendelenburg position vs the supine (p<0.0001, one-way Anova). Table 1 lists the mean values for IAP and Cdyn in the different positions. The Cdyn was lowest in the upright position vs the supine (p<0.0001, 2-tailed paired student's t test). There was only in the upright position a poor but slightly significant correlation between IAP and Cdyn (p=0.05 

Putting a patient in different body positions has significant effects on IAP. This is in contradiction with the hypothesis that the abdominal compartment is fluid in character, since IAP would then remain constant. Assessment of IAP should therefore be done in the complete supine position. The upright position significantly increases IAP, and lowers Cdyn compared to the supine. The effects on IAP are more pronounced in obese patients and the effects on Cdyn more in the non-obese. Putting a patient upright may deteriorate respiratory function, caused by the acute increase in IAP.

Aim of this study was to investigate the effects of a protective ventilatory strategy during one lung ventilation for open lobectomies. Arterial oxygen deterioration, due to partial dependent and total nondependent lung collaps, may be counteracted by alveolar recruitment manoeuvres.

Six patients(4M/2F), undergoing elective open lobectomies, were studied, after informed consent, at five time points: T 0 = during two-lung ventilation (TLV), T 1 = during one-lung ventilation(OLV)before and T 2 ,T 3 = after two consecutive alveolar recruitment manoeuvres (ARM), performed at one hour interval; T 4 = at the end of the procedure, before extubation. At these time points blood samples were drawn for blood-gas analysis with the patient turned on one side. All the patients were ventilated using a Servo 900 C ventilator in a volume controlled mode at a VT of 10ml.Kg -1 , at a rate of 10-12 bpm, with no PEEP and with a Ti/Ttot= 0,33. Minute ventilation and FiO 2 were adjusted in order to maintain etCO 2 within 32-35 mmHg and SaO 2 above 0,95, respectively. ARM consisted of a PEEP application (10 cm H 2 O) after a brief pulmonary inflation of the dependent lung at 35-40 cm H 2 O peak pressure.

Mean values(±SD)of PaO 2 /FiO 2 resulted: 449(±146)at T 0 , 221(±124)at T 1 , 299(±110)at T 2 , 450(±66)at T 3 and 515 (±104) at T 4 . These results were analysed with ANOVA test for repeated measures. Oxygenation values were normalized to the different FiO 2 levels used; p<0.05 was taken as significant. T 1 vs T 0 PaO 2 /FiO 2 values showed a significant decrease (p<0.05), while no significant change was found when comparing T 2 with T 1 values. On the contrary a statistically significant difference (p<0.05) was found between T 3 and T 2 oxygenation parameters testifying the usefulness of the ARM in terms of oxygenation improvements.

Arterial oxygen deterioration demonstrated to ameliorate at least at T 3 thanks to ARM; the improvement was also further confirmed, in terms of PaO 2 , during OLV at constant FiO 2 (T 2 ,T 3 ).

Lower and Upper Inflection Points (LIP and UIP, respectively) determined by Pressure Volume Loops have been utilized by some clinicians to guide setting PEEP and tidal volumes during ventilator management. However, current methods for determining inflection points are cumbersome and technique dependent. A new technique utilizes delivery of small consecutive pulses of gas to the lungs for 10-15 seconds and simultaneous measurement of pressure changes. The LIP and UIP are determined by pressure-time analysis. The aim of this study was to compare the inflection points obtained by the Pulse Pressure-Time (PPT) method to the Super Syringe on a previously validated mechanical lung model of ARDS.

Three separate measurements each of the LIP and UIP were conducted with the PPT method and the Super Syringe utilizing three different PEEP levels to simulate different lung conditions. Compliance was held constant at 0.05 L/cmH20. Data are reported as mean ± SD. Pearson product-moment correlation coefficients were calculated to determine the strength of association between measurements. Results were considered significant if p< 0.05. 

Piagnerelli M 1 , Su F 1 , Zouaoui Boudjeltia K 2 , Wang Z 1 , Vanhaeverbeek M 2 , Vincent J 1 1 Dpt of Intensive Care, Erasme University Hospital, Brussels, 2 Experimental Medicine Laboratory, ISPPC A. Vesale Hospital, Montigny-le-Tilleul, Belgium

We recently observed that sialic acid (SA), a carbohydrate included in the terminal position of glycoproteins, rapidly decreases in septic patients both on red blood cell (RBC) membranes (1) and on circulating proteins (2) . This was associated with an increased in free SA in the sera (2) . One possible mechanism of the SA metabolism alterations could be an increased effect of circulating sialidase produced by white blood cells or by the endothelium. To confirm the rapid alteration of SA metabolism in sepsis, we studied the evolution of total (free Sa and SA bound by inflammatory proteins) and free SA in a sheep model of septic shock induced by cecal ligation and perforation.

METHODS. 10 mature female sheep were anesthetized, mechanically ventilated and invasively monitored. Baseline measures were performed before surgery (T-1) and after cecal perforation (T0). Sheep were fluid resuscitated for 20 hours (T0-T20) and there after sacrificed. Total serum SA was measured each hour by an enzymatic colorimetric assay (SA, Roche) adapted for microdeterminations. Free SA concentrations were obtained by the same method after ultracentrifugation to delete proteins. Blood lactate concentrations were also determined each hour.

Temperature increased rapidly (T4; p = 0.02). Both mean arterial pressure and cardiac index decreased rapidly and after increased with fluids resuscitation (T5; p < 0.01). Lactate also rapidly increased (T1; p < 0.01). Total SA decreased (0.55 ± 0.11 versus 0.45 ± 0.07 mg/dL at T1;p <0.05) and free SA significantly increased at T 15 ( 0.018 ± 0.015 mg/dL versus 0.067 ± 0.04; p < 0.05).

In septic shock, an increased sialidase activity is probably responsible of the early increase in free SA concentration. As SA plays an important role in the half-life of RBCs and proteins, blockade of sialidase activity may represent an interesting therapeutic option in sepsis. 

Bracht H 1 , Hauser B 1 , Ivànyi Z 1 , Kelbel I 1 , Radermacher P 1 1 Department of Anesthesiology, University clinic Ulm, Ulm, Germany INTRODUCTION. NF-kappaB plays a key role in inflammatory processes. The cyclopentenone prostaglandine 15-Deoxy-Delta12,14-PGJ2 can inhibit NF-kappaB activation through inhibition of IKK.We tested the effects of 15-Delta-PGJ2 in a porcine model of hyperdynamic long term endotoxemia.

Up to now 12 anesthetised, ventilated and instrumented pigs were randomized to receive either placebo or 15-Delta-PGJ2 started at 12h of i.v. LPS (1 g/kg/min for 1h, 0,25 micg/kg/min for the remainder 11h). Before and at 12, 18 and 24h of LPS we assessed systemic and hepato-splanchnic hemodynamics and metabolism. Hydroxyethyl starch was infused to maintain mean arterial blood pressure above 60 mmHg. A Friedman rank sign analysis of variance and a subsequent Dunnett test and the Mann-Whitney test were used to analyse intra-and intergroup differences, respectively.

Before endotoxin there was no intergroup difference. 15-Delta-PGJ2 infusion did not significantly affect any of the hemodynamic and metabolic parameters assessed. T75 ->  T75 ->  T135 ->  T90 ->  T90 ->  T180 ->  T120 ->  for  end  T90  end  end  T120 end end

The systemic and metabolic responses to LPS are extensive, involving the overall energy metabolism in a complex fashion. The full metabolic response is delayed until 120-180 min after LPS and is sustained thereafter. The increased metabolic rate is mainly fuelled by a stimulation of fat oxidation, like in septic patients.

Hyperlactatemia is a hallmark of hypermetabolic and hyperdynamic sepsis in critically ill patients. It is correlated with the development of multiorgan failure and patients outcome. The aim of our study was to determine the contribution of whole body and skeletal muscle lactate production after lipopolysaccharid (LPS) administration in healthy subjects.

We investigated the metabolic and inflammatory responses in 14 healthy male volunteers in an experimental model simulating an acute inflammatory syndrome. Subjects were investigated twice: once with 2ng/kg bacterial LPS and another with saline (control). In the first group (n=6) we assessed the contribution of skeletal muscle to lactatemia from the gradient between interstitial skeletal muscle (microdialysis) and arterialized blood lactate concentrations.

In the second group (n=8), during continuous infusion of lactate, lactate clearance (LC) and endogenous lactate production (ELP) were calculated from plasma lactate levels, using a pharmacokinetic model. The measurements were performed before and 270 minutes after LPS (T270). Results were analysed by ANOVA for repeated measurements; significance-level p<0.05.

Body temperature, heart rate, energy expenditure and lipid oxidation increased starting at T60-T120, reaching plateau at T240-T360 and remaining significantly elevated till the end of test. In the first group without exogenous lactate infusion, the muscle-systemic lactate gradient was positive but remained unchanged after LPS, being not different from values observed in the controls. LPS induced a peak of plasma lactate and TNF-alpha at T90. IL-6 peaked at T120; all values returned to baseline within 60 min following the peak. In the second group, combining LPS and lactate infusion, hyperlacatemia persisted till the end of test, while returning to baseline in controls. Compared to initial values, both, LC (+31%, p=0.03) and ELP (+75%, p=0.04) were increased after LPS. Exogenous lactate infusion without LPS did not change LC (p=0.32) or ELP (p=0.55).

Muscular microdialysis results suggest that LPS-induced hyperlactatemia does not originate from muscle. The main mechanism explaining hyperlacatemia is increased ELP rather than decreased LC. 

Sepsis is a major concern in intensive care units. Strategies to block or control inflammatory mediators such as cytokines and biologically active lipids are currently being investigated as possible new adjuvant therapies for sepsis. In the present work, we investigated if inactivation of platelet-activating factor (PAF) and PAF-like lipids by administration of recombinant PAF-acetylhydrolase (rPAF-AH) protects mice from endotoxemia and/or sepsis induced by cecal ligation and puncture (CLP). We also investigated the activity of PAF-AH in the plasma of septic animals and patients.

In order to do that we used the CLP (cecal ligation and puncture) model or an endotoxemia model in Swiss mice and treated those animals with PAF-AH (1mg/kg, i.p. 

In this study we observed significant decreases in the PAF-AH activity in the plasma of patients diagnosed with septic shock and in mice that had been either submitted to the CLP procedure or that had been injected with LPS. Administration of rPAF-AH increased PAF-AH plasma activity and significantly protects mice submitted to CLP from death (from 40% survival rate in the control group to 80% in the group treated with PAF-AH). Similarly, when animals were injected with LPS the survival rate increased in PAF-AH treated group. A combination therapy that included both the antibiotic Imipenem and rPAF-AH was more protective of animals undergoing experimental sepsis than single treatments with either of these agents. In addition, the combined therapy resulted in reduced IL-6 levels in mice subjected to CLP.

We conclude that inactivation of PAF and/or PAF-like lipids by exogenous administration of rPAF-AH is an effective way to reduce hiperinflammation and mortality during sepsis and endotoxemia. 

Cationic antimicrobial protein of 37 kDa (CAP37) is a multifunctional inflammatory mediator mainly stored in polymorphonuclear leukocytes and able to induce endothelial hyperpermeability (1). It was questioned whether an alleged increase of CAP37 during cardiopulmonary bypass could be causative in systemic inflammatory response syndrome sometimes seen in these procedures. We report both on the time course of CAP37 plasma concentration during coronary artery bypass grafting and the CAP37 releasing effect of unfractionated and fractionated heparin.

METHODS. CAP37 plasma levels in 10 patients undergoing coronary artery bypass grafting surgery were determined by ELISA. In the first protocol, CAP37 concentrations in 5 patients were measured during and up to 24 hrs after the surgical procedure. In a slightly modified second protocol, the impact of routine antagonisation of heparin by protamin on plasma CAP37 concentration after cessation of the cardio-pulmonary bypass was investigated in another 5 patients. To evaluate whether fractionated heparin has the same CAP37 releasing effect of (unfractionated) heparin, CAP37 plasma levels in 14 consecutive patients of an internal medicine ward anticoagulated with subcutaneously administered low molecular weight heparin were studied.

Median CAP37 plasma concentration was raised significantly from 8ng/ml at baseline to 428ng/ml 3 min. after routine administration of heparin. With the onset of cardiopulmonary bypass, CAP37 plasma concentrations further increased to 3470ng/ml 5min. after opening of the aortic clamp. 15min. after administration of protamin, CAP37 plasma levels dropped significantly to a median of 241ng/ml corresponding to 8.9% of the CAP37 concentration measured immediately before protamin had been given. Moreover, a significant increase of the median CAP37 plasma concentration from 3.3ng/ml at baseline to 11.7ng/ml 3 hrs after subcutaneus administration of low molecular weight heparin was seen.

Beside the cardiopulmonary bypass procedure, unfractionated and fractionated heparin represent a stimulus for CAP37 release in humans. Protamin seems to be a strong enhancer of CAP37 clearance from human plasma. 

To test the hypothesis that activated neutrophil-?@endothelial cell interactions in disseminated intravascular coagulation (DIC) cause endothelial injury contributing to multiple organ dysfunction syndrome (MODS) and poor outcome in sepsis.

Forty-five patients with sepsis, severe sepsis, and septic shock were studied. The patients were subdivided into two groups, 27 with DIC and 18 without DIC.?@Serial levels of soluble L-, P-, and E-selectins, intercellular adhesion molecule-1 (sICAM-1), vascular cell adhesion molecule-1 (sVCAM-1), thrombomodulin (sThrombomodulin), and neutrophil elastase were measured within 12 hours after the diagnosis of sepsis (day 0), and on days 1-4 after the diagnosis. The numbers of systemic inflammatory response syndrome (SIRS) criteria that patients met and DIC scores were determined simultaneously.

Acute Physiology and Chronic Health Evaluation II scores were identical between the two groups. In the DIC patients, higher DIC scores, lower platelet counts, and more maximum numbers of SIRS criteria were observed compared with the non-DIC patients. The incidence of MODS and the number of the dysfunctioning organs were higher in the patients with DIC than those without DIC, and the DIC patients had poor outcome. Soluble L-selectin (sL-selectin) levels in both groups tended to be lower than those in the control subjects. All other parameters both in the two groups were continuously higher than those in the control subjects during the study period. The levels of sE-selectin, sICAM-1, sVCAM-1, neutrophil elastase, and sThrombomodulin were elevated in the DIC patients than those in the non-DIC patients. There were no differences in the sP-selectin levels between the two groups, however, more increased sP-selectin levels per platelet were found in the DIC patients compared with the non-DIC patients. Maximum DIC scores in the DIC group positively correlated with the peak levels of neutrophil elastase and sThrombomodulin, and the number of the dysfunctioning organs.

CONCLUSION. Systemic inflammation caused by neutrophil-endothelial cell interactions contributing to endothelial injury in DIC give rise to MODS and poor outcome in patients with sepsis, severe sepsis, and septic shock. The result suggests a strong critical link between inflammation and thrombosis in septic patients INTRODUCTION. Antithrombin (AT) is an important endogenous factor to control coagulation, fibrinolysis and inflammation. We already suggested that AT could inhibit the acceleration of coagulopathy and polymorphonuclear leukocyte elastase production in major vascular surgery 1 . The purpose of this study was to investigate the effects of AT on cytokine production in major vascular surgery using the analysis of cytokine levels in the blood samples obtained in the previous study 1 .

Twelve patients (aged 55 to 80 years) scheduled for Y graft replacement for abdominal aortic aneurysm were divided into two groups (AT group and Control group) at random after informed consent and institutional approval. Heparin 2,000 units were administered before clamping the aorta. In the AT group, AT 3,000 units were infused before clamping the aorta and 24 hours later (Total dose was 6,000 units). Concentrations of AT, interleukin (IL)-1Beta, -6, and -8 and tumour necrosis factor (TNF)alpha were measured before surgery, at the end of surgery, one and two days after surgery.

Results are summarized in the table as mean values (AT group/Control group). *: P < 0.05 between the two groups, $: P < 0.05 vs. the value before surgery. IL-1eta and IL-8 could not be analyzed statistically because of the values less than the detection limit.

End 

Accumulation of asymmetrical dimethylarginine (ADMA) has been linked to endothelial dysfunction, and is an important risk factor for cardiovascular disease. Its elimination from the body is dependent on urinary excretion and degradation by the enzyme dimethylarginine dimethylaminohydrolase. This enzyme is highly expressed in the liver, and in rat studies a high net hepatic uptake of asymmetrical dimethylarginine was found. In critically ill patients, we investigated the relation between indicators of renal and hepatic dysfunction and plasma ADMA concentration, and tested the association between ADMA concentration and outcome.

We prospectively collected blood samples from a cross-section of critically ill patients (n=52) with clinical evidence of dysfunction of more than two organs. We identified correlates of plasma ADMA concentration with laboratory values, organ failures score and outcome by univariate and multiple regression analyses. 

Drotrecogin alfa (activated) [DrotAA] administration to adult severe sepsis patients in the PROWESS trial significantly decreased 28-day and hospital mortality. We explored regional differences, both overall and by treatment group, in hospital survival and resource use in the US, Europe (EU), and other (OT).

We determined intensive care unit (ICULOS) and hospital length of stay (HLOS), hospital discharge location, and hospital mortality in a long-term follow-up study of PROWESS. Follow-up at hospital discharge was 98% (n=1658). TISS-28 (T28) scores were recorded while the patient was in ICU during PROWESS. All comparisons were performed with Wilcoxon rank-sum or Kruskal-Wallis tests. 

A survival benefit with DrotAA was observed at hospital discharge regardless of region. Regional differences in overall total resource use likely reflect differences in health care systems and reimbursement mechanism in the US, EU, and OT regions participating in PROWESS and were not due to increased number of survivors with DrotAA. 

The cellular immune function is closely related to patient outcome in sepsis (1). It can be studied using the ex vivo whole blood stimulation with LPS which includes the capacity of all leukocytes. The method had been shown to be associated with mortality and morbidity (2) . In this study we investigated the capability of the ex vivo stimulation to predict outcome in comparison to flow cytometric parameters.

We investigated 40 septic patients (ACCP/SCCM) from 2 operative and a medical ICU and 25 healthy volunteers for determination of normal values. Ex vivo stimulation was performed using 500 pg/ml LPS and an incubation time of 4 hours. Production of TNFalpha was detected by a semiautomatic ELISA system (DPC Biermann, Bad Nauheim, Germany). Human Leukocyte Antigen (HLA)-DR expression was assessed by flow cytometry as quantitative analysis (MFI). Daily differential blood counts were performed using a Sysmex XE2100 cell counter (Sysmex, Norderstedt, Germany). Patients were followed up for a maximum of 14 days on ICU and for ICU-mortality.

Overall mortality was 15%(6/40). Mean APACHE II on admission was 16.5(SD: 5.4) and differentiated well between survivors (15.5(4.5)) and nonsurvivors (21.8(6.7))(p<0.05). A reduced capability to produce TNFalpha after stimulation below 200 pg/ml for more than 48h was associated with a 4.5 fold relative risk to die. Association with mortality was significant for ex vivo stimulation and for HLA-DR expression (ex vivo: p<0.05, HLA-DR: p<0.05; Pearson chi 2 ). Sensitivity and specificity for the ex vivo stimulation were 0.67(4/6)and 0.77(23/30) and for HLA-DR expression 1.0(6/6) and 0.47(16/34), resp. The parameters showed a low correlation coefficient (CC)(0.38; Spearman). Regarding the differential blood count, TNFalpha ex vivo production was significantly correlated with monocyte numner: percentage (CC=0.5), count (CC=0.4); p<0.001).

CONCLUSION. In our study cellular immune function as assessed by stimulated TNFalpha ex vivo production showed a close correlation to survival. The same was true for monocytic HLA-DR expression. TNFalpha ex vivo values are positively correlated with the number of monocytes. The data give further evidence for the importance of the monocyte system in human sepsis. 

The lowered expression of HLA-DR expression on monocytes has been well documented as a feature in sepsis. The precise mechanism by which this occurs is still not fully understood. We have previously described that shedding of HLA-DR from the surface of monocytes may be one mechanism by which a lowered surface expression. A recent report has demonstrated that a post-translational mechanism may exist where HLA-DR is retained within the cell or endocytosed from the surface 1 . In this study, we attempted to ascertain if this mechanism of HLA-DR regulation was a feature in our septic patients and also in healthy volunteers. We also investigated whether the rate of transcription of HLA-DR was different in healthy controls compared with septic patients.

After consent was obtained 7.5ml of arterial blood was taken from patients who fulfilled the ACCP [n = 5]. Peripheral blood mononuclear cells were isolated by Ficoll Paque isolation and stained for HLA-DR expression in both permeabilised and non-permeabilsed cells for intracellular and surface HLA-DR expression respectively. Immunofluorescent microscopy was used for detection of both intracellular and surface HLA-DR on monocytes. Reverse transcription PCR was used to investigate the levels of HLA-DR mRNA in septic patients and healthy controls. A numerical value was generated from the bands on the gel using densitometry software, which allowed statistical analysis between septic patients and healthy controls.

In the septic patients we studied it was observed that there was a higher intracellular presence of HLA-DR in the septic patients (n=5) compared to the healthy controls (n=5). This was observed in septic patients with reduced surface expression of HLA-DR. When the mRNA of septic patients was compared to healthy controls, on days 0 and 3 the mRNA of the septic patients were statistically significantly lower than that of healthy controls, p=0.02 and p=0.05 respectively.

We have shown that in our septic patient the mechanism which may contribute to the lowered surface expression are shedding of HLA-DR from the surface (previously shown), transcriptional and post translational regulation. Future work needs to be done to investigate the mediators that may cause the intracellular HLA-DR to be expressed on the surface. 

The human leukocyte antigen-DR expression on monocytes (HLA) significantly characterizes the immune function. The clinical impact of the HLA has been proven before. Common methods to measure the HLA are either not standardised or not evaluated clinically. We therefore investigated the performance of two common and of one standardised method in septic patients. 

Overall ICU mortality was 24% (14/59). Mean APACHE II was 17.5(SD:6.5). Prolonged decrease of HLA was associated with increased mortality using all 3 methods. MFI<200 over a period of at least 48h (>48h) showed a 5.6-fold relative risk to die. HLA-DR+ (<75%>48h) and ABC (<10,000>48h) were associated with mortality. 

Depression of monocytic HLA-DR is strongly associated with mortality in this population of patients. Although there are different methods to assess this parameter with varying reliability, the MFI method showed the highest values for sensitivity and specificity for ICU mortality. The results of the different methods correlated well with each other. Determining an optimal and standardized method to assess HLA-DR expression needs to be subject of further clinical studies.

Haveman J W 1 , Verhoeven E L G 1 , Van den Dungen J J A M 1 , Van den Berg A P 2 , The H T 2 , Zwaveling J H 1 1 Surgery, 2 Clinical Immunology, Groningen University Hospital, Groningen, Netherlands

Mortality in patients after surgical repair of ruptured abdominal aortic aneurysms (RAAA) remains high. Downregulation of HLA-DR expression on monocytes is associated with sepsis and mortality in surgical patients. The prognostic role of HLA-DR expression in non-infectious disease after ischemia-reperfusion injury remains unknown. We studied the role of HLA-DR expression on monocytes on outcome in RAAA patients.

Patients with RAAA were prospectively analysed. Blood samples were collected on day 1, 3, 5, 7, 10 and 14 and analysed the same day. The percentage of CD-14 positive monocytes expressing HLA-DR was measured by flow-cytometry. 

Studies have shown that lipoprotein abnormalities exist in critically ill patients versus normals. Lipoproteins play a key role in the transport of lipids, and have been shown to be involved in lipopolysaccharide sequestration and clearance, scavenging of reactive oxygen and nitrogen species, as well as interacting with plasma proteins involved in coagulation.

In this study we evaluated lipoproteins and other biochemical parameters upon ICU enrollment in relationship to outcome in a critical care setting.

Observational study of 20 consec. adult ICU patients w/ expected stay of >72 hours. Analysis of lipids, markers of acute phase and coag parameters was performed. NMR analysis was carried out to quantify lipoprotein particle numbers, sizes and surface areas. SOFA scores were calculated daily. Diagnosis of sepsis was by physician adjudication.

. 15 of 20 patients were adjudicated as septic. Upon enrollment, there was a significant decrease (P<0.05) in total plasma protein, total, LDL and HDL cholesterol, number of HDL particles, HDL and total lipoprotein surface area and in LDL size from patients who were septic versus those not septic. 10 of 20 patients experienced 2 or greater organ failures at some time during their ICU stay. Upon enrollment, there was a significant decrease (P<0.05) in total plasma protein, total triglycerides, surface area from beta lipoproteins, HDL particle size and in ATIII values from patients who experienced >= 2 organ failures versus those with one or less organ failures during the ICU stay.

In this pilot study, enrollment lipoprotein abnormalities between critically ill patients were observed in relationship to the outcome measures of sepsis and organ failure. For septic and >=2 organ failures, significant decreases were seen in certain lipoprotein classes and with serum lipid concentrations. Though all 20 patients showed significant hypolipidemia in comparison to a normal population, this study has shown that even among the critically ill population, differences between lipoprotein measurements may indicate or may potentially show the role played by lipoproteins in the occurrence of specific clinical conditions. A study involving a larger cohort of patients more representative of the total ICU population is necessary to validate and to expand upon these observations. 

Brunkhorst F M 1 , Burgett U 1 , Engel C 2 , Reinhart K 1 1 Anaesthesiology and Intensive Care Medicine, Friedrich-Schiller-University, Jena, 2 KKSL, University Leipzig, Leipzig, Germany

Lipopolysaccharide binding protein (LBP) has been described as an inflammatory acute phase protein, contributing to the onset and progression of infection, sepsis, and organ dysfunction. We explored the relationship between LBP levels and the presence and development of organ dysfunction according to the ACCP/SCCM criteria in a large ICU-cohort.

We performed a 7-months prospective study in a 28-bed surgical ICU in 328 pts with an ICU stay >24 hours. Serum LBP was assayed daily using a commercially available immunoassay system (DPC Biermann, Bad Nauheim, Germany). The ULN for LBP is 12 ug/ml. .0)ug/ml; p<0.0001 for differences between SIRS and sevSep-SS; p<0.008 between sepsis and sevSep-SS. Differences in procalcitonin (PCT) levels were significant between all categories (p<0.0001), C-reactive protein (CRP) levels differ only between SIRS and sevSep-SS. Median LBP levels in pts with a max total SOFA Score of 1-6 points were 23.8 (3.9-109.7) mg/ml, compared to 29.5 (3.8-134.0) in pts with SOFA Scores of 7-12, 41.5 (3.8-200.0) in pts with 13-18 points and 45.16 (14-204) in pts with 19-24 points thereby far beyond the ULN. The most pronounced association with increases of SOFA were found for PCT and IL-6 levels, (p<0.0001). LBP but not CRP levels were higher in non-survivors of sevSepsis-SS, than in survivors (p<0.001). Max PCT levels correlated most with sepsis-related death (p<0.00001). The AUC for LBP to discriminate SIRS from sevSep-SS was 0.53, comparable to C-reactive protein (0.55). The accuracy was highest for serum procalcitonin (0.80) and platelet count (0.71).

CONCLUSION. LBP serum levels are found to be elevated in nearly all critically ill pts and are not restricted to infectious etiologies. The acute phase response mediated by LBP seems to be associated more strongly to the severity of organ dysfunction than the response mediated by CRP, whereas an increase of PCT levels indicates more specifically the degree of organ dysfunction and prognosis in pts with sevSep-SS. Grant acknowledgement: Thuringian Government, grant B-309-00014

Surbatovic M 1 , Jovanovic K 1 , Romic P 1 , Filipovic N 1 1 Clinic of anesthesiology and intensive therapy, Military Medical Academy, Belgrade, Yugoslavia

Severe trauma and sepsis are the major sources of morbidity and mortality despite the rapid development of intensive therapy. Studies have indicated that there are marked alterations in immune response in patients exposed to major trauma or prolonged surgical procedures, including altered pro-and anti-inflammatory mediator/cytokine release (1). Traumatic injury results in profound immunosuppression which predisposes the patients to sepsis and/or multiple organ dysfunction syndrome (MODS). Aim of this study was to assess the prognostic value of anti-inflammatory cytokines: interleukin (IL)-1 receptor antagonist (IL-1ra) , IL-4, IL-10 and transforming growth factor (TGF)-beta 1 regarding severity and outcome in patients with blast trauma complicated with secondary sepsis.

Thirty patients with blast trauma (air-strikes and bomb explosions) who developed sepsis, 15 patients with blast trauma without sepsis and 15 patients with severe sepsis were enrolled in this study. Forty-seven patients developed MODS), 42 died. Blood was drown on the day 1, 3 and 5. Concentrations of IL-1ra, IL-4, IL-10 and TGF-beta 1 were determined in plasma using ELISA assays.

When compared MODS group with group without MODS, we found significant difference (p<0.01) in IL-1ra and IL-10 concentrations; mean values of IL-1ra were 8-fold and IL-10 90-fold higher in MODS patients. When compared non-survivors with survivors, we found significant difference (p<0.01) in IL-1ra and IL-10 concentrations; mean values of IL-1ra were 5.3-fold and IL-10 8.4-fold higher in non-survivors; When compared trauma+sepsis group with trauma group, we found significant difference (p<0.01) in IL-1ra and IL-10 concentrations, they were higher in trauma+sepsis group (IL-1ra 7.7-fold, IL-10 53-fold). IL-4 and TGF-beta 1 were not different (p>0.05) in any case.

CONCLUSION. IL-1ra and IL-10 are excellent predictors of severity and outcome of severe trauma and sepsis; higher concentrations were found in MODS group and in non-survivors. IL-4 and TGF-beta 1 had no significance as predictors of severity and outcome what so ever. 

Early identification of patients with systemic inflammatory response syndrome (SIRS) who are at increased risk for sepsis would permit earlier intervention and potentially result in more favourable outcomes. However, early detection of sepsis is problematic and the disease may be relatively advanced before sepsis is clinically suspected. Highly specific markers that distinguish patients with SIRS who will or will not develop sepsis are therefore particularly desirable.

Fifty surgical intensive care unit (SICU) patients presenting with SIRS were studied. Most patients were admitted for cardiovascular or peripheral vascular surgery, or as a result of trauma. Cell surface HLA-DR on monocytes as well as CD11b and CD64 on polymorphonuclear cells (PMN) were assessed daily by flow cytometry with QuantiBRITE quantisation. Plasma levels of IL-6, IL-10, CRP, and procalcitonin were assessed by ELISA or luminometric assays. Patients were retrospectively classified based on whether they did (Group 2) or did not (Group 1) develop sepsis over a maximum 14-day observation period. Markers with premonitory value on a specific study day were identified by comparing data from Group 1 patients with data from preseptic Group 2 patients. RESULTS. PCT levels were significantly higher in patients with sepsis (3,81+6,19) , severe sepsis (33,21+59,5) and septic shock (84+163,93) than those of SIRS (1,39+1,50). There was no correlation between PCT and age, IL6, TNF, WBC, iCa , but significant correlation was found between PCT , CRP (r=0,471 p<0,001) and, APACHE II score (r=0,275 p=0,01) There was no significant difference in PCT levels between survivors and nonsurvivors and also between culture positive and negative patients in sepsis, severe sepsis and septic shock patients.

Increase in plasma levels of PCT are closely associated with the severity and type of systemic inflammation.

Blairon L F 1 , Wittebole X 1 , Laterre P 1 1 Intensive Care and Emergency Medicine, St-Luc University Hospital, Brussels, Belgium INTRODUCTION. Procalcitonin (PCT) has often been considered as a good marker of sepsis, with a higher specificity than CRP. In this study, we have evaluated PCT serum level as a predictor of outcome in critically ill patients (pts).

All consecutive pts admitted in a 7-bed ICU for 5 months were included in the study. Pts have been sorted in 4 groups: severe sepsis (group A), medical, non infected pts (group B), "long" surgery (>6h) (group C) and "short" surgery pts (<6h) (group D). CRP, PCT, LBP, and IL-6 serum levels have been determined on the day severe sepsis criteria were met (group A), or on the admission day (groups B, C, D). Markers serum values have been displayed as median (SD). PCT is a sensitive and a specific marker for prediction of sepsis, as previously described, but also a better predictor of mortality than CRP and LBP in critically ill patients. PCT serum level could guide clinicians to detect patients with severe sepsis at high risk of mortality.

Boelke E 1 , Orth K 2 , Gruenert A 3 , Steinbach G 3 1 Thoracic and Vascular Surgery, University of Ulm, Ulm, 2 General Surgery, Community Hospital, Emden, 3 Clinical Chemistry, University of Ulm, Ulm, Germany

Setting: Procalcitonin is a 116-aminoacids propeptide of calcitonin and has been found in high concentrations in septic conditions. With a low molecular weight of 13600 Dalton procalcitonin is ultrafiltrable. Therefore procalcitonin elimination could be impaired and levels might be elevated in renal insufficiency.

We determined procalcitonin and C-Reactive-Protein in patients with different degrees and treatments of renal insufficiency RESULTS. Procalcitonin in serum (reference value in healthy controls: <1microg/l) is moderately elevated in patients with continuous ambulatory peritoneal dialysis (median of 1.18 microg/l). Short and long term-hemodialysis is not associated with elevated procalcitonin levels (median of 0.25 microg/l and 0.61microg/l) and does not limit the use of procalcitonin for the diagnosis of infections. In contrast, C-Reactive-Protein (reference value under healthy conditions: <5mg/l) is markedly increased in patients undergoing short-and longterm haemodialysis patients (medians of 14.5 and 51.1 mg/l).

Continuous Ambulatory Peritoneal Dialysis influences serum procalcitonin levels; however, the underlying mechanisms remain unknown

Brunkhorst F M 1 , Engel C 2 , Russwurm S 3 , Wahlers T 4 , Reinhart K 1 1 Anaesthesiology and Intensive Care Medicine, Friedrich-Schiller-University, Jena, 2 KKSL, University Leipzig, Leipzig, 3 Director, SIRS-Lab, 4 Cardiothoracic and Vascular Surgery, Friedrich-Schiller-University, Jena, Germany INTRODUCTION. Procalcitonin (PCT) levels have been described as being superior to Creactive protein (CRP) in the diagnosis and monitoring of patients with severe sepsis and septic shock (sevSep-SS). However, the diagnostic accuracy of PCT may be limited in the early course of pts undergoing CPB due to non-specific PCT-release.

We performed a 21-months prospective study in a 28-bed surgical ICU in 996 critically ill patients for 9275 days, 504 (50.6%) of whom had undergone CPB previously. Procalcitonin (PCT) serum concentrations were measured in daily routine using a commercially available assay (Brahms AG, Berlin, Germany). The LLN for PCT is 0.1 ng/ml in healthy humans, and <1.0 ng/ml for critically ill patients.

Severe sepsis(SS) is associated with a significant cost. The cost of SS may vary from one region to another, is often estimated and influenced by the way healthcare is financed. Hospital costs of SS patients (pts) were reviewed from one academic and one nonacademic hospital.

Individual hospital bills from pts admitted to the ICU with SS, and an expected survival > 3 months, were collected. Cost was determined for the entire hospital stay, from the day SS was diagnozed till discharge. Total cost, hotel, medical honorarium, pharmacy, radiology, antimicrobials, dialysis and blood products costs were analyzed. Pts were divided in 3 groups based on the primary source of SS (CAP, Intra-abdominal (IAI), Other 

Our aim was to elucidate how we used our resources on patients who died in the ICU. We studied length of stay (LOS), severity of illness, therapy limitations (withheld/ withdrawn), ressource usage.

Retrospective study of 236 patients who died over a 44 months period, median age 68 (range 3-91), 62% men, 56% unscheduled and 20% scheduled surgery, 24% medical. Severity of illness judged by Simplified Acute Physiology Score, SAPS II. Resources measured by Nine Equivalents of nursing Manpower use Score, NEMS.

The 73% who died within one week used 25% of the resources. The 6% who stayed five weeks or longer used 30% of the resources. SAPS were greater than 45 for most patients. Median LOS was 3 days, maximum 71 days. NEMS/patient/day did not change with LOS. Therapy limitations were used on 50% of the patients (33% withdrawn, 17% withheld). 72% of the limitations were used on patients who died within one week. Limitations were used on 11 of the 13 patients with the longest LOS. CONCLUSION. All patients who died in the ICU were severely ill the first day (SAPS II). Most resources and therapy limitations were used on patients who died within one week. The use of resources per patient per day was independent of LOS. The 6% with LOS of five weeks or more used 30% of the total resources. Reliable predictors for death should be searched for as they could enable us to earlier limit futile therapy and thus lessen the burden for patients and relatives.

Several commonly used classification methods exist for assessing critically ill patients and have been found to correlate well with observed outcome (e.g. ICU and/or hospital mortality). Few investigators have examined how well these scoring systems explain variation in ICU costs [1]. We investigated the relationship between three measures of severity of illness based on daily assessment: I) total maximum SOFA score; II) total SAPS II score and III) total SOFA score in a sample of high cost patients. 

Clark J R 1 , Hutchinson S P 1 1 Critical Care Unit, Royal Hallamshire Hospital, Sheffield, United Kingdom

Over the past twelve months the number of requests for assistance from other specialties in sitting central venous catheters (CVCs) has increased dramatically. This may be related to the recommendation of the National Institute for Clinical Excellence (UK)_ that all elective CVCs should be sited under 2D Ultrasound guidance. Our aim was to make an assessment of this workload and the cost to our specialty.

We carried out a prospective survey over a period of three months to examine i) the number of requests for CVC insertion made to critical care staff from other specialties, ii) the reason for these requests, iii) the specialty and iv) the time the request was made. A costing calculation for this additional workload describes, i) a costing for each insertion of a CVC, ii) time spent sitting CVCs for other specialties and iii) overall cost to the department. A multi-step risk-adjustment process was planned. ICU volumes corresponding both to overall (Odds Ratio -OR 0.966) and to 3,838 high-risk (OR 0.830) pts were both negatively correlated with mortality of the whole patient population. The relative mortality decreased by 3.4% for every 5 pts per bed per year total volume increase and by 17.0% with the same increase in high-risk volume. The model identified and adjusted for several explicative variables of hospital mortality (demographic, clinical characteristics, process of care, regional, hospital and ICU characteristics). We found a direct relationship between mortality and ICU occupancy-rate over 80% (OR 1.324 and 1.351 respectively).

A measure of volume is suitable also in an environment with high variability like Intensive Care Medicine. Intensive care pts, whatever their level of risk, are best treated where many high-risk pts are treated (at least > 13 per bed per year). Moreover, the higher the ICU occupancy-rate, the greater mortality is. They showed that patients resuscitated with HES (Haesteril 10%, Fresenius, France) had a significantly higher incidence of acute renal failure (defined as a doubling in the serum creatinine concentration), and higher peak creatinine concentrations. As a result of this paper, we made the decision to switch from using HES as our standard resuscitation fluid, to using modified fluid gelatin (Gelofusine, B. Braun, Germany).

Usage and cost data for both gelatins and HES were obtained from our pharmacy information system for the year preceding, and the two years following the publication of Schortgen's paper. Demographic and case-mix data were obtained from our ICU information system. (14), or other major procedures (6). cTnI was measured in the morning of postoperative days 1 and 2. In accordance to the manufacturer of the test kit, a cTnI value of > 2.0 mcg/l was considered as elevated. All-cause mortality was evaluated by follow-up interviews after 30 days, and after 12 and 24 months. Logistic regression was used for data analysis. 

Data from 2800 patients hospitalized more than 48 hours in nine ICUs were prospectively collected (1997-2001) for the OUTCOMEREA® database. 984 patients met the inclusion criteria: normal platelet counts at admission (>150000/mm3), and ICU stay>6 days. Six thresholds of platelet drops on day 4 were tested in univariate and multivariate analysis, from -10% to -60% of the highest count obtained on day 0 or 1, by 10% steps.

In univariate analysis, any platelet drop, whatever the threshold considered (even 10%) is associated with death in ICU, (p< 0.001 in all groups), as well as the absolute platelet count on day 4, the SAPS II score, the presence of a chronic disease as defined by Knaus, admission for severe sepsis, and transfer from another hospital or ward (as opposed to admission from home or emergency room). In multivariate analysis, four variables are independently associated with death : SAPS II score ( 

Among 44,194 ICU admissions from 1994 through 2002, 15,063 who stayed in the ICU for 3 or more days were included in the study. Race, gender, first and third day Acute Physiology Score (APS) and probability of hospital death, and actual hospital mortality were obtained. The area under the receiver operating characteristic curve (AUC) and Hosmer-Lemeshow statistic were used to determine the performance of the mortality predictions.

Most of the patients (96%) were Caucasians; 58% male. The first and third ICU day mean predicted mortality rates were 16.5% and 23%, respectively. The observed hospital mortality rate was 15.1%. The differences between survivors and non-survivors are listed in Table 1 . The AUC was 0.786 (95% CI, 0.780 -0.793) for the first day predicted mortality compared to 0.845 (95% CI, 0.839 -0.850) for the third day predicted mortality (P < 0.0001). The calibration for both predictions was poor (P < 0.0001) with Hosmer-Lemeshow statistic of 292 for the first and 213 for the third day predictions. 

Samalavicius R S 1 , Misiuriene I 1 , Norkiene I 1 , Juozaitis M 1 1 II department of anesthesia, Vilnius University Hospital Santariskiu Clinics, Vilnius, Lithuania

The aim of the study was to determine patient characteristics and operative and postoperative variables responsible for the need of prolonged mechanical ventilation following coronary artery bypass grafting.

We retrospectively reviewed the records of 1428 consecutive patients who underwent CABG at our institution between Jan 1, 2000 and Jun 31, 2002. Patients were assigned to two groups (postoperative mechanical ventilation < 12 h and > 12 h). Demographic data, risk factors, data of operative and postoperative period were collected and compared between the groups. CONCLUSION. According to our data both patient preoperative data and surgery related factors affect the lengh of postoperative ventilation. Identification of risk factors for prolonged mechanical ventilation can help in designing preemptive strategies for improving outcomes in these patients. 

Prospective observational study of 1187 cardiac surgery patients in a tertiary referral centre. Probabilities of hospital death for patients were estimated by applying the four models and were compared with actual mortality rates. Performance of the four systems was assessed by evaluating calibration with the Hosmer-Lemeshow goodness-of-fit test, and discrimination with receiver operating characteristic (ROC) curve. CONCLUSION. In our experience several cardiac surgery risk models developed in North America perform well. Higgins and Parsonnet models demonstrate to be appropriate tools to assess mortality in this set of patients, with excellent calibration and discrimination, better than O'Connor and Tu models. Predictive models originally developed in another country should be validated in the population to which they are finally applied.

McCaffrey J 1 , Symington S 1 , McCarthy G 1 , McLoughlin C 1 1 Anaesthesia, Belfast City Hospital, Belfast, United Kingdom

A controlled trial in mostly postoperative cardiac patients, has demonstrated that control of blood sugar to below 6.1 mmol/l using insulin and carbohydrate improves survival in those staying longer than 5 days(1). In our non-cardiac ICU, however, the trigger for insulin therapy is 10 mmol/l. Hypothesis: There may be an increased mortality in all patients with poorer control of blood sugar using the 10mmol/l threshold

This was a retrospective case review of 154 consecutive admissions of more than 5 days to an adult general ICU over 1 year. Patient characteristics, type of admission, diabetic history, serial blood glucose (BM) measurements to unit discharge, and unit outcome were recorded. The area under the serial measurements polygon of glucose against date/time was measured for each patient and adjusted for the duration of stay (AUC) to give a summary value. Patient characteristics, AUC values for medical and surgical survivors and non-survivors were then compared using t-tests, Wilcoxon, or ƒÓ-squared testing as appropriate

The main results are in table 1. There were 79 medical and 75 surgical patients with 14 and 12 deaths respectively. There were no differences in the age, sex distribution, or proportion of diabetics or mean duration of stay between survivors vs non-survivors in either group. BM values were similar in surgical (7.35, 1.12 vs 6.8, 1.33) but not medical (7.67, 1.3 vs 8.55*, 1.2) (mean, SD) survivors and non-survivors. The AUC for medical patients also showed a significantly greater value in non-survivors.

AUC values are Median with 95% CI, duration of stay in days with SD Hosptial

Duration/stay Duration/stay Survivors 6.64(6.5-7.45) 6.8(6.32-7.64) 13.8(9.31) 11.7(9.3) Non-Survivors 8.52*(6.99-9.78) 6.9(5.83-9.83) 9.2(5.67) 16.6(7.5) . * Denotes significance at the 5% level CONCLUSION. We were able to demonstrate a mortality rise in long stay medical patients depending on BM despite using our current treatment trigger of 10mmol/l. A finding not repeated in our general surgical patients. 

Soares M 1 , Rocco J R 2 1 Intensive Care Unit, Instituto Nacional de C, 2 Intensive Care Unit, Hospital Universitário Clementino Fraga Filho -UFRJ, Rio de Janeiro, Brazil

Cancer patients requiring mechanical ventilation (MV) have a high mortality. Our objective was to identify factors related to in-hospital death in patients receiving MV for at least 24h admitted to the intensive care unit (ICU) at the Instituto Nacional de Câncer, Brazil (INCA).

Data were prospectively collected during the first 24h of ICU stay and assessed by univariate and multivariate analysis. Patients were stratified based on type of admission, main reasons for MV and cancer diagnosis. In multivariate analysis, variables were expressed as odds ratio (CI 95%, p-value).

From 1216 ICU patients, 443(36.4%) required MV. ICU and hospital mortality rates were, respectively, 19.7% and 29.0% for overall patients, and 49.7% and 63.0%, for patients receiving MV. Mean age was 56.9±17.4 years. There were 282(63.7%) medical admissions and 99(22.3%) patients had haematologic malignancies. SOFA score 1 was 7.8±3.9 points and predicted mortality assessed by Cancer Mortality Model 2 was 73.3%±20.6%. The main reasons for MV were: sepsis (57.8%); ALI/ARDS (35.2%); postoperative recovery (13.5%); central nervous system involvement (11.5%); advanced tumour (9.9%); cardiac arrest (7.4%); and pulmonary embolism (4.5%). In septic patients, pneumonia was the source of infection in 58.6%. In multivariate analysis, the following variables were selected: age (years)=1. CONCLUSION. Both factors related to cancer status and acute illness were associated with poor outcome in our patients. We believe knowledge of such easily available clinical features can help health professionals in providing a better use of ICU resources, reducing inappropriate aggressive support, such as MV. 

Martínez-Alario J 1 , Villegas J 1 , Diaz De Tuesta I 2 , Mora M 1 , Malaga J 1 , Galvan R 1 , Garcia C 1 1 Medicina Intensiva, 2 Cirugia Cardiaca, Hospital Universitario de Canarias, Tenerife, Spain INTRODUCTION. The aim of this work was to assess the mortality in septic shock patients and the performance of Multiple Organ Dysfunction (MOD) score, Sepsis-related Organ Failure Assessment (SOFA) score, and Logistic Organ Dysfunction System (LODS) score, to predict this mortality.

Prospective observational study of 254 patients who meet the criteria for septic shock defined by the American College of Chest Physicians and Society of critical Care Medicine Consensus Conference, in a tertiary referral centre. Predicted probabilities of hospital death for patients were estimated by applying the three models and were compared with actual mortality rates. Performance was assessed by evaluating calibration with the Hosmer-Lemeshow goodnessof-fit test, and discrimination with receiver operating characteristic (ROC) curve. CONCLUSION. In our experience MOD, SOFA and LODS scores perform very well, with calibration and discrimination very high, and they are appropriate tools to assess mortality in septic shock patients.

Alberti C 1 , Brun Buisson C 2 , Chevret S 3 , Le Gall J 4 , For the European Sepsis Group 5 1 Unité d'épidémiologie clinique, Hopital Robert Debré, Paris, 2 Intensive Care Unit, Hopital Henri Mondor, Créteil, 3 Service de Biostatistiques, 4 Intensive Care Unit, Hopital Saint Louis, Paris, 5 ICUs, Worldwide, Worldwide, France INTRODUCTION. Criteria for the systemic inflammatory response syndrome (SIRS) are not specific enough to identify infected patients at risk of worsening to severe sepsis or shock. The objective of this work was to examine the incidence of and variables associated with progression of infected patients to a more severe sepsis stage and to propose a new objective definition of SIRS variables METHODS. Prospective inception cohort of patients (n=3,443) admitted to 28 intensive care units in academic centers throughout Europe during a 1-year period, and having a first infection at admission or during the ICU stay. Main outcomes measured were thirty-day cumulative incidences of severe sepsis or septic shock and hospital mortality RESULTS. The cumulative incidence of progression from infection or sepsis (n=1531 patients) to severe sepsis or shock -accounting for the competitive risk of death without worsening -was 22% at day-14 and 24% at day-30. Twelve variables were independently associated with worsening: temperature (>38.2°C), heart rate (>120/min), systolic blood pressure (<110 mHg), platelet count (<150.109/L), serum sodium (>145 mMol/L), bilirubin (>30Mol/L), mechanical ventilation, and 5 variables characterizing infection (pneumonia, peritonitis, primary bacteraemia, and infection with gram positive cocci or aerobic gram negative bacilli). The risk of progression from infection to severe sepsis score (RISSC) developed from these variables ranged from 0 to 49, and four classes of risk, ranging from 9% to 55%, were defined CONCLUSION. Using objectively defined criteria for SIRS and impending organ dysfunction should help physicians stratify patients according to their level of risk, to design management strategies or testing new interventions 

Highly Active Antiretroviral Therapy (HAART) has produced a significant decrease in mortality and morbidity from Human Immunodeficiency Virus infection. Whether this therapy resulted in changes in outcome in HIV-infected patients admitted to ICU is still controversial (1).

We reviewed the clinical notes of HIV positive patients admitted to our ICU from 2000 through 2002. On each patient the following was collected (i) demographics (ii) admission diagnosis, (iii) 1st 24 hours APACHE II score, (iv) HIV infection epidemiological and clinical characteristics. Statistical analyses were performed using SPSS (SPSS inc. Chicago IL). For continuous variables Mann-Whitney non-parametric test was used, whereas Chi-squared test for categorical ones. A Binary Logistic Regression Model performed to estimate the effect of each considered risk factor on a death (yes/no) outcome for HIV-patients. P values < 0.05 were considered statistically significant.

There were a total of 54 admissions for 51 HIV-infected patients (76.5% males) with a median (IQR) 1st 24 hour APACHE II score of 21(13-24) and ICU stay of 6 days (2-12).Drug abuser was the main risk factor (51%), while respiratory failure (58.8%) the most frequent reason of ICU admission. Median CD4+ value was 150/mmc (IQR 30-400) median viral load was 52,000 (IQR 1,100-207,000). A Binary Logistic Regression Model was used, considering HIV-infected patient ICU outcome (dead/alive) as the categorical dependent variable and (i) presence/or not of organ failure (kidney, liver, CNS, heart and respiratory system); (ii) taking HAART/or not; (iii) CD4+ absolute value < or > 200/mmc; (iv) HIV status (HIV-positive/AIDS), as the independent ones. We found that taking HAART (OR = 2.7, 95% CI: 1.4-3.2, p = 0.0242)is a predictive factor of positive ICU outcome of HIV-infected patients.

In summary, our data, although retrospectively collected, suggest that a HIV patient on HAART, admitted to ICU has a significantly higher chance of survival if compared with those who are not. Future prospective studies should be designed to validate both our observation and whether starting HAART in a HIV positive critically ill patient improve his survival. 

Van der Spoel J I 1 , Bosman R J 1 , Oudemans-van Straaten H M 1 , Wester J P J 1 , Zandstra D F 1 1 ICU, Onze Lieve Vrouw Gasthuis, Amsterdam, Netherlands INTRODUCTION. It is "common wisdom" that the number of infusion pumps correlates with severity of illness and outcome. These relations were investigated.

Daily SOFA scores and maximum number of infusion pumps were registered in all patients between 1/1/99 and 1/1/2003 with a length of stay >24 h. Cumulative number of pumps and cumulative SOFA scores were computed for the total ICU stay. Data are expressed as mean (SD). CONCLUSION. Non-survivors had both higher daily and higher cumulative numbers of pumps and SOFA scores, and there is a good correlation between them. However, the number of pumps nor the SOFA score is indicative for outcome in the individual patient.

Hofhuis J 1 , Rommes H 1 , Spronk P 1 , Bakker J 2 1 Intensive Care, Gelre Hospital Location Lukas, Apeldoorn, 2 Intensive Care, Isala Hospital Location Weezenlanden, Zwolle, Netherlands

Survival is the primary goal of intensive care. As severe sepsis remains a prominent cause of mortality, morbidity and costs we studied the impact of severe sepsis on long term survival compared with non-sepsis patients (1).

From September 2000-January 2003, patients admitted to the ICU for > 48 hours were eligible for inclusion. Survival was followed at discharge ICU, discharge Hospital and 3 and 6 months after discharge ICU. Severe sepsis was defined by the presence or highly likely suspicion of infection, two or more SIRS criteria and one or more dysfunctioning organ systems. Severity of disease was measured by APACHE II. Survival was analysed by Kaplan-meyer survival statistics and by the use of the log-rank test to study differences between the two groups.

. 339 patients of whom 111 met the criteria for severe sepsis were included in the study. Eighteen patients were lost to follow up. Patients with severe sepsis (age 67 ± 12.9 ) stayed in the ICU longer (18.7 ± 11.3 p=0.1) than patients without sepsis (age 69.3 ± 11.7) Small differences were found in APACHE II scores of sepsis and non-sepsis patients (21.2 vs 18.3 respectively). Six months after discharge ICU, survival of sepsis patients was lower compared with non-sepsis patients, but this result was not significant (p = 0.15). 

Andrews F J 1 , Mostafa S M 2 1 Emergency Medicine, 2 Intensive Care, Royal Liverpool University Hospital, Liverpool, United Kingdom

Scoring systems predict mortality in critically ill patients but none show the relative importance of their components. Studies on febrile medical patients [1] have previously demonstrated that some components of SIRS predicted mortality more accurately than the presence of SIRS or sepsis itself. We evaluated the criteria for Systemic Inflammatory Response Syndrome (SIRS) and sepsis [2] combined with infection site and organ dysfunction for mortality prediction in patients admitted to our Intensive Care Unit (ICU).

This study was performed prospectively in a university teaching hospital general intensive care unit. Two hundred and thirty patients were included who met at least three of the following criteria: (1) Expected survival for at least 24 hours, (2) Expected stay on ICU for at least 72 hours, (3) evidence of inflammation or infection and (4) currently receiving antibiotics.

Logistic regression was used to determine which variables significantly contributed to mortality prediction on admission (day 1) and day 3. The predictive values for mortality for patients with SIRS and sepsis were compared with the new model. SIRS was present in 97% of patients and 51.7% had confirmed sepsis. On day 1 only soft tissue infection or coagulation dysfunction significantly predicted death. However, day 3 variables significantly predictive for death were circulatory and respiratory dysfunction, leucocytosis and soft tissue infection. These significant variables on days 1 and 3 predicted mortality better compared to categorical SIRS and sepsis.

CONCLUSION. SIRS and sepsis are poor mortality predictors in a general ICU cohort. The predictive power is markedly increased when logistic regression is used to analyse the variables of SIRS along with infection and organ dysfunction. The derived models demonstrate which variables significantly contribute to mortality. 

In critically ill patients a number of factors may result in altered pharmacokinetics. Potential changes in renal clearance, protein binding and volume of distribution may lead to significant changes in pharmacokinetics. This is important because clinical efficacy is thought to relate to pharmacokinetic parameters. Only one previous study has examined levofloxacin pharmacokinetics in critically ill patients [1].

Ten critically ill adult patients with a normal serum creatinine and a clinical indication for administration of levofloxacin were studied. Levofloxacin 500 mg IV was administered as a 60 minute infusion. Arterial blood samples were taken at 0,60,120, 180, 300, 540 and 720 minutes after the start of the infusion. Levofloxacin concentrations were measured using high pressure liquid chromatography. suggest that values of Cmax and AUC achieved following administration of levofloxacin to critically ill patients are at least as high as those achieved in less severely ill patients and in human volunteers. These data suggest that there is no need for an increased dose to ensure efficacy in critically ill patients. 

High percentages of patients received 3rd generation cephalosporins, glycopeptides, fluoroquinolones and metronidazole, presented underlying diseases or had a magor surgery. Careful use of antimicrobials in ICU patients with major problems and a long ICU stay could probably protect patients from the emergence of such multi -resistant strains.

Palomar M 1 1 ICU, H Vall d´Hebron, Barcelona, Spain

Antibiotic (ATB) policy in an individual unit is influenced by patients characteristics and epidemiological factors that can change overtime. Antibiotic pressure depends on the spectrum and also the duration of the therapy.

The aim of our study was to assess the pattern of ATB prescriptions and the evolution in a 5 years period. Prospective study 2 months/year from 1998 to 2002 in a 32 beds medical-surgical ICU in a reference hospital. In all the patients staying>1day, the administered antibiotics were recorded. 

Routine microbial surveillance cultures are a cost-intensive part of the strategy of selective digestive decontamination (SDD). This microbiological monitoring is regarded as necessary for early identification of pathogens responsible for pneumonia, but also provides information on colonization of patients and changes of bacterial resistance during prophylactic antimicrobial therapy. This study was conducted to evaluate the value of routine microbial investigations performed during SDD on a university surgical intensive care unit (ICU).

Data of 786 patients admitted to the surgical ICU for more than 24 hours during a two-years period were analyzed. Pneumonia was diagnosed according to Centers of Disease Control guidelines. Bacterial and mycological cultures of tracheal aspirates, nasal, pharyngeal and rectal swabs were collected on admission and then twice weekly. Both topical (polymyxine, tobramycine, amphotericine B) and systemic (cefotaxime) antimicrobial agents were routinely administered to ventilated patients who were expected to remain in the ICU for more than 24 hours.

A total of 75 episodes of pneumonia were diagnosed in 69 patients. Pathogenic organisms were isolated in 56 episodes, but only in 10 episodes the pathogen was isolated due to routine cultures prior to the onset of pneumonia. An almost complete picture of bacterial colonization was provided when bacterial monitoring was restricted to regular tracheal aspirates in combination with a pharyngeal swab on admission. Candida was isolated from tracheal aspirates in 27% of patients and from pharyngeal swabs in 49% of patients. Only three episodes of pneumonia due to Candida spp. were diagnosed.

CONCLUSION. Sufficient bacterial surveillance during SDD could have been provided by routine tracheal aspirates in conjunction with a pharyngeal swab on admission. Mycological cultures -for its low positive predictive values -can be restricted to high risk patients and to diagnostic cultures during infections.

Montravers P 1 , Lepape A 2 , Gauzit R 1 , Martin C 3 1 DAR, CHU Jean Verdier, Bondy, 2 DAR, Lyon Sud, Pierre Benite, 3 DAR, CHU Nord, Marseille, France

Frequency, timing and causes of the changes of antibiotic therapy (AB) in the course of intra-abdominal infections have been minimally evaluated, both in community-acquired (CIA) or nosocomial infections (NIA). The aim of the current study was to assess the AB changes and their link with clinical success.

In a prospective (06/00 to 01/01) multicentric (176 centres) study, all the patients (pts) undergoing surgery for intra-abdominal infections (except postoperative peritonitis) were evaluated. Clinical and microbiological parameters were collected. Changes and duration of AB were decided by the attending physician. The consequences of AB changes performed before or after identification (ID) of the pathogens were assessed on the outcome. The efficacy of AB was evaluated over 30-days of follow-up. Chi-2 test and ANOVA were used, p<0.05 significant.

Among the 761 CIA pts and 247 NIA pts, AB changes were performed in 304 (47%) and 123 (62%) pts respectively (p<0.01). AB was given for 9±15 and 12±18 days in CIA and NIA groups (p<0.01). AB changes were decided in 427 pts (including persistant infection n=88, organisms resistant to initial AB n=101, superinfection n=28). Success was obtained in 92% and 82% of CIA and NIA, with contrasted results (table1). Early AB changes and repeated changes were mostly observed in situations of clinical worsening or absence of improvement. 

Montravers P 1 , Gauzit R 1 , Martin C 2 , Lepape A 3 1 DAR, CHU Jean Verdier, Bondy, 2 DAR, CHU Nord, Marseille, 3 DAR, CHU Lyon Sud, Pierre Benite, France

A large number of antibiotic regimens (AB) have been recommended in the empirical treatment of community-acquired peritonitis (CP)(1). However, none of them have been clearly assessed in open epidemiological studies. The aim of the current study was to describe the type of patients (pts) receiving the most prescribed AB in CP.

In a prospective (06/00 to 01/01) multicentric (176 centers) study, 761 pts undergoing surgery for CP were collected. Demographic parameters, severity criteria and prognosis of the pts receiving one of the four most prescribed AB were analysed: amoxycillin/clavulanic acid (A) alone or combined with aminoglycosides (AA), third generation cephalosporins+imidazoles (CI) alone or combined with aminoglycosides (CIA). AB choices and management were left to the attending physician. Results are expressed as % and compared by Chi-2 test.

. 468 pts (61%) were analysed (table1). Demographic data were similar in all groups (Male 59%, mean age 49±20 years, non underlying disease 92%). Proportions of success were similar in all groups (A:96%, AI:90%, CI:95%, CIA:89%). Pts without any AB changes had similar proportions of success (A:96%, AI:96%, CI:94%, CIA:97%). CONCLUSION. AB selection depends directly on the clinical severity of CP. Although AA, CI and CIA regimens were selected by the french consensus(1), no recommendation has been edited based on clinical severity, an issue probably to take into account. 

Binning A R 1 , Buttigieg M 1 , Carmichael S J 2 , Thomson A H 2 1 Intensive Therapy Unit, 2 Pharmacy Department, Western Infirmary, Glasgow, United Kingdom

Antibiotics that require a prolonged time above MIC can be given by a large loading dose and short dosing intervals to maintain therapeutic levels. Since optimal therapy with vancomycin depends on maintaining serum concentrations above the MIC, it has been suggested that continuous intravenous infusion may be a suitable mode of administration, as it would achieve a flat concentration-time profile. This study evaluates if vancomycin therapy is more effective with the use of intravenous administration by continuous infusion compared to our previous practice of intermittent dosing.

Dosing guidelines for continuous infusions were developed from a simulation study. A one-compartment pharmacokinetic model was used to simulate concentration-time profiles following the administration of a range of loading doses and maintenance infusions to simulated patients.

(1) Dosage regimens that maintained concentrations between 15-25 mg/L were chosen. (2) The characteristics of the first 20 patients treated with vancomycin following the introduction of these guidelines were compared to the previous 25 patients treated with intermittent dosing. Demographic data, diagnosis, required duration of therapy, time to MRSA infection resolution(as assessed by microbiological analysis) and pharmacokinetic data were analysed.

The two groups were comparable in terms of age, M:F ratio, concomitant antibiotics and diagnoses. Both groups received vancomycin for a mean of 8.8 days but more drug concentrations were measured in the continuous group 152 versus 114 in the intermittent group. 110 (72%) of the samples were in the target range of 15-25mg/l for the continuous group but only 57 (50%) of the predicted troughs were in the target range of 5-15mg/l for the intermittent dosing group. The continuous group demonstrated a trend towards faster resolution of MRSA infection by 2 days (p=0.4). No patients in either group experienced toxic side effects.

Vancomycin can be safely and effectively administered by infusion using appropriate guidelines developed from known pharmacokinetic variables. The higher achieved plasma concentrations may be associated with increased efficacy. 

Skeletal musle accounts for 80% of insulin mediated glucose uptake. It seems presumable that muscle wasting prevention by glutamine administration could attenuate insulin resistance in multiple trauma patients.

. 40 multiple trauma patients randomised into 2 groups, both of 20 patients. They were receiving isocaloric nutrition. Control group (C) were receiving 1.5 g of AA/kg bw/24 h (standard parenteral AA solution without glutamine), tested group (AG) were receiving 1.1 g of AA/kg bw/24 h as standard AA solution and 0.4 g/kg bw/24 h of parenteral dipeptide alanylglutamine (AG). Euglycemic hyperinsulinemic clamps were performed on the 4th and 8th day.

On the 4th day, the better insulin-mediated glucose disposal was found in Group AG, but without significant difference from Group C. On the 8th day after injury, insulin sensitivity in Group C was significantly smaller than in Group AG (p < 0,01). General characteristics of subjects at baseline METHODS. 25 patients after acute pancreatitis were enrolled into the study. CCK, gastrin, insulinlike growth factor-1 (IGF-1), and glycaemia levels were estimated in every patient under different types of n utrition: total parenteral nutrition, total enteral nutrition into the jejunum, total enteral nutrition into the stomach, and oral bolus sipping. Oligopeptide and whole protein formulas were used for enteral feeding. CCK was estimated by EIA, gastrin and IGF-1 were estimated by RIA. The results were statistically evaluated by ANOVA.

Administration of oligopeptide formula represented the highest trigger on CCK levels in bolus (sipping) feeding (average 452.9 pg/mL) in comparison with continuous enteral feeding into the stomach (average 390.3 pg/mL), jejunum (average 384.2 pg/mL), and parenteral nutrition (average 379.9 pg/mL). Similarly, administration of whole protein formula represented the highest trigger on CCK levels in sipping (average 498.7 pg/mL) in comparison with continous enteral feeding into the stomach (average 385.5 pg/mL), and jejunum (average 388.9 pg/mL). The results were statistically significant (p<0.05).

The results indicate that either parenteral nutrition, and enteral nutrition with oligopeptide and whole protein formula administered continuously into the stomach or into the jejunum have a very low impact on exocrine pancreatic stimulation in comparison with oral intake. These findings support the hypothesis that use of enteral nutrition even in early stages of acute pancreatitis and in treatment of its late symptomatic complications. 

Rümelin A 1 , Humbert T 1 , Lühker O 1 , Fauth U 1 1 Klinik für Anästhesiologie, Klinikum der Johannes Gutenberg Universität, Mainz, Germany

Reduced ascorbic acid (AA) concentration in plasma in the postoperative period has been well documented (1-4). The underlying reason is not clear but only an altered postoperative AA consumption would justify a perioperative substitution.

The AA consumption was investigated pre-and postoperatively on 18 patients undergoing major maxillofacial surgery. The metabolic clearance (Clmeta) of AA in plasma was calculated on the first postoperative day and compared to the preoperative values subsequent to bolus-injection of AA, 6 mg/kg body weight. Blood specimen were taken before and 5, 15, 30, 45, 60, 90, 120 and 240 minutes after injection. Urine was collected. AA in plasma and urine was analyzed using a high performance liquid chromatographic technique.

The preoperative concentration of Clmeta was 4.0 + 1.9 l/h (mean/SD). On the first postoperative day Clmeta (6.3 + 2.5 l/h (mean/SD)) increased significantly (p<0.001). A dose of approximately 1155 mg/d AA should be necessary to compensate the losses.

There is a significantly increased postoperative metabolism of AA that should be considered for future dosage recommendations in perioperative patients. 

Ulusoy H 1 , Akyol A 1 , Ozen I 1 1 Anesthesiology and Reanimation, Karadeniz Technical University, Trabzon, Turkey

To evaluate the effects of erythromycin, domperidone, and metoclopramide on postpyloric placement of nasoduodenal feeding tubes, this randomized, double-blind, placebocontrolled trial was performed.

Eighty critically ill, mechanically ventilated patients were involved in this study. Patients enrolled were groupped according to the Glasgow Coma Scale as being below eight points (GCS<8) or above eight points (GCS >8). The two groups, GCS<8 or GCS >8 were then randomized independently to receive four different medications via nasoenteric feeding tube; erythromycine (E Group), domperidone (D Group), metoclopramide (M Group), and placebo (water) (P Group) followed by blind placement of a feeding tube .Then patients were given right semi-decubitus and 30°fowler position for an one hour.Tube placement was verified by an abdominal radiograph.

Overall, the rates of postpyloric placement were significantly better in E Group (17/20), D Group (14/20), and M Group (11/20) than P Group (2/20) (p<0.05; p=0.0000). In patients with GCS<8, the success rates with erythromycin (9/10), domperidone (6/10), and metoclopramide (4/10) were significantly higher than with placebo (0/10) (p<0.05 ; p=0.0007). In the GCS >8 group the success rates with erythromycin (8/10), domperidone (8/10), and metoclopramide (7/10) were significantly higher than with placebo (2/10) (p<0.05; p=0.0144).On the otherside, no difference was noticed in the success rates of all these drugs with comatose and noncomatose patients.

CONCLUSION. Erythromycin, domperidone, and metoclopramide are effective in facilitating placement of a nasoenteric feeding tube into the duodenum in both comatose (GCS<8) or noncomatose (GCS >8), mechanic ventilated critically ill patients. . The modified glucose oxidase test monitors aspiration by both glucose enrichment of standard feed and semi-quantitative measurement of glucose concentration in the tracheal secretions, improving both sensitivity and specificity [2] . Using this test we performed a pre-and post-intervention observational study to examine the effect a multidisciplinary educational program supported by a clinical information management system (CIMS)(Centricity QS, GEMS-IT, Annapolis, USA) aimed at reducing the incidence of aspiration of enteral feed in critically ill intubated patients.

There was a 6 month multidisciplinary educational period during which time the CIMS was also updated with fields to prompt the user with regard to semi-recumbence and monitoring of tracheal and pharyngeal secretions. The educational program consisted of formal tutorials and daily ward round reinforcement aimed at improving interventions related to the reduction of pneumonia and pulmonary aspiration. Protocols for measuring gastric residual volumes in fed patients were enforced and feeds withheld or pro-kinetic agents introduced as needed. Pulmonary aspiration of enteral feed was monitored 3-4 times daily. Data collected was analysed using the Fisher's Exact test.

The pre-intervention group (n=25) had an incidence of feed aspiration of 44% and a prevalence of 69/1000 intubation days. The post-intervention group (n=43) had an incidence of feed aspiration of 19% (p<0.05) and a prevalence of 30/1000 intubation days (p<0.05).

With appropriate education and interventions, pulmonary aspiration of enteral feed can be reduced in the critically ill intubated patient. 

The preferred method of nutritional support in critically ill patients is via the enteral feeding route (1). Patients who have repair of ruptured abdominal aortic aneursym (AAA) are particularly prone to gastroparesis, and often require post-pyloric feeding (2) . The aim was to review our enteral feeding practices in patients who have AAA repair.

All patients who had AAA repair and received enteral nutrition during a twelve month period were reviewed retrospectively. In addition, patients who had AAA repair and nasojejunal tube placement in the two years prior to this period were also reviewed.

A total of 59 patients had repair of AAA during the twelve months, 41% as an emergency procedure. Nutrition intervention and review by a dietitian was requested in 36. Enteral tube feeding was commenced in 25 patients and was initiated during their critical care stay (14 nasogastric and 11 nasojejunal). All tubes were placed during surgery and were used on average for 7 days. Four of the patients who were fed via the NG tube had gastroparesis and were referred for endoscopic NJ placement. However, their clinical condition deteriorated prior to placement (< 7 days from operation). In the two years prior to this review 19 procedures were undertaken in 15 patients who failed to tolerate NG feeding post repair of ruptured AAA. The mortality was 80%.

Gastroparesis can be a problem in AAA patients. Nasojejunal feeding tubes should be placed in patients undergoing repair of ruptured AAA, at the time of theatre to allow early enteral feeding. Continued failure to tolerate NG feeding after 7 days often relects the severity of the clinical condition, and endoscopic placement of an NJ should be considered. 

Artificial nutrition (AN) is an important link between the response to injury and recovery in critically ill patients. A hypocaloric start of AN is recommended with a progressive increase over several days after ICU admission. However, an isocaloric start would assure adequate energy supply from admission. We have, therefore, undertaken this prospective study to compare isocaloric and hypocaloric starting of AN in critically ill patients.

On the day after ICU admission (day 1), 41 consecutive, critically ill medical patients were randomised into two groups (group A: n=21, group B: n=20). Study period was 5 days. Energy requirements were calculated as 25 kcal/kg/d. Patients of group A were planed to receive isocaloric supply during the study period. Patients of group B were intended to obtain hypocaloric energy supply (50% on day 1), which was increased progressively until day 3 to isocaloric levels and remained unchanged until day 5. Patients underwent daily metabolic monitoring and the incidence of hyperglycemia as well as the number and causes of interruptions of AN were documented.

Energy supply (kcal/d) was different between the two groups on days 1 and 4 ( CONCLUSION. An isocaloric start of AN did not meet energy requirements of critically ill patients during the first 5 days after admission to a medical ICU, and it was associated with higher complication rates. A hypocaloric beginning of artificial nutrition, therefore, seemed to be more suitable for nutritional therapy in these patients. 

When a critically ill patient is taken into ICU their relatives face many difficulties. We from the ICU feel quite unprepared to deal with the emotional aspects of a relationship with patients and their relatives. In an effort to overcome this problem we tried educating professionals (i.e.ourselves) who are in contact with patients' relatives.

In 2002 the psychiatrist who works in our department organized several 60 minutes meetings with our patients' relatives and us. The families had been chosen because the patient's illness was very serious or because the daily encounters with them had been unsuccessful. During these meetings the psychiatrist led a practical training exercise based on experience. The professional at an arranged time and place, in a harmonious atmosphere, becomes a really active listener and focuses not only on what to say but also on how and when to say it. After each meeting we discussed the results with the psychiatrist who helped us to find the key to a deeper understanding of the relatives.

During this practical training experience we learned how to wait and listen. We noted the importance of gestures and silent pauses, while trying to understand beyond the word actually spoken. There are messages which are hidden behind certain banal remarks or behind the patients' relatives' aggressive behaviour. Though active listening we learned how to find the right moment to break bad news. These meetings led to a great mutual understanding so passing from a formal level to a humane relationship. The improved relationship and greater understanding led us to longer visiting hours and relatives' active part in the patient treatment as soon as condition would allow. Relatives were invited to take an active role in weaning the patient from artificial nutrition. In this way they felt useful and gradually the intimate relationship between family members was renewed after the brusque interruption.

CONCLUSION. This experience showed that you cannot draw up guidelines which would be useful in communicating with families. The relatives would immediately be aware of a formal, distant approach to their suffering. We think that relationship with the patients' family can only be improved though real listening and a real acceptance of their suffering. The changes benefit not only the relatives but also the patients. In fact they improve life and well-being for everyone involved.

Howitt R 1 1 Project Worker, UK Transplant, Leeds, United Kingdom

A year long project was undertaken to research the UK's major religions' perspectives on organ donation. A multifaith conference in the UK in 2000 showed that despite religious leaders' public and strong consensus in favour of organ donation, this message was not being disseminated into the public arena and into some areas of health care.

The project looked at various religious issues and enlisted the help of religious authorities, both UK and worldwide, in producing a variety of educational material. The response was very positive and a wealth of resources developed. Written leaflets and audio cassettes have been produced about organ donation and religious views of the Jewish, Buddhist, Sikh, Hindu, Christian and Islamic faiths. Important issues such as life after death and the body are discussed and various languages used. Information has been distributed to ICUs nationally with particular emphasis on areas with large ethnic minority populations. The purpose is to inform staff and to support potential donor families for whom religious opinion is important. Extra material and information is available on UKT's website.

RESULTS. An initial survey of ICU staff had reported a definite need for staff education about religious perspectives. A teaching pack was also produced for transplant co-ordinators and donor liaison nurses to use in teaching programmes. A supportive CD Rom is also under development.

The project was successful in getting information into the health care setting for staff and relatives. As a consequence to the religious authorities involvement plans are now underway for both a chaplaincy conference and a Muslim religious leaders conference to continue increasing awareness and discussion of organ donation.

Plasma levels of natriuretic peptides may serve as important markers for acute coronary syndromes and left ventricular performance. The present study was designed to determine whether systematic heart rate (HR) control achieved by continuous perioperative administration of the short-acting beta-blocker esmolol might influence on perioperative cardiac neurohormonal activation and myocardial ischemia.

With approval of the local ethics committee 34 patients scheduled for elective abdominal aortic surgery were randomly assigned to receive strict HR control (n=17; esmolol) to a target HR of 20% below baseline measurements by continuous perioperative treatment with esmolol or standard therapy (n=17; control). HR control began after induction of anesthesia and continued for 48 hours thereafter. Monitored variables included invasive arterial and pulmonary artery pressures, cardiac index, and automated ST-segment analysis of leads II and V5 of the electrocardiogram. Plasma samples were analyzed for atrial (ANP) and brain (BNP) natriuretic peptide, and cardiac troponin T (TnT) after induction of anaesthesia, 20 minutes after aortic clamping, after declamping, at the end of surgery, and in the morning of the first and second postoperative day.

There was no incidence of esmolol-related adverse events such as congestive heart failure, bronchospasm, or prolonged haemodynamic instability. Target 

We assessed the native cardiac function in a patient with dilatative cardiomyopathy who was under mechanical assistance with a Novacor -LVAD (Baxter Health Care Corp.) for six months. The patient was intubated for an abdominal wall repair, and postoperative, we tested him for a potential weaning from Novacor.

The assessment of the adequacy of the native cardiac function was done with the NICO 2 device (Novametrix). The weaning test was done by switching the pump operating mode from fill rate mode to fixed rate mode(FR)at 70 beats min -1 ,lowered than with 10 beats min -1 at every 30 minutes, down to 30 beats min -1 .

While the left ventricle was totally assisted (fill rate mode) by the Novacor, the cardiac output measured by NICO 2 (CO 2 CO = native CO + pump output) was higher than the pump output (PO) because of the still existing native cardiac output, echocardiographic revealed by the native aortic valves openings. During the weaning protocol we observed that the CO 2 CO remain quite stable, despite the diminishing of the PO. (see table) . That lead us to the conclusion that the patient native cardiac function had recovered enough and we could start the conventional weaning protocol. 

Manolakoglou N D 1 , Papadopoulou N N 1 , Hatjis K D 1 , Dourgiou D A 1 , Stanopoulos I T 1 1 Respiratory failure unit aristotle universtity, general hospital g. Papanikolaou, Thessaloniki, Greece

The aim of our study was to pharmaceutically reduce preload and afterload in a group of difficult to wean MICU patients (pts) during transition from mechanical ventilation (MV) to spontaneous breathing.

METHODS. 8 male pts were studied. All of them were catheterized with a Swan Ganz catheter, after following at least 2 unsuccessful T-piece trials on 2 consecutive days. Hemodynamic and blood-gas parameters were monitored before (pressure-support mode) and during the first 10' of the T-piece trial. In 4 pts (2 COPD with CHF and 2 IPF with Cor Pulmonale -called N group) an IV infusion of nitroglycerin was started, titrated to maintain a normal blood pressure, while in 4 pts (4 COPD with CHF), a 24-hour infusion of levosimendan was started (L group). Measurements were repeated in 60' of the T-piece trial. 

Mas Font S 1 , Bisbal E 1 , Abizanda R 1 , Ferrándiz A 1 , Sánchez F 1 , Cubedo M 1 , Mateu L 1 1 Intensive Care Department, Hospital General de Castellón, Castellón de la Plana, Spain INTRODUCTION. Less invasive monitoring through PICCO system allows continous assessment of haemodynamic function with lower degrees of morbidity and mortality than than using conventional pulmonary artery catheters. Our paper´s aim is to analyze the effect of changes in vasopressive drugs infusion doses on certain haemodynamic parameters.

We obtained 103 pairs of measurements in 15 mechanically ventilated severily ill patients (12 males) monitorized through PICCO system approach. Arterial systolic BP, CI, systolic volume and dp max were registered after changes in infusion doses of dopamine, dobutamine and / or norepinephrine (alone or associated) used for achieving an adequate circulatory and haemodynamic status according to physician opinion. Reasons for admission of these in the ICU were: 7 cases of sepsis (48 pairs of measurements), 2 severe cardiac failure (23 pairs), 1 pancreatitis (5 pairs), 2 multiple trauma (6 pairs) , 1 torathic trauma (13 pairs) and 2 pneumonias (8 pairs). Drugs doses were increased in 36 pairs, were decreased in 29, and were not modified in 38. Analysis of data was performed by SPSS/PC 11.0 with a significance level of p <0.05.

Only BP values changed globally and significantly after changing pressive drugs doses. In those cases with which infusion rates were increased, correlations between dp max and CI and systolic volume showed correlation coefficients of 0.267 and 0.115 respectively, while dp max-BPs correlation was 0.755. When we decreased doses, R values for same correlations were 0.204 and 0.093 in front of R for dp max-BPs of 0.797. When doses were not changed, the three R values were 0.026, 0.237 and 0.538 respectively.

In spite of global changes in BPs, CI, and dp max, as indicators of systolic function, when pressive drugs doses were modified , BPs is (through PICCO monitoring) the most reliable indicator of the induced haemodynamic effect.

Sramek V 1 , Stetka P 1 , Dadak L 1 , Suk P 1 , Pavlik M 1 1 Department Anesthesia and IC, St. Annas University Hospital, Brno, Czech Republic

Optimal MAP in critically ill is still discussed. Little is known how manipulation of MAP within physiologic range with different vasoactive drugs influences static (PAOP, CVP, ITBV) and dynamic (SVV) preload parameters [1].

Six normotensive patients on CMV with low/no catecholamines and without significant cardiac dysfunction (SVI > 35 ml/m 2 ) were studied. All had PAC in situ and besides PiCCO system was introduced. After baseline measurement MAP was manipulated +20 mmHg at random with stepwise titration of either NE, phenylephrine (PHE) and nipride (NIP,4 pts)/or NE decrease (2 pts). After 30 minutes minimum new measurement was performed. Study was approved by local EC. Values as median (range). Statistics: Friedman ANOVA, Wilcoxon matched paired test; p<0.05 considered significant. 

The effects of thoracic epidural anaesthesia (TEA) on the microvascular oxygenation of the gastric mucosa (mHbO2) are unclear. TEA should, by splanchnic sympathicolysis, increase mHbO2, but undesired systemic side effects of TEA may counteract these desired regional splanchnic effects. Ventilation with high positive end-expiratory pressures (PEEP) -as model of compromised hemodynamics [1]-decreases mHbO2 by depression of systemic O2-transport (DO2) and also by regional effects, e.g. sympathetic mediated vasoconstriction [2] ; the latter should be prevented by TEA. Since effects of a TEA on mHbO2 may depend on the respective DO2, we studied the TEA-effects also after restoration of DO2 with infusion of HES.

On 6 healthy, chronically instrumented dogs (flowmeter for cardiac output measurement, sevoflurane, mechanical ventilation, permission of local government) we repeatedly studied mHbO2 (spectrophotometry,EMPHO-II [3] ) and DO2 under the following conditions: Baseline, TEA, TEA+HES, TEA+PEEP, TEA+PEEP+HES. TEA was performed randomly with lidocaine or saline(controls).Data are mean±sem.Statistics:ANOVA,p<0.05.

The TEA maintained mHbO2 (47±3% before and 49±4% with TEA) despite significant reduction of DO2. The HES-infusion, titrated to restore DO2 to the baseline level, even increased mHbO (49±4% to 57±4%). In contrast, TEA aggravated the PEEP-induced reduction of mHbO2 (lidocaine-TEA: 32±1% vs saline-TEA: 44±4%), in parallel with an aggravated depression in DO2. HES-infusion completely restored the depression of mHbO2 (to 45±2%).

CONCLUSION. TEA per se maintained mHbO2 -despite significantly reducing DO2-possibly by redistribution of perfusion towards the gastric mucosa. The aggravated reduction of mHbO2 by PEEP was paralleled by an aggravated depression of DO2, which together supports the concept of increased systemic sensibility, but does not indicate selective depression of regional splanchnic oxygenation during TEA. The compensation of the mHbO2-depression by HES-infusion, titrated to restore DO2, also speaks against a selective depression of mHbO2 during TEA. However, during TEA there is a higher sensitivity of regional and systemic oxygenation during compromised cardiovascular conditions. 

Ischaemia and reperfusion of the liver may occur perioperatively during surgical procedures. In these cases especially impaired microvascular flow contributes to perioperative liver injury. This animal study was designed to investigate, wheter systemic sympathicolyis with clonidine (C) or regional sympathicolysis with peridural anaesthesia (PDA) affect hepatic sinusoidal width (SW) as one marker for sinusoidal perfusion during and after haemorrhagic shock (HS).

Following approval by the local animal ethics committee 36 Sprague-Dawley rats were randomly assigned to one of six groups: Controls, animals with PDA, animals with C, animals with HS, animals with PDA and HS, animals with C and HS. Instrumentation included PDA, tracheotomie and catheter in A. carotis. HS was induced for 60 min at a mean arterial pressure of 40 mmHg, reperfusion time was 5 hours. Sinusoidal width was measured by intravital microscopy at baseline, after 60 min HS and after 5 h reperfusion, and at corresponding times in groups without HS.

During baseline there was no difference in SW in all groups (10.6 mum). In animals with PDA and HS the SW was 8.39 mum vs. 6.36 in HS animals after 60 min HS, and 7.58 mum in C plus HS animals. After reperfusion SW was 7.95 mum in HS animals, 9.36 mum in PDA plus HS and 9.00 mum in C plus HS animals.

In animals with sympathicolysis during HS, the sinusoidal width was greater than in animals without sympathicolysis. With PDA the SW was maintained better than with C. Sympathicolysis might have a protective effect on hepatic microperfusion during haemorrhagic shock due to an increase in sinusoidal width.

Vagts D A 1 , Iber T 1 , Roesner J P 1 , Mutz C 1 , Kurzweg V 1 , Harkner C 1 , Radis K 1 , Nöldge-Schomburg G F E 1 1 Clinic of Anaesthesiology and Intensive Care Medicine, University of Rostock, Rostock, Germany

Clonidine is an alpha-2 adrenoceptor agonist with many interesting properties, which is used more and more in intensive care medicine for treatment of perioperative delirium and alcohol withdrawal. However, clonidine is accused to severely impair gastrointestinal function by reducing gut motility and and mucosal perfusion [1,2], but there is still a lack of conclusive data. This sudy was designed to investigate the effects of intravenously applicated clonidine on intestinal perfusion and oxygenation in an acutely instrumented pig model.

Following approval by the local animal ethics committee 19 anaesthetized, ventilated and acutely instrumented pigs (catheter in femoral artery, pulmonary artery, sup. mesenteric vein, ultrasound transit-time flowmeter around the sup. mesenteric artery, intestinal tonometer and pO 2 -electrode onto intestinal serosa and mucosa) were randomly assigned to 2 groups: Group 1: (9 pigs) received 2 mug kg -1 i.v. clonidine after baseline measurement and an infusion of 2 mug kg -1 h -1 clonidine 45 min after bolus injection. Group 2 (10 pigs) served as controls. Measurements of systemic haemodynamics and regional parameters of perfusion and oxygenation were repeated 1.5 h and 4.5 h after bolus injection of clonidine/vehicle. Statistics: Wilcoxon signed rank test, Mann-Whitney-U-test.

Clonidine induced reduction of mean arterial pressure (MAP) (10%), heart rate (20%), cardiac output (CO) (30%) and systemic oxygen transport (30%). Superior mesenteric arterial blood flow (16.5 ml min -1 kg -1 at baseline vs. 15.8 at 4.5 h), intestinal oxygen supply and intestinal oxygen uptake were not affected in both groups. Serosal (56 vs. 50 mmHg) and mucosal (31 vs. 28 mmHg) surface oxygen and mucosal carbondioxide partial pressure did not change.

CONCLUSION. Despite reduction of CO, MAP and systemic oxygen delivery after application of clonidine, intestinal oxygen supply and mucosal oxygenation were maintained. Our data do not support, that clonidine impairs gastrointestinal perfusion. Clonidine did not affect mucosal oxygenation in acutely instrumented pigs.

Providing adequate nutritional support in children with impaired oral feeding remains a provocation for physicians. When possible most authorities recommend the enteral route as opposed to parenteral route because it is cheaper and safer and has a known superiority in preserving gut integrity and immune function.

Evaluation of enteral nutrition with Frebini was performed in 16 patients aged between 1 and 10 years that received enteral nutrition for a period of minimum 14 days. The cases were admitted in our Intensive Care Department between 2001 and 2003. The reasons for prolonged nasogastric enteral nutrition were: severe oral-motor incoordination with impaired deglutition (5 cases), gastrointestinal surgery (3 cases), neurological impairment (5 cases) and severe malnutrition with refusal of oral feeding (3 cases). The parameters used to evaluate the nutritional therapy were: clinical and neurological examination, gastric residue, body weight, number and consistency of stools, hemoglobin and hematocrit, serum protein concentration and protein fractions, nitrogen balance, blood glucose, serum electrolytes, liver enzymes, direct and indirect bilirubin, serum creatinine and uric acid, blood gases and pH.

Enteral nutrition with Frebini was well tolerated and no withdrawal of enteral feeding was necessary. High gastric residue and vomiting were notice in 20% of cases, especially at the beginning of enteral nutrition. Supplementation with a prokinetic agent and switching from bolus to continuous feeding was needed in these patients. Body weight and body composition improved in all patients with no disturbances in serum electrolytes, liver or renal function. Blood gases and pH were in concordance with the underlying disease.

Administration of Frebini through a nasogastric tube in children proved to be safe and well tolerated even in patients with malabsorption. Patients with malnutrition showed a marked improvement in their nutritional status. Implementation of an organized home enteral nutrition program for handicapped child is required in order to augment the results achieved in the hospital. 

Blazek D 1 , Cvachovec K 1 , Dlask K 1 , Pavlicek P 1 , Hechtova D 1 1 Anaesthesiology and Intensive Care Medicine, Motol University Hospital, Prague, Czech Republic

In the last years there has been seen the ever increasing number of the patients awaiting an organ transplant as the only curative treatment of their disease shifts the offage limits of suitable cadaver organ donors to both higher age limits (above age of 60) as well as lower ones in the case of paediatric donors.

This report presents authors' experience based on the care of 36 cadaver donors since 1993.

The authors suggest some principles of the care of the potential donor in the paediatric intensive care unit, specifically in comparison to the care of the adult donor. A special attention is given to the issues of assessment of the "brain death", the tests prior to the PGA (atropine test, fluorescein test, etc.) and the actual arrangement of the cerebral PGA. The importance of the continuity of the intensive care before and also after establishing the diagnosis of brain death is emphasized. The stress is put on the ventilation, cardiovascular system, arrhythmias, haemostasis, body temperature, infection prevention and intensive continuous monitoring of the vital functions of both the recipient and donor. Specific details of the preparation of the donor for organ harvesting and anaesthesia during the procedure are discussed.

In the last section the authors address variety of sensitive issues regarding legislation, previous informed consent, the consent of the paediatric patient's family and the influence of their involvement on the number of successful donations. They also discuss the maintenance of the medical record and overall approach to the management of the paediatric cadaver donor in the settings of intensive care unit.

REFERENCE(S). 1. Bellomo R et al: Acute renal failure: Time for consensus

PVC as a biomedical polymer -plasticizer and stabilizer toxicity

Intrahospital transport of critically ill patients

Intrahospital transport of neuro ICU patients

Mishaps during transport from the intensive care unit

RESULTS. The way we have measured intra-abdominal pressure was through an indwelling foley catheter placed in stomach. Before measuring we injected 10 to 15 ml of saline into the patients stomach and catheter connect by tube with pressure connection to the bedside monitor. The pressure was measured every two hours. Pressure above 15 mmHg was pathologic and above 30 mmHg was an indication for opening the abdomen. All babies were operated when pressure was nearly 30 mmHg. General condition was much better after the operation.

Measuring an intra-abdominal pressure is very useful method in newborns with malformations of GIT.

Rozhko Y V 1 1 Neonatal intensive care, Motherhood and childhood defence, Minsk, Belarus INTRODUCTION. For the last decade the positive tendency of neonatal mortality decrease is observed, however it is accompanied by increase of cerebral palsy (CP) and other neurodevelopment disorders (1). The role of perinatal factors and intensive care volume on frequency of CP among very low birth weight (VLBW) infants were studied METHODS. The retrospective antenatal, intrapartum and neonatal events and therapies analysis of 130 VLBW infants was performed. All survived VLBW newborns who were admitted into neonatal intensive care unit of Minsk city hospital during 1997-1999 were enrolled in the study.RESULTS. 14 of 130 survived neonates suffered from CP to the age of 36 month. The VLBW neonates born in vaginal delivery compared with those born in caesarean section (18% vs. 2,3% respectively) as well as hypocapnia and metabolic acidosis (100% vs. 65% and 36,4% vs. 10,4% accordingly) were at higher risk for CP (2) . The VLBW neonates with severe periventricular/interventricular haemorrhage (PVH/IVH) and periventricular leukomalacia suffered later from CP more often then those without such pathology (42.8% vs. 4.7%).

Vaginal delivery, hypocapnia, metabolic acidosis and severe PVH/IVH are identified as risk factors for CP development (3) . 

Poojara L R 1 1 Anesthesiology And Critical Care, Sri Ramachandra Medical College And Research Centre, Chennai, India INTRODUCTION. : Organophosphate Compund (OPC) poisoning often has two phases. 1) An immediate phase of respiratory depression, cholinergic symptoms and coma requiring ventilatory support, oximes and anticholinergics. 2) After initial recovery some patients have resurgence of weakness, paralysis, twitching and increasing cholinergic symptoms due to redistribution of the poison. This lasts 4-9 days requiring continued ventilatory support and is termed intermediate syndrome (IMS) 1 . OPC poisoning results in irreversible blockade of the enzyme acteylcholinesterase(AChE) with excess acetylcholine causing persistent depolarisation of the motor end plate. During IMS, prolonged suppression of AChE could progress to phase II block 2 . However peripheral nerve stimulation using TOF/ Tetanus has failed to show a characteristic fade. We elected to study whether electro physiological monitoring might show a decremental response during IMS.

This was a prospective blinded study done from April 2002 to March 2003 in our ICU. 43 consecutive patients of OPC poisoning admitted during this period were included in this study. Repetitive nerve stimulation (RNS) using train of ten at 3 Hz at the ulnar nerve was done on all patients on day 1, 4, 7 and every 4th day thereafter until discharge. Patients were ventilated until weanable as per our usual protocol. The results of the RNS study were not revealed to the intensivist.

. 7 out of 43patients required ventilation for more than 6 days and showed overt signs of intermediate syndrome -proximal muscle weakness, twitching and respiratory weakness. Only 1 out of 7 had a decremental response on RNS. This patient had symptoms of severe poisoning and was deeply comatose requiring controlled ventilation for 15 days after which he had full recovery of both symptoms and RNS. All other patients with IMS showed no changes on RNS.CONCLUSION. RNS is a specific but poorly sensitive marker in diagnosing intermediate syndrome after OPC poisoning. Other markers like red cell cholinesterase levels may be studied for better specificity and sensitivity. 

Cosgrove J F 1 , Nesbitt I 1 , Snowden C P 1 , Thomas P 1 , Shenfine S 2 , Cowen W 3 1 Perioperative and Critical Care, Freeman Hospital, 2 Accident and Emergency, Newcastle General Hospital, 3 Accident and Emergency, Newcastle general Hospital, Newcastle upon Tyne, United Kingdom INTRODUCTION. Delayed airway protection in severe acute poisoning had previously been linked with respiratory morbidity in the Newcastle upon Tyne Hospitals NHS Trust. Recommendations for intubation (based on previously published guidelines) were therefore introduced to facilitate the management of such patients. They emphasised the need to assess the airway, ventilatory sufficiency and level of consciousness as judged by the Glasgow Coma Scale.

To determine whether intubating severe acute poisoning patients according to established guidelines resulted in reduced duration of intubation and length of stay on intensive care in a prospective analysis of severe acute poisoning patients aged over 16 years, between September 2000 and August 2001. Statistical analysis was performed using the Mann-Whitney Utest for unmatched parametric data.

Thirty-two patients met specific intubation criteria, (age range: 16 to 74 years; median age 38.5 years.) Failure to intubate on initial assessment in Accident and Emergency (n=6) resulted in longer average intubation times (28 hours versus 14 hours: p=0.125) and a longer average length of stay on intensive care (52 hours versus 23 hours: p=0.04.)

The study had limitations: small sample number, lack of controls, uncertainty over the types of poisons ingested and the possibility of medical co-morbidities contributing to illness severity. However, early intubation in severe acute poisoning (according to specific criteria) resulted in a significant reduction in length of stay on intensive care in a UK patient population. 

Ecstasy (MDMA) is very popular designer drug in discotheque attendees. It is an amphetamine derivative with mescaline like effects and is considered by users as safe and without serious side effects. Substantial risks of toxic effects on nervous system with permanent psychic and psychiatric damage is rarely known. In addition, may occur potential life-threatening rhabdomyolysis, disseminated intravascular coagulation, liver cell necrosis, renal failure, cardiac arrhythmias, disorders in water and electrolyte metabolism and aplastic anemia. World wide there are already about a hundred fatal casualties reported. A threshold for these problems is unknown.

We report about two cases in our cohort with toxic serums levels and different outcome.

Case 1: 23 year old male was admitted in deep unconsciousness after he was found in front of a discotheque. He had after a drug-free period taken ecstasy again. When the emergency crew arrived blood pressure was 80/60 mmHg, pulls frequency was 170 beats per min with narrow QRS-complexes in ECG. Oxygen saturation was 99%, he was breathing normally. Blood sugar was 76 mg/dl, Glasgow-coma scale 3 points, eye pupils were mid wide without reflexes to light. In the ambulance the patient developed seizures which stopped after injection of diazepam. The ecstasy serum peak was 1.4g/ml. Thereafter, he had gasping with the necessity of mechanical ventilation and resistance to therapy multiple organ failure with DIC, lever and kidney malfunction. Exitus letalis occurred on day 4. Case 2: 19 year old male had take nearly every two hours since two days one tablets of ecstasy. In addition, he ingested 20 tablets of diazepam a 5 mg, 20 caffeine tablets, two liters of "Red bull" plus Vodka, 3 g cocaine and 5 g "Speed". Apart of an impressive psychic alteration there were found normal vital signs and laboratory values and a high toxic potential lethal ecstasy serum peak of 1.5g/ml.

In both cases, a serious ecstasy intoxication was found with potential lethal MDMA-Serum peaks. The serum levels seem to be just one parameter with prognostic impact. Adaptation processes in chronicle abuse may have potential protecting properties against toxically effects.

Guven M 1 , Sungur M 1 , Eser B 2 , Sari I 1 , Coban O 1 1 Intensive Care, 2 Hematology, Erciyes University Medical School, Kayseri, Turkey INTRODUCTION. The treatment of organophosphate (OP) poisonings, including atropine and oximes, may be failed to improve mortality and morbidity in some cases (1). Human plasma may be potentially source of acetylcholinesterase (AChE) (2) .We want to determine the effects of human plasma on AChE levels in patients with OP poisoning.

The study was performed at our ICU. 28 patients were admitted for OP poisoning. The diagnosis was based on history and plasma AChE levels. Atropine was given until control of hypersecretion occurred, and was discontinued 24 hours after all signs of atropinisation. Pralidoxime was administered as 4 g daily divided to four doses. We obtained pralidoxime from the offices of the Ministry of Health, but it is sometimes not available. Frozen plasmas were given 9 patients who received pralidoxime and 1 patient did not receive. Plasma therapy was started after day 2 in 8 patients and after developing IMS in a patient. AChE levels were measured before and after therapy. Two bags of plasma (300-400 ml/bag) were given daily until the patients were not needed to atropine. The study was approved by the Ethical Committee of Medical School, and verbal informed consent was obtained from the patients or their relatives.

Seventeen patients received pralidoxime, 8 patients pralidoxime+plasma after day 2, one patient plasma after developing IMS, one patient only atropine and 1 patient atropine+plasma. While IMS had developed in 29.4% (5/17) of patients received pralidoxime, there was no any IMS cases in patients received plasma (0/8 cases). The mortality rates were 11.8 % in pralidoxime group, and 0%in plasma group. Patient receiving atropine+plasma therapy did not have IMS, and this patient survived. Mean amounts of plasma given were 4.9±3.5 bags, and AChE levels of human plasma were 3944.1±538.1 IU/L. Every two bag of plasma provided an increasing in plasma AChE approximately 414.0±85.6 IU.

We found that human plasma could increase AChE levels in patients with OP poisoning. Plasma+atropine with/without pralidoxime therapy appear to be more effective than atropine+pralidoxime therapy. This approach may also prevent the developing of IMS and related mortality. Thus, plasma therapy may be suggested a treatment method in OP poisoning. But, this must be supported by other study. 

Soltner C 1 , Dubé L 1 , Bouhours G 1 , Berton J 1 , Colbus L 1 , De Sevin F 1 , Granry J J C 1 , Beydon L 1 1 Département d'Anesthésie Réanimation, CHU d'Angers, Angers, France INTRODUCTION. The TRISS method, developed from a cohort of 25,000 North American patients, allows expected mortality to be estimated within a population of trauma patients in an Anglo-Saxon management system. It is based on the Revised Trauma Score, the Injury Severity Score (ISS), and the age of patients. The purpose of the present study was to compare the mortality predicted by the TRISS method with that observed in a population of French trauma patients managed in a care system involving medical treatment before hospitalisation.

A retrospective study was conducted from 1997 to 2000 in 608 trauma patients. Overall mortality was analysed, and three degrees of severity were then considered as well as severe cranial trauma (SCT). Statistical analysis was based on the Z score of the TRISS method (a score of less than -1.96 was indicative of a significant difference in terms of mortality).

Predicted mortality was 15%, observed mortality 12%, and the Z score -2.39 (p<0.05); Ninety percent of deaths occurred in the first 24 h. The M score was 0.91, indicating that the population studied was sufficiently similar to that used to elaborate TRISS. The mean age of patients was 34 years, and the M/F sex ratio 3/1. SCT was found in 19% of cases 

Observed mortality in the population studied was less than predicted mortality, and this difference was more marked for patients with SCT and those with moderate severity. These results argue in favour of medical treatment before hospitalisation of polytrauma patients, as this factor constituted the main difference between the two management systems. 

In several studies the mortality in polytraumatic ICU patients was correlated with PCT levels, especially those with abdominal trauma. The aim of the study is to find out if PCT values at the first three days after injury are correlated with tissue hypoxia in several trauma.

Twenty-seven multiple injuried, mechanically ventilated, not infected patients, (21 men, 6 women, mean age 27 years) were included in the study. After haemodynamical stabilization arterial blood gases were taken at days 0, 1 and 2, for measurements of PCT, Lactate, blood gases and leukocytes count. MAP was also recorded at the same days. For statistical analysis the Paired-samples T test was used. 

Since the incidence of chest trauma is projected to mount and the number of patients with chest trauma reaching the hospital alive will continue to grow. Thus, the management of these patients will become an increasingly large part of critical care practice (1). Remifentanil is a new opioid agent which is able produce a predictable and easily level of sedation in post-cardiac patients (2) . Aim of this study is to report our experience with remifentanil alone in sedating patients with chest trauma, over an one year period.

In 2002 patients affected by chest trauma requiring sedation and ventilation were sedated with remifentanil (0.1 mcg/kg/min).The following data were recorded: age, sex, Simplified Acute Physiology Score II (SAPS), needed of supplemental analgesia, ICU stay and mortality. Level of sedation was evaluated using Cook and Palma and Ramsay scores. 

Remifentanil is a safe sedative agent and given alone is well indicated for sedation of patients affected by chest trauma and could offer advantages over many other drugs. 

Taneva E K 1 1 Clinic of Cardiology, Intensive Care, Medical University, University Hospital Alexandrovska, Sofia, Bulgaria

In patients with thermal shock thermoregulation is changed through peripheral and central mechanisms. Bigger temperature gradients are found in different parts of the human body. The aim of this study is to establish the clinical values of measured temperature gradients in patients with thermal shock.

The study was conducted in the Clinic of burns, Emergency Institute "Pirogov", Sofia, Bulgaria. The study includes 25 patients in the phase of thermal shock, 14 of them are children (from 2 to 9 years olds) and 11 adults (from 18 to 50 years) with burned area from 15 to 50%. The patients were separated in 2 groups according to development of thermal shock -light or severe shock. Every 3 hours oral, rectal, skin temperature of the toe, and also temperature and relative humidity of the room have been investigated for 3 days after the incident. Following parameters were measured: oral-skin temperature gradient (OSTG), rectal-skin temperature gradient (RSTG) and rectal-oral temperature gradients (ROTG).

During the light thermal shock the internal temperatures (oral and rectal) and skin temperature are normalized on the 12 hours of the first day. During the severe thermal shock the internal temperatures remains under 37oC, but the mean skin temperature is between 24 -26oC. In light development thermal shock OSTG and RSTG return to referent values as early as within 24 hours, when medical and infusional treatment has been started. But these gradients in patients with severe thermal shock remain from 10oC to 14oC till the end of the 3-rd day. There are no statistically significant differences in ROTG in the two groups.CONCLUSION. 1. Tracing the temperature gradients is fast, painless and reliable supplementary method that will help the diagnostic process.2. The scale of temperature gradients shows the real possibility of evaluation of the haemodynamics and the results of the treatment.3. The prognostic values of these gradients are to show the issue of the thermal shock, but not the prognosis of the disease. good correlation between high PCT levels and the risk of mortality in inflammatory conditions. We previously identified PCT as a predictor of early graft failure-related mortality after HTx (1). We now for the first time evaluated PCT as a prognostic marker in the early postoperative period after heart (HTx) and lung (LTx) transplantation.

Blood was prospectively collected at daily intervals from 100 consecutive patients (age 47 ± 13 years) for up to 28 days following HTx (n =73) and LTx (n = 27). PCT serum concentrations were measured by immunoluminometry. In addition, C-reactive protein (CRP) was measured and WBC counts performed.

From the 1st until the 28th postoperative day (POD) PCT levels were significantly higher in non-survivors as compared to survivors, whereas CRP levels only from the 7th and WBC counts from 5th POD became significantly higher in non-survivors. Areas under the receiver operating characteristic curve for the risk of mortality on the 1st POD was 0.74 for PCT and increased continuously thereafter, compared to 0.47 for CRP and 0.48 for WBC counts on the 1st POD. A PCT value of > 2 ng/mL on the 7th POD as a predictor for mortality had a specificity of 65.8% and sensitivity of 75%.

In comparison to CRP and WBC counts PCT was identified as the best predictor of mortality in patients after HTx and LTx. PCT measurements in the early postoperative period after HTx and LTx may be helpful in early therapeutic decision making by identifying patients at high risk of mortality. 

Merlani P 1 , Chenaud C 1 , Cottini S 1 , Reber G 2 , De Moerloose P 2 , Ricou B 1 1 Surgical Intensive Care, 2 Hemostasis Unit, University Hospital, Geneva, Switzerland INTRODUCTION. Coagulation management after cardiopulmonary bypass (CPB) need measurements of activated partial thromboplastin time (aPTT) to avoid excess bleeding and start anticoagulation. Usual practice refers to results from the central laboratory, but they take too long. We hypothesized that bedside assessment of Point of care (POC) aPTT would reduce time to achieve the desired coagulation state, resulting in reduced blood loss and transfusions.METHODS. 126 patients planned for surgery with CPB [42 valves(VV) and 84 coronary artery bypass grafting (CABG)], were randomized in 2 groups. The Lab-group had the postop coagulation management guided by central laboratory aPTT (Lab-aPTT) values and the POCgroup by POC-aPTT obtained by CoaguCheck pro. Prophylactic/therapeutic heparin was applied according to a guideline, accounting for the difference between Lab-and POC-aPTT (1).

Patients were similar in the two groups. POC-aPTT results were available earlier after venipuncture (3±2 vs 125±68min in VV and 3±4 vs 114±62min in CABG, both p<0.00001). This resulted in earlier introduction of heparin in the POC-group (7±23 vs 13±78h in VV and 11±57 vs 14±101h in CABG, p=0.01 and p=NS respectively). During ICU more Lab-aPTT results were in the desired range in VV of the POC-group (47% vs 34%, p=0.007) and they received more heparin/24h (15±8 vs 9±6 [1000 UI], p=0.01). The cumulative mediastinal blood loss during ICU was higher in the Lab-group (992±647 ml) compared to the POC group (630±300ml) (p<0.01, Bonferroni) in VV but not CABG (1099±621 vs 1076±868 ml, p=NS). At ICU discharge, the hemoglobin was similar in the POC-and Lab-group (9.8±1.2 vs 9.3±1.4g/L in VV and 9.7±1.3 vs 9.7±1.0.5g/L in CABG, both p=NS), but significantly more VV in the Lab-group needed blood transfusions during ICU compared to the POC-group (8/21 vs 1/21; p=0.03) (CABG: p=NS). Mortality and length of ICU and hospital stay of the 2 groups were similar in this small cohort. Agreement between POC-and Lab-aPTT was poor (-17±33sec) as already reported by our group (1).CONCLUSION. POC-aPTT permitted a better management of coagulation after valvular surgery. POC-aPTT allowed decreasing mediastinal blood loss and transfusions after valvular surgery but not after CABG. 

Various infusion fluids can be used during resuscitation from hypovolemia, after cardiac surgery. Cardiac surgery can lead to an increase of endothelial permeability and eventually the adult respiratory distress syndrome (ARDS). Colloids are believed to maintain the colloid osmotic pressure (COP) and would result in less propensity for development of pulmonary oedema than crystalloids. HES and albumin might have additional benefits in ameliorating an increased permeability by 'plugging the leaks'. We tested these hypotheses by comparing fluid challenges with NaCl 0.9%, Gelofusine, Hemohes and albumin 5% in 39 patients with reduced filling pressures, after cardiac surgery.

In this prospective, non-blinded, clinical trial, patients were randomly assigned to receive Hemohes (6% hydroxyethyl starch, MW 200.000, substitution 0.45-0.55), Gelofusine (40g/L), albumin 5% or NaCl 0.9%. Before and after the fluid regimen, we recorded demographics, haemodynamics and ventilator variables. We measured the pulmonary microvascular permeability by calculating the 67 Ga-transferrin Pulmonary Leak Index (PLI) and the extravascular lung water (EVLW, with thermal/dye indicator dilution). Blood samples were taken for gas exchange parameters, protein levels and the COP. A chest radiograph was made and a lung injury score was calculated. Fluids were dosed during 90 minutes on the basis of the response within predefined pressure limits.

In the albumin group, there was a significant increase in the albumin concentration (p=0.003) and a decrease in the two other colloid groups (p<0.007), related to similar changes in total protein concentrations, except for the Gelofusin group, were the decrease in total protein did not reach statistical significance. The colloids were effective in elevating the COP (p<0.015). Twelve patients had an EVLW of more than 7 ml/kg and 19 patients had a PLI of more than 14·10 -3 . min -1 at baseline (= upper limit normal). Groups did not differ with respect to (changes of) EVLW, PLI and gas exchange after the fluid resuscitation compared to baseline measurements.

Despite a significant increase of the COP in the colloid groups, we could not find important differences in gas exchange, EVLW, radiographic abnormalities and pulmonary permeability between the different fluids used to correct hypovolemia after cardiac surgery. Grant acknowledgement: B. Braun Melsungen, Germany

Albers J 1 , Schmidt T 1 , Vahl C F 1 , Hagl S 1 1 Cardiac Surgery, University of Heidelberg, Heidelberg, Germany

Atrial fibrillation (Afib) as the most frequent postoperative arrhythmia causes high comorbidity and increased mortality. Delivery of cardioplegia in on-pump coronary artery bypass grafting (CABG) patients is not standardized. High volumes of cardioplegia, however, may impair ventricular function and cause arrhythmia. The purpose of this study was therefore, to analyze the possible influence of cardioplegic volume on evolvement of Afib in the intensive care unit (ICU).METHODS. In 2,251 consecutive patients operated on three-vessel coronary artery disease, distribution of cardioplegic volumes (crystalloid cardioplegia, Custodiol TM) was determined. Two groups matched by demographic data (age, weight, sex, body surface area, history of myocardial infarction or previous Afib, NYHA status) and separated by the median cardioplegic volume, were further analyzed for development of Afib in the ICU. Statistical testing used Maentel-Haeuszel Chi-Square, Fisher's exact test and logistic models.

Volumes of cardioplegia used during CABG were: 1789.0, 601.9, 1900.0, 12.7 ml (mean, standard deviation, median, SEE) representing 0.09, 0.03, 0.09, 0.0006 ml/cm2 (mean, standard deviation, median, SEE) body surface area. Incidence of Afib was significantly higher in the high-volume-cardioplegia group (n=56 vs. 43 at the first, n=59 vs. 38 at the second postoperative day, p=0.01). Respective overall relative risk (Odds ratio) for development of Afib was 0.85 (logit 0.85).

In comparable CABG patients, the use of a higher amount of crystalloid cardioplegia seems to result in higher incidence of Afib. In the ICU setting, those patients at risk should therefore be closely monitored for development of Afib. Mofidied protocols may enhance early detection and therapy of Afib in the future, reducing comorbitiy and mortality following primarily successful CABG.

Chacko J 1 , Raghunath T 2 1 Intensive Care Unit, N H I M S, 2 Biochemistry, NHIMS, Bangalore, India

Cardiac troponin I (cTnI) is useful in the diagnosis of perioperative myocardial infarction (MI) after coronary artery bypass (CABG) (1) and high levels may be a marker for longer ICU stay and ventilator days, arrhythmias, ST changes and cardiac failure (2) . We hypothesised that postop cTnI levels due to perioperative ischaemia would predict adverse outcomes such as prolonged ventilation, ICU stay, use of multiple inotropes and intraaortic balloon pump (IABP) support. We also tried to correlate high cTnI levels with intra and postop ST changes and regional wall motion abnormalities (RWMAs) in the postop period.

We prospectively enrolled 193 consecutive patients undergoing elective CABG. cTnI levels were measured soon after arrival in the ICU (T0), at 12 hours (T12) and 24 hours (T24). New intra or postop ST changes or q waves, the number and duration of postop inotrope usage, time to extubation and the use of IABP were recorded. RWMAs were assessed by echocardiogram on the day after surgery.

. cTnI levels correlated with time to extubation, ICU days, duration of inotropic support and number of inotropes used (Table1). Signifcantly higher mean cTnI levels were seen in patients with intra and postop STchanges, IABP support (p<0.0001) and new RWMA (p<.001) by unpaired t test. cTnI levels of >6.0ng/ml were significantly associated with ventilation for >24 hours (Odds ratio = 6.325, Chi-square=16.04, p <0.0001). 

Acute postoperative renal changes were more frequent for higher preoperative risk category. Increasing degrees of renal impairement were associated with longer ICU stay without affecting mortality. Adverse outcome may be explained by more severe levels of multi organ impairement. These definitions may be helpful to describe the different impact of renal injury and RF on use of resources and outcome.

Sacre P 1 , Senard M 2 , Ledoux D 1 , Monchi M 1 , Canivet J 1 , Damas P 1 1 General Intensive Care Unit, 2 Anesthesia and Reanimation, University Hospital of Liege, Liege, Belgium INTRODUCTION. Troponin I has a complete cardiospecificity and is used as a marker of myocardial damage after cardiac surgery. A recent study showed that Troponin I was a predictor of post cardiac surgery complications rate and outcome. The aim of this study was to assess the ability of the peak troponin I level (pTnI) to predict outcome in cardiac surgery patients and to compare its discrimination power with peak CKMB level (pMBM).

We carried out a retrospective study on all patients admitted for cardiac surgery over a 12 month period. Three groups were considered: coronary artery bypass (CABG), valve replacement (Valve) or both valve replacement and CABG (Valve + CABG). We used the Receiver Operating Characteristic (ROC) curve to assess the discrimination power of the peak Troponin I level and peak CKMB level on ICU outcome.

Over the 12 month period, 707 patients were included (demographic data are presented in Table 1 ). The mean pTnI level according to the type of surgery was not statistically different although there was a trend for a higher level in patients who underwent a Valve+CABG surgery (pTnI : 33,5 ± 78,7 g/l, 35,5 ± 83,8 g/l and 55,6 ± 85,9 g/l for CABG, Valve and Valve+CABG respectively, p = 0.07). Thirty one patients died in the ICU (4.4 %). The pTnI level was higher in the patients who died (pTnI : 31,6 ± 68,2 g/l versus 146,3 ± 189,4 g/l respectively for survivors and non-survivors patients, p < 0,001). The discrimination power of the pTnI for the outcome was good (area under the ROC = 0.81) and significantly better than the discrimination power of pMBM (area under ROC = 0.63)(p = 0.004).

The pTnI level has a good discrimination capacity to predict ICU outcome in patients undergoing cardiac surgery. It is superior to pMBM which has a poor discrimination power. This easy method offers a reliable tool to early distinguish patients at high risk and in turn to adapt therapeutic strategies.

Oczenski W 1 , Jilch R 2 , Köller U 2 , Krenn H 1 , Schwarz S 1 , Fitzgerald R D 1 1 Anesthesia and Intensive Care, 2 Clinical Chemistry, Lainz Hospital, Vienna, Austria

Decreased expression of the monocytic human leukocyte antigen (HLA)-DR was shown to correlate with the occurrence of SIRS and septic complications (1). To evaluate the predictive value of a decreased HLA-DR levels for postoperative SIRS and septic complications we compared HLA-DR levels of patients with postoperative complications to those with an uncomplicated course. We hypothesised that a decrease of HLA-DR levels within the first 24 hours after cardiac surgery is not related to postoperative SIRS or septic complications.

In a prospective, observational study the expression of monocytic HLA-DR of 85 consecutive cardiac surgery patients was analysed by flow cytometry. Arterial blood samples for determination of HLA-DR were collected before induction of anaesthesia, immediately after admission at the ICU and on the first postoperative day. Monocytic HLA-DR expression was analysed by the quantitative test kit QuantiBRITETM as average number of bound anti-HLA-DR antibodies expressed per monocyte (mAb/cell).

. HLA-DR expression was significantly decreased in all patients. No significant differences of HLA-DR expression within the first 24 hours after surgery were found in patients with uncomplicated course compared to those developing SIRS or septic complications.

In cardiac surgery patients the predictive value of decreased HLA-DR levels within the first 24 seems to be low in predicting an increased risk for postoperative SIRS or septic complications. 

Mendonça-Filho H F 1 , Campos L A 1 , Gomes R V 1 , Fagundes F E S 1 , Nunes E M 1 , Gomes R 2 , Bozza F 2 , Bozza P T 2 , Castro-Faria-Neto H C 2 1 Surgical Intensive Care Unit, Pro Cardiaco Hospital, 2 Laboratory of Immunopharmacology, Oswaldo Cruz Institute, Rio de Janeiro, Brazil

Cardiac surgery (CS) with cardiopulmonary bypass (CPB), is a recognized trigger of systemic inflammatory response, usually related to postoperative acute lung injury (ALI). As an attempt to dampen inflammatory response, steroids have been perioperatively administered to patients. Macrophage migration inhibitory factor (MIF), a regulator of the endotoxin receptor, is implicated on the pathogenesis of ALI. We have previously detected peak circulating levels of MIF, 6h post CPB. Experimental data have shown that steroids may induce MIF secretion by mononuclear cells. This study aims to correlate levels of MIF assayed 6h post CPB with the intensity of postoperative pulmonary dysfunction, analysing the impact of perioperative steroid administration.

We included patients submitted to CS with CPB, electively started in the morning, performed by the same team under a standard technique, except for the addition of methylprednisolone (15mg/kg) to the CPB priming solution for patients from the group MP (n=37) but not for the remaining patients -group NS (n=37). MIF circulating levels were assayed at the anesthesia induction, 3, 6, and 24h after CPB. We adopted a standard weaning protocol with fast track strategy, and registered indicators of organ dysfunction and terapeutic intervention along the first 72h postoperative.

Ryan D W 1 , Topin G 2 1 General ITU, 2 Recovery, Freeman Hospital, Newcastle upon Tyne, United Kingdom

The impact of ICU bed shortage due to excessive demands places the hospital under daily siege. We use the recovery room as a short term solution but the impact on elective work has not been recognised.

In a 30 month period from July 2000-December 31st 2002 a prospective study of all admissions to ICU requiring initial management in the recovery rooms of the central operating suite, due to lack of availability of an ICU bed were recorded.

In the 30 month study period the recovery room cared for 38,601 patients of which 186 were critically ill patients (0.5%), and there were 2 deaths (1%). The ICU in the same time period looked after 1,456 patients, of which 11% were initial treated in recovery. There was no statistical difference in delay to transfer to ICU (mean 3 hours) whether the patients were a planned admission to ICU or if this was a new emergency arising, but the delay was reduced if they arrived after 5pm, p>0.05. The principle source of critically ill patients nursed in recovery was from operating theatres 66% (123), the emergency admission suite 19% (35), and Level 2 HDU/CCU 8% (15). The majority of these cases were transferred to our own unit in 82% (153), or our own HDU 5% (9) but a sizeable minority 10% (19) had to be transferred to another ICU in another hospital.

The shortage of critical care resources required an interim strategy to solve the immediate problem of where to resuscitate and ventilate critically patients when there is no ICU bed. The use of routine recovery facilities causes considerable disruption to elective surgical lists. The lack relatives accommodation plus problems with privacy, noise, documentation and patient safety have been highlighted by others [1, 2] . We are fortunate to have senior ICU trained nurses in charge of the recovery which may explain are low death rate in such critical patients. The problem is yet to be solved but it is at least recognised that our shortfall in resources is requiring us to compromise other aspects of our clinical commitment. 

Gyves H 1 , Manji M 1 , Young I 2 , Rosser D 1 1 Critical Care Medicine, University Hospital Birmingham, 2 Wolfson Computer Laboratory, University of Birmingham, Birmingham, United Kingdom

In April 2002 CCIPS, an electronic prescribing and clinical information management system was introduced to 26 critical care beds in the Trust. The system is capable of using information from the patient's clinical details, laboratory results, drug prescriptions and administrations to provide real time decision support. This report represents the views of nurses on the system's impact on them and their patients.

A questionnaire was developed asking the following 4 questions and for suggestions to improve the system: 1) Does CCIPS improve safety of drug and fluid administration? 2) Does CCIPS protect your professional accountability more than paper drug charts?; 3) Would you prefer to stop CCIPS and return to paper charts?; 4) Does electronic prescribing have a future in critical care? The questionnaire was anonymous and distributed to 110 qualified nurses of all grades over a 3 week period 6 months after CCIPS was introduced.RESULTS. 100 responses were received (91%). The results are tabulated below.

80 73 16 95

These results must be seen in the context that they represent the views of a large group of staff who have been exposed to CCIPS for a maximum of 6 months. Many of these staff were not computer literate at the start of the training period and were highly apprehensive about the move away from paper records. CCIPS is a highly flexible system which continues to develop. The suggestions received from staff in this survey have been used to guide this process. It is likely that the staff's feeling that they can influence the way the system behaves helps create ownership which is a major contributor to the success of its introduction. Data presented elsewhere demonstrate a number of benefits of CCIPS on quality of patient care, safety and reduction in drug errors. These data demonstrate clearly that the system is popular with the vast majority of staff who believe that it delivers benefits to them and their patients.

Yagen I 1 , Benbenishty J 2 1 Intensive Care PACU, 2 ICU PACU, Hadassah Medical Organization, Jerusalem, Israel

The issue of permitting young children the opportunity to visit their critically ill parent in the ICU, is a controversial issue. Very often our nursing staff is asked by the well parent how to cope with the young children's anxiety to their critically ill parent and how to deal with the children's demands to visit their parent. Our ICU believes in holistic practice, we felt a strong need to develop a program to enable child visitation in the ICU.

During a staff meeting, we distributed a short questionnaire in order to assess the staffs' knowledge, and willingness to accept the ICU policy change. A child psychologist at the meeting gave the staff pertinent information. At this point we developed structured guidelines to implement this new practice. When a parent is admitted to the unit, the nursing staff introduces the issue to the well parent. After the well parent approaches the staff with an affirmative response we then photograph the sick parent in the unit. We then meet with the child and we discuss with him the situation and answer questions. At this point if the child requests to visit his parent, we show him the photograph, explaining the visual elements in the picture. We assure the child that even though the parent can't talk with them, they can touch and talk ,we discuss the time limit of the visit. We accompany the child into the unit continually answering all questions that arise. Afterwards we conduct a short debriefing with the child to process the impact of the visit on him.

We have implemented a new nursing strategy. 10 children in the age range of 3-teens have been introduced to this program. The feedback is varied, for some one visit was enough, for others they needed more. The well parents have reported a significant reduction in anxiety symptoms. They expressed relief that the subject has been opened. They themselves were in great need for support and guidance on how to cope with their children. There was also a significant improvement in the confidence that the children felt when their well parent had to leave the home for many hours at a time.CONCLUSION. This ongoing project has been beneficial to the children and the entire family in the coping process. Also beneficial to the staff by addressing a challenge to an intervention that became a unique process to our professional growth.

Cox F J 1 , Townsend R 1 1 Pain Management Service, Royal Brompton and Harefield NHS Trust, London, United Kingdom

Managing acute pain effectively requires appropriate prescribing 1 , in addition to accurate assessment and documentation of pain, adverse effects and analgesia related interventions by skilled nurses 2 . We describe a multidisciplinary approach to promoting best practice in the Intensive Care Unit (ICU) -outlining the development and piloting of a document that streamlines epidural prescribing, combining troubleshooting guidelines with assessment tools and standard infusates.

In response to an unfavourable audit of epidural infusion prescribing in the ICU a multidisciplinary working party critically appraised the published literature and reviewed the function and content of current epidural related documentation. This led to the development and introduction in March 2003 of a pilot colour epidural prescription, troubleshooting and assessment document. A new presentation of local anaesthetic (LA) to complement the existing low dose opioid-LA mixture was introduced. A targeted education programme was provided. Evaluation of this pilot project continues using a questionnaire and unstructured interviews to evaluate staff response and adverse incident reporting.

Offering prescribers pre-printed infusate choices has streamlined ICU stock and decreased the number of reported epidural analgesia adverse effects. The positive effects related to infection control are difficult to quantify. Initial findings of the ongoing evaluation suggest that the majority of prescibers are in favour of the new document as it reduces time and provides adequate choice. All nurses questioned welcomed legible prescriptions, easily accessible troubleshooting guidelines and the professional design of the document.CONCLUSION. Improving epidural care, documentation and prescribing practices related in the ICU is complex and time consuming. Our initial results suggest that comprehensive epidural documentation can facilitate safe, streamlined, evidence-based epidural analgesia practice in the ICU. 

Harada M J C S 1 , Marin H F 2 , Carvalho W B 3 , Pedreira M L G 2 1 Nursing, Federal University of São Paulo, Guarulhos, 2 Nursing, 3 Pediatric, Federal University of São Paulo, São Paulo, Brazil

Pediatric Intensive Care Units (PICU) are environments with complex patient management. They provide a fertile ground for critical incidents that can produce serious patient morbidity and mortality 1 . Several studies have been demonstrated that iatrogenic complications can cause many deaths and disabilities resulting in low confidentiality about the health care system, increasing health care costs 2 . The objective of this study was (a) to identify the critical incident and (b) analyse its relation among PRISM and NEMS in a PICU.

This descriptive and correlative study was conducted in an university hospital in the city of São Paulo, Brazil. The sample comprises 113 critical incidents identified in 38 children of the 76 in PICU. The ethics committee approved the research execution. The parametric test (Ttest, ANOVA) was used for statistical analyses. It was also used non parametric Kruskal-Wallis, with confidence interval of 95%.

The results showed an average of 2.9 incidents per child. It was verified that 32.74% occurred due to medication procedures, 29.20% due to mechanical pulmonary air way/ventilation, 16.81% related to nursing procedures, 14.16% due to catheters, tubes and drains, 4.43% due to equipment and materials and, 2.66% related to other reasons. The children's PRISM average was 4.04 with a predictive mortality risk ranging from 0 to 5% (64.6%). There was no relation with the type of occurrence (p=0.85) and PRISM. The average of NEMS was 35.10 points and there was no relation to the type of incidence (p=0,18) indeed.

The data demonstrated that there was no relation to the different types of incidents with severity of the children's health and the nursing workload of those admitted in the investigated PICU. 

The relationship between organ dysfunction and prognosis is valuable in identifying potential early changes in treatment that could improve outcome. This work aimed to assess the applicability of the sequential organ failure assessment (SOFA) score (1) to a heterogeneous population of ICU patients carried out by nurse staff.

From March to November 2001, patients admitted in the setting of a tertiary medical (24 beds) were enrolled and 193 patients with an ICU stay more than 48 hours were prospectively evaluated in a daily basis. Organ function was evaluated daily according to the sequential organ failure assessment (SOFA) score. Organ failure was considered on the presence of SOFA score > 3. It has been used a colored graphical display at bedside to illustrate to the staff the changing pattern of organ dysfunction.

The mean age was 53 years old (12-89). The length of ICU stay was 9 days (1-68). Neurologic failure was the most prevalent organ dysfunction occurring in 25% of the patients at admission and 23% at day 3, followed by pulmonary and cardiovascular failure (24%, day 0 and 25%, day 3; 16% day 0 and 20% day 3, respectively). Non-survivors patients had significantly higher SOFA score at admission and on day 3 compared to survivors (7.7±3.0 vs 5.6±2.9, 8.2±3.9 vs 5.4±3.3, respectively, p<0.05). Importantly, patients in whom Total SOFA score decreased (n=58) or did not change (n=49) within 48h mortality rates were 25% and 26%, respectively. Patients with increasing values of Total SOFA (n=86) had significantly higher mortality rate (51%) (RR 1.89 CI 95% 1.28 -2.78, p<0.05).

Early changing patterns of organ dysfuction are related to mortality. Evaluating SOFA score daily, especially in the first 3 days, may direct earlier strategies to prevent organ failure. Nurse evaluation of SOFA score may be a valuable tool for bedside care. Incompatibility between certain drugs and PVC also causes extraction of the DEHP into the patient's blood circulation. 1 The objective of this study was to identify IV drugs administered to children that are incompatible with PVC.

This was a descriptive study performed in an university hospital in the city of São Paulo, Brazil. The data collection took place through the observation of 519 administrations of IV drugs in 178 children, in which 444 IV therapy accessories made of PVC were identified. For the identification of incompatibility, the lengths of contact time between the PVC and the drugs identified were considered. 

A working group of seven nurses in the ICU was formed to develop the competence model which was based on recommendations for specialised educated intensive care nurses issued by the National Board of Health and Welfare. The group met regularly over two years. All staff nurses in the ICU (n=40) participated continuously in discussion and meetings.RESULTS. The competence model was implemented in 1999 and includes the following parts: 1) Introductory program, 2) Five-level "ladder" of competence, 3) Criteria for assessment of competence 4) Dialogue once a year with the head nurse.The competence ladder contains of critical care nursing, leadership, teaching/coaching and quality improvement/nursing research.

Today, the competence model is integrated as a natural part in the organisation of the ICU. All nurses are positioned according to their level of competence. This facilitates to allocate patients to a nurse with adequate competence. It is also possible to give training to nurses in a specific competence level. The competence model highlights the competence level of each individual nurse as well as the whole unit, which is used for example in the recruiting process of nurses. In the future we would like to connect the competence model with the salary discussion. 

Peripheral intravenous catheterization is the invasive procedure most frequently performed in hospitalized children. The use of peripheral intravenous catheters (PIC) may increase the incidence of disorders such as phlebitis and extravasation. 1 The objective of this study was to verify the incidence and severity of peripheral intravenous catheter-related phlebitis in children.

Prospective, controlled study, whose sample was composed by 68 children. A protocol was implemented for the standardization of care in the PICU and surgical ward of a university hospital in Brazil. There were selected control variables related to children, professionals, and intravenous therapy characteristics. Phlebitis was analyzed according to four severity levels. 2 The ethics committee approved the research execution. To the statistical analyses were applied Chi-Square, ANOVA and Kruskal-Wallis tests, with a significance level of 5%. 

Clayton J A 1 , Rogers J 2 1 Acute Intensive Care Unit, South Manchester University Hospitals NHS Trust, 2 Acute Intensive Care Unit, South Manchester University Hospitals NHS Trust, Manchester, United Kingdom

This teaching programme was developed on a 14 bedded general intensive care unit/high dependency unit (ICU/HDU) in a South Manchester teaching hospital in response to national and local issues, namely the Workforce Development Confederation and Intercollegiate Boards initiatives on competency based training, and locally changing shift patterns and workload which were affecting our ability to provide quality teaching.

The authors reviewed the training needs of medical and nursing staff with particular attention paid to core competencies. Multi-disciplinary key topics were identified and a twelve week rolling programme developed. An ICU consultant and senior nurse were allocated to each topic and encouraged to develop and review the sessions. These sessions are designed to be flexible to reflect the learning needs of the individuals present. A range of interactive teaching methods are used.

Seven of the rolling teaching programmes have now been completed. An audit was undertaken after the second programme. All staff who had attended agreed that multi-disciplinary teaching was of benefit to both medical and nursing staff. The sessions ensure that medical staff is guaranteed their protected teaching time. However, there is some inequality as the nurses can only attend the session if the workload on the ICU permits this. Feedback from visiting speakers and participants suggests that there is improved communication between themselves and the ICU staff. In addition, there is some anecdotal evidence that sharing teaching sessions has promoted a greater understanding of the respective roles of the medical and nursing staff and improved teamwork. Interestingly, team working features as a key topic.CONCLUSION. The development of a multi-disciplinary teaching programme does not meet all our educational needs but does go some way towards them. The topics cover a wide range of subjects which can be offered to other staff groups and may be used to develop an outreach facility. Finally, despite being innovative, the multi-disciplinary teaching programme is cost neutral.

Hass H J 1 1 Intensive Care Unit, St Thomas' Hospital, LONDON, United Kingdom

There is a paucity of research into the use of agency nurses in intensive care.With the increasing numbers of nurses leaving full time employment, and the lack of critically care qualified nurses in intensive care units, it is timely to understand alternate forms of nursing employment.METHODS. This Heideggerian hermeneutic phenomenolgical study explored the lived experience of eight full time agency nurses working in the environment of intensive care in London. In-depth interviews were used to ascertain the participants' experience of full time agency nursing in intensive care. Colaizzi's (1978) method of data analysis was utilised. Thematic analysis identified four principal themes and these provided the framework for presentation of the results. The four identified themes were: Performance Issues, Attitudes of Staff, Allocation, and Education.

For the theme of performance issues, participants in this study identified the importance of support and understanding of agency nurses in the environment of intensive care. The theme of attitudes of staff identified that often agency nurses did not feel as though they 'belonged to a team'. Under the theme of allocation of agency nurses, the participants expressed concern at not having the opportunity of looking after higher acuity patients. Finally, the theme of education addressed the need for agency nurses to attend annual Basic and Advanced Life Support updates and the positive impact of on-going education on nursing care.

The findings of this study identify the need for managerial recognition of potential agency nurse deskilling and also maximising the agency nurse's clinical ability in the intensive care unit. The findings support the need for the implementation of clinical protocols within the intensive care setting, the provision of performance feedback for the agency nurses and on-going education for agency staff. This untapped area of nursing warrants further investigation.

Dinarés C G 1 , Fernandes S B 1 , Rocha P L 1 1 Polivalent Intensive Care Unit, Hospital Fernando Fonseca, Amadora, Portugal

Sleep is a vital component of health that affects quality of life. Adults spend approximately 1/3 of their lifes sleeping. Critical patients, admited into the ICU, frequently require for intensive treatment and monitorization during the 24 hours of the day. This can be one of the reasons for sleep disturbances. The incapacity to rest and to sleep can induce negative outcomes in critical patients. ICU nurses can and must promote the quality of treatment and the healing of patients providing them the opportunity to sleep.

Retrospective study, during the first term of 2003, with data analysis from two questionnairies: one made to ICU staff nurses, that worked in rounds, and another to patients admited in the same setting. The objective of this study is to determine the existence of disturbances in the quality of sleep on ICU patients.

In this survey a total of 60 patients, admited into the ICU during the same period, answered a questionnaire, as well as 26 nurses working in the same ICU. 79,5% of patients refers to have suffered from sleep disturbances during the time they were being treated on the ICU; 90% of nurses refered that a significant portion of patients have sleep disturbances. Agitation and desorientation were observed on most of the patients as the night falls.

Our data analysis supports that psychological stress and fear are associated with the ICU enviroment and severity of illness, causing in patients a difficulty on relaxation and falling assleep. The majority of patients referred that the quality and duration of sleep were deficient during the "ICU period". Nightmares, hallucinations and agitation were refered on 60% of the observed patients. Sleep is a state of relaxation necessary to all human beings; it is a universal and natural process of life. It is a state of serenity that restablishes the body and mind.