S132

17th Annual Congress – Berlin, Germany – 10–13 October 2004

Oral Presentations Therapeutic approached in acute lung injury: Experimental results – 504-509 504

506

TEZOSENTAN-INDUCED ATTENUATION OF LUNG INJURY IS ASSOCIATED WITH BLOCKADE OF PROTEIN KINASE C Kuklin V N1, Kirov M Y1, Sovershaev M A2, Ingebretsen O C3, Andreasen T4, Ytrehus K4, Bjertnaes L J1 1Department of Anaesthesiology, 2Department of Biochemistry, 3Department of Clinical Chemistry, 4Department of Physiology, Medical Faculty, University of Tromsø, Tromsø, Norway INTRODUCTION: Endothelin (ET-1), a potent vasoconstrictor peptide, is released by endotoxin in sepsis (1). When infused under experimental conditions, ET-1 enhances pulmonary microvascular pressure and permeability (2,3). Binding of ET-1 to its receptor activates the translocation of protein kinase C (PKC) from the cytosole to the membrane of pulmonary endothelial cells contributing to acute lung injury (ALI) (4). We recently demonstrated that the ET-1 receptor antagonist tezosentan (Actelion, Alschwil, Switzerland) ameliorates ALI in endotoxemic sheep, but the mechanism remains unclear (5). Our aim was to evaluate whether blockade of ET-1 receptors prevents activation of PKC, in addition to its reduction of lung microvascular pressure and permeability. METHODS: Seventeen yearling sheep were anesthetized and instrumented with vascular catheters. After one week of recovery, the animals were randomly assigned to a sham-operated group (n=3), a lipopolysa?charide (LPS) group (n=7) receiving intravenously (I.V.) infused E.coli LPS 15 ng/kg/min for 24 hrs, and a tezosentan group (n=7) subjected to LPS and from 4 hrs, tezosentan 3 mg/kg I.V. followed by infusion of 1 mg/kg/h for 24 hrs. Extravascular lung water index (EVLWI) was assessed by thermal-dye dilution (Cold Z-021, Pulsion Medical Systems, Germany). Pulmonary micro-occlusion pressure (Pmo), plasma concentrations of ET-1, TNF-a and IL-8 were determined every 4 hrs. Translocation of a and e isoforms of protein kinase C (PKC) was assessed by Western blotting technique. RESULTS: In both LPS groups, the plasma concentration of ET-1 correlated with a gradual increase in Pmo (n=19, r=0.47, P<.05). Pmo also correlated with the increase in EVLWI (p<.01). In the LPS group, we found a 65 – 75% reduction of the cytosole fraction of PKC compared to samples from sham operated sheep. Tezosentan reduced Pmo and EVLWI compared with the LPS group (p <.05), and completely prevented the reduction of PKC (?, ?). At the end of the experiment, plasma concentrations of ET-1, TNF-? and IL-8, all increased in the tezosentan group as compared with LPS alone (p <.05). CONCLUSION: In endotoxemic sheep, blockade of ET-1 receptors ameliorates ALI by reducing lung microvascular pressure and preventing the activation of protein kinase C. The increase in ET-1 is thought to result from ET-1 receptors that are saturated by tezosentan, and that further escalates cytokine production. REFERENCES: 1. Yanagisawa M et al: Nature 1988; 332:411-415. 2. Horgan MJ et al. Circ Res 1991; 69(1):157-64. 3. Helset E, et al. Circ Shock 1993; 39(1):15-20. 4. Siflinger-Birnboim A et al. Lab Invest 1992; 67(1):24-30. 5. Kuklin VN et al. Critical Care Medicine 2004; 32:766-773. Grant acknowledgement: Helse Nord project no 4001.721.132

INTRODUCTION: Assessment of recruitment manoeuvres (RMs) using increases in PaO2 and compliance in combination with a decrease in PaCO2 have been proposed as criteria for successful recruitment. We have used EIT to evaluate changes in global and regional ventilation during RMs.

505

507

INHALED AEROSOLIZED ACTIVATED PROTEIN C IMPROVES GAS EXCHANGE AFTER ENDOTOXIN-INDUCED LUNG INJURY

SPONTANEOUS BREATHING FAVORS DEPENDENT LUNG VENTILATION AND COUNTERS CYCLIC COLLAPSE IN PORCINE ALI

Waerhaug K1, Kuzkov V1, Kuklin V N1, Foss T2, Kirov M Y2, Bjertnaes L J1 1 Department of anesthesiology, 2Department of Radiology, Faculty of Medicine, University of Tromsø, Tromsø, Norway

Muders T1, Wrigge H1, Neumann P2, Zinserling J1, Varelmann D1, Hedenstierna G3, Putensen C1 Anesthesiology and Intensive Care Med., University Hospital, Bonn, 2Anesthesiology and Intensive Care Med., University Hospital, Göttingen, Germany, 3Clinical Physiology, University Hospital, Uppsala, Sweden

INTRODUCTION: We recently demonstrated in sheep that endotoxin (LPS)-induced lung injury is improved by infusion of activated protein C (APC). APC reduced lung microvascular pressure and extravascular lung water index (EVLWI), normalized blood coagulation and gas exchange and blocked the translocation of protein kinase C in lung tissue (1). We speculated that APC might act more selectively on gas exchange when inhaled. Our aim was to investigate if aerosolized inhaled APC improves gas exchange and reduces lung parenchymal densities in sheep subjected to LPSinduced lung injury. METHODS: Sheep were instrumented with intravascular introducers and underwent after 1 -5 days recovery, a second anaesthesia induced with propofol, fentanil and midazolam for endotracheal intubation and mechanical ventilation that was maintained throughout the experiment. A thermal dilution catheter (131HF7; Baxter, Irvine, CA) was inserted into the pulmonary artery for haemodynamic monitoring and another one into the thoracic aorta for determination of EVLWI employing PICCO (PV2014L16; Pulsion Medical Systems, Munich, Germany). After obtaining stabile baseline conditions, sheep were randomly assigned to two LPS groups with- and without APC (Drotrecogin alfa, Eli Lilly & Co, USA), 48 mc/kg/h dissolved in isotonic saline (LPS+APC, group, n=8 and LPS group, n=8), and a sham-operated group (n=3) was exposed only to nebulized saline for the 6 hrs experiment. APC and isotonic saline were aerosolized via a nebulizer (Servo Ultra 145; Siemens-Elema) from 30 min after the start of LPS until lung computed tomography at cessation of the experiment. Data analysis: ANOVA and Scheffe’s test; p< .05 was regarded as significant. RESULTS: Preliminary data show that APC significantly improves gas exchange, as judged by prevention of PaO2 from falling, reduction of the alveolar-arterial PO2 difference, increase in the PaO2/FiO2 ratio and reduction of venous admixture (p<.05). APC also prevented a rise in the right atrial pressure and tended to preclude an increase in lung microvascular pressure at 6 hrs (p<0.05t-test, one-sided). Moreover, lung computed tomography displayed less densities in the treatment group; a quantitative analysis will be presented. CONCLUSION: Aerosolized inhaled APC improves gas exchange in ovine LPS-induced ALI. However, further investigations are warranted to find out if inhaled APC could represent a beneficial adjuvant therapy for the patient with sepsis-induced lung injury. REFERENCE: 1. Kuklin VN, Kirov MY, Waerhaug K, Kuklina NA, Langbakk, B, Bjertnaes LJ. (Abstract) Acta Anaesthesiologica Scandinavia 2003 suppl 16, volume 47 page 32 Grant acknowledgement: Helse Nord project 4001.721.132

LUNG RECRUITMENT ASSESSED BY REGIONAL VENTILATION DISTRIBUTION USING ELECTRIC IMPEDANCE TOMOGRAPHY Odenstedt H1, Lindgren S1, Erlandsson K1, Olegård C1, Söndergaard S1, Lundin S1, Stenqvist O1 1 Anesthesia and Intensive Care, Sahlgrenska University Hospital, Gothenburg, Sweden

METHODS: In 6 pigs lung injury was induced by lavage. With 16 electrodes around the thorax impedance changes was monitored (Dräger/GOE MFII) and calibrated vs known lung volume changes (LV). Global and regional ventilation in ventral, upper mid, lower mid and dorsal regions of the lungs were assessed by impedance changes. In random order, three RMs; (1) sustained inflations at 40 cmH2O and (2) pressure control with peak airway pressure 40 and PEEP 20 cm H2O, rate 20, I:E 1:1 were repeated 3 times during 30 sec and (3) a slow RM with PEEP elevation to 15 and inspiratory pauses for 7 sec twice a min during 15 min was performed. Prior to each RM, baseline, “derecruited”, measurements were performed during volume control (VC) ventilation, tidal volume (TV) 10 ml/kg and PEEP 5. Following RMs VC was resumed with PEEP 10. Regional compliance was calculated at baseline and 10 min after RM as fraction of ventilation times TV divided by end-inspiratory minus end-expiratory tracheal pressure. RESULTS: Global impedance and LV increased significantly with the three RMs. Respiratory compliance for all RMs increased from 14±3 (mean±SEM) prior to RM to 22±1 10 min after RMs (p<0,0001). Following RMs compliance increased significantly in upper mid, lower mid and dorsal regions but was unchanged in the ventral region of the lungs.

CONCLUSION: Successful RMs were associated with redistribution of ventilation towards more dorsal and less aerated lung regions. Compliance increased due to increases in compliance in middle and dorsal parts of the lungs. EIT together with airway pressure gives information on global and regional lung compliance and is a promising tool for assessment of RMs in ALI.

1

INTRODUCTION: Experimental and clinical studies have shown reduction in intrapulmonary shunt with improved oxygenation by spontaneous breathing during airway pressure release ventilation (APRV) in acute lung injury (1,2). Oxygenation improvement can partially be explained by increased end-expiratory lung volume (3). We hypothesized that spontaneous breathing will result in better aeration and ventilation of dependent lung areas and less cyclic collapse during the tidal breath. METHODS: Two hours after induction of oleic acid-induced lung injury, 22 pigs were randomly assigned to APRV with (+SB) or without (-SB) spontaneous breathing at comparable airway pressures. Four hours after randomisation, spiral computed tomograms (CT) of the chest were performed at end-expiration and end-inspiration. Analyses of regional density were performed separately for dependent and non-dependent halfs of the lungs in end-expiratory and endinspiratory images. Tidal changes were assessed as differences between in- and expiration. RESULTS: Spontaneous breathing resulted in improved end-expiratory aeration in dependent lung regions (distribution of air: +SB: ventral 52%, dorsal 48%, -SB: ventral 59%, dorsal 41%, P<0.05). With SB there was also less non-aerated tissue in dependent regions at end-expiration (+SB: 175 ml, -SB: 324 ml, P<0.05) and less cyclic collapse (+SB: 25 ml, -SB: 55 ml, P<0.05) as compared with APRV without spontaneous breathing. During +SB, tidal ventilation was preferably distributed to dependent lung regions (distribution of tidal volume: +SB: ventral 35%, dorsal 65%, -SB: ventral 46%, dorsal 54%, P<0.05). CONCLUSION: Spontaneous breathing during APRV redistributes aeration and ventilation to dependent, usually well perfused lung regions and may thereby contribute to improved arterial oxygenation. Spontaneous breathing also counters cyclic collapse which is a risk factor for ventilator associated lung injury. REFERENCES: 1. Putensen C et al. Anesthesiology 81(1994) 921. 2. Putensen C et al. Am. J. Respir. Crit. Care Med. 159 (1999) 1241. 3. Wrigge H. et al. Anesthesiology 99 (2003) 376

17th Annual Congress – Berlin, Germany – 10–13 October 2004

508

S133

Oral Presentations From bench to bedside: Evaluation of therapy in sepsis – 510-515 510

REGIONAL COMPLIANCE CHANGES AFTER OPEN SUCTIONING USING ELECTRIC IMPEDANCE TOMOGRAPHY (EIT) Lindgren S1, Odenstedt H1, Olegård C1, Söndergaard S1, Lundin S1, Stenqvist O1 Anesthesia and Intensive Care, Sahlgrenska University Hospital, Gothenburg, Sweden

1

INTRODUCTION: To minimise negative side-effects of open suctioning (OSS), such as desaturation and lung collapse, closed system suctioning (CSS) has been proposed. CSS has been shown to have little suctioning efficacy (1). Thus, in patients with heavy airway secretion there may still be an indication for OSS. The time course of lung collapse and the regional effects of disconnection and OSS were studied. METHODS: In 7 anesthetised pigs weighing ~28 kg acute lung injury (ALI) was induced by lung lavage. Regional impedance changes (~aeration), in a ventral, upper mid, lower mid and a dorsal slice of the lungs were monitored using EIT by 16 electrodes around the thorax (Dräger/GOE MFII). Airway pressure was measured continuously via a pressure line inserted through the endotracheal tube in the trachea. Regional compliance (Creg) was calculated, using the regional impedance values as representative for each region for the whole lung in ventral-dorsal direction. Creg was calculated at base line, 30 seconds, 2 minutes and 10 minutes after suctioning as fraction of ventilation times tidal volume (TV) divided by the end inspiratory minus the end expiratory tracheal pressure. OSS was performed with a 14 Fr catheter at 10 cmH2O PEEP, RR 20, I:E 1:2 and TV 10 ml/kg during volume and pressure control ventilation (VCV, PCV). RESULTS: Lung collapse is pronounced immediately after disconnection and increases further during OSS. Compliance of the lower mid lung region decreased significantly more (p<0,01) than in the other three regions. Restitution of lung aeration was significantly slower during PCV as compared to VCV 30 s after suctioning (p<0.01). ANOVA. (n=7). See fig.

HES 130/0.4 AND GELATIN 4% IMPROVE INTESTINAL MICROVASCULAR PERFUSION DURING NORMOTENSIVE SEPSIS Adolphs J1, Schäfer T1, Ahmed R1, Mousa S A1, Habazettl H2, Schäfer M1, Welte M3 1 Anesthesiology and Critical Care, 2Physiology, Charité - Universitätsmedizin Berlin, Campus Benjamin Franklin, 3Anesthesiology and Critical Care, Klinikum Darmstadt, Berlin, Germany INTRODUCTION: Colloid solutions, such as hydroxyethylstarch (HES) and gelatin, have been suggested to improve rheology1 and to preserve plasma volume compared to crystalloid solutions in the presence of capillary leakage2. Additionally, in vitro3,4 and in vivo5 experiments have shown beneficial effects of HES on endothelial activation and leukocyte adhesion after an inflammatory stimulus. We, therefore, tested the hypothesis that HES and gelatine attenuate the impairment of gut microvascular perfusion and influence mesenteric leukocyte-endothelium interaction in a rat model of normotensive endotoxemia. METHODS: After approval of the animal care committee rats (n=7 per group) were anesthetized (urethane/ketamine) and normotensive sepsis was induced by a continuous infusion of LPS (1.5 mg/kg/h). After 1 h of LPS infusion HES 6% 130/0.4 (16 ml/kg), gelatin 4% (16 ml/kg) or NaCl 0.9% (66 ml/kg) were infused over a 1 h period. Using intravital microscopy, functional capillary density and erythrocyte velocity were measured in the mucosa of the terminal ileum at baseline and 3 h after fluid resuscitation. In another set of experiments the number of rolling, adhesive and transmigrated leukocytes was determined in mesenteric venules 3 h after fluid resuscitation. RESULTS: Macrohemodynamic parameters were comparable between groups throughout the experiment. 3 h after fluid resuscitation functional capillary density decreased in NaCl (808 [749/843] cm-1) (median [quartiles] but not in HES (995 [945/1036] cm-1) and gelatin (988 [867/1193] cm-1) groups (P<0.05, ANOVA), whereas erythrocyte velocity did not differ between groups. Leukocyte-endothelium interaction was not notably influenced by either treatment. CONCLUSION: During early sepsis colloid solutions have beneficial effects on microvascular perfusion of the gut, which were comparable between HES and gelatin in our study. Modulation of the immune response to LPS infusion was not observed after either treatment.

CONCLUSION: The lower mid region of ALI lungs is most affected by disconnection and suctioning with marked decreases in compliance. As restitution of lung aeration after OSS is better during VCV it may be advisable to use VCV after OSS as an easy way of recruitment of suctioning induced lung collapse.

REFERENCES: 1 Pries et al. Cardiovasc Res 1998; 37:225-235. 2 Marx et al. Int Care Med 2002; 28:629-635. 3 Collis et al. Int Care Med 1993; 20:225-235. 4 Nohé et al. Eur Surg Res 2002; 34:364-372. 5 Hoffmann et al. Anesthesiology 2002; 97:460-70

REFERENCE(S): (1) Lindgren S, Almgren B et al. (2004). Intensive Care Med On-line publ. feb. 2004.

Grant acknowledgement: Financial support was provided by FRESENIUS Kabi Deutschland GmbH.

509

511

EFFECT OF SPONTANEOUS BREATHING ON VA/Q MISMATCH IN PRIMARILY PULMONARY AND EXTRAPULMONARY ALI

CAN ALBUMIN PROTECT THE LUNG IN SEPSIS?

Varelmann D1, Wrigge H1, Zinserling J1, Muders T1, Hedenstierna G2, Putensen C1 1 Dept. of Anesthesiology, University of Bonn, Bonn, Germany, 2Dept. of Clinical Physiology, University of Uppsala, Uppsala, Sweden INTRODUCTION: Acute lung injury may be caused by direct (pulmonary) damage originating from the alveolar epithelium with decreased lung compliance or by indirect (extrapulmonary) injury with damage originating from the vascular endothelium which is often associated with decreased chest wall compliance e. g. due to sepsis with elevated abdominal pressure (1). Reopening of consolidated lung regions by mechanical ventilation strategy seems to be more difficult in direct lung injury (2).We hypothesized, that beneficial effects of spontaneous breathing on gas exchange are not dependent on type of lung injury. Therefore, we examined effects of spontaneous breathing on VA/Q mismatch in pigs with different types of lung injury. METHODS: 44 pigs of which 35 reached the study’s endpoint were randomized either to direct pulmonary injury resulting from acid aspiration or to primarily extrapulmonary lung injury realized by increasing abdominal pressure in combination with oleic acid injections. Both groups were treated with airway pressure release ventilation (APRV) either with or without spontaneous breathing. Spirometric variables were obtained using pneumotachography, and VA/Q distributions were determined using multiple inert gas elimination technique (MIGET). RESULTS: PaO2/FiO2, cardiac output (CO), shunt fractions (in %) and deadspace ventilation (Vds) are presented in table below. Pulm: pulmonary lung injury, extra: extrapulmonary lung injury. There was no significant effect of lung injury type on presented variables. Subgroup analysis revealed that animals with a PaO2/FiO2<150mmHg benefited even more from spontaneous breathing (p<0.05) as compared with animals with PaO2/FiO2>150mmHg. *: p<0.05, Baseline (BL) vs Treatment (Tr). #: p<0.05, effect ventilation, spontanous breathing (sb) vs controlled ventilation (cv) PaO2/FiO2mmHg PaO2/FiO2mmHg CO l/min CO l/min Shunt % Shunt % Vds % Vds %

pulm extra pulm extra pulm extra pulm extra

Sb BL 132 ± 18 145 ± 51 4.1 ± 0.3 4.2 ± 0.7 27.1 ± 6.2 32.6 ± 18.9 33.0 ± 5.5 39.1 ± 6.6

Sb Tr 150 ± 51 184 ± 11 4.6 ± 0.8 4.3 ± 0.8 23.3 ± 12.7 26.0 ± 17.9 45.1 ± 11.8 44.9 ± 12.8

Cv BL 151 ± 58 149 ± 73 4.2 ± 0.9 5.0 ± 0.5 27.7 ± 7.9 32.4 ± 12.4 34.4 ± 5.9 39.0 ± 6.0

Cv Tr 137 ± 10 107 ± 49 4.8 ± 0.9 5.4 ± 0.7 37.4 ± 17.4 47.2 ± 17.1 38.7 ± 3.9 46.2 ± 12.2

# # * * # # * *

CONCLUSION: Spontaneous breathing results in improved oxygenation and a reduction in shunt fraction independently of the lung injury model. These beneficial effects were stronger in more severe lung injury. REFERENCES: (1) Gattinoni L, Am J Respir Crit Care Med 1998; 158(1): 3-11. (2) Gattinoni L, Crit Care Med 2003; 31(4 Suppl): S300-S304.

Su F1, Wang Z1, Cai Y2, Nguyen N D u c3, Rogiers P4, Jean-Louis V1 1 Intensive Care Department, Erasme Hospital, 2Diabetes Center, 3Intensive Care Department, Vrije University, Brussels, 4Intensive Care Department, Middelheim General Hospital, Antwerp, Belgium INTRODUCTION: Whether fluid resuscitation with albumin can protect the lung in sepsis remains unclear: Zhang et al showed beneficial effects of albumin resuscitation in acute lung injury induced by hemorrhagic shock but not by endotoxic shock in rats (1), while Powers et al showed attenuated lung injury in a rat model of endotoxemia (2). The aim of this study was to investigate whether albumin protects the lung in poly-bacteria fecal peritonitis-induced septic shock in a ewe model. METHODS: Fourteen fasted, anesthetized, invasively monitored, mechanically ventilated 3.5 kg) received 0.5 kg/kg body weight of faeces into the±female sheep (29.3 abdominal cavity to induce sepsis. During the surgical operation and for 4 hours after surgery operation, Ringer’s lactate (RL) was administered. The animals were then randomized to receive continuous infusions of either RL alone or in combination with 20% albumin (1:10; the albumin was kindly provided by Baxter Company). Fluid resuscitation was titrated to maintain pulmonary artery occlusion pressure (PAOP) at baseline levels throughout the experiment without administration of antibiotics or vasoactive drugs. All of the animals were monitored until their death. RESULTS: The volume of RL required was significantly lower in the albumin group than in the RL group (20 ± 2 vs 14 ± 2 L, p<0.05). Serum albumin and total blood protein concentrations were maintained significantly higher in the albumin group than in the RL group (p<0.05). Colloid oncotic pressure (COP) was maintained 10 mmHg higher in the albumin group than in the RL group (p<0.05). There were no differences between groups in plateau pressure, respiratory system compliance, resistance, PaO2/FiO2, and lung wet/dry ratio. No difference was found in haemodynamic values or arterial blood lactate concentration between groups. 2.2± 4.0 vs 17.7 ±Survival times in the albumin group and RL groups were 19.7 hours, respectively (p=0.26). CONCLUSION: In this clinically relevant model of septic shock, albumin administration did not significantly influence lung function, haemodynamics, or survival time despite better maintaining serum albumin and total blood protein concentrations and COP. REFERENCES: 1. Zhang H, Voglis S, Kim CH, Slutsky AS. Effects of albumin and Ringer’s lactate on production of lung cytokines and hydrogen peroxide after resuscitated hemorrhage and endotoxemia in rats. Crit Care Med 2003; 31(5): 1515-22. 2. Powers KA, Kapus A, Khadaroo RG, et al. Twenty-five percent albumin prevents lung injury following shock/resuscitation. Crit Care Med 2003; 31(9): 2355-63.

S134

17th Annual Congress – Berlin, Germany – 10–13 October 2004

512

514

EFFECTS OF DOPEXAMINE, DOPAMINE OR PLACEBO ON RENAL FUNCTION IN PATIENTS WITH SEPTIC SHOCK

IMMUNOPARALYSIS IN SEVERE SEPSIS RESOLVES AFTER GM-CSF: A DOUBLE BLIND, RANDOMIZED, CONTROLLED TRIAL

Schmoelz M1, Irlbeck M1, Dunker M1, Schelling G1 Department of Anaesthesiology, Ludwig Maximilians University, Munich, Germany

Montag B1, Zapfe A1, Steigemann T1, Frings D P1, Staude H2, Jung R3, Pothmann W1, Nierhaus A1. 1Anesthesiolgy, 2Pharmacy, 3Clinical Chemistry, University Hospital Eppendorf, Hamburg, Germany

1

INTRODUCTION: To maintain adequate systemic perfusion pressure, patients with septic shock routinely require the infusion of vasopressors, such as norepinephrine, which induces vasoconstriction and may result in a decrease of renal blood flow. Additional vasoactive agents, especially dopexamine and renal-dose dopamine, may have a role in increasing renal perfusion, thereby protecting the kidneys from further injury. In a prospective, randomized, double-blinded, controlled trial, we compared the effects of the ß2- and DA1/DA2-receptor-agonist dopexamine, renal-dose dopamine and placebo in 61 Patients with vasopressor therapy and sepsis. Endpoints were creatinine-clearance, and urine-output. METHODS: Patients received either dopexamine (2 µg/kg/min), dopamine (3 µg/kg/min) or normal saline in a double blinded fashion in addition to standard therapy of septic shock. Creatinine-clearance, urine-output and total daily diuretic dose were measured at baseline (T0), and after 24 h (T1), 48 h (T2) and 72 h (T3). All patients required treatment with norepinephrine in a dose of at least 0.05 µg/kg/min. In a post-hoc analysis, creatinine-clearance was calculated in those patients, who had a baseline clearance under 85 ml/min. RESULTS: There was no significant difference between groups in the number of patients requiring renal replacement therapy. Creatinine-clearance increased significantly in the dopexamine group from T0 to T1 by 15.4% from 78.9±38.3 ml/min to 91.1±45.0 ml/min (p=0.02, n=16) and in the dopamine group by 17.5% from 86.9±55.0 ml/min (T0) to 102.2±58.5 ml/min (T1)(p=0.02, n=18). No significant changes were found in the placebo group. Quantitative urine output increased significantly over the entire study period only in the dopamine group from 4831.7±2294.2 ml/24h (T0) to 5336.4±2703.4 ml/24h (T3)(p=0.03, n=18), while there was no significant change in the dopexamine group (n=17) and placebo group (n=15). Applied doses of diuretics were not significantly different between the three groups at any point of time (p>0.05). In patients with a baseline creatinine-clearance under 85 ml/min, both, dopexamine and dopamine induced a significant increase in creatinine-clearance (52.11±24.24 ml/min (T0) to 70.4±44.8 ml/min (T3) (p=0.01, n=9) and 56.1 ±16.7 ml/min (T0) to 80.7±45.0 ml/min (T3) (p=0.01, n=10), respectively). We found no significant changes in the placebo group. CONCLUSION: In norepinephrine-dependent patients with septic shock, treatment with renaldose dopamine increases urine output and creatinine-clearance. Dopexamine improves only creatinine-clearance but has no effect on urine output and serum-creatinine. Especially in patients with impaired renal function, both drugs have been shown to markedly improve creatinineclearance. Renal-dose dopamine or dopexamine in addition to standard therapy of septic shock may have beneficial effects on renal perfusion and improves renal function.

INTRODUCTION: Immunoparalysis as defined by a low monocytic HLA-DR expression is associated with a high mortality rate and severity (1). Preliminary studies had suggested that granulocyte-macrophage colony stimulating factor (GMCSF) could restore the HLA-DR expression and enhance monocyte function (2,3). This study was performed to evaluate whether GMCSF in patients with severe sepsis and immunoparalysis can counteract monocytic anergy. METHODS: After ethical approval 2x16 consecutive patients with severe sepsis and a HLA-DR expression below 200 mean fluorescence intensity (MFI) over 48h were included in a double blind RCT. GMCSF (4µg/kg) was continued until HLA-DR expression increased above 500 MFI >48 h and restarted when MFI<=200. Observation period was 21 days. HLA-DR expression was assessed daily by the MFI method, QuantiBrite® (ABC) and in percentage as described before (3,4). Additionally we evaluated plasma levels of GMCSF and IL6, TNFa ex vivo production, APACHE, MODS, ICU mortality, appearance of infection (CDC) and adverse reactions to therapy. RESULTS: Main diagnoses were peritonitis, pneumonia or pancreatitis (31%, 28%, 21%, resp.). Overall ICU mortality was 10/32 (33%)(3 patients still in ICU). The groups showed no differences in mortality, severity, gender, age, length of stay, respiratory function or incidence of mycosis. The treatment group (TG) showed a significant increase of MFI values and white blood count 24h after treatment start until day 6 and 20, resp., compared to the placebo group (PG). This was accompanied by a significant increase of GMCSF plasma levels. All patients of the TG reacted within 48h to treatment, only one patient reacted delayed. MFI values showed a significant increase between 24h, 48h and 72h ((mean±SEM)121±24, 300±41, 443±52; p<0.05, resp.) after treatment start. Values of PG increased spontaneously over time. HLA-DR expression, measured as ABC and percentage, as well as TNFa ex vivo production were also increased in the TG. There was no increase in IL 6 plasma levels following application of GMCSF. Values for MODS tended to be higher in TG. There were no serious adverse events in the TG. CONCLUSION: In this study the application of GMCSF could rapidly and permanently restore HLA-DR expression in all patients of the treatment group. Due to the small sample size, no definite conclusion can be made concerning mortality. However, the treatment was well tolerated and the result should stimulate a larger clinical trial. REFERENCES: 1. Saenz et al.(2001)Intensive Care Med 2. Nierhaus et al. (2003) Intensive Care Med 3. Flohé et al.(1999)Shock 4. Montag et al. (2003) Intensive Care Med (abstract) Grant acknowledgement: This study was supported in part by grants of DPC Biermann and Sysmex Int..

513

515

SAFETY AND EFFICACY OF DROTRECOGIN ALFA (ACTIVATED) IN ADULT SURGICAL PATIENTS WITH SEVERE SEPSIS

RANDOMISED CONTROLLED TRIAL TO THE EFFECTS OF LAXATIVES IN A MIXED ICU POPULATION

Payen D1, Williams M D2, Sarwat S2, Janes J3 1 Service d’Anesthésie-Réanimation, Hôpital Lariboisière, Paris, France, 2Lilly Research Laboratories, Eli Lilly and Company, Indianapolis, United States, 3Lilly Research Laboratories, Eli Lilly and Company, Erl Wood, United Kingdom

Van der Voort P H J1, Van Roon E N2, Van der Linden C T M3, Van der Spoel J I4 1 Intensive Care, 2Clinical pharmacology, Medical Centre Leeuwarden, Leeuwarden, 3Clinical pharmacology, 4Intensive Care, OLVG, Amsterdam, Netherlands

INTRODUCTION: Severe sepsis is an important and serious potential complication of surgery. Drotrecogin alfa (activated) [DrotAA] is indicated for the reduction of mortality in patients with severe sepsis and multiple organ failure. Using data from 2 global clinical trials of DrotAA, we compared clinical characteristics, efficacy and incidence of serious bleeding in surgical (SURG) severe sepsis patients (defined by recent surgery within 30 days of severe sepsis onset or postoperative patients, based on APACHE II) versus non-surgical (NONSURG) severe sepsis patients. METHODS: Enrolment criteria were similar for patients from PROWESS (randomized placebocontrolled trial, n=840 placebo patients, n=850 DrotAA patients) and ENHANCE (single-arm, open-label trial, n=2378 DrotAA patients). Severe sepsis was defined as a known or suspected infection, 3 or 4 criteria for systemic inflammatory response syndrome, and >1 sepsis-induced organ dysfunction (OD). DrotAA or placebo was administered as a continuous 96-hour infusion at 24µg/kg/hr, n=1505 for SURG, n=2563 for NONSURG. Efficacy and safety data were collected during an identical 28-day follow-up period in each trial. The primary endpoint was all-cause mortality at 28 days. RESULTS: SURG patients were significantly older than NONSURG patients (mean: 63 vs. 58yrs, p<0.0001), and had a greater number of OD at baseline (mean: 2.7 vs. 2.5, p<0.0001). A higher percentage of SURG patients also required vasopressors (75% vs 65.5%, p<0.0001) and mechanical ventilation (90% vs 73%, p<0.0001), compared to NONSURG. The most common site of infection was significantly different between the two groups: intra-abdominal infection was the predominant site for SURG patients (SURG: 54%, NONSURG: 4.4%) versus lung for NONSURG patients (SURG: 24%, NONSURG: 64%). A treatment effect consistent with that of the overall population was seen in both SURG and NONSURG patients, the odds ratio for death for DrotAA versus placebo was 0.81 (95% CI 0.61-1.08) for SURG patients and 0.72 (95% CI 0.59-0.88) for NONSURG patients. The rate of serious bleeding during the infusion period in patients receiving DrotAA was higher for SURG than NONSURG (% patients experiencing a serious bleeding event, SURG: 4.9%, NONSURG: 2.6%), and was higher in both groups compared to placebo (SURG: 0.4%, NONSURG: 1.2%). CONCLUSION: Surgical patients are an important subgroup of severe sepsis patients. SURG patients with severe sepsis tended to be older and have greater disease severity relative to NONSURG patients. The survival benefits of DrotAA were consistent across SURG and NONSURG patients, although bleeding rates were generally higher in SURG patients.

INTRODUCTION: Constipation in ICU patients is a frequent clinical problem, but no comparative studies on effectiveness of medication are available. METHODS: Ventilated patients without defecation during the first two days of ICU-stay in two hospitals were eligible for inclusion into this prospective, randomised, double-blind, placebocontrolled study. Medication in study arms consisted of placebo tid, lactulose 13.3 g tid or macrogol 13.81 gr (Movicolon®) tid, all dispensed as 100 ml dilutions in water. Patients were treated with study medication for 88 hours (11 administrations). End point was (time to) defecation of at least one stools of >100 mL. The study was approved and informed consent was obtained. Chi-square, Fisher-exact test, one-way ANOVA and Kaplan-Meier estimates were used where appropriate. RESULTS: In 26 months 317 patients were included. 57 patients (18%) were censored due to death, early discharge or withdrawal from treatment for other reasons than defecation, leaving 260 patients for evaluation (53% male, mean age 67 years (range 14-92)). Overall, 31, 75 and 66% of patients had defecation on placebo (n=84), lactulose (n=89) and macrogol (n=87), respectively. Lactulose and macrogol both showed significant differences versus placebo (p< 0.0001), but no mutual difference (p=0.19). Mean time to defecation in those patients passing stools during the study period was 79, 47 and 57 hours after start of medication in the placebo, lactulose and macrogol group, respectively. Lactulose (p<0.0001) and macrogol (p=0.003) both showed significant differences versus placebo, but no mutual difference (p=0.37). The figure shows the Kaplan-Meier estimate for time to defecation.

CONCLUSION: Patients on lactulose and macrogol defecated more frequently and earlier compared to placebo. No significant differences between the two active treatments were observed.

17th Annual Congress – Berlin, Germany – 10–13 October 2004

Oral Presentations Prognostic markers of the circulation – 516-521 516 NEUROLOGIC INJURY AFTER CARDIOPULMONARY BYPASS SURGERY: THE ROLE OF PAI-1 4G/5G POLYMORPHISM SIRGO G1, Perez Vela J J L1, Renes E E1, Garcia A A1, Hernández E E1, Morales P P2, Del Rey M M2, Perales N N1 1 Critical Care Unit, 2Departament of Immunology. Genetics Laboratory., 12 Octubre University Hospital, Madrid, Spain INTRODUCTION: Molecular Biology has revolutionized Medicine by increasing our understanding of the pathophysiological mechanism of disease and our ability to assess genetic risk. Neurological injury is a significant risk for patients undergoing cardiac surgery. The outcome of cardiac surgery is influenced by several factors, but the impact of specific genetics variants has no been systematically explored. METHODS: Prospective observational study. A total of 150 patients undergoing cardiac surgery with cardiopulmonary bypass were enrolled. DNA extraction and 4G/5G PAI-1 genotyping was carried out using published techniques. We hypothesized that 4G/5G polymorphism is as significant risk for cardiac surgery neurological injury: stroke, encephalopathy, delirium and confusion. RESULTS: The 4G/4G genotype was significant associated with neurological injury (4G/4G 38.5% vs No-4G/4G 14.55, p<0.05). (OR 3.5 CI 95% 1.39-8.8). The 4G/4G geotype was also associated with postoperative prolonged mechanical ventilation (> 12h)(p<0.05).This genotype was no associated with other complication (renal dysfunction, postoperative myocardial infarction). CONCLUSION: This study showed that patients carrying the functional PAI-1 4G/4G are at increased risk of developing neurological injury and prolonged mechanical ventilation. Identification of genotypes predictive of adverse perioperative events may not only further our understanding of pathophysiological response to surgery but also potentially decrease surgical morbidity and mortality via preoperative risk assessment. Grant acknowledgement: MAPFRE CARDIOVASCULAR (2003)

S135

518 GASTRIC TONOMETRY AND PREDICTION OF WEANING FAILURE Guven M1, Baysal O F2, Sungur M1, Buyukoglan H3 Intensive Care, 2Internal Medicine, 3Pulmonary Medicine, Erciyes University, Medical Faculty, Kayseri, Turkey

1

INTRODUCTION: The aim of study was to measure the changes in pHi and gastric intraluminal PCO2 (PiCO2) during weaning, and to determine the value of variables to predict weaning failure. METHODS: Forty (20 m, 20 f) patients considered ready to be weaned from mechanical ventilation (MV) were studied. The study was approved by the Institutional Review Board. An informed consent was obtained. All patients were on pressure support MV for at least 2h before weaning trial. PiCO2 was determined with a tonometer and monitor. pHi was calculated by monitor. At the beginning of the weaning, MV was discontinued and extubation was performed after 2h of T-piece trial. Patients were followed for at least 24h. RESULTS: While 28 patients were successfully weaned, weaning was failed in 12 patients. All initial values were similar in both groups. Failure developed at 2ndh after weaning in 4, at 10th hour in 2, at 20th hour in 4 and 24th hour in 2 patients. Four patients died. Initial pHi (7.39±0.1 vs 7.37±0.1;NS) and PiCO2(47.01±11.0 vs 46.87±20.3;NS) values were similar in both groups. At postweaning 2ndh, there was an increment in PiCO2 in successful (49.20±18.8) and failure(55.80±20.5) groups. pHi (6.57±1.9 vs 7.319±0.1; p=0.04) were lower and PiCO2 (57,41±23,0 vs 46,17±13,7; p=0,09) were higher at postweaning 6thh in failure group. The changes after 6thh were similar. Initial and final pHi and PiCO2 values before reintubation were compared. pHi decreased from 7.39±0.1 to 7.25±0.2 and PiCO2 increased from 47.01±11.0 to 57.53±19.4 during the weaning in failure group(p=0.038, P=0.107; respectively). At the same time, pHi and PiCO2 showed a nonsignificant change from 7.37±0.1 to 7.33±0.2 and 39.70±13.0 to 46.87±20.3 (p=0.26, p=0.08;respectively) in success group. Final threshold value for pHi was <=7.32 (sens=83.3%; spec=57.1%). These value for PiCO2 was > 47 mm Hg (sens=75%; spec=67.9%). The best parameters that showing weaning failure were found as, final pHi and final PiCO2 levels. However, there were no any significant differences between post-weaning 6thh pHi (0.64) or PiCO2(0.67) and final pHi (0.63)or PiCO2 (0.69) AUCs (p=0.86, p=0.80,respectively). CONCLUSION: The results in both groups may be similar first 2h at postweaning period, since respiratory effort, oxygen consumption and production of CO2 increased. Thereafter, these parameters improve in patients who adapt spontaneous respiration and have adequate gas exchange. However, negative events continue in failure group. This leads to difference at following hours. Weaning failure was associated with gastric intramucosal acidosis. pHi and PiCO2 may be helpful to predict weaning failure. The results obtained after 6thh do not give more information. Thus, a follow-up period consisting 6h may be adequate to asses weaning failure. REFERENCE: 1-Ann Intern Med 1993;119:794.2-Crit Care Med. 2001;29:70.

517

519

IMPACT OF ELEVATED HEART RATE ON INCIDENCE OF MAJOR CARDIAC EVENTS IN CRITICALLY ILL PATIENTS

INCREASED L-LACTATE IN THE RECTAL LUMEN IN PATIENTS WITH SEPTIC SHOCK IS RELATED TO SHOCK SEVERITY

Sander O1, Welters I D2, Foex P1, Sear J W1 1 Nuffield Dept of Anaesthetics, University of Oxford, Oxford, United Kingdom, 2Surgery, Anaesthesiology and Surgical Intensive Care Medicine, Justus-Liebig-University, Giessen, Germany INTRODUCTION: Heart rate is a key determinant of myocardial oxygen consumption. Critically ill patients are prone to develop tachycardia due to increased sympathetic tone, arrhythmogenesis or inadequate sedation and pain control. However, no information is available yet on the critical duration of tachycardia associated with an increased cardiac complication rate. Therefore, we hypothesised that the incidence of major cardiac events is increased in critically ill patients with high risk of cardiac complications presenting with elevated heart rate for a prolonged period. METHODS: In this prospective observational cohort study over a period of one year, we investigated patients with high risk of cardiac complications according to the Revised Goldman Index (Lee et al. 1999). Patients were selected when they were treated for at least 36 hours in the 15-bed medical/surgical Intensive Care Unit at the John Radcliffe Hospital, Oxford. Those patients developing prolonged elevated heart rate defined as heart rate > 95 bpm for > 12 hours in at least one 24 hour period of their ICU stay were investigated. Cardiac high risk patients not presenting with this criterion served as control. The number of patients with major cardiac events, defined as myocardial infarction (Alpert et al. 2000), non-fatal cardiac arrest and cardiac related death, was the primary outcome parameter which was statistically analysed by Fisher’s Exact test. Secondary outcome parameters were length of stay, ICU mortality and 30 days post ICU discharge mortality. The sample size required to test the above defined study hypothesis was calculated as being 2 x 27 patients assuming a statistical probability of alpha = 0.05 and beta = 0.2.

Reiter N1, Jørgensen V L2, Perner A2 Dept of Intensive Care, Rigshospitalet, 2Dept of Intensive Care, Herlev Hospital, Copenhagen, Denmark

1

INTRODUCTION: In a small study of patients with septic shock, the concentrations of L-lactate was observed to be markedly increased in the rectal lumen and strongly related to the dose of norepinephrine (1). It is, however, unclear if disease severity or norepinephrine in it self caused the observed metabolic changes. METHODS: Twenty-one patients with established septic shock (>24 h) and a normal rectum were included. All patients were fluid resuscitated guided by central venous O2 saturation and systolic pressure variation and received antibiotics, shock reversal steroid and norepinephrine (range 0.05 - 0.64 µg/kg/min) to maintain a MAP > 70 mmHg. The concentration of L-lactate in the rectum was estimated by luminal equilibrium dialysis. A bag of dialysis tubing (cut-off 12 kDa; Sigma, St. Louis, MO), containing 4 ml of 10% Dextran 40, was placed in the rectal lumen for 4 h to obtain full equilibrium. Dialysate concentrations of L-lactate were determined by autoanalyser (ABL 725, Radiometer, Copenhagen). The severity of disease was graded by the simplified acute physiology score (SAPS) II and the sequential organ failure assessment (SOFA) score. RESULTS: In patients with septic shock (mean SAPS II 56 and SOFA 15), the mean concentration of dialysate L-lactate was well above that previously observed in healthy subjects (mean [SD], 3.4 [2.5] vs 0.8 [0.5] mM; p<0.001). The dialysate concentration of L-lactate was related to the dose of norepinephrine (linear regression analysis, r2= 0.29, p=0.01, see Fig), but not to the SAPS II (p=0.7) or SOFA score (p=0.40).

RESULTS: From a total of 791 patients 69 were assessed as cardiac high risk patients. From 39 patients with prolonged elevated heart rate 19 (49%) sustained major cardiac events, whereas in the control group with 30 patients only 4 (13%) were concerned (p = 0.002, Odds ratio 6.2). Patients with elevated heart rate were treated 3.5 days longer at ICU (p = 0.01), whereas ICU and 30 day post ICU discharge mortality did not differ at a significant level. CONCLUSION: In this study, we provide data for an increased incidence of major cardiac events in critically ill cardiac high risk patients presenting with a prolonged elevated heart rate during their ICU stay. Additionally, elevated heart rate was associated with a significant longer ICU stay. A further study is planned to investigate if strict heart rate control can reduce the number of cardiac events in this subset of patients. REFERENCES: 1. Alpert, J. S. et al. (2000). J Am Coll Cardiol 36: 959-69. 2. Lee, T. H. et al. (1999). Circulation 100: 1043-9

CONCLUSION: In septic shock, concentrations of L-lactate are markedly increased in the rectal lumen, but the weak correlation to norepinephrine dosage suggests that other factors contribute at least in fluid resuscitated patients. In any case, the results indicate that metabolic dysfunction observed in the rectum is related to the severity of shock rather than the severity of disease. REFERENCE(S): 1. VL Due et al. BJA 2002;89:919.

S136

17th Annual Congress – Berlin, Germany – 10–13 October 2004

520 POSTOPERATIVE COMPLICATIONS AND INTRAOPERATIVE HAEMODYNAMIC PARAMETERS IN LIVER TRANSPLANTATION Ther G1, Fazakas J1, Mándli T1, Tóth T1, Gondos T2, Járay J3 1 Department of Transplantation and Surgery, Semmelweis University, 2Department of Anaesthesiology and Intensive Care, National Medical Center, 3Department of Transplantation and Surgery, Semmelweis University, Budapest, Hungary INTRODUCTION: Blood volume monitoring technique allows measurement of intrathoracic blood volumes, cardiac index (CI) and lung water during liver transplantation (OLTx). The aim of the study was to analyse the early major postoperative complications from the point of view of intraoperative haemodynamic parameters. METHODS: The relationships between intravascular filling pressures (CVP), intrathoracic blood volume (ITBI) and flow parameters (CI) were analysed in 74 liver transplanted patients underwent volumetric haemodynamic monitoring (PiCCO, Pulsion). Measurements were performed during liver transplantation: before incision (BI), during hepatectomy (H1-H2), in the anhepatic phase (AH), after portal and arterial reperfusion (PVR, HAR) and at the end of surgery (ES). The patients were divided into two groups: group A (without complications, n=29), and group B (with major postoperative complications including: artery thrombosis, bleeding, renal, graft and respiratory failure, n=45). Statistical analysis was performed with Statview program. RESULTS: There weren’t significant differences between the groups in the values of ITBI, CVP, MAP, CI measurements. The EVLWI levels were significantly higher in Group B then in Group A in the anhepatic phase and after venous reperfusions (AH: 6.27±2.57 vs. 5.29±2.46 ml/kg; PVR: 6.03±2.15 vs. 5.2±1.94 ml/kg, p<0.05). The pHi values were found significantly higher in group A compared to group B from anhepatic phase until the end of surgery (AH: 7.36±0.2 vs. 7.28±0.08; PVR: 7.3±0.18 vs. 7.24±0.1; HAR: 7.36±0.11 vs. 7.3±0.1; ES: 7.4±0.17 vs. 7.33±0.08, p<0.05). The DO2I levels were significantly higher in Group A then in Group B after both reperfusions (PVR: 679±172 ml/min/m2 vs. 601±223; HAR: 645±157 vs. 556±196 ml/min/m2, p<0.05). CONCLUSION: The haemodynamic parameters showed an adequate volume replacement therapy during OLTx. The worsening of the regional perfusion (pHi) during OLTx could be related to the occurrence of the early postoperative complications.

Oral Presentations Antibiotic policy – 522-527 522 INFLUENCE OF ANTIMICROBIAL RESTRICTION ON THE ANTIBIOTIC RESISTANCE PATTERN IN AN ICU Paramythiotou E1, Ntagiopoulos P1, Antoniadou A2, Poulakou G2, Tsaganos T2, Katsala D2, Koukoulitsios G1, Giamarellou H2, Karabinis A1 1ICU, G.Gennimatas general hospital, 24th department of internal medicine, Attikon University hospital, Athens, Greece

INTRODUCTION: Resistance of gram - negative microorganisms to a large spectrum of antimicrobials is a well known problem especially in ICUs, often related to antibiotic use and misuse. Antibiotic policies promoting prudent antimicrobial use are suggested measures to combat microbial resistance. Our purpose was to study the influence of an antibiotic policy program combining restriction of antibiotics and empiric therapy algorithms according to unit’s continuous colonization surveillance resistance data. METHODS: The study was conducted in the 12 - bed ICU of our hospital. After a 3 month (1 – 1 to 31 -3 - 2003) surveillance of nosocomial ICU infections and antimicrobial resistance – pattern of implicated microorganisms, we applied a restriction protocol according to which no quinolone (especially ciprofloxacin) or any third generation cephalosporin (especially ceftazidime) was used as empiric therapy in any case. An effort was made not to use empirical imipenem and algorithms for empiric therapy according to resistance data were used. Only patients hospitalized > 48 hours were included. After a 6 - month intervention period antimicrobial resistance pattern was recorded from 11 to 31- 3- 2004. RESULTS: During the “before” trimester, 38 patients were included (mean age: 55.3y, mean ICU stay: 23 d). Forty infections were recorded. Number of positive cultures: 50 (Acin. baum. 50%, Pseud. aerug. 30%, Klebs. pn. 20%). Sensitivity pattern: for imipenem/meropenem 38%, 44% and 42% for the three microorganisms correspondingly, ceftazidime 22% for Pseudomonas aerug. and 23% for Klebsiella pn., gentamicin 25%, 22% and 42%, ampicillin/sulbactam 62% for Acinetobacter baum. Finally ciprofloxacin presented 0% sensitivity for these three most often encountered gram - negative microorganisms. During the “after” trimester 36 patients were included (mean age 50.7 y, mean ICU stay 16.4d). A total of 75 infections were recorded. The same microorganisms were isolated (Acin.baum 54.7%, Pseudom.aerug. 30%, Klebsiella pn.15%). Sensitivities to microorganisms were modified as follows: for restricted antibiotics, a slight increase in sensitivity was observed for ceftazidime (36%, 25% and 5% correspondingly) but it was not statistically significant. For ciprofloxacin we noticed a statistically significant change (p <=0.05) for the first two microorganisms (22%, 35% and 4%). Sensitivity of imipenem ameliorated: 54%, 85% (p< = 0.05), 42%. On the contrary for antibiotics used more often as empiric therapy, a decrease in their sensitivity was noticed. CONCLUSION: Our results are encouraging about the favourable influence of restriction on the antibiotic resistance pattern, but they also show that interventions of longer duration are needed and probably a rotation antibiotic policy could be of value.

521

523

PROBNP LEVELS IN CARDIAC SURGERY: A NEW TOOL IN POSTOPERATIVE CARE?

PROSPECTIVE SURVEY OF THE FOLLOW-UP OF EMPIRIC ANTIMICROBIAL THERAPY IN 41 FRENCH ICUS

Forés Moreno G1, Reyes Copa G2, Rodriguez-Abella H3, Cuerpo Caballero G2, Romero C4, Lopez Lopez J2 1 Anestesiology, Gregorio Marañón, 2Cardiac Surgery, Gregorio Marañon, 3Cardiac Surgery, Gregorio Marañón, 4Clinical Chemestry, Gregorio Marañon, Madrid, Spain

Lepape A1, Veber B2, Gauzit R3, Bedos J4, Montravers P3 Réanimation Nord, Chu Lyon Sud, Pierre Benite, 2Anesthésie-Réanimation, Hôpital Charles Nicolle, Rouen, 3Anesthésie-Réanimation, Hôpital Jean Verdier, Bondy, 4Réanimation, Hôpital André Mignot, Versailles, France

INTRODUCTION: ProBNP levels increase in patients with ventricular dysfunction. The objective was to determine proBNP levels in patients undergoing cardiac surgery and to explore its correlation between the peptide and patients postoperative evolution in order to be used as treatment monitor and predictor of outcome.

INTRODUCTION: Empiric antibiotics therapy is widely used in ICU due to the severity of patients. However, it is a good practice to adapt this antimicrobial therapy to microbiological data when they exist or to the clinical evolution when indicated. This study is a part of a prospective survey aiming at describing the whole course of empiric therapy in ICU patients.METHODS: A prospective survey of 1 month was realized in 41 voluntary ICUs (November 2003). All patients were included and, if an empiric antimicrobial therapy was prescribed, the subsequent changes according to the reception of microbiological information described, as well as modifications for non-microbiological reasons.

METHODS: Prospective cohort analysis. All patients undergoing elective cardiac surgery in a 4 month period in 2003 were included. Preoperative, intraoperative and postoperative data were collected. Nine determinations of proBNP were measured:the first sample was obtained preoperatively, second sample the evening after surgery, next two samples on the first postoperative day and then every 24 hours until sixth day. RESULTS: A total of 86 consecutive patients were included. A Euroscore >6 was found in 30%of patients. NYHA classification was as followed: I: 27.7%;II 47%; III: 25.3 %.Preoperative AF was seen in 20.5% of patients. In the early postoperative period 18.1% of patients required inotropic agents. 34.9% of patients suffered complications, being the AF the most frequent (20.5%). Only one patient died during the study period. A statistically significant higher level of preoperative proBNP was observed in patients with, Euroscore>6(p=0.021),worse NYHA class(p=0.020)and preoperative AF (p=0,001).A statistically significant proBNP elevation was observed in postoperative day one up postoperative day six (p< 0,001). Highest levels of proBNP were found on day four (p< 0,001).These levels were significantly higher in patients who required inotropic agents after surgery (p<0,012).Complications rate was not associated with higher proBNP levels (p=0.59). CONCLUSION: Preoperative proBNP levels depend on preoperative patient status, with a significantly increase after cardiac surgery. There is a statistically significant correlation between patients who required inotropic agents and higher postoperative proBNP levels. There is no association between proBNP levels and rate of complications. Monitoring proBNP levels could be a useful tool to optimize the management of patients after cardiac surgery.

1

RESULTS: Overall, 1702 patients in 41 ICUs were included with 509 (29.9 %) receiving an empiric antimicrobial therapy (241 (47 %) for lung infections, 105 (21 %) for intra-abdominal infections). According to the final microbiological data, 83 % were at the end considered as adapted. Streamlining of initial antimicrobial therapy : in 204 patients (43 %), direct examination of the samples were available, leading to adaptations in 19/204 patients (9%). In 255 (58%), the identification of the microorganism was transmitted to the physician leading to adaptations in 58/255 patients (23 %). Finally, complete identification and susceptibility data were obtained in 260 (59 %) of empiric therapy. Adjustments were performed in 102/260 (41 %) of patients. In 27 % of these patients, changes were dictated by non microbiological reasons, such as narrowing spectrum (8 %), stopping aminoglycosides (7 %), clinical deterioration (4 %), and adverse events (1 %). CONCLUSION: The rate of adequate initial antimicrobial treatment is important in this study. The follow-up of empiric antimicrobial prescriptions reveals a high proportion of changes based on results of microbiological results and clinical course. Grant acknowledgement: Unresctricted grant from Wyeth-Lederlé

17th Annual Congress – Berlin, Germany – 10–13 October 2004

S137

524

526

FACTORS INFLUENCING SENSITIVITY AND SPECIFICITY OF THE SURVEILLANCE ON ICU-ACQUIRED INFECTIONS

INITIATION OF EMPIRIC ANTIBIOTIC THERAPY: A FRENCH EPIDEMIOLOGIC STUDY

Morales I1, Versporten A1, Suetens C1 1 Unit of epidemiology, Scientific Institute of Public Health, Brussels, Belgium

Montravers P1, Gauzit R1, Lepape A2, Colin-Mercier F3, Veber B4, Conscience G5, Pourriat J6 1 DAR, CHU Jean Verdier, BONDY, 2DAR, CHU Lyon Sud, Pierre Benite, 3Societe, Stat Process, Paris, 4DAR, CHU Charles Nicolle, Rouen, 5-, Wyeth Lederle France, Paris La Defense, 6DAR, CHU Hotel Dieu, Paris, France

INTRODUCTION: The Belgian surveillance of nosocomial infections in intensive care units (ICU) started in 1996. The validation followed in 2001 to evaluate the surveillance sensitivity, specificity, exhaustiveness and accuracy. The workload and factors influencing SS and SP were also studied. METHODS: Hospital data about surveillance policies and preventive measures were collected. Linking validation data to surveillance data permitted to study factors influencing sensitivity and specificity using a multiple logistic regression for the prediction of false negatives (FN) or false positives (FP) with correction for the ICU level. Profiles of participating and refusing units were compared. RESULTS: 35 units were included in the analysis. 906 patient records were validated. Overall surveillance sensitivity was 53.3% for pneumonia (PN) and 61.9% for bacteraemia. Overall specificity was 98.6% and 99.4% respectively. Global exhaustiveness was 81.2% for patients staying more than 48 hours in the ICU. 70% of persons involved in data collection were ICU staff (44% doctors and 26% nurses) and 25% were infection control practitioners. On average, 2 persons were involved in data collection which required 20 minutes per patient, depending on the type of patient and mode of surveillance. When the decision to participate in the surveillance was taken outside the ICU and the intensivist was not involved in data collection, the number of FN and FP was much higher (OR 6.4; p-value <0.001) compared to a reference situation of decision taken in the ICU with intensivist involved in data collection. Other characteristics in the model associated with higher numbers of FN and FP were the use of other PN diagnostic criteria, zero infection rates, presence of discordant values, and an overall high risk score for PN. Inclusion of early nosocomial PN, reporting only „high risk patients“ and PN without micro-organism were associated with less FN/FP. 10 units did not participate. They were not significantly different from participating ones on ICU characteristics like mean LOS, missing or discordant data, invasive device use rates and bacteraemia rates. PN incidence density was significantly higher among nonparticipating units (9.0 v 12.3 PN per 1000 patient-days p=0.04). CONCLUSION: ICU surveillance is mainly performed by ICU staff. Their involvement is a key determinant for good quality data. The use of too specific PN diagnostic criteria had a marked effect on sensitivity. The study showed that a zero infection rate suggests non-reported infections. Based on this study, a tool was developed for an “at distance” evaluation of data validity in surveillance programmes, without performing a time-consuming field validity study.

INTRODUCTION: The initiation of empiric antibiotic therapy (EA) in ICU patients is a source of debate. Epidemiological studies on this topic are scarce, mostly retrospective, focusing on an entire institution or an unique disease. However, prescription of EA varies largely with the type of hospital, patients (pts), underlying diseases and signs of severity. An observational prospective multicenter study was performed in order to assess the frequency of administration of EA, its timing, and type of pts. METHODS: 41 ICUs participated to this one-month survey. Every pt was recorded and those receiving EA initiated at any time during the ICU stay were analysed (gender, age, reason for admission, underlying disease using the Mc Cabe score, signs of severity, SAPSII and SOFA scores). The timing of EA (day or night, week or week-end), identification of the prescriber (resident, attending physician, decision of the team) and the suspected source of infection were noted. Results are expressed as proportions and mean±SD. RESULTS: Among the 1700 recorded pts, 39% did not receive any antibiotic during their ICU stay, 28% were already treated on admission, 3% had already microbiologic identification and susceptibility testing, and 509 (30%) pts received EA. These EA pts (58±19 y, male 64%) were admitted for medical reason (57%), scheduled (7%) or emergency (36%) surgery. Underlying diseases were frequently reported (including immunosuppression (12%), respiratory and cardiovascular comorbidities (12%), cirrhosis (6%)) and scored as ultimately (24% of the pts) or rapidly (11%) fatal. Their mean SAPSII score was 41±18 (mean SOFA score 6.5±4.6) and signs of severity were reported in 42% of the cases (including hypoxemic pneumonia 20%, septic shock 24%, and multiple organ failure 9%). EA was started on day work in 44% of the cases, on day in the week-end in 15% and at night in 41%. EA was decided by the team in 16% of the cases, or by a single prescriber in 74 % (including the attending physician in 62%). Lung (47%), abdomen (21%), and urinary tract (16%), were the most frequently suspected sources of infection. CONCLUSION: Our results demonstrate that initiation of EA is a frequent situation in ICUs, mostly due to suspicion of lung infection. Signs of severity and debilitated patients are frequent characteristics which could influence the therapeutic decision. EA appears to be often a solitary choice rather than a team attitude, even on day work. This apparent loneliness could be explained by written protocols of EA largely used in the most common circumstances such as peritonitis or pneumonia. Grant acknowledgement: Wyeth Lederle France

525

527

SELECTIVE DIGESTIVE DECONTAMINATION IN SEVERELY BURNED PATIENTS: A RANDOMIZED, DOUBLE BLIND TRIAL

ANTIBIOTIC DE-ESCALATION STRATEGY IN CRITICALLY ILL PATIENTS WITH BLOODSTREAM INFECTIONS

Cerdá E1, De la Cal M A1, Abella A1, Alía I1, García-Hierro P2, Aranguren A3 Intensive Care Unit, 2Department of microbiology, 3Department of pharmacy, Hospital Universitario de Getafe, Getafe, Spain

Zaragoza R1, Artero A2, Sancho S1, Camarena J3, González R3, Nogueira J3 1 Intensive care Unit, 2Internal Medicine, 3Microbiology, Hospital Universitario Dr. Peset, Valencia, Spain

INTRODUCTION: Pneumonia has been shown to be a risk factor for mortality in patients with severe burns (1). Selective digestive decontamination (SDD) is a manoeuvre that reduces pneumonia and mortality in other populations of critically ill patients. The objective of this study was to evaluate the efficacy of SDD in severely burn patients to reduce the incidence of pneumonia and mortality

INTRODUCTION: De-escalation of antibiotic therapy (DAT) could be considered as a strategy to balance the need to provide adequate initial antibiotic treatment of high risk patients with the avoidance of unnecessary antibiotic utilization, which promotes resistance. The aims of this study were to determine the prevalence of DAT in ICU bacteraemic-patients, to describe the main clinical, epidemiological and microbiological features of such patients and to know the impact of this strategy on global and related mortality to bacteraemia in critically ill patients.

1

METHODS: Prospective, randomized, double blind, placebo controlled trial conducted at a 6 bed burn ICU in a tertiary hospital between May 1997 and January 2000. All patients >=14 years with burns >= 20% of total body surface area and/or suspected inhalation injury and an interval between injury and burn admission <= 3 days were included. Exclusion criteria were: stay in the burn unit < 3 days, withdrawal of treatment within 3 days, immunosuppression, pregnancy and inhalation injury not requiring mechanical ventilation within the first 3 days. SDD was administered as soon as possible after patient’s admission and included four components 1) i.v administration of cefotaxime (1g q 8h) for 4 days, 2) topical application in the oropharynx using non absorbable polymyxin E, tobramycin and amphotericin B 0.5g of 2% past qid, 3) digestive administration of a 10 ml solution containing 100 mg of polymyxin E, 100 mg of tobramycin and 500 mg of amphotericin B qid, 4) surveillance samples of throat and rectum on admission and twice a week. The control patients received isotonic 0.9% saline i.v and placebo paste and digestive solution The diagnosis of pneumonia was made taking account radiologic criteria, clinical signs and quantitative culture of tracheal aspirate. It was done by two independent physicians not involved in the care of patients. Other infections were diagnosed according to CDC definitions for nosocomial infections. RESULTS: 117 patients were randomised and 107 were analysed (53 in the SDD group and 54 in the control group). The ICU mortality was 27.8 % in the control group and 9.4 % in the SDD group. Treatment with SDD was associated with a significant reduction in mortality both in the burn unit (risk ratio 0.25, 95 % CI 0.08 to 0.76) and in the hospital (risk ratio 0.28, 95 % CI 0.10 to 0.80), following adjustment for predicted mortality. The incidence of pneumonia was significantly higher in the control group: 30.8 versus 17.0 pneumonias per 1,000 ventilation days (p=0.03). CONCLUSION: Treatment with SDD reduces mortality and pneumonia in patients with severe burns and should be considered by those involved in the care of this type of critically ill patients. REFERENCE: 1-de la Cal MA et al. Pneumonia in patients with severe burns: a classification according to the concept of the carrier state. Chest.2001;119:1160-5. Grant acknowledgement: Funded partially by FIS 02-1883

METHODS: During a four years and a half period, from 1999 to 2003, 129 ICU-bacteraemic patients were prospectively evaluated. Initially a broad-spectrum antibiotic combination regimen was administrated until the agent of infection was identified, then patients were switched to spectrum-adjusted regimen following the patient´s physician criteria. Clinical and microbiological variables were recorded. Logistic regression analysis was performed to determine the factors associated with DAT and the impact of this strategy on global and related mortality to bacteraemia. RESULTS: DAT was applied in the 23.25% (n = 30). The mean age of these patients was 63.1±13.6 years and its mean APACHE II and SOFA score were 18.8±7.5 and 7.7±3.6 respectively. The incidence of inadequate empirical antibiotic treatment was 20.6 % in this group. The most frequent source of bacteraemia in DAT group was the respiratory focus (41.3%) and their principle aetiologies were: Acinetobacter baumannii (n = 8), and Streptococcus pneumoniae (n = 3). There were no significant differences in these variables with the rest of the episodes of bacteraemia. Only the presence of septic shock (30% vs. 59.5%; p=0.004) and multiorganic failure (33.3% vs. 62.6%; p=0.005) were significantly less frequent in DAT group although these findings were not confirmed in a multivariate analysis. Related mortality was 23.3% in DAT group. Only the global mortality rate was significantly lower in DAT group (40% vs. 62.6%; p=0.02). Logistic regression analysis showed that DAT was not related to neither global (OR 0.54; 95% CI 0.181.57) nor related mortality to bacteraemia (OR 1.34; 95% CI 0.38-4.7). CONCLUSION: DAT appears to be an appropriate, frequent and safe strategy to treat critically ill patients with bloodstream infections and is not associated with an increased global or related mortality to bacteraemia. Further studies are needed to assess its influence on development of resistances. REFERENCE: 1.- Álvarez-Lerma F, Álvarez B, Rello J et al. Antibiotic De-escalation Therapy for nosocomial Pneumonia in ICU(ADANN study. Intensive Care Medicine 2003;29(S1):S68

S138

17th Annual Congress – Berlin, Germany – 10–13 October 2004

530

Oral Presentations Cellular biology and genomics of acute brain injury – 528-533 528

EVALUATION OF CENTRAL REGULATION OF HAEMODYNAMIC IN SEVERE BRAIN INJURY Bubnova I D1, Astakhov A A1 1 Department of anaesthesiology, Reanimatology and Neuroreanimatology, Urals Medical Academy of postgraduate education, Chelyabinsk, Russian Federation

NEUROPROTECTION OF ERYTHROPOIETIN IN IN VITRO AND IN VIVO RAT MODELS OF MECHANICAL CEREBRAL TRAUMA 1,

L1,

A2,

E1,

R 1,

E2

Cozzi Gramigni De Gaudio Pellegrini-Giampietro D Adembri C Venturi 1 Dept. Critical Care Medicine, 2Dept.Pharmacology, University of Florence, Florence, Italy

INTRODUCTION: Recombinant human erythropoietin (rhEpo) is a potent neuroprotective agent in a number of ischemic/excitotoxic experimental models. In a small clinical trial, rhEpo was also effective in improving the outcome after acute ischemic stroke (1). Because data on the protective effects of rhEpo against trauma are few, we examined whether rhEpo could attenuate cell death in models of traumatic brain injury in vitro and in vivo. METHODS: In the in vitro experiments, organotypic rat hippocampal slices were subjected to traumatic injury by impacting the CA1 area with a specially devised apparatus (energy of impact 6 muJ). Damage was identified and evaluated 24 and 48 h later with the fluorescent dye propidium iodide (PI). In vivo, head injury in rats was induced by inserting a 2F Fogarty balloon catheter under the left parietal skull. The catheter was rapidly inflated with 100 µL of saline for 10 min and then removed. After 96 h, rats were neurologically examined, sacrificed and the brains were frozen, sectioned and stained with toluidine blue. Damage was quantified by measuring and integrating the unstained areas. RESULTS: In organotypic slices, trauma induced a significant and progressive increase in the mean PI fluorescence that was nearly doubled after 48 h (p<0.01 vs. control). Addition of rhEpo (1-100 U/ml) after the traumatic injury significantly reduced CA1 hippocampal damage at 24 h by approximately 30% (p<0.05 vs. control). After 48 h, the effect of rhEPO was enhanced, the damage being reduced by approximately 45% (p<0.01 vs. control). The same pattern of neuroprotection was observed even when rhEpo administration was delayed up to 30 min after trauma. Sham operated rats had no neurological impairment and only minimal damage of the left hemisfere. Trauma induced an impairment in neurological test scores and increased the volume of the injury to 51.4 ± 7.2 mm3 (mean ± SEM). Treatment with rhEpo (5000 U/Kg i.p./die for 4 days) improved the neurological outcome and reduced post-traumatic damage by approximately 35% (p<0.05) CONCLUSION: Our results suggest that rhEPO attenuates post-traumatic injury in vitro as well in vivo.

INTRODUCTION: Central regulation of haemodynamic (CRH) disturbances is one of the reason of high mortality in severe brain injuries (SBI). SBI are multifocal as a rule. There are two or more focuses of injury, haematomas combination and secondary brain damages. The evaluation of CRH topical level, CRH heaviness and adaptive reactions in haemodynamic system are most significantly under the given circumstances. METHODS: 125 patients with SBI were examined. All of them had 8 and less GCS score and were under artificial ventilation. Multimodal monitoring included registration of EEG amplitude and haemodynamic parameters (HP) such as heart rate (HR), blood pressure (BP), stroke volume (SV), ejector fraction (EF) and pulsatory amplitude of peripheral vessels (PAPV) by bioimpedans method. By means of spectral analyses we determined the fluctuation power (FP) of EEG amplitude and HP in diapason of 0-0.5 Hz and its four ranges- UVLF (0 – 0,025 Hz – metabolic influences), VLF (0, 025 – 0,075 Hz – humoral activity), LF (0,075 – 0,15 Hz – baroregulation), HF (0,15 - 0,5Hz -parasympatic activity). In addition we examined 42 patients with hemispheric stroke (30) and brain stem stroke (12) and 24 patients with global ischemia. RESULTS: It was revealed that patients with SBI might have 4 types of haemodynamic regulation. Type 1 was characterized by FP increasing in UVLF and VLF ranges of all HP and EEG amplitude. Type 2- extremely high FP in VLF range, especially of BP and SV. At the same time the significant FP decreasing was noticed in all ranges of PAPV. These types we also determined in patients with hemispheric stroke and in case of mass-effect. It was evidence, that patients with SBI might have an adaptive hypothalamic reactions, which lead to rather good outcome (type 1), or disadaptive hypothalamic reactions (“splash activity”) (type 2), which was associated with poor outcome, expressive catabolism and multiorgan failure. In most severe SBI as well as in cases of the brain stem stroke the substantial FP reduction of EEG amplitude and all HP were revealed. The baroregulation was absent. Only HF power was present as a result of artificial ventilation. It was the 3rd type of regulation. Permanent reducing of FP would result in poor outcome. The 4th type was characterized by moderate reducing of FP, especially in UVLF of such parameters as EEG amplitude, HR, EF and PAPV. It was the overall sign for SBI and global ischemic damage. CONCLUSION: Thus variability of EEG amplitude and HP are markers of CRH disturbances because they reflect the degree and topic level of brain regulatory system’s damage and might be useful for determination of SBI complications and outcome prediction.

REFERENCE: (1) Eherenreich H. et al: Molecular Med (2002); 8:495-505

529

531

CIRCULATING LEVELS OF CALCITONIN-GENE RELATED PEPTIDE IN PATIENTS WITH ACUTE BACTERIAL MENINGITIS

ROLE OF CODON 72 POLYMORPHIC VARIANTS OF P53 IN OUTCOME AFTER TRAUMATIC BRAIN INJURY

Moller K1, Strauss G I2, Tofteng F2, Qvist T1, Sahl C3, Edvinsson L4 1 Infectious Diseases, 2Hepatology, 3Intensive Care, University Hospital Rigshospitalet, 4Experimental Medicine, University Hospital Glostrup, Copenhagen O, Denmark

Martinez Lucas P1, Moreno Cuesta J1, Picazo Martinez M1, Cuartero del Pozo A1, Garcia Olmo D C1, Guardiola Salmeron M1, Garcia Olmo D2 1 Anesthesia Department, Hospital General, Albacete, 2Surgery Department, Universidad Autonoma, Madrid, Spain

INTRODUCTION: The endogenous vasodilator, calcitonin-gene related peptide (CGRP), has been implicated in the pathophysiology of bacterial meningitis (1). In the present study, we measured arterial plasma concentration and cerebral net flux of CGRP in patients with acute bacterial meningitis. Because meningitis patients are often mechanically hyperventilated, septic, and exposed to vasopressor support, we compared the results to those found during similar interventions in healthy controls. METHODS: The study was preceded by Institutional Review Board approval and informed consent from the subjects or their families. We included seven mechanically ventilated adult patients with severe bacterial meningitis and sepsis, five of whom were infused with dopamine or norepinephrine for blood pressure control, and sixteen healthy young volunteers (controls). Controls were subsequently studied during acute voluntary hyperventilation (N=8), 90 minutes after an intravenous bolus of endotoxin (2 ng/kg, N=8), and during an infusion of norepinephrine (N=8). Cerebral plasma flow (CPF) was measured by the Kety-Schmidt technique. The plasma concentration of calcitonin-gene related peptide (CGRP) was measured by radioimmunoassay in blood samples drawn from arterial and jugular bulb catheters. The net cerebral flux was calculated by multiplication of the arterial to jugular plasma concentration difference and CPF. Statistical analysis was performed using non-parametric testing and P<0.05. RESULTS: CPF did not differ between patients and controls. The arterial plasma concentration of CGRP was significantly higher in patients than in controls (P<0.01). There was a trend towards a lower cerebral net flux of CGRP in patients than in controls (P=0.08), but the net flux did not differ from zero in either group at baseline. In controls, CGRP plasma levels decreased significantly during hyperventilation (P=0.01) and after endotoxin injection (P=0.01), and increased significantly during norepinephrine infusion (P=0.01). Cerebral net flux was significantly larger than zero during norepinephrine infusion, i.e., an influx was observed. No flux was present during hyperventilation or after endotoxin injection. CONCLUSION: Patients with severe bacterial meningitis exhibit elevated circulating levels of CGRP. This is in line with experimental studies; however, infusion of sympathomimetics may be an alternative explanation. In the present study, no cerebral net flux of CGRP was observed in patients, possibly due to the limited sample size. REFERENCE: 1. Hoffmann et al. J Cereb Blood Flow Metab. 2002;22:988-96 Grant acknowledgement: The Danish Health Research Council (file # 52-00-0778)

INTRODUCTION: P53 is believed to be a major genetic control for the apoptotic mechanism in cells1.It has been described changes in P53 related proteins expression with DNA fragmentation in central nervous system (CNS) after experimental traumatic brain injury (TBI) and ischemic events2.The most prevalent P53 polymorphism is exon 4 codon 72 in which a guanine changes to cytosine. This gives to a substitution of a arginine(A) for proline(P) in the related protein(AA,AP,PP)3.The A variant is five times better at inducing apoptosis than are the P variants4. Our objective is to determine whether the alleles are associated with the outcome of TBI in humans. METHODS: Prospective study conducted from January 2001 to December 2003 in Albacete University Hospital Surgical ITU. The protocol was approved by the Research Board of Albacete Hospital. Informed consent was obtained from relatives.The TBI group included 92 patients(16 females,76 males).The mean(SD) age was 34(19) years.To be eligible for enrolment the patients had suffered a TBI within the previous 24 hours with a GCS<=8.The exclusion criteria was a previous deficit in the CNS. The main outcome measure was the Glasgow Outcome Scale (GOS) on the discharge moment(GOSD) and 6 months later(GOSL). A blood sample was taken and a template for Polymerase Chain Reaction(PCR) was prepared. DNA was measured by a spectrophotometer.A PCR amplification and a ethidium bromide stained agarose gel electrophoresis was used for the detection of the p53 exon 4 codon 72 polymorphysm.P53 frequencies were tested for their association with GOSD and GOSL using X2 test. A 2-tailed p<0.05 was considered significant. RESULTS: Mortality among patients with TBI was 38%,consistent with the predicted mortality from the APACHE II value 32%.The polymorphism frequencies were AA 59.7%,AP 32.6% and PP 7.6%.The AA frequency was greater among those who had a bad outcome on the discharge moment (Bad outcome: AA 65.6%,AP-PP variants 34.3%. Good outcome: AA 46.4%,AP-PP variants 53.5%, p=0.08; OR 2.20, 95% CI 0.89-5.44). The TBI polymorphism frequencies were not different between those who had a bad or a good outcome on GOSL (Bad outcome: AA 60.4%, AP-PP 39.5%. Good outcome: AA 59%,AP-PP 40.9%. p=0.8; OR 1, 95% CI 0.45-2.4). CONCLUSION: In this study the p53 exon 4 codon 72 AA variant frequency was greater among the patients who had a bad outcome on the discharge moment. The difference was in the limit of the significance and the strength of the relation is consistent with the molecular data about the protective effect over the apoptotic mechanism of the proline variants. REFERENCES: 1. Fisher DE.Apoptosis 2001;6(1-2):7-15. 2. Lu J et al. Neurosci Lett 2000;290:89-92. 3. Ara S et al. Nucleic Acids Res 1990;18:4961. 4. Dumont et al. Nat Genet 2003;33:357-365.

17th Annual Congress – Berlin, Germany – 10–13 October 2004

532 PROTECTIVE EFFECTS OF MYOBLASTS PRODUCING BRAIN DERIVED NEUROTROPHIC FACTOR ON ISCHEMIC BRAIN Hong X1, Zheng L2, Liu C3, Qu S4, Chen X1, Sun Q5 Pediatrics, Fuzhou Dongfang Hospital, 2Gene Engineering, Fujian Medical University, 3Pediatrics, Fujian Provincial Maternal and Child Health Hospital, Fuzhou, 4Neurobiology, Second Military Medical University, Shanghai, China, 5Internal Medicine, University of Michigan, Ann Arbor, United States 1

INTRODUCTION: Hypoxia ischemia brain injury (HIBI) in survivors of perinatal asphyxia is a severe brain condition frequently encountered in the clinic for which there is currently little effective therapy except symptomatic treatment. BDNF has been regarded as a putative option in the treatment of asphyxia. METHODS: Because of difficulties hindering the application of exogenous BDNF, e.g., short period of half-life in the brain and inability of penetration through the blood brain barrier, we developed a genetically engineered rat myoblast cell line (BDNF(+)/L-6TG) that was confirmed to express biologically active BDNF at high level in vitro. We investigated the neuroprotective effects of BDNF by transplantation of BDNF(+)/L-6TG into the cortex of neonatal rats subjected to HIBI. RESULTS: Two days after the transplantation, brain oedema of the left hemisphere, the prominent lesion side for HIBI of BDNF(+)/L-6TG treated group was markedly reduced compared to the vehicle cell treated group. Twenty-one days after the transplantation, the left brain atrophy was significantly alleviated in the BDNF(+)/L-6TG treated group compared to the vehicle cell treated group with an increase of 15.6% left/right weight ratio. The left/right ratio of the cortical, hippocumpal and striatal volumes, and neuronal counts of the hippocampal CA1 pyramidal layer were significantly increased whereas mortality rate significantly decreased in the BDNF(+)/L-6TG treated group than in the vehicle cell treated one. CONCLUSION: Intracerebral transplantation of genetically modified myoblasts producing BDNF exerted favourable effects on the recovery of HIBI thus to reduce the mortality rate. Grant acknowledgement: This study was supported in part by „9•5“ General Projects of Medical Science and Research Program of PLA (#98D022, type B, Part II).

S139

Oral Presentations Markers of preload and volume responsiveness – 534-539 534 ERRORS IN ESTIMATING INTRATHORACIC BLOOD VOLUME USING TRANSPULMONARY THERMODILUTION IN SHOCK Nirmalan M1, Willard T2, Edwards D J3, Little R A3, Dark P M4. 1Critical Care Unit, Manchester Royal Infirmary, 2Medical Physics, South Manchester University Hospitals, 3MRC Trauma Group, University of Manchester, Manchester, 4Critical Care Unit, Hope Hospital, Salford, United Kingdom INTRODUCTION: The transpulmonary thermodilution technique is used to estimate intrathoracic thermal volume (ITTV) and global end diastolic volume (GEDV). Intrathoracic blood volume (ITBV) however is derived from measured GEDV values [1,2]. ITBV is critical for estimating extravascular lung water (EVLW) and the validity of deriving ITBV from GEDV is not established in hypovolaemic subjects. METHODS: Anaesthetised pigs were allocated to control (C:N=9) and shock (S:N=10) groups. S was subject to hypovolaemic shock (>30% blood volume loss) for 60 minutes followed by volume resuscitation. Mean arterial pressure (MAP), cardiac output (CO) and mixed venous saturation were measured at baseline (Time 0), early shock (30 minutes), late shock (90 minutes) and post resuscitation (120 minutes). Thermal-dye dilution (COLD, PULSION) technique was used to measure ITBV (ITBVm), GEDV and EVLW.The linear regression between GEDV and ITBVm in C was used to derive an indirect estimate of ITBV (ITBVd) from measured GEDV in both groups. The derived value (ITBVd) was compared to true ITBV (ITBVm) using within subject correlation (r) and the %prediction error at each of the time points was compared using repeated measures ANOVA. RESULTS: Haemorrhage was associated with a significant reduction in CO, MAP and ITBV (P<0.05). There was a significant correlation between GEDV and ITBVm in group C [ITBVm=1.25*GEDV+88.7;r 2 =0.93). The percentage errors in deriving ITBVd [mean(SD)] using this equation at the four time points are shown in Table 1. The errors for S at 30 & 90 minutes were then corrected by using the equation obtained by plotting GEDV/ITBVm values for S at 30 and 90 minutes only (20 sets of values).The corrected errors were similar to the errors for C at similar time points and the baseline/post resuscitation errors for S. [(ITBVm-ITBVd)/ITBVm]*100 Baseline 0min Shock 30 min Shock 90min Post resus 120min C(n=9) 3.4(1.8) 3.6(2.2) 4.8(3.6) 4.3(3.3) S(n=10) 3.2(1.8) 13.2(6.2)* 12.8(9.1)* 5.3(3.2) * Significant increase in error CONCLUSION: The linear and predictable relationship between GEDV and ITBV is preserved in the presence of gross hypovolaemia (r=0.95). The quantitative relationship between the two variables however is different due to compensatory pulmonary vasoconstriction. Hence applying the regression equation obtained in normovolaemic subjects (as in current practice) may lead to significant errors in EVLW measurements. These findings have considerable implications in the clinical applications of the thermodilution technique (PiCCO; PULSION). REFERENCES: 1. Neumann P. Intensive Care Med 1999;25:216-219. 2. Sakka SG et al. Intensive Care Med 2000;26:180-187 Grant acknowledgement: 1. MRC & Maelor(UK)

533

535

ENOXIPARIN OR NADROPARIN AND THE FUNCTION OF THE BLOODCEREBROSPINAL FLUID BARRIER IN FEMALES

DO LEG RAISING OR DOPPLER AORTIC FLOW VARIATION PREDICT FLUID RESPONSE IN VENTILATED PATIENTS?

Pakulski C1, Mikulski K1, Badowicz B1, Kwiecieñ K2 1 Department of Emergency Medicine, 2TraumaCentre, Pomeranian Medical University, Szczecin, Poland

Monnet X1, Rienzo M1, Osman D1, Anguel N1, Richard C1, Pinsky M R2, Teboul J L1 Service de réanimation médicale, CHU de Bicêtre - Assistance Publique Hôpitaux de Paris Université Paris XI, Le Kremlin-Bicêtre, France, 2Department of critical care medicine, University of Pittsburgh, Pittsburgh, United States

INTRODUCTION: The aim of the study was to find, if any, differences between the function of the blood-cerebrospinal fluid barrier (BCSFB) with regard to optional administration of anticoagulant therapy, and whether BCSFB function depended on the type of the anticoagulant drug used in prophylaxis. METHODS: The study enrolled 108 females aged 50 years or more, at least one year after menopause, without hormonal replacement therapy in previous history. The study was approved by the Pomeranian Medical University Ethic Committee (BN-001/35/02). The eligible females were divided into 3 groups, each group containing 36 women. The inclusion criteria included: for the I group- administration of enoxiparin; for the II group- administration of nadroparin, and for the III group - lack of anticoagulant therapy. Enoxiparin in the group I and nadroparin in the II group were administered subcutaneously 12 hours before the elective surgery at equivalent doses, 20 mg of enoxiparin and 0,3 ml nadroparin, respectively. The study protocol consisted of single sampling of 10 ml of venous blood and 2 ml of cerebrospinal fluid (CSF). In the collected samples the concentrations of albumin and immunoglobulin G (IgG) were measured in the blood serum while concentrations of microalbumin and IgG were measured in the CSF. The concentrations of blood serum albumin were evaluated using the colorimetric method with application of bromcresol green, while the concentrations of CSF microalbumin and CSF and blood serum IgG were evaluated using the kinetic immunonephelometry. The following ratios were calculated: the albumin permeability index (QALB), the IgG permeability index (QIgG). Finally, basing on the calculated permeability indexes the functional condition of BCSFB was evaluated. RESULTS: The studied groups did not differ statistically significantly with regard to age and body mass index (BMI). The groups differed with regard to QALB values (group I- 4,62 ±1,3; group II- 3,55±1,4; group III- 3,66 ±0,7). Similar tendency was noted for QIgG values (group I2,9±0,9; group II- 2,21 ±0,7; group III- 2,04±0,5). The statistically significant difference was found between the group of females receiving enoxiparin and the control group (groups I and III), and between the groups receiving different heparin preparations (groups I and II) with regard to QALB and QIgG ratios (p<0,001). No such differences were found between the group of females receiving nadroparin and the control group (groups II and III). CONCLUSION: Anticoagulation therapy with enoxiparin administered during the preoperative period increased permeability of BCFSB for microalbumin and immunoglobulin G. Nadroparin does not alter the function of BCSFB. In females aged 50 years or more who are at risk of brain dysfunction and require anticoagulation preventive therapy, nadroparin is the low molecular weight heparin of choice.

1

INTRODUCTION: The aim of this study was to test whether response to volume expansion (VE) could be predicted by aortic blood flow (ABF), monitored with transesophageal Doppler in patients adapted to mechanical ventilation and with sinus rhythm. METHODS: In 36 patients mechanically ventilated with sinus rhythm and fully adapted to the ventilator who were considered to receive VE, a transesophageal Doppler probe was introduced (Hemosonic 100, Arrow). This device monitored both aortic diameter and velocity, allowing calculation of aortic blood flow. The 10-second averaged mean ABF was measured at baseline, after passive leg raising (PLR) and after VE (500 mL NaCl 0.9N infusion over 10 minutes). The variation of ABF (∆ABF) over a respiratory cycle between its minimal (ABFmin) and maximal (ABFmax) values (∆ABF = (ABFmax-ABFmin)/[(ABFmax+ABFmin)/2]) was calculated. RESULTS: The change in mean ABF with PLR correlated with that induced by VE (r=0.84, p<0.001). The mean ABF increased by ≥15 % after VE in 18 patients (defined as responders). A PLR-induced ABF increase ≥8% predicted the response to VE with a sensitivity of 94 % and a specificity of 89%. Before VE, ∆ABF was higher in responders than in non-responders (27±14% vs. 12±7%, p<0.05). The ∆ABF significantly decreased with PLR (22±15%) and after VE (18±11%) in responders, while DABF did not change in non-responders. A ∆ABF ≥18% predicted fluid responsiveness with a sensitivity of 89% and a specificity of 94%. Receiver operating characteristic curves analysis showed that PLR-induced changes in mean ABF was as predictive of fluid responsiveness as ∆ABF and more predictive than respiratory variation of aortic blood velocity. CONCLUSION: PLR-induced changes in ABF measured by transesophageal Doppler, as well as ∆ABF, can be reliably used to predict fluid responsiveness in patients with sinus rhythm and fully adapted to mechanical ventilation.

S140

17th Annual Congress – Berlin, Germany – 10–13 October 2004

536

538

RESPONSE TO LEG RAISING PREDICTS FLUID RESPONSIVENESS IN CASE OF SPONTANEOUS BREATHING OR ARRHYTHMIA

EJECTION FRACTION CORRECTED END-DIASTOLIC VOLUME AS A BETTER PRELOAD INDICATOR

Monnet X1, Rienzo M1, Osman D1, Anguel N1, Richard C1, Pinsky M R2, Teboul J L1 Service de réanimation médicale, CHU de Bicêtre, Le Kremlin-Bicêtre, France, 2Department of critical care, University of Pittsburgh, Pittsburgh, United States

De Potter TJR1, Deeren D1, Nijs J1, Libeer C1, Dits H1, ÏMalbrain MLN1. 1 Department of Intensive Care, Ziekenhuis Netwerk Antwerpen, Antwerp, Belgium

1

INTRODUCTION: In ventilated patients in whom the analysis of respiratory variations of systolic ejection cannot be correctly interpreted, prediction of haemodynamic response to volemic expansion (VE) is difficult. We tested the hypothesis that in patients with spontaneous respiratory activity and/or arrhythmia, the haemodynamic response to passive legs raising (PLR), noninvasively evaluated by transoesophageal Doppler, could predict fluid responsiveness. METHODS: In 24 mechanically ventilated with spontaneous respiratory activity and/or arrhythmia who were considered for VE, a transesophageal Doppler probe was introduced (Hemosonic 100, Arrow) to record the aortic blood flow (ABF) signal. Mean ABF value was measured at baseline, after PLR and after VE (500ml NaCl 0,9N infusion over 10 minutes). RESULTS: The ABF increased after volume expansion by ≥15 % in 12 patients (responders) and by <15 % in 12 patients (non-responders). In responders, ABF increased by 21±12% (p<0.05) after PLR and 35±16% after VE, while it was unaltered in non-responders. Considering all patients, the increase in ABF induced by PLR correlated with that induced by VE (r=0.78, p<0.05). An increase >8% induced by PLR predicted the response to VE with a sensitivity and a specificity of 92%. CONCLUSION: Thus, in ventilated patients with spontaneous breathing and/or arrhythmia, the non-invasive analysis of changes in ABF induced by PLR using transesophageal Doppler is valuable to predict fluid responsiveness.

INTRODUCTION: Volumetric monitoring with right ventricular end-diastolic volume, indexed (RVEDVi) and global end-diastolic volume, indexed (GEDVi) is increasingly being suggested as a better preload indicator than traditional filling pressures (CVP, PAOP) (1). Static monitoring of these volumetric parameters has not consistently been shown to correlate with cardiac index (CI). This study aims 1° to evaluate the influence of right ventricular ejection fraction (REF) and global ejection fraction (GEF) on RVEDVi and GEDVi respectively, 2° to study the effect of RVEDVi and GEDVi changes (Delta) adjusted for their corresponding REF and GEF on DeltaCI and 3° to identify optimal resuscitation target volumes. METHODS: Complete haemodynamic profiles with VoLEF and PiCCO catheter (Pulsion, Germany) were obtained in 9 mechanically ventilated medical ICU patients. Age 63.4±12.8 years, SAPS-II 47.2±19.9, APACHE II 20.9±6.2, and SOFA 5.6±4.4. In total 181 paired measurements of RVEDVi, REF, GEDVi, GEF, CVP, PAOP and CI were performed. RESULTS: Figure 1-B shows the correlation of DeltaRVEDVi corrected for REF (DeltaRVEDViC) with CI (R? = 0.67, p<0.0001). A similar correlation was found between DeltaGEDVi corrected for GEF (DeltaGEDVi-C) with DeltaCI (R? = 0.63, p<0.0001). Adjustment of volumes was achieved by correction for EF deviation from normal EF values in an exponential fashion. Suggested volumetric resuscitation targets corrected for EF are given in Table 1. We found an absolute lack of correlation between (Delta) filling pressures (CVP,PAOP) or (Delta) static volumes (RVEDVi, GEDi; figure 1-A) and (Delta) CI. Ejection fraction 10% 20% Target RVEDVi 265-350 200-250 Target GEDVi 950-1150 750-900 EF-corrected target resuscitation volumes

30% 150-170 600-675

40% 110-115 475-525

50% 80 400

CONCLUSION: Filling pressures and static volumes are unreliable preload indices. The EF corrected target volumes better reflect preload status. We suggest to use EF-adjusted RVEDVi and GEDVi as resuscitation endpoints. Volumetric parameters should not be interpreted independently of their corresponding ejection fraction. REFERENCE(S): (1) Michard ea. Global end-diastolic volume as an indicator of cardiac preload in patients with septic shock. Chest 2003; 124: 1900-8.

537

539

ITBV-GUIDED VOLUME THERAPY MAINTAINS LIVER PERFUSION AND OXYGENATION DESPITE PEEP VENTILATION IN ALI

PLASMA VOLUME, HEMODYNAMICS AND FLUID SHIFTS DURING ACCIDENTAL HYPOTHERMIA. AN EXPERIMENTAL APPROACH

Iber T1, Vagts D A1, Mutz C W1, Roesner J P1, Radmann S1, Noeldge-Schomburg G F E1 Anesthesiolgy and Intensive Care, University of Rostock, Rostock, Germany

Hammersborg S M1, Farstad M1, Haugen O1, Onarheim H1, Husby P1 Anesthesia and Intensive Care, Haukeland University Hospital, Bergen, Norway

1

1

INTRODUCTION: The detrimental effects of PEEP on liver perfusion and oxygenation are well described. Furthermore it has been demonstrated in an animal model of normal lung compliance that itbv-guided volume loading diminishes these effects. But it remains controversial, whether this applies for situations with reduced lung compliance. Therefore we investigated the effects of PEEP in an animal model of oleic acid induced lung injury with itbv-guided volume loading.

INTRODUCTION: The mechanisms behind the transcapillary fluid shifts during accidental hypothermia are not fully understood. The purpose of this study was twofold. 1:To create a model of accidental hypothermia. 2: To describe the effects of surface cooling on microvascular fluid exchange, vascular capacity and haemodynamic variables.

METHODS: 17 anesthetized and ventilated pigs were studied. Ultrasonic flow probes were placed around the hepatic artery and portal vein, catheters were inserted into the femoral artery, portal and hepatic veins. Animals were randomly assigned to two groups: group1=ali (n=9) animals received basic fluid with 15 ml/kg/h cristalloids and group 2=itbv (n=8) animals received basic fluid and colloid substitution to maintain itbv at baseline level. After preparation, animals were allowed to stabilize for 2h before ALI was induced with oleic acid (0,1 ml/kg). Then baseline values at PEEP 8 were taken. Further measurements were made at PEEP 12, 16 and 20 cm H2O. ITBV was measured by the COLD®-System. RESULTS: All data are median, # p<0.05 vs. PEEP8, § p<0.05 vs. ALI. ALI paO2/fiO2 mmHg CO ml/kg*min ITBV ml/kg THBF ml/kg/min DO2_liver ml/min tpO2_liver mmHg Hb mg/dl

PEEP8 143 94 28 28 55 62 9,1

PEEP12 191 78 24 24 51 59 9,9

PEEP16 225 81 24 20# 51 53 10,4

PEEP20 189 69# 22# 17# 47# 48# 11,2

ITBV paO2/fiO2 mmHg CO ml/kg*min ITBV ml/kg THBF ml/kg*min DO2_liver ml/min tpO2_liver mmHg Hb mg/dl

PEEP8 151 96 28 26 56 58 9,9

PEEP12 195 108 29§ 28 58 65 8,7

PEEP16 233 114#§ 29§ 28§ 62 60 8,2§

PEEP20 218 111#§ 29§ 27§ 55 63§ 7,7§

CONCLUSION: ITBV-guided volume loading maintains hepatic blood flow up to a PEEP level of 20 cm H2O. Furthermore oxygen delivery to the liver and liver tissue pO2 remained unchanged. Nevertheless these effects would be more pronounced when haemodilution is avoided. In conclusion, ITBV-guided volume loading is a helpful concept for the safe application of PEEP in acute lung injury.

METHODS: 7 piglets were anesthetised. After 60 min of normothermia, they were cooled by immersion in ice sludge for 60 min to 29 degree C. Thereafter they were moved to environmental temperature (21.0 C) The animals remained at stable hypothermia ((27.8 ± 1,6) (mean ± SEM)degree C) for 120 min. Haemodynamic parameters were measured or calculated. Plasma volume (PV)(CO-method), organ perfusion (dyed microspheres), fluid I/O, colloid osmotic pressure in plasma/interstitial fluid, Hct, s-albumin/protein and total tissue water content were measured. Albumin/protein masses and fluid extravasation rates (FER)were calculated. RESULTS: 60 minutes after start of cooling, as the animals were transferred to room temperature, an increase was observed in Hct (0.31 ± 0.00 to 0.35 ± (0.01), (Mean ± SEM) (p<0.01) concomitantly with a decrease in plasma volume (1093.6 ± 58.7 to 918.9 ± 70.0 ml)(p<0.05); albumin mass (36.5 ± 2.3 to 28.7 ± 2.6 g)(p<0.05) and protein mass (51.8 ± 4.1 to 42.1 ± 4.1 g) (p<0.05). FER increased in the same period (P<0.05). See figure.

CONCLUSION: The corresponding losses of fluid and proteins from circulation to the interstitial space indicates a transcapillary shift of whole plasma. Inflammatory fluid leakage usually results in protein rich exudate. Hence a general inflammatory response is suspected to be a possible mechanism behind the cold induced fluid extravasation during surface cooling. Further studies are needed. The model seems suitable to mimic accidental hypothermia. Grant acknowledgement: This study was financially supported by the Frank Mohn Foundation, The Lærdal Foundation for acute Medicine and the University of Bergen.

17th Annual Congress – Berlin, Germany – 10–13 October 2004

Oral Presentations Paediatrics II – 540-545 540 ACCURACY OF PREDICTIVE EQUATIONS FOR RESTING ENERGY EXPENDITURE CALCULATION IN CRITICALLY ILL CHILDREN Cotting J1, Depeyre J2, Ranson M3, Maeder A2, Biolley R2, Roulet M3, Stucki P1 1 PICU, CHUV, Lausanne, 2School, Of Dietetics, GENEVE, 3Clinical Nutrition Unit, CHUV, LAUSANNE, Switzerland INTRODUCTION: According to the ASPEN’s recommendation (2002), Resting Energy Expenditure (REE) in critically ill children must be measured by indirect calorimetry. However when this procedure is not available, predictive equations should be used as an alternative way to adapt the nutritional support. Objectives: to determine the more accurate predictive equations of REE in critically ill children. METHODS: REE was measured each day during artificial ventilation period by indirect calorimetry and compared with the most used predictive equations of REE: Harris-Benedict, FAO/OMS, Schofield [weight (W) and Weight + Height (WH)], Talbot, Liu, Altmann, Nelson, Kleiber and 60% of RDA. For each child predicted energy expenditure was estimated for children of the same age at P50 and P3 for body weight and for height using Eurogrowth charts. Thus, the Z-score was calculated for each child and for each equation by using the mean value of his REE measurements. Population’s Z-scores were compared to a theoretical value of zero using a twotailed t-test. Significant p values will represent an inefficient predictive equation for our population. RESULTS: 18 children were included in this study and among them, there were 3 girls, they were measured but were excluded for this evaluation. 116 measures of REE (7 ± 4, range 4 to 16 per patient) were performed in 15 boys. Population characteristics (mean ± SD) were: age: 22 ± 20.9 months, body-weight 9.8 ± 5.6 kg, height 77 ± 22 cm, PRISM score 11.5 ± 5.4, Measured REE: 541 ± 270 kcal/ day. The table shows five of ten evaluated equations with the lowest mean Zscores. Harris Benedict FAO/OMS Schofield W Schofield WH Talbot Predicte EE cal/d 601±20 565±32 576±32 573±30 639±46 Z-score 5.1±9.4 0.8±2.8 1.2±2.8 0.9±2.8 2.4±1.7 95% CI min-max -0.7_10 -0.7_2.4 -0.3_2.8 -0.6_2.5 1.5_ 3.3 t-test 2.11 1.14 1.68 1.30 5.66 p value 0.053 0.27 0.11 0.21 <0.0001 Significance S NS NS NS S

S141

542 INFECTION AND ABNORMAL FLORA INCREASES MORTALITY IN PICU Lubjuhn R1, Taylor N2, Thorburn K1, Van Saene H K F2 Paediatric Intensive Care, Alder Hey Children’s Hospital, 2Medical Microbiology, University of Livrepool, Liverpool, United Kingdom

1

INTRODUCTION: Normal flora is defined as the absence of aerobic Gram-negative bacilli [AGNB] in the oropharynx and the presence of E.coli in gut only. Carriage of AGNB in throat and/or gut is considered to be abnormal. Objectives [i] association between infection on admission and abnormal flora; [ii] association between abnormal flora and infection sites; [iii] AGNB causing abnormal flora; [iv] impact of abnormal flora on mortality. METHODS: Critically ill children requiring 4 or more days on PICU were prospectively enrolled in a 4 year study. Surveillance cultures of throat and rectum were obtained on admission and twice weekly afterwards. Diagnostic samples including blood, tracheal aspirate, urine and pus were taken on clinical indication only. All samples were cultured using standard microbiological techniques. RESULTS: Data are presented on a total of 1,241 children accounting for 1,443 admissions to the unit. Median PIM was 0.063 [IQR 0.025 – 0.131] and the actual mortality was 9.6%. There were a total of 305 children admitted 320 times with infection. 67% carried abnormal flora on admission. Amongst the 1,124 non infectious admissions, abnormal flora was present in 55.5%. 80.2% of children with RSV bronchiolitis were admitted with abnormal flora. 68.4% of septicaemic children imported abnormal flora. Klebsiella, Enterobacter, Pseudomonas and Citrobacter were predominant amongst AGNB. The mortality in the abnormal group was three times that of the normal flora group [9.2% vs 3.3%] CONCLUSION: Abnormal carriage is common when admitted to PICU with infection. In this group of children the mortality was higher in those with abnormal flora compared to those with normal flora. Grant acknowledgement: Chris Stoutenbeek Foundation

CONCLUSION: Only 3 of the 10 predictive equations (FAO/OMS, Schofield W and Schofield WH ) demonstrate an accurate calculation of Resting Energy Expenditure in critically ill children under sedation and artificial ventilation. Further studies including more patients should be made to confirm these results for each age group.

541

543

THE VALUE OF TRACHEAL ASPIRATE MICROBIOLOGIC INVESTIGATION IN DIAGNOSING VAP

OPERATING ROOM EXTUBATION IN PAEDIATRIC LIVER TRANSPLANTATION

Chuyko M M1, Dobryanskyy D O1 1 Faculty and Hospital Paediatrics, L’viv National Medical University, L’viv, Ukraine INTRODUCTION: Ventilator-associated pneumonia (VAP) is a common and severe complication in patients of intensive care units. The lack of specimens other than secretions suctioned through an endotracheal tube makes the diagnosis of VAP in newborns much more challenging in comparison with children and adults. METHODS: We evaluated the value of complex microbiologic investigation of tracheal aspirate (microscopic evaluation, qualitative and quantitative culture) in diagnosing of VAP in preterm newborns. The data of 56 infants with confirmed diagnosis of VAP (VAP group) were compared with the investigation results of 54 neonates without pneumonia (control group). The mean birth weight and gestational age were 1716,97±487,71 g and 1915,09±733,78 g (p>0,05) and 30,40±3,4 wks and 32,97±3,93 wks (p<0,05), respectively in the infants from VAP and control groups. All newborns included into the study had a primary non-infectious lung disease and were treated with mechanical ventilation for at least 24 hours before inclusion. The diagnosis of VAP was based on the CDC criteria for the infant < 1 year old. RESULTS: The tracheal aspirate (TA) was obtained on the 1st and on the 4th days of ventilation using sterile catheter and special technique. On the 4th day of ventilation 48 (86%) infants with VAP and 10 (19 %) newborns from the control group demonstrated increased production of tracheal secretion changed in colour and consistency (p<0,05). There was a significant correlation (r = 0,37; p<0,05) between these findings and an increase of polymorphonuclear leukocyte count in the TA smear (>10 cells in the field). On the first day of ventilation the TA cultures were positive in 49 (88%) neonates with VAP and 13 (24%) infants form the control group (p<0,05). At the same time a colony count in all cases was not higher than 10(2) colony-forming units (CFU) per millilitre indicating rather colonization than infection. By the 4th day of ventilation a colony count from the cultures of 37 (65%) newborns with VAP exceeded 10(3) CFU/ml in comparison with a count of <10(3) CFU/ml in all positive TA cultures of the control infants. The bacteria most commonly associated with the quantitative cultures of >10(3) CFU/ml were Pseudomonas aeruginosae, Enterobacter species, Klebsiellae pneumoniae and Acinetobacter lwofii. It was a statistically significant correlation (r = 0,54; p<0,05) between a colony count of >10(3) CFU/ml from the TA culture obtained on the 4th day of ventilation and the development of VAP. CONCLUSION: Our results suggest that the complex approach using physical assessment, microscopic evaluation and determination of quantitative cultures of TA is of value in diagnosing VAP in preterm newborns treated with mechanical ventilation shortly after birth.

Schellino M1, Cerutti E1, Pironti M1, Fop F1, Ballaris M1, Brunati A1, Gennari F1, Salizzoni M1 1 Centro Trapianto di Fegato, Az. Ospedaliera S. Giovenni Battista, Torino, Italy INTRODUCTION: To date, factors related to operating room (OR) extubation in paediatric liver transplantation (LT) have not been explored. Aims of the study were 1) to evaluate feasibility and safety of OR extubation in paediatric LT recipients; 2) to analyse pre-transplant and intra-operative factors related to OR extubation; 3) to evaluate the influence of this practice on postoperative complications, intensive care unit (ICU) and hospital length of stay. METHODS: Data from paediatric LT in our centre from October 1999 to February 2004 were analysed. Considered variables were: a) Pre-transplant: weight, age, bilirubin, INR, haemoglobin, platelets, creatinine, Child, UNOS status; b) Graft: type (whole liver, split, reduced, living donor), donor’s age and ICU stay, cold and warm ischemia time; c) Intra-operative: anaesthesia and surgical time, packed red blood cell (PRBC) and plasma transfusions, PaO2/FiO2, pH, lactate, INR, sodium, haemoglobin, temperature, heart rate, central venous pressure, and need of vasopressors at the end of surgery. Anaesthesia was routine. Piggy-back technique was used in all procedures. Patients were transfused to maintain haemoglobin level above 8 g/dl. Statistical analysis was performed using Mann-Whitney test for continuous variables and Chi-Square and Fisher exact test for discrete variables. RESULTS: 23 patients were transplanted. Two patients were excluded because of pre-operative mechanical ventilation due to encephalopathy and haemodynamic instability. Of the 21 recipients available for analysis, 11 (52,4%) had OR extubation (group A) (mean time for extubation 1,1 hour); 10 (47,6%) were extubated later (group B) (mean ventilation time 44.9 hours). In the latter group, 6 children were extubated during the first postoperative day (mean ventilation time 18 hours). No patient was reintubated for respiratory failure and no post-operative death occurred. Significant differences between groups were: (see table) Weight (kg) Age (years) UNOS 2A PRBC (ml/kg) Inotropic support ICU stay (days) Infectious complications Medical complications

Group A (n=11) 30 (15-55) 7,9 (2,9-14,2) 1/11 (9%) 17,8 (0-70) 0/11 (0%) 0 (0-4) 2/11 (18%) 4/11 (36%)

Group B (n=10) 8,45 (5,4-15) 0,9 (0,5-7,3) 7/10 (70%) 52,4 (0-199) 3/10 (30%) 3 (2-9) 7/10 (70%) 10/10 (100%)

p <0,01 <0,01 0,008 0,023 0,05 0,0015 0,03 0,04

CONCLUSION: Immediate extubation can be performed safely in about 50% of paediatric LT recipients. In addition to factors already recognized in adults (bad pre-transplant conditions, intraoperative blood consumption and haemodynamic instability at the end of surgery), small age and low weight are specific obstacles to the procedure in paediatric patients. Immediate extubation decreases ICU stay and is related to fewer infectious and medical complications.

S142

17th Annual Congress – Berlin, Germany – 10–13 October 2004

544 QUALITY ASSURANCE OF PAIN TREATMENT FOR SURGIAL INFANTS IN AN INTENSIVE CARE UNIT Poulsen M1 1 Dept. Anaesthesiology and Intensive Care V-BRITA, University Hospital of Southern Denmark Odense, Odense, Denmark

Oral Presentations Assessment and prognosis in neuro-intensive care – 546-550 546 NEUROLOGICAL SCORINGS SYSTEMS IN BRAIN INJURY: GCS VS. RLS85

INTRODUCTION: In the intensive care unit a pain assessment scheme was introduced to 1) improve the quality of the pain treatment given to recently operated infants. 2) to work out a schedule to improve the evaluation whether the infants are in pain or not. 3) to increase interest and create a debate about children and pain on the intensive care unit. 4) to increase knowledge by the staff who works with children in pain. METHODS: The pain assessment scheme was adapted from the validated Neonatal Infant Pain Scale (1) and transformed into Danish and called “SIS-skemaet”. There was a control- and a project group. To avoid influence of intervention the data of the control group were collected retrospectively, while the data in the project group were collected prospectively. The infants in the control group were not pain assessed by “SIS-skemaet”, while the infants in the project group was pain evaluated by “SIS-skemaet”. The to groups were compared with respect to number of infants receiving analgesics and amount of analgesics used in the first postoperative 24 hours. Staff evaluation was done by a question scheme after the project period. RESULTS: 46 infants were included in the control group and 55 in the project group. The groups were comparable by demographic data (age, weight, sex, age of gestation etc.). In the control group (N=46) 10 infants did not receive analgesics after surgery, while in the project group (N=55) only 2 infants did not receive analgesics after surgery. So significantly more infants received pain relieving treatment in the project group and there were used significantly more paracetamol in this group. There was a tendency to a higher use of morphine, but this difference was not significant. In the project group there was a clear relation between fall in pain score and administration of analgesics. The nurses evaluation showed, that their knowledge to pain treatment in infants have been increased during the project period. CONCLUSION: There has been a change and possible improvement of the pain treatment in the project period. Significantly more infants have received a pain treatment. There has been given significantly more paracetamol in the project group indicating a lower “pain threshold of detection” when using a pain assessment scheme. The nurses evaluation shows that they would like to use a pain assessment scheme to recently operated infants in intensive care unit. REFERENCE(S): (1) Lawrence J. et al. The development of a tool to Assess Neonatal Pain, Neonatal Network 1993; 12 (6): 59-66.

Starmark J1 Unit of Neuropsychology and Neuropsychiatry, Dept of Neurology, Institute of Clinical Neuroscience, Göteborg, Sweden

1

INTRODUCTION: Scoring of acute impaired brain function is considered important. The Reaction Level Scale (RLS85) (1,2) was developed as an alternative to Glasgow Coma Scale and was recommended for use in Sweden 1989. The Swedish clinical experience was that unacceptable problems could occur with GCS sum score and scoring of intubated patients as well as patients with swollen eyelids (after frontal brain trauma). Recently, the value of the GCS as predictor of outcome after severe traumatic brain injuries has been questioned. METHODS: Literature review and published studies. Four prospective cohort studies of aneurysmal SAH collected 1989-2001 (n>300). RESULTS: The clinical experience was documented in a review of the literature of the international use of the GCS (3). RLS85 was found to have superior qualities compared to GCS(4,5,6,7,8). Preliminary results from four prospective cohorts of patients after aneurysmal SAH and the changing weight of RLS85 as predictor of prognosis will be presented. In these studies, it is possible to observe the effect of the introduction of the concept of neurointensive care as well as of the coiling technique. CONCLUSION: RLS85 compared to GCS shows no systematic bias for assessment and has better reliability compared to GCS Sum Score or the EMV profile(7).The implications for predictive studies will be discussed (7,9). REFERENCES: 1) Starmark J-E. (1988) Analysing „Coma Scales“. Thesis. Göteborg. 148 pp. 2) Starmark J-E., Stålhammar D., Holmgren E. (1988) Acta Neurochir (Wien) 91: 12 - 20. 3) Stålhammar D., Starmark J-E., Holmgren E. et al(1988) Acta Neurochir (Wien) 90: 73 - 80. 4) Starmark J-E., Heath A. (1988) Human Toxicol 7: 551-555. 5) Starmark J-E., Holmgren E., Stålhammar D. (1988) J Neurosurg 69: 692 - 698.6) Starmark J-E., Stålhammar D., Holmgren E. et al (1988) J Neurosurg 69: 699 - 706. 7)Starmark , J-E: The Reaction Level Scale (RLS85): An Update. In ( Editor. Bo Risberg ): Trauma Care. Pharmacia & Upjohn, Göteborg 1996, ISBN: 9185794-32-5, pp 27-44. 8) Rödholm M, Starmark J-E, Ekholm S, von Essen C. Acta Neurol Scand. 2002; 105:1-11. 9) Svensson, E, Starmark, J-E. J Rehabil Med 2002; 34:251-259. Grant acknowledgement: Sahlgrenska Academy at Göteborg University

545

547

A RETROSPECTIVE STUDY ON PEDIATRIC DROWNING AND NEAR-DROWNING A GROUP OF 107 PATIENTS

INDUCTION OF HYPOTHERMIA USING LARGE VOLUMES OF ICE-COLD INTRAVENOUS FLUID: A FEASIBILITY STUDY

Blazek D1, Dlask K1, Cvachovec K1, Pavlicek P1, Hechtova D1 Anaesthesiology and Critical Care Medicine, University Hospital Motol, Prague, Czech Republic

Rijnsburger E R1, Girbes A R1, Spijkstra J J1, Peerdeman S M2, Polderman K H1 1 Department of Intensive Care, 2Dept of Neurosurgery, VU university medical center, Amsterdam, Netherlands

1

INTRODUCTION: Drowning and near-drowning is an event of submersion under a liquid surface. The death in the case of drowning is caused by asphyxia either immediately or within 24 hours after the submersion. In developed countries drowning still represents one of the leading cause of traumatic injury death in children and adolescents. METHODS: The authors studied retrospectively a group of 107 patients after drowning and neardrowning hospitalized at PICU, UH Motol, Prague in a period of 1991 - 2002. The time of CPR, depth of undercooling, presence of MOF and ARDS, ICP monitoring and other parameters were followed. Patients, who survived, were observed for the ability of school attendance, self control and neurological findings. RESULTS: From the initial group of 107 patients 84 have survived. CPR was applied in 71 patients, in 52 of them the time of it was exactly known - 38´ in average (max. 180´). 89 patients were undercooled, mostly to the deepest degree (32°C and lower). ARDS was diagnosed in 87 patients and MOF in 82 patients. ICP was monitored in 37 cases, 14 patients had IC hypertension. CONCLUSION: The long-term prognosis in paediatric patients after drowning and neardrowning is generally poor. It depends mainly on the period of the submersion and on the time of CPR. Undercooling is on the contrary to the common conviction not a good prognostic sign. Its protective character can manifest itself only at extremely cold water, when undercooling is faster than hypoxic changes of the brain and other organs tissue appear. Prevention still remains the best treatment. REFERENCES: 1. Quan L, Kinder D: Pediatrics submersion: Pre-hospitals predictors of outcome. Paediatrics 90:909, 199. 2. Rabinovich BA, Lerner ND, Huey RW: Young children ability to climbe fences. Hum Factors 36:733, 1994. 3. Wintemute GJ: Childhood drowning and near-drowning in the United states. Am J Dis Child 144:663, 1990. 4. Zuckerman GB, gregory PM, Santos-Damiani SM: Predictors of death and neurologic impairment in paediatric submersion injuries. Arch Pediatr Adolesc Med 152:134, 1998.

INTRODUCTION: Induction of hypothermia can be achieved using methods based on the four basic mechanisms for heat loss: convection, conduction, evaporation and radiation. A range of cooling techniques have been used in clinical studies, including include ice-water circulating blankets, ice bags, air mattresses, cooling catheters, cooling helmets and caps, cold peritoneal lavage and cardiopulmonary bypass. However, most non-invasive techniques lead to relatively slow decreases in core temperature. Results from animal models and some clinical studies strongly suggest that speed of induction of hypothermia is an important factor in determining outcome [1]; therefore, quicker techniques for lowering of body temperature are needed. The core compartment can be directly cooled using intravenous cooling catheters; however, early studies suggest that reaching temperatures £34oC still takes ±2 hours, plus the time required for catheter insertion. One preliminary study in 22 patients reported that hypothermia could be induced by intravenous infusion of ice-cold ringers lactate [2]. No studies using other types of refrigerated fluids have so far been performed. We therefore assessed the feasibility, speed and complication rates of induction of hypothermia through infusion of large volumes of refrigerated saline or geloplasma in combination with ice-water cooling blankets. METHODS: 62 patients with various types of neurological injury (mostly post-anoxic encephalopathy, subarachnoid haemorrhage, and traumatic brain injury) were included in the study. Hypothermia was induced through infusion of refrigerated fluids (4oC) in combination with ice-water cooling blankets. RESULTS: An average of 2370 ± 860 ml of refrigerated fluids was infused within 50 minutes. Core temperatures decreased from 36.9 ± 1.9 to 34.1 ± 1.3oC at t=30 minutes and to 32.9 ± 0.9oC at t=60 minutes (target temperatures: 32 or 33oC). Continuous monitoring of blood pressure, heart rhythm, CVP, arterial blood gasses, and serum levels of electrolytes, platelets and WBCs revealed no adverse effects. No severe arrhythmia’s occurred in our patients during infusion of refrigerated fluids, although 23 patients received amiodarone because of previous arrhythmias (mostly during or following CPR). MAP increased by 15 mmHg, with the largest increases in blood pressure seen in haemodynamically unstable patients. No patients developed pulmonary oedema. CONCLUSION: Induction of hypothermia with a combination of cold fluid infusion and use of ice-water rubber cooling blankets is a safe, efficacious and quick method to induce hypothermia in patients with various types of neurological injury. REFERENCES: 1.Polderman KH (2004) Intensive Care Med 30:556-75 (review) 2.Bernard S et al. (2003) Resuscitation 56:9-1