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

S143

548

550

ENDOCRINE ABNORMALITIES IN SURVIVORS OF ANEURYSMAL SUBARACHNOID HAEMORRHAGE (SAH)

A DYSPNEA DIFFERENTIATION INDEX (DDI) IN THE ELDERLY IN EMERGENCY

Dimopoulou I1, Kouyialis A2, Kopteridis P1, Tzanella M3, Korfias S2, Armaganidis A1, Thalassinos N3, Sakas D E2, Tsagarakis S3 1 Critical Care Medicine, Attikon Hospital, 2Neurosurgery, 3Endocrinology, Evangelismos Hospital, Athens, Greece

1

INTRODUCTION: Aneurysmal SAH poses a risk for hypothalamic-pituitary dysfunction, given the proximity of these structures to the arterial cycle of Willis. The long-term effects of SAH on endocrine abnormalities have received little systematic attention and discrepant results have been reported regarding the incidence and pattern of hormonal impairments. METHODS: To further clarify this, 21 survivors of SAH (12 men) ranging in age from 20 to 71 years underwent basal hormone measurement and dynamic assessment with the low-dose (1 mcg) corticotropin stimulation test. Hunt & Hess score on admission in the hospital ranged from I to IV and Fisher grade from II to IV. Treatment included embolization (n=12) or surgery (n=8). Endocrine abnormalities were considered if hormone concentrations were below the local reference ranges and, in addition, cortisol hyporesponsiveness was defined if stimulated cortisol was below 20 mcg/dl. RESULTS: Of the 21 patients screened, 12 patients (57%) had 17 endocrine abnormalities. These included low dehydroepiandrosterone sulphate (n=5), high prolactin (n=5), low insulin-like growth factor I (n=3), cortisol hyporesponsiveness (n=2), subclinical hypothyroidism (n=1) and hypogonadism (n=1). There were no correlations between hormone levels and clinical state on admission or severity of bleeding on brain CT scan. CONCLUSION: Endocrine abnormalities are common in survivors of aneurysmal subarachnoid haemorrhage. It remains to be seen whether these should be considered as a potential cause of impaired quality of life and general health.

Borsali-Falfoul N1, Ben jrad G1, Hachicha Y1, Aouina H2, Zaghdoudi I1, Zouari B3, Jerbi Z1 Emergency and critical care, Habib thameur hospital, tunis, 2Pneumology, Charles nicolle hospital, 3Epidemiology department, Medicine Faculty, Tunis, Tunisia

INTRODUCTION: Peak Flow (PEF) is used for rapid separation of cardiac versus pulmonary dyspnoea (1). Ailani (2) used a DDI based on the PEF and PaO2 measurement. In this study we assess the DDI and try to improve it by using the PaO2/FiO2 ratio. METHODS: A one year prospective study including patients aged >ou= 65 YO with dyspnoea seen at an emergency department of a general university hospital. Cardiac Insufficiency (CI) and Pulmonary Insufficiency (PI) diagnosis is established by echocardiography and forced spirometry. Post traumatic and metabolic dyspnoea are excluded. Clinical parameters, blood gaz and PEF are done on admission. We calculate following parameters: DDI= PEF x PaO2/1000 (lmmHg /min) (PaO2 on room-air), DDIr = PEF x (PaO2/FiO2)/1000. %PEF = PEF/predictedPEFx100. %DDI = %PEFxPaO2/1000. %DDIr = %PEFxPaO2/FiO2)/1000. Predicted PEF for each patient estimation is based on height and age. Statistical analysis: chi 2 test with significant statistical difference(SSD) if p<0.05. Sensitivity (Sv), specificity (Sp) and positive predictive value (PPV) for each parameter are calculated. RESULTS: 35 CI and 27 PI with respectively a mean age of 72.2 ± 7.8 years and 70.9 ± 5.4 years. The main results are on table 1. ROC curves analysis showed the optimal cut-off point, Sv, Sp and PPV of the different parammeters in reference of pulmonary cause of dyspnoea at the optimal cut-off points (table 2). Main parameters in patients with CI and PI PaO2/FiO2 PEF l/mn %PEF DDI %DDI DDIr %DDIr (lmHg/mn) lmmhg/mn CI 255±14 155±64 40.2±1 13±10 3.3±2.5 41.8±2 10.9±7 PI 237±7. 109±94 26.9±2 8.6±8. 1.9±1.8 28.6±3 6.6±7. p 0.7 0.002 0.001 0.01 0.02 0.003 0.008 ROC curves analysis PEF l/mn Optimal point Sv % Sp % PPV %

125 73.5 70.8 78.1

%PEF (lmm Hg/mn) 26.9±20 59 79.2 80.8

DDI 8.6±8.4 78.1 72.5 81.7

%DDI (lm mHg/mn) 1.9±1.8 73.3 69.6 78.3

DDIr

%DDIr

28.6±33 81.4 82.3 85

6.6±7.6 83.6 84.4 88.2

CONCLUSION: for fast separation of CI versus PI, DDI has better Sv, Sp and PPV than PEF. The DDIr has higher Sv, Sp and PPV than DDI and doesn’t need to stop oxygenotherapy. REFERENCE: (1)McNamara RM and al. Chest 1992;101:129-32. (2)Ailani RK and al. Chest 1999;116:1100-4. Grant acknowledgement: to Professor Bechir ZOUARI

549 OUTCOME OF PATIENTS WITH DIFFUSE AXONAL INJURY CLASSIFIED BY MAGNETIC RESONANCE IMAGING Ortiz P1, Arruego M A1, Sirvent J M1, Bonet A1, López de Arbina N1, Joly C2, Pedraza S3 1 Intensive Care Unit, 2Neurosurgery, 3Radiology, Hospital Universitario de Girona Dr. Josep Trueta, Girona, Spain INTRODUCTION: Diffuse axonal injury (DAI) is the most common cause of persistent vegetative state and severe disability in trauma brain injury (TBI) patients. Computed tomography (CT) underestimates DAI diagnosis. Magnetic resonance (MR) imaging provides a better knowledge of the DAI lesions and also a way of classification. It is still not clear if the MR is a predictor of the outcome in patients with DAI. In this study we evaluated the functional outcome at the discharge from intensive care unit (ICU), after six months and after a year of TBI in patients with DAI classified by MR imaging. METHODS: This is a cohort study with patients admitted in ICU with TBI diagnosis from June 2.001 to December 2.002. Patients with Glasgow Coma Score (GCS) < 9, TCDB types II and III of the CT scan classification and impairment of consciousness after stopping sedation were enrolled on a MR protocol of DAI study in the initial two weeks of the TBI. In the group of patients with a DAI diagnosis a classification was made, considering DAI 1 as white matter lesions, DAI 2 as white matter lesions and corpus callosum and DAI 3 as white matter, corpus callosum and brainstem. Glasgow outcome score (GOS) was obtained by a personal clinical evaluation at the discharge from ICU, after six months and after a year of TBI. The data obtained was compared statistically by t-test, chi-square test and a Friedman test (p<0.05). RESULTS: One hundred seventy-five patients with TBI were studied. The demographic data were, age, mean(SD): 37.5(19.4) years; sex, n: M/F: 101/74; GCS, mean(SD): 8.2(3.9), Injury severity score (ISS), mean(SD): 25.9(13.4) and mortality rate: 20%. Twenty-five patients were included in the MR protocol of DAI study. The data of this population were, age: 31.8(21.5) years; M/F: 20/5; GCS: 6(2.4), ISS: 31.6(14.8) and mortality rate: 12%. ISS was higher in the MR population with significant differences (p=0.04). Finally, 23 DAI diagnostics were made, DAI 1: 3, DAI 2: 9 and DAI 3: 11 patients. The median of GOS at discharge from ICU was 2, after six months 3 and after a year 4, (p=0.001). Improvement of the GOS was found either after 6 months (p=0,001) and after a year post injury (p=0,0001) respect from the GOS at the discharge from ICU. No correlation was found between the types of DAI and GOS in the three periods studied. CONCLUSION: DAI is associated with higher ISS. Patients with DAI have an improvement in the functional outcome after six months of TBI. Outcome is even better up to a year post injury compared with the functional outcome at the ICU discharge. REFERENCE: Paterakis K, et al. Outcome of patients with diffuse axonal injury: the significance and prognostic value of MRI in the acute phase. J Trauma 2000; 49:1071-5.

Oral Presentations Clinical studies on pathogenesis of sepsis – 551555 551 SIGNIFICANT SUPPRESSION OF MSNA DURING EXPERIMENTAL ENDOTOXEMIA IN NORMAL HUMANS Sayk F1, Vietheer A1, Wellhöner J P1, Schaaf B2, Dodt C1 Medizinische Klinik 1, 2Medizinische Klinik 3, UKSH Campus Luebeck, Luebeck, Germany

1

INTRODUCTION: Sepsis is known to inflict significant reduction of vascular resistance. Muscle sympathetic nerve activity (MSNA) directly influences vascular tone via the baroreceptor-reflex (BRR) thereby regulating blood pressure. Never before MSNA had been investigated in experimental endotoxemia in humans. METHODS: Young healthy male volunteers randomly either received a bolus of lipopolysaccharide (LPS; n=8) or placebo (saline; n=7). TNF-? and IL-6 documented the systemic immune response. MSNA was recorded microneurographically at rest from the peroneal nerve and oscillometric blood pressure and heart rate were monitored prior to (baseline) and 90 min postinjection. Additionally, MSNA and blood pressure were correlated at both phases during BRRstimulation with nitroprusside and phenylephrine. RESULTS: Endotoxin but not placebo induced flu-like symptoms and elevated cytokine levels. Compared to baseline resting MSNA (mean +/- SEM in bursts/min) was profoundly suppressed 90 min after LPS injection (27.5±3.3 vs 12.1±2.9; p=0.003) whereas MSNA was slightly increased in the placebo group (20.0±3.4 vs 27.3±4.5; p=0.214). Heart rate was elevated following LPS (60.6±2.0 vs 78.4±3.1; p=0.0002) in contrast to placebo (63.0±3.3 vs 63.3±2.9). Systolic or diastolic blood pressure did not differ significantly (LPS: sys 129.4±2.7 vs 135.8±3.4; dias 71.6±1.5 vs 70.5±2.1; placebo: sys 125.6±2.5 vs 126.9±1.9; dias 71.4±1.9 vs 71.7±1.8). BRRcorrelation of MSNA and blood pressure was significantly altered in the endotoxin-group but not following placebo. CONCLUSION: Experimental endotoxemia with an immunologic response equivalent to early sepsis induces a significant reduction in sympathetic outflow to the muscle vascular bed 90 min after LPS administration. Furthermore, BRR-function was altered. These unexpected findings could be explained in part by central nervous cytokine effects and give new insight into the pathophysiology of the cardiovascular response to septic insults.

S144

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

552

554

DISTURBANCES OF PROTEIN C ACTIVITY IN ABDOMINAL SEPSIS: DIAGNOSTIC AND PREDICTIVE IMPORTANCE

IMPROVEMENT OF MUSCLE TISSUE DEOXYGENATION DURING STAGNANT ISCHEMIA IN SURVIVORS FROM SEVERE SEPSIS

Karamarkovic A R1, Radenkovic D B1, Popovic N N1, Sijacki A N1, Stefanovic B B1 1 Dpt. for surgical infection and sepsis, University Center for Emergency Surgery Belgrade, Belgrade, Yugoslavia

1

INTRODUCTION: Despite potent antibiotics, modern intensive care, and aggressive surgical treatment, up to 30% of patients still die of severe secondary peritonitis(1-2).Systemic inflammatory process is often well under way before the clinical signs and symptoms of sepsis are present (3). Protein C system plays a crucial role in control of microvascular coagulation and inflammation (4).Septic cascade are initiated before clinical manifestations which indicates the significance of determination of early biological markers of sepsis (5). METHODS: This prospective study refers to diagnostic and prognostic importance of protein C detection during abdominal sepsis. We treated surgically 22 patients with severe generalized peritonitis with sepsis syndrome vs. 15 patients with hernia repair(control group). RESULTS: Protein C: As for the group treated for abdominal sepsis, preoperative values were decreased in 55% of the patients, while more severe protein C activity disorders (below 60%) were evidenced in 18%. Preoperatively, average protein C activity was significantly lower in the group of the patients who developed septic shock in the late course of the disease and died, instead the patients in the group with severe diffuse peritonitis and sepsis who survived (mean protein C activity 57.2%,vs 81.1%; p-0.037). Difference in postoperative protein C values is confirmed as statistically highly significant on the postoperative days 7(52.5%,vs 81.1%; p-0.032) and 10(56.0%,vs 95.5%; p-0.021). As for the control group, protein C activity is permanently maintained within normal range. Difference in protein C activity between the control and study groups during the whole studied period was highly statistically significant (p-0.001). CONCLUSION: The results and multivariate regression statistical analysis revealed the protein C as very sensitive biological marker of septic cascade (p-0.0001). According to our results the low level of protein C consistently correlate with disease severity and suggest the development of septic shock and poor outcome. REFERENCES: 1.Teichman W, Herbig B. Scheduled Reoperations for Diffuse Peritonitis. Dig Surg 1996;13: 396-9. 2. Wittmann DH. Operative and nonoperative therapy of intra-abdominal infections.Infection 1998; 26: 335-51. 3.Grunau G, Heemken R, Hau T. Predictors of outcome in patients with post-operative intra-abdominal infection. Eur J Surg 1996; 162: 619-25. 4.Bernard GR, Vincent JL, Laterre PF. Efficiacy and safety of recombinant human activated protein C for severe sepsis. N Engl J Med 2001; 344: 699-709. 5.Macias W, Derchak PA.: Coagulopathy may occur before the clinical diagnosis of organ dysfunction. Crit Care Med 2003; 31 (1 Suppl): S100S110.

Pareznik R1, Voga G1, Kne?eviè R1, Podbregar M1 Department for internal intensive medicine, General hospital Celje, Celje, Slovenia

INTRODUCTION: Sepsis is characterized by defect in tissue oxygen consumption (1,2). Tissue deoxygenation in ischemic animal muscles was lower after treatment with lipopolysaccharide (3). In patients the effect of sepsis on rate of muscles tissue deoxygenation during stagnant ischemia is still controversial (4). METHODS: The aim of our study was to determine the rate of thenar muscles tissue deoxygenation during stagnant ischemia in haemodynamically stabilized septic patients (8 severe sepsis, 6 septic shock) compared to 17 healthy controls. Thenar muscles tissue oxygen saturation (StO2) was measured non invasively by near-infrared spectroscopy (NIRS) (InSpectra ™, Hutchinson Technology Inc., USA) before and during uper limb ischemia induced by automatic cuff inflation to 260 mmHg, until StO2 decreased to 40%. Measurements were made: after haemodynamic stabilization (during 48 hours after admission), at day 7 and at ICU discharge. RESULTS: Baseline StO2 did not differ between different groups (controls 79±9%, severe sepsis at admission 78±7%, severe sepsis at ICU discharge 82±9%, septic shock at admission 80±9%, septic shock at 1 week 77±17%, septic shock at ICU discharge 76±9%). The rate of StO2 decrease in stagnant ischemia at admission was lower in septic shock compared to severe sepsis and controls (-5±2%/min vs. -12±2%/min vs. -37±7%/min, respectively; p<0.001). At 1 week the rate of StO2 decrease in septic shock was comparable to severe sepsis at admission (-11+1%/min vs. 12±2%/min, respectively). At ICU discharge StO2 decrease did not differ between septic shock and severe sepsis group (-22±3 vs. -21±2 %/min, respectively), but it was still slower compared to controls (p<0.001). ICU stay was longer in septic shock group compared to severe sepsis group (28±8 vs. 8±3 days, respectively, p<0.05). CONCLUSION: At admission thenar muscles tissue deoxygenation provoked by stagnant ischemia was significantly slower in septic shock patients compared to severe sepsis and healthy volunteers. After improvement of sepsis muscle tissue deoxygenation rate increased in both septic shock / severe sepsis patients, but was still slower compared to healthy volunteers. The rate of muscle tissue deoxygenation during stagnant ischemia could potentially be used as a marker of oxygen uptake capability in septic patients. REFERENCES: 1.Ince C, Sinaasappel M. Microcirculatory oxygenation and shunting in sepsis and shock. Crit Care Med 1999; 27: 1369-77. 2.Fink MP. Cytopathic hypoxia: Is oxygen use impaired in sepsis as a result of an acquired intrinsic derangement in cellular respiration? Critical Care Clinics 2002; 18: 165-75. 3.Boushel R, Piantadosi CA. Near-infrared spectroscopy for monitoring muscle oxygenation. Acta Physiol Scand 2000; 168: 615-22. 4.Sair M, Etherington PJ, Winlove P, Ewans TW. Tissue oxygenation and perfusion in patients with systemic sepsis. Crit Care Med 2001; 29: 1343-49.

553

555

INCREASED PERMEABILITY AND LACTATE CONCENTRATIONS IN THE RECTAL MUCOSA IN PATIENTS WITH SEPTIC SHOCK

SEVERE METABOLIC ALTERATIONS IN ADIPOSE TISSUE DURING EARLY ENDOTOXEMIA IN HUMANS

Jørgensen V L1, Nielsen S L2, Espersen K3, Perner A1 1 Dept. of Intensive Care, Herlev Hospital, Copenhagen, 2Dept. of clinical physiology, Odense Hospital, Odense, 3Dept. of Intensive Care, Rigshospitalet, Copenhagen, Denmark

1

INTRODUCTION: Gut barrier dysfunction may contribute to morbidity and mortality in septic shock, but little is known about the mucosal barrier in these patients. Therefore, we estimated the paracellular permeability and L-lactate concentration in the rectal mucosa in healthy subjects and critically ill patients with or without septic shock. METHODS: Seven patients with septic shock, six ICU controls and six normal subjects with an intact rectum and left colon were included. All patients were fluid resuscitated, sedated, and mechanically ventilated and were stabile during study. The patients with septic shock (median SAPS II 49 and SOFA 16) had an abdominal focus of infection and were treated with norepinephrine (0,04-0,47 ?g/kg/min) to MAP>70 mm Hg. The ICU controls (SAPS II 42 and SOFA 10) had sepsis with pulmonary focus but normal cardiovascular function and blood gas values. The rectal permeability was determined by the systemic recovery after 1 hour of 99-TcDTPA in 50 ml of isotonic saline instilled into the rectal lumen. The volume of distribution of DTPA was estimated by an i.v. bolus of 51-Cr-EDTA, which is an analogue of DTPA. 99-Tc and 51-Cr-activities were assessed by gamma-spectrometry (Packard Instrument, Meriden, LT). The concentration of L-lactate in the rectal lumen was subsequently measured by a 4 hour equilibrium dialysis as previously described.1 RESULTS: In patients with septic shock, the systemic recovery of Tc-DTPA and luminal concentration of L-lactate were increased compared to controls (see Table). The Tc recovery was dependent (after log transformation) of the luminal concentration of L-lactate (linear regression analysis, r2=0.82, p<0.01) and the dose of norepinephrine (r2=0.66, p=0.03) in the shock patients. RECTAL PERMEABILITY AND LACTATE LEVELS Tc-DTPA recovery, ‰ Normal control medians, n=6 0.03 Ranges (0.01-0.07) ICU control medians, n=6 0.07 Ranges (0.01-1.2) Septic shock medians, n=7 0.14* ranges (0.05-1.4) P<0.05 vs. *normal or **ICU controls

Luminal L-lactate, mM 1.1 (0.6-1.5) 1.5 (0.4-4.0) 5.3* ** (1.3-7.8)

CONCLUSION: In patients with septic shock, the permeability and concentration of L-lactate in the rectal mucosa may be increased. The results suggest that the barrier defect may be induced by metabolic dysfunction secondary to norepinephrine treatment or shock severity. REFERENCE: 1. Due VL et al. BJA 2002;89:919

Wellhoener P1, Sayk F1, Vietheer A1, Weitz G1, Dodt C1 Internal Medicine I, UKSH- Campus Luebeck, Luebeck, Germany

INTRODUCTION: Endotoxemia caused by infection with Gram-negative bacteria is one of the major causes for the sepsis syndrome and the early symptoms of gram-negative sepsis like fever, rigors and cytokine release can be mimicked by intravenous injection of purified bacterial endotoxin. During endotoxemia metabolism is accelerated leading to a decrease of lean body mass despite high caloric intake. Although, in patients with sepsis fat is the preferred fuel for oxidation, little is known about the metabolic changes in adipose tissue that accompany early sepsis. METHODS: In order to investigate the effect of early endotoxemia on adipose tissue metabolism and blood flow we performed a randomized, double-blind experiment in 20 male, healthy and non-smoking volunteers (LPS group: n=11, mean age 28.7 ± 2.6, placebo group: n=9, mean age 26.3 ± 5.3). Participants received either 4ng/kg LPS (E.coli O:113) or placebo iv. Cardiovascular parameters were digitally recorded; skin- and adipose tissue blood flow was monitored with laser doppler flowmetry (LDF). We used microdialysis in femoral adipose tissue to measure interstitial lactate, glycerol, pyruvate and glucose. Statistical analysis of the effects of LPS versus placebo relied on ANOVA with a repeated measure factor „time“ and the condition „treatment“. RESULTS: As expected, endotoxin induced flu-like symptoms starting 60 min. after injection. Body temperature rose 180 minutes after LPS iv from 36.1°C to 37.3°C (p<0.05). Interstitial pyruvate concentrations increased significantly after 90 min (p>0.05) while interstitial glucose and lactate were not affected by LPS treatment (p>0.7). Glycerol release from adipose tissue after LPS rose slowly and was significantly different from controls 180 min after injection (p<0.05). Furthermore, laser doppler flow revealed a marked increase in subcutaneous adipose tissue blood flow of 243 ± 46% but a 36 ± 7% decrease in skin perfusion 180 min after LPS. CONCLUSION: Low dose endotoxemia in otherwise healthy humans causes profound changes in adipose tissue metabolism that can be detected by microdialysis. Especially interstitial pyruvate which is a marker of high glycolytic flux and increased fat oxidation seems to be a sensitive and early marker of those metabolic alterations. Although adipose tissue is a main source of lactate production under basal conditions, it does not contribute significantly to lactic acidosis that is commonly seen during endotoxic shock or sepsis. Furthermore, endotoxemia causes a vasoconstriction of skin vessels, and a parallel vasodilatation of subcutaneous tissue.

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

S145

558

Oral Presentations Outbreaks and multiresistance – 556-560 556

INTRAICU INFECTIONS AND ISOLATION POLICIES: THEIR INFLUENCE ON ACTIVITY AND COSTS OF A GENERAL ICU Abizanda R1, Reig R1, Belenguer A1, Sánchez F1, Mateu L1, Vidal B1, Iranzo J1 1 ICU Department, Hospital Univesitario Asociado General de Castelló, Castelló, Spain

CONTROL OF MRSA ENDEMIA WITH ENTERAL VANCOMYCIN IN A BURN INTENSIVE CARE UNIT Abella A1, De la Cal M A1, Cerdá E1, López L1, Alía I1, García-Hierro P2, Aranguren A3. 1 Intensive Care Unit, 2Department of microbiology, 3Department of pharmacy, Hospital Univesitario de Getafe, Getafe, Spain

INTRODUCTION: Nosocomial infections in ICU represent an important increase in severity, in length of stay, in pharmaceutical and other resources consumption, and in mortality rates. Our aim is to analyze the influence of these aspects on our ICU activity.

INTRODUCTION: The usual recommendations to control MRSA endemia have often failed in different settings. The efficacy of the use of enteral vancomycin has been demonstrated in the control of MRSA without adverse effects in multidisciplinary Intensive Care Units (ICU) without vancomycin-resistant Enterococcus (VRE). The objective of this study is to evaluate the impact of digestive administration of vancomycin in the control of MRSA endemia in a burn ICU, without VRE.(1)

METHODS: Retrospective analysis of our ICU information data set during the time period between January 2002 and September 2003. Identification of infections appearing during the ICU stay and patients in which isolation measures were set on. Analysis of ICU stays, nursing workload (NEMS system), comparing infected patients with overall indicators for the study period. Admissions due to coronary disease and atrioventricular blocks as reasons for admission were excluded of the analysis.

METHODS: Two consecutive periods were compared (from January-1995 to January-2000 and from February-2000 to December-2002). During the first period, the traditionally recommended methods were applied: hygiene, surveillance sampling (nose, pharynx, rectum and burn wound) and carriers’ isolation. During the second period, besides the aforementioned procedures, all admitted patients were given selective digestive decontamination (tobramycin, colistin and amphotericin B) and vancomycin (nasal and oropharynx 4% paste and digestive solution, 500 mg).every 6 hours.

RESULTS: One thousand two hundred and eighty-eight (1288) patients were admitted to ICU during the study period. Mean age was 61 years; SAPS 2: 30.5 ± 15.9; estimated mortality risk (PI SAPS 2) 40 ± 25, with a global hospital mortality of 12.3 %, with an standardi CONCLUSION: Estimation of IAP via IGP or IBP is feasible. The COVA for these parameters and APP in sedated mechanically ventilated patients is around 15 to 20% (ranging from 5 up to 50%) during a 24h period and thus varies substantially. These variations may even be more pronounced in nonsedated patients. Therefore IAP and APP are continuous variables like any other pressure and should be monitored as often as possible during the day to adapt treatment accordingly.

RESULTS: The RR of acquiring MRSA (period 2 versus period 1) after adjusting for age, body surface burn, full-thickness burn and inhalation injury was 0.19 [CI 95 % (0.11 to 0.32)]. During period 1, 6 % of patients carried MRSA at admission to the ICU; in period 2, 3 %. There was no case neither of VRE nor vancomycin intermediate-resistant S aureus. Period 1 (n=402) Age (years), mean 46 Body surface burn (%), mean 31 Full-thickness burn (%), mean 15 Inhalation injury, % 41 Mechanical ventilation, % 49 Patients with ICU-acquired MRSA in 16.5 surveillance or diagnostic samples, per 1000 days Patients with ICU-acquired MRSA in diagnostic 10.9 samples, per 1000 days

Period 2 (n=274) 47 24 11 39 41 4.1

p 0.59 0.002 0.006 0.34 0.14 0.0000

1.9

0.0000

REFERENCE: (1) Malbrain MLNG. Intensive Care Med 2004; 30(3): 357-371.

CONCLUSION: The administration of enteral vancomycin is effective in the control of MRSA and is not associated to the apparition of resistance in a burn ICU free form VRE. REFERENCE: 1- de la Cal MA et al. Effectiveness and safety of enteral vancomycin to control endemicity of methicillin-resistant Staphylococcus aureus in a medical/surgical intensive care unit.J Hosp Infect. 2004;56:175-83. Grant acknowledgement: Funded partially by FIS 02-1883

557

559

DECREASING THE INCIDENCE OF EXTENDED SPECTRUM BETA-LACTAMASES A MULTIPRONGED STRATEGY

CONTROL MEASURES FOR MRSA AND THEIR IMPACT ON THE PREVALENCE OF A. BAUMANNII (Ab) IN AN ICU Arvaniti K1, Lathyris D1, Adamopoulos C1, Tsagourias M1, Synnefaki E1, Matamis D1 Intensive Care Medicine, Papageorgiou General Hospital, Thessaloniki, Greece

1 1,

T1,

S1,

U2,

V1

Kumar A A Sekar Kamat Thodur V M Rao V anesthesiology And Intensice Care, 2microbiology, SRI Ramachandra Medical College And Research Institute, Chennai, India

1

INTRODUCTION: Plasmid mediated spread of ESBL strains is a worldwide concern. It may be because of a clonal spread or the result of an antibiotic selection pressure due to indiscriminate use of third generation cephalosporins.our aim was to analyze if a strict antibiotics policy and scrupulous infection control practice will help reducing an already high ESBL rate. METHODS: This was a prospective observational study conducted in a 20-bedded intensive care unit of a tertiary care university teaching hospital. In 2003, we reported a high incidence of ESBL strains (45.07%) among the gram-negative infections in our ICU. Plasmid profiling and gene sequencing showed that this was a combination of both a clonal spread and antibiotic selection pressure. Hence, in the month of march 2003 we embarked upon a strict antibiotic policy aimed at cutting down the usage of third generation cephalosporins. In addition, scrupulous infection control policies were put in place which involved strict hand washing, use of alcohol based handrubs before examining patients and strict isolation practices. Subsequently , we analysed blood, urine and respiratory samples on all patients with nosocomial infections. ESBL strains were identified using E-Test and their MIC’s were obtained. We also analysed the use of third generation cephalosporins and carbapenams during this period. RESULTS: The incidence of ESBLproducing strains decreased from 45.07% to 30.77%,as the absolute use of ceftazidime decreased from 153.54 gms/100 patient days to 56.20 gms/100 patient days.The piperacillin tazobactam usage increased from 113 to 290.35 gms/100 patient days. As the usage of carbapenems increased the incidence of metallobetalactamases producing organisms increased from 8 to 14%.This is in accordance with other studies1,2. CONCLUSION: A high incidence of ESBL strains can be reduced by strict infection control and antibiotic restriction policies. It remains to be seen if we can further reduce the incidence. However the increasing incidence of MBL is alarming. REFERENCES: 1. Thodur.v.m. patterns of infection in a multidisciplinary Indian icu over time intensive care medicine 2003;29:278(suppl). 2. Patterson JE et al. Infect Control Hosp Epidemiol. 2000;21:455-458.

INTRODUCTION: In our ICU, during a 8-month period, screening for nasal-MRSA revealed 16,3% MRSA-carriers (22/135), 41% admitted-carriers (9/22) and 59% acquired-carriers (13/22). MRSA-Nosocomial Infections (NIs) and Ab-NIs were 14,5% (8/55 NIs) and 28,9% (16/55 NIs), respectively. For the next 18 months (control period-CP), aggressive control measures were applied to decrease the transmission of MRSA. Comparison between the two periods, showed significant decrease of MRSA-carriage in the CP (8,2%-30/367, p<0,05). MRSA-acquisition fell to 43% (13/30, p<0,01) while admitted-carriage remained stable (17/30, 57%, p=NS). In the CP, MRSA-NIs didn’t change (12,8%-19/148 NIs) while Ab-NIs remained high (41,89%-62/148 NIs). METHODS: An observational study (study period-SP) was conducted in the next 12 months to evaluate the impact of additional measures on the prevalence of Ab-NIs. Standard measures in the two periods included: prophylactic contact-isolation of suspected MRSA-carriers, contactisolation according to the antibiogram, geographical isolation when possible, Sterilium bedside and education of the personnel. Additional measures applied only in the SP were: Ab-screening by oropharyngeal, cutaneous and rectal specimens at admission and once-weekly and isolation of the Ab-carriers. MRSA-carriage, MRSA-NIs and Ab-NIs were compared between CP and SP. RESULTS: We compared 367 patients in the CP with 301 patients in the SP. MRSA-carriage didn’t change (8,2%-30/367 and 4,98%-15/301 resp, p=NS). Neither admitted MRSA-carriage (57% vs 46,67% resp, p=NS) nor acquired MRSA-carriage changed significantly (43%-13/30 vs 53,33%-8/15 resp, p=NS). MRSA-NIs remained also stable (12,8%-19/148 NIs vs 8,27%-11/133 NIs, p=NS). Despite our additional measures, Ab-NIs didn’t decrease but raised from 41,89%62/148 NIs in the CP to 50,38%-67/133 NIs in the SP, p=NS. Ab-carriage during the SP was 21%65/301 patients and acquired-carriage quite high (84,6%-55/65 patients or 18,9% of the SPpopulation). Colonization pressure differed between MRSA and Ab in the SP (15 MRSA-carriers vs 65 Ab-carriers). Mean age (43,3+/-18,4), APACHE II score (18,33+/-5,8), antibiotic policy and mean nurse:patient ratio (1:2 to 1:3) were similar in the two periods. CONCLUSION: In our ICU, early detection and isolation of carriers was not effective against Ab, even though effective against MRSA. Possible explications are: different epidemiology of the two pathogens, eventual inanimate reservoirs of Ab and continuously elevated colonization pressure due to an important number of Ab-carriers during the SP. An investigation of this Aboutbreak is already in progress

S146

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

560

562

CONTACTS BETWEEN ICU PATIENTS AND NON-UNIT EMPLOYEES DURING AN MRSA OUTBREAK

EXTRAVASCULAR LUNG WATER ASSESSED BY TRANS-PULMONARY SINGLE THERMODILUTION AND GRAVIMETRY IN SHEEP

Spanjersberg R1, Ligtenberg J J1, Tulleken J E1, Werf van der T S1, Zijlstra J G1 1 Intensive and Respiratory Care Unit, University Hospital Groningen, Groningen, Netherlands

Kirov M Y1, Kuzkov V V1, Kuklin V N1, Waerhaug K1, Bjertnaes L J1 1 Department of Anesthesiology, Institute of Clinical Medicine, University of Tromsoe, Tromsoe, Norway

INTRODUCTION: During a period of MRSA outbreak at the University Medical Centre Groningen, one of the four ICU’s was closed. To minimize the risk of MRSA transmission by hospital employees to our ICU, restrictive visiting rules were introduced. We studied the adherence to the protocol and examined how often consulting specialists visited our ICU. METHODS: The prevention protocol introduced was characterized by the following: First, a hospital employee was allowed to enter the ICU only if necessary for patient’s treatment. Second, the hospital employee was asked to fill in his/her name on a special form and to contact the involved nurse by phone. Third, at entrance, all non unit hospital employees were confronted with strict infection precaution rules. RESULTS: The number of beds in our medical ICU was extended from 12 to 14 during week days; bed occupancy was > 95% and the admitted number surgical and trauma patients increased. During February 2004 there were 1020 consulting visits. All hospital employees agreed to adhere to the protocol. MRSA was not found in any of the obtained cultures of the admitted patients to our ICU in this period.

physicians physical therapy radiology ward nurses patient related material related

weekdays median (range) 16 (7-31) 8 (6-11) 5 (2-14) 6 (0-18) 3 (0-6) 3 (1-7)

weekend median (range) 6 (5-13) 1 (1-6) 5 (2-7) 3 (0-10) 1 (0-5) 1 (0-5)

P-value 0.001 0.0001 0.31 0.14 0.07 0.04

CONCLUSION: A substantial decrease in the number of visits of physicians and physical therapists during weekends was noted, while the patient population didn’t change. This may indicate that there were too many visits during week days. Whether this number can be structurally reduced remains to be established and warrants further evaluation of visiting reasons. The measures we introduced to minimize MRSA transmission may be of additional value in preventing the further spread of other resistant pathogens throughout the hospital.

INTRODUCTION: Acute lung injury (ALI) has been associated with accumulation of extravascular lung water (EVLW). Transpulmonary single thermodilution (ST) was recently introduced as a bedside technique for assessment of EVLW (1). However, this new method has not been validated against postmortem gravimetry (EVLWG), which is supposed to be the gold standard for quantification of EVLW. Thus, our aim was to find out in two ovine models of ALI, whether a correlation exists between EVLW as determined by ST (EVLWST) and by gravimetry (EVLWG). METHODS: Eighteen sheep were instrumented with vascular catheters. To induce ALI of different severity, awake animals received intravenously Escherichia coli lipopolysaccharide (LPS) 15 ng/kg/min for 6 h (n=7) or oleic acid (OA) 0.06 ml/kg for 30 min (n=7). A third group (n=4) was sham-operated. All groups received Ringer’s lactate 5 ml/kg/h. EVLWST was measured by using a PiCCO plus monitor (Pulsion Medical Systems), and the last measurement was compared with EVLWG (2). EVLWST was estimated as a difference between intrathoracic thermal volume and intrathoracic blood volume (ITBV). For calculation of EVLWST and ITBV, we introduced a specific correction coefficient for sheep based on ITBV = 1.34 x global end-diastolic volume. To evaluate the relationship between EVLWST and EVLWG, we used linear regression and BlandAltman analysis. A p value of < 0.05 was regarded as statistically significant. RESULTS: At the end of experiment, EVLWST was 8.9±0.6, 11.8±1.0, and 18.2±0.9 ml/kg in sham-operated, LPS, and OA groups, respectively (mean ± SEM, p < 0.05). The regression analysis between in vivo EVLWST and post mortem EVLWG resulted in EVLWST = 1.30 x EVLWG + 2.32 (n = 18, r = 0.85, p < 0.0001). The mean bias±2SD between EVLWST and EVLWG was 4.91±5.08 ml/kg (p < 0.001). CONCLUSION: The determination of EVLWST in sheep correlates closely with gravimetric measurements over a wide range of changes. Thus, despite a moderate overestimation of EVLW, we consider the single transpulmonary thermodilution to be a useful tool for assessment of pulmonary oedema during ALI. REFERENCES: 1. Sakka SG et al. Intensive Care Med 2000;26:180-187. 2. Pearce ML et al. Circ Res 1965;16:482-488 Grant acknowledgement: Supported by Helse Nord (Norway) and Pulsion Medical Systems (Germany).

Oral Presentations Monitoring I – 561-565 561 ESTIMATION OF TOTAL PEEP DURING ONGOING MECHANICAL VENTILATION BY USING ARTIFICIAL NEURAL NETWORKS Perchiazzi G1, Rylander C2, Vena A3, Dello Russo M1, Brienza N1, Giuliani R1, Fiore T1, Hedenstierna G3 1 Emergency and Transplant, Bari University, Bari, Italy, 2Anaesthesia, Sahlgrenska University Hospital, Göteborg, 3Clinical Physiology, Uppsala University, Uppsala, Sweden INTRODUCTION: Artificial neural networks (ANN) are universal function approximators and can be used to extract information from complex signals. Aim of the present experiment is to test the capability by ANN of estimating the static total positive end-expiratory pressure (PEEPst) from the tracings of flow and pressure during ongoing mechanical ventilation, in absence of an end-expiratory hold manoeuvre (EEHM). PEEPst derives from the effects of applied PEEP (PEEPe) and intrinsic PEEP (PEEPi). In order to create the array of examples of PEEPst to train the ANN, we shortened progressively the expiratory time (TE): this procedure determines the formation of intrinsic PEEP. METHODS: Nine pigs were anaesthetized, tracheostomized and ventilated in a volume control mode. The time constant of the respiratory system (TC) was measured at the beginning of each experimental session. Keeping respiratory rate and tidal volume constant, we progressively shortened TE in proportion to multiples of TC, obtaining different respiratory patterns (Rpat), by imposing a TE of 4.0, 3.0, 2.5, 2.0, 1.5, 1.0 TC. Pressure (P) and flow (F) at airways opening during ongoing mechanical ventilation were recorded at the different implemented Rpat. The last breath had an EEHM, that was used to compute the reference PEEPst. In order to increase the number of different conditions creating PEEPst, we added an external resistance (Rext) in series with the endotracheal tube during the application of two levels of PEEPe. After that, the entire procedure was repeated, after having induced acute lung injury by injection of oleic acid (OA). 382 tracings were obtained: 305 were used for training, 77 for testing the ANN.The ANN was implemented on a computer via software and was composed of 40 input, 18 hidden and 1 output neurons. The ANN had to extract PEEPst from inspiratory P and F signals during ongoing mechanical ventilation. The results by the ANN were compared with the manually calculated PEEPst, using the EEHM. Linear regression (LR) and Bland and Altman analysis were used to test the ANN performance. RESULTS: LR between ANN output (y) and expected values (x) was y=1.005x-0.011 with a correlation coefficient R=0.984. According to Bland & Altman method, bias and standard deviation of the error were 0.04±1.88 cmH2O. CONCLUSION: These results show that ANNs can estimate PEEPst even under dynamic conditions that cause different levels of intrinsic PEEP. This achievement may open new possibilities in monitoring technology. Grant acknowledgement: The Swedish Medical Research Council (5315)

563 VALIDATION OF 2 NOVEL INTRA-ABDOMINAL PRESSURE MEASUREMENT DEVICES Malbrain M L N G1, Libeer C1, Nijs J1, Deeren D1, De Potter T1. 1Intensive Care Unit, Ziekenhuisnetwerk Antwerpen, site Stuivenberg, Antwerpen, Belgium INTRODUCTION: Definitions for intra-abdominal hypertension and abdominal compartment syndrome stand or fall with the reproducibility and accuracy of the intra-abdominal pressure (IAP) measurement (1). The aim of this study is to validate a novel fully-automated continuous technique to measure IAP via a balloon-tipped catheter connected to an IAP monitor (Spiegelberg, Hamburg, Germany) versus a bladder FoleyManometer (Holtech Medical, Copenhagen, Denmark) and to look for the coefficient of variation (COVA, defined as the SD divided by mean) for IAP and abdominal perfusion pressure (APP) during different 24h periods. METHODS: IAP was estimated using 2 different METHODS: via the bladder (IBP) and via a balloon-tipped gastric catheter (IGP) connected to an IAP monitor. In total 2029 IGP and 705 IBP measurements were performed in 22 sedated mechanically ventilated ICU patients. Correlation between IGP and IBP was studied in 705 paired samples. The M/F ratio was 1/1, age 63±11.6, APACHE-II 23±9.5, SAPS-II 55±16.8. RESULTS: The values for IAP (mmHg) were 9.7±3.3 (IGP) vs 9.9±3.2 (IBP). There was a good correlation between both measurements (Figure 1): IBP= 0.86 x IGP + 1.6 (R2=0.81, p<0.0001). Bland and Altman analysis showed good agreement: IGP was almost identical to IBP with a mean bias of –0.2±1.5(SD) mmHg (95%CI –0.3 to 0); the limits of agreement (LA) were –3.1 to 2.8 mmHg (95%CI –3.3 to –2.9 for the LLA and 2.6 to 3 for the ULA). The COVA was 18.6±7.6% for IGP (range 4-46), 16.8±12.2% for IBP (range 0-58) and 14.4±5.9 for APP (range 5-48).

CONCLUSION: Estimation of IAP via IGP or IBP is feasible. The COVA for these parameters and APP in sedated mechanically ventilated patients is around 15 to 20% (ranging from 5 up to 50%) during a 24h period and thus varies substantially. These variations may even be more pronounced in non-sedated patients. Therefore IAP and APP are continuous variables like any other pressure and should be monitored as often as possible during the day to adapt treatment accordingly. REFERENCE: (1) Malbrain MLNG. Intensive Care Med 2004; 30(3): 357-371. Grant acknowledgement: ESICM 2003 Chris Stoutenbeek Award

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

564 ACCURACY OF NEW “MOTION-RESISTANT” PULSE OXIMETERS DURING CARDIAC ARRHYTHMIA AFTER CARDIAC SURGERY Lutter N1, Kozma E1, Mell J1, Schuettler J1 Department of Anesthesiology, University of Erlangen-Nuremberg, Erlangen, Germany

1

INTRODUCTION: Within the first five days after cardiac surgery low perfusion, decreased cardiac output, and cardiac arrhythmias represent common findings which affect the arterial pulse waveform (1, 2). This study was designed to determine the impact of cardiac arrhythmia on the accuracy of saturation estimates (SpO2) and pulse rate readings (PR) of four new “motionresistant” pulse oximeters. METHODS: After IRB approval and informed consent, 110 ICU patients (ASA II-IV, 28 to 89 yr) suffering from cardiac arrhythmia after cardiac surgery were enrolled into this study. Patients with low cardiac output were precluded from this study as were patients with inadequate signal strength (perfusion index of Philips CMS < 0.5). All patients were connected simultaneously to a Nellcor N-595, a Philips CMS, a Masimo Radical (Masimo SET V3), and a Dolphin Medical 2001 pulse oximeter utilizing randomly placed proprietary finger probes. Prior to and at the end of each measurement an in vitro blood sample analysis (Radiometer ABL615 and ABL625) was performed. SpO2, PR and referential heart rate (HR) were recorded continuously and alarm events were classified immediately into technical/physiological and false/correct to calculate sensitivity [TP/(TP+FN)] and specificity [TN/(TN+FP)] (TP = true positive, FP = false positive, TN = true negative, FN = false negative). RESULTS: Correlating the mean of initial and terminal SaO2 values with the corresponding SpO2 data returned adequate correlation coefficients (R): R=0.95 for N-595, R=0.87 for CMS, R=0.84 for Dolphin, and R=0.82 for Masimo. Accordingly, Bland Altman-testing of SaO2 vs. SpO2 yielded standard deviations (SD) of ±1.4% for N-595, ± 2.2% for CMS, ± 2.8% for Dolphin, and 2.9% for Masimo. However, comparing the pulse oximeters to each other during the entire measuring period returned significantly decreased correlation coefficients (p<0.05) for SpO2 vs. SpO2 ranging from R=0.59 to R=0.77. Correlation of PR and HR produced low correlation coefficients (CMS. R=0.37, Masimo: R=0.41, N-595: R=0.42, Dolphin: R=0.45) including SDs of 20 to 28 bpm. Mainly caused by PR alarms, sensitivity (Masimo: 0.91, N-595: 0.47, CMS: 0.50, Dolphin: 0.32) and specificity (Masimo: 0.57, N-595: 0.80, CMS: 0.46, Dolphin: 0.56) appeared mostly low with all devices. During a total measuring time of 55 hrs, the dropout time was short for each device: 2.1% with CMS, 1.0% with Masimo, 3.3% with N-595, and 1.7% with Dolphin.

S147

Oral Presentations Postoperative I – 566-570 566 AN AIR-CAPSULE PROBE FOR THE DIRECT MEASUREMENT OF IAP IN PATIENTS AFTER ABDOMINAL SURGERY Schachtrupp A1, Biermann A1, Toens C2, Schwab R3, Becker H3, Schumpelick V1 Dpt. of Surgery, Technical University, Aachen, 2Surgical Clinic, Marien Hospital, Duesseldorf, 3 Surgical Clinic, Military Hospital, Koblenz, Germany

1

INTRODUCTION: Intravesical bladder pressure (IVP, the gold standard for the assessment of the intraabdominal pressure (IAP)is indirect, discontinuous, time and staff consuming and bears the risk of urinary tract infection. Aim of the underlying clinical study was to evaluate an aircapsule method (ACM) – established in the measurement of intracerebral pressure– for the direct measurement of IAP regarding agreement with IVP. METHODS: With approval of the local ethical committee and with informed consent 30 patients were studied. In these patients oesophageal, gastric, pancreatic, liver or colonic resection as well as other abdominal operations (e.g. incisional hernia repair) were performed. IVP was measured with 50 and 100 ccm measurement volume using a modified Kron-technique. Direct measurement of IAP was performed with an air-capsule mounted catheter (Probe 3, Spiegelberg, Hamburg) connected to an automatic, self-calibrating monitoring device (HDM 13.3, Spiegelberg). During the operation the air-capsule was placed on the greater omentum. Then the catheter was passed through the abdominal wall together with the routinely inserted drainages and was left in-situ for up to 5 days. Measurements of IVP and AKM were performed parallel. To assess agreement according to Bland and Altman mean difference (±SD) and limits of agreement (LA) were calculated (±1.96 SD). Additionally, paired t-test was performed. Finally, probe related morbidity or malfunctioning was recorded RESULTS: 168 measurements were performed. Mean ACM result was 8.1±0.9 mmHg while IVP with 50 ccm measurement volume resulted in 8.5±1.1 mmHg (ns). Mean difference (IVP50–LKM, ±SD) was –0.4±3.3 mmHg (LA: –6.7 to 6.2 mmHg). Mean IVP with 100 ccm measurement volume was 10.5±1.3 which was higher than ACM readings (p< 0,001). Mean difference (IVP100–LKM) was 2.2 mmHg±3,7 (LA: –5.2 to 9.6 mmHg). Mean insertion period was 4.4 days (min-max: 1-5 days). There was no malfunction or probe related morbidity.

CONCLUSION: Since cardiac arrhythmia considerably modifies the arterial pulse waveform, PR readings of “motion-resistant” pulse oximeters differ grossly from HR indication of the ECG. With this condition present, only N-595 meets the accuracy specifications for SpO2 of the manufacturer.

CONCLUSION: ACM and IVP agreed when 50 ccm were used as measurement volume which is comparable to previous reports. Direct Measurement of IAP via AM days does not increase morbidity and may be used in patients after abdominal surgery. In order to replace IVP however, further clinical studies first have to confirm its value in being related to organ dysfunction or the abdominal compartment syndrome.

REFERENCE: 1. Anaesthesia 1991, 46: 207-212. 2. Anesthesiology 2003; 95: A624.

Grant acknowledgement: Supported by the German Armed Forces

565

567

A NOVEL KNOWLEDGE BASED SYSTEM FOR AUTOMATED WEANING FROM MECHANICAL VENTILATION

VOLUME RESUSCITATION PRESERVES CARDIAC OUTPUT BUT CANNOT PREVENT ORGAN DAMAGE IN A MODEL OF ACS

Schaedler D1, Tonner P H1, Pulletz S1, Zick G1, Steinfath M1, Scholz J1, Weiler N1 Department of Anaesthesiology and Intensive Care Medicine, University Hospital Schleswig Holstein Campus Kiel, Kiel, Germany

Schachtrupp A1, Toens C2, Afify M1, Lawong G1, Schumpelick V1 Dpt. of Surgery, Technical University, Aachen, 2Surgical Clinic, Marien Hospital, Duesseldorf, Germany

INTRODUCTION: Knowledge based systems (KBS) may be of benefit in weaning from mechanical ventilation (1). However, limitations still exist. Most systems control only a limited number of parameters of ventilation. We developed a KBS for automated weaning from mechanical ventilation (MV), based on four aspects of MV (alveolar ventilation, oxygenation performance, spontaneous activity, invasiveness). The present investigation was designed to study a) the applicability and safety of the KBS and b) the reduction of invasiveness using a newly developed score system.

INTRODUCTION: Intraabdominal hypertension (IAH) can reduce organ perfusion and cardiac output. It has also been shown to lead to the abdominal compartment syndrome (ACS) and morphological organ impairment. Aims were firstly to evaluate whether additional crystalloid volume can preserve cardiac output (CO) in the presence of an intraabdominal pressure (IAD) of 30 mmHg lasting for 24h. Secondly, would this regimen prevent morphologic organ damage?

1

METHODS: After IRB approval and written informed consent eight haemodynamically stable and mechanically ventilated patients who were considered to be ready for weaning were enrolled in a prospective selected case study. The KBS was implemented on a PC which remotely controlled an EVITA 4 respirator (Dräger Medical, Lübeck, Germany). Ventilator settings chosen by the responsible physician (t0) were compared with the settings 10 minutes after start of the KBS (t1) and with the settings at the end of the session (t2) using one- way ANOVA for repeated measurements (p<0.05). The session was finished a) when extubation criteria were fulfilled, b) after 4 hours, or c) when a technical failure occurred. RESULTS: The mean time on the KBS was 165 ± 56 min. No unsafe ventilator settings and no failure of the KBS occurred. All patients were transferred from controlled ventilation to assisted spontaneous breathing in 37 ± 22 min. Initial ventilator settings (t0) and early ventilator parameters applied by the KBS (t1) did not differ significantly. However, during the session the KBS succeeded in reducing the invasiveness in all patients (Table1). At the end of the session four of the eight patients were successfully extubated. Comparison of the ventilator settings t0 50.6±14.3 11.9±1.9 20.1±3.6 6.1±2.4 8.1±2.46 8.3±1.2 3,9±0.4

FiO2 [%] fmech min-1 Pinsp [mbar] PEEP [mbar] MV [l*min-1] Vt/kg BW [ml*kg-1] Invasiveness Score

t1 41.6±10.5 12±2.7 16±4.6 7±2.8 7.3±1.47 6.6±1.2 3.9±0.6

t2 38.8±6.8* 0.8±0.3* 15.1±4.8* 7.3±3 8.58±3.2 7.2±2.8 3.1±1*

CONCLUSION: The KBS is able to find safe ventilator settings and to decrease the invasiveness of MV. Randomized controlled studies are necessary to prove the efficacy of the new KBS. REFERENCE: (1) Dojat et al. Am J Respir Crit Care Med (1996); 153(3): 997-1004 Grant acknowledgement: Supported by a grant from Dräger Medical (Lübeck, Germany)

1

METHODS: We investigated 12 domestic pigs with a mean bodyweight of 45.5±2.3 kg. Animals were ventilated volume controlled and were analgo-sedated continuously using ketamine and pentobarbitale. Via CO2 pneumoperitoneum, the IAD was increased to 30 mmHg in 6 animals while in another 6, the IAP remained unchanged (control). Investigation period was 24h. Additional to basic fluid replacement (2 ml/kg/h) crystalloid volume was given whenever continuously monitored CO was below control values. Systemic haemodynamic parameters (central venous pressure (CVP), heart rate (HR) and mean arterial pressure (MAP) as well as peak inspiratory pressure and urine output (UO) were recorded. Serum parameters e.g. ALT, lipase, AP, lactate, creatinine were investigated. In the end an overdose of pentorbarbitale was given and a histological examination of the liver, kidney, bowel, and lung was performed. Damage of bowel specimen was classified according to Park whereas damage to other organs was classified as none, low, intermediate and high. Statistic analysis was performed using ANOVA and post-hoc analysis together with paired tests. A p < 0.05 was considered significant and was adjusted (Bonferroni) when multiple paired analysis was performed. RESULTS: In the study group, CO was between 73.5±8.5 to 100.9±28.4 ml/kg/min and was not different to control values. Total amount of infused fluid was 10570±1928 vs. 3918±1042ml (control, p<0.01). CVP and PIP increased significantly while UO and serum parameters did not differ. In the study group, liver, kidney, bowel, and lung displayed intermediate histomorphological damage (p<0.01). CONCLUSION: Additional large volume administration can preserve CO in the presence of IAH over a 24h period. Under this condition, UO and serum parameters remain unchanged. In contrast to a former porcine study (1), some symptoms of the ACS were not present but histological organ damage occurred. In the clinical regard an intensified volume management may diminish major symptoms of ACS but an immediate decompression is recommended in presence of critical IAH. REFERENCE: Toens C, Schachtrupp A, Hoer J et al: A porcine model of the abdominal compartment syndrome. Shock 2002; 18: 316-321 Grant acknowledgement: Supported by the Deutsche Forschungsgemeinschaft

S148

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

568

570

A META-ANALYSIS OF HEMODYNAMIC OPTIMISATION: RELATIONSHIP WITH THE METHODOLOGICAL QUALITY

ORGAN IMPAIRMENT RESULTS AS EARLY AS 6H AFTER THE ONSET OF INTRAABDOMINAL HYPERTENSION

Poeze M1, Greve J M1, Ramsay G1 1 Department of Surgery, University Hospital Maastricht, Maastricht, Netherlands

Schachtrupp A1, Afify M1, Lawong G2, Schwab R3, Henzler D4, Schumpelick V2 1 Dpt. of Surgery, Technical University, 2Dpt. of Surgery, RWTH Aachen, Aachen, 3Surgical Clinic, Military Hospital, Koblenz, 4Dpt. of Anesthesiology, RWTH Aachen, Aachen, Germany

INTRODUCTION: The objective was to systematically review the effect of interventions aimed at the haemodynamic optimisation of both peri-operative patients and patients with sepsis and/or multiple organ failure and to relate this to the methodological quality of the individual trials. METHODS: A systematically computerised bibliographic search of published studies and citation review of relevant studies was performed including all randomized clinical trials, in which adult patients were included into a trial deliberately aiming at an optimized haemodynamic profile. An a priori division was made between the studies including peri-operative patients and studies on patients having organ failure and sepsis. A total of 28 studies were selected for independent review. Two reviewers abstracted data on population, intervention, outcome and methodological quality. Agreement between reviewers was high: differences were eventually resolved by third party decision.

INTRODUCTION: Trauma, vascular lesions or peritonitis can cause an increase of the intraabdominal pressure (IAP) or an abdominal compartment syndrome (ACS). Maintaining the cardiac output (CO) cannot prevent impairment of organ function(1,2). However, it is unknown, whether an IAP of 30 mmHg can lead to organ damage within 6 and 12h despite an unchanged CO.

RESULTS: Although the methodological quality of the studies was only moderately adequate, the outcome was not related to the quality of the randomised clinical trial for both peri-operative and septic patient populations. Aiming at an optimised haemodynamic profile resulted in a decreased mortality rate (relative risk ratio 0.45; 95% CI 0.33 - 0.56) in peri-operative patients. In contrast, in patients with sepsis and organ failure no difference in mortality rate was identified (relative risk 0.92; 95% CI 0.75 - 1.11).

METHODS: We examined 16 male pigs (50.6±3.8 kg). Animals were ventilated volume controlled and analgo-sedated using ketamine and pentobarbitale. Via CO2-pneumoperitoneum, IAD was increased to 30 mmHg for 6 and 12h (each n= 6). In the control (n=4) the IAP remained unchanged. Heartrate (HR), mean arterial pressure (MAP), central venous pressure (CVP), CO and peak inspiratory pressure (PIP) were continuously recorded. Additional to basic fluid replacement (2 ml/kg/h) crystalloid volume was given whenever continuously monitored CO was below control values. In the end, animals were given an overdose of pentobarbitale and specimen were taken from lung, heart, kidney, liver, stomach, pancreas (in vivo) as well as small and large bowel. Damage of bowel specimen was classified according to Park whereas damage to other organs was classified as none, low, intermediate and high. Statistic analysis was performed using ANOVA and post-hoc analysis together with paired tests. A p<0.05 was considered significant and was adjusted (Bonferroni) when multiple paired testing was performed.

CONCLUSION: Peri-operative interventions aimed at the haemodynamic optimisation of highrisk surgical patients reduce mortality. No benefit could be found in studies in which the therapy is used in patients with sepsis and/or organ failure.

RESULTS: A significant increase of CVP and PIP was found whereas the other haemodynamic data was not changed. Compared to control, an increased morphologic damage was found in lung, stomach (12h), liver, pancreas, bowel and kidney. The myocardium was unchanged. CONCLUSION: Even a short-termed increase of the IAP for 6h results in morphologic organ impairment despite an unchanged CO. In the clinical regard, these results imply an immediate decompression in the presence of a critically increased IAP. REFERENCES: 1. Harman PK, Kron IL, McLachlan HD et al: Elevated intra-abdominal pressure and renal function. Ann Surg 1982; 196: 594-597. 2. Bloomfield GL, Blocher CR, Fakhry IF et al: Elevated intra-abdominal pressure increases plasma renin activity and aldosterone levels. J Trauma 1997; 42: 997-1004 Grant acknowledgement: Supported by the German Armed Forces

569 PREVALENCE OF LONG-TERM POSTTRAUMATIC STRESS DISORDER RELATED SYMPTOMS AFTER SECONDARY PERITONITIS Mahler C W1, Unlu C1, Boer K R2, Lamme B1, Till v J W O1, Vroom M B3, Borgie de C A J2, Boermeester M A1 1 Surgery, 2Clinical Epidemiology and Biostatistics, 3Intensive Care Unit, Academic Medical Center, Amsterdam, Netherlands INTRODUCTION: Symptoms related to posttraumatic stress disorder (PTSD) may occur after exposure to a traumatic event and are clustered in re-experiencing, avoidance, and hyperarousal. To what extent peritonitis patients experience traumatic memories and long-term PTSD-related symptoms is unknown. The aim of this study was to screen patients for persisting long-term PTSD-related symptoms after surgical treatment for secondary peritonitis. In addition it is unclear whether there is a difference in prevalence of PTSD-related symptoms in patients with ICU admission and those without (hospital ward only).

Oral Presentations Nutritional support to prevent organ failures – 571-575 571 UTILIZATION OF PARENTERAL NUTRITION IN CANADIAN INTENSIVE CARE UNITS: OPPORTUNITIES FOR IMPROVEMENT? Drover J W1, Dhaliwal R2, Day A3, Jain M1, Keefe L4, Heyland D K2 Department of Surgery, 2Department of Medicine, 3Clinical Research Centre, 4Clinical Nutrition, Kingston General Hospital, Kingston, Canada 1

INTRODUCTION: Parenteral Nutrition (PN) is not without risks and recent guidelines have been developed to maximize the benefits and minimize the risks associated with PN. The purpose of this project is to compare current PN support practices in ICUs across Canada to the recommendations from the recently developed CPGs (1).

METHODS: An existing database of 278 patients who were surgically treated for secondary peritonitis between January 1994 and January 2000 revealed that by October 2003 131 patients were alive of which 125 patients could be tracked. A standardized validated questionnaire was mailed and this mailing was repeated twice. PTSD-related symptoms were measured by the PostTraumatic Stress Syndrome 10-Questions Inventory (PTSS-10) and the Impact of Events Scale Revised (IES-R). PTSS-10 scores >35 or IES-R scores >24 were regarded as fitting with PTSD symptoms.

METHODS: We conducted a prospective, observational study in 59 ICUs in Canada. In May of 2003, each ICU recorded nutrition support practices on a consecutive cohort of mechanically ventilated patients in their ICU that stayed for a minimum of 72 hours. The sites enrolled an average of 10.8 (range 4-18) patients for a total of 638 patients and 6819 patient days. Patients were observed for an average of 10.7 days (range 3-12). The median age was 65 (range 15-94), the most common admission diagnosis was medical (52.0%), and the median length of stay in ICU was 15.4 days.

RESULTS: The response rate was 81%; 101 questionnaires were suitable for analysis. Mean age of the responders (60 men) was 58.5 (± 14.4) years. Mean follow-up was 7.2 (4.0-10.4) years. Mean APACHE II score on admission and Mannheim Peritonitis Index score at operation were 9 (± 5) and 22 (± 8), respectively. Sixty percent of the patients had been admitted to the ICU for a mean of 19 (± 15) days. Overall PTSS-10 scores and IES-R scores were 25.3 ± 13.6 and 14.5 ± 15.3, respectively. According to the PTSS-10 questionnaire, PTSD-related symptoms had an overall prevalence of 23% and were found in 28% of former ICU patients and in 15% of non-ICU patients. The IES-R revealed PTSD-related symptoms in 28% of ICU patients and in 20% of non-ICU patients, an overall prevalence of 25% (Normal population 1.3-9.2%). In univariate analysis, age, ICU days and mechanical ventilation days were significantly related to the prevalence of PTSD symptoms. In multivariate analysis age was independently related to PTSD symptoms.

RESULTS: Of all the patients that were fed, 64/598 (11 %) were fed via PN only, 98/598 (16 %) were fed via combination EN (enteral nutrition) + PN and 436/598 (73 %) were fed via EN route only. In patients that were fed via the PN route only, PN was started 2.3 days (median) (range 0-11 days) after admission to ICU. In patients on combination EN + PN, PN was started 2.8 days (median) (range 0-12 days) while EN was started later i.e. 3.0 days (median) (range 0-10.5 days) after admission to ICU. PN was started before EN in 47 % of the patients on combination EN + PN. The average adequacy of calories from PN (defined as % calories received/calories prescribed) over the 12 days of observation was 69%, compared to adequacy from EN (60%) and to the adequacy from combination EN + PN (65%). Intravenous lipids were used in 88% (999/1138) of PN days and propofol was used in 11% (122/1138) of PN days. Supplemental enteral glutamine was used in only 1% (11/1138) of PN days. The median daily blood sugars in patients receiving PN was 8.6 (range 6.0-13.4).

CONCLUSION: Long-term follow-up of patients after admission for secondary peritonitis showed a high prevalence of symptoms fitting with PTSD (more than 23%). Univariate analysis showed a significant relation between age, days of mechanical ventilation, ICU days and the presence of PTSD-related symptoms as registered by the questionnaires. In peritonitis patients with or without ICU admission special attention should be given to early measurements of this presumably underestimated problem. More research is needed into ways to achieve an effective preventative approach.

CONCLUSION: Although utilization of PN (alone or EN in combination with PN) amongst Canadian ICUs is not excessive, PN is being initiated before delivery of EN is being optimized. Overfeeding by the parenteral route is not occurring. Utilization of intravenous lipids and supplemental glutamine and control of blood sugars seems to be sub-optimal. Significant opportunities to improve parenteral nutrition support practices in critically ill patients exist, when compared to the clinical practice guidelines. REFERENCE: (1) Heyland DK, Dhaliwal R, Drover JW, Gramlich L, Dodek P et al. Canadian clinical practice guidelines for nutrition support in mechanically ventilated, critically ill adult patients. J Parenter Enteral Nutr. 2003;27(5):355-73

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

S149

572

574

EVALUATION OF THREE DIFFERENT STRATEGIES FOR POST-PYLORIC PLACEMENT OF ENTERAL FEEDING TUBES

ELECTROPHYSIOLOGICAL PROOF OF CRITICAL ILLNESS MYOPATHY IN MECHANICALLY VENTILATED ICU PATIENTS

Samis A J1, Drover J W1, Heyland D K2 Department of Surgery, 2Department of Medicine, Kingston General Hospital, Kingston, Canada

1

INTRODUCTION: Enteral nutrition has many advantages for critically ill patients, but can be complicated by volume intolerance in those who are gastrically fed. Post pyloric placement of the distal end of the feeding tube is considered more advantageous because: 1) there is better success at reaching nutritional targets and they are reached sooner when feeding into duodenum versus gastric feeding, and 2) there is a decreased incidence of pneumonia. We evaluated three different strategies for post-pyloric placement of feeding tubes. In the first method, a standard feeding tube (Entriflex, Kendall Inc.) was placed blindly by two experienced clinicians (A.S., J.D.) using a standardized protocol. The second method involved tube placement assisted by an external magnet (MagnaFlow Tube, Arrow International Inc.). The third method was frictional placement, where small plastic tabs allow the tube to be captured by peristalsis and carried passively (Tiger Tube, Cook Inc). METHODS: We prospectively evaluated the success in placing 20 Entriflex tubes, 15 MagnaFlow tubes and 10 Tiger Tubes. Success was determined by a radiograph showing the tip of the tube beyond the pylorus. RESULTS: Thirteen of 20 Entriflex tubes (65%), ten of 15 MagnaFlow tubes (67%) and nine of 10 Tiger Tubes (90%) were successfully placed past the pylorus. The only Tiger Tube which did not traverse the pylorus was one in which initial gastric placement was not possible, such that all Tiger Tubes which were placed into the stomach were captured by peristalsis and carried beyond the pylorus. The final location of the distal tip of the tiger tube was farther beyond the pylorus than that of the Entriflex or MagnaFlow tubes, with a much greater proportion ending up in the jejunum (7 jejunal of 13 post pyloric entriflex tubes, 2 jejunal of 10 post-pyloric Magnaflow tubes, and 8 jejunal of 9 post pyloric Tiger Tubes). The average physician time was considerably less with the frictional feeding tube (<10 min for all tubes) versus the entriflex (25 min, range 6-45), the magnetically guided feeding tube (26 min, range 5-60 min). The only complication with any tube was minor nasal bleeding. CONCLUSION: This prospective study suggests that the frictional feeding tube achieves a high success rate in post-pyloric placement and may show promise as a simple, relatively non-invasive method of bedside feeding tube placement.

Weber-Carstens S1, Koch S1, Bercker S1, Heyn B1, Bubser F1, Behse F2, Kaisers U1, Deja M1 Anesthesiology and Intensive Care Medicine, 2Dept. of Neurology, Charité, Campus VirchowKlinikum, Berlin, Germany

1

INTRODUCTION: Results of histological studies suggest, that myopathy is a common cause of muscular fatigue and paresis in the critically ill patient. In electrophysiological testing direct muscle stimulation (DMS)in addition to conventional electromyography(EMG) and electroneurography (ENG) offers an electrophysiological method to determine critical illness myopathy (CIM) from critical illness polyneuropathy (CIP) in unconscious patients. The purpose of this clinical study was to determine the prevalence of electrophysiologically proven critical illness myopathy (CIM) in mechanically ventilated ICU patients receiving analgesics and sedatives and showing electrophysiological signs of CIP/CIM in conventional EMG/ENG. METHODS: In addition to conventional EMG/ENG direct muscle stimulation(DMS) was prospectively recorded in mechanically ventilated ICU patients receiving analgesics and sedatives with SAPS II > 20 on 3 successive days. When patients were considered to be satisfactory awakened according to the study of de Jonghe et al.(2) neuromuscular function was assessed using the medical research council (MRC) score. Additionally quantitative electromyography (QEMG) was performed when voluntary muscle contraction was possible to assess myopathy. RESULTS: Complete data could be sampled from 16 survivors, who met inclusion criteria (SAPS II 38±11, mean±SD). These patients revealed reduced compound muscle action potential and positive sharp waves diagnosing CIM or CIP in conventional EMG/ENG. 13 patients (81%) additionally showed a reduction of muscle fiber excitability during DMS indicating myopathy. In 11 patients (85%) with positive proof of DMS QEMG revealed myopathy during voluntary muscle contraction. In 2 patients with positive proof of DMS myopathy could not be reassured during QEMG. The 3 patients with normal muscle fiber excitability during DMS also did not show electrophysiological signs of myopathy during QEMG. In all patients clinical paresis was apparent, although 4 patients only showed mild paresis with MRC-Score of 4,4±0,1 (mean±SD), the other 12 patients showed moderate to severe paresis with mean MRC-Score of 1,6±0,7. CONCLUSION: (1) Using electrophysiological testing, in our study critical illness myopathy(CIM) but not critical illness polyneuropathy(CIP) could be recognized to be the predominant cause of muscular fatigue and paresis in the critically ill patient. (2) Electrophysiological investigation of muscle fiber excitability in mechanically ventilated ICU patients receiving analgesics and sedatives predicts occurrence of myopathy with a high sensivity and specifity. REFERENCE: (1) De Jonghe B, Sharshar T, Lefaucheur JP, et al. Paresis acquired in the intensive care unit: a prospective multicenter study. Jama 2002; 288:2859-67.

573

575

ENTERAL FEEDING IN THE CRITICALLY ILL PATIENT: CAN PROKINETIC AGENTS IMPROVE ABSORPTION?

IMPACT OF FLUID CHOICE ON RENAL FUNCTION IN CRITICALLY ILL PATIENTS WITH SHOCK

Pretorius R D1, Findlay G P1, Roberts R1, Weaver C A1 1 Critcal care services, University Hospital of Wales, Cardiff, United Kingdom INTRODUCTION: We assessed the efficacy of 2 well known prokinetic agents (Metoclopramide, Erythromycin) in a randomised controlled trial, in their ability to improve absorption of enteral feed in the critically ill patient using the paracetamol absorption test. METHODS: 32 Ventilated patients on ITU who were nasogastrically fed and not absorbing feeds as defined by aspirates of > 200 ml on 2 occasions in a 12 h period were entered into the study. Patients were randomised to Metoclopramide, Erythromycin or placebo and absorption assessed using the paracetamol absorption test at the start and at 48 hours as well as the measurement of nasogastric aspirates during this period. The results of the paracetamol test were expressed in terms of T max (time to max serum concentration) and Cmax (max serum concentration of paracetamol) and at the start as AUC (area under curve). The Cmax change and Tmax change (difference between 48h and start of study) was calculated from this. Gastric aspirates were expressed as % absorbed at 24 h and 48 h. RESULTS: The difference in % of gastric absorption measured in ml gastric aspirates at 48 h for all agents was measured using ANOVA (p= 0.77) and a paired t-test with Fisher’s extract for each prokinetic (Erythromycin, Metoclopramide p=0.79, Erythromycin, Placebo p=0.63, Metoclopramide, Placebo p=0.43). The % absorbed favoured the placebo (mean = 68.6 ml, St Dev =54.6). ANOVA for Tmax difference and Cmax difference for all agents were p=0.09 and p=0.15 respectively. Tmax change and Cmax change favoured Metoclopramide in both instances. A correlation was drawn between 48 h % feed absorbed and Cmax change, Tmax change respectively with significance values (p=0.56 and p=0.98).

Schortgen F1, Deye N2, Bastuji-Garin S3, Brochard L2 Medical ICU, Bichat-Claude Bernard Hospital, Paris, 2Medical ICU, 3Public health, Henri Mondor Hospital, Créteil, France

1

INTRODUCTION: Fluid resuscitation is the most important treatment to restore renal perfusion and prevent acute renal failure (ARF) in hypovolemic patients. Because colloid-related ARF has been described, the best fluid to use remains a matter of debate. METHODS: This prospective cohort study included patients admitted in 114 ICUs during 4 weeks in 2002 and needing fluid loading. Severity scores, indication for fluids, risk factors for ARF, fluids administered, and occurrence of ARF were prospectively recorded. ARF was defined by either a two-fold increase in the serum creatinine level over the value at study inclusion or the need for dialysis. Several models of logistic regression were tested to determine whether the use of colloids might be associated with a higher incidence of ARF. RESULTS: 687 patients with shock were included. The most frequent indications for fluids were severe sepsis (52%), hemorrhagic shock (23%) and trauma (13%). 119 patients (17%) received crystalloids alone and 83% colloids, alone (16%) or in association with crystalloids (67%). Starches were used in 47% of patients, gelatins in 38%, plasma in 28%, albumin in 15% and dextrans in 3%. 142 patients developed ARF, 61% of them needed dialysis. Variables significantly associated with ARF are indicated in the table. The use of colloids was an independent risk factor for ARF. The type of colloid was entered in the same model for multivariate analysis; plasma [OR:1.93 (CI 95%:1.20-3.11), p=0.007] albumin [1.74 (1.01-3.00), 0.05], and starches [1.55 (1.02-2.37), 0.04] remained independently associated with ARF. The use of gelatins was not associated with ARF, and the number of patients receiving dextrans was too low to be analysed.

CONCLUSION: The % of feed as measured in ml of aspirate that was absorbed after 48 h of prokinetic use favoured the use of placebo, but with no statistical significance between any prokinetic. The paracetamol absorption test showed that there was no statistical significant increase of Cmax (max concentration of serum paracetamol levels) or reduction in Tmax (time for paracetamol to reach max concentration) after 48h use of prokinetics. There was no correlation between changes in paracetamol levels and the % of feed absorbed. We conclude that this study showed no benefit for the use of prokinetics in critically ill patients that fail to absorb nasogastric feed and that there is no correlation between the results of the paracetamol absorption test and residual gastric aspirates after 48 h REFERENCE: 1. Sturm A, Prokinetics in Patients with Gastroparesis; A Systemic Analysis. Digestion 1999;60:422-27

CONCLUSION: The results of this observational study suggest that the use of crystalloids or gelatins might be associated with a lower incidence of ARF than the use of starches, albumin and plasma in patients suffering from shock. Grant acknowledgement: Assistance Publique-Hôpitaux de Paris (PHRC 2000)

S150

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

Oral Presentations Meeting the needs of the patient – 576-580 576 DESIGNING OF AN ICU-ROOM BY ART-THERAPY: EVALUATION BY PATIENTS AND MEDICAL PROFESSIONALS Sünderhauf M1, Lippock D1, Kaiser H2, Richter P1, Ragaller M2 1 Work- Organisation and Social-Psychology, Technical University, 2Anaesthesiology and Intensive Care Medicine, University Hospital Dresden, Dresden, Germany INTRODUCTION: The design of patient rooms in the ICU has to follow specific requirements: it has to respect the demands of the critical ill patients as well as to create a functional and comfortable environment for the medical staff. Due to the hypothesis that colours and pictures have positive impact on human emotions and health conditions (1) an art-therapist created a 2 bed room of an ICU (light blue ceiling, a chain of coloured pictures on the walls and the ceiling). The effects of this new created room on patients and medical professionals were evaluated in a standardized interview. METHODS: After approval of the local ethic-committee and of informed consent from the patients and the volunteers a standardized interview was performed. The new created room was compared with a conventional designed room relating to wellbeing and pleasure. Factors like age, gender, personal expertise and actual health conditions were included into the assessment. A number of 40 patients (20 pro room) and 20 medical professionals (ICU-nurses and intensivists) were interviewed. The interview included 12 items by using a stepwise “Likert-scale” (6 steps) (polarity profile). RESULTS: The evaluation of the two rooms by the patients showed no general difference. Due to the light situation the medical staff preferred the conventional styled room. Compared to the medical staff the patients preferred the new created room in five out of 12 items. The actual emotional conditions of the patients had a significant influence on the judgment of the room. Individuals with a positive mood preferred the new room whereas individuals in a negative mood preferred the conventional room. The intensity of the light, especially of natural day light had significant influence on the assessment of the patients as well as of the medical staff.

578 HOW DO PATIENTS LOOK BACK ON INTENSIVE CARE? Orsini M1, Van Dijk H1, Breederveld P1, Couprie I1, Van der Hoeven M1, Dawson L1, Salm E1, Meynaar I1 1 Intensive Care Unit, Reinier de Graaf Hospital, Delft, Netherlands INTRODUCTION: Intensive care treatment can be a life-saving and terrifying experience at the same time. The object of this study was to evaluate how survivors of intensive care remember intensive care treatment. METHODS: The study was done in our 10 bed closed format mixed surgical and medical ICU from January 1st, 2002 until December 31st, 2003. All patients who were discharged from the ICU to the ward and who had been treated in the ICU for at least 3 days, or who had been on a ventilator, or who had experienced cardiopulmonary resuscitation were entered in the study. ICU nurses visited study patients on the ward between the second and fifth day after ICU discharge and used a questionnaire to evaluate the patients’ ICU experience. RESULTS: Out of 156 patients who fulfilled the study criteria, 125 were entered in the study. 23 patients were not entered because of high workload in the ICU, 8 patients were not entered because they were either too confused or were re-admitted to the ICU. All patients Memory of feeling safe in the ICU Memory of noise disturbance Dissatisfied with ICU stay Patients who were on a ventilator No memory of being on a ventilator Memory of anxiety or nightmares Memory of not being understood Memory tube discomfort Memory of choking

125 63 15 6 65 37 21 14 7 6

100% 50% 12% 5% 100% 57% 32% 22% 11% 9%

CONCLUSION: In the first week after discharge from the ICU the majority of patients have no memory of time spent on a ventilator in intensive care. Patients who do remember, often have memories of discomfort or anxiety.

CONCLUSION: The perspectives of patients and medical professionals relating to a room design with art elements seems to be different. Lack of natural daylight may offset the postulated positive effects of art therapeutic elements in patients. To improve the effects of art therapeutic elements and to improve the wellbeing of patients and medical professionals natural daylight in the ICU seems to be a substantial precondition. REFERENCE: 1. Rodeck B. , Meerwein G., Mahnke F. Mensch – Farbe- Raum. (1Aufl.)Verlagsanstalt Alexander Koch, Leinfelden Echterdingen, 1998

577

579

MEETING THE NEEDS OF INTENSIVE CARE UNIT PATIENT FAMILIES

MEMORIES AND PERCEPTIONS IN ICU SURVIVORS: A MULTIDIMENSIONAL QUESTIONNAIRE

Teixeira C1, Teixeira C1, Silva I2, Keating J2, Veloso A2, Moura F1, Carneiro A1 1 Intensive Care Unit, Hospital Geral de Santo António, Porto, 2Departamento de Psicologia, Universidade do Minho, Braga, Portugal INTRODUCTION: The identification of critically ill patients’ family needs has been the focus of several studies. Those studies have defined that family needs are varied and cover different domains (Leske,1991). Many of these investigations used the Critical Care Family Needs Inventory (CCFNI) of Molter (1979) and Leske (1991). Leske’s classification of needs, in turn, has been transformed in measurements of the extent to which these needs are met by Intensive Care Units. The present study, based on an adapted version of CCFNI by Johnson(1998), has two objectives: to report on the process and results of the assessment of family satisfaction in one ICU (the UCIP of HGSA)and to assess the quality of the questionnaire used both in terms of the needs assessed and its reliability. METHODS: Data collection spread over a period of three months. Forty nine relatives of ICU patients, one for each patient, were subjected to the questionnaire. The questionnaires were administered between 48 h and 168 h following admission. RESULTS: The questionnaire items related to the communication strategy adopted by ICU staff and the confidence in the quality of healthcare provided, were those who received higher ratings. Those items related to the direct interaction between the family and the ICU environment, namely understanding the equipment used, were those rated lower. The Cronbach’s alpha coefficient for the questionnaire is 0.72. The facial validity of the questionnaire is discussed using the results of several focus groups with relatives. CONCLUSION: This study is a part of a larger project about improvement in the quality of health care services. In this context, the utility of this instrument, or a revised version, as an adjunct in assessing the quality of critical care services provided is discussed. REFERENCES: 1. Johnson, D.and al(1998). Measuring the ability to meet family needs in an intensive care unit. Critical Care Medicine, 26(2): 266-271. 2. Leske. J. S. (1991). Internal psychometric properties of the Critical Care Family Needs Inventory. Heart & Lung, 20(3): 236244. 3. Molter, N. C. (1979). Needs of relatives of critically ill patients: A descriptive study. Heart & Lung, 8(2): 332-339.

Hernandez G1, De la Fuente R1, Romero C1, Naranjo M E1, Zanolli M1, Barticevic N1, Castillo L1, Bugedo G1 1 Anesthesia and Critical Care, Universidad Catolica, Santiago, Chile INTRODUCTION: Few studies have examined patients’ memories of the ICU after hospital discharge (1-3). Some patients have little recall but others describe very unpleasant experiences in the emotional and communication areas and/or in factual or delusional memories. It is unknown if this experience can induce changes in perceptions of life value, religious faith or acceptance of future ICU treatments. This has not been addressed in previous studies. METHODS: We applied a multidimensional questionnaire addressing memories and perceptions in a face to face interview to 44 ICU patients (age 52±12 y/o.; APACHE II 17±3; mechanical ventilation 10.6±8 d; septic shock 43%, ARDS 29%; ICU LOS 17.5±14 days) three months after ICU discharge. Also, ICU stress level was assessed by a 10 cm VAS. RESULTS: Thirst (76%), sleep deprivation (66%) and isolated pain (52%) predominated in factual memories, while 60% recalled paranoid delusions. Only 3 patients had none memories. At emotional level, 28% expressed fear of getting permanently disabled and 23% of death. Anxiety was reported by 47% and feeling lonely by 33%. In relation to communications, 85% believed staff to be lying about prognosis, 57% felt unable to communicate effectively with staff or family, 90% received no explanation about monitor alarms and 72% did not recall any staff or nurses name. ICU stress level assessed by VAS was 7.0±3.7. Nevertheless, 80% would accept a new unrestricted ICU admission, 85% felt equal or more optimistic about life than before and 72% maintained or increased religious faith. CONCLUSION: Although most of ICU survivors report unpleasant memories, especially at the communication level and consider the ICU experience as very stressful, they would accept new ICU treatments and express no negative changes in life or religious perceptions. REFERENCES: 1. Jones C, et al. Crit Care Med 2001;29:573-580. 2. Capuzzo M, et al. Crit Care 2004; 8: R48-R55. 3. Rotondi AJ, et al. Crit Care Med 2002;30:746-752

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

S151

580

582

POST-TRAUMATIC STRESS DISORDER AFTER MEDICAL INTENSIVE CARE

HYPOXIA ACCELERATES ENDOTOXIN-INDUCED BIOENERGETIC FAILURE

Wehler M1, Riek A2, Stolle M2, Loew T2, Reulbach U3, Hahn E G1, Strauss R1 1 Medicine I, 2Psychosomatic Medicine, 3Medical informatics, University of Erlangen, Erlangen, Germany

1

INTRODUCTION: Few studies have analyzed the prevalence of post-traumatic stress disorder (PTSD) in survivors of medical intensive care. METHODS: From June 1997 to December 1999 all adult admissions to our medical intensive care unit (ICU) who were treated for at least 24 hrs were eligble. On admission pre-ICU functional status and subjective well-being were assessed by interview (1). Six months after admission survivors memory of the ICU stay was assessed (none, positive, negative). At 18 month follow-up a standardized interview at the patients home was performed using the PTSD-10 Questions Inventory (PTSD-10) (2), the 90-item Revised Symptom Checklist (SCL-90-R), the Hamilton Anxiety (HAMA) and Depression (HAMD) Scale, the 57-item Giessen Subjective Complaints List and a 28-item quality of life (QOL) scale. RESULTS: 444 patients were enrolled. Cumulative mortality rates were 23% in the ICU, 33% in the hospital, 42% at 6 month and 53% at 18 month follow-up. From the 209 survivors, 22% were lost-to-follow up, 27% were unable to be interviewed due to physical or cognitive reasons and 13% declined the interview. The remaining 80 study patients had a mean age of 46±12 (±SD) yr; 69% were male, mean ICU length of stay was 12±17 days, mean APACHE II score after 24 hrs was 19±9 and mean SOFA total maximum score was 6.3±4.7. According to PTSD-10 criteria 10 patients (12.5%) had a diagnosis of PTSD. PTSD was more frequently diagnosed in patients who had reported poor pre-ICU subjective well-being compared to patients with good subjective wellbeing (8/41 vs 2/39 pts; 20% vs 5%; p=0.05), in patients with multiple organ dysfunction (MOD) compared to patients without MOD (8/38 vs 2/42 pts; 21% vs 5%; p=0.03), and in patients who had negative or no memories of their ICU stay compared to patients with positive memories (7/30 vs 3/50 pts; 23% vs 6%; p=0.02). Patients with PTSD had significantly (p<0.0001) worse scores on the SCL-90-R global index of psychopathology, showed a significantly (p<0.0001) higher degree of somatic and psychic anxiety, major depression, bodily complaints and mental exhaustion, and reported poorer self-perceived QOL. CONCLUSION: A small subgroup (12.5%) of our medical ICU survivors developed PTSD. Subjective well-being before ICU admission, MOD, and ICU memories were associated with PTSD and related psychopathologic symptomatology. These criteria could be used to identify survivors at risk for developing PTSD. REFERENCES: 1. Fernandez RR et al. Intens Care Med 1996, 22:1034. 2. Stoll C et al. Intens Care Med 1999, 25:697

Frost M T1, Karyampudi S1, Singer M1 Wolfson Institute for Biomedical Research, University College London, London, United Kingdom INTRODUCTION: Nitric oxide (NO) inhibits mitochondrial respiration, leading to decreased production of ATP. The effect of endogenously generated NO in a low PO2 environment is not known but may be clinically pertinent as tissue hypoxia is a common accompaniment of sepsis and may amplify the inflammatory response. METHODS: J774 macrophages were incubated with/without endotoxin/interferon gamma; in either 21% (normoxia) or 1% (hypoxia) O2 over a 24 hr period. Measurement of O2 consumption (Clark electrode), nitrite, NO and peroxynitrite (DAF-FM and DHR-123, respectively), and ATP (luciferase assay) was made at predefined timepoints. RESULTS: Cell viability was >90% in all conditions. Hypoxia alone decreased VO2 and ATP by 35-50% after 24 hrs (p<0.05, ANOVA, Tables 1+2). A more rapid fall in VO2 was seen in hypoxic compared to normoxic activated cells (p<0.05), although NO, ONOO- and nitrite levels were 6080% lower than in normoxia. After 12 hrs’ activation, ATP increased in both O2 environments compared to controls, even though mitochondrial respiration had fallen by 50-70%. By 24 hrs, ATP levels had fallen by 60-90%. O2 consumption (n = 6) (umol/min) Normoxia + LPS Hypxoxia Hypoxia + LPS

(n = 6) (umol) Normoxia + LPS Hypoxia Hypoxia + LPS 0.00*

12 hr 6.2 ± 1.6 9.7 ± 1.3 3.6 ± 0.7 *

24 hr 2.9 ± 1.5 8.6 ± 1.0 2.3 ± 0.7*

0 hrs 0.15 ± 0.01 0.14 ± 0.09 0.13 ± 0.02

12 hrs 0.18 ± 0.01^ 0.13 ± 0.01^ 0.21 ± 0.02*

24 hrs 0.03 ± 0.01 0.06 ± 0.00 0.02 ±

* p<0.05 vs. hypoxia at same time point, ^ p<0.05 vs. 0 hrs time point CONCLUSION: Though less NO is produced by activated macrophages in a hypoxic environment, its effects on inhibiting mitochondrial respiration are greater. The initial rise in cellular ATP, despite a large fall in mitochondrial respiration, implies up-regulation of glycolytic activity and/or reduction of energy-utilising cellular processes.

583

SYSTEMIC AND TISSULAR PRODUCTION OF OXIDATIVE STRESS IN AN ANIMAL MODEL OF SEPTIC SHOCK

1

MITOCHONDRIAL DYSFUNCTION IN CRITICALLY ILL PATIENTS WITH MULTIPLE SYSTEM ORGAN FAILURE

Klouche K1, Delbosc S1, Abdelrahman M2, Cristol J P1, Thiemermann C3 1 Intensive Care Unit, Lapeyronie University Hospital, Montpellier, France, 2Department of experimental Medicine, 3Department of Experimental Medicine, William Harvey Institute, London, United Kingdom INTRODUCTION: During sepsis, the activation of leucocytes and the overproduction of reactive oxygen species (ROS) may lead to tissue damage and organ dysfunction. While systemic overproduction of ROS has been studied by several authors, tissular oxidative stress production has not been extensively assessed. This study was therefore undertaken to establish, in a rodent animal model of septic shock, an overproduction of ROS in the plasma and in different sites especially myocardium, aorta, liver and the kidney and to specify its time course. METHODS: Twenty Sprague-Dawley rats weighing between 200-250g have been tracheotomized after ip pentothal. A dose of 10 mg/kg of LPS Pseudomonas aeruginosa was administered through the internal jugular vein to 15 animals who developed a septic shock. Four groups (5 rats each) have been individualized: groupe 1: non LPS sacrificed after the surgical procedure, and groups 2,3,4 : LPS + sacrificed respectively at 90 min, 180 min and 360 min after LPS injection. The measurements of the production in the plasma: of thiobarbituric acid reactive substances (Tbars nmol/ml) by fluorometry ; and the production in tissus (crush of myocardium, aorta, liver and kidney) of superoxide anion (SA mv/mg tissu) by chemiluminescence technique have been investigated at different times. RESULTS: Results are represented in Table 1. group 2 LPS 90 0.71±0.02* 0.31±0.02* 0.91±0.02* 0.34±0.02 0.25±0.02

6 hr 10.6 ± 1.1 11.3 ± 0.9 6.3 ± 1.4*

* p <0.05 compared to hypoxia at same time point ATP

Oral Presentations Oxidative stress – 581-583 581

group 1 control Plasma Tbars 054±0.03 S A heart 0.21±0.01 aorta 0.78±0.06 liver 0.30±0.01 kidney 0.20±0.02 *: p<0.05 vs group 1 control

0 hr 13.3 ± 1.9 13.2 ± 1.6 12.2± 0.5

group 3 LPS 180 0.88±0.11* 0.36±0.02* 0.96±0.04* 0.44±0.01* 0.28±0.01

group 4 LPS 360 0.82±0.08* 0.28±0.01* 0.89±0.06* 0.40±0.01* 0.29±0.02

CONCLUSION: An overproduction of oxidative stress in the plasma and in several organs: heart, aorta, liver and kidney, was observed in this rodent animal model of septic shock. This production was consistent at 90 minute in myocardial and vascular sites, and at the third hour in the liver while it was not significantly increased in the kidney until the sixth hour. The tissular oxidative stress production preceded organ dysfunction especially myocardial dysfunction (sixth hour in this model). An early inhibition of oxidative stress production may therefore prevent tissular damage and myocardial dysfunction during septic shock.

Gosling P1, Cooper D2, Chinn K F2, Bowen D2, Waring R H2, Manji M3 Clinical Biochemistry, University Hospital Birmingham NHS Trust, 2School of Biosciences, University of Birmingham, 3Critical Care, University Hospital Birmingham NHS Trust, Birmingham, United Kingdom INTRODUCTION: In critically ill patients with multi-system organ failure (MSOF) secondary to sepsis, ATP generated by mitochondrial oxidative phosphorylation (OP) is compromised (1). To study the effects of sepsis on mitochondrial function, platelet mitochondria from MSOF patients and healthy controls were used. An in vitro laboratory model using human hepatic cells (G2) in the presence of lipopolysaccharide (LPS) was also developed. In both these models OP efficiency was assessed as complex I and IV activities. METHODS: Platelet mitochondria in EDTA blood were harvested from MSOF patients (n=12) and healthy controls (n=12). Hep G2 (human hepatic) cells were incubated with 1 microgram/mL LPS. Homogenised cells were centrifuged to harvest functional mitochondria. Complex I activity was measured by incubation of ubiquinone and sodium azide before addition of NADH. The decrease in absorbance at 340nm due to the NADH/NAD conversion was measured. Assay specificity was confirmed using rotenone as inhibitor of Complex I. Complex IV activity was measured by incubation of reduced cytochrome C with the mitochondrial sample in platelet buffer. Oxidation of cytochrome C was assessed by the decrease in absorbance at 550nm. Assay specificity was confirmed using KCN as an inhibitor of Complex IV. RESULTS: There were no significant differences in Complex I activity between MSOF patients and controls (144.56 ± 20.07 v. 132.18 ± 17.2 nmol NADH oxidised/min/mg protein respectively). However, MOSF patients had significantly lower Complex IV activities than controls (7.35 ± 6.43 v. 58.07 ± 22.71 nmoles cytc/min/mg protein; p < 0.01). Treatment of Hep G2 cells with LPS for 48 hours, reduced complex I and complex IV activities to 12.0% and 20.1% of control values respectively. Following incubation for a further 24 hours in additive-free medium, complex I and complex IV activities were 75.0% and 5.8% of control values respectively. CONCLUSION: Hep G2 Complex I inhibition following 48h exposure to LPS largely recovered after 24 hours incubation in fresh medium. In contrast inhibition of Complex IV activity was irreversible. Although previous studies suggest complex I activity is impaired in septic patients,(2) only complex IV activity was reduced in platelet mitochondria from MOSF patients in this study. These results suggest that in MOSF, complex IV activity is compromised not Complex I. REFERENCES: 1. Szabo G et al Liver in Sepsis and systemic inflammatory response syndrome. Clinics in Liver Disease, 2002;6:1045-66. 2. Brealey D et al Association between mitochondrial dysfunction and severity and outcome of septic shock. Lancet 2002;360:219-23. Grant acknowledgement: Partly funded by West Midlands Intensive Care Group.

S152

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

Oral Presentations NO-dependent pathways in sepsis – 584-586 584 ENDOTOXIN PRODUCES AN INOS-DEPENDENT INHIBITION OF COMPLEX I VIA S-NITROSYLATION AND NITRATION Frost M T1, Wang Q1, Moncada S1, Singer M1 Wolfson Institute for Biomedical Research, University College London, London, United Kingdom

1

INTRODUCTION: Excess nitric oxide (NO.)is implicated in the development of sepsis-induced MOF, with a putative mechanism involving direct mitochondrial inhibition, predominantly affecting complex I. NO effects on complex I may be mediated via S-nitrosylation and/or persistent inhibition by nitration. METHODS: We studied the temporal contribution of these chemical modifications on the inhibition of respiration using endogenous NO generated by endotoxin/interferon gamma in J774 macrophages over a 24 hr period. RESULTS: Oxygen consumption and complex I activity fell progressively over time (in activated macrophages (p<0.001, ANOVA) [Table] though cell viability remained in excess of 95% at 24 h. No changes were seen in control cells. Co-incubation with the NO synthase (NOS) inhibitor LNIO resulted in 70-90% reversal of these effects, and a 84% decrease in nitrite formation (Griess reaction), confirming involvement of the NOS pathway. Initial reversal achieved by addition of glutathione indicated that the early inhibition of complex I is related to S-nitrosylation. This reversibility diminished over time (p<0.05), coinciding with an 81% increase in peroxynitrite formation and a 170% increase in mitochondrial tyrosine nitration over 24 hr (measured by GCmass spectrometry). 21% O2 + LPS (n=6) O2 consumption (umol/min) Complex1activity (nmol min/mg pr) Complex1activity (nmol min/mg pr)

0 hrs 13.3 ± 1.9 8.1 ± 0.4 8.0 ± 0.6

6 hrs 10.6 ± 1.1 8.1 ± 0.3 7.8 ± 1.3

12 hrs 6.2 ± 1.6 4.0 ± 0.4 6.4 ± 0.8

24 hrs 2.9 ± 1 1.9 ± 0.6 2.9 ± 1.2

586 RESEVERATROL REDUCES UPREGULATED ARGININE TRANSPORT IN SEPTIC SHEEP, BUT NOT HUMAN, PBMCS Reade M C1, Clark M F2, Boyd C A R2, Young J D1 Nuffield Department of Anaesthetics, 2Department of Human Anatomy and Genetics, University of Oxford, Oxford, United Kingdom

1

INTRODUCTION: Arginine transport is rate limiting for NO production, and so may be important in the pathogenesis of sepsis. We recently reported increased arginine transporter (y+ ) activity in PBMCs from endotoxic sheep, and a similar increase in y+ activity after in vitro exposure to endotoxin 1. We also found increased y+ transport in PBMCs from patients with septic shock 2. In sheep PBMCs exposed to endotoxin, cyclo-oxygenase inhibition with ibuprofen prevented the increase in y+ transport 1. Resveratrol reduces NO production and selectively inhibits COX-1. We hypothesized that resveratrol might reduce y+ transport in sheep endotoxaemia and human sepsis. METHODS: Resveratrol or diluent were added to sheep blood, and after incubation with endotoxin or saline for 18 hours, PBMCs were isolated. Arginine transport was determined by incubating PBMCs with 3H-labelled lysine for 3 minutes, then collecting the cells by centrifugation through oil. Lysine and arginine have indistinguishable transport kinetics. Radiolabelled lysine influx was measured in the presence of leucine, which quantified the contribution of y+ to total transport. (The other arginine transport system, y+L, is almost completely inhibited by leucine.) Blood was also obtained from patients meeting the SCCM definition of septic shock and from healthy controls, and immediately exposed to resveratrol or diluent. After PBMC separation, the activity of the y+ transporter was similarly assessed. RESULTS: Endotoxin produced the expected increase in sheep PBMC y+ activity. Co-incubation with resveratrol prevented this increase. Resveratrol added at the end of endotoxin incubation also returned y+ activity to control levels. In contrast, while septic human PBMCs had raised y+ activity, resveratrol did not return y+ function to that of non-septic controls. Control 1.0

y+ activity in sheep PBMCs ETx ETx+Res ETx+Res at end 1.7 ± 0.12 1.0 ± 0.09 1.0 ± 0.08 y+ activity in human PBMCs

y+ activity in human PBMCs Control Septic Septic+Res 1.0 ± 0.5 4.5 ± 3.3 5.0 ± 5.7 mean ± SEM

CONCLUSION: We describe a predominantly NOS-dependent inhibition of mitochondrial respiration in this endotoxin/cytokine-activated macrophage cell line, acting through inhibition of complex I. The mechanism of action appears related to an initial, reversible S-nitrosylation, followed by progressively persistent inhibition by nitration. This carries potential significance to the temporal ability to reverse mitochondrial inhibition during in vivo sepsis.

CONCLUSION: The difference between sheep and human PBMC responses to resveratrol may reflect species variation or differences between clinical sepsis and in vitro endotoxin exposure. Nonetheless, it appears resveratrol lacks promise as an inhibitor of the upregulated (y+) component of arginine transport in human sepsis.

585

Oral Presentations Improving outcomes in bacteremia – 587-589 587

IN VIVO NITRIC OXIDE PRODUCTION IS ELEVATED BOTH IN MODERATE AND SEVERE CLINICAL INFLAMMATION Poeze M1, Luiking Y C1, Steens L2, Ramsay G1, Deutz N E P1 1 Department of Surgery, 2Department of Intensive Care Medicine, University Hospital Maastricht, Maastricht, Netherlands INTRODUCTION: L-Arginine (L-Arg) is the precursor of nitric oxide (NO) synthesis. L-Arg is produced in the body (mainly kidney) by de novo production from Citrulline (Cit) and by protein breakdown. Although elevated levels of the NO metabolites nitrate/nitrite have been described in sepsis, this measurement may not represent true NO production. This study was aimed at comparing Arg and NO metabolism in septic patients versus non-septic controls in the ICU by using stable isotope techniques. METHODS: 13 ICU patients were studied: 8 with severe sepsis/septic shock (sepsis) within 48h of diagnosis and 5 non-septic controls (2 recovered from pulmonary failure, 2 exacerbations COPD, 1 neurotrauma; non-sepsis). Arg and NO metabolism were examined using primedcontinuous (2h) infusion of L-[guanidino-15N2-2H2]Arg and L-[ureido-13C]Cit and subsequent measurements of arterial amino acid concentrations and tracer-tracee ratios using LC-MS. NO production was measured as Arg to Cit conversion; de novo Arg production as Cit to Arg conversion. CRP levels were used to indicate the degree of inflammation. Statistical analysis by 2way Anova; data are means ± SEM RESULTS: CRP levels indicated that non-septic controls had moderate inflammation (CRP: 64 ± 7 mg/L) and septic patients had severe inflammation (CRP: 242 ± 35 mg/L). Plasma Arg levels were lower during sepsis (47.7 ± 4.4 vs. 68.2 ± 5.1 ?M in non-sepsis; P<0.01), but whole body Arg and NO production were not different between both groups (NO: 2.1 ± 0.4 and 2.0 ± 0.5 ?mol/kg BW/h in sepsis and non-sepsis, respectively (NS)). In contrast, de novo Arg production was lower during sepsis (5.8 ± 2.9 and 17.0 ± 3.0 ?mol/kg BW/h in sepsis and non-sepsis, respectively (P<0.05)). Compared with healthy subjects, de novo Arg production is increased in our non-sepsis controls, while it is decreased in sepsis. NO production is 2-fold higher in both ICU-patient groups. CONCLUSION: Lowered plasma Arg levels in severe inflammation do not coincide with increased de novo Arg production, higher NO production or altered total Arg production. This unresponsiveness to increase endogenous Arg production in severe inflammation compared with moderate inflammation results in an Arg deficiency state, and may limit a further increase of NO production.

REFERENCES: 1. Clark MF, Reade MC, Young JD, Boyd CAR. Effects of endotoxin exposure on cationic amino acid transporter function in ovine peripheral blood mononuclear cells. Exp Physiol 2002; 88: 201-208. 2. Reade MC, Clark MF, Young JD, Boyd CAR. Increased cationic amino acid flux through a newly expressed transporter in cells overproducing nitric oxide from patients with septic shock. Clin Sci (Lond) 2002;102:645-50.

SOURCE CONTROL AND BLOOD STREAM INFECTIONS FROM ABDOMINAL ORIGIN De Waele J J1, Blot S1, Hoste E1, Decruyenaere J1, Colardyn F1 1 Intensive Care Unit, Ghent University Hospital, Ghent, Belgium INTRODUCTION: Inadequate source control of an intraabdominal infection after abdominal surgery may lead to blood stream infections (BSI) in the postoperative period, but establishing the diagnosis of inadequate source control and thus the need for surgical intervention, is often difficult. We studied characteristics of BSI from abdominal origin after previous abdominal surgery, and tried to identify factors that are associated with inadequate source control at the moment of BSI. METHODS: We analyzed all patients with BSI from abdominal origin in an 11-year period (1992-2002) in the intensive care unit of the Ghent University Hospital. Patients who developed BSI after a abdominal surgical operation were identified. Demographics, disease severity as determined by the Acute Physiology and Chronic Health Evaluation (APACHE) II score, the incidence of organ failure, and outcome were recorded. In all patients, the need for surgical intervention, the interval between the onset of BSI and the surgical intervention, and the presence of source control at the moment of the surgical intervention preceding the BSI, and at the moment of BSI itself, were retrieved from the patients file. “Source control” was defined as the elimination of an infectious focus, and control of ongoing contamination. Mortality was defined as in hospital mortality. Mortality was defined as in hospital mortality. RESULTS: In 74 patients, 80 episodes of BSI occurred after abdominal surgery. The mean age of the patients was 57 years (±14.3). Mean APACHE II score on admission was 22.8 (±9.9). The median interval between surgical intervention and BSI was 8 days (IQR 3 - 19). In 64 out of 74 patients, the source of the intraabdominal infection was considered “controlled” at the moment of the last surgical intervention, but when the BSI occurred, the intraabdominal source appeared “uncontrolled” in 38 of these patients (51%). Compared to patients with controlled intraabdominal infections, the patients with uncontrolled intraabdominal infections had a higher number of surgical interventions before BSI (median 2 (IQR 1-3 vs. 1 (IQR 1-2) p=0.009), a longer hospital stay before BSI (median 25 (IQR 18-36 vs. 11 (IQR 4-26) p=0.021), and a longer interval between surgery and the moment of BSI (median 10 (IQR 6-19 vs. 3 (IQR 1-19) p=0.011). Overall mortality in the patients was high (45/74, 61%), and was not different in patients with or without source control at the moment of BSI (55 vs. 65%, p=0.47). CONCLUSION: The occurrence of a BSI in patients who recently underwent abdominal surgery should raise suspicion of inadequate source control. Factors associated with a higher risk for inadequate source control were the length of stay before BSI, the number of abdominal surgical interventions before BSI and the duration of the interval between abdominal surgery and BSI. Mortality in these patients is high, irrespective of source control.

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

588 BACTEREMIA IN ICU PATIENTS: A MATCHED, RISK ADJUST, MULTICENTER STUDY OF ATTRIBUTABLE MORTALITY Garrouste-Orgeas M M G O1, Tafflet M M T2, Soufir L L S1, Zahar J J Z3, Mourvillier B B M4, Adrie C C A5, Jamali S S J6, Darmon M M D7, De Lassence A A L8, Costa M M C9, Dumenil A A D10, Carlet J J C1, Timsit J J T11 1 Intensive Care, Hôpital Saint Joseph, 2Biostatistical Departement, Outcomerea, 3Microbiology, Hôpital Necker, Paris, 4Intensive Care, Hôpital Aulnay, Aulnay, 5Intensive Care, Hôpital Delafontaine, Saint Denis, 6Intensive Care, Hôpital Dourdan, Dourdan, 7Intensive Care, Hôpital Saint Louis, Paris, 8Intensive Care, Hôpital Louis Mourier, Colombes, 9Intensive Care, Hôpital Tenon, Paris, 10Intensive Care, Hôpital Clamart, Clamart, 11Intensive Care, Hôpital Bichat Claude Bernard, Paris, France INTRODUCTION: Nosocomial bacteremia (NB) is one of the leading nosocomial infections in ICU patients (pts). The importance of attributable mortality is controversial. The objective of the study was to determine attributable mortality taking into account the evolution of patient severity during ICU stay. METHODS: 3954 pts included in the Outcomerea database in 12 French ICUs were followed for the occurrence of NB. A matched (1: n) exposed-unexposed study was used. Each exposed pt with a NB after day 4 of ICU admission was matched with unexposed control pts with at least the same duration of exposure to the risk and a probability of death of ± 5% (TRIO score (1) assessing the severity of illness calculated between day 1 and 3 after ICU admission). Only the first episode of NB was included in the analysis. High risk micro-organisms were MRSA, P. aeruginosa, Acinetobacter sp, multi resistant Enterobacteriaceae sp and yeasts. RESULTS: 3247 (82%) pts stayed more than four days and constituted the cohort study. 237(7.3%) pts had an ICU-acquired NB. Matching was successful for 217 pts with 990 unexposed pts. Gram positive, Gram negative bacteria and yeasts were 142/243 (58.4%), 80/243 (33%) and 18 (7.4%) respectively. High risk micro-organisms were 68 (28%). Crude ICU and hospital mortality rates for exposed and unexposed pts were 51.1% and 27.8% (p<0.0001) and 60.8% and 36.1% (p<0.0001) respectively. After adjustement on the probability of death of the Trio score, the Logistic Organ Failure (LOD) 4 days before the NB and the type of the patient, NB was responsible of a 3-fold increase of hospital death: 2.98 (2.12-4.2), p<0.0001. The over-risk of death was higher for high risk micro-organisms NB (4.79 (2.62-8.75, p<0.0001), for NB not related to catheter or of unknown origin (OR: 3.91 (2.29-6.71, p<0.0001), and for NB inappropriately treated initially (OR: 3.32 (2.14-5.14, p<0.0001). CONCLUSION: When taking into account the severity of the patient before the nosocomial event, NB is associated with a 3-fold increased risk of hospital death. The estimate vary considerably according to the micro-organisms, site of infection and appropriateness of treatment.

S153

Oral Presentations Organ support – 590-592 590 PERFORMANCE OF TRANSPORT VENTILATORS Maeda Y1, Fujita Y1, Nishimura M2, Uchiyama A1, Mashimo T1, Fujino Y1 1 Intensive Care Unit, Osaka University Hospital, Osaka, 2Emergency and Critical Care Medicine, University of Tokushima Graduate School, Tokushima, Japan INTRODUCTION: Transport of mechanically ventilated critically ill patients is risky. Recently, several transport ventilators, which do not need compressed air tubing and are equipped with internal battery, have become commercially available. These ventilators can provide patienttriggered ventilatory assist. We evaluated the inspiratory and expiratory triggering performances, gas consumption and battery duration of these ventilators. METHODS: We compared the performances of four ventilators; LTV1000 (Pulmonetic Systems), iVent201 (VersaMed), VELA (VIASYS Healthcare) and Savina (Draeger Medical). A model lung simulating spontaneous breathing was set at a rate (RR) of 10/min, inspiratory time of 1sec, and peak inspiratory flow (IF) of 30 or 60 L/min. Each ventilator was set at positive end-expiratory pressure (PEEP) of 5cmH2O, pressure support of 10cmH2O and flow triggering was used. Sensitivity was set at the most sensitive setting without self-triggeriing. Airway pressure was measured and recorded. To evaluate inspiratory and expiratory trigger functions, we measured the inspiratory delay time (DT, the time between the start of inspiration to minimum airway pressure) and supra-plateau expiratory pressure (PE). A tank containing 500L of oxygen was connected to each ventilator. The model lung was ventilated using volume-controlled ventilation, tidal volume (VT) of 500mL, RR of 20/min, fraction of inspiratory qxygen (FiO2) of 1.0 and PEEP of 5cmH2O. The time required for the oxygen tank to be emptied was measured. Battery duration was also evaluated by measuring time for each ventilator to stop working when it was ventilating the model lung with VT of 500mL and RR of 20/min using an internal battery as its power source. RESULTS: See table. Ventilator DT (msec) IF 30

DT (msec) IF 60

PE (cmH2O) IF 30

LTV1000 64±5 72±4 1.36±0.12 iVent201 81±7 79±4 0.51±0.08 VELA 61±6 61±4 0.11±0.00 Savina 58±5 67±8 0.34±0.08 *Bias flow 10L/min**large internal battery

PE (cmH2O) Oxygen IF 60 consumption (min) (min) 5.20±0.20 35 * 6.03±0.08 51 1.18±0.13 32 * 1.94±0.16 57

Battery duration 103 89 403 ** 79

REFERENCE: (1): Timsit et al for the Outcomerea group. Intensive care Medicine 2001 27: 1012-1021

CONCLUSION: Inspiratory trigger with all ventilators were less than 100msec, which is clinically fast enough. Expiratory trigger deteriorated with LTV and iVent at higher inspiratory flow. Gas consumption should be paid attention with LTV and VELA when transport takes more than 30 min and FiO2 is high.

589

591

THE USE OF APACHE II AT ONSET OF BACTEREMIA IN CRITICALLY ILL PATIENTS AS A PREDICTOR OF MORTALITY

PROGNOSIS FACTORS OF THROMBOTIC MICROANGIOPATHIES (TMA) TREATED BY PLASMA EXCHANGE THERAPY (PET)

Artero A1, Zaragoza R2, Camarena J3, Sancho S2, Alberola J3, Nogueira J3 1 Internal Medicine, 2Intensive Care Unit, 3Microbiology, Hospital Universitario Dr. Peset, Valencia, Spain

1

INTRODUCTION: The aim of this study was to determine if APACHE II at the onset of bacteraemia is a predictive factor of mortality in critically ill patients. METHODS: From 1998 to 2004, 129 patients with a significant bacteraemia were prospectively evaluated in an intensive care unit. Clinical and microbiological variables were studied. The APACHE II score was recorded on the day of admission to the ICU and the day when blood cultures were obtained. A multivariate analysis was performed to determine the influence of these scoring systems on global and related mortality to bacteraemia. RESULTS: Seventy-nine men and 50 women, with a mean age of 63.6 ± 13.8 years were studied. Septic shock was present in 52.7%. Inadequate empirical antimicrobial treatment was given in 34.1% of cases. High risk focus of bacteraemia and high risk microorganisms were present in 39.5% and 41.9%, respectively. APACHE II recorded at the onset of bacteraemia was related to global mortality (OR 1.20, 95% CI 1.09 – 1.32) and related mortality to bacteraemia (OR 1.14, 95% CI 1.03 - 1.26). Septic shock was associated with related mortality to bacteraemia (OR 7.36, 95% CI 2.21 – 24.50). APACHE II recorded on the day of admission to the ICU was not related to neither global nor related mortality to bacteraemia. CONCLUSION: APACHE II at the onset of bacteraemia was associated with global and related mortality to bacteremia in critically ill patients. For the assessment of mortality in bacteraemic critically ill patients APACHE II at onset of bacteraemia was a better indicator than APACHE II on the day of admission to the ICU.

Klouche K1, Federici L1, Massanet P1, Amigues L1, Rossi J F2, Beraud J J1 Intensive Care Unit, 2Hematologic Unit, Lapeyronie University Hospital, Montpellier, France

INTRODUCTION: The use of PET in adults with TMA has dramatically improved outcome. Resistance to PET, which is observed in 1/3 of such patients and may affect mortality, remains however incompletely understood. We retrospectively studied 25 adults with TMA treated by PET in our unit to evaluate short and long term outcome and to identify predictive factors of mortality and of resistance to PET. METHODS: All records of adults with TMA treated by PET between 1999 and 2003 were reviewed. TMA associated with bone marrow transplantation were excluded from the study. Age, sex, cause of TMA were collected. Glasgow and SOFA scores were estimated at the admission. Clinical data including: neuroligical or pulmonary disorders with mechanical ventilation (MV), renal failure, and therapeutic delay (TD) to PET; biological data including: haemoglobinemia, platelet count, and LDH; plasmatic volume exchange per procedure and number of plasmapheresis sessions were also collected. Mortality was assessed at one month and at one year follow-up. All data were analyzed and compared between survived/deceased and between responders/nonresponders (R/nR) patients. RESULTS: 19 females and 6 males were included. Mean age: 46.8±16.3 yo, mean Glasgow coma score: 11±3, mean SOFA score: 5.8±2.8. Aetiologies of TMA: post-immunologic 4, postinfectious 4, post-neoplastic 4, drugs associated 4, idiopathic 9. Two patients were in MV, 3 underwent haemodialysis and 10 had at least two organ dysfunction. The mean TD for PET was 2.7±5.2 days and the mean plasmatic volume exchange per procedure was 35.9 ±8.6 ml/kg. 19 patients (76%) partially or fully responded to PET. 20 patients (80%) survived after 1 month and 19 (76%) after 1 year follow-up. The comparison between survived and deceased patients showed that response to PET (18/20 vs 1/5 R respectively) was the only significant determinant parameter. The comparison between R and nR showed that a longest TD(2.5±1.8 vs 3.3±2.0) and neoplastic cause of TMA (3/19 vs 3/6 neoplastic TMA) were significantly discriminant for nR. Almost all of the R exhibited a positive response to PET(95%) before the tenth plasmapheresis session. In a median follow-up period of 23,6 months, relapses episodes occurred in 7 patients(35 %). CONCLUSION: Adults with TMA, characterized by a mild to important severity, treated by PET have a relatively good outcome since survival reached 80% at 1 month and was maintained at 76% after 1 year follow-up. Among the parameters studied, lack of response to PET was the only predictive factor of mortality. Two factors were predictive of resistance to PET: neoplastic aetiology of TMA and a longer TD to PET. Adult with TMA non-responding to PET after the tenth plasmapheresis session could be considered as totally nR and should benefit shortly from another therapy.

S154

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

592

594

USE OF THE HUMAN RESPIRATORY SYSTEM MODELS TO EXPLAIN HFV EFFECTS OBSERVED IN CLINICAL PRACTICE

LAPAROSTOMY FOR INTA-ABDOMINAL SEPSIS: A CASE CONTROL STUDY

Roubik K1, Rozanek M1, Kopelent V1, Pachl J2 1 Dept. of Radioelectronics, Czech Technical University, 2Dept. of Anaesthesiology and CCM, Charles University, 3rd School of Medicine, Prague, Czech Republic INTRODUCTION: Different effects of artificial ventilation can be observed when conventional ventilation (CV) and high frequency ventilation (HFV) are used. An additional technique of artificial lung ventilation improving alveolar oxygenation and CO2 removal without increasing tidal volume or pressure amplitude in the respiratory system is represented by tracheal gas insufflations (TGI). Deriving model of the respiratory system exactly corresponding with the reality is one of few possibilities how to study effects of mechanical lung properties, ventilatory parameters, TGI, etc. METHODS: Two models have been developed to model the real situations in the human respiratory system (RS) during CV and HFV. The modelling techniques respect the real geometrical proportions of the human RS in both models. The RS is considered as an acoustic system in the first case. An electro-acoustic analogy was used to develop an electric model of the respiratory system respecting its exact anatomical structure. The final model has 23 airway generations and it employs 67 108 859 individual components. Concerning the second model, the RS is modelled as a diverging tube with a growing diameter. The core of this model is a convection-diffusion equation (partial differential equation), which is solved numerically under proper initial conditions given by the model structure and gas concentrations in the RS. RESULTS: Changes of alveolar compliance have a significant effect on total lung impedance (TLI) during CV while TLI changes during HFV are not essential. Contribution of airway resistance changes is significant mainly during HFV contrary to CV. It results in better functioning of HFV in patients with a reduced alveolar compliance, which is the main symptom of adult respiratory distress syndrome. Nearly 95% of proximal pressure amplitude is present inside the lung structure when using CV. On the other hand between 5-10% of the input pressure amplitude is transferred deep to the structure of the RS during HFV. It suggests that HFV is a protective ventilatory strategy. Combination of HFV+TGI is better in oxygenation than combination of CV+TGI (by 14.4% for TGI flow = 24% of MV). Gas flow modelling shows a strong dependence of alveolar PAO2 on VT during HFV while it is preserved constant during CV. Therefore a proper VT adjustment is also essential for control of oxygenation during HFV contrary to CV.

Macnaughton P D1, Edwards T J2, Finlay I2, Lambert A W2 1 Critical Care Unit, 2General Surgery, Derriford Hospital, Plymouth, United Kingdom INTRODUCTION: Damage control surgery (DCS) involves initial temporising surgery, laparostomy, stabilisation in the intensive care unit (ICU) followed by delayed definitive surgery. Its incorporation into the management of critically ill patients with non-traumatic pathology may improve outcome. This retrospective case control study investigated the efficacy and impact on ICU resources of DCS in the management of severe intra-abdominal sepsis. METHODS: A consecutive series of patients presenting with non-traumatic intra-abdominal sepsis, managed by DCS, laparostomy and ICU admission, were incorporated into a case control study. The control group comprised all patients with matching diagnoses who underwent standard surgical interventions and ICU admission during the same time period. Demographics, APACHE II score and predicted mortality, ICU and total post op stay, and hospital mortality were recorded prospectively. RESULTS: 94 patients (17 cases; 77 controls) were analysed. Median ages were 55 and 75, and mean APACHE II scores were 14.29 and 16.25 for cases and controls respectively. The crude ICU and hospital mortality for cases and controls was 6% and 18% (cases) and 12% and 31% (controls) respectively. The odds ratio of death for cases versus controls was 0.39 (95% CI: 0.1 to 1.48). The case mix adjusted predicted mortality was calculated using the APACHE II model. The ratio of observed to predicted deaths was 1.07 (95% CI: 0.22 to 3.13) for the cases and 1.62 (95% CI: 1.05 to 2.37) for the controls. The mean length of ICU stay was significantly longer for cases than controls (13.8: 7.1 days; p=0.025), although the overall hospital stays were equivalent (28.3: 29.2 days; p=0.89). CONCLUSION: There is a trend towards an improved outcome when DCS and laparostomy is used in the treatment of severe intra-abdominal sepsis but this appears to be associated with a significantly longer ICU length of stay when compared to standard management. Larger studies are required to establish whether this approach to the management of severe intra-abdominal sepsis is justified.

CONCLUSION: It is possible to use results of mathematical simulations for explanation of differences between CV and HFV. Some important effects observed in the clinical practice can be explained and studied by these modelling techniques. Grant acknowledgement: The work has been supported by MSMT VZ:J04/98:210000012 and GA CR 102/03/H086.

595

Oral Presentations Postoperative II – 593-595 593

REDUCED INTRAVASCULAR VOLUME AND CARDIAC OUTPUT IN THE PRESENCE OF INTRA-ABDOMINAL HYPERTENSION

PREDICTIVE MARKERS FOR POSTOPERATIVE SEPSIS Swaraj S1, Breen A1, Ritchie G1, Toh C2, Shenkin A3, Leuwer M1, Marx G1 Department of Anaesthesia, 2Haematology, 3Clinical Chemistry, University of Liverpool, Liverpool, United Kingdom 1

INTRODUCTION: Early recognition of sepsis is of particular interest, because it could expedite the initiation of early specific treatment and potentially improve patient outcome. The aim of our prospective clinical observational study was to determine the value of procalcitonin (PCT), white blood cell count (WCC), interleukin 6 (IL-6), terminal complement complex (C5b-9) and Creactive protein (CRP) in the prediction of postoperative sepsis. METHODS: 72 non-infected patients admitted postoperatively to a general Intensive care unit (ICU) who required >24 h ICU-treatment were recruited . Blood samples were collected on the 1st, 2nd, 3rd, 5th, 7th and 10th day. PCT was determined by an immuno-luminometric assay (Lumitest PCT Brahms Diagnostica). Other parameters were measured with conventional methods. Sepsis/severe sepsis/septic shock were defined according to the ACCP/SCCM criteria and severity of sepsis-related organ failure by SOFA score. Values were given as median and interquartile range. Sensitivity and specificity were calculated. Diagnostic accuracy of determined parameters was expressed as area under curve (AUC). RESULTS: Of 72 recruited patients 11 developed postoperative sepsis and 18 severe sepsis or septic shock. Infection was microbiology proven in 22 of 29 patients (76%). The patients enrolled had an age of 68 (21.5) and an initial SOFA score of 5 (4). After 28 days, 12 patients had died, 60 were survivors. Diagnostic performance of markers to predict sepsis on inclusion day; *p<0.05 Cut-off value Sensitivity (%) Specificity (%) AUC (95%CI)

PCT 2 ng mL-1 52 86 0.70*

IL-6 150 pg mL-1 90 40 0.80*

C5b-9 600 ng mL-1 31 65 0.51

WCC 12000 mm3 41 65 0.54

CRP 150 mg mL-1 55 60 0.64

CONCLUSION: Our results indicate that procalcitonin may be a specific and IL-6 a sensitive predictive marker for sepsis in postoperative patients. Grant acknowledgement: Supported by R&D RLBUH NHS Trust fund & R&D fund University of Liverpool.

Schachtrupp A1, Graf J2, Toens C3, Lawong G1, Fackeldey V1, Schumpelick V1 1 Dpt. of Surgery, 2Dpt. of Internal Medicine, Technical University, Aachen, 3Surgical Clinical, Marien Hospital, Duesseldorf, Germany INTRODUCTION: Trauma, peritonitis or vascular lesion may lead to a pathological increase of the intraabdominal pressure (IAP) which is also known as intraabdominal hypertension (IAH). In the presence of IAH, a reduction of cardiac output (CO) together with an increase of so called filling pressures e.g. central venous pressure (CVP) has been noted (1). In this regard generous volume resuscitation has been observed to lead to a higher incidence of the abdominal compartment syndrome (ACS) (2). However, measurement of the intravascular volume has not been performed under the condition of long term IAH. The purpose of the study was to examine hemodynamic parameters and intravascular volume in a porcine model in the presence of intraabdominal hypertension (IAH) lasting for 24 hours. METHODS: Twelve pigs (52.5 ± 4.9 kg) were studied over a period of 24h. Animals were ventilated volume controlled and were analgo-sedated continuously using ketamine and pentobarbitale. In six animals the intraabdominal pressure (IAP) was increased to 30 mmHg via CO2-pneumoperitoneum. The others served as control. Using transpulmonary double-indicator dilution technique, intrathoracic blood volume (ITBV), total circulating blood volume (TBV) and cardiac output (CO) were measured. Standard parameters, e.g. central venous pressure (CVP), were also recorded. Statistic analysis was performed using ANOVA and post-hoc analysis together with paired t-tests. A p<0.05 was considered significant and was adjusted (Bonferroni) when multiple paired analysis was performed. RESULTS: In the presence of IAH, ITBV and TBV were significantly reduced to 55% and 67% of control values. CO decreased to 27% while CVP increased fourfold. CONCLUSION: IAH leads to significant reduction of the cardiac output and the intravascular volume. The letter is not reflected by the CVP. Assessment of CO and ITBV in the presence of a critically increased IAP is therefore recommended. REFERENCES: 1. Ridings PC, Bloomfield GL, Blocher CR et al: Cardiopulmonary effects of raised intra-abdominal pressure before and after intravascular volume expansion. J Trauma 1995; 39: 1071-1075. 2. Balogh Z, McKinley BA, Cocanour CS et al: Supranormal trauma resuscitation causes more cases of abdominal compartment syndrome. Arch Surg 2003; 138: 637-643

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

Oral Presentations Continuous renal replacement 596-598 596 UNFRACTIONED VERSUS LOW-MOLECULAR-WEIGHT HEPARIN (ENOXAPARIN) FOR ANTICOAGULATION IN CVVH Joannidis M1, Kountchev J1, Grote A1, Bellmann R1, Mayr A2, Wiedermann C1 1 General Internal Medicine, Medical Intensive Care Unit, 2Anaesthesiology and Critical Care, University Hospital, Innsbruck, Innsbruck, Austria INTRODUCTION: Heparin is the most frequently used anticoagulant for CRRT in European ICUs. Nevertheless, low-molecular-weight (LMW) heparins appear to possess several advantages over unfractioned heparin, such as causing less bleeding, platelet activation, AT III consumption and thrombocytopenia, as well as affecting lipid profile to lesser extent. We have reported interim results of our study comparing unfractioned heparin versus enoxaparin previously. This is the final analysis after completion of the study. METHODS: Forty consecutive adult medical and surgical ICU patients with normal anticoagulation parameters requiring CRRT were included in this randomised, prospective, crossover study. CVVH was performed with pre-filter fluid replacement at 2500 mL/hr and blood flow rates of 180 ml/min. Filters were primed with normal saline containing anticoagulant (5000 IE heparin or 25 mg enoxaparin, respectively). Heparin-treated patients received a initial pre-filter bolus of 30 IU/kg and a maintenance infusion at 7 units/kg/hr, titrated to achieve a systemic activated partial thromboplastin time (aPTT) of 40 - 45 seconds. Enoxaparin-treated patients received an initial pre-filter bolus of 0,15 mg/kg and a maintenance infusion starting at 0,05 mg/kg/hr, which was subsequently adjusted to maintain systemic anti-factor Xa activity (anti-Xa) at 0,25-0,30 IU/ml. To adjust anticoagulation blood samples were drawn at baseline, 0.5, 1, 2, 4, 12, 24 and 48 hours after initiation and at the end of CVVH. Maximum treatment duration for each set was 72 hours. RESULTS: Patients included showed a mean APACHE II score of 20 (range 10 to 35). Mean filter life span was 21,5 h (±16,9 SD) for heparin and 30,8 h (± 25,3 SD) for enoxaparin (p=0.018, paired t-test). No correlation could be established between filter life span and either peak aPTT or steady state aPTT for heparin treatment. On the other hand, a significantly positive correlation was found between filter life span and both peak systemic and post-filter anti-Xa activity (p=0.029 and 0.008, respectively). Two major and one minor bleeding episodes occurred during heparin treatment. One major and two minor bleedings were observed during enoxaparin treatment.

S155

598 PLASMA EXCHANGES: THE BEST THERAPEUTIC APPROACH FOR ICU PATIENTS WITH THROMBOTIC MICROANGIOPATHY Darmon M1, Thiery G1, Moreau D1, Ciroldi M1, Le Gall J1, Schlemmer B1, Azoulay E1 1 Medical ICU, Hôpital Saint-Louis, Paris, France INTRODUCTION: Thrombotic microangiopathy (TMA) is a rare life-threatening disease characterised by peripheral thrombocytopenia, microangiopathic haemolytic anaemia, fever, and various stages of renal and neurological dysfunction. Only few studies focused on critically ill patients with TMA, in whom the optimal therapeutic management remains controversial. METHODS: We studied a cohort of consecutive patients with TMA admitted between January 1, 2000, and June 30, 2003. Patients characteristics, treatment and evolution of organ failure (LOD score) where collected within the 10 first days of ICU-admission. RESULTS: Among the 36 patients , 22 were treated by plasma infusion (PI) and 15 by Plasma exchanges (PE). All patients had anaemia or thrombopaenia at ICU admission, 50% had neurological symptoms and median creatinine clearance was 55 ml/min. Median LOD at admission was 5 (2-7), platelet count was 25500 (11500-48000) and schizocytes were present in every patient. 14 patients had microbiologically documented infection, 7 were recipients of allogenic BMT and 7 had systemic disease. Six patients had AIDS, 5 had drugs-associated TMA, 2 were pregnant and 2 had cancer. No predisposing factors were found in 10 patients. Median ICU LOS was 7 days (4-13), ICU mortality was 5.6% (two patients) and Hospital Mortality was 13.9% (seven patients). PE was associated with a decreased mortality (no in-hospital death versus 7 in the PI group. p< 0.001). Moreover, LOD score became significantly lower in the PE group from day-3 when compared to patients assigned to the PI group. Plasma volume day-9 (ml/kg) <0.0001 LDH Day-3 LOD Day-0 LOD Day-3 LOD Day-5 LOD Day-9 Weight gain (Kg)Day-0 to 9 Hospital mortality

PI Group 242 ± 89

PE Group 604 ± 81

P

1000 ± 600 5.88 ± 4 4.93 ± 4 5.1 ± 4.8 7.11 ± 4 +2.4 ± 0.8 7 (31.8 %)

608 ± 200 4.06 ± 2 2.47 ± 2 2±2 2.36 ± 2 -3.1 ± 0.3 0

0.04 0.40 0.05 0.05 0.01 0.03 0.005

CONCLUSION: PE should be the first therapeutic option in critically ill patient with TMA.

CONCLUSION: Enoxaparin can be safely used for anticoagulation during CVVH resulting in higher filter life span compared to unfractioned heparin. Both post-filter and systemic anti-Xa activities correlate with filter life span in patients treated with enoxaparin.

597 LESS BLEEDING COMPLICATIONS DURING CITRATE CVVH COMPARED TO NADROPARIN Van der Voort P H J1, Postma S1, Van Roon E N2, Koopman M1, Kingma W P1 Intensive Care, 2Clinical pharmacology, Medical Centre Leeuwarden, Leeuwarden, Netherlands

1

INTRODUCTION: The anticoagulant used to prevent clotting of the extracorporeal circuit during continuous veno-venous haemofiltration (CVVH) has impact on filter run time and bleeding complications in the patient. Citrate is a regional whereas nadroparin is a systemic anticoagulant. METHODS: We studied prospectively 304 consecutive CVVH sessions in 60 ICU patients with multiple organ dysfunction syndrome. For each session we observed the anticoagulant used, the filter run time (FRT), the reason to stop the CVVH, need for transfusion and bleeding complications. RESULTS: 148 sessions used citrate as an anticoagulant, 124 nadroparin, 13 heparin, 2 danaparoid and 17 no anticoagulant. The mean FRT for the citrate group was 23 hours (SD 16), for the nadroparin group 30 (SD 17) hours (p=0.002, figure), for the heparin group 36 (SD 20) and for the sessions without anticoagulant 25 hours (SD 19). During the CVVH session, a bleeding complication occurred in 22/148 (15%) patients in the citrate group and in 31/124 (25%) patients of the nadroparin group (p=0.045, Fisher exact 2 sided). The kind of bleeding was catheter site 21 vs 13, digestive tract 4 vs 4, other bleeding sites 6 vs 5. The number of transfused blood cell concentrates was not significantly different.

Oral Presentations Cost issues – 599-601 599 PATTERNS OF SURVIVAL, DEATH, AND RESOURCE UTILIZATION IN AN INTENSIVE CARE UNIT Meinich P1, Stokland O2, Haagensen R3, Olafsen K2 1 Anaesthesia, Ulleval University Hospital UUS, 2Intensive Care, Ulleval University Hospital, Oslo, 3Intensive Care, Akershus University Hospital, Akershus, Norway INTRODUCTION: UUS is a tertiary level reference university hospital. Our aim was to elucidate how we used our resources on patients who survived vs. patients who died in the ICU. METHODS: Retrospective study of 1214 patients over a 44 months period, median age 58 years, 67% men, 56% unscheduled and 21% scheduled surgery, 23% medical. Severity of illness judged by Simplified Acute Physiology Score, SAPS II. Resources measured by Nine Equivalents of nursing Manpower Score, NEMS. RESULTS: Overall mortality was 19%. Median SAPS was 32 for survivors and 51 for those who died. 68% of the patients were either discharged from the ICU or dead within the first week. These patients used 23% of the resources. 21% of these patients died (median SAPS 53), while 79% survived (median SAPS 29). Only 4% of the patients stayed longer than four weeks. However, these 4% used 22% of the total resources. 27% of these patients died (median SAPS 61), and 73% survived (median SAPS 38.5). There was a tendency towards higher NEMS among those who died for all lengths of stay (LOS), and patients who died after lengthy stays had higher NEMS also during the first period of their stay. Length of Stay Patients Alive / Dead Days Total Alive / Dead SAPS Median Alive *) SAPS Median Dead *) NEMS/Pas/D Alive / Dead

0-7 Days 658 / 172 1818 / 422 29 (21-38) 53 (32-71) 27.3 / 35.9

29-79 Days 36 / 13 1491 / 569 38,5 (31-48) 61 (49-72) 30.3 / 32.3

Total 978 / 236 7487 / 1739 32 (23-42) 51 (42-61) 29,9 / 35.2

*) Interquartile Range in Parenthesis

CONCLUSION: Citrate CVVH was safer as less bleeding complications occurred compared to nadroprin CVVH. However, filter run time was shorter leading to a greater use of haemofilters.

CONCLUSION: Most patients stayed less than one week in the ICU, but 4% stayed longer than 4 weeks and used 22% of all the resources. More resources were used per patient per day on patients who died. Resource usage per patient per day was fairly constant, so that total resource usage per patient was closely correlated to length of stay. Patients with a prolonged ICU stay require large resources, and one should pay attention to the additional risks and complications of such treatment so as to limit futile treatment.

S156

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

600 INTENSIVE CARE UNIT FUNDING ACCORDING TO SPECIFIC PROCEDURES AND SAPS 2 LEVEL Guidet B R1, Boumendil A2, Aegerter P3 Réanimation médicale, Hôpital saint Antoine, 2U444, INSERM, Paris, 3U444, Hopital Ambroise paré, Boulogne, France 1

INTRODUCTION: Intensive Care Units (ICU) funding through the DRG system is unfair since it does not recognize the specificity of ICU activity. We obtained from French authorities an activity-driven ICU funding. Only stays fulfilling two conditions will be funded independently of DRGs. 1- Stay in an authorized ICU structure. Without strict definition of ICU, SAPS II will be used as a surrogate marker of ICU stay. 2- Specific ICU procedures, such as mechanical ventilation or haemodialysis, performed during the ICU stays. The objective of the study is to identify the percentage of stays with ICU procedures according to SAPS II values. METHODS: Data were extracted from a multicenter database including 36 ICUs in the Paris area (France) with 20,445 stays during year 2001 (age 55.8; SAPS II : 37.3; LOS: 7.2; ICU mortality 17.3%). ICU stays were distributed in five SAPS II categories. RESULTS: The percentage of ICU stays with ICU procedures is 51.3%. Among the 2503 stays with a SAPS II value lower than 15 (12.3% of all stays), there were 675 stays with ICU procedures accounting for 6.4% of all stays with ICU procedures. TABLE 1: Number and percentage of ICU procedures according to SAPS categories SAPS II n stays n stays with procedures % stays with procedures

0 - < 15 2503 675 27

>=15 - < 25 4190 1259 30

>= 25-< 35 4243 1765 41.6

>= 35 - < 50 4614 2737 59.3

> 50 4865 4085 83.1

CONCLUSION: A combination of two criteria : specific ICU procedures performed during an ICU stay for a patient with a SAPS II value greater than 15 enable identification of “real ICU stays” that deserve special funding. A SAPS II threshold of 15 induces a loss of 6.4% of stays with ICU procedures but avoid to include to many stays with low severity. The financial impact of this new funding method is estimated to be close to 900 per ICU day of “ real ICU stays” . Grant acknowledgement: CUB-REA is funded by AP-HP and INSERM

601 A SIMPLIFIED METHOD OF ISOFLURANE ADMINISTRATION IN THE ICU ENVIRONMENTAL AND COST ASPECTS Sackey P V1, Martling C1, Nise G2, Radell P J1 Anesthesiology and Intensive Care Medicine, Surgical Sciences, 2Occupational and Evironmental Medicine, Environmental Medicine, Stockholm, Sweden

1

INTRODUCTION: Inhalational agents represent an alternative to intravenous sedatives for ventilator-dependent intensive care unit(ICU)patients. Excessive agent consumption, lack of vaporizers in the ICU and environmental concerns may partly explain why sedation with inhalational agents has not become established in the ICU (1,2). The Anaesthetic Conserving Device(ACD, AnaConDa,Hudson RCI,Upplands Väsby,Sweden)is a modified heat-moisture exchanger(HME)fitted at the endotracheal tube with features permitting direct infusion of isoflurane to the ACD and rebreathing of exhaled anaesthetic agent to the patient. Consequently only a small amount of anaesthetic gas escapes the system. The degree of ambient pollution and isoflurane consumption of ACD-delivered isoflurane sedation in the ICU setting is unknown. We studied 13 patients sedated with isoflurane via the ACD at the General ICU of the Karolinska University Hospital, Stockholm, to evaluate ambient isoflurane pollution and isoflurane consumption. METHODS: Thirteen ICU patients were sedated with isoflurane infused via the ACD for 12-96 hours. Changing of the ACD, isoflurane syringe and opening of the respiratory circuit was performed according to investigators standardized instructions. Active scavenging of waste gas from the respirator was performed in eight patients, in the remaining five no active scavenging was performed. Ambient isoflurane pollution was measured with stationary continuous spectrophotometry near the patient and with passive dosimeter sampling for ten staff nurses over eight-hour shifts. Isoflurane requirement was registered. RESULTS: Spectrophotometry readings(0,1 ± 0,2 parts per million)were well below internationally recommended working limits in all patients. Peaks during interventions were short, infrequent and of low amplitude. There was no significant difference between isoflurane trace levels with or without an active scavenging system. Dosimeter values were also low, ranging from undetectable to 0,1375 ppm. Mean isoflurane consumption was 2,2 ± 1,0 ml per hour. This is approximately one fourth of predicted and previously reported consumption of isoflurane with vaporizer-administered sedation in the ICU setting. CONCLUSION: Isoflurane via the ACD is an environmentally safe method of sedation provided users follow standardized instructions in situations with potential spillage of isoflurane.This method of sedation requires less isoflurane than traditional vaporizer technique. REFERENCES: 1.Millane TA, Bennett ED, Grounds RM: Isoflurane and propofol for long-term sedation in the intensive care unit. A crossover study.Anaesthesia.1992;47:768-74. 2. Kong KL, Bion JF: Sedating patients undergoing mechanical ventilation in the intensive care unit-winds of change? Br J Anaesth.2003;90:267-9