key: cord-0033863-cqtg2196
authors: nan
title: Poster Sessions 555 - 719
date: 2003-09-23
journal: Intensive Care Med
DOI: 10.1007/s00134-003-1975-9
sha: 13adcaf2b074661bc80c7fd31d75bfda3212b763
doc_id: 33863
cord_uid: cqtg2196

nan

After recruitment, gas in lungs (CT, n=4) amounted to 1462(1356/1558)ml and PaO2 to 532(290/628)mmHg [mean(min/max)]. The suction flow of 20l/min effected a dP along the SL-ETT of 8.0(7.5/8.5)mbar and a relevant reduction of Pa02 and lung volume not seen with DL-ETT. 

The DL-ETT technique is promising for ventilating patients with acutely injured lungs: Reduction of lung volume (= alveolar collapse) and subsequent re-recruitment, potentially damaging to lung tissue, can be reliably avoided even during suctioning.

Gendo A 1 , Bauer E 2 , Zauner C 1 , Kramer L 1 , Holzinger U 1 , Madl C 1 1 Department of Internal Medicine IV, ICU, 2 Department of Internal Medicine III, Nephrology, General Hospital, Vienna, Austria

A 30% reduction of mortality and an increase of the number of ventilatorfree days were described in patients with acute lung injury (ALI) treated with permissive hypercapnia. However, according to recent animal studies, hypercapnia leads to a temporary rise of the cerebral blood flow as well as cerebral blood volume and to an increase of the intracranial pressure. Other studies showed the disappearance of the cerebral autoregulation if the pCO2 was acutely rised to 80 mmHg. All changes described may negatively influence cerebral function. The effect of permissive hypercapnia on cerebral function in men has to be investigated.

METHODS. 11 patients (8 males, mean age 58±17 years) with ALI on account of bilateral pulmonary infiltrates were enrolled. All patients were ventilated in pressure control mode (tidal volume: normocapnia: 9-13 ml/kg body weight, hypercapnia: 4-6 ml/kg body weight). Cerebral function was assessed by short-(N13-N20 interpeak latency) and long-latency (N70 peak) sensory evoked potentials (SEP), an objective and sensitive method of cerebral integrity. Regional cerebral oxygen saturation (rSO2) was measured by near infrared spectroscopy. After randomisation SEP were assessed by a doctor blinded for ventilation mode in a cross-over design at baseline and 1, 3 and 6 hours after enrollment.

During normocapnia SEP peak latencies and rSO2 remained unchanged. Data during permissive hypercapnia are shown in the There was a significant impairment of the long latency-SEP after 1 hour of treatment with permissive hypercapnia. After 6 hours of treatment, the N70 peak latencies returned to baseline values.These results indicate a cerebral adaptation to a pCO2 which is increased for a longer time period. 1 Icu, ABC medical center, Mexico city, Mexico INTRODUCTION. sevoflurane(sevo) is an inhaled anesthesic of short acting period brief elimination time and minimal hepatic biodegradation.Thus it´s an attractive agent for intensive care (IC) patients (P) sedation(S) who are under mechanical ventilation(MV)the scope of this work was to determinate its uselfullness in medium term sedation METHODS. Prospective and longitudinal study approved by the hospital ethical committee were performed. P under mechanical ventilation for longer than 24 hrs in the ICU were included The two exlusion criteria were intracraneal pathology and malignant hyperthermia Demographic data simplified acute physiological score (SAPS),s time length Ramsay score (RS)percent of inhaled anesthesic qualiti and time to wakeup were recorded All of the patients were respiratory and haemodynanamically monitored Fluids inotropics and vasoactive amines were used for haemodynamic stability Respiratory supportwas achieved improving MV parameters ALL patients received 2 mcg/kg/hr fentanyl analgesia.Ventilator in its inspiratory flow branch with 1litter of O2 flow was used RESULTS. eigth male (67%)and 4 female(33%)with a mean age of 61±18yearsand initial SAPS of 16±3 were included.Sevo concentration varied betwen 0.9±0.8 and 1.5±1.3%.Initial an final RS mode were 4and 5.Total S time was 70±34 hrs. The wakeup time defined as the time elapsed for open eyes once sevo was stopped was 17±18 min. One p die after the first 24 hr due mesentheric trhombosis.Another P presented agitation after suspension of sevo. Sevo administration dose not modified hepatic or renal functions. Intrapulmonary shunt started in 28.6% and finished in 19.4% CONCLUSION. Sevo pharmacokinetics and pharmacodinamics allows its administration in IC P without detrment in haemodynamic,respiratory,hepatic and renal functions.P present an adecuate grade of S and rapid wakeup time.We conclude that sevo is usefull in short term and medium term S in P under MV.

16th Annual Congress -Amsterdam, Netherlands -5-8 October 2003 S153 Poster Session Neuroendocrinological aspects of sepsis -583-596 583

Su F 1 , Nguyen N D u c 2 , Wang Z 1 , Rogiers P 3 , Vincent J 1 1 Intensive care department, Erasme Hospital, 2 Intensive care department, AZ-VUB Hospital, Brussels, 3 Intensive care department, AZ Middelheim Hospital, Antewep, Belgium INTRODUCTION. Fever is the primary host defense mechanisms of life and an energydependent process. Heat shock proteins (HSP) may have protective effects. Without knowing whether fever is blessing or curse, antipyretics are widely used. The aim of this study is to investigate whether utilization of acetaminophen and external cooling to control fever in ewe septic shock model is beneficial and influence HSP70.

Twenty-four fasted, anaesthetized, invasively monitored, mechanically ventilated female sheep (27.0±4.6 Kg) received 0.5 g/kg body weight of feces into the abdominal cavity to induce sepsis. Ringer¢s lactate (RL) was titrated to maintain pulmonary artery occlusion pressure (PAOP) at baseline level throughout the experimental period without any antibiotics and vasoactive drugs utilization. After surgical operation, randomization was performed as following: if temperature< 36.0°C, the animal was placed in the hypothermia group; the other animals were randomized to three groups: high fever (T>39.0°C); mild fever(37.8°C<T<38.2°C) and normothermia(36.0°C<T<37.0°C) group. Acetaminophen 25 mg/ kg/ 4~6 hours combined with external cooling (ice pad) was used to control core temperature in the expected range. Hemodynamic, mechanical ventilation parameters, and gas exchange values were obtained every hr. Plasma samples were obtained every four hrs for HSP70 measurement (Hsp70 ELISA Kit, Stressgen, Canada).

Survival time was longer in the fever group (25.2 ± 3.0 hrs) than in the mild fever group (17.7 ± 3.5 hrs), normothermia group (16.0 ± 1.9 hrs) and hypothermia group (18.5 ± 2.5 hrs) (p < 0.05). There was no significant difference in homodynamic parameters except that DO2 (oxygen delivery) was higher in the two fever groups then the other two groups (p < 0.05). PaO2/FiO2 ratio was highest and blood lactate level was lowest in the high fever group than the other three groups (p < 0.01 and 0.05 respectively). Plasma HSP70 level was higher in the two fever groups than in the other groups (p<0.05).

In this septic shock model, febrile response had beneficial effects on the respiratory function, blood lactate level and survival time. Antipyretic interventions including acetaminophen and external cooling were associated with lower circulating HSP70 levels.

Myrianthefs P 1 , Karatzas S 1 , Venetsanou K 1 , Boutzouka E 1 , Evagelopoulou P 1 , Zidianakis V 1 , Fildissis G 1 , Baltopoulos G 1 1 Athens University, School of Nursing, ICU, KAT General Hospital, Athens, Greece INTRODUCTION. Biologic rhythms are found in the immune system (circadian, circaseptan, circannual), which has important clinical applications in medical practice, including the establishment and interpretation of reference values (1). Sepsis is associated with a refractory state characterized by a reduced capacity of monocytes to produce pro-inflammatory cytokines after restimulation with LPS ex vivo ("LPS tolerance") (2) . Monocytic ex vivo capacity to produce proinflammatory cytokines in response to LPS (whole blood and ELISA) has been suggested as a measurement of immunocompetency in patients with sepsis (3).

We examined seasonal differences in whole blood cytokine production after LPS stimulation in 17 healthy volunteers. We first established a dose and time response curve for TNFa production after LPS stimulation of whole blood. We selected 500 pg/ml of LPS for incubation period of 4 h to stimulate 100 microL of whole blood of the same subjects in June, September, February, and March.

We found no differences in the circulating total WBCs and differentials between different seasons. Cytokine levels are shown on the CONCLUSION. In early summer there is a predominating pro-inflammatory response, which is counterbalanced in autumn. These results may have significant implications in the determination of reference values, in exploration of immune response between different seasons, in determining LPS tolerance and planning clinical trials and immunomodulary therapies. 

Nam D D u c 1 , Su F 2 , Spapen H 1 , Malderen C 1 , Opdenaecker L 1 , Huyghens L 1 1 Intensive care department, AZ-VUB Hospital, 2 Intensive care department, Erasme Hospital, Brussels, Belgium INTRODUCTION. Encephalopathy (SE) is a common complication of sepsis and associates with a higher mortality. The aim of this study is to investigate the two brain specific proteins markers in serum: Neuron-Specific Enolase (NSE) and S100B protein (S100) as indicators of brain damage in septic shock patients.

Serum measurement of NSE and S100 was performed in 91 patients with septic shock, in the first 24 hours after haemodynamics (HD) stabilization and for 4 consecutive days. NSE and S100 were considered elevated if their values were above 12.5 ƒ_g/L and 0.5 ƒ_g/L respectively. Patients were not included in the study if they had a primary cerebral infection, or a secondary sepsis after brain injury, cardiac arrest, coronary by-pass surgery, as well as those that had primary renal and liver failure at their admission. Chi square test was used for statistics.

The mean age of these 91 patients was 64.4¡Ó13.6 years old. 67 patients had elevated NSE (73%), but only 32 had elevated S100 (35%). Compared with no difference in mortality and multi-organ failure (MOF) between NSE positive and negative group, S100 positive was correlated with higher mortality and morbidity of MOF (p<0.005). In the subgroup of patients with an elevated S100 in the first 24 hours and normalized value in the following 24 hours, mortality and morbidity of MOF was significantly lower (p<0.005).

Elevation of NSE and S100 in more than one third patients with septic shock showed that brain damage was an on-going process and occurred early during the evolution of sepsis. The elevation of S100, but not NSE, was associated with a higher incidence of MOF and had a predictive value for adverse outcome in these patients.

Septic shock is the reason of human body organ dysfunctions including the hormonal system. There are abnormal thyroid hormones releasing as well. Unfortunately clinical value of these changes is unknown yet. It is also noticed that sepsis caused serious disturbances in pituitary-thyroid axis function. This is called euthyroid sick syndrome (ESS). The aim of this study was to qualify the prognostic value of thyroid hormones serum concentrations changes in patients with septic shock. We also hope that thyroid hormone mechanism of action recognition in the septic shock could help in treatment of these patients.

30 patients with septic shock were included into the study. Septic shock was diagnosed according to AACP/SCCM criteria. The study group was divided into two subgroups: survivors (n=15) and non-survivors (n=15). 20 healthy volunteers served as controls. Blood for analysis was taken at the moment of septic shock recognition and on the 1st, 2nd, 5th and 10th day of the observation between 8.00 a.m. and 9.00 a.m. We studied thyrotropin (TSH), free triiodothyronine fraction (fT 3 ) and free thyroxin fraction (fT 4 ) serum concentrations, APACHE II and APACHE III scores, acute lung injury (ALI) or acute respiratory distress syndrome (ARDS) appearance.

We noticed significant decrease of fT 3 and TSH serum concentrations (respectively 2.36±0.79 pg/ml and 0.76±1.12mU/l) according to the control group (respectively 3.28±0.61 pg/ml and 0.95±0.46mU/l). Non-survivors had significantly lower TSH serum level (0.37±0.62 mU/l) in comparison to survivors (1.27±1.45 mU/l) in spite of very similar fT 3 serum level (respectively 2.45±0.87 pg/ml and 2.22±0.66 pg/ml). It could mean that there were disturbances in the pituitary-thyroid axis function in patients who did not survive. Our study did not show any correlations between thyroid hormones serum concentrations and APACHE II and APACHE III scores, ALI or ARDS .

CONCLUSION. This study shows that low TSH serum level may be useful as a significant prognostic factor of death in patient with septic shock especially with low fT 3 serum level. The results also suggest that ESS could be a consequence of pituitary TSH realising disturbances.

Eklund A 1 , Leppäniemi A 2 , Kemppainen E 2 , Pettilä V 1 1 Intensive Care Unit, Division of Anesthesiology and Intensive Care Medicine, 2 Division of Gastroenterological Surgery, Department of Surgery, Helsinki University Hospital, Helsinki, Finland

A few studies have reported that low doses of hydrocortisone (HC) may rapidly improve hemodynamics and reduce the time to vasopressor cessation in septic shock, but none has focused on this effect in acute pancreatitis. We therefore performed this study to assess the effects of HC on catecholamine-dependent shock among patients with severe acute pancreatitis.

Retrospective, case controlled study among 10 patients with severe acute pancreatitis. The control group comprised 11 patients with the same severity of circulatory shock according to the norepinephrine support required.

The patients in the HC group were weaned off norepinephrine in a significantly shorter time (61 h in HC group vs. 141 h, p=0.016). The HC group received significantly less norepinephrine (area under curve of norepinephrine dose, p=0.041). The reduction in norepinephrine dose was comparable at 24 h being -0.051 (-0.208 to 0.022) mg/kg/min in the HC group vs. -0.026 (-0.150 to 0.030) mg/kg/min in the controls (p=0.307), and at 48 h with respective figures of -0.206 (-0.317 to -0.102) mg/kg/min and -0.103 (-0.178 to 0.029) mg/kg/min (p=0.072), from the start of HC administration. The reference point for the control group was the time point at which the dose of norepinephrine exceeded 0.3 mg/kg/min.

Low doses of HC seem to shorten the time to vasopressor cessation and rapidly reduce the need for norepinephrine support in patients with hemodynamic shock associated with severe acute pancreatitis.

Huettemann E 1 , Bernhardt A 2 , Schelenz C 1 , Reinhart K 1 1 Anaesthesiology and Intensive Care, University hospital, Friedrich-Schiller-University Jena, Jena, 2 Surgery, Bundeswehrkrankenhaus, Leipzig, Germany

In patients with septic shock low doses of hydrocortisone(HC)have been shown to reduce vasopressor requirements and improve survival (1). However, there are only limited data whether myocardial function is affected by administration of low dose HC.

After approval by the local ethical comittee 10 patients with septic shock (mean age: 46,9 ± 19,7 (18 -74) years, APACHE II score: 21,5 ± 2,4 (18-25) and a mean norepinephrine (NE) dose of 1,4 ± 0,7 (0,6 -2,5) ug/kg/min were prospectively studied before and after administration of HC(100 mg bolus followed by continuous infusion of 200 mg/24 h) and reduction of NE requirement of at least 50 % (control). Following parameters were assessed by transesophageal echocardiography: Fractional area change (FAC), preload adjusted maximal power (PAMP), tei-index, transmitrale flow profile (E/A), and isovolumetric relaxation time (IVRT). Furthermore, cardiac index (CI), intrathoracic blood volume index (ITBVI), mean arterial pressure (MAP) and heart rate were measured.

After administration of HC, NE requirement fell significantly from 1,4 ± 0,7 to 0,2 ± 0,2 mug/kg/min (Wilcoxon test, p<0.01) after 45,5 ± 28,6 hours. Parameters of myocardial function and haemodynamic parameters did not change significantly except for heart rate (109 ± 6 vs. 89 ± 17 ; p<0.05)( Table 1) .

Before 

At our surgical ICU, HCN was mainly administered to patients with severe or prolonged dependency on noradrenalin. In our protocol HCN was started in a dose of 3x100 mg iv. Fludrocortisone was not used. The start and stop dates of HCN, as well as the HCN dose were recorded. All serum sodium (normal range 132-144 mmol/l) and 24-hour urinary sodium values were collected. No special protocol for preemptive correction of serum sodium was used. In the case of hypernatriemia the patients infusion was adjusted to glucose 5% or glucose 2.5%/NaCl 0.45%. Means were compared with Students t-test.

Over a 2-year period 39 patients with a mean age of 63 +/-16 years were studied. 15 patients (38%) died at the ICU. Hydrocortisone was administered over a median period of 5 days (IQR 3-8). In the 5 days following HCN administration, 33 (85%) patients displayed a serum sodium above 144 mmol/l. At the end of HCN administration serum sodium had risen to a mean of 144.8 mmol/l. It gradually decreased to 136.5 mmol/l over 12 days (P<0.01) after discontinuation of HCN. Urinary excretion of sodium displayed the opposite pattern with an initial decrease of excretion after the initiation of HCN and later restoration of excretion to levels > 150 mmol/day.

In accordance with its mixed mineralocorticoid and glucocorticoid action, HCN administration indeed induces mild hypernatremia. The mechanism appears to be sodium retention, because sodium excretion decreases at the same time. In order to avoid hypernatriemia as a result of HCN, sodium levels must be monitored and timely restriction of sodium administration should be considered (2) . . Intensive insulin treatment reduced mortality in critically ill patients admitted to a surgical intensive care unit [2] . The aim of the present study is to describe a possible association between hyperglycaemia and mortality in critically ill patients without known diabetes.

All adult patients admitted to the multidisciplinary ICU in a 6-month period were consecutively included. Patients with < 2 days in the ICU or with known diabetes were excluded. Altogether, the study included 135 patients, 97 surgical and 38 medical. Insulin was administrated when blood glucose level >12 mmol/l. A registration form was fulfilled for each patient, including demographic data, reason for admission, first day APACHE II score, daily maximum blood glucose level and death. Patients were classified into three groups according to maximum blood glucose level during the stay. The possible association between blood glucose level and mortality were described with an odds ratio from a multivariate logistic regression model. 

There is increasing evidence that low dose hydrocortisone (HC) is beneficial in patients with septic shock, especially in those with adrenal insufficiency. On the other hand HC may be harmful in patients (pats) with normal adrenal reserve. The exact mechanisms on how HC may improve the outcome of pats with hyperdynamic septic shock is not known. Recently HC was shown to attenuate the immune response. We conducted a single centre randomised controlled trial to study the effect of low dose HC on hemodynamics and cytokine response.

METHODS. 41 pats with hyperdynamic septic shock according to the consensus criteria were included in the study and randomised to receive hydrocortisone (0.18 mg/KG body weight/hr) or placebo. After a short synachten test study medication was started. Time to cessation of vasopressor support was documented. The SOFA (sequential organ failure assessment) score was performed daily. Blood for cytokine measurements was drawn before medication was started and consecutively on day 1, 3, 5, 7, 9, 14 and 28.

Of all pats 72% fulfilled criteria of adrenal insufficiency (i.e. a rise in cortisol plasma level of < 200 nmol/l after 0.25 mg ACTH). Time to shock reversal was significantly (sig.) shorter in the HC group compared to placebo (53 hrs.vs. 120 hrs; p< 0.02). After 48 hrs sig. less pats were in septic shock in the HC group (9 pats) compared to placebo (19 pats) (p< 0.05). This hemodynamic effect was more marked in the pats with adrenal insufficiency compared to pats with normal adrenal reserve, though not statistically sig.. Also, morbidity (as assessed by the SOFA score) was improved in the HC treated pats (after 48 hrs: 8 vs. 12; p< 0.05). Mortality was not different between the two groups (p= 0.

2) The level of the proinflammatory cytokine IL-6 was sig. lower during days 1 and 5in the pats treated with HC compared to placebo (p< 0.05).

Although mortality was not different between the two groups, low dose HC did sig. improve hemodynamics (time to shock reversal) in pats with hyperdynamic septic shock. Also, morbidity and cytokine response were positively influenced. This beneficial effect was more evident in those pats with impaired adrenal reserve, however, not sig. Therefore, larger multicentre trials including cytokine measurements are needed.

Angstwurm M W A 1 , Rashidi-Kia A 1 , Bidlingmeier M 1 , Schopohl J 1 , Gaertner R 1 1 Medizinische Klinik, Intensive care unit, Munich, Germany

In patients with severe illness adrenal insufficiency is often suspected and treatment with hydrocortisone and fludrocortisone has been shown to decrease mortality. However the pathophysiology of an adrenal failure is not understood.

Using commercially available essays, the steroid hormones progesterone, 17-OH progesterone, cortisole, testosterone, dehydroepiandrostenedione and 17-estradiol were determined before, 30 and 60 minutes after stimulation with 250g cosyntropin. The underlying admission diagnosis grouped patients in septic (n=43, 5 women), cardiogenic (n=22, 9 women) shock or control (n=34).

At baseline septic and cardiogenic patients showed similar cortisol levels (21 and 21g/dl) higher than control (15g/dl, p<0.05). Progesterone was increased 4-fold (p<0.001) in septic (1.2ng/ml) and cardiogenic shock (1.1ng/ml) compared with control (0.3ng/ml). In addition 17-OH progesterone was increased in both groups of patients compared to control (p<0.05). There were no correlations between steroid hormones and scoring systems or laboratory signs of infections. After stimulation, testosterone, 17-estradiol and DHEAS remained constant, whereas progesterone and 17-OH progesterone increased (p<0.001) in all groups without significant difference. In control or cardiogenic patients stimulation leads to significantly increasing values of cortisol (p=2,15E-12 and p=0.04), in patients with sepsis the increase of cortisol (p>0.1) was blunted, however. This diminished cortisol stimulation was independent of the use of sedatives. In cardiogenic patients the increase in cortisol levels after stimulation was similar to control (7g/dl) not influenced by increasing dosage of catecholamines but in septic patients the increase was blunted espcially in patients with high catecholamines. The increment of serum cortison after stimulation in septic group was correlated inversely with baseline progesterone and 17-OH progesterone but not with baseline cortisol levels.

At baseline, patients with septic or cardiogenic shock had higher progesterone, higher 17-OH progesterone but only slightly elevated cortisol levels compared to control. Septic patients showed diminished response to cosyntropin stimulation regarding cortisol levels despite a normal increase of cortisol precorsors progesterone and 17-OH progesterone. This impairment of cortisol synthesis at the level of the enzymes 21-hydroxylase or 11-hydroxylase should impair the aldosterone synthesis as well.

We enrolled 105 consecutive MODS patients into the study and assessed heart rate variability (HRV), baroreflex sensitivity (BRS) and chemoreflex sensitivity (CRS) as markers of AD according to the international standards. The cohort of patients was divided into 3 subcohorts CONCLUSION. According to our results we conclude that the AD in MODS is mainly attributed to the disease serverity that superimposes upon the potential effects of age. These results are in keeping with those from lost of age effects after myocardial infarction. The elderly patients seem not to have an additional risk by an especially pronounced AD. can induce hypothermia in certain conditions, although the mechanism is poorly understood. We investigated the changes in extracellular concentrations of serotonin (5HT) and dopamine (DA) at the level of the pre-optic region of the hypothalamus in freely moving rats during endotoxemiainduced hypothermia.

Pyrogen-free Wistar rats weighing 250 to 300 g were anesthetized for the stereotactic placement of a microdialysis probe in the left pre-optic area and of a thermistor probe in the right frontal lobe. A subcutaneous thermistor was implanted in the interscapular region to measure skin temperature. The following day the freely moving animals were randomized into two groups: the LPS group and the control group. Dialysates were collected every 15 minutes for a period of 6 hours. Temperatures were noted every 15 min. After baseline collections during 90 min the LPS group received 4 mg/kg LPS of E. coli type 055.B5 in 0.5 ml saline i.p. The control group received 0.5 ml of saline i.p. Levels of 5HT and DA were analyzed using reversed phase liquid microbore LC with electrochemical detection. In a second experiment haloperidol was administered in the 3rd ventricle 1 hour before LPS injection RESULTS. All animals receiving LPS developed mild hypothermia (average -1.4 °C) and a more pronounced drop in skin temperature. Serotonin levels were not influenced by the LPS administration. A significant increase of extracellular DA was seen starting 30 min after LPS injection and lasting for 210 min. The average maximal increase was 1295 percent of the basal level. The LPS induced hypothermia was partially antagonised by the DA D2 antagonist.

We found an increase in the extracellular concentration of dopamine but not of serotonin in the pre-optic region of the anterior hypothalamus during LPS induced hypothermia. The dopaminergic neurotransmission is probably not the sole player in this process since a DA D2 receptor antagonist only partially reversed the observed effect on brain temperature.

Lewejohann J C 1 , Muhl E 1 , Bruch H P 1 1 Surgery, Universitaetsklinikum Schleswig-Holstein Campus Luebeck, Luebeck, Germany

. Autonomic dysfunction appears frequently in patients at the ICU, especially in sepsis. Heart rate variability (HRV) is a window on central autonomic regulation which enables to investigate the activity of the autonomic nervous system. This phenomenon is caused by oscillation in the interval between consecutive heart beats. Sedoanalgesia results in a decrease of HRV and is proposed as method to estimate the depth of sedoanalgesia. The aim of our study was to estimate the extent of the autonomic dysfunction in patients with septic shock.

We investigated 22 patients (mean age 46), 9 with severe sepsis and catecholamine administration and 13 without (control group) in an observational study who received a continuous sedoanalgesia (Midazolam®, Fentanyl®) to achieve tolerance for mechanical ventilation corresponding to a Ramsay-Score between R2 and R3 at a SICU. Dosage of sedoanalgesics was unchanged during 24h-registration with a flash-memory recorder (Elamedical, Munich) . T-test was used for statistical analysis (SPSS®). 

Sedoanalgesia results in a well known decrease of the autonomic tone corresponding to a downregulated HRV. Patients with severe sepsis show a marked impairment of HRV in comparison to a control group with an equal level of sedation and moreover seems to loose their day-night-rhythm. The plain impairment of HRV and in particular the loss of circardian rhythm is not only explainable with the administration of catecholamines. Reduced HRV as a marker of autonomic dysfunction reflects the loss of biological oscillation in patients with severe sepsis. TBARs exhalation has been reported in numerous inflammatory lung disorders in adult patients, including adult respiratory distress syndrome, pneumonia and chronic obstructive pulmonary disease. TBARs are recognized as end products of polyunsaturated fatty acid peroxidation, however, they are also formed during oxidative injury of DNA, proteins and carbohydrates (1). In this study we tried to investigate the outcome of very low birth weight (VLBW) neonates suffered from respiratory distress syndrome (RDS)using the concentrations of exhaled TBARs.

METHODS. 21 intubated VLBW neonates suffered from RDS were enrolled into the study. Mean gestation age was 28,3 weeks (26 -30), mean birth weight 1076 grams (780 -1430). All neonates were ventilated using IMV or PRVC mode of ventilation. EBC was collected by cooling the additional exipiratory tube for 60 minutes. During the collection of EBC, humidification of inspiratory gas was switched off. TBARs were measured by a spectrophotometric assay; readings were expressed in micromoles using the regression equation.

RESULTS. 6 of total 21 patients developed bronchopulmonary dysplasia (BPD). TBARs concentrations in EBC obtained from these patients were higher then in 15 neonates who did not suffered from BPD, and the difference was statistically significant (p<0,01). Mean concentration of TBARs in EBC from patients who developed BPD was 0,825micromole, in contrast, TBARs levels of 11 from 15 patients who did not developed BPD were below the method sensitivity.

The results show a correlation between elevated TBARs concentrations in EBC of VLBW neonates suffered from RDS, who developed BPD. We speculate that TBARs concentrations may be useful as a biomarker of lung injury in newborn infants suffered from RDS, but further sudies are needed. 

The multiple organ dysfunction syndrome (MODS) is the consecutive failure of several organ systems after a trigger event like sepsis or cardiogenic shock with a high mortality of up to 70%. Autonomic dysfunction (AD) may substantially contribute to the development of MODS [1]. Our study aimed to characterize the AD by several techniques and to check the accuracy of AD in predicting 28 day mortality (28DM) and in-hospital mortality (IHM).

We enrolled 90 consecutive MODS patients into the study and assessed heart rate variability (HRV), baroreflex sensitivity (BRS) and chemoreflex sensitivity (CRS) as markers of AD according to the international standards (summary in [2] ). The patients were followed up for 28DM and IHM. A MODS was definied by an APACHE score of 20 or above. 5 patients were excluded on account of technically inappropriate Holter-ECG´s.

Total mortality after 28 days was 34% (29/85, range 3-28 days) and IHM 47% (40/85). Mean hospital survival time was 31.1±23.8 days (range 3 -127 days). HRV parameters lnTP (28DM: CHI squared [CHI]=5.1, p=0.03/ IHM: CHI=4.6, p=0.04), lnSDNN (CHI=4.0, p=0.04/ CHI=3.4, p=0.04) and lnVLF (CHI=7.6, p=0.006/ CHI=6.4, p=0.01) were the best predictive parameters in univariate analysis. Sensitivity (SE) and specifity (SP) for VLF to predict 28DM and IHM were assessed across a range of cutoff values using ROC curves and the best prognostic cutoff for survival status was defined as that whith the highest product of SE and SP. Using the optimal cutpoint of VLF (3.85 lnms2) for illustrating the cumulative survival (CM, Kaplan-Meier-survival analysis) there was a CM of 0.8 vs. 0.56 (28DM, p=0.006, first group: values above and second below the cutpoint). Kaplan-Meier-analysis for IHM also revealed a significant difference in CM (0.57 vs. 0.31, p=0.01).

CONCLUSION. This is the first study providing evidence that the autonomic function of MODS patients is blunted according to the "uncoupling" hypothesis of MODS development and that this attenuation has prognostic implications. 

Saladi S M 1 , Taylor N 2 , Thorburn K 1 , Van Saene H F K 2 1 Paediatric Intensive Care Unit, 2 Department of Microbiology, Royal Liverpool Children's Hospital -Alder Hey, Liverpool, United Kingdom

The control of infections caused by methicillin-resistant Staphylococcus aureus (MRSA) has become a global and difficult task. The MRSA carrier state has been identified as an independent risk factor for endogenous infection. Screening for and treating gut carriage of MRSA may control secondary endogenous MRSA infections. The aim of this prospective observational single centre study is to control secondary endogenous infections due to MRSA, by detecting and eradicating MRSA carriage using enteral vancomycin in critically ill children.

On our 20 bed paediatric intensive care unit(PICU), all microbiological data on patients staying >4 days is collected prospectively. All children in the PICU have surveillance (oral and rectal) swabs on admission and then twice weekly. The children who are identified to carry MRSA are treated with enteral vancomycin for 5 days along with conventional infection control measures.

Over a 3 year period (Mar 99 -Feb 02) data was recorded on 948 children staying >4 days in the PICU. 26 children were MRSA positive (2.7%); 9 cardiac, 8 medical, 5 burns, 4 surgical. Median age and stay for MRSA children was 5 months and 11 days respectively. 23 children were carriers and 10 received enteral vancomycin. 16 patients imported MRSA into the unit, whilst 7 patients acquired it, with a median time to acquisition of 7 days. MRSA carriage was abolished in 80% receiving enteral vancomycin after a median of 5.5 days. 8 children (0.8%) had 19 infections -the majority were wound infections. Of the 19 infections 4 were primary endogenous, 14 exogenous, 1 unable to evaluate. There were no secondary endogenous infections. There was no difference in the median Paediatric Index of Mortality [PIM] score between the children with MRSA and those without. 2 children who had MRSA died. Mortality was similar between MRSA children and those without. Neither vancomycin resistant enterococci nor vancomycin intermediate resistant Staphylococcus aureus were isolated.

Eradication of MRSA carriage, in a minute 1% of the population, controlled transmission and resulted in the absence of secondary endogenous infections. There were no side effects in terms of resistance to enteral vancomycin.

16th Annual Congress -Amsterdam, Netherlands -5-8 October 2003 S157

Stark M J 1 , Ferguson S 2 , Madar J 1 1 Child Health, 2 Intensive Care, Derriford Hospital, Plymouth, United Kingdom

The recognition and management of the critically ill child outside the intensive care setting is of critical importance.Delay leads to potentially avoidable morbidity and mortality. This abstract describes the development of a paediatric early warning system (PEW) within our hospital.

PEW comprises a group of key physiological observations. These are classified along the A,B,C, and D approach. In addition a number of clinical conditions activate the system irrespective of physiological state. Activation leads to senior paediatric/intensive care review and initiation of treatment. A prospective evaluation of PEW through case note review for all consecutive admissions occured over a 4-month period. Children attaining one or more PEW criteria were selected to determine if critically ill children were being correctly identified with subsequent activation of the PEW system, and if this led to a change in clinical management RESULTS. All 808 admissions in the time period were reviewed. 50 (6.1%)attained 1 or more PEW criteria. PEW was activated in 35 (70%). The mean time from attaining PEW criteria to activation of the system was 0.8 hours (0-6 hours). Commonest site of activation was Children's High Dependency 23/35 (65.7%). Median time from PEW activation to senior clinician revue was 0.9 hours (0.1-7 hours). In 20 (57.1%) of the activated cases, management changed following senior review. The remaining 15 (42.8%) where merely highlighted as potential problems with no intervention.

CONCLUSION. PEW is a model for the inproved recognition and early management of the sick child. Although only a small proportion of paediatric admissions led to its activation, in more than half, early senior clinician involvement led to a change in clinical management and may have prevented morbidity or mortality 601 PEDIATRIC INTENSIVE CARE OUTCOME:BEYONT MORTALITY. 2) .We examined the alterations in the scales after PICU stay and their relationship with other parameters of pediatric critical care illness.

We examined prospectively 120 consecutive PICU patients, 70M/50F, aged 1mo to 17y. Data collected: Demographics, PCPC and POPC scores at admission and discharge, Pediatric Risk of Mortality Score (PRISM III 24h), Mechanical Ventilation days (MV), Length Of Stay (LOS), Hospital Stay (HS). DPCPC and DPOPC scores (the decline in scales scores between discharge and admission, group 0 = no alteration, group 1 = 1 scale decline, group 2 = 2 to 5 scales decline) were calculated and related to the above parameters. Statistical analysis: Kruskal Wallis test, p<0,05. Values are mean ± SD.

. 15 patients died in the PICU.72 pts were in DPCPC group 0, 25 pts in DPCPC group 1 and 23 pts in DPCPC group 2. Groups 0,1 had lower PRISM scores (6,4 ± 4,09 and 8,08 ± 6,5 vs 19,1 ± 12,2 p <0,001), and shorter PICU stay(LOS 5,7 ± 5,5 and 9,6 ± 11,9 vs 22 ± 40,4 p <0,01) compared to group 2. Group 0 spend less days in MV compared to group 2 (4,1 ± 4,3 vs 16,4 ± 31,65). 36 pts were in DPOPC group 0, 57 pts in DPOPC group 1 and 27 pts in DPOPC group 2. Groups 0,1 had lower PRISM scores (7,02 ± 4,5 and 6,1 ± 3,4 vs 18,5 ± 12,5 p<0,001)), and shorter PICU stay (LOS 7,7 ± 8,1 and 6,1 ± 7,8 vs 20,4 ± 38,26 p<0,01). Group 1 spend less days in the hospital compared to group 2 (15,7 ± 16,38 vs 31,46 ± 39,9, p<0.01).

CONCLUSION. DPCPC and DPOPC groups 0 and 1 had lower illness severity and stayed in the PICU shorter compared to DPCPC and DPOPC group 2. Cognitive impairement (DPCPC) seemed to have more impact on MV whether functional morbidity (DPOPC) seemed to had more impact on HS. 

To evaluate the diagnostic value and to delineate serum profiles of interleukin-6 (IL-6), interleukin-1 receptor antagonist (IL-1ra) and procalcitonin (PCT) we analyzed their concentrations in newborns with confirmed congenital pneumonia (CP).

METHODS. IL-6, IL-1ra and PCT were measured in 10 newborns with CP[gestational age (median and interquartile range): 35(31-36) weeks] and in 35 others [33(30-34)weeks] without symptomatic infection admitted to NICU. The studied markers were analyzed together with CRP, leukocyte(WBC) and neutrophils count at admission and after 24 and 72 hrs. The diagnosis of CP was based on radiological examination and clinical symptoms.

At admission the values of analyzed parameters did not differ between the groups with the exception of the higher neutrophils count in newborns with CP (p=0.02). The difference in the neutrophils count was also noted on the following days. The concentration of IL-6 has increased after 24 hrs only in the patients with CP while in the other patients has declined. The difference at this time was significant [135(103-315) vs 24(10-80) ng/ml respectively; p<0.02)]. 72 hrs after admission IL-6 concentrations were low and of the same level in the both groups. Similar profile of changes was noticed for IL-1ra [15.7(6.1-36.0) vs 2.9(1.2-8.5) ng/ml after 24 hrs -p<0.01; 8.0(3.5-14.4) vs 2.3(1.1-4.4) ng/ml after 72 hrs -p<0.02]. PCT values did not differ between the two groups in all time points and showed transient increase 24 hrs after admission. No difference in respect to serum CRP levels was found after 24 and 72 hrs. 9.6(6.4-33) 1.4(.7-3.6) 13.7(10-17) 9.1(5.1-12) 0.7(.6-.8) others 87(52-131) 7.4(3.1-22) 1.2(.4-3.7) 9.9(7-13) 5.0(3.3-8) 0.6(.4-1.2)

CONCLUSION. The predictive usefulness of serum IL-6, IL-1ra and PCT measurement seems to be limited shortly after delivery, but could be helpful -especially IL-6 and IL-1ra -on the following days for the diagnosis of CP.

Hanekamp M N 1 , Prins S A 1 , Van Dijk M 1 , Houmes R J 1 , Tibboel D 1 1 Pediatric Surgical Intensive Care, Erasmus MC-Sophia, Rotterdam, Netherlands

Membrane Oxygenation (ECMO)treatment protocols. However the pharmacodynamics of the applied drugs may be altered due to an expanded circulating volume and drug absorption by circuit materials together with altered drug elimination. Assessment of the effect of sedation and analgesia in ECMO patients remains limited and standards are lacking. We assessed efficacy of standardised protocols for sedation and analgesia in neonates on ECMO using the COMFORTscale and Bispectral Index Monitoring (BIS).

During the canulation procedure for VA-ECMO, patients receive morphine (50mcg/kg/hr) and midazolam(0.2mg/kg/hr). Directly after canulation morphine and midazolam infusion is stopped, to asses the neurological status of the neonate. Morphine and/or midazolam are restarted after standardised behavioural re-assessment. Neonates were monitored using a validated postoperative pain and sedation score, the COMFORT-scale, together with a BISmonitor. BIS measures the effect of sedative agents on the brain and computes the patients EEG to a single number from 0 to 98. Paired COMFORT and BIS values were assessed at regular intervals during 48hrs post-canulation for ECMO in 18 newborns.

The median postnatal age was 1 day(range 0-6). The median time at restart of midazolam (dosage 0-0,2mg/kg/hr) and/or morphine(dosage 0-20mcg/kg/hr) was 22hrs(range 3-72). 15min.before restart, 86% of the BIS values were>60. There was no difference (Wilcoxon rank test) between the BIS 15min. before and 15min. after restart, resp. a median of 75(range 20-98) and 65(range 32-98). 130 paired observations showed a median BIS of 56(IQR 44-73) and a COMFORT of 9(IQR 8-10). Looking specifically at body movements, in only 7% of all observations frequent small movements were observed. BIS values were>60. The correlation of BIS v.s. COMFORT(pearson), regardless of medication was 0.3(p=0.001,n=130). The withinsubject correlation between BIS and COMFORT varied from -0.42 to 0.99.

There is a significant but moderate correlation between the COMFORT and BIS in neonates on ECMO. Despite adequate analgesia and sedation, BIS values were highly variable. The intra-individual correlation between COMFORT and BIS shows a great variability. These preliminary data do not support the hypothesis that BIS can be used as a primary parameter to assess the effect of sedation and analgesics in neonates on ECMO. 

Sensitive, reliable and early parameters of bacterial infection are extremely valuable in diagnosis of nosocomial infections in neonatal intensive care unit. In this study prokalcitonin (PCT) and C-reactive protein (CRP) were evaluated for their diagnostic relevance in neonatal late onset sepsis. The acute-phase reactant CRP is the most common used biochemical inflammatory marker in neonatology, however its use has important disadvantage: it does not increase significantly until 24-48 hour after onset of inflammatory response. Prokalcitonin is a highly sensitive and specific early marker of bacterial infection used in neonatology from late nineties (1).

In this study I have analysed inflammatory parameters in 48 newborn infants admitted to the Neonatal Intensive Care Unit in University Hospital in Lodz, who suffered from nosocomial sepsis. 17 of them had Gram negative infection and 31 had Gram positive sepsis. They were sampled for PCT and CRP levels at the time of onset of signs and 24 hours later. CRP was determined by a nephelometric method and PCT by an immunoluminometric assay.

At the onset of Gram negative sepsis 14 from 17 contaminated newborns had significantly increased CRP levels and 15 of them had increased levels of PCT. After 24 hours 100% of them had elevated CRP and PCT levels. At the onset of Gram positive sepsis only 18 from 31 neonates with positive blood culture had increased CRP levels and 28 of them had elevated concentrations of PCT. This difference was statistically significant. After 24 hours 26 of them had elevated CRP and 31 (100%) had increased PCT concentrations -this difference was not significant.

Measurement of prokalcitonin concentrations is useful for early diagnosis of late onset sepsis in neonates and its diagnostic relevance may be superior to that of C-reactive proteine. 

The debate over colloid versus crystalloid as the best solution for intraoperative fluid resuscitation in neurosurgical paediatric patients is not resolved. The quality of postoperative recovery between colloid and crystalloid in paediatrics with neurosurgical diseases has not been well investigated. In addition, experience of Voluven use during neurosurgical operations in paediatrics is absent.

We investigated the effects of colloid hydroxyethyl starch Voluven 130/0.4 and crystalloid (0,9% NaCl) resuscitation on nausea and vomiting and on the postoperative patient recovery profile. 40 paediatrics from 3 to 12 y.o. undergoing major neurosurgical operations without clinical signs of intracranial hypertension were randomized to receive 6% Voluven in saline (group 1) and 0,9% NaCl solution only (group 2) on the basis of a fluid administration algorithm. The anaesthetic was standardized.

Hemodynamic targets included maintenance of arterial blood pressure, heart rate,and urine output within a predefined range. A postoperative mobidity survey was performed at baseline and daily after surgery. The amounts of study fluid (mean SD)administered were 750 410 mL in group 1 and 1650 840 mL in group 2, respectively (P<0.05, Voluven group versus 0,9%NaCl group).The colloid group had a significantly less frequent incident of nausea and vomiting,use of rescue antiemetics,severe pain,periorbital edema, and double vision.

We conclude that intraoperative fluid resuscitation with colloid solution 6% Voluven, when compared with crystalloid administration is associated with an improvement in the quality of postoperative recovery.

Helbling C 1 , Schmidt D 1 , Hassan C 2 , Bindl L 2 1 General pediatrics, 2 Oncology, Univeristy Childrens Hospital, Bonn, Germany

The survival in paediatric patients with acute hypoxemic respiratory failure (AHRF) has substantially improved with optimized management of ventilator strategies and standartized sepsis therapy. We tested the hypothesis, that also in immunocompromised patients, the subpopulation with the highest mortality risk outcome improved.

Retrospective data on etiology, clinical course and therapeutic interventions in immunocompromised patients with AHRF treated at our PICU since 1991 were systematically reviewed and stored to an ACCESS data base. Two groups were compared according to the treatment period 1991 -1996 (Group A, N=19) and 1997 -2002 (Group B, N=12) using the Chi-Square-test modified by Pearson and T-test as appropriate. 

It has been reported that ultrasound-guided cannulation of the internal jugular vein (IJV) is useful in infants 1) . However, ultrasound transducers with high frequencies for paediatric superficial vessels are not yet widely available. When the three-dimensional position of the IJV from the carotid artery (CA) is estimated, the success rate may increase even without ultrasonography. In this study, we measured the distance between the IJV and the CA (Dis), the width of the IJV (W) and depth of it from the skin (Dep) with an ultrasound scanner and evaluated whether they could be estimated by age, height and body weight (BW).

After obtaining institutional approval and parental informed consent, 50 paediatric patients (0-34 months, 47-94 cm, 2.5-12 kg,) undergoing congenital heart surgery were prospectively studied. After the trachea was intubated, the lungs were ventilated with pressure control (15 _} 1mmHg). Dis, W and Dep were measured using a 12MHz transducer with a SONOS 5500 ultrasound system (Philips Medical Systems, Andover, MA, USA). The measurement was performed at the end-inspiratory period at the level of the cricoid ring (0 o 

Zisovska E M 1 , Pehcevska N P 1 1 Neonatology, Obstetric and Gynecology Clinic, Skopje, Yugoslavia

Literature data show high incidence of Acute Respiratiry Distress Syndrome (ARDS) in septic newborns due to hipoproteinemia and low oncotic pressure, increased microvascular permeability and oedema formation. AIM: the goal of this study was to determine the correlation of hypoproteinemia with ARDS occurence and to compare the outcome of the septic newborns depending on the serum total protein levels.

In the study were included term newborns born on O&G Clinic in Skopje, with clinically and laboratory proven sepsis. Serum total protein levels were obtained second-daily for each baby with sepsis, beginning from the day two. The signs of ARDS were confirmed clinically, with blood gasses, and X-rays. As referral levels for all parametrs were taken those from Neonatology, Roberton 1999.

. 32 newborns with proven sepsis met the criteria for the study, 14 of which had low serum total protein levels. These formed the examined group, and the others (18) had normal serum total protein levels. In the examined group, in 10/14 ARDS occured with worse weight gain, longer oxygen needs, even mechanical ventilation, (mean 11±2 days), and higher mortality rate (2/14) . In the control group, septic newborns without hypoproteinemia, there was lower ARDS incidence (2/18), better weight gain, shorter period of oxygen needs (mean 7±2 days), and none of the newborns died.

Although these results show high predictive and prognostic value of hypoproteinemia in septic patients, prospective randomized trials of serum total protein levels are needed to establish whether this difference is statistically significant, and whether there is a causeeffect relationship between these two entities. 

Although acute renal failure due to acute tubular necrosis is common in ICU's, its pathogenesis is unclear and the site of tubular injury (proximal/distal) is unknown. The absence of reliable markers of tubular cell injury has hampered intervention studies. Glutathion S transferases (GST) are cytosolic enzymes. The alfa isoform is present only in proximal tubular cells, whereas the pi isoform is confined to distal tubular cells. Aim of the present study was to determine the extent and site of tubular injury as reflected by urinary GST enzyme excretion after cardiac surgery.

Urinary enzyme excretion and endogenous creatinine clearance were determined 0-4 hours and 20-24 hours after cardiac surgery in 43 consecutive patients. Urinary GST-alfa and -pi were measured by an ELISA as previously described (1). Data are expressed as mean±sem.

We have previously shown that GST alpha-and pi/creatinine ratios in healthy volunteers range from 0.12 to 0.75 ng/mmol and from 0.19 to 1.08 ng/mmol, respectively (2) . The patients in our study all had uneventful recovery after surgery, and none had evidence of acute renal failure (increase of serum creatinine > 25%). Urinary excretion of GST alpha was 1.1±0.3 and 0.6±0.1 ng/mmol, whereas excretion of GST pi was 2.5±0.7 and 1.7±0.7 at the early and late time point.

There was no major increase in the urinary excretion of GST alpha in our patients after cardiac surgery compared to healthy controls. In contrast, urinary excretion of GSTpi was elevated. Our study indicates that after cardiac surgery there is evidence of tubular injury, predominantly at a distal site, even in patients without evidence of renal failure. We hypothesize that measurement of these specific enzymes might be useful for detecting subtle cell injury, and might allow to define high risk groups and enable future intervention studies. 

Acute renal failure (ARF) is a frequent complication of critically ill patients in the intensive care unit (ICU) often leading to renal replacement therapy through haemodiafiltration (HDF). The aim of this study was to evaluate the epidemiology, outcome and prognostic factors in critically ill patients with severe ARF requiring HDF.

We retrospectively studied 197 consecutive patients treated with HDF over a 7-year period in a 16-bed adult multidisciplinary ICU. Demographic, biochemical, clinical, and outcome data were collected at ICU admission. ARF onset, initiation and duration of HDF were also recorded.

The incidence of severe ARF requiring HDF was 5.9% in the ICU. The mortality rate was 71.6%, clearly higher than overall ICU mortality during the study period (25.5%). Univariate analysis found numerous prognostic factors significantly associated with death such as Simplified Acute Physiology Score at admission, ARF onset and HDF initiation, number of Organ System Failure at ARF onset, delayed onset of ARF (ARF onset > 6 days after ICU. admission), mechanical ventilation, sepsis, shock and poor haemodynamic tolerance of HDF. Chronic renal failure, urine output > 1000 mL/day at HDF initiation, high serum creatinine concentration and high variation of serum creatinine concentration during HDF predicted favourable outcome. Multivariate analysis found 3 independent factors associated with fatal outcome: mechanical ventilation, sepsis and shock requiring vasoactive medication. In contrast, 2 independent factors predicted favourable outcome: Urine output > 1000 mL/day at HDF initiation (nonoliguric ARF) and serum creatinine concentration over 34mg/L (300micro mol/L) at ARF onset. An algorithm using the Chi-square Automatic Interaction Detector (CHAID) statistical method allowed the identification of patient groups with very different mortality rates ranging from 25 to 100%.

Severe ARF requiring HDF in our ICU was associated with a high overall mortality rate casting doubt over the effectiveness of HDF. However, our prognostic algorithm identified sub-groups of haemodiafiltrated patients with low mortality rates. In such patients, HDF appeared as an essential therapeutic contribution. 

Renal function was studied in experimental septic shock with different results (1,2). We were interested if urine biochemical changes reflect changes in blood.

Six pilot pigs were studied for 12 hrs after induction of experimental Gneg volume resuscitated sepsis (continuous live Pseudomonas aeruginosa intravenous infusion). Urine output, pH, osmolarity and Na/K extretion were studied on top of original experimental protocol (splanchnic perfusion/metabolism). Hourly diuresis was measured and urine sampled at T0, T2,T6 and T12. Data are median(range). Statistics: Friedman ANOVA and Wilcoxon matched pair test; p < 0.05 considered significant.

Main results are summarized in Table. Time 

Acute renal failure (ARF) is seldom a community-acquired disease but usually develops in hospitalized patients. Critically ill patients have the highest incidence (>20%), and is associated with a persistent high mortality in intensive care units (ICU) Objective:to assess the incidence and the outcome of Acute Renal Dysfunction /Failure in ICU.

Prospective study. All patients admitted to our ICU, were included, during 3 years. Basic demographic data were collected. We used renal SOFA score to evaluate renal dysfunction/failure.

A total of 823 patients were admited in our ICU during 3 years. 90 patients(11%) had ARF at admission, as diagnosed by a serum creatinine of 300 micromol/l (3.5 mg/dl) or more and/or a urine output of less than 500 ml/day. 121 patients had renal dysfunction (SOFA 1,2). Anesthesia and Intensive Care Department, 2 Endocrinology, Khmelnitsky Regional Clinic Hospital, Khmelnitsky, Ukraine

The hemodynamic derangements present in diabetic ketoacidosis are the results not only of profound volume depletion but also of the effects of increased production of vasodilating prostaglandins (PGs), principally PGI2. In animal and in vitro models, prostaglandin synthesis is increased during insulin deficiency [1].

We assessed the effects of short-term ketosis on the metabolic and hemodynamic variables of 10 IDDM patients free from long-term complications and of 9 normal control subjects after a 7-day randomized double-blind indomethacin (INDO) (50 mg q.i.d.) or placebo treatment period. Calf blood flow (CBF), postocclusive reactive hyperemia (PORH), and recovery half-time (an index of overall perfusion) after PORH were measured by plethysmography. Left ventricular and myocardial functions were also studied in each different condition during placebo and INDO treatment in IDDM patients.

During placebo treatment, the increase in CBF during ketosis was higher (1.75±0.29 ml / min / 100 ml muscle) than during INDO (0.85±0.17 ml / min) / 100 ml muscle; P = 0.007 

Van der Voort P H J 1 , Feenstra R A 1 1 Intensive Care, Medical Centre Leeuwarden, Leeuwarden, Netherlands

Maintaining a normal glucose level in ICU is beneficial [1] . The regulation of the glucose level and not the administration of insulin may be responsible for the beneficial results [2] . However, the long-term effects of high parenteral glucose intake are unknown. We studied the metabolic regulation of long-stay ICU patients and correlated metabolic regulation with outcome parameters. In addition we studied the effect of glucose regulation, insulin doses, and the amount of infused glucose on outcome parameters.

We performed a retrospective analysis of all the patients admitted to our ICU in 45 months. Patients treated for 7 to 30 days in the ICU, and availability of a complete medical record were included. We collected baseline characteristics, the amount of parenteral glucose and insulin and biochemical results. Plasma glucose level (PGL)was measured 4 times daily in most patients.

In patients with stable PGL the measurements were reduced to twice daily. The daily mean value was therefore a representative value of that day. T test was used to compare groups after logarithmic transformation to obtain a normal distribution.

. 273 patients were eligible. The mean age was 66 years; the mean APACHE II was 24.6. The mean daily PGL of all patients was 9.0 mmol/l. Mean PGL and mean insulin dose were related (r=0.66, p<0.001). Hospital survivors showed a lower mean PGL compared to nonsurvivors (p=0.04). For ICU survivors, such a relationship was not found. PGL lower than 8 mmol/l was associated with a lower mortality rate (p=0.023). In a multivariate linear regression analysis, none of the parameters of mean daily PGL, mean daily insulin dose or mean daily glucose infusion were related to duration of mechanical ventilation, or duration of ICU treatment. Both ICU mortality and hospital mortality were correlated, in a logistic regression analysis, to mean daily glucose infusion (p=0.001) and APACHE II score (p=0.001) but not to mean daily PGL or mean daily insulin dose.

High dose glucose infusion and APACHE II showed a relation with hospital mortality in contrast to glucose regulation and insulin doses. Future studies to the effect of glucose regulation on outcome parameters should include analysis of the amount of parenteral glucose administered. 

Meijering S 1 , Vogelzang M 2 , Ligtenberg J J M 1 , Nijsten M W N 2 , Zijlstra J G 1 , Van der Horst I C C 3 1 Intensive and Respiratory Care, 2 Surgical ICU, 3 Internal Medicine, Groningen University Hospital, Groningen, Netherlands

Hyperglycemia at admission predicts mortality in various stress situations. It has been proven that strict regulation of hyperglycemia is beneficial, especially in patients who need intensive care for more than 5 days [1]. Thus glucose regulation over a prolonged period is important. We recently found that mean hyperglycemia (MHG), i.e. the mean glucose level above a cut-off point divided by time, predicts mortality more precise than hyperglycemia at admission.

We performed a retrospective analysis of all eligible patients admitted to a medical ICU of an University Hospital to determine if MHG predicts mortality.

Over a two year period 678 men (56%) and 531 women (44%) were included. Mean age was 56±18 years. Median (IQR) ICU stay was 3 (2-5) days with a maximum of 51 days. Reason for admission was respiratory insufficiency in 27%, sepsis or multi-organ failure in 13% and post surgery in 23%. Median (IQR) hospital stay was 15 (6-31) days. Mean glucose on admission was 7.9±4.5 mmol/l, mean glucose was 7.5±2.9 mmol/l, MHG was 7.7±2.6 mmol/l. 235 patients (19%) died in the medical ICU and 298 patients (25%) died during hospital stay. In patients who died the MHG was significantly higher than in patients surviving the medical ICU, 7.5±1.9 mmol/l versus 8.8±3.7 mmol/l (p<0.01). The mortality rate in the quartiles of MHG were 16%, 20%, 25%, and 40% (p for trend<0.01).

Mean hyperglycemia is related with mortality in patients admitted to a medical ICU. MHG in the highest quartile is associated with a 40% mortality rate. If strict regulation of glucose metabolism in this subgroup of patients is feasible, a large beneficial effect could be obtained. 

Sramek V 1 , Dadak L 1 , Mickalova K 1 , Sneidr J 1 , Babakova I 1 1 Department Anesthesia and IC, St. Annas University Hospital, Brno, Czech Republic

Normoglycaemia seems to be a very important factor influencing ICU patients outcome (1). We evaluated the impact of implementation of therapeutic protocol for glycaemia control in the critically ill.

Evaluation of glycaemia and insulin therapy during five day test in 8-bed general ICU. Evaluation of compliance with the therapeutic protocol and glycaemic values obtained by two methods (biochemical analyser and glucometer).

Close correlation between analyser and glucometer glycaemic values were found (n = 109, r2 = 0.86; p < 0.01). Glycaemia measured with an analyser was higher by 0.46 + 1.12 mmol/l. Compliance of nurses with the protocol was 67% (71 out of 106 cases). In case hyperglycaemia was measured a significant decrease in glycaemia has been found in the following 6 hour period ((from 9.2 (8.0 -10.7) to 6.0 (5.0 -9.3); p<0.00001)). The most significant drop was measured between the 2nd and 6th hours after an insulin dose change (( from 8.0 (6.4 -11.2) to 6.0 (5.0 -9.0); p<0.01)). In case normoglycaemia (4-6 mmol/l) was measured (n=34) this was present only in 44% (15 cases) at the following check after 6 hours. Severe hypoglycaemia (< 3 mmol/l) occured in 5 cases during the study and only 1 of these values was measured with the analyser. The most frequent deviation from the protocol was caused by further correction of insulin dose after one and expecially 2 hours after previous insulin dose change. Glycaemia during the study was 6.3 (6.0 -8.7) and did not differ from glycaemia seen before and after the study ((7.3 (6.1 -9.1) and 6.0 (5.0 -8.7) mmol/l, respectively; NS).

The implementation of normoglycaemia protocol in the critically ill increases interest in this topic among ICU personnel. Seminars explaining the protocol including its importance and longer test period are necessary to reach significant therapeutic results. 

Strict blood glucose regulation in ICU-patients decreases both morbidity and mortality (1). Since blood glucose (BG) can change quickly in ICU patients, frequent BG samples are necessary to achieve timely normalisation of BG. Measurements of urinary glucose (UG) might be useful for early detection of hyperglycaemia if blood glucose is above the 'renal threshold'. In healthy subjects this threshold is 10 mmol/l (2). However, the relationship between BG and UG in critically ill patients is not known. We investigated urinary glucose concentrations in surgical ICU patients. We also compared BG with UG to determine the renal glucose-threshold. 

Over the past decades hyperglycaemia in critically ill patients has been regarded as normal and even beneficial to patients. The publication by van den Berghe et all (1)changed this perception and has created an awareness within the critical care society that stricter regulation of the patients blood glucose is needed. After the publication our unit implemented stricter rules for glucose management.

These, unwritten, rules where that patients glucose should be maintained between 4.5 and 7 mmol/L. The use of a strict protocol was not considered a necessity. 

Arterial blood samples were drawn from 30 patients of a MICU over the course of one week. All together 556 blood samples were analysed. Metabolic acid-base status was assessed by means of the model by Gilfix with base excess of free water (BE Na+ ), base excess of chloride (BE Cl-), base excess of albumin (BE Alb ) and base excess of unmeasured anions (BE UMA ) accounting for metabolic acid-base disorders caused by changes of free water, serum chloride, serum albumin and unmeasured anions, respectively. Standard base excess (SBE) was used as an overall measure of metabolic acid-base state. Data were analysed using linear regression assuming autocorrelation of error terms. Partial R 2 was computed for each variable, in order to compare the influence of BE Na+ , BE Cl-, BE Alb and BE UMA on variations of SBE.

Mean daily increase of SBE, BE Cl-and BE Alb was 0.95, 0.58 and 0.16 mmol/L, respectively. BE Na+ and BE UMA remained unchanged. Partial R 2 of BE Na+ , BE Cl-, BE Alb and BE UMA were 6%, 41%, 4% and 22%, respectively.

Progressive hypochloremic alkalosis is the main cause of a developing metabolic alkalosis in critical illness. 41% of the overall metabolic acid-base changes can be assigned to changes of serum chloride. Assessment of chloride-related acid-base disorders might be helpful in diagnosis, prevention and treatment of metabolic acid-base disorders in critically ill patients.

TSOLAKIDIS G F 1 , Drivakou A 1 , Anthopoulos G 1 , Sevastos N 2 1 ICU, 251 HELLENIC AIR FORCE GENERAL HOSPITAL, 2 University of Athens,Internal Medicine Dept, " Hippokrateion" HOSPITAL, ATHENS, Greece

There is a significant relation between the difference in serum and plasma potassium values and platelet counts (1) (2) . As a result factitious potassium serum levels can be measured. This is observed in ICU patients during everyday practice, since serum potassium is measured by routine biochemistry and plasma potassium is measured by the gas analysis machine. We related the difference in serum and plasma potassium levels of ICU patients to their platelet counts and compared them with normal volunteers.

We compared serum potassium (SK-a), plasma potassium (PK-a), serum/plasma potassium difference (SPD-a) and platelet count ( 

One hundred critically ill patients (75 men), having diverse admission diagnoses, with a median age of 50 years were enrolled in the present study. First, a morning blood sample was taken to determine baseline cortisol. Then, a LDST was performed: 1 mcg of synthetic ACTH was injected as a bolus through a central venous line and 30 min later a second blood specimen was obtained to measure stimulated cortisol. Patients having stimulated cortisol levels below 18 mcg/dL were defined as non-responders to the LDST.

Median values for baseline and stimulated plasma cortisol were 17.0 mcg/dL (range: 4.5-79.0 mcg/dL) and 23.5 mcg/dL (range: 6.2-80.0 mcg/dL) respectively. The median increment in cortisol was 4.9 mcg/dL (range: 0-21.3 mcg/dL). There was a significant correlation between baseline cortisol and stimulated cortisol (r=0.73, p<0.001). Overall, 20/100 patients (20%) were non-responders to the LDST. There were no differences between responders and non-responders in gender or age. Non-responders had lower baseline cortisol (12.1 vs. 18.2 mcg/dL, p<0.001), along with lower stimulated cortisol levels (15.3 vs. 25.4 mcg/dL, p<0.001).

Adrenal cortisol production following dynamic stimulation is inadequate in a substantial number of critically ill patients.

Until recently the treatment of liver failure consisted mainly of supportive care, or in selected cases, of liver transplantation. The elimination of accumulating hepatotoxic substances by albumin dialysis (MARS®) can theoretically lead to improvement of liver function. We report the treatment of 4 patients with liver dysfunction as evidenced by severe hyperbilirubinemia.

One patient had severe itching and cholestasis as a result of toxic hepatitis of unknown origin; 2 patients had hyperbilirubinemia due to small-for-size syndrome after adult living donor liver transplantation (ALDLT), and 1 patient had cholestasis secondary to a steatotic cadaveric liver transplantation. The patient with toxic hepatitis was initially treated unsuccessfully with steroids; in the transplant patients rejection and mechanical bile duct obstruction were excluded. Treatment was considered when bilirubin exceeded 15 mg/dL. Biochemical parameters of liver function were evaluated before and after each session.

A total of 17 MARS® sessions was performed. Apart from membrane leakage during 1 session no adverse events were noted. MARS® treatments resulted in a decrease of bilirubin with a median of 28 % per session (interquartile range 17-33%). The patient with toxic hepatitis had 6 sessions during which bilirubin decreased from 42.6 to 17.0 mg/dL, with improvement of complaints. After the last session, bilirubin again gradually increased to 28.3 mg/dL. In contrast to the situation before MARS® treatment, this was successfully treated with steroids. Two days after the last session 1 ALDLT patient had a rise of bilirubin, attributed to acute cellular rejection. The other ALDLT patient had a progressive recovery of liver function after MARS® treatment, and the patient with the cadaveric transplant was retransplanted because of persistent liver failure.

CONCLUSION. MARS® albumin dialysis was easy to perform and safe. It was effective in lowering bilirubin levels, and may therefore have lead to decreased hepatotoxicity and pruritus associated with cholestasis. These data suggest that therapy with MARS® may contribute to the management of patients exhibiting small-for-size-syndrome after ALDLT. The highly effective albumin dialysis possibly resulted in decreased levels of immunosuppressives and acute cellular rejection in 1 patient.

Patroniti N 1 , Bellani G 1 , Curto F 1 , Galbiati L 1 , Amigoni M 1 , Foti G 1 , Pesenti A 1 1 Anaethesia and Intensive Care, University of Milano-Bicocca, S.Gerardo Hospital, Monza, Italy

Aim of this study was to assess the measurement of pressure time product (PTP) obtained by the rapid interrupter technique, performed by means of a commercial ventilator, compared to PTP derived from esophageal pressure (P es ) measurement, during PSV.

We studied 13 no-COPD patients undergoing PSV by an Evita4 ventilator (Drager -Lubeck). We recorded: Flow, airway opening pressure (P aw ), and P es . To perform inspiratory and expiratory occlusions, the ventilator was driven by a PC (Medibus serial protocol). Occlusions, lasting 2 seconds, were performed at different inspiratory volumes (25 ml steps) in random order, along tidal volume range. Immediately after the occlusion P aw equals the alveolar pressure (P alv ). Further changes in P aw are due to patient's respiratory effort; finally the patient relaxes his muscles and a plateau is seen on the P aw trace, representing elastic recoil of respiratory system. The difference between P alv at the time of occlusion and plateau represent the pressure generated by the inspiratory muscles (P mus ). However, since the occlusion gives raise to a noise in the signal for a period of 50-100 ms, the value of P alv at the time of occlusion must be back extrapolated from the most linear portion of P aw after the noise ceases. A mean of 31.6 occlusions for subject were performed, for a total of 411 occlusions. In each patient PTP occl was computed as the area under the time course of P mus,occl . The measures of P alv , P mus , PTP, and maximal inspiratory pressure (P max ) obtained from the P es by standard computations were used as controls. data were analysed according to Bland and Altman, and by linear regression.

Rapid interrupter technique provided a good estimate of both P alv (P alv, occl = P alv, pes *1.05-2.5, r=0.84; 95% CI: +10.2 -5.9 cmH 2 O) and P musc (P musc, occl = P musc,pes *1+0.39, r=0.87; 95% CI:+5.7 -5.1). From the analysis of P musc time course, PTP (PTP occl =PTP pes *0.87+1.02, r=0.90; 95% CI: +2.3 -2.8 cmH 2 O*s) and maximal P musc value (P max,occl = P max,pes *1.3+0.23 + ; r=0.95; 95% CI +2.93 9.98) could be estimated.

Faltlhauser A 1 , Combe V 2 , Werner P 2 , Thomas A 1 , Wellnhofer E 2 1 Dept. for Anesthesiology, Central Armed Forces Hospital, Koblenz, 2 Dept. for Cardiology, German Heart Center, Berlin, Germany

For a valid data base of the in arterial waves contained informations about hemodynamic parameters an exact pressure signal is essential. A new tool which is able to correct system induced errors in resonance and phase shift (Online Correction System, "OCS") shall be evaluated in a lab investigation.

Analysis of seven clinically used pressure measurement kits. All systems were diagnosed with the Gabarith1)-technique, for absolute and phase error. We developed a special fourier analysis based correction data set. Now synthetic arterial wave forms (BIOTEC 601A) were applicated to the measurement kit with different frequencies (40-150 bpm) . The resulting measured wave forms were now modified online with the correction data set. Evaluation was done comparing original, measured and corrected wave form -especially for systolic, diastolic and dp/dt differences.

The kits showed typical reproducible specifications. All errors could be eliminated with the correction data set. It is interesting, that the absolute error and changes in wave form (dp/dt) is increasing with the heart rate.

Example 3 Systems at 90 bpm -error in comparison to original wave form Error (% to systolic systolic diastolic diastolic dp/dt dp/dt reference) corrected corrected corrected System 1 0,6 0,3 0,3 0,0 26,5 2,5 System 2 3,1 0,3 0,3 0,2 42,2 3,2 System 3 2,2 0,2 1,1 0,4 15,3 1,4

CONCLUSION. Even with big differences in signal conduction quality, OCS® was able to reduce error probability to levels below 3%. The frequency dependent errors result from configuration dependent phase shift in the systems and probably follow s certain algorithm. With OCS an unique system for online bedside correction of data sets from liquid filled pressure measurement kits is now available. This make further pulse wave analysis more reliable.

Human colonic bacteria produce gaseous methane, which is absorbed into the blood through the colonic mucosa and exhaled. This might form the basis of a test for colonic blood flow, as reduced pulmonary methane excretion could indicate colonic underperfusion. However pulmonary methane excretion also depends on a range of other factors. We undertook a small observational study to determine what influences pulmonary methane excretion in critically ill patients before proceeding to more detailed studies.

Daily measurements of pulmonary methane excretion were made on unselected Caucasian patients while they were artificially ventilated in our ICU. Mixed exhaled gas was analysed using a gas chromatograph with flame ionisation detection. The methane concentration was converted to excretion rate using the minute volume with appropriate corrections for temperature and water vapour content. Factors likely to affect the methane excretion (number of days ventilated, sex, age, weight, enteral feeding, surgical status, antibiotic use, vasoactive drug use and mean blood pressure) were recorded. Multiple logistic regression was used to determine the model which best explained pulmonary methane excretion.

The within-subject coefficient of variation was 10.3%. Twenty-six patients were studied for 53 patient-days. Pulmonary methane excretion ranged between 0.63 and 20.8 umol.min -1 (mean 2.72, SD 2.93). The best fit model to predict pulmonary methane excretion is given in the table 

Pulmonary methane excretion can be measured in critically ill patients. The rate of excretion may be altered by blood pressure and hence by inference colonic blood flow. However, the major determinant of methane excretion appears to be the presence or absence of enteral feed. This will have to be addressed if a practical test of colonic perfusion is developed.

Boret H B 1 , Meaudre E 1 , Goutorbe P 1 , Carré E P 2 , Cantais E 1 1 Intensive Care Unit, 2 IMNSSA, Sainte-Anne, Toulon, France

Haemodilution and hypertension more or less hypervolaemia are recommended to prevent and/or treat delayed ischemic neurological deficit after cerebral vasospasm. The purpose of this study is to determine the effect of increased cardiac index (CI) on ischemic status assessed by lactate/pyruvate ratio (L/P ratio).

METHODS. 6 patients hospitalized following high grade aneurysmal subarachnoïd haemorrhage (SAH) were monitored with bedside microdialysis (MD). The catheter was inserted into the brain parenchyma of the vascular territory most likely to be affected by vasospasm. MD levels of glucose, glycerol, lactate and pyruvate were analyzed (MD probe CMA 70, micro-injection pump CMA 106, CMA 600 analyzer, CMA/Microdialysis AB, Sweden). Continuous CI was measured using a PiCCO system (Pulsion®). Vasospasm, suspected on daily transcranial Doppler, was diagnosed on angiography. The hypertension-hyperdynamic-haemodilution therapy was then initiated to maintain a CI > 3.5 L.min-1.m-2. The occurrence of biochemical ischemia, defined by two consecutives measurements of L/P ratio > 30, was treated by increasing the CI with dobutamine to obtain normalization of this ratio.

3 of the 6 patients developed cerebral vasospasm. Among 2 of them, 6 episodes of biochemical ischemia were successfully treated by increasing CI. Even if a CI superior to 4 L.min-1.m-2 was associated with a significant decreasing of L/P ratio (p=0.0005), the threshold of 6 L.min-1.m-2 was the only one associated with normalization of L/P ratio, with a mean of 28.35 (p=0.0031).

These preliminary results seem to show that the recommended CI threshold of 3.5 L.min-1.m-2 to prevent focal ischemia after vasospasm is too low (1). Even if additional studies are mandatory, threshold for CI adjustment appears to be above 4 L.min-1.m-2. In addition, future prospective prognostic studies will have to define the acceptable upper limit of L/P ratio. In conclusion, one way to manage cerebral vasospasm after SAH might be to optimize CI using bedside biochemical monitoring. This monitoring is feasible in an ICU, and might provide a valuable tool for clinicians in charge of patients presenting SAH with vasospasm and focal ischemia. 

Klaus S 1 , Bahlmann L 1 , Heringlake M 1 , Gliemroth J 2 , Schmucker P 1 1 Anesthesiology, 2 Neurosurgery, University of Luebeck, Luebeck, Germany

Acute hypovolemia with subsequent tissue hypoperfusion has been suggested to be of major importance in the development of multi organ dysfunction. In critical staes of shock disorders on the macro-and microcirculatory level are responsible for subsequent cellular malnutrition. In intensive care treatment hemodynamic parameters still are the "targets" of treatment while it remains unclear, if complete restoration of the hemodaynamic situation represents an adequate treatment to the tissue. Aim of the present investigation was to correlate biochemical tissue monitoring with hemodynamic observation after acute hemorrhage shock and catecholamine therapy.

After approval of the local ethics committee 14 German landrace pigs (31,4+5,6 kg b.w.) under general anaesthesia and normoventilation were observed for hemodynamic parameters, global oxygenation and blood gas values (MAP, HR, CO, SvO2) for an observation period of 180 min. A CMA 60 microdialysis catheter was inserted into the adductor muscle for continuous measurement of interstitial lactate concentration. 7 animals were exposed to an acute blood loss (25 ml/kg b.w.) to a MAP of 30 mmHg without any therapy (C), while 7 animals after 60 min of shock were treated with continuous adrenaline infusion to a MAP of 60 mmHg (A).

. MAP in both groups during the shock period decreased from 64+11 to 32+6 mmHg, accompanied by a decrease of CO (3,6+1,2 l/min to 1,6+0,5 l/min) and the SvO2 from 75+10 % to 56+14%. Under therapy with adrenaline significant higher values for MAP (60+6 vs. 30+15 mmHg), CO (4,2+2,1 vs. 1,5+0,5 l/min) and SvO2 (75+11% vs. 57+21%) were observed compared to control (p<0,05). Interstitial lactate concentrations of the muscle did not show significant differences between both groups during the whole observation period increasing to 6,1+2,2 mmol/l without therapy, while 6,4+2,1 mmol/l were measured under adrenaline therapy (n.s.).

Clinically "successful" treatment of hemodynamic deficites with adrenaline did not show any advantage in lower tissue lactate accumulation compared to animals without treatment. Biochemical tissue monitoring as possible with microdialysis should be introduced into the observation of critical ill patients to additionally adjust therapeutical interventions to the effects on tissue metabolism.

Ramnarain D 1 , Braams R 1 , Leenen L P H 1 1 Surgery-Intensive Care, UMCU Utrecht, Utrecht, Netherlands

The primary goal in the treatment of patients with shock is to restore the tissue perfusion and oxygenation. Clinical endpoints of resuscitation i.e. bloodpressure, heart rate, urine output and oxygen saturation give an incomplete or even misleading picture. Tissue oxygen tension (ptO 2 ) reflecting tissue oxygenation may be a usefull endpoint of resuscitation, however in clinical setting ptO 2 measurements in subcutaneous tissue (ptO 2 sc) and muscle (ptO 2 im) are both used although a comparative study has never been performed.

In five critically ill patients ptO 2 sc and ptO 2 im were simultaneously and continuously measured using polarographic Clark-type electrodes (LICOX Catheter Measurement System, GMS) placed subcutaneous and in the m. biceps brachii of the upper arm. Collected data were stored in a bedside computer.

Three men and two women with septic shock n= 3 , trauma n=1 and non septic ARDS n=1 were included. The median age was 56(range 37-72), median APACHE-score on admission was 21 (range 18-31). Median duration of tissue oxygen measurements was 5 days (range 3-7). In three patients mean ptO 2 sc was higher than mean ptO 2 im ( 34 vs 28, 51 vs 40 and 42 vs 38 mmHg) while in two patients the opposite was the case (38 vs 28 and 42 vs 30 mmHg). PtO 2 sc as well as ptO 2 im values showed variation around the mean, but the variation was greater in ptO 2 im values ( Variance ptO 2 im:199 vs Variance ptO 2 sc:164). Curves reflecting changes in ptO 2 sc and ptO 2 im during the course of the illness showed an identical pattern and run in a parallel fashion.

Although differences between mean ptO 2 sc and ptO 2 im were found in individual patients a clear pattern could not be established. Based on the findings in our patients we conclude that because of the lesser variation around the mean, absolute values of tissue oxygenation are more reliably measured in subcutaneous tissue, while the measurement of trends in tissue oxygenation may be performed in subcutaneous as well as muscle tissue.

Blanch L 1 , Bernabé F 2 , Vatua S 2 , Lucangelo U 2 , López-Aguilar J 1 , Villagrá A 1 , Saura P 1 , Romero P V 3 

In early Acute Respiratory Distress Syndrome, elevated values of dead-space fraction are associated with an increased risk of death (1). The fraction of tidal volume corresponding to the exhalation of alveolar gas (V AE /V T ) is a computerized, physiologically based new index, easy to measure at the bedside and not influenced by different values of tidal volumes or positive end-expiratory pressure (2) . The objective is to evaluate the prognostic value (association with mortality) of different outcome and respiratory variables in patients with acute lung injury (ALI) receiving mechanical ventilation.

Twenty-five patients were prospectively studied. Simplified Acute Physiologic Score II (SAPS II), PaO 2 /FiO 2 , respiratory system compliance (Crs), and capnographic indices (Bohr's dead space, expired CO 2 slope and V AE /V T ) were measured at the admission and after 48 hours. Data were expressed as mean ± SD. Risk of death was assessed with receiver operating characteristic (ROC) curves. 

Rodrigues M G 1 , Chindamo A 1 , Salgado D R 1 , Campos T T P 1 , Paiva R N 1 , Verdeal J R 1 , Martins L C 1 1 Intensive care unit, Barra D'or, Rio de Janeiro, Brazil

Arterial cannulation is a very useful tool in the management of patients in mechanical ventilation or haemodynamic instability. However local complications are always a concern. Objective: To describe complications of different arterial cannulation sites correlating them with line obstruction, local and distal ischemia , infection and thrombosis.

A prospective, observational study of the arterial cannulations performed in a clinical and surgical intensive care unit from October 2001 to August 2002. Daily evaluations for catheter obstruction (dumping of waves, difficulty in draining blood) or local and distal ischemia (livedo reticularis, pale or cyanotic extremity) were done. Arterial Doppler scans were obtained 24h after catheter removal searching for partial or total obstructive thrombosis. 

The PiCCO monitor (PULSION Medical Systems) allows the assessment of the maximum quantity of blood contained in the heart, called the global end-diastolic volume index (GEDI). The PiCCO monitor also estimates the intrathoracic blood volume index (ITBI = GEDI + pulmonary blood volume) and the extravascular lung water index (ELWI), assuming that ITBI is 25% greater than GEDI. Several anatomical, mechanical and physiological factors may affect the relationship between the volume of blood contained in the heart and in the pulmonary circulation.

METHODS. The first 4 transpulmonary thermo-dye dilution measurements (COLDsystem) done in 48 surgical ICU patients were analyzed to compare the reference ITBI and ELWI to the estimated ITBIpicco (1.25 x GEDI) and ELWIpicco (intrathoracic thermal volume -ITBIpicco) and to investigate factors that may influence the relationships between reference and estimated parameters.

A total of 192 measurements were available for analysis. Overall, ITBI and ELWI were closely correlated with ITBIpicco (r = 0.94) and ELWIpicco (r = 0.96), respectively. The bias were not influenced by the weight, the body surface area, the body mass index, the ITBI, the cardiac output and the PvO2, but significantly correlated with the ELWI, the level of PEEP, the intrapulmonary shunt, and the PaO2/FiO2 ratio. 

The degree of sedation in critically ill patients is usually determined using clinically derived, subjective sedation scores [e.g. Ramsay Score (RS)]. We hypothesized that event-related potentials (ERPs) to auditory stimuli could reflect the electrophysiological analogue to clinical assessment of sedation.

In ten healthy volunteers ERPs were measured during stepwise increasing, clinical relevant levels of sedation (RS 2-4), induced by randomly either propofol (Pro) or a combination of propofol and remifentanil (Pro/Remi). Effects of remifentanil infusion alone (Remi) were tested during target controlled infusion (1, 2 and 3 ng/ml). Auditory evoked potentials at about 100 ms after the stimulus (N100) were measured by using a paradigm consisting of 40 trains of 4 stimuli separated by 12 s intervals.

Remifentanil did not affect ERP amplitudes and latencies. During both Pro and Pro/Remi-induced sedation, amplitudes of N100 decreased significantly and similarly as the level of sedation increased from Ramsay score 2 to 4 ( , prompting the development of a 4Fr 50 cm radial artery catheter for use with the PiCCO system (Pulsion Medical Systems). We have compared this catheter with a pulmonary artery catheter (Aortech Critical Care) and investigated whether the use of a shorter catheter might be possible.

We studied 18 patients undergoing coronary artery surgery. Measurements were made post-operatively. TPCO was determined using 20ml of iced injectate. PCCO was then recorded. Simultaneously, PACO was determined using 10ml of room temperature injectate. After 3 measurements the catheter was withdrawn by 5cm and the above measurements repeated. Further withdrawals were made until no measurement of TPCO was possible. Statistical analysis was by the method of Bland and Altman. 

Setting: A 12-bed, community-based, university affiliated closed surgical ICU. A formal 6-week long CC training program was established for newly hired PAs. The methods used included 1)PA Manual containing state-of-the art current journal articles related to surgical CC, 2)1-1 teaching at the bedside by the intensivist and the senior PA, 3)Completion of Fundamental Critical Care Support Course of SCCM, and 4) Protocols in the disease management.The evaluation process inluded 1) Competency Assessment forms completed by the attending intensivist and senior PA, 2) Live Performance Competency Assessment completed by the hired PA during last week of training, and 3)Successful completion of FCCS course. Only after successful completion of all of the above 3 evaluations that the PA was credentialled for designated procedures and case management, and allowed to work under gerneral supervision of the intensivist and take night call in the ICU. All PAs were recredentialled on an annual basis.

The training program in CC greatly improved hired PA's cognitive and procedural skills. The PAs rated the overall experience highly valuable in preparing them for the clinical practice of CC.

Utilizing PAs in CC delivery provides for innovative ways to provide costeffective and safe delivery of CC in the US. It represents a feasible solution to the current shortage of intesivists and likely to attract PAs to CC. CONCLUSION. An important predictor of sickness is RR2. RR is seldom recorded1. Using MEWS has made improvements on the observations recorded in all patient groups. It is only in areas where the Outreach service is regularly accessed that virtually all patients will have vital signs recorded.

Details of ME were collected over 9 days for HW charts and 17 days of CIS. Total prescriptions numbers were recorded. A novel QOP Scale was developed: A-C Excellent to adequate, D Quite poor e.g. an abbreviation used for drug name, signature omitted, E Moderately poor e.g. lack of information on prescription to adequately give drug appropriately, F Extremely poor e.g. wrong patient, wrong dose, wrong drug etc. ME were coded D, E or F. Prescriptions without errors were coded A-C.

. CIS was associated with a significant improvement in the incidence of A-C prescribing (p<0.01 Chi squ) (table 1). CIS was associated with a reduction in D and E prescribing (p<0.01 Chi squ). The difference in F prescribing was not significant. 

Prescribing is a key function of patient management in the ICU. A novel way to assess prescribing was used -focusing on Q rather than patient outcome from ME. QS CIS improved the QOP, with less 'quite poor' and 'moderately poor' prescribing. 

The study of the performance of one ICU is still setting up. The usual criteria, nowadays, is not enough to assure quality and safety. To analyse the performance of the ICU and to benchmark it with other ICUs could be an experience of great value METHODS. We register, in a daily basis, variation of the data studied, the mean value, the mean value of benchmarking (the same subject in 7 other ICUs),control graphics (with upper limit and lower limit of control covering a confidence interval of 99,7%);these information is in a ICU computer covering a period of 16 weeks; each three months we receive a report with more complete analyses and after 12 months there are a special report with consolidated information. The data are divided in management, clinical and risk indicators. Each group of indicators are composed of some data; for example risk indicators are composed of rate of patients with some infection, rate of accidental extubation, rate of pneumothorax by barotraumas, rate of pneumothorax by venous puncture, rate of pressure ulcers. The system has confidentiality, you only know your data and the benchmarking, but no the data from one specific ICU. The databank is domain of AMIB (the Brazilian Society of Intensive Medicine).We start using the QuaTI at the final of year 2001.There are 30 ICUs in the system at end of year 2002 RESULTS. We compare the performance of our ICU with ourselves along the time, each three months, each year and with data from benchmarking ICUs. With this information we have to know our comparative performance and we can decide if it's necessary to change some process in real and comparative basis CONCLUSION. The study of the performance of one ICU is difficult, complex and expensive. The better way to know one specific performance is to compare it with similar ICUs that have similar resources and case-mix. To compare a Brazilian with European or American ICU is not ideal and could give equivocal information that generates erroneous decisions. With QuaTI, we start a very interesting project that will create a Brazilian databank on ICU performance

In spite of new technologies and sophisticated monitoring, the intensive care has been distant from humanization. To know anxieties and the perception of the family members of patients at the ICU can optimize the improvement of the intensive care quality.

Medical students interviewed 45 family members of patients at the ICU of a university hospital, using a 12 query questionnaire the quality of care ( physicians and nursing ).

. 45 families answered the questionnaire; mean length of stay by the time of the interview was varying from 3 to 120 days; the most frequent complaint was the noise and reduced visit period ( only 30 minutes per day ) in 26%(12)of the answers. 12(26%) of the families reported that the patients have not complained of pain and 9(20%) did report pain, most of all of minimum intensity (66,6%) with quick relief after medication. There were 2(4,4%) complains of pain during blood exam sampling. Variations of temperature troubled 7(15,5%) of the patients. Visit period (30 minutes) was considered to be unsatisfactory by 18 (40%) of the families; 20(44,44%) suggested to amplify visit period and number of visitants. The medical attendance was qualified as "very good" by 24(53%) and "good" by 21(47%); the nursing attendance of qualified as "very good" by 12(26,6%) and "good" by 27(60%); unsatisfactory information provided by nursing was the complaint of 14(31,11%) families. 38(84,4%) families reported great hope in the treatment instituted at the ICU.

To know the anxieties of the families in the regard of treatment instituted in the ICU allows correction of mistakes and improvement the quality and humanization.

Gianesello L 1 , Pavoni V 1 , Paparella L 1 , Pavarin P 1 , D'Urso G 1 , Gritti G 2 1 U.O of Anaesthesia and Intensive Care, University-Hospital, Padova, 2 Dept. of Critical Medical-Surgical Area, Section of Anaesthesia and Intensive Care, Firenze, Italy

Early diagnosis of any alterations of quality of life (QOL) post-injury and specification of the type of alteration facilitating treatment of problems through rehabilitation, psychological support, occupational therapy is mandatory in all critically ill patients and particularly in burn patients(1).In recent years, the studies regarding QOL have proliferated and a number of questionnaires have been developing; one such instrument is the EuroQol-5D(EQ-5D) (2) .

This study was carried out on severe burn patients who were discharged from a polyvalent intensive care unit (ICU) of a University teaching Hospital from 1999 to 2001. All patients at admission were intubated a cause of inhalation or head burn involvement. The following data were collected: age, sex, percentage of total body surface area burned (% TBSA), degree of burn, location of burns, length of ICU stay (LOS). The QOL was evaluated using EQ-5D questionnaire that was administered to survivor patients by telephone by the same clinician six months after injury.

Sixteen adult burn patients were evaluated (9 male and 7 female); their mean age was 50.1±19.4; they had a mean of percentage of TBSA burned of 39.6±19.2, of III degree. Most patients were burned to the upper and lower extremities 12 (75%); burns to the head, face, neck were present in 9 (56.2%) patients. The mean of LOS was 15.6±14.9 days. Three patients died in the ward after ICU discharge, two died within six months follow-up period and two were lost to follow-up. Nine patients were interviewed. At time of interview the level of health of all patients was worse than previously to injury. EQ visual scale (VAS) score median was 50. Moderate and extreme problems in the five dimensions studied were present as follow: mobility (moderate 44.4%; extreme 0%), self-care (moderate 22.2%; extreme 33.3%), usual activities (moderate 66.6%; extreme 22.2%), pain/discomfort (moderate 66.6%; extreme 11.1%), anxiety/depression (moderate 44.4%; extreme 33.3%).

CONCLUSION. In our severe burn patients population, QOL is influenced by consequences of injury both in psychological and physical health. In this preliminary report, EQ-5D seems to be reasonably valid, reliable and responsive in burn patients. 

Wester J P J 1 , Bosman R J 1 , Oudemans-van Straaten H M 1 , Van der Spoel H I 1 , Zandstra D F 1 1 Intensive Care Medicine, Onze Lieve Vrouwe Gasthuis, Amsterdam, Netherlands

Traditionally, physicians are divided in headworkers and handworkers. Generally, the medical staff in a department of Intensive Care Medicine consists of physicians of different origins. We describe and analyse the contribution of headwork and handwork by intensivists, fellow-intensivists and residents in our Intensive Care Unit (ICU).

In the setting of an 18-bed level I closed format medical-surgical ICU in an university-affiliated teaching hospital, data on all procedures and notes performed by our medical team are prospectively collected in the Intensive Care Database (ICDB) and through an Intensive Care Information System (MetaVision®, iMDsoft, Tel Aviv, Israel) since March 1, 2001. We performed a descriptive analysis of the total number of procedures and notes. Unit of analysis was the mean number of procedures per day of employment per physician and the mean number of notes per day of employment per physician. Physicians were classified as headworkers or as handworkers. From March 1, 2001 until March 21, 2003 physicians worked in our ICU: 5 intensivists (permanent staff-members), 12 fellow-intensivists, and 12 residents. Twenty-one physicians were classified as headworkers (14 internists, 1 pulmonologist, 1 cardiologist, 5 emergency physicians), and 8 were classified as handworkers (7 anesthesiologists, 1 surgeon). The group consisted of 15 males and 14 females. The number of procedures per physician per day varied from 1.13-2.64. The number of notes per physician per day varied from 0.24-1.56. The scatterplot showed no real difference between the main activities of the two classical stereotypes of physicians.

CONCLUSION. The number of procedures and the number of notes per unit of time per physician were equally distributed between the two stereotypes of headworkers and handworkers. Physicians in Intensive Care Medicine have gone beyond the classical frontiers.

This study aims to determine the extent of correlation between arterial and venous pH; Base Excess and Lactate with a view to identifying whether venous samples could be used as an alternative to arterial values in the clinical management of acutely ill patients.

We prospectively compared 210 pairs of simultaneously obtained central venous and arterial blood gas values for pH, Base Excess and Lactate in a group of 31 surgical intensive care patients. Data were analysed using Pearson correlation and Bland-Altman plots.

For the entire group (n=210), the correlation between venous and arterial pH, Base Excess and Lactate were r_=0.878, 0.866, and 0.856 respectively (p<0.05). Median values for this group were pH, venous 7.32 (7.059-7.48) arterial 7.36 (7.1-7.52), Base Excess, venous -1.3 (-16.5-5.1), arterial -2.0 (-16.8-4.6 ) and Lactate, venous 1.0 (0.4-8), arterial 0.9 (0.3-8). Bias and limits of agreement are shown in Table 1 . 

Venous blood gas measurements provide a reliable reflection of acid base variables as compared to arterial samples. This suggests that they could be used to estimate the severity of acid base derangement in patients' without arterial access and also as a simple method of quantifying severity of illness.

Anglès O 1 , Baixas C 2 , Sanchez-Verlaan P 1 , Mucke F 2 , Galinier M 2 , Génestal M 1 1 Intensive care unit, Hôpital Purpan, 2 Cardiology, Hôpital Rangueil, Toulouse, France

We sought to assess the diagnostic value of BNP and NT-proBNP levels to identify cardiogenic origin in patients admitted in shock.

Twenty-one patients with shock were prospectively enroled in this multicentric study. All of them underwent clinical examination, EKG, chest X ray, cardiac echo and conventional blood tests, BNP and NT-proBNP dosage and Swan Ganz if necessarly. Patients were evaluated by 2 independant clinicians blinded to BNP and NT-proBNP results. Practitioner divided patients in 3 groups : no cardiac dysfunction (group 1), patient with cardiac dysfunction with a shock not related to their cardiac condiction (group 2) and cardiogenic shock (group 3).

For all conditions the correlation between BNP and N-T pro-BNP was strong (r=0,83; p<0,00001). These two markers were significantly elevated in patients with abnormal cardiac echo findings (BNP: 849 +/-428 pg/ml vs 199 +/-428 pg/ml, p<0,005; NT-proBNP: 28953 +/-54349 pg/ml vs 1106 +/-1055 pg/ml, p<0,01) and strongly correlated to increased capillary pulmonary wedge pressure (BNP: r=0,69, p<0,01; NT-proBNP: r=0,62, p<0,05). We noticed a proportional increase in both markers associated to the severity of cardiac dysfunction (BNP: 338 +/-345 pg/ml in group 1, 775 +/-375 pg/ml in group 2, 1063 +/-425 pg/ml in group 3, p<0,01; NT-proBNP: 1729 +/-1503 pg/ml in group 1, 7186 +/-3503 pg/ml in group 2, 58922 +/-74184 pg/ml in group 3, p<0,001). To detect cardiac abnormalities in patients with shock, sensibility (Se) of BNP levels higher than 163 pg/ml was of 93% and specificity (Sp) of 57%, with a positive predicting value (PPV) of 81% and negative predictive value (NPV) of 80% (AUC of 0,85). For a NT-proBNP level > 3669 pg/ml we found Se=92%, Sp=100%, PPV=100% and NPV= 88% (AUC of 0,97). Finally, in a patient admitted for shock a BNP level = 889 pg/ml allows to diagnose cardiogenic shock with a Se of 86%, Sp of 86%, PPV of 75% and NPV of 92% (AUC of 0,82), a NT-proBNP level =16541 pg/ml had a Se=86%, Sp=100%, PPV= 100% and NPV= 93% (AUC of 0,90).

CONCLUSION. This study shows that BNP and NT-proBNP levels assessment is a powerful, fast and non invasive tool to diagnose cardiac participation in patients admitted for shock. Moreover NT-proBNP appears to be a stronger predictive marker than BNP in assessing the reality of cardiogenic shock.

Butrov A V 1 , Gubaidullin R R 1 1 Department of Anesthesiology and Intensive Care, Russian Peoples' Friendship University, Moscow, Russian Federation

Reduction of cardiac index (CI) and gastrointestinal ischaemia occurs particularly because of intraabdominal hypertension (IAH) (1). Analysis of CI and intraabdominal pressure (IAP) may determine their participation in gastrointestinal ischaemia.

As intestinal failure can be accompanied by IAH (2) we studied 24 patients with diagnosis of intestinal obstruction. IAP and PgCO2-PaCO2 (kPa) were measured by TRIP NGS catheter and Tonocap monitor. CI was fixed by rheographic device. All measurements were carried out preoperatively and on the second day after operation. Cluster analysis was performed on the base of differences between IAP, CI and PgCO2-PaCO2.

Before operation in 16 patients the higher level of IAP was combined with relatively low CI and significant Pg-aCO2. On the second day after operation (table 2) IAH was accompanied higher Pg-aCO2 in 7 patients of the first cluster. There was no significant difference between CI in the clusters on the second day after operation.

Differences between IAP, CI and PgCO2-PaCO2 before operation (n=24) 

The aim of this clinical trial is to study the influence of daily interruption of sedative infusions in ICU patients (pts) with head injury on: duration of mechanical ventilation (DMV), duration of weaning (DW), length of stay (LS), number of brain CT scans required, number of extubations due to agitations when pts were awake and prognosis.

We studied retrospectively 69 ICU pts with head injury, 45 men (65.2%) and 24 women (34.8%). All of them received sedation and mechanical ventilation. Mean age: 34.2±18.3 years. The pts were divided in 2 groups: in group A (30pts, 43.5%), sedation was regularly interrupted every 24h until the pts were awake and examined neurologically or until the absolutely needed sedation again because of agitations and hemodynamic instability. In group B (39 pts, 56.5%) sedation was not interrupted until the onset of weaning. Sedatives used were propofol or midazolam in combination with opiates in 56 pts (81.2%) and paralytic drugs in 45 (65.2%).

In groups A and B were respectively observed: Mean initial Glasgow Coma Scale: 6.7±1.7 and 6.6±1.3. Mean DMV: 11.2±3.7 and 16.8±3.4 days. Mean DW 1.8±0.9 and 4.7±1.3 days. Mean LS: 14.9±5.1 and 21.1±3.8 days. Mean number of brain CT scans required 2.2±0.2 and 2.7±0.5. Mortality rates: 6/30=20% and 9/39=23.1%. Overall mortality rates: 15/69=21.7%. From pts of groups A, 1 needed brain MRI and 3 lumbar punction, while in group B 3 and 6 pts respectively. In group A the mean duration of infusion was 22.5±0.6h per day for propofol and 19.4±2.3h for midazolam.

Statistical analysis showed that: 1) Pts of group A had shorter DMV (p<0.05), DW (p<0.01) and LS (p<0.05). 2) Paradoxally the mean number of brain CT scans required had no significant difference between the two groups (p<0.1); however, this slight decrease may reduce the rate of complications related to the transport of pts and the risks due to irradiation. We mention that additional information was provided, especially during the last year, by the frequent use of transcranial Doppler in all pts with head injury. 3) No increased episodes of extubation by the pts were noticed in group A. 4) Prognosis was not influenced; mortality rates were similar in both groups (p<0.1).

Dimopoulou I 1 , Douka E 1 , Tsagarakis S 2 , Zervou M 1 , Thalassinos N 2 , Roussos C 1 1 Critical Care Medicine, 2 Endocrinology, Evangelismos Hospital, Athens, Greece

Disagreement between studies is evident regarding the incidence of gonadal dysfunction following traumatic brain injury (TBI). Moreover, factors posing a risk to gonadal deficiency in head injury victims are incompletely understood.

To clarify these, 27 male patients with moderate-to-severe TBI (median GCS 7), having a mean age of 37 years, were studied after weaning from mechanical ventilation (10-60 days following physical injury). Head injury was due to motor vehicle collissions (n=20) or accidental falls (n=7). Initial brain CT-scans were graded according to Marshall Computerized Tomographic Classification (MCTC)(scores I-VI). Intracranial pressure was determined by collecting hourly measurements. Endocrine assessment included measurement of testosterone (T), prolactin (PRL), follicle-stimulating hormone (FSH) and luteinizing hormone (LH). Hypogonadism was considered when serum levels of T were low along with low or normal concentrations of gonadotropins, in the presence of normal PRL concentrations.

T levels ranged from 66-535 ng/dl and PRL concentrations ranged from 3.3-42.0 ng/ml. Six of the 27 patients (22%) had hypogonadism. There were no differences in age, GCS score on admission in the ICU, presence or magnitude of intracranial hypertension between patients with hypogonadism and those with normal gonadal function. In contrast, patients with hypogonadism had a higher MCTC score on initial brain CT-scan compared to subjects with normal gonadal function (V vs. III, p=0.006).

In male patients treated in the ICU for TBI gonadal abnormalities are relatively frequent and depend upon radiological measures of head injury severity.

Ferrari C 1 , Colombo A 1 , Franchi S 1 , Losappio S 1 , Marchesi R 1 , Stocchetti N 1 1 Anaesthesia and Intensive Care Medicine, University of Milan, Milano, Italy INTRODUCTION. our objective is to describe intracranial pressure (ICP) in traumatic brain injured patients (TBI) admitted to our ICU between 1997 and 2001 and its association with therapeutic level and outcome.

METHODS. 176 TBI (mean age 39 years, 71% males, 67% pure TBI, median post resuscitation motor GCS 4) with ICP digitally recorded. Therapeutic level was defined as: 1. standard (L1): sedation, mannitol, CSF (cerebral spinal fluid) drainage, PaCO2 35-30 mmHg 2. strengthened (L2): induced arterial hypertension, PaCO2 29-25 mmHg, myorelaxants 3. extreme (L3): barbiturate, PaCO2 < 25 mmHg. Outcome was assessed at ICU discharge with a simplified score (D1: performing simple orders; D2: not performing simple orders; D3: dead). For each patient the percentage of monitoring time with ICP > 20 mmHg has been calculated. Intracranial hypertension (HICP) was defined as ICP> 20 mmHg for more than 20% of recording time.

. 20112 hours of ICP monitoring (median 96 hours; min 12-max 240), mean ICP 15 mmHg (SD 5). 79% of patients with at least one episode of ICP > 20 mmHg, 43% with HICP. Therapeutic level and outcome at discharge are shown in the next table:

Wellnhofer et al: JClinMonitComput

Appendini L 1 , Barberis L 2 , Micelli C 3 , Racca F 3 , Donner C F 1 , Ranieri V M 3 1 Divisione di Pneumologia, Fondazione Salvatore Maugeri, IRCCS, Istituto Scientifico di Veruno, Veruno, 2 Servizio di Anestesia e Rianimazione, Osp. Maria Vittoria, 3 Dipartimento di Anestesia e Terapia Intensiva, Università di Torino, Torino, Italy

Detection of the beginning of inspiratory effort (IEstart) by the expiratory flow pattern analysis (AJRCCM 2002; 166:21-30, online suppl.) , combined with the evidence that airway (Paoslope) and esophageal (Pplslope) pressure slopes during the first 100 ms of an occluded inspiratory breath are equal to Pplslope at IEstart in patients with PEEPi (AJRCCM 1996; 154:907-12) , allow to compute noninvasively PEEPi,dyn by multiplying Paoslope by the time interval between IEstart and the onset of inspiration (DtPEEPi) (AJRCCM 2003; abstract in press) . We wondered whether this noninvasive estimation could be automated.

We developed algorithms to detect both IEstart and Paoslope. Then, we measured in 18 intubated / tracheostomized, mechanically ventilated patients with COPD PEEPi,dyn both conventionally (Ref) (AJRCCM 1994; 149:1069-76) , and noninvasively with manual (Manual) and automatic (Auto) detection of IEstart and Paoslope during a trial of spontaneous breathing. CONCLUSION. We conclude that it is possible to obtain a clinically acceptable noninvasive automatic estimation of PEEPi,dyn in spontaneously breathing intubated-tracheostomized patients with COPD. Grant acknowledgement: Drager, Germany

Bos M E 1 , Ramnarain D 1 , Braams R 1 , Leenen L 1 1 Surgery, University Medical Centre Utrecht, Utrecht, Netherlands

To compare an alternative technique, the Pulse Contour Cardiac Output (PiCCO), a combination of transpulmonary thermodilution and arterial pulse contour analysis, with the artery pulmonary catheter (PAC) for measurement of circulating blood volume and cardiac preload.

Prospective study on a Surgical Intensive Care Unit. All patients were mechanically ventilated and had a distributive shock. In patients with a PAC, a PiCCO catheter was inserted and hemodynamic variables consisting of CVP, CI, CO, wedge, Intrathoracic blood volume (ITBV) and systemic vascular resistance (SVR), were collected simultaneously every 4 hours. The correlation between variables was evaluated with a linear regression model.

The median APACHE II score of the included patients was 17 (range 10-24). All patients were mechanically ventilated with positive end expiratory pressure (PEEP), ranging from 5 to 15. A total of 139 paired data sets in 13 patients were collected. We found a reasonable correlation between the PAC and The PiCCO, for the CO (0.84), the CI (0.73) and the SVR (0.76). When comparing the ITBVI and the CI (0.1), no correlation was found, and only moderate between the wedge and the CI (0.41). There was no correlation between the wedge and the ITBV (0.12). There was a correlation between the wedge and the PEEP (0.6) while hardly any between the ITBV and the PEEP (-0.12) was found.

There is an acceptable correlation between the CO measured by the PAC and the PiCCO. The same applies for the CI and the SVR. In contrast to the results in recent literature (1, 2) , there was hardly any correlation between the CI and the ITBVI. There is, as expected, no correlation between the ITBV and the wedge. There is a discrepancy between pressure derived measures (wedge) and volume derived measures (ITBV) regarding the volume status of hemodynamically unstable patients. This is also influenced by positive end expiratory airway pressure. These observations are of major influence when performing goal directed volume therapy in patients with distributive shock. 

Fernández-Mondéjar E 1 , Castaño J 1 , García Delgado M 1 , Chavero M 1 , Machado J 1 1 Cuidados Críticos y Urgencias, Hospital Universitario Virgen de las Nieves, Hospital de Traumatologia, Granada, Spain INTRODUCTION. The accuracy of the transpulmonary method to detect large increments in extravascular lung water (EVLW) is very high in normal lung and somewhat less so in oedematous lung (1). However, it may be more clinically relevant to measure small EVLW variations, for which other diagnostic measures such as chest X-ray are of little utility. Objective: to analyze the accuracy of the transpulmonary thermodilution method to detect small increases in EVLW METHODS. Eleven EVLW determinations were performed by transpulmonary thermodilution (PICCO®) in four pigs weighing 29-32 Kg before and after the intratracheal introduction of 50 ml of saline serum. Six determinations were performed in normal lung and five in oedematous lung.

Navarrete-Navarro P 1 , Garcia-Delgado M 1 , Chavero-Magro M 1 , Rivera-Fernández R 1 , Rincón-Ferrari M 2 , Fernández-Ortega F J 3 1 Intensive Care Unit, Virgen de las Nieves Universitary Hospital, Granada, 2 Intensive Care Unit, Virgen del Rocío University Hospital, Seville, 3 Intensive Care Unit, Carlos Haya Regional Hospital, Malaga, Spain

The objective is to identify mortality and quality of life (QL) outcomes obtained in a population with severe trauma presenting adult respiratory distress syndrome (ARDS)

Prospective observational study. Setting: 24 ICUs in Andalusia. Period: 1/7 to 31/12/00. Inclusion criteria: severe trauma, defined by Injury Severity Score (ISS) > 15 and/or Revised Trauma Score (RTS) < 12; and ARDS, defined by American European Consensus Conference (AECC) criteria. Measurement instruments. Severity measured by Apache II, ISS and RTS. QL was evaluated by the PAECC Project questionnaire for critical patients (range: 0-29 pts) with 3 subscales: subscale 1, basic activities (range 0-9 pts); subscale 2, normal daily activities (range 0-15 pts); and subscale 3, emotional state (range 0-5 pts) 

Patients with ARDS of traumatic origin present a worsening in their QL at 2 years, mainly due to inability to perform normal daily activities and residual emotional disorders. Recovery to pre-accident levels is only produced in a minority of these patients. Grant acknowledgement: Authors are representing GITAN group.

Lebuffe G 1 , Solus H 1 , Fleyfel M 1 , Tavernier B 2 , Vallet B 1 1 Anesthesiology II, 2 Anesthesiology I, University Hospital of Lille, Lille, France

Major hepatic surgery leads to variations in volume status and venous return. The aim of the study was to assess the use of six hemodynamic parameters in predicting fluid responsiveness in major hepatic surgery. Two static parameters measuring cardiac preload were studied through pulmonary artery catheterization: right atrial pressure (RAP) and pulmonary capillary wedge pressure (PCWP). Four dynamic parameters studied ventricular preload dependency : respiratory pulse pressure variations (∆PP) (1) of radial arterial pressure (∆PPart), pulmonary arterial pressure (∆PPpap), infrared photo-plethysmographic capillary pressure (Finapress®) (∆PPfina) and respiratory variations of the pulse oximetry curve (∆PPsat). Fluid responsiveness was assessed through increase in stroke volume index (SVI) measured by thermodilution.

With institutional ethics review board approval, 8 patients undergoing major hepatic surgery were prospectively enrolled. A 250 mL colloid fluid challenge (FC) was systematically performed for heart rate rising and/or systolic blood pressure falling up to 20% from baseline. Each hemodynamic parameter was measured before and after FC. The FC was repeated if SVI increased over 10% (responder: R). An increase in SVI < 10% was classified as non responder (NR). To assess the ability of indexes to predict increase in SVI to FC, Receiver Operating Characteristic (ROC) curves were generated and the areas under the ROC curve (AUC) were calculated and compared for each parameter. 

Pirat A 1 , Candan S 1 , Zeyneloglu P 1 , Özgür S 1 , Arslan G 1 , Haberal M 2 1 Anesthesiology, 2 General Surgery, Baskent University, Ankara, Turkey

To identify the major reasons why renal transplant recipients (RTRs) at our centre are admitted to the intensive care unit (ICU); to isolate prognostic factors and determine the ICU mortality rate in this group; and to assess graft function in ICU survivors.

The medical charts of all RTRs who were admitted to our ICU from January 2000 through December 2002 were retrospectively studied. Demographic data, interval from transplantation to ICU admission, indications for ICU admission, Acute Physiology and Chronic Health Evaluation (APACHE) II score, length of ICU stay, need for and duration of mechanical ventilation, renal graft function at discharge from ICU, and ICU mortality were recorded.

. Twenty-six patients had a total of 31 ICU admissions. The mean patient age and mean time since transplantation were 37±12 years and 38±43 months, respectively. Twelve (38.7%) admissions were due to infection (7 sepsis and 5 pneumonia cases), and 8 (25.8%) were due to neurological disorders. The other ICU admissions were due to cardiopulmonary arrest (4 cases, 12.9%), postoperative care (3 cases, 9.7%), metabolic disorders (2 cases, 6.5%), pulmonary oedema subsequent to hypervolaemia (1 case, 3.2%), and dissecting aortic aneurysm (1 case, 3.2%). The mean APACHE II score and mean ICU stay were 25±10 and 6±11 days, respectively. The overall ICU mortality rate for the group was 42%. Compared to the ICU survivor admissions, the non-survivor admissions required mechanical ventilation more frequently (p=0.001) and for longer periods (p=0.004). The non-survivor group also had a lower mean white blood cell count at the time of ICU admission (p=0.16), a higher mean number of organs that became dysfunctional during the ICU stay (p=0.03), and a higher mean APACHE II score (p=0.001). In the 18 survivor admissions, 44% of the patients showed adequate renal graft function at the time of hospital discharge.

Infections and neurological disorders are the main reasons for ICU admission in our RTRs. Need for and duration of mechanical ventilation, white blood cell count on admission to ICU, number of organs that dysfunction during the ICU stay, and APACHE II score were identified as important prognostic factors in this patient group. In more than half of the ICU survivors, the patient required haemodialysis at the time of hospital discharge.

In apparantly adequately resuscitated shock patients microvascular failure may be severely underestimated. If not recognized timely, intestinal microvascular blood flow becomes jeopardized inducing organ dysfunction and increasing the risk of death. If intestinal hypoperfusion is reflected in abnormal sublingual microvascular flow, the latter parameter could possibly be used as a guideline for treatment.

Sublingual and mucosal microcirculatory flow was determined in patients with a stoma admitted to our mixed 10 bed ICU by orthogonal polarization spectral (OPS) imaging. For analysis, digital screen images were evaluated as described before[1] and microvascular flow index (MFI) was calculated in small, medium, and large-sized microvessels (0=no flow; 1=sludging, 2=moderate flow, 3=high flow).

During the study period of 6 months, 23 paired measurements (sublingual and stoma) in 17 patients (12 ileostoma, 5 colostoma) were evaluated. The mean age was 59 years with a mean APACHE score at admission of 18.6; mean predicted mortality 41%. Fourteen patients left the hospital alive. Two patients died due to persisting multiple organ failure. Mean sublingual MFI varied in small (2.2; range 0-3.0), medium-sized (2.5; range 0.3-3.0), and large-sized microvessels (2.9; range 0.8-3.0). Mean intestinal MFI was 2.4 (range 0-3). Intestinal microvascular flow was related to sublingual microvascular flow in small microvessels (P=0.001), medium sized microvessels (P=0.001), but not to the flow in large microvessels (P=0.199).

This observational data suggests that sublingual flow patterns in smaller microvessels might reflect intestinal mucosal blood flow. Sublingual OPS imaging may be a valuable and easy bed-side tool for optimizing intestinal blood flow. Further studies should corroborate this finding. 

The appearance of new television and computer facilities has made possible processing of large video information content and obtaining of the quantitative characteristics of such dynamically varying processes as capillary blood flow.

We used computer capillaroscope "Capillar" (Russia) for measurement capillary blood velocity (CBV), perivascular zone (PZ) size and some other parameters during treatment by diuretics and captopril in 17 patients with coronary artery disease (CAD) complicated with congestive heart failure (CHF).

At the beginning of the patient's treatment PZ size was 143,5±16,5 m , CBV was 265,3 ± 123 m s. After first week PZ size reduced to 126,9±12,1 m, CBV increased to 320,6±120 ms. PZ size declined to 115 ±12,6 m, CBV increased to 342,7±116,8 ms after 3 weeks from a beginning of treatment .

Computerized capillaroscopy could be use for noninvasive quantitative assessment of microcirculation parameters in cardiology, intensive care management, clinical and pharmacological trials.

Gómez-Martinez E 1 , Borrás-Pallé S 1 , Jarabo-Bueno M M 2 , Díaz-Ruiz M 2 , Moltó-Guillamón L 1 , Gómez-Casals V 2 , Valentín-Segura V 1 1 Intensive Care Unit, 2 Biochemistry Laboratory, University Hospital Dr Peset, Valencia, Spain

Inflammation plays a role in the development and maintenance of atrial fibrillation (AF).High-sensitivity C-Reactive Protein (hs-CRP) is increased in these patients. Some pro-inflammatory cytokines, such as TNF-alpha, IL-6 and IL-1beta are elevated in patients with heart failure. We aimed: 1) To evaluate inflammation in patients with AF by measuring these markers 2) To analyse them, after elective electrical cardioversion (ECV), at one-month followup, and the relationship with cardiac rhythm.

We included 34 consecutive patients with AF planned to undergo electrical cardioversion (ECV). Levels of hs-PCR, TNF-alpha, IL-6 and IL-1beta were measured in all the patients previously to ECV and compared with a control group of 16 healthy volunteers. Patients with malignancy or active infectious process were excluded. At 1-month follow-up, rhythm and inflammatory markers were reassessed in 24 patients.

. 34 patients (mean age, 63+/-11 years; 8 (23 %) women) with AF (mean duration 2 +/-4 years) were included.All the patients underwent programmed ECV, which was successful in 30 (88.2 % ) cases. hs-PCR was significantly higher in patients compared to controls (mean 5.29 +/-7.63 vs. 1.56 +/-2.45 mg/l, p 0.015). TNF-alpha levels were elevated in 10 (29 %) patients (mean 2.07 +/-3.39 pg/ml) as compared with controls, whose levels were undetectable in all of them (p=0.017). Neither IL-6 nor IL-1beta were significantly different in both groups. Sinus rhythm (SR) was maintained in 10 (41.7 %) patients at 1 month. hs-PCR in these patients was significantly decreased compared with their baseline hs-PCR levels (baseline 4.59 +/-4.58 vs 2.97 +/-2.20 mg/l at 1 month, p= 0.03). We found no difference for TNF-alpha. Patients with recurrence of AF did not show this change.CONCLUSION. 1. hs-CPR and TNF-alpha are increased in patients with AF compared with a control group of healthy volunteers. Other pro-inflammatory cytokines, such as IL-6 and IL-1beta, are not different in both groups. 2. Patients maintaining in sinus rhythm 1 month after elective ECV show a significant decrease in hs-CPR compared with their baseline levels, in contrast with patients with recurrence of AF. These are preliminary results and further data are needed to confirm these results in the long-term follow-up.

Heringlake M 1 , Bahlmann L 1 , Klaus S 1 , Leptien A 1 , Misfeld M 2 , Kraatz E 2 , Schmucker P 1 1 Klinik fuer Anaesthesiologie, 2 Klinik fuer Herzchirurgie, Universitaetsklinikum Schleswig-Holstein -Campus Luebeck, Luebeck, Germany

To determine the relationship between interstitial lactate -determined by myocardial microdialysis (1) -, hemodynamic variables, and (right ventricular) myocardial performance during coronary artery bypass grafting (CABG) -surgery .METHODS. 20 patients undergoing CABG -surgery with cardiopulmonary bypass (CPB) were studied. Measurements of myocardial lactate levels were performed with a microdialysis probe inserted into the apical region of the beating heart. Right ventricular function was determined with a fast-response pulmonary artery catheter. Measurements were performed before (t1) and after CPB (t2) during periods of 15 min. Patients were grouped according to the median interstitial lactate at t1 (high lactate group: HL-group; low lactate group: LL-group; n = 10, respectively). Statistical analysis were performed nonparametrically.*: p < 0.05 for between group differences -Mann-Whitney-test.

Due to grouping, myocardial lactate at t1 was significantly higher in the HL-group. Lactate levels after CPB were not different between both groups. Heart rate, mean arterial pressure, and cardiac index were not different between both groups. Right ventricular ejection fraction before CPB was lower and central venous and pulmonary artery pressures after CPB were higher in the HL-in comparison with the LL-group (p < 0.05, respectively) CONCLUSION. Increased myocardial lactate levels -before CPB -are associated with decreased right ventricular contractility before and increased filling pressures after CPB in pts. during CABG -surgery. It is suggested that myocardial microdialysis may be useful to identify pts. at risk for developing postbypass myocardial dysfunction and, hence, for optimizing pharmacological treatment during CABG-surgery.

Schmidt H 1 , Müller-Werdan U 1 , Tymiec P 1 , Werdan K 1 , Witthaut R 1 1 Med.Klinik III, Martin-Luther-University Halle-Wittenberg, Halle, Germany INTRODUCTION. Brain natiuretic peptide is an established marker of left ventricular dysfunction in different disease entities. Its release characteristics are potently influenced by the sympathetic activity. Our present study aimed to investigate whether the BNP plasma level (BNP) correlates with myocardial damage and whether BNP is influence by autonomic dysfunction (AD) seen in MODS.

We enrolled 45 consecutive MODS patients who were admitted to a university Medical/Cardiological ICU. MODS was definied by an APACHE score (AP2)>=20 and sepsis as a sepsis score according to Elebute and Stoner (SeSc) >=12. BNP was measured by RIA. Troponin I (TI) reflected the myocardial damage and was measured by routine hospital analysis (ELISA). Heart rate variability (HRV, time-domain and frequency domain analysis), baroreflex sensitivity (BRS, phenylephrine method) and chemoreflex sensitivity (CRS, hyperoxic method) were used as markers of AD and assessed according to the international standards (summary in [1]). HRV parameters pNN50 and HF as well as BRS were used for assessement of parasympathetic activity, LF/HF ratio as a measurement of sympathetic modulations and mean heart rate (24 h) as an indicator of sympathetic-parasympathetic balance.

. 16 female and 29 male MODS patients (mean±sd, age 59±15 y, AP2 29.0±8.9 [for comparison SOFA Score 11.4±3.9], SeSc 12.5±5.3, 84% on mechanical ventilation, 56% with catecholamine application, 62% sedated) into the study. The mean BNP was (ln-transformed) 0.8±0.2 pg/ml and that of TI (ln-transformed) 2.2±2.8 ng/ml. We found a significant correlation of TI with the BNP (r=0.54, p=0.017). The parameters of AD were characterized as follows (mean±sd): pNN50 5.1±8.4, HF 141.4±370.5 ms2, LF/HF 1.1±1.1, BRS 1.5±1.2 ms/mmHg, CRS 0.5±0.4 ms/mmHg, mean heart rate 93.0±18.5 beats/min. We found no significant correlation of AD markers with BNP (pNN50 r=-0.1, p=0.5; HF r=-0.2, p=0.3, LF/HF r=0.01, p=0.9; BRS r=0.1, p= 0.4, CRS r= 0.2, p=0.4, mean heart rate r=-0.3, p=0.1). There was no difference between nonseptic and septic patients (lnBNP 0.76±0.77 vs. 0.77±1.2 ng/ml, p=0.5) CONCLUSION. According to our results we conclude that BNP is correlated to the myocardial damage in MODS patients. The AD was blunted in the observed cohort of patients. Otherwise than in healthy subjects BNP seems not to be linearly correlated to sympathetic activity in MODS. 

Piacevoli Q 1 , Azzeri F 1 , Muller T 1 1 Anesthesia and Intensive Care, San Filippo Neri Hospital, Rome, Italy INTRODUCTION. The technique of cerebral microdialysis has been used on patients with cranial trauma, subarachnoidal haemorrhage and epilepsy.Cerebral energetic metabolism was monitored by analysing glucose, lactate, pyruvate and glutamate molecules collected by microcuvettes.

The aim of our study was to observe the sensitivity of cerebral metabolites as glucose, lactate, pyruvate and glycerol in severe head trauma using microdialysis in predicting outcome.In literature four different types of lactate build-up mechanisms are described: 1.temporary build-up (< 1 hour), immediately after the probe is positioned. It is a localised phenomenon, probably due to the fitting of the catheter and does not reflect a true change in brain tissue metabolism. 2. a prolonged initial lactate increase, which gradually decrease over 24 to 48 hours, as illustrated in figure 2. Such conditions are typical of patients with significant primary brain injury, further complicated by massive cerebral oedema and pharmacologically resistant intracranial hypertension, which is later resolved by surgical decompression. 3. the third type of lactate increase is directly related to lack of cerebral oxygenation and extracellular glucose decrease.An increase in the lactate/pyruvate ratio is, therefore, one of the main markers for brain hypoxia/ischemia.

During the initial stages which follow cerebral trauma, high levels of extracellular glutamate and other excitatory aminoacids (EAA) have been observed. Using microdialysis, we have demonstrated EAA increases, especially glutamate, from 500 to 700% for basal values. We found two types of trends. Prolonged increases in concentration. Temporary increases in concentration. 

Intensive care patients with severe brain injury (SBI) suffer from cerebral ischemia and dysoxia. Both low and high SjvO 2 values are associated with poor outcome. SBI affects thyroid function and secondary hypothyroidism correlate with neurological impairment. The purpose of this study was to compare cerebral oxygen metabolism (COM) in patients who survived or died of SBI and to correlate this changes with thyroid function.

COM and thyroid hormones were monitored in 56 patients with SBI in the NICU. Patients were managed by a standard protocol that emphasized prompt evacuation of intracranial hematomas and prevention of secondary brain insults. They all were ventilated and treated aggressively to keep ICP 70 mm Hg. Patients were divided in two groups according to GOS at discharge from the hospital: group I -severe disability and vegetative states (GOS 3-2, 16M/5F,38 yo) and group II -dead (24M/11F, 39 yo). SjvO 2 was measured intermittently in blood obtained from jugular bulb catheters. Plasma thyroid hormones and prolactin level were evaluated in 28 patients using RIA.

Multiple jugular venous desaturations were found in 67% and 36% and elevated SjvO 2 were found in 44% and 59% of patients of group I and II respectively. Parameters of COM during first week after insult were different between groups, with higher SjvO 2 , lower CEO 2 and AVDO 2 in dead patients. Low TSH, T 3 , T 4 and prolactin level were very common, correlated with severity of diencephalic syndrome and were more pronounced in Group II. We found a very close correlation (r=0.9267, p<0.0001) between two methods of CEO 2 calculation: 1.S(a-jv)O 2 and 2. AVDO 2 /CaO 2 x 100%.

These results demonstrate that both low and high SjvO 2 are very common events in neurosurgical comatose patients with poor outcome. Decreased COM and symptoms of secondary hypothyroidism due to severe diencephalic dysfunction are very poor prognostic factors. We postulate that high level of SjvO 2 is associated with impaired cerebral mitochondrial function due to severe secondary hypothyroidism. Severe diencephalic dysfunctions are independent factor of unfavorable outcome. We recommend to use S(a-jv)O 2 as more practical method of calculating CEO 2 .

Rodrigues M G 1 , Salgado D R 1 , Resende V M 1 , Negri M P 1 , Paiva R A N 1 , Verdeal J R 1 1 Intensive care unit, Barra D'or, Rio de janeiro, Brazil

Neurological monitoring is very important to change the prognosis of critical neurological patient. The jugular venous bulb saturation is extremely important to evaluate the consumption and delivery of oxygen. To describe complications during insertion and permanence period of jugular venous bulb catheter.

Prospective, observational study of 21 patients from June 2000 to September 2002 in an intensive care unit. All patients were monitored with intracranial pressure device (ICP-CAMINO). Jugular venous line was cannulated independent of which side. The catheter flow was sustained by continuous saline infusion (rate-3ml/h). The monitor used was VIGILANCE (Baxter). Complications were observed during insertion such as arterial puncture, bleeding, and misplacement; during catheter permanence obstruction and infection (daily examination); after decannulation: thrombosis detected through Doppler, which was performed after 24 hours. All the catheter tips were sent to bacteriological examination.

The meaning time of cannulation was 5 days. The thrombosis rate detected by Doppler was 31,6% (without clinical compromise). The catheter obstruction rate was 15,8% and infection rate 10,5%.

Strict control with Doppler examination is very important to warrant optimal flow. The catheter must be changed every five days in order to avoid infection. were related to a significant higher grade of disability measured with DRS (p< 0,05). The existence of any type of DAI was also related to higher disability to the discharge of ICU (p< 0,04). 

Albanese J 1 , Bourgoin A 1 , Kaya J 2 , Antonini F 1 , Leone M 1 , Martin C 1 1 Anesthesiology and ICU, 2 Neurosurgery, CHU Nord, Marseille, France

Glasgow Outcome Scale (GOS) is the more often assessment used as prognosis index for head trauma patients. That prognosis scale does not seem specific enough to have a good assessment of the quality of life of these patients and overrates probably the real recovery. The aim of the present study was to investigate the long-term outcome in survivors among patients with severe head trauma admitted in intensive care unit (ICU).

Among 2641 trauma patients admitted from 1992 to 2000, a severe head trauma, defined as GCS ≤ 8 and indication of ICP monitoring, was diagnosed in 381 patients. 269 (70%) were alive after the ICU stay. All these patients or next of kin were contacted in order to determine their real independence, motor defects, and behavioral troubles. The visits were scheduled for the follow-up examinations at least one year after the trauma.

Among the 269 patients, 199 (74%) were got in touch. Their mean age was 32 +/-15 yrs, with 77% of men. Among these 199 patients, GOS was as shown in CONCLUSION. The prognosis of severe head trauma patients is gloomy, since 50% of them died between admission and follow-up examinations (> one year). Only few patients (< 20%) can have a quality of life similar to that they had before the occurrence of trauma. The determination of the long-term prognosis of severe head trauma patients with GOS overrates the patient quality of life. Most of them have severe cognitive disorders, harmful to their professional reintegration. Patients are classified in two groups: group I for patients that present a GCS less or equal than 8, one volume of the haemorrhage greater than 65, 50, 20 and 5 cc if the location of the ICH is lobar, basal ganglia, cerebellum or brain stem*. Group II for the rest of the patients. Statistical analysis is done by the ANOVA test (analysis of variance). Location is analyzed by Fisher´s exact test and GCS by Chi-square method.

. 25 patients progress to encephalic death. 22 present lobar haemorrhage, 36 of basal ganglia, 12 of the cerebellum and 8 of brain stem. In the group I, 22 of the 31 patients progress to encephalic death. Only three patients of the group II present encephalic death. Patients with a GCS less or equal than 8 have a greater risk of evolution to encephalic death (P<0´00001) with a relative risk of 46´6 (95% confidence limits 5´83-373´38). Probability of progress to brain death with this method owns a sensibility of 88% and a specificity of 83%. Among lobar haemorrhages with a volume greater than 65 cc the probability of brain death is 75% and the specificity 64´3%. In basal ganglia haemorrhages with a volume greater than 50 cc sensibility is 90´9% and specificity 92%. In haemorrhages of the cerebellum (volume >20 cc) and of the brain stem (volume > 5 cc) sensibility is 100%. Brain death is in the first five days of stay.

Volume of the cerebral haemorrhage (depending on the location) measured by CT scan at admission, together with the Glasgow Coma Scale score are very good, feasible and easy prognostic indicators of the evolution of ICH to brain death. 

Cerebral vasospasm remains a significant source of morbidity and mortality in patients with subarachnoid hemorrhage (SAH) after an aneurysmal rupture (1) . Interventions have to be quick and aggressive. We aimed to define the clinical outcome and factors related to cerebral vasospasm after SAH.

From August 2000 to December 2002, we retrospectively studied 101 patients consecutively admitted to our twenty-four-bed ICU (700-bed primary hospital) who had a computed tomography (CT) scan revealing SAH. Ruptured aneurysm was verified using angiography.

Out of 101 patients (66 male, 35 female, mean age 51±11 years, Hunt and Hess grade 2.1±0.8, Fisher 2.4±1.0), 32 patients (32%) developed symptomatic vasospasm within 14 days after SAH. Presence of hypotension (mean arterial pressure < 70 mm Hg at any moment) was associated with vasospasm (RR 3.94 CI 95% 1.22-9.72). Patients who developed vasospasm had significantly higher values for Hunt and Hess (1.96 ± 0.76 vs 0.81 ± 0.14, p<0.05) and Fisher grades (2.0 ± 1.0 vs 2.8 ± 0.9, p<0.05) as well longer hospital stay (16±11 vs 11±7 days, p<0.05) compared to patients who did not. The observed hospital mortality rate was 34% (n=34) and was significantly higher in patients with vasospasm (63% vs 20%, RR 3.08; CI 95% 1.79-5.28).

The results of this study suggest that after SAH very close attention should be directed to hemodynamic stability that may be essential to optimize cerebral hemodynamics and thus to minimize secondary injuries.

Otter H E 1 , Martin J 2 , Bäsell K 1 , Behnisch I 1 , Jänsch P 1 , Kutun S 1 , Kox W J 1 , Spies C D 1 1 Dept. of Anesthesiology, University Hospital Charité, Berlin, 2 Dept. of Anesthesiology, Klinik am Eichert, Göppingen, Germany

Withdrawal symptoms are observed in more than 60% of all patients requiring long term sedation [1, 2] . Prolonged agitated states are associated with a longer ICU stay and poor outcome [2] . We investigated if a new score, the Modified Agitation and Sedation Score (MASS), can predict length of ICU stay.

After ethical approval and written informed consent, 1073 patients in the ICU were assessed using the MASS together with the Ramsay Sedation Scale (RSS) 3 times per day. The MASS is composed of eight criteria (orientation, hallucination, agitation, anxiety, seizures, tremor, paroxysmal sweating, altered sleeping waking rhythm) and for each criterion 0 to 7 points can be allocated. Earlier studies showed that a MASS > 7 is sensitive and specific for withdrawal symptoms [3] . A clinical diagnosis of withdrawal symptoms was documented as well as total ventilation time, over-all length of ICU stay and TISS-28. Statistical analysis: non-parametric variance analysis, rater operating characteristics (ROC).

Patients with a MASS > 7 (n=93) had a significant longer ventilation time (p<0.001), a significant longer ICU stay (p<0.001) and a significant higher TISS-28 (p<0.001) than patients with a MASS <= 7. The ROC for a MASS > 7 versus length of ICU stay showed an AUC of 0.718 (CI: 0.666-0.770; p<0.001).

With the MASS a screening of patients at risk for withdrawal symptoms and prolonged ICU stay is possible and enables the clinician to start an intervention therapy immediately. A consequent screening of withdrawal symptoms, i.e. applying the MASS, is very reasonable in order to treat withdrawal symptoms early and avoid subsequent cost. 

Activation of opioid receptors is common after surgery because the stress of surgery provokes the release of endogenous opioids, but also because opioids remain the most common treatment for pain in patients undergoing surgery. Delays in gastric emptying and prolongation of bowel transit time are well documented side effects in these patients. Opioid antagonists like MNTX or naloxone can be used to antagonize these undesirable gastrointestinal side effects (1). The aim of this study was to evaluate the effect of MNTX or naloxone in combination with sufentanil on peristalsis.

Guinea pig small bowel segments of 8 cm length were set up in parallel organ bathes containing oxygenated Tyrode's solution. Peristalsis was elicited by luminal perfusion (0.5 ml/min) against an aboral resistance of 400 Pascal. Peristaltic pressures were recorded at the aboral end of the segments. Perfusion of the segments resulted in an increase of the intraluminal pressure until a pressure threshold (PT), at which peristaltic contractions were triggered, was reached. An increase of the PT is interpreted as inhibition of peristalsis, while a decrease of the PT is interpreted as stimulation of peristalsis. Increasing concentrations of sufentanil were added to the organ bath, after MNTX (1muM) or naloxone (0.5muM) had been added to the organ bath. Each drug or drug combination was tested on 8 different segments. One way and two way ANOVA for repeated measures were used for statistics, p < 0.05 was considered statistically significant.

Sufentanil showed dose-dependent inhibitory effects on peristalsis, with a complete block of peristalsis at a concentration of 1 nM. Naloxone was able to abolish the inhibitory effect of sufentanil almost completely. MNTX on the other hand, only shifted the dose-response curve of sufentanil to the right, but was not able to prevent a complete block of peristalsis at 10 nM.

Naloxone is a potent antagonist of opioid induced inhibition of intestinal motility, while MNTX notedly attenuates the inhibitory effect on peristalsis. In contrast to naloxone, MNTX does not cross the blood-brain barrier. Therefore, in addition to a sufficiently maintained intestinal function, the analgesic effect of opioids is completely preserved. A et al. NEJM 2001; 345: 935-940 

Leiner T 1 , Mikor A 1 , Szakmány T 1 , Molnár Z 1 1 Anaesthesiology and Intensive Care, University of Pécs, Pécs, Hungary INTRODUCTION. The aim of our trial was to evaluate the ability of microalbuminuria as an indicator of outcome and to investigate its' relationship with the postoperative respiratory dysfunction in the initial postoperative period in a high-risk patient group (1).

In a prospective observational study, 153 patients were consecutively recruited following elective oesophagectomy, total gastrectomy, pancreas and liver resection due to tumor removal. Microalbuminuria (expressed as urine albumin:creatinine ratio) was measured before (tp), and after surgery (t0,t6,t24,t48,t72). To assess the patients clinical progress, Multiple Organ Dysfunction Scores (MODS) were calculated on ICU admission than daily (t1,t2,t3). For statistical analysis Wilcoxon rank sum test, Mann-Whitney U test and Spearman's rho test were used as appropriate.

. 130 survivors, 23 nonsurvivors were investigated. Significantly higher MODS were observed in non-survivors throughout the study period (p<0.001). Microalbuminuria increased significantly (p<0.01) on admission to ICU (t0) compared to the preoperative levels, but levels returned to normal within 6 hours and remained so for the rest of the study. There was a significant difference between survivors and non-survivors at t0 (p<0.01). Comparison of M:Cr values with the PaO2/FiO2 ratio showed an inverse relationship on admission, which remained so for t24 and t48 ( 

The adage 'See one, Do one, Teach one' is often quoted as the method of acquiring practical skills training. If trainees are to acquire the high number of necessary skills in a shortening timescale the opportunity to 'See one' must be maximized. We have developed a method of transferring video clips of practical procedures onto two versions of PDA (Pocket PC TM and the Sony Clié TM NX70V). Trainees can see procedures, learn key points and ask questions prior to undertaking them, without patients becoming unduly concerned by the inexperience of the operator.

Following consent we have videoed key elements of several practical procedures. Images were converted into either Sony specific *.mqv files or *.wma files and viewed using Microsoft Windows Media TM or Sony Movie Player 1.0 TM . Procedures included central line and PA catheter insertion, percutaneous tracheostomy and thoracic epidural. We conducted an audit with the staff on the ICU. They stated whether they were naive, novice or experienced in the procedure, viewed the teaching package and assessed it using the following questionnaire: The quality was good, it has increased my knowledge of the procedure, and I learned more than I would have anticipated from a verbal presentation or a practical demonstration. Each question was assessed: I strongly agree, I agree, neither agree nor disagree, I disagree, and I strongly disagree.

We audited 10 junior staff. Most claimed to be amateurs for the procedures covered, with only one graded as a novice. All of them stated that the quality of the presentation was good, with the majority strongly agreeing (70%). Only one trainee thought that the presentation had not increased his knowledge of the procedure shown. All stated that they had learnt more from the presentation than they would have anticipated from a verbal presentation. 3 (30%) felt that a practical demonstration would have had the same teaching impact and 3 (30%) thought they would have learned more from a practical demonstration.

We have successfully used modern digital technology to add a new approach to the teaching of invasive procedures on our ICU. Our early audit confirms that the quality of the images is good and that it can be used with significant advantage, particularly if a practical demonstration is not available to highlight important teaching points.

Maraví-Poma E 1 , Jiménez I 1 , Martínez J M 1 , Escuchuri J 1 , Izura J 1 1 ICU, Virgen del Camino Hospital, Pamplona, Spain 

Viana W N 1 , Garcia N C 1 , Spinoza R E A 1 , Costa M A 1 , Rotman V 1 , Oliveira L E 1 , Castro J E C 1 1 Surgical Intensive Care Unit, Copa D´Or Hospital, Rio de Janeiro, Brazil

Bariatric Surgery (BS) is a relatively safe therapeutic approach for severe obesity. Complications are not completely known because there are scarce patient follow-up review in the literature. Our purpose is to describe the complications that took place and the final outcome of the patient admitted into our surgical ICU.

We revised the data on BS in the last 31 months. Complications were divided in surgical and medical ones. Surgical complications were arbitrarily divided in early (< 7 days) and late (> 7 days) and these were subdivided in general (related to any surgery) and specific (related to BS). Medical complications refer to the need for mechanical ventilation, hemodynamic support, infectious, nutritional and metabolic features.

In that period, one hundred and sixty-nine patients were admitted with the diagnosis of BS in our ICU. Thirteen patients (7%), 3 women (23.1%) and 10 men (76.9%), had one or more type of complications. The mean age of the complicated group was 42±13.7 years-old and the mean BMI was 44.8±4.2 Kg/m2. Previously known comorbidities were: hypertension in 6 patients (46.2%), diabetes in 1 (7.7%) and coronary artery disease in one. The surgical technique used was Scopinaro: 10 patients (76%), Higa: 2 (15%) and Capella: 1 (7%). Fistulae in the anastomosis between stomach and jejunum (3 patients, 23%), mechanical obstructions due to internal hernia, volvulus or adherences (4, 30%), gastrointestinal bleeding or hemoperitoneum in anastomotic sites (2, 15%) and malnutrition due to excessive weight loss one year after BS (2, 15%) were the main specific surgical complications. Important medical complications were ARDS in 4 patients (30%), shock in 6 (46.1%), who were managed with Swan-Ganz catheter and vasopressors. Peritonitis occurred in 4 patients (30%), treated with antibiotics, peritoneostomy and periodic surgical reinterventions. Seven patients (53.8%) had parenteral nutrition support. None of our patients developed pressure ulcers although one had decubitus rhabdomiolysis. One patient (9.1%) died from massive venous mesenteric thrombosis. Twelve patients were discharged home and resumed well. Mortality in the whole group of BS was 0.6%.CONCLUSION. In our sample, severe complications occurred in a small subset of patients. Most of them succeeded well despite difficulties in handling heavy patients, intrinsic high mortality rates and a heterogeneous group of surgeons.

Usichenko T I 1 , Groh A 2 , Gruendling M 1 , Rothe K F 2 , Wendt M 1 , Lehmann C 1 1 Anaesthesiology and Intensive Care Department, University of Greifswald, Greifswald, 2 Anaesthesiology and Intensive Care Department, Hospital Dresden-Friedrichstadt, Dresden, Germany

Six-year experience of perioperative management of infrarenal aortic reconstruction (IAR) was evaluated. The purpose was to identify the intraoperative factors that could be associated with the postoperative complications (PC) in the ICU patients after elective IAR.

Intraoperative and intensive care records of 250 consecutive patients that underwent IAR under standardised anaesthetic procedure were reviewed. The following variables were studied: body-mass index (BMI), duration of surgery (OP) and clamping (CT), blood loss volume (BL), fluid replacement volume (FR), minimal body temperature (Tmin) during surgery, experience of anaesthetist and Sequential Organ Failure Assessment (SOFA) score (1) taken on the first postoperative day. The PC in the ICU were defined as one or more of the following: vital organ ischemia or failure, coagulation disorder, bleeding, requiring additional surgery, or death. Cases of severe intraoperative bleeding with subsequent hemodynamic instability and intraoperative ischemia of vital organs were excluded. The association of studied variables with PC in the ICU was evaluated by the logistic regression.

The overall in-hospital mortality was 5.2 % (13 cases). The PC rate in the ICU was 15.6% (39 cases), including 15 cases of bleeding, 12 vital organs ischemia, 8 cases of multiple organ failure and 4 coagulation disorders. BMI, BL, OP time and Tmin during the IAR were found to be strongly associated with the outcome (P<0.05). Patients' age, CT, FR and anaesthetist experience were not associated with increased PC rate. OP time and BL showed significant association with the SOFA score.CONCLUSION. BMI, OP time, BL and Tmin might serve as predictive factors of ICU morbidity and mortality after elective infrarenal aortic reconstruction. Our findings are in agreement with the previously published data (2) . REFERENCE(S). 1. Vincent et al. Intensive Care Med 1996; 22:707. 2. Dardik et al. J Vasc Surg 1999; 30:985. 

Waarsenburg N 1 , Van Meer O 1 , Schoonderbeek J 1 , Bonjer J 1 , Van der Hoven B 1 1 Department of Intensive Care, Erasmus University Medical Center, Rotterdam, Netherlands INTRODUCTION. Percutaneous Dilatational Tracheotomy was already described by Ciaglia in 1985 but is not the standard method for performing a tracheostomy in the Netherlands. This method, however, has many advantages compared to the Surgical Conventional Tracheotomy (SCT), which takes usually more than 30 minutes. Percutaneous Dilatational Tracheotomy is an easy procedure, which can also be performed by other clinicians than surgeons. It is easily performed at the bed-side and causes less bleeding. In 1999 this technique was modified in order to simplify the procedure with concurrent decrease in operating time and complications. The objective of this study is to document the time required to perform a modified Percutaneous Dilatational Tracheotomy (mPDT) and complications associated with the procedure.

Since 2000 eighty-four patients in a surgical intensive care unit of a tertiary referral center underwent modified Percutaneous Dilatational Tracheotomy. All had prolonged mechanical ventilation with expected long duration of weaning. All tracheostomies were performed under supervision of a surgeon or staff intensivist. Airway management was performed by an anaesthesiologist.

The group consisted of 65 men and 19 woman, age 59 +/-15 years. The time needed to perform the procedure was only 6 minutes (median) and varied between 6 and 30 minutes. Mean ventilatory time before tracheostomy was 9 days (1-43); after mPDT 20 days . Complications were seen in 7 patients (8%). In 5 patients superficial bleeding occurred, which could be managed by temporary compression, and in one patient mediastinal emphysema was seen. In only one patient the procedure needed to be converted into an open tracheotomy due to a bleeding venous plexus. No late complications were encountered. Procedure-related mortality was 0%. Mean post-procedure follow-up was 3.7 months (0-25).CONCLUSION. The modified Percutaneous Dilatational Tracheotomy is a safe and quick procedure. It is easily performed at the bed-side, also by non-surgically trained clinicians. It is a safe procedure with low morbidity.

Miliaras S E 1 , Papadopoulos A C 2 , Bakogiannis K S 1 , Karakoulas K 2 , Grossomanides V 2 , Vassilakos D 2 , Kiskinis D 1 1 1st Surgical Dept, 2 Anesthesiology, Aristotele University of Thessaloniki, Thessaloniki, Greece INTRODUCTION. The safety and efficacy of a long-term constant-dose iv infusion of Remifentanil (R) is not well established, based on patient satisfaction and lack of adverse effects. The aim of our randomized, double-blind study was to compare two constant dose continuous infusion of R in ICU patients following major abdominal surgery.

Twenty-eight patients, 62±6 y.o., ASA II-IV were studied. All patients received TIVA with propofol and R, admitted in ICU and randomly assigned in two equal groups: 0.05mug/Kg/min (Group A) or 0.1mug/Kg/min (Group B). After extubation BP, HR, SpO2, RR, Pain score (0-3), and PONV score (0-2) were evaluated for 24 hours. Meperidine 0.25mg/Kg in bolus IV was given for pain score 2 or 3 as rescue analgesia.

The two groups were similar in respect of demographic data, surgical procedures, duration of anesthesia, and time of extubation in ICU. The percentages of patients with adequate analgesia (Pain score 0-1) at measured intervals are shown in figure. Rescue analgesia was significantly less in group B (18%) than in group A (43%)(p<0.05).There were no hypoxemia and respiratory depression.CONCLUSION. The use of long term infusion of remifentanil at 0.05mug/Kg/min or 0.1mug/Kg/min, provided adequate analgesia in patients following major abdominal surgery, although patients in the former group required more rescue meperidine.This approach of remifentanil analgesia represents an effective and safe regiment in ICU patients.

Haji-Michael P 1 , Calderwood R 2 , Halka T 2 , Columb M 1 , Welch M 2 1 Acute block intensive care unit, 2 Vascular Surgery Unit, Wythenshawe Hospital, Manchester, United Kingdom

Emergency ruptured aortic aneurysm (AAA) has a high mortality. Many of those who do die do so after a protracted admission to critical care. The aim of this study is to ask if there is any way of identifying this patient group at an early stage.

All patients admitted to a regional vascular unit over a 3 year period (1999) (2000) (2001) (2002) with a ruptured AAA were identified using theatre records, critical care database and surgeons' personal logbooks. Patients' notes were examined for pre-operative, intra-operative and postoperative events, demographics and physiological data.

Theatre records, databases and logbooks identified 138 patients. There were 77 deaths (55.8% 90 day mortality), with 37 occurring in the first 48hr and 40 occurring after 48hr. A total of 69 patients were still in an ICU at 48hr and this data was analysed. Significant predictors of subsequent mortality are seen on table 1. Many factors had no significant association. Outcome is most significantly related to age >76yr, sepsis and respiratory, cardiovascular and renal failure at 48hrs ( CONCLUSION. Late mortality seems unrelated measures of hypovolaemic shock and reperfusion injury at the time of operation. It also seems unrelated to most common premorbid conditions with the exception of age. Late mortality does seem associated with the development of organ failure and sepsis; no patients over 76yr old survived failure of all 3 organs and sepsis.

Corradi F 1 , Mas A 1 , Escorsell A 1 , Bombui E 1 , García-Valdecasas J C 1 , Rodes J 1 1 Liver Unit, IMD, Hospital Clínic, Barcelona, Spain

We designed a study aimed at assessing the early postoperative period (intra-ICU) of living-donor liver-related transplantation (LDLT) in comparison with contemporaneous cadaver liver transplantation (CLT).

Analysis of 60 pre, intra and postoperative variables in our first 23 consecutive LDLT patients and in 46 CLT recipients (the immediate pre and post of each LDLT).

Preoperative characteristics were similar in the 2 groups regarding demographics, etiology of cirrhosis, presence of hepatocarcinoma and degree of liver impairment (LDLT: Child A 33.3%, B 38.1%, C 28.6%; CLT: A 37%, B 32.6%, C 30.4%). Surgical procedure was similar although the operation time was longer for LDLT (7.7±1.4 vs 6±1.3h; p=0.002) and packing was only required by 14% of CLT recipients (p=0.058). As expected, the ischemic period was shorter for LDLT (85±38 min) than for CLT (343±125 min; p=0.001). In the intra-ICU period there were no differences in the presence of renal, metabolic, infectious, neurological and respiratory complications in the 2 groups. The AST peak was significantly lower in the LDLT (313±181 vs 865±881 UI/l; p=0.001). Other parameters of liver function were similar as were the presence of technical surgical problems (reintervention and vascular problems). Plasma requirements were significantly lower in LDLT (55±235 vs 708±1131 ml; p=0.001). The median ICU stay and the readmission rate were also similar in the 2 groups. Five patients died in the intra-ICU period (LDLT: 2; CLT: 3). The causes of death were infectious in 4 cases and pulmonary embolism in the remaining patient.

LDLT has a similar ICU evolution than CLT although the liver graft is smaller and the surgical procedure is larger and more difficult than in CLT. The low transaminase peak and plasma requirements probably reflect a lower ischemia-reperfusion injury due to a shorter ischemic time in LDLT.

Shimizu M M S 1 , Honda O O H 1 , Yokokawa N N Y 1 , Hiraga K K H 1 1 Anesthesiology, National Cancer Center Hospital, Tokyo, Japan

In recent years,radical cancer surgery is performed after induction chemotherapy or irradiation therapy(CXR).But there is no evidence about the effects CXR for clinical course. In this study, we examined the effects of CXR for perioperative course in radical esophagectomy.

PartA:We retrospectively collected data on 21 patients undergoing radical esophagectomy which were performed CXR before surgery over31 months.We analyzed patients profile, postoperative complications, prognosis, perioperative managements, intraoperative hemodynamics. PartB: Prospectively, we examined the effects of CXR in radical esophagectomy. Eleven patients (CXR group) undergoing induction chemotherapy and irradiation therapy prior to radical esophagectomy from 2001 April to October. Patients in the control group (N=10) did not receive CXR before surgery. Anesthesia was maintained with isoflurane in oxygen and nitrous oxide and differential lung ventilation was done. An intravenous infusion of acetate ringer was maintained 8ml/kg/hr. A postoperative pain was controlled with epidural morphine. During operation, we used the routine monitoring ( heart rate, blood pressure, body temperature, urine volume, blood loss volume, blood sugar, saturation of oxygen, end-tidal CO2. Besides these, cardiac output (co), extra volume of lung water (EVLW) and central venous pressure (CVP) were monitored contiuously. From 1postoperative day(OPD)to4opd CO, CVP, EVLW, PaO2/FiO2, renal function,liver function,body weight balance were wxamined. And also,C-reactive protein, interlukine-1(IL-1),IL-8,IL-10 and IL-6level were measured.

In the CXR group, mortality rate was higher than control group. A preoperative white blood cell count was much lower than control group. EVLW was increasing earlier than control group after operation. Both the serum level of IL-6 and IL-8 were maintained high level until 3opd.CONCLUSION. Compared to control group, the mortality rate of radical esophagectomy with CXR is very high. This is probably due to that systemic inflammation response syndrome (SIRS) is frequent in patients undergoing CXR therapy prior to radical esophagectomy. This hypothesis is shown that serum cytokine level of CXR group is higher than that of control group. And according to peri-operative body weight balance and EVLW, postoperative refilling of water is occurred earlier than control group. 16th Annual Congress -Amsterdam, Netherlands -5-8 October 2003 S187

Spatenkova V 1 , Fric R 1 , Suchomel P 1 , Dienelt J 1 , Treuchel A 1 1 Neurocentre, Hospital, Liberec, Czech Republic

Cerebral vasospasm is a serious complication of subarachnoid hemorrhage (SAH) whose therapeutic options are still limited. Intraventricular administration of nitropruside as a donor of nitric oxide (NO) molecules for the treatment of cerebral vasospasm after SAH has been described in the literature. Although intravenous administration of nitropruside bears the advantage of direct action on the vessel wall, there is the risk of systemic hypotension.

We retrospectively evaluated data from 12 patients with cerebral vasopasm detected by transcranial Doppler sonography (TCD) after surgical treatment of ruptured intracranial aneurysms and subarachnoid hemorrhage. Isosorbid dinitrate (ISDN) was administrated intravenously in all patients, hemodilution was evoked by use of hydroxy-ethyl-starch and all patients received nimodipine. There were seven patients with rupture of ACoA aneurysms, five with MCA, two with ACA and one with PCoA aneurysm, the mean age 46,8 years (range 25-61), Hunt-Hess Grade at admission was 1 to 4 (median 2). Vasospasm was diagnosed on the 1st to the 16th day after SAH (median on the 6th day). Effect of administartion of ISDN was evaluated by regular TCD monitoring on a daily basis.

Administration of ISDN was initiated between the 1st and 4th day (median on the 1st day) following the onset of significant vasospasm unresponsive to other medical treatment. Dose varied between 2 to 15mg per hour. Total time of administration was 5 to 21 days (median 9,5 days). Statistically significant change in end-diastolic-velocity (EDV) on TCD was observed on the 3rd day of administration. Mean values of these parametres with confidence intervals are shown in Table 1 . In seven patients, the decrease of blood pressure was simultaneously treated by norepinephrine.CONCLUSION. Intravenous administration of ISDN seems to be a clinically advantageous alternative of NO donor in the treatment of cerebral vasospasm after subarachnoid hemorrhage. Although we found statistically significant change in end-diastolic-velocity (EDV) on TCD on the 3rd day of administration of ISDN, this finding could not be distinguished from the natural course of cerebral vasospasm. Therefore, the efect of ISDN should be further evaluated by a prospective randomized study.