The critical role of heme oxygenase in neutrophil migration impairment in polymicrobial sepsis

During severe sepsis a marked impairment of neutrophil migration into the infectious focus occurs, which is associated with dissemination of infection resulting in high mortality. We recently showed that heme oxygenase (HO) products, carbon monoxide and biliverdin, downregulate neutrophil recruitment by reducing the neutrophil/endothelium rolling and adhesion in a noninfectious inflammatory model. This study aimed to investigate a possible role of the HO-1 pathway on the failure of neutrophil recruitment in mice subjected to severe (S-CLP) polymicrobial sepsis induced by cecal ligation and puncture (CLP).


Background
The pathogenesis of endotoxemic tubular dysfunction with failure in urine concentration is poorly understood. Urea plays an important role in the urinary concentrating mechanism and expression of the urea transporters UT-A1, UT-A2, UT-A3, UT-A4 and UT-B is essential for tubular urea reabsorption. The present study attempts to assess the regulation of renal urea transporters during severe inflammation in vivo. Materials and methods By agreement of the animal protection committee C57BL/6J, mice were injected with lipopolysaccharides (LPS, 10 mg/kg) or proinflammatory cytokines. Hemodynamic, renal parameters and the expression of renal urea transporters were investigated. To clarify the role of cytokines and renal ischemia in the regulation of renal urea transporters, experiments were performed with cytokine knockout mice, mice treated with low-dose LPS (1, 5 mg/kg) as a sepsis model without induction of hypotension, glucocorticoid-treated mice, and mice with renal artery clipping serving as a model for renal ischemia. Results and discussion LPS-injected mice (10 mg/kg) presented with reduced glomerular filtration rate, fractional urea excretion and inner medulla osmolality associated with a marked decrease in expression of UT-A1, UT-A2, UT-A3, UT-A4 and UT-B ( Figure 1). Similar alterations were observed after application of TNFα, IL-1β, IFNγ or IL-6. LPS-induced downregulation of urea transporters was not affected in knockout mice with deficient TNFα, IL-receptor-1, IFNγ or IL-6. Glucocorticoid treatment inhibited LPS-induced increases of tissue TNFα, IL-1β, IFNγ or IL-6 concentration, diminished LPS-induced renal dysfunction and attenuated the downregulation of renal urea transporters. Injection of low-dose LPS (1, 5 mg/kg) also led to renal dysfunction paralleled by a downregulation of renal urea transporters without alterations in blood pressure. Renal ischemia induced by renal artery clipping did not influence the expression of urea transporters. Conclusion Our findings demonstrate downregulation of renal urea transporters that probably accounts for tubular dysfunction during sepsis. Furthermore, they suggest that downregulation of Effect of lipopolysaccharide (LPS) (10 mg/kg), dexamethasone (10 mg/kg) and the combination of both on UT-A1, UT-A2, UT-A3, UT-A4 and UT-B mRNA in the kidney 6, 12 and 24 hours after intraperitoneal injection. Values are related to signals obtained for β-actin mRNA and presented as the percentage of vehicle control. Mean ± SEM of six animals per group. *P < 0.05 versus control, # P < 0.05 versus LPS treatment.
to evaluate the efficacy of using the AS reconvalescent donor plasma for passive immunotherapy of AS.

Materials and methods
The study was conducted experimentally on 775 Wistar line rats and 38 inbreed dogs. A total of 296 patients with AS were also involved in the study; 58 formed the control group; 26 patients were selected as reconvalescent donors of plasma. Serum concentrations (ELISA) of major antibodies were determined against most significant pathogens (Escherichia coli, Staphylococcus spp., Staphylococcus aureus, Bacteroides spp., Klebsiella pneumoniae, Pseudomonas aeruginosa). Results Changes of serum antibody concentrations were time dependent and fluctuating during the current of AS forming the waveform curve. Most remarkable decreases were found during 24-72 hours of AS. Serum antibody titers to the main pathogens were slightly higher due to antibiotics and detoxification therapy. Operation by itself decreased titers from 4.42 ± 0.28 to 3.49 ± 0.25 (E. coli), and from 5.41 ± 1.02 to 3.0 ± 0.58 (P. aeruginosa). Antistaphylococcal antibody titers decreased from 7.22 ± 0.9 before surgery to 4.83 ± 0.47 after. Repeated operations alter antibody concentrations even more significantly. The highest levels of antibodies were found in patients who underwent successful treatment of AS 1-2 months prior to investigation. Their plasma was used in treatment of AS patients. Intravenous administration of two-dose 100-200 ml hyperimmune plasma per day prevented a following decrease of antibody levels, and in 98% of cases increased them (21.39 ± 1.47%). The cost of treatment was 15-37% lower if compared with traditional methods (control group). Conclusion There is exact evidence of efficacy for using hyperimmune plasma in patients with abdominal sepsis; it is more costeffective if compared with traditional methods of immunotherapy.
Background Based on the available evidence, professional societies have published practice guidelines on severe sepsis and septic shock. We started using a paper order set based on the guidelines in our medical intensive care unit (ICU) in October

Patients
Patients before Patients with Patients with with sepsis Volunteers surgery smooth recovery pneumonia Aromatic acid (ng/ml) (n = 12) (n = 16) (n = 36) (n = 9) (n = 10) 2005. This prospective study aims to describe the impact of the order set on the process of care.

Materials and methods
We included patients with severe sepsis/ septic shock treated in our ICU between October 2005 and April 2007. We collected Acute Physiology and Chronic Health Evaluation (APACHE) III derived severity data, compliance with six elements of early goal-directed therapy and hospital mortality. Compliance with each element was defined as the use of the following within 6 hours of severe sepsis/septic shock: use of central venous pressure, central venous oxygen saturation measurement, adequate fluid resuscitation and appropriate use of vasopressors, inotropes and transfusion of red blood cells. The ICU admission severity of illness and sepsis stage (severe or shock) were entered in a logistic regression model to determine the independent impact of the order set on mortality. P < 0.05 was considered significant. Results Of 561 patients (168 severe sepsis and 373 septic shock), 31 were excluded for not authorizing research. The order set was utilized in 328 (61.9%) of 530 patients. There were no significant differences in gender, age, race, and severity of illness at ICU admission between the order set and nonorder set groups.
The order set was more likely to be used in patients with septic shock than in those with severe sepsis (67.3% versus 51.4%; P = 0.0004). Compliance with all six elements occurred in 130 (39.6%) of the order set group compared with 50 (24.8%) of the nonorder set group (P = 0.0004). Although mortality did not change, compliance with five of the six elements improved significantly with the order set. Logistic regression analysis showed that shock (odds ratio (OR) = 2.384, 95% confidence interval (CI) = 1.431-3.970; P = 0.0008) and predicted APACHE III mortality (%) (OR = 1.040, 95% CI = 1.031-1.050; P < 0.0001) were associated with mortality, not the order set (OR = 0.742, 95% CI = 0.476-1.157; P = 0.1881).
Conclusions This study showed that a protocolized order set improves compliance with the standard of care in patients with severe sepsis and septic shock. However, it did not resolve some of the noncompliance problems and did not improve survival.
Background Cardiac surgery with cardiopulmonary bypass (CPB) is associated with a number of adverse effects due to systemic inflammatory response syndrome, a physiologic reaction to tissue injury. Because of this response, conventional clinical and biological signs may be misleading in the diagnosis of postoperative complications, particularly infections. The aim of the study was to evaluate the impact of autologous centrifuged shed mediastinal blood procedures in attitude of infection, estimating the predictive role of procalcitonin (PCT) and C-reactive protein (CRP) changes during the postoperative period.

Materials
We have analysed data on 90 patients, who had been subjected to cardiac surgical procedures on CPB: there are 41 patients in Group I, who were reinfused with the centrifuged autologous mediastinal blood 4 hours after the end of surgery; and 49 patients in Group II, whose shed mediastinal blood was not reinfused (control group).
Methods We studied the quantity of haemoglobin, haematocrit and leucocyte counts, and the value of CRP and PCT concentrations before the surgery (baseline), 4 and 20 hours after the end of surgery and during 5 days after surgery. Preoperative patient conditions, intraoperative and postoperative periods, were recorded. Statistical significance was accepted at a level of P < 0.05. Results In Group I, patients who were reinfused with the centrifuged autologous shed mediastinal blood, requirement for the allogenic blood transfusion procedures was significantly lower (14.6% versus 38.8%, P < 0.05). The CRP concentration was greater, but there were no significant differences between the groups in all postoperative periods. At 20 hours after the end of surgery and the second postoperative day, the increase of the PCT concentration was significant and often observed in group II (33.3% versus 58.3%), where there were significantly more complications of infection (2.4% versus 10.2% P < 0.05) and a significantly longer length of postoperative hospital stay (9.32 ± 2.55 versus 14.38 ± 4.27 days, P < 0.05).
Conclusions Our data suggest that the early reinfusion of autologous centrifugated shed mediastinal blood procedures did not increase bleeding and statistically significantly reduced the requirement of allogenic blood transfusion procedures, reduced the number of infection complications, and significantly shortened the length of postoperative hospital stay. In evaluation of postoperative infection rates, PCT is highly suggestive as a marker of postoperative complications.

P8
System approach to the diagnosis and treatment of septic arthritis in newborns Background Bone and joint sepsis occurs in 20% of newborns having perinatal sepsis. The feature of septic arthritis against the background of age-specific functional immunodeficiency gives us a reason to consider this decease as the manifestation of immunodeficiency. Septic arthritis in newborns is lethal in 10% of cases and gives orthopedic complications in 20% of cases. The starting moments of septic arthritis development are: mother's intrauterine infection and nosocomial infecting in maternity hospitals and intensive care departments, and purulent omphalitis.

Materials and methods
We present the experience of diagnostics and treatment of 180 newborns aged 3 days and older having septic affection of the hip, knee, shoulder and other joints. The crucial role in diagnostics of septic arthritis is played by: cytology Background Glomerular and microvascular thrombosis due to the activation of the inflammation and coagulation pathway contribute to the occurrence of acute renal failure in sepsis. The proteaseactivated receptors (PARs) have been shown to play an important role in the interplay between the inflammation and coagulation.

Materials and methods
We hypothesized that PAR2 blocking would improve glomerular and vascular thrombosis by attenuating the inflammation and coagulation, leading to the prevention of acute renal failure, and we assessed the effects of the PAR2 blocking peptide (PAR2 BP) in a rat model of lipopolysaccharide (LPS)-induced acute renal failure. Results Levels of TNFα were significantly expressed 1 hour after LPS administration, followed by: (i) an increase in levels of tissue factor, factor VIIa, factor Xa, thrombin and plasminogen activator inhibitor-1; (ii) unchanged levels of tissue factor pathway inhibitor; and (iii) subsequent deposition of fibrin in kidney tissues, which led to the elevation of creatinine and blood urea nitrogen. Timedependent PAR2 expression was observed at both the gene and protein levels. Immunoreactivities of PAR2 and fibrin were colocalized in the glomerulus and the other kidney tissues. PAR2 BP suppressed TNFα elevation, and attenuated activation of the coagulation, thus leading to a decrease in fibrin formation and its deposition in the glomerulus. However, the levels of creatinine and blood urea nitrogen remained unchanged.
Conclusions These results show that PAR2 plays a crucial role in the inflammatory and coagulation process of LPS-induced renal failure and may in part participate in the pathogenesis of the disease. Methods We performed a 'before and after' evaluation of the critical pathway concerning 184 critically ill patients sequentially admitted throughout a 16-month period. Results A total of 184 patients with severe sepsis or septic shock entered the study. Ninety-four patients had their analysis performed before the implementation of the standardized protocol (the 'before' group), and 90 patients were managed following the implementation of the protocol (the 'after' group). Basal demographic variables and the severity of illness score (APACHE II) were similar for both groups. Patients in the 'after' group had statistically more cultures obtained before institution of antibiotics and more patients received antibiotics in a due time (2 hours from diagnosis). In addition, those patients received more corticosteroids and activated protein C ( Table 1). The intensive care unit length of stay and the hospital length of stay were similar in both groups. Remarkably the hospital mortality rate was significantly lower (34.4%) in the 'after' group in septic shock patients (67.7% versus 44.4%, P < 0.04) (Figure 1).

Conclusions
The implementation of the Surviving Sepsis Campaign guidelines through a standardized protocol was associated with a 34% reduction in septic shock-related hospital mortality. Background Sepsis is a common source of morbidity and mortality among critically ill patients. Targeting measures to reduce the incidence of and to promote early recognition and treatment of sepsis is at the forefront of many critical care initiatives. Advances in the management of severe sepsis have evolved over recent years in an attempt to combat the spiraling mortality trends. The Surviving Sepsis Campaign (SSC) is a worldwide initiative promoting the evidence-based treatment of sepsis, with the explicit goal of reducing both the morbidity and mortality associated with sepsis. This study was conducted to assess the clinical relevance of the early physiologic screening criteria advocated by early goaldirected therapy for sepsis, and the Surviving Sepsis Campaign guidelines.

Materials and methods
The Project IMPACT ® database was used to obtain a sample of patients (n = 363) with an ICU admission diagnosis of sepsis and a random acuity-matched comparison sample of patients with an admission diagnosis (n = 364) other than sepsis. Results Significant group differences were found on all physiologic monitoring variables tested (high temperature, P = 0.000; low Mortality rate reduction in septic shock patients (42 deaths in 'before' group versus 32 deaths in 'after' group). P < 0.04. S7 temperature, P = 0.001; heart rate, P = 0.004; respiratory rate, P = 0.005; and mean arterial pressure, P = 0.000). In the logistic regression model, high temperature and mean arterial pressure functioned as significant predictors, with odds ratios of 2.12 for temperature at or above 38°C and 3.87 for MAP less than 70 mmHg. The odds ratio of having sepsis was 4.63 if both of these predictors were present. Conclusions It is important to understand the value of common monitoring parameters in the early identification of sepsis, since those parameters are continuously monitored and readily available. It is the responsibility of bedside clinicians to assure that the parameters chosen for monitoring provide the most accurate reflection of the patient's clinical status. These results provide some support for the use of the currently recommended criteria for physiologic monitoring in the early identification of patients at risk for developing sepsis. Furthermore, if this could be done automatically, it would probably shorten the recognition time and thus speed up the initiation of sepsis treatment. ProtocolWatch is a tool that offers an electronic version of the SSC guidelines, screens the physiologic criteria automatically, and is resident on a bedside patient monitor. Development of tools such as ProtocolWatch will probably be an important adjunct to sepsis identification and treatment in the future.

P14
The role of CC-chemokine receptor 4 in murine polymicrobial sepsis Background As part of the innate immune system, antimicrobial peptides (defensins and cathelicidins) are produced by both circulating and epithelial cells. Cathelicidins have been reported as an essential component against group A streptococcal (GAS) skin infections. However, bacterial factors such as streptococcal pyrogenic exotoxin B (SpeB) may inactivate these peptides. We have studied the interaction between GAS and the human cathelicidin LL-37, by use of patient tissue material. Methods and materials Thirty-seven biopsies from 17 patients suffering from GAS skin and soft-tissue infection were obtained and graded according to disease severity (erysipelas, cellulitis, necrotizing fasciitis). Three additional biopsies served as negative controls. Tissue sections were immunostained for LL-37, GAS, SpeB and specific cell markers. Sections were investigated by light and confocal microscopy, and results were quantified by in situ imaging.
Results High expression of LL-37 was detected in erysipelas and severe soft-tissue infections, and showed a significant positive correlation to bacterial load (P < 0.001 and P = 0.042, respectively). Confocal microscopy identified neutrophils as the main source of LL-37 at the epicenter of infection, and the degree of neutrophil infiltration showed a significant positive correlation to LL-37 levels (P < 0.001). LL-37 and SpeB were detected in the same biopsy areas, and colocalization was confirmed by confocal microscopy.
Conclusions Despite the high expression of LL-37 in close proximity to streptococci at the local site of infection, there seems to be a significant lack of antimicrobial effect, as evident by the bacterial load. The colocalization of SpeB and LL-37 suggests that this streptococcal factor probably contributes significantly to a resistance mechanism towards antimicrobial peptides at the local tissue site. Background High mobility group box-1 (HMGB1) is an intracellular protein that is secreted by activated immune cells during inflammation, or is passively released by cells undergoing necrosis. HMGB1 acts both as a proinflammatory cytokine and a chemokine, and has been identified as a late mediator of sepsis. We have studied HMGB1 expression in streptococcal skin and soft-tissue infections, and its relation to bacterial load as well as to infiltration of inflammatory cells. Materials and methods Thirty-seven biopsies from 17 patients suffering from streptococcal skin and soft-tissue infection were obtained and graded according to disease severity (erysipelas, cellulitis, necrotizing fasciitis). Three additional biopsies served as negative controls. Tissue sections were immunostained for HMGB1, group A streptococcus, and specific cell markers. Sections were investigated by light microscopy, and results were quantified by in situ imaging. Results HMGB1 was found both intracellularly and secreted in the tissue. Its expression increased in parallel to disease severity and was significantly higher in necrotizing fasciitis than in erysipelas (P = 0.023). HMGB1 showed a positive correlation to neutrophils (P < 0.01) in erysipelas, but not in severe infections. A correlation to bacterial load was not found.

P16
Conclusion In contrast to erysipelas, large amounts of necrotic tissue are present in severe skin infections, which probably contribute considerably to the expression of HMGB1. The high values may disturb a statistical correlation to the degree of inflammatory cell infiltration in the tissue. However, our results suggest that the massive HMGB1 expression at the local site of infection is probably an important mediator and enhancer of inflammation in skin tissue and soft-tissue infections, as evident by its expression in correlation to disease severity.

P17
Can procalcitonin reflect the etiology of the bacteremia? Background Hyperglycemia is common among patients admitted to intensive care units (ICUs) and carries risk for various complications, especially sepsis, and prolonged ICU stay. Some studies have demonstrated that intensive insulin control of blood glucose reduced morbidity and mortality. The objective was to determine whether intensive or conventional insulin control of blood glucose in hyperglycemic septic and nonseptic ICU patients is correlated with the prognosis. Materials and methods Septic and nonseptic ICU patients with hyperglycemia were randomly assigned to a group treated intensively with insulin targeting glucose at 6.6-8.4 mmol/l or to a conventional insulin therapy group where glucose upon exceeding 12 mmol/l was controlled at 8.4-12 mmol/l. Rates of morbidity and mortality, hypoglycemic episodes and required insulin dosage were compared. Results A total of 89 patients were enrolled, including 27 patients with sepsis: 11 patients were treated with insulin intensively with a mean glucose of 8.3 mmol/l, while 16 patients received conventional insulin treatment with a mean of 10.3 mmol/l. Thirty nonseptic patients received intensive insulin treatment with a mean of 8.46 mmol/l and 32 nonseptic patients were treated conventionally with a mean of 10.4 mmol/l. Among septic patients, both groups were similar with respect to age and Acute Physiology and Chronic Health Evaluation scores. There was no significant difference between groups in the morbidity, including rates of new infection, renal and hepatic damage. There was a somewhat shorter ICU stay in the intensive treatment group. Both groups had similar ICU, inhospital and 28-day follow-up mortalities and similar rates of hypoglycemic episodes. The daily dosage of insulin was higher with the conventional treatment. Similar results were obtained among nonseptic patients between both groups, but septic patients had a longer total ICU stay and higher mortality. Conclusions Intensive insulin control of blood glucose at 8.4 mmol/l does not affect the mortality or morbidity of septic and nonseptic patients in intensive care, except for a somewhat shorter ICU stay. An increased insulin dosage in the conventional treatment group was attributed to the group's higher initial blood glucose, probably due to a higher prevalence of diabetes and associated insulin resistance and toxicity hyperglycemia. Background Vasopressin agonists, such as arginine vasopressin (AVP) and terlipressin (TP), are increasingly used to stabilize haemodynamics in catecholamine refractory hyperdynamic septic shock. However, it is still not fully understood if and how low-dose infusion of both drugs impacts on mesenteric blood flow (Qma) and outcome. The present study was conducted as a prospective, randomized, controlled laboratory experiment to compare the effects of AVP and TP on Qma and mortality in an established model of ovine septic shock. Materials and methods Fifteen ewes were anaesthetized and instrumented for chronic haemodynamic monitoring. A median laparotomy was performed to place a flow-probe around the superior mesenteric artery and to take faeces from the caecum under sterile conditions. After the gut and abdomen had been closed and baseline measurements (BL1) taken, the faeces were injected into the peritoneal cavity. After the onset of septic shock (defined as mean arterial pressure (MAP) < 60 mmHg), a second set of measurements (BL2) was taken. The animals were then randomly assigned to receive either AVP (0.5 mU/kg/min) or TP (1 μg/kg/hour). The control group received only the vehicle (normal saline). Norepinephrine was titrated to maintain MAP at 70 ± 5 mmHg in all groups. Systemic haemodynamics, global oxygen transport including arterial lactate concentrations, gas exchange, electrolytes and Qma were determined at baseline and following every hour after the onset of septic shock. Animals surviving the 12-hour study period were deeply anaesthetized and killed by an overdose of saturated potassium solution. Mortality was analyzed by the Kaplan-Meier survival analysis. All the other variables were compared using two-way analysis of variance with appropriate post-hoc comparisons.

Effects of vasopressin and terlipressin in ovine septic shock on mesenteric blood flow and survival
Results The Qma and electrolytes were similar between groups. However, systemic haemodynamics and global oxygen transport were stabilized more effectively in both treatment groups versus control animals. Notably, continuous infusion of AVP and TP significantly prolonged survival as compared with the control group (P < 0.05 each; Figure 1). There were no differences between the treatment groups.
Conclusions In this clinically relevant large animal model of septic shock, low-dose infusion of AVP and TP did not impair the Qma, but stabilized systemic haemodynamics and prolonged survival. Our data suggest that early infusion of AVP or TP may be beneficial in catecholamine-refractory septic shock. Methods We studied PMN CD64 in three clinical groups: neonates (390 specimens), hospitalized patients (236 specimens), and ambulatory outpatients (96 specimens). PMN CD64 was measured as an index using the Leuko64 (Trillium Diagnostics). Samples were also processed in parallel for the measurement of CD64 using the Leuko64 assays for flow cytometry on a FACScan (Becton Dickinson) and the blood cell counter Cell Dyn Sapphire (Abbott Diagnostics). Data were analyzed using Leuko64 software (Trillium Diagnostics). Results from both platforms were expressed as the PMN CD64 index, the monocyte CD64 index, and the monocyte CD163 index. CRP was measured in parallel. Plasma samples were stored at -70°C for subsequent measurement of procalcitonin (B.R.A.H.M.S.) and soluble CD163 (Trillium Diagnostics).
Results PMN CD64 correlated best with CRP, closely followed by PCT, but the degree of correlation varied among the clinical groups. The correlation was best in neonates (r = 0.592 for CRP and r = 0.391 for PCT), followed by hospitalized patients Available online http://ccforum.com/supplements/11/S4

Figure 1 (abstract P19)
Survival of animals over time.

S10
(r = 0.345 for CRP and r = 0.349 for PCT), and less so in outpatients (r = 0.251 for CRP and r = 0.257 for PCT). The correlation between PMN CD64 and the soluble markers was higher than that between CRP and PCT (r = 0.331 for hospitalized patients, r = 0.305 for neonates, and r = 0.196 for ambulatory patients Background Severe sepsis and septic shock are common disease processes in the critically ill and are associated with substantial morbidity and mortality. The importance of the early identification and diagnosis of severe sepsis has been highlighted by the Surviving Sepsis Guidelines with the aim to provide early and aggressive management in order to improve outcome. In contemporary practice, all clinicians have the responsibility of identifying severe sepsis. Therefore the objectives of this study were to determine whether emergency department and intensive care clinicians could identify and diagnose severe sepsis in those patients in their care within the first 24 hours of admission, and to identify predictors of failing to diagnose sepsis. Methods The patient cohort were prospectively screened and enrolled on admission to intensive care within the first 24 hours. Severe sepsis was defined as new-onset acute organ dysfunction, using consensus criteria. Clinical data and physiological parameters were collected prospectively. Diagnosis was based on microbiologically confirmed clinical findings. Clinicians caring for each patient were prospectively surveyed. Results All 402 subjects had infection. Infection sites included 52% pneumonia, 17% urinary, 15% abdominal, 6% wound and skin, and 10% isolated organs and bone. Single-organ failure was evident in 21%, 42% had two-organ failure, 29% had three-organ failure and 8% had four-organ failure. Nurses identified sepsis in 141 of the 402 patients (P < 0.001) whereas physicians did so in 265 of the 402 patients (P < 0.05). Misdiagnosis of severe sepsis by the attending nurse or physician was more likely to be associated with pneumonia (odds ratio (OR) = 4.2 (95% confidence interval (CI) = 3.6-4.2), P < 0.01), urinary sepsis (OR = 2.9 (95% CI = 2.6-3.4), P < 0.5), less than three-organ failure (OR = 3.1 (95% CI = 2.4-3.7), P < 0.01), Gram-negative infection (OR = 2.3 (95% CI = 1.6-3.5), P < 0.5) and presenting without fever (OR = 3.5 (95% CI = 3.1-3.9), P < 0.05). Thirty-two percent of clinicians did not know the criteria for severe sepsis and 57% missed the patient diagnosis in their care at that time.
Conclusion In this study, misdiagnosis of severe sepsis is still an acknowledged problem in meeting the goals of early resuscitation.

Materials and methods
Plasma from preconditioned (Caniplas ® ) and normal dogs (FFP) was provided blind to the study by a commercial supplier (Plasvacc Pty Ltd). In vitro anti-TNFα activity in canine donor plasma was determined by a L929 murine cell TNFα inhibition bioassay using recombinant murine TNFα. In vivo effects were tested by a rat subcutaneous skin pouch model. Rats were pretreated for 3 days with either Caniplas ® , FFP, saline (2 ml/day, s.c.) or carprofen (5 mg/kg, s.c.) and inflammation was induced by injecting monosodium urate crystals into the pouch (5 mg/ml in 5 ml saline). Fluid was taken from pouches at specified intervals for cell count, TNFα and IL-6 analysis. Data analysis: normalized data were fitted to a four-parameter logistic curve. The fitted midpoints were compared statistically for datasets using an F-test and calculated fitted hill slopes.

Results
In the rat skin pouch model, both Caniplas ® and FFP reduced TNFα levels and Caniplas ® was a more potent antagonist. The heightened anti-TNFα activity of Caniplas ® compared with FFP was confirmed in the in vitro cell bioassay ( Figure 1). Neither Caniplas ® nor FFP reduced inflammatory cell infiltration or the levels of IL-6. Conclusion While we need to confirm the mechanism, we report that preconditioning with endotoxin does illicit specific anti-TNFα activity and that this observation has been confirmed in both in vitro testing and in vivo animal models. It is plausible that preconditioning animals with endotoxin induces an increase in the concentration of soluble TNFα receptors I and II in donor plasma, and that this is the probable source of TNFα antagonism. This report suggests that preconditioned plasma may be a beneficial treatment where inflammation causes increased expression of TNFα.

P24
The Background Septic shock includes blood vessel dilatation and activation of innate immunity. The activation causes release of antimicrobial peptides such as LL-37. It has been shown that LL-37 can attract leukocytes via the lipoxin A4 receptor (ALX). ALX is also present in vascular endothelial cells. To explore possible ways of pharmacological intervention in septic shock, we investigated whether LL-37 can affect vascular tone. Materials and methods Human omental arteries and veins were obtained during abdominal surgery. The circular smooth muscle activity, in the wall of the vessel segments, was studied in organ baths. Gene expression was studied using reverse transcriptase PCR.
Results LL-37, at micromolar concentrations, induced a concentration-dependent and endothelium-dependent relaxation in vein segments but not in artery segments precontracted by endothelin-1. The relaxation was profoundly reduced by potassium chloride (30 mM) to inhibit endothelium-derived hyperpolarizing factor (EDHF), while it was less affected by the NOS inhibitor L-NAME and not at all by indomethacin. The ALX agonist, WKYMVm, did also induce a relaxation, and both the relaxations induced by LL-37 and WKYMVm were inhibited by the ALX antagonist WRWWWW. ALX was expressed in the endothelium. Conclusion We demonstrate for the first time that the human antimicrobial peptide LL-37 induces endothelium-dependent Available online http://ccforum.com/supplements/11/S4

Figure 1 (abstract P23)
In vitro TNFα dose-response to canine sera. relaxation in human omental veins mediated via an effect on endothelial ALX. The relaxation involves the release of nitric oxide and EDHF but not prostanoids. LL-37 released from white blood cells could contribute to the blood vessel dilatation during sepsis and treatment with ALX antagonists such as WRWWWW might be successful.

Materials and methods
We report the effects of PC in six HUS patients. HUS is characterized by a simultaneous occurrence of hemolytic anemia, thrombocytopenia and acute renal failure. Postdiarrheal HUS is often based on an infection with enterohemorrhagic Escherichia coli producing Shiga toxins. Our current pathogenetic understanding is that Shiga toxins cause endothelial injury, leading to thrombotic microangiopathy. There is still a 5% rate of mortality particularly caused by cerebral involvement. We treated six children with a severe cerebral manifestation, five of them suffered from a multiple organ dysfunction syndrome (MODS), of HUS with PC over 7-10 days. All patients suffered peritoneal dialysis, one patient a plasmapheresis. In addition to the treatment of the MODS, all of the patients received 100-200 U/day PC. Results All of the patients showed signs of disseminated intravascular coagulation. We found typical hypodense lesions in basal ganglia and edema of the brain in computed tomography. During the therapy with PC, MODS was remarkably improved and abnormal D-dimer and plasminogen activator inhibitor 1 levels could be normalized. All of the patients recovered a nearly normal kidney function. Two patients persisted in a severe reduced neurological status. The others showed only slight or no neurological disabilities on discharge. No adverse effects were observed with the PC concentrate administration. Conclusion There is no generally accepted therapy regimen to treat HUS in case of neurological involvement. Mortality in HUS accompanied with cerebral microangiopathy is high and difficult to alter. This is the first trial of human PC concentrate administration to anticipate thrombotic microangiopathy in HUS. All of our patients showed rapid clinical improvement under PC administration. Four of six patients were discharged in a healthy condition despite their severe disease. The containment of the severe neurological involvement and the lack of side effects in the treatment with human PC concentrate administration in our patients yield hope that PC treatment may be an effective therapy regimen in the treatment of severe HUS.

Background
The clinical picture of sepsis is varying, and the severity of the disease is influenced by numerous factors including the infectious agent and the host genetics. P. aeruginosa and S. aureus, the frequently sepsis-causing agents, can switch from a planktonic to a biofilm lifestyle. The cell wall of P. aeruginosa contains lipopolysaccharide (LPS), which can detach and trigger the immune response in the patient. We investigated the effects of live planktonic or biofilm bacteria on human monocytes in terms of the production of proinflammatory cytokines, such as TNF, in order to determine whether these two bacterial lifestyles influenced the immune response during sepsis. Materials and methods Clinical isolates of P. aeruginosa and S. aureus, bacterial culture plates and media (Luria-Bertani and tryptic soy broth), RPMI media, ELISA kits, a spectrophotometer and biofermentors were used. The same bacterial concentration of planktonic and biofilm forming P. aeruginosa and S. aureus strains was obtained by optical density measurements and confirmed by colony counting. Blood samples were collected from healthy donors and monocytes were isolated by adherence and further incubated together with live planktonic or biofilm forming strains of P. aeruginosa or S. aureus for 5 hours; the cytokine content was measured by ELISA. LPS was extracted from P. aeruginosa in order to investigate the structural differences between planktonic and biofilm derived LPS. The limulus amebocyte lysate test, SDS-PAGE gel electrophoresis and mass spectrometry of LPS were performed to analyze the LPS structure. Results The production of TNF, IL-6 and IL-1α was increased in monocytes incubated with biofilm forming P. aeruginosa as compared with planktonic ones, whereas no difference between cytokine responses was observed in monocytes incubated together with planktonic or biofilm S. aureus. Two predominant forms of rough LPS were detected in planktonic P. aeruginosa by SDS-PAGE and one of the rough LPS bands was absent in the biofilm. Fatty acids differed by their level of hydroxylation in the two bacterial growth conditions as seen by mass spectrometry. Conclusions Biofilm forming P. aeruginosa induces an enhanced inflammatory response in human monocytes compared with the planktonic bacteria, and LPS structures were found to be different. No difference was seen in response to S. aureus planktonic or biofilm bacteria. The biofilm P. aeruginosa was more immunostimulatory than the planktonic form. LPS from biofilm forming bacteria may increase the immune response during sepsis. Background Coagulase-negative staphylococci (CoNS) are important causes of bloodstream infection, especially in immunocompromised patients, but they are also the most common contaminants of blood cultures; discrimination between these two is often difficult. The time necessary for microbial growth to appear (time to positive) has been used as a laboratory marker for assessing clinical significance of CoNS bacteremia. However, with the use of continuously monitoring blood culture systems, the previous study results are controversial. The aim of the present study was to assess microbiological laboratory markers that are suggestive of true CoNS bacteremia.

Materials and methods All blood cultures positive for CoNS between 1 October 2006 and 30 April 2007 in Tartu University
Hospital were included in this analysis. Blood specimens were monitored with the BACTEC 9240 system. Microbes were identified using the VITEK 2 system and antibacterial susceptibility pattern was tested according to CLSI standards. The CDC definition for bloodstream infection to determine the clinical significance of CoNS was used. Results A total of 109 CoNS blood isolates from 86 patients (51 male; median age 22 years, range 0-94 years) were identified. According to the CDC criteria, 81 isolates were contaminants and 28 were true causes of bacteremia. Fifteen of the patients with infection were from the intensive care unit. The time to positive for blood cultures in infected patients was shorter than that in contaminated subjects: median 30 hours (range 20-68 hours) versus 42 hours (range 16-116 hours), respectively (P = 0.029). A total of 10 different CoNS species were identified; S. epidermidis was most commonly isolated in both groups -21/28 in infection and 49/81 in contamination (OR, 1.87; 95% CI, 0.77-4.56). S. hominis comprised a higher proportion in the contamination group than in the infection group (17/81 versus 2/28 respectively; OR, 2.59; 95% CI, 0.84-8.01). Conclusions The majority of CoNS isolated from blood cultures were contaminants. The time necessary for microbial growth is an important laboratory marker in differentiating between true bacteremia and contamination. To induce septic shock, mice were injected intraperitoneally with a dose of 50 mg LPS/kg body weight. To induce respiratory inflammation, mice were exposed for 20 minutes to aerosolized LPS (500 μg/ml saline) in a chamber connected to an air nebulizer (Miko, CA-MI s.n.c., Italy). To investigate the effects of TEL, mice received a single dose of 20 mg ketolide/kg body weight by intraperitoneal injection, 1 hour prior to the administration of LPS. Heparinized blood samples were centrifuged, and plasma was stored at -20ºC until cytokine determination. To obtain bronchoalveolar lavage (BAL) fluids, the lungs were lavaged with 1 ml phosphate-buffered saline through an intratracheal catheter. Enzyme immunoassays kits were use to determine TNFα (Pierce Endogen, Rockford, IL, USA) and macrophage inflammatory peptide 2 (MIP-2) (R&D Systems, Minneapolis, MN, USA). The differences in cytokine levels were analyzed using Student's t test. P < 0.05 was considered significant. Results When mice were intraperitoneally challenged with a lethal dose of LPS, TEL protected 50% of animals and significantly reduced the plasma levels of TNFα at 2 hours after LPS administration. In the respiratory inflammation model, the treatment with TEL significantly reduced the BAL levels of TNFα and MIP-2 at 4 hours post endotoxin (Table 1). Conclusions TEL exerts anti-inflammatory activity in vivo that may contribute to its pharmacological effectiveness in the treatment of respiratory infections and their possible progression to septic shock.  Background Migration inhibitory factor (MIF), a critical proinflammatory mediator in sepsis, has a profound affect on cardiovascular function. Our animal studies show that the lungs release MIF into the systemic circulation during late sepsis. MIF released in this way has direct and immediate access to cardiac cells. The purpose of our study was to assess the lung as a source of MIF in human septic shock patients and to further study the MIF-associated pathways involved in cardiovascular depression. Materials and methods Nine adult patients with septic shock in the medical intensive care unit were studied. Blood samples were collected pre-lung (pulmonary artery or central venous catheter) and post-lung (arterial line catheter) at 12, 24 and 48 hours from the time of diagnosis. MIF was measured using an ELISA, and differences in plasma MIF concentration pre-lung versus post-lung were assessed using a paired t test on signed ranks. Furthermore, since inhibition of caspase 3 activity during sepsis reduces myocardial apoptosis and cardiac dysfunction, we examined in vitro the effect of MIF on cardiomyocyte apoptosis.

Results
The mean age of patients was 57.6 years (range 25-82 years). Bloods from six patients were culture positive. There was a wide variation in plasma MIF concentrations in both pre-lung (0.2-64.7 ng/ml) and post-lung (0.2-76.4 ng/ml). However, there was a significant increase in the median MIF level of the post-lung blood (3.9 ng/ml) compared with pre-lung blood samples (2.9 ng/ml, P = 0.005). At 48 hours post diagnosis, eight out of nine individuals had increased MIF concentration post-lung, with a mean increase of 64 ± 49%. These findings were independent of the source or nature of the infection. The basal level of apoptosis in primary cardiomyocytes cultured in defined, serum-free medium was 3.7 ± 0.9% (% TUNEL positive of total nuclei). This increased significantly (P < 0.001) to 15.5 ± 3.9% and 26.0 ± 7.1% when treated with MIF 20 and 30 ng/ml, respectively. Simultaneous treatment of cardiomyocytes with MIF and specific MIF inhibitor ISO1 (50 mM) resulted in significant attenuation of the apoptotic effect, returning apoptosis to a basal level at 4.1 ± 0.6%. Conclusion The study demonstrates for the first time in humans that the lung is a major source of MIF in septic shock. The study further suggests that MIF released from the lung may reduce cardiac function by increasing cardiomyocyte apoptosis, and that blocking MIF released from the lung to prevent cardiocirculatory deterioration may provide new strategies for the treatment of sepsis. Background Sepsis induces an increase in vasopressin (AVP) secretion and massive production of inflammatory mediators, including leukotrienes (LTs) and nitric oxide (NO), which are also produced in the brain and may affect the neuroendocrine system. Our aim was to study the role of central LTs and NO in AVP secretion during polymicrobial sepsis induced by cecal ligation and puncture (CLP). Methods Male Wistar rats (250-280 g) received an intracerebroventricular injection of MK-886 (1 μg/kg), an LT biosynthesis inhibitor, or L-NAME (250 μg), an NO synthase inhibitor. Controls received a carrier injection before CLP, or were only sham operated. In one group of animals, survival was monitored for 5 days. In another group, the animals were decapitated at 4, 6 or 24 hours after surgery and blood was collected for measurements of hematocrit, plasma nitrate (NOx) and AVP levels.

Inhibition of central leukotrienes and nitric oxide production affects vasopressin secretion induced by polymicrobial sepsis
Results CLP caused an increase in plasma AVP levels in the initial phase of sepsis, which was blocked by the central administration of MK-886 and was elevated by intracerebroventricular-injected L-NAME (P < 0.05). This contrasted with the final phase of sepsis, when plasma AVP remained at basal levels and was not altered by the administration of LT and NO blockers. The gradual increase in NOx levels was reduced by MK-886 and blocked by L-NAME. The increase in hematocrit caused by CLP was diminished (P < 0.05) by L-NAME injection, but was not modified by central administration of MK-886. The low survival rates observed after CLP were improved by the central administration of both drugs (P < 0.05).
Conclusion These results suggest that central LTs and NO differentially affect AVP secretion during CLP-induced experimental sepsis. Acknowledgement Supported by FAPESP. Background As fungi are considerably more slow growing than bacteria and are not affected by broad-spectrum antibiotics, too often conventional diagnostics fail to identify fungal bloodstream infection and confirmation is made postmortem. We describe the initial tests of a Fountain Flow™ cytometer (FFC) to detect pathogenic fungi in human blood, rapidly and cost-effectively. This innovative technique is based on fluorescence in which a stream of solution containing the cells, labeled with a fluorescent stain, is illuminated with an LED (Figure 1). The resulting fluorescence is detected with a CMOS imager.

Materials and methods
The FFC used here was used by us in previous published research to detect individual bacteria and amoebae in river water at concentrations ≥200 bacteria/ml and 0.06 amoebae/ml. An optical stimulant of blood was made by mixing red blood cells (RBCs) in saline to obtain a RBC concentration comparable with that of human blood. This was further diluted 1:20 in distilled water to reduce the opacity. Then the solution was inoculated with Candida albicans from the American Type Culture Collection to obtain samples of varying concentration. To label the Candida, FUN1, a fluorescent, fungusspecific dye (Invitrogen), was used. The dye Trypan was used to suppress background, especially from RBCs, which absorb small amounts of FUN1. Three 0.1 ml samples at each concentration were flowed into the FFC in 200 seconds (although a much greater rate was possible), while 500 images were taken continuously. The images were then analyzed (in minutes) with custom software to count the fluorescent spots in the images corresponding to Candida. Plate counts of Candida on YM agar (18 hours incubation) were used for comparison. Results Figure 2 shows the comparison of the FFC and plate counts.
Conclusions The FFC has potential as a system for low-cost, realtime detection of low concentrations of fungi in human blood. Further work is required to lower the false detection rate, to increase the detection efficiency, and to perform tests on human blood with its additional complexity. Background The optimal duration of antibiotic therapy for hospitalacquired pneumonia (HAP) and ventilator-associated pneumonia (VAP) is not clear. A multicentre randomised controlled trial indicated similar clinical efficacy for 8 versus 15 days of antibiotic therapy for VAP, with less emergence of multiresistant organisms following the shorter course [1]. Pseudomonas aeruginosa is difficult to eradicate, however, and American Thoracic Society guidelines for treatment of HAP due to P. aeruginosa recommend 14 days of therapy [2]. Our aim was to study the rate of recurrence following treatment of Gram-negative HAP in a critically ill population with short-course (typically 5 days) antibiotic therapy.

Materials and methods
We retrospectively reviewed 50 patients treated consecutively with short-course (typically 5 days) antibiotic therapy for Gram-negative HAP in a UK teaching hospital critical care unit from 2004 to 2007. Pneumonia was defined as semiquantitative respiratory culture (≥2+) of a single Gram-negative isolate, clinical pulmonary infection score ≥6 and initiation of antibiotic therapy. Recurrence of HAP was defined either as relapse (pure growth of the organism causing the initial infection) or reinfection (due to a different organism). Patients were studied until hospital discharge or death. Results Demographic and outcome data are summarised in Table 1. The commonest causative organisms were P. aeruginosa Available online http://ccforum.com/supplements/11/S4

Figure 1 (abstract P32)
The Fountain Flow™ cytometer. Fluorescently stained cells flow through the flow cell toward the CMOS camera, illuminated by an LED or laser. Cell(s) in the CMOS camera focal plane are imaged by the camera through the flow cell window at the wavelength of fluorescence.

Figure 2 (abstract P32)
Fountain Flow™ cytometer counts of C. albicans in 0.1 ml red blood cells. Comparison of Fountain Flow™ cytometer (FFC) counts and plate counts of C. albicans spiked into 1:20 diluted red blood cells (RBCs). The line of best fit gives a 52% counting efficiency. S16 (42%), Enterobacter species (14%) and Klebsiella species (14%). Two patients died before completing the initial course of antibiotics.
Conclusions Treatment of Gram-negative HAP in the critically ill patient with short-course antibiotic therapy is associated with a low rate of recurrence (10.4%). This compares favourably with reported recurrence rates of 18-26% following ≥2 weeks of antibiotic therapy for VAP [1,3,4]. Background Early diagnosis of sepsis improves outcomes for trauma patients. The aim of this study was to detect additional markers that help to recognize sepsis before manifestation in patients with severe multiple trauma and blood loss. It is known that depletion of fibrinolysis promotes microvascular alterations after massive blood loss. During inflammation, C-reactive protein may inhibit fibrinolysis by inducing plasminogen activator inhibitor 1. These alterations are correlated with an increased risk of multiple organ failure and death. In accordance with the above-mentioned purpose, we investigated fibrinolysis at the early post-traumatic period in patients with and without sepsis.

Materials and methods
We examined a database of 58 intensive care unit patients with trauma and shock, Acute Physiology and Chronic Health Evaluation II score (first day in intensive care unit) of 27.4 ± 0.5. Group 1 consisted of 23 patients who developed sepsis (day after admission, when severe sepsis or septic shock manifested, was 6.1 ± 0.28). Group 2 consisted of 35 patients without sepsis within the early post-traumatic period (1-14 days after injury). The plasma fibrinolytic potential was evaluated by the euglobulin lysis time, determined by a semiautomatic method.
Results Data are presented as the mean ± median (Table 1). It was established that fibrinolytic potential was significantly lower in patients with sepsis versus patients without sepsis since the fifth day after admission. At the same time, for white blood cells, a significant difference was observed just after the seventh day following admission. Platelets did not differ between groups. Conclusion Plasma fibrinolytic potential is closely related to inflammation and may be used as a marker of sepsis. Moreover, plasma fibrinolytic potential may be a useful test for early identification of the high-risk sepsis patient.

P35
A functional microsatellite in the macrophage migration inhibitory factor gene influences susceptibility to meningococcal sepsis

S17
Methods A case-control study of 1,106 children and adults with meningococcal sepsis and 626 control subjects and a familybased study (106 families with one afflicted child) to analyze transmission of MIF alleles using the transmission disequilibrium test. The -794 CATT5-8 microsatellite and -173*G/C single nucleotide polymorphism were detected by PCR. MIF promoter alleles cloned into a luciferase reporter construct were tested in resting and stimulated THP-1 human monocytic cells. DNA binding activity was assessed by EMSA.
Results Compared with control subjects, the frequency of the CATT5-5 genotype was markedly reduced in meningococcal sepsis patients (5.3% versus 8.8%; OR = 0.6, 95% CI = 0.4-0.9, P = 0.01 for CATT5-5 versus all other CATTx-x). The frequencies of the -173*G/G, -173*G/C or -173*C/C genotypes was comparable in meningococcal sepsis patients and control subjects (P = 0.75). The transmission disequilibrium test in families with one afflicted child revealed that the CATT5 allele was preferentially not transmitted (P = 0.002) to meninogoccal sepsis offspring, while the opposite situation occurred for the CATT6 allele (P = 0.024).
In vitro studies showed lower activity of CATT5 promoter compared with all other CATT promoters in resting and N. meningitidisstimulated THP-1 cells. Consistently, DNA binding activity to the CATT region of the MIF promoter increased with an increasing number of the CATT repeats.
Conclusion The frequency of the -794 CATT5-5 genotype was reduced in children with meningococcal sepsis and was less frequently transmitted from parents to affected offspring, suggesting that it may confer protection against severe sepsis due to N. meningitidis. Reduced transcription factor DNA-binding activity to CATT5 and weaker CATT5 promoter transcriptional activity provided a functional relevance of the CATT5 polymorphism. MIF polymorphisms might help to identify patients who may benefit from anti-MIF treatment strategies. The aim of the study was to establish a standardized and reproducible model of MRSA-related ovine septic pneumonia and to compare the pathophysiological responses to those seen in a previously established Pseudomonas aeruginosa-induced ovine sepsis model.

Materials and methods
Twenty-four sheep were operatively prepared for chronic study. After 5 days recovery, tracheostomy was performed under anesthesia and injury was given. Sheep were randomly allocated into four groups: (1) Sham group, no injury (n = 6); (2) Smoke group, exposed to smoke inhalation (n = 6); (3) PS group, exposed to smoke inhalation and instilled with P. aeruginosa (2.5 × 10 11 CFU) into the lungs by bronchoscope (n = 6); and (4) MRSA group, exposed to smoke inhalation and instilled with MRSA (2.5 × 10 11 CFU) into the lungs by bronchoscope (n = 6  Table 1). The excessive nitric oxide in the MRSA group was associated with a significant increase in lung tissue vascular endothelial growth factor mRNA and its protein expression. Conclusion The severe vascular leakage syndrome seen in the MRSA group may be due to excessive production of nitric oxide and vascular endothelial growth factor, a potent permeability factor. MRSA is largely responsible for these pathological changes rather than smoke inhalation alone.

Materials and methods
In 11 patients fulfilling the ACCP/SCCM criteria for severe sepsis, we analyzed serum levels of GM-CSF and IL-8 at days 1, 2 and 7 after admission. Patients (10 males/one female) had a median age of 68 years (range 22-87 years). The primary infection had urinary (seven patients) and respiratory (four patients) origins. Seven healthy volunteers served as controls. Results Four out of 11 patients had positive blood cultures (Gramnegative). At admission, the median procalcitonin was 2 ng/ml (0.5-10 ng/ml). All patients survived except one patient that died 5 days after admission following acute myocardial infarction. GM-CSF was always below the detection limit (7.8 pg/ml) in septic patients and controls. In seven patients, serum IL-8 was below the detection limit (31.2 pg/ml) at all time points. The median value for IL-8 for the group of patients was 0 pg/ml (0-264 pg/ml), but while IL-8 was absent in the sepsis of respiratory origin it had a median value of 32 pg/ml in urinary sepsis. Surprisingly, IL-8 was detectable in five out of seven controls, with a median of 68.7 pg/ml (0-145.2 pg/ml).
Conclusions We failed to identify detectable circulating levels of GM-CSF in healthy individuals or in severe sepsis. The circulating levels of IL-8 were lower in septic patients than in healthy controls. As previously suggested, locally produced GM-CSF might have a role in this condition. Initial low levels of IL-8 found might be a component of the immunodepression seen in severe sepsis. Also, it is possible that the IL-8 levels correlate with the primary site of infection and the severity of the immunodepression related to that. Background IL-17 is a proinflammatory cytokine predominantly produced by T cells, which is involved in the innate immune responses to various physiologic and pathophysiologic processes including bacterial host defense. The neutralisation experiments showed that the lack of IL-17 leads to decreased neutrophil emigration and systemic granulopoietic responses to pulmonary bacterial pathogens and allergens. The aim of our study was to determine the kinetics of IL-17 in plasma and lungs of animals intratracheally infected with Klebsiella pneumoniae. Materials and methods In our experiments we used 8-12-weekold BALB/c male mice. Mice were intratracheally inoculated with 150 CFU K. pneumoniae strain Caroli. At different time points, mice were sacrificed and the lungs and blood were aseptically removed and prepared for the cytokine determination. Cytokine determination was performed by commercial ELISA kit (Bender-Med Systems, Vienna, Austria). Results The IL-17 concentration in lung homogenates slightly increased in the first 2 hours of infection. Then it slightly decreased and again started to increase 24 hours after the infection. The concentration in the lungs reached the maximal value 48 hours post infection. These results are consistent with data previously published by others. On the other hand, we also found increased plasma values of IL-17. Its concentration in plasma started to increase 12 hours after the infection and reached the peak value 24 hours post infection. These results are in contrast with the results of others who reported no changes in systemic IL-17 production after the intratracheal K. pneumoniae challenge. Conclusions IL-17 in local host defenses against the Gramnegative pathogens is undoubtedly important for the clearance of microorganisms, but its importance in the systemic host response is still not resolved. Its maximal concentration in plasma correlates with the appearance of the bacteria in the blood after 24 hours, so we speculate that its role is also to stimulate systemic proinflammatory cytokines to combat release of bacteria and/or their toxic products into the blood system. Background A rapid microbiological diagnosis permits one to undertake an appropriate therapy against bacterial sepsis and contributes both to improve the healing possibilities and to lower the high costs of hospitalisation. Materials and methods Blood samples from hospitalised patients were inoculated into Bactec Plus Aerobic/F and Anaerobic bottles (standard inoculum; Becton Dickinson) and into an EDTA tube (3 ml). The Septifast Lys kit, Septifast Prep kit and LightCycler Septifast kit (Roche) were used in disrupting bacterial walls, and in extracting and amplifying bacterial DNA, respectively. Blood agar plates were from Becton Dickinson. Strains were identified by Vitek2 (Biomerieux) panels. Real-time PCR results were analysed by Septifast identification software. Blood cultures were incubated into a Bactec 9240 instrument and, if positive, aliquots were taken, plated onto blood agar plates and identified according to routine procedures. EDTA-blood samples were processed according to the LightCycler Septifast test. Results Among 147 samples from 51 patients with sepsis clinical signs, 30 were positive by Bactec and 27 by Septifast (which did not contain probes for three species isolated by cultural method). Concordance between the two methods was 76% for species identification. Different results between the two methods included 10 Bactec-positive and Septifast-negative samples (Acinetobacter baumannii was not amplified seven times, Gram-positive cocci were not determined by Septifast identification software three times) and 12 Bactec-negative and Septifast-positive samples (five Gram-positive cocci, three Pseudomonas aeruginosa, three Escherichia coli, one Klebsiella spp.), while four samples were positive by both methods but gave different species. The time response was evaluated: Septifast was faster than the traditional method in 97% cases.

Conclusions
Although the impossibility to determine antibiotic susceptibility represents a great limit of the Septifast technique, the short response time may contribute to undertake a rapidly tuned therapeutic regimen. Septifast Test appears to be a valid help in an early assessment of sepsis, along with cultural method.

P42
Community-acquired bacteraemia with sepsis in adults admitted to the intensive care unit: a prospective multicentre study

Background
The main objectives of this study were to describe the epidemiology and microbiology of community-acquired bacteraemia (CAB) in patients with sepsis and to determine the associated crude mortality. Materials and methods A prospective, multicentre, cohort, study, was performed on community-acquired bloodstream infections with sepsis admitted to Portuguese intensive care units (ICUs), from 1 December 2004 to 30 November 2005, with a follow-up until discharge. CAB was defined as an infection that was present on admission or within the first 48 hours, with positive blood cultures drawn in the same period. Bacteraemia was considered to be healthcare related (HCRB) if: (a) patients had wound dressing or intravenous treatments in the previous 30 days, (b) patients were observed at a hospital or haemodialysis centre or had chemotherapy in the previous 30 days, (c) patients were admitted to an acute care hospital for 2 days or more in the previous 90 days, (d) patient lives at a nursing home or institution. Results Seventeen units entered the study from north to south Portugal, corresponding to 41% of all mixed national ICU beds. Over this period 4,142 patients were admitted to the study -897 (22%) had community-acquired sepsis, and of these 804 (90%) had blood cultures done in the first 48 hours. Significant isolates were grown from samples of 160 patients (20%). The following focus of infection was identified: respiratory (37%), intra-abdominal (19%), urinary (16%), endovascular (14%), skin and soft tissue (7%), central nervous system (5%) and other (2%). Of these 31% were health-care related (HCRB). Comparing patients without HCRB with those who had HCRB, an inversion in the microbiological profile was found with Gram-positive dominating the first (55% versus 29%) and Gram-negative dominating in the latter (29% versus 54%). When compared with patients without bacteraemia, the 28-day crude mortality was higher for patients with positive blood cultures (39% versus 29%, P = 0.014); however, when adjusted to other variables, it was not found to be an independent determinant of 28-day mortality. Conclusion CAB with severe sepsis accounts for 4% of all ICU admissions with a crude 28-day mortality rate of 39%, although not representing an independent risk factor for 28-day mortality. The microbiological profile is quite different in healthcare-associated bacteraemia, making their identification extremely important to prompt an initial adequate antibiotic approach. Acknowledgement Presented on behalf of the SACiUCI study group.

Materials and methods
The Alteco LPS Adsorber contains polyethylene discs with a specific polypeptide that binds LPS. The priming volume is 80 ml and the recommended blood flow is 100-150 ml/min. Heparin is used for anticoagulation. Fifteen patients scheduled for elective surgery for coronary artery disease and/or valvular surgery were included in the study. Nine patients had the LPS adsorber included in the ECC circuit whereas six patients served as controls with no adsorber. Blood flow through the adsorber was initiated at the time of aortic clamping, adjusted to 150 ml/min, and terminated when weaning from ECC started. Samples for LPS, TNFα, IL-6, and IL-1β were taken before anaesthesia, 10 minutes after aortic clamping, at skin suture and 6 hours after skin suture. Results There were no differences between the groups in age, perfusion time and time of aortic clamping ( Table 1). The time of adsorber treatment is also shown. The use of the adsorber was uneventful and flow was easily maintained at 100-150 ml/min. LPS was found in two patients, one in each group at skin suture. Both patients had a long clamping time, 113 versus 123 min. Cytokines and complement results are presented in Tables 2 and  3. There were no significant changes in TNFα or IL-1β while IL-6 increased in both groups. Complement factors C3 and C4 decreased in the control group while C1q increased in the adsorber group.
In vitro studies have shown that the Alteco LPS adsorber can reduce LPS in whole blood. LPS was found in only two patients in our study, probably due to the short duration of perfusion and clamping. Thus the possibility to evaluate the efficacy of the adsorber in this setting was limited. Only IL-6 increased during the study but no difference between the groups was seen. The decrease in C3 and C4 seen in the control group may indicate activation of inflammatory system, more pronounced in the control group. Conclusion The adsorber can be used safely in the extracorporeal circulation. Further studies in patients with more pronounced endotoxic response are needed to evaluate the effects of the adsorber on clinical outcome.    S21 system for an exaggerated response to subsequent inflammatory stimuli such as lipopolysaccharide (LPS): the so-called 'two-hit hypothesis'. The mechanisms of oxidant-induced cell priming, however, remain poorly elucidated. Our group has previously reported a role in this priming process for translocation of the LPS receptor TLR4 to the plasma membrane [1]. We have also previously shown that oxidant priming reprograms LPS signaling in macrophages to a Src-dependent pathway leading to PI3k activation [2]. Taken together, we hypothesized that Src activation may play a role in oxidant-induced TLR4 translocation.

Materials and methods
Wild-type (WT) and triple Src (hck/fgr/lyn -/-) knockout (KO) mice were subjected to hemorrhagic shock and resuscitation to generate oxidative stress in vivo. Alveolar macrophages (AMs) were then retrieved by bronchoalveolar lavage and analyzed for TLR4 translocation by immunofluorescence staining and flow cytometry. In a separate in vitro experiment, AMs from WT and KO mice were exposed to 200 μM hydrogen peroxide for 60 minutes and similarly analyzed by flow cytometry for surface expression of TLR4. Results In vivo, S/R induced visible translocation of TLR4 to the plasma membrane of AMs from WT mice as seen by immunofluorescence. This effect was completely inhibited in the Src KO animals, and the observation was corroborated quantitatively by flow cytometry (see Figure 1). In order to confirm the role of oxidative stress in S/R-induced TLR4 translocation, AMs from WT and KO mice were exposed to hydrogen peroxide in vitro.
Hydrogen peroxide caused an increase in surface expression of TLR4 in WT AMs by a ratio of 1.23 compared with WT control. This effect was once again inhibited in the Src KO AMs, where the ratio was 0.95 compared with WT control. Conclusion Oxidative stress induces TLR4 translocation to the cell surface of macrophages in a Src-dependent manner. Since cellular responsiveness to LPS is known to correlate with surface levels of TLR4, this novel finding may direct future therapies in modulating oxidant-induced cellular priming and subsequent organ failure.
Background Sepsis is a systemic inflammatory response resulting from the inability of the host to restrict local infection. During severe sepsis occurs a marked failure of neutrophil migration into the infectious focus, which is associated with dissemination of infection resulting in high mortality. Recently, we showed that heme oxygenase (HO) products, carbon monoxide and biliverdin, downregulate neutrophil recruitment by reducing the neutrophil/ Available online http://ccforum.com/supplements/11/S4

Figure 1 (abstract P44)
Shock/resuscitation (S/R)-induced TLR4 translocation is inhibited in Src knockout (KO) mice. WT, wild type. endothelium rolling and adhesion in a noninfectious inflammatory model. This study aimed to investigate a possible role of the HO-1 pathway on the failure of neutrophil recruitment in mice subjected to severe (S-CLP) polymicrobial sepsis induced by cecal ligation and puncture (CLP).

Materials and methods
To evaluate the role of HO-1 in S-CLP, the Balb/c mice were pretreated with a specific HO-1 inhibitor, zinc protoporphyirin IX (ZnPPIX 30 mg/kg, subcutaneously) 30 minutes before severe sepsis induction. Mice were killed 6 hours after CLP and HO-1 expression in mesenteric tissue and in blood neutrophils were determined. In another set of experiments, mice were killed 6 hours and 12 hours after CLP, and neutrophil migration to the peritoneal cavity, bacteremia, lung neutrophil sequestration assessed by myeloperoxidase activity and lung histology, cytokines, liver, kidney and cardiac function and mean arterial pressure were determined. The survival rate of animals was assessed every 12 hours up to 120 hours after surgery.
Results A significant increase in HO-1 expression was observed in mesenteric tissue and in blood neutrophils after S-CLP. The ZnPPIX pretreatment prevents the failure of neutrophil endothelium rolling, adhesion and migration observed in vehicle-pretreated mice subjected to S-CLP. As consequence, these mice presented reduced bacteremia, low levels of seric TNFα and lung neutrophil sequestration, reduced liver, kidney and cardiac injury and improved mean arterial pressure, resulting in an increase of the survival rate.
Conclusions These data suggest that during an infectious process HO-1 displays a crucial role on the failure of neutrophil migration to infectious focus and consequently on the susceptibility in severe sepsis. Acknowledgement Supported by FAPESP/CAPES/FAEPA. Background There is a marked defect in neutrophil migration into the infectious focus during severe sepsis, which is associated with the severity of disease. We recently demonstrated that this phenomenon is a consequence of the downregulation of chemokine receptor CXCR2 on the surface of circulating neutrophils. Toll-like receptors (TLRs) are pattern-recognition receptors that are important in innate immune responses to bacterial infection. TLR activation in phagocytes produces proinflammatory cytokines and chemokines that contribute directly to elimination of infectious agents. However, a sustained inflammatory response can result in tissue damage and sepsis. Here, we address the role of TLR2 in the downregulation of CXCR2 and establishment of neutrophil migration impairment in severe sepsis. Materials and methods TLR2-deficient (TLR2 -/-) and C57BL/6 (WT) mice were subjected to severe polymicrobial sepsis by the cecal ligation and puncture (CLP) model, and neutrophil migration, bacteremia, CXCR2 expression and cytokines levels were evaluated.
Results It was observed that TLR2 is critical for downregulation of CXCR2 expression on circulating neutrophils during severe sepsis, since this event was prevented in TLR2 -/mice. In accordance, TLR2 -/mice did not present failure of neutrophil migration into the infectious focus and, consequently, they presented lower bacteremia and did not display systemic inflammation determined by reduced levels of circulating cytokines, showing an improved survival rate. Furthermore, in vitro, TLR2 agonist (lipoteichoic acid) was able to downregulate CXCR2 expression and to markedly inhibit neutrophil chemotaxis induced by CXCR2-ligand. The downregulation of CXCR2 was associated with enhanced expression of G-protein-coupled receptor kinase-2 (GRK-2), which is known to play an important role in the desensitization and internalization of this chemokine receptor. Finally, we showed that in vitro lipoteichoic-acid-stimulated neutrophils adoptively transferred into naïve WT mice display a significantly reduced competence to migrate into the perinoneal cavity in response to thioglycolate. Conclusion Altogether, these findings suggest that TLR2 through GRK-2 signaling downregulates CXCR2 expression on the surface of circulating neutrophils, which is a critical determinant of impairment of neutrophil migration into the infection focus during severe sepsis. Acknowledgements Supported by FAPESP and FAEPA.

P48
The activity of the endothelial Tie-2 receptor modulates ventilation-induced lung injury in septic mice Background Patients requiring mechanical ventilation, particularly in the presence of sepsis, are at risk for ventilation-induced lung injury (VILI), which is characterized by increased pulmonary vascular permeability and inflammation. Inflammation in VILI is initiated through interactions between endothelial and inflammatory cells. The vasoprotective protein angiopoietin-1, a ligand for the endothelial-selective tyrosine kinase receptor Tie-2, has recently been recognized to play a critical role in regulating endothelial inflammation and vascular permeability. In contrast, angiopoietin-2, which binds with equal affinity, has been described as a Tie-2 antagonist, inhibiting receptor activation in response to angiopoietin-1. We therefore examined the role of the angiopoietins/Tie-2 pathway in mechanically ventilated, septic mice. Materials and methods Wild-type (WT) or Tie-2 +/mice were anaesthetized and sepsis was induced by cecum ligation and puncture (CLP). After 24 hours, animals were ventilated for 6 hours to induce VILI. Animals were ventilated in a pressure-controlled mode, at a fractional inspired oxygen concentration of 0.5, an inspiration to expiration ratio of 1:2, a peak inspiratory pressure of 14 cmH 2 O and a positive end-expiratory pressure of 2 cmH 2 O. Animals were sacrificed and bronchoalveolar lavage (BAL) was performed for total cell count, differential, and permeability parameters. Inflammatory cytokine IL-6 levels were measured in BAL and plasma. Results There was no difference between sham-operated or CLPoperated WT mice in total cell numbers in BAL. Sham operated Tie-2 +/mice had similar cell numbers, while ventilation in CLPoperated Tie-2 +/mice resulted in higher cell numbers (P = 0.06 versus WT), with an increase mainly in macrophages. IL-6 plasma levels were increased in septic Tie-2 +/mice subjected to mechanical ventilation compared with the WT mice, whereas IL-6 levels in BAL were not significantly different. Despite the higher total cell counts, permeability parameters measured in Tie-2 +/-mice were lower than in WT mice (albumin, P = 0.08; total proteins, not significant; IgM, P = 0.05).
Conclusions Pulmonary vascular permeability in mice with knockdown of endothelial Tie-2 receptor is preserved in ventilated septic mice. However, there was an increase in the number of inflammatory cells and systemic inflammation (IL-6) in these animals. These results further emphasize the importance of Tie-2 in regulating vascular permeability in the lungs. Background Our previous studies indicated that nicotine attenuated systemic inflammation, and improved survival in both lethal endotoxemia and cecal ligation and puncture [1].

Materials and methods
Here we analyzed the pathophysiological and pharmacological implications of the spleen in experimental sepsis. Acetylcholine and nicotine inhibit the production of TNF and HMGB1 from lipopolysaccharide-treated RAW264.7 macrophage-like cells in a concentration-dependent manner. Nicotine inhibits the production of these cytokine macrophage-like cells through a mechanism dependent on the α 7 -nicotine acetylcholine receptor (nAChR). From a physiological perspective, our studies indicate that the spleen has a significant expression of the α 7 nAChR. To study the implications of the splenic α 7 nAChR in the therapeutic potential of nicotine in sepsis, we analyzed the effects of treatment with nicotine in splenectomized mice. Adult male 8-12-week-old BALB/c mice (25-30 g) were splenectomized and 7 days later subjected to cecal ligation and puncture to induce polymicrobial peritonitis. Control sham animals underwent laparotomy without splenectomy. Administration of nicotine (400 mg/kg, i.p.) or vehicle injection (1 × PBS, i.p.) began 24 hours after surgery, and thereafter twice a day for 3 days. Animals were euthanized after 44 hours for serum cytokine analysis.
Results Treatment with nicotine improved survival in sham mice but it worsened the survival of splenectomized mice. Nicotine inhibited the production of TNF and HMGB1 from lipopolysaccharide-treated macrophages differentiated from peripheral blood mononuclear cells. However, in vivo, nicotine attenuated serum TNF levels in sham mice but not in splenectomized mice with polymicrobial peritonitis. There were no differences in serum TNF, IL-12, IFNγ or IL-10 levels between vehicle-treated or nicotine-treated splenectomized animals. Unlike systemic TNF, circulating HMGB1 levels correlated with the increased mortality induced by nicotine in splenectomized mice. Treatment with nicotine increased circulating serum HMGB1 levels in splenectomized mice with experimental sepsis.
Conclusion These results suggest that the spleen has major pharmacological implications for the treatment of sepsis. Materials and methods Sheep were surgically instrumented ahead of the study. After obtaining baseline data, they were anesthetized and insufflated with 48 breaths of cotton smoke and 1,011 colony forming units of Pseudomonas aeruginosa instilled into their airways via a tracheostomy. The sheep were then placed on a ventilator and awakened for the study. They were resuscitated with Ringer's solution to maintain left atrial and central venous pressures and hematocrit at baseline levels. If, despite fluid management, the mean arterial pressure (MAP) fell by 10 mmHg from baseline, a continuous intravenous infusion of AVP or FE 202158 was initiated and titrated to keep MAP within this limit, except in a septic control group and in a nonseptic sham group.

Results
In the septic control group, MAP fell by ~30 mmHg and the systemic vascular resistance index (SVRI) by ~50% over the 24-hour study. In the AVP group, although the MAP and SVRI were maintained above septic control group levels, they could not be brought back to sham group levels. In contrast, in the FE 202158 group, the MAP and SVRI were maintained at sham group levels.
The septic control group accumulated ~7 l fluid over the 24-hour study (~20% body weight), while this was reduced by ~50% in the AVP group, and was totally prevented in the FE 202158 group. The fluid accumulation was associated with a constant hematocrit, suggesting that it was extravascular. It was also correlated with a fall in plasma total protein and oncotic pressure, suggesting protein leakage from the vasculature and probably reflecting sepsis-induced vascular leak syndrome. S25 patients with at least one ICU-acquired infection was 47.6% for trauma males and 43.1% for trauma females (P = 0.13); and was 44.2% for nontrauma male patients and 45.8% for nontrauma female patients (P = 0.63). The inhospital mortality of patients with at least one ICU-acquired infection was 12.5% for trauma males and 7.3% for trauma females (P = 0.06); and was 30.8% for nontrauma male patients and 28.0% for nontrauma female patients (P = 0.55). Logistic regression analysis did not show gender to be a significant variable in the acquisition of an ICU infection. The results of a logistic regression analysis of predictors of inhospital mortality in patients with ICU-acquired infections are presented in Table 1. Only variables significant by univariate analysis were entered into logistic models. Empty cells represent variables that were significant for only one group (for example -only trauma, but not nontrauma).
Conclusion Gender does not appear to play a role in acquisition of an ICU infection in either trauma or nontrauma patients. In contrast, female gender appears to provide a strong survival advantage in trauma patients with ICU-acquired infections. This phenomenon was not observed in the nontrauma group. Further investigation into the hormonal and cytokine differences between the genders in critically ill trauma patients and their response to infection may prove beneficial.

Background
We evaluated the immunomodulatory effect of three antifungal agents, amphotericin B (AmB), micafungin (MF) and itraconazole (ITZ), in the aspect of cytokine production and expression of NF-κB in Aspergillus fumigatus conidia-treated RAW264.7 cells, a murine alveolar macrophage cell line.

Materials and methods
We evaluated the cytotoxic effect of antifungal agents using a commercial cell proliferation assay. TNFα and IL-10 production according to stimulation (control, A. fumigatus conidia only, conidia + antifungal drug, conidia + antifungals + GM-CSF) was evaluated and each compared using a commercial ELISA method. NF-κB activation was evaluated by western blot analysis. Results AmB, MF and ITZ showed a dose-dependent cytotoxic effect on the tested cells. Stimulation of cells by A. fumigatus conidia induced TNFα production. Pretreatment of AmB at concentrations not affecting cellular survival did not change the production of TNFα compared with conidia-treated cells, but pretreatment of MF or ITZ showed a reduced amount of TNFα production compared with conidia-treated cells. AmB also showed a synergistic effect on TNFα production when simultaneously treated with GM-CSF. IL-10 production was markedly increased when the cells were treated with AmB with conidia. MF and ITZ induced a smaller increase of IL-10 production. AmB also showed a synergistic effect on the production of IL-10 when treated with GM-CSF simultaneously. A. fumigatus conidia enhanced expression of NF-κB. The degree of NF-κB expression was associated with the amount of TNFα and IL-10 produced.
Conclusion The antifungal agents we used in this experiment showed decreased TNFα production and increased IL-10 production from the RAW264.7 cells stimulated by A. fumigatus conidia after pretreatment of antifungal agents. But more studies such as the association between the immunomodulatory effect and antifungal activity and the difference of the signal pathway of cellular activation according to drugs should be performed. Background Inadequate antibiotic therapy predicts poor outcome from sepsis, but there is no simple test of adequacy. We suggested that a time-to-positivity assay (T pos ) might act as a surrogate for antimicrobial activity and predict outcome from sepsis in the intensive care unit (ICU) [1]. We are conducting a prospective clinical trial to test this hypothesis. Methods We studied 35 sequential ICU patients with onset of sepsis who had not had antibiotics for at least 3 days. Sepsis was defined as clinical evidence of infection plus at least three criteria for systemic inflammatory response syndrome. All patients received standard empiric therapy. Sera taken at 24 hours post antibiotics were inoculated into blood culture bottles containing standardised bacteria, incubated in an automated microbial detection system and the time to positivity noted. The primary clinical endpoint was days in the ICU.

Results
Cultures that are negative after 5 days of incubation indicate adequate antimicrobial therapy; cultures that become positive in <5 days indicate inadequate antimicrobial therapy. Patients with negative T pos (n = 11) were associated with a stay in the ICU of less than 6.5 days (P = 0.052). See Figure 1.
Conclusion These data show that some septic patients on standard antimicrobial regimens are receiving inadequate therapy. The T pos is a new, simple assay that might be used effectively to monitor antibiotic use. Average length of stay for patients with negative and positive time-topositivity 1 (taken at 24 hours after first dose of antibiotics) with 95% confidence intervals.