Critical care

Published by BioMed Central
Print ISSN: 1364-8535
Publications
Red blood cell storage lesion
During the past 20 years, the perceived value of blood transfusions has changed as it has become appreciated that transfusions are not without risk. Red blood cell transfusion has been associated with disease transmission and immunosuppression for some time. More recently, proinflammatory consequences of red blood cell transfusion have also been documented. Moreover, it has become increasingly evident that stored red blood cells undergo time-dependent metabolic, biochemical, and molecular changes. This 'storage lesion' may be responsible for many of the adverse effects of red blood cell transfusion. Clinically, the age of blood has been associated with multiple organ failure, postoperative pneumonia, and wound infection. The relationship between age of blood and clinical adverse effects needs further study.
 
Baseline characteristics of the patients
Primary and secondary outcomes
Post hoc analysis 90-day mortality and need for renal replacement therapy
Hct, hematocrit; MAP, mean arterial pressure; ScvO
2
, central venous oxygen saturation; UO, urinary output.
Study flow diagram.
There are limited data on the efficacy of early fluid resuscitation with third generation hydroxyethyl starch (HES 130) in burn injury. Adverse effects of HES on survival and organ function have been reported. In this randomized, controlled, double-blind trial 48 patients with severe burn injury were assigned to receive either Lactated Ringer's solution plus 6% HES 130/0.4 in a ratio of 2:1 or Lactated Ringer's solution with no colloid supplement for the first 72 hours. Primary outcome parameter was the group difference of administered total fluid from intensive care unit (ICU) admission up to day 3. Secondary outcomes included kidney and lung injury and failure, length of stay, and mortality. 3 days total of administered resuscitation fluid (medians) was 21,190 ml in the Lactated Ringer's group and 19,535 ml in the HES group (HES: -1,213 ml; P = 0.39). Creatinine levels day 1 to 3 (HES: +0.4 mumol/l; 95% CI -18.7 to 19.5; P = 0.97) and urinary output day 1 to 3 (HES: -58 ml; 95% CI -400 to 284; P = 0.90) were not different. 6 patients in each group developed acute respiratory distress syndrome (ARDS) (risk ratio 0.96; 95% CI 0.35 to 2.64; P = 0.95). Length of ICU stay (HES vs. Lactated Ringer's: 28 vs. 24 days; P = 0.80) and length of hospital stay (31 vs. 29 days; P = 0.57) were similar. 28-day mortality was 4 patients in each group (risk ratio 0.96; 95% CI 0.27 to 4.45; P = 0.95), in-hospital mortality was 8 in the HES group vs. 5 patients in the Lactated Ringer's group (hazard ratio 1.86; 95% CI 0.56 to 6.19; P = 0.31). There was no evidence that early fluid resuscitation with balanced HES 130/0.4 (6%) in addition to Lactated Ringer's solution would lead to a volume sparing effect in severe burn injury. Together with the findings that early renal function, incidence of ARDS, length of stay, and mortality were not negatively influenced by HES in this setting, balanced HES 130/0.4 (6%) plus Lactated Ringer's solution could not be considered superior to Lactated Ringer's solution alone.Trial registration: ClinicalTrials.gov NCT01012648.
 
Number of patients by the AKIN and RIFLE classifications
Study design.
Evolution of mean serum creatinine (SCr) levels over time.
Inadequate initial treatment and delayed hemodynamic stabilization (HDS) may be associated with increased risk of death in severe sepsis patients. In order to compare the hemodynamic efficacy and safety of 6% HES 130/0.4 and NaCl 0.9% for HDS in patients with severe sepsis, we designed a prospective, multicenter, active-controlled, double-blind, randomized study in intensive care units. 174 out of 196 patients reached HDS (88 and 86 patients for HES and NaCl, respectively). Significantly less HES was used to reach HDS vs. NaCl (1,379 ±886 ml in the HES group and 1,709 ±1,164 ml in the NaCl group (mean difference = -331± 1,033, 95% CI -640 to -21, P = 0.0185). Time to reach HDS was 11.8 10.1 hours vs. 14.3 ±11.1 hours for HES and NaCl, respectively. Total quantity of study drug infused over four consecutive days, ICU and hospital LOS, and area under the curve of SOFA score were comparable. Acute renal failure occurred in 24 (24.5%) and 19 (20%) patients for HES and NaCl, respectively (P = 0.454). There was no difference between AKIN and RIFLE criteria among groups and no difference in mortality, coagulation, or pruritus up to 90 days after treatment initiation. Significantly less volume was required to achieve HDS for HES vs. NaCl in the initial phase of fluid resuscitation in severe sepsis patients without any difference for adverse events in both groups.
 
Preoperative hemodilution is an established practice that is applied to reduce surgical blood loss. It has been proposed that polyethylene glycol (PEG) surface decorated proteins such as PEG-conjugated human serum albumin may be used as non-oxygen-carrying plasma expanders. The purpose of this study was to determine whether there is any difference in survival time after severe hemorrhagic shock following extreme hemodilution using a conventional hydroxyethyl starch (HES)-based plasma expander or PEG-albumin. Experiments were performed using the hamster skinfold window preparation. Human serum albumin that was surface decorated with PEG was compared with Voluven 6% (Fresenius Kabi, Austria; a starch solution that is of low molecular weight and has a low degree of substitution; HES). These plasma expanders were used for a 50% (blood volume) exchange transfusion to simulate preoperative hemodilution. Exchange transfusion was followed by a 60% (blood volume) hemorrhage to reproduce a severe surgical bleed over a 1 hour period. Observation of the animal was continued for another hour during the shock phase. The PEG-albumin group exhibited significantly greater survival rate than did the HES group, in which none of the animals survived the hemorrhage phase of the experiment. Among the treatment groups there were no changes in mean arterial pressure and heart rate from baseline after hemodilution. Both groups experienced gradual increases in arterial oxygen tension and disturbance in acid-base balance, but this response was more pronounced in the HES group during the shock period. Mean arterial pressure remained elevated after the initial hemorrhage period in the PEG-albumin group but not in the HES group. Maintenance of a greater mean arterial pressure during the initial stages of hemorrhage is proposed to be in part due to the improved volume expansion with PEG-albumin, as indicated by the significant decrease in systemic hematocrit compared with the HES group. PEG-albumin treatment yielded higher functional capillary density during the initial stages of hemorrhage as compared with HES treatment. The ability of PEG-albumin to prolong maintenance of microvascular function better than HES is a finding that would be significant in a clinical setting involving preoperative blood management and extreme blood loss.
 
We argue that assessing the hemodynamic efficacy of hydroxyethyl starch (HES) versus NaCl in patients with severe sepsis requires an algorithm to direct the timing and amount of fluid resuscitation. Such an algorithm may include hemodynamic flow parameters. In a recent issue of Critical Care, Guidet and colleagues [1] reported that a smaller amount of 6% HES 130/0.4 versus 0.9% NaCl was required to achieve hemodynamic stability (HDS) during the initial phase of fluid resuscitation in patients with severe sepsis. The target parameters indicating HDS included central venous pressure (CVP) (8 to 12 mm Hg), a poor indicator of fluid responsiveness [2], and a large urine output (>2 mL/kg per hour), and therefore pose a risk of over-infusion. Other authors have reported that over-infusion, elevated CVP, and excessive fluid resuscitation with HES are associated with increased mortality in patients with sepsis [3,4]. In contrast, after initial HDS was achieved, no such target parameters were defined, and so the cumulative volumes of study drug infused over the course of four consecutive days in the intensive care unit (ICU) were similar for the HES (2,615 mL) and NaCl (2,788 mL) groups. No differences in mortality, hospital length of stay, or kidney function were found. This study may be showing only that, in the absence of an algorithm to guide fluid resuscitation, intensivists use an unvarying amount of fluids, but it is impossible to know whether these fluids were, in fact, indicated. In patients undergoing major abdominal surgery, hemodynamic algorithms that guide the timing and amount of fluid administration have helped determine the clinical efficacy of fluid therapy [5]. The negative results reported by Guidet and colleagues suggest that hemodynamic algorithms for patients with sepsis are urgently required to accurately compare the hemodynamic efficacy, safety, and outcome of HES versus NaCl fluid replacement.
 
We read with great interest the study by Chen and colleagues highlighting interest in hydroxyethyl starch (HES) 130/0.4 in treatment after hemorrhagic shock to ameliorate oxidative stress and the inflammatory response in a rat model. Compared with HES 200/0.5 and succinylated gelatin, the authors showed that infusions of HES 130/0.4 significantly reduced malondialdehyde levels and myeloperoxidase activity and also inhibited about 50% of TNF-α production in the intestine [1]. However, we regret the lack of assessment of another resuscitative fluid: the hypertonic saline solution (HTS). In our level 1 trauma center, we chose to use HTS because we have some concerns about HES safety. Indeed, HES may induce coagulopathy and increase risk of renal-replacement therapy [2]. HTS has several advantages due to its osmotic effects. Firstly, it leads to restoration of circulating volume with a smaller volume of fluid. Secondly, it reduces intracranial pressure in case of associated traumatic brain injury [3]. In addition, HTS attenuates the increase in plasma concentration of IL-1β, IL-6, IFN-γ and TNF-α, suggesting that HTS may also limit the inflammatory response to hemorrhage and reperfusion [4]. One of its inconveniences may be the increased risk of acute kidney injury due to hyperchloremic metabolic acidosis decreasing renal blood flow; however, this effect was especially demonstrated when using large amounts of 0.9% saline solution [5]. We suggest that, in 2013, studies on fluid resuscitation should compare all the available resuscitative fluids, and not just HES, currently under concern for safety reasons. Abbreviations HES: hydroxyethyl starch; HTS: hypertonic saline solution; IFN: interferon; IL: interleukin; TNF: tumor necrosis factor. Competing interests The authors declare that they have no competing interests.
 
Acute kidney injury (AKI) in the ICU is associated with poorer prognosis. Hydroxyethylstarch (HES) solutions are fluid resuscitation colloids frequently used in the ICU with controversial nephrotoxic adverse effects. Our study objective was to evaluate HES impact on renal function and organ failures. This observational retrospective study included 363 patients hospitalized for more than 72 hours in our ICU. A hundred and sixty eight patients received HES during their stay and 195 did not. We recorded patients' baseline characteristics on admission and type and volume of fluid resuscitation during the first 3 weeks of ICU stay. We also noted the evolution of urine output, the risk of renal dysfunction, injury to the kidney, failure of kidney function, loss of kidney function and end-stage kidney disease (RIFLE) classification and sepsis related organ failure assessment (SOFA) score over 3 weeks. Patients in the HES group were more severely ill on admission but AKI incidence was similar, as well as ICU mortality. The evolution of urine output (P = 0.74), RIFLE classification (P = 0.44) and SOFA score (P = 0.23) was not different. However, HES volumes administered were low (763+/-593 ml during the first 48 hours). Volume expansion with low volume HES 130 kDa/0.4 was not associated with AKI.
 
ABSTRACT: It is heavily debated whether or not treatment with hydroxyethyl starch 130/0.4 contributes to the development of acute kidney failure in patients with severe sepsis. In the previous issue of Critical Care, Muller and colleagues report no association between initial resuscitation with hydroxyethyl starch 130/0.4 and renal impairment in a cohort of septic patients. Can we then consider hydroxyethyl starch 130/0.4 a safe intervention? The answer is no - observational data should be interpreted with caution and should mainly be used to identify risks, while safety must be assessed in randomised clinical trials. With these factors in mind, Muller's data associate the use of vasopressors with poor outcome, underlining the need for further randomised clinical trials to assess the potential harmful effects of common interventions in the critically ill.
 
Effects of hydroxyethyl starch (HES) 130/0.4 and HES 200/0.5 on in vitro clot formation in comparison with saline after haemodilution rates of 10% (white) or 40% (grey) with either HES solution or saline (Control). Clotting time (a), clot formation time (b) and maximum clot firmness (MCF) in EXTEM (c) as well as MCF in FIBTEM (d) were determined by ROTEM. Data are given as mean ± standard error of the mean obtained from six identical experiments. Significant differences between saline and HES 130/0.4 or HES 200/0.5: *P < 0.05, **P < 0.005, ***P < 0.001.
Effects of hydroxyethyl starch (HES) 130/0.4 and HES 200/0.5 on platelet activation markers and formation of platelet-leukocyte conjugates. After haemodilution of 10% (white) or 40% (grey) with either HES solution or saline (Con), platelets were activated by adenosine diphosphate (ADP) (5 μM) or thrombin receptor-activating peptide (TRAP) (25 μM) for 5 minutes or remained inactivated (Basal). CD62P expression (a) and fibrinogen binding to platelets (b) as well as platelet-neutrophil (P-N) (c) and platelet-monocyte (P-M) (d) conjugates were determined by flow cytometry. The results are expressed as the relative numbers of platelets or leukocytes that were positive for CD62P or fibrinogen or the platelet-specific antigen CD42a. Data are given as mean ± standard error of the mean obtained from eight identical experiments. Significant differences between saline and HES 130/0.4 or HES 200/0.5: *P < 0.05, ** P < 0.005, *** P < 0.001. Significant differences between HES 130/0.4 and HES 200/0.5: §P < 0.05, §§P < 0.01.
Hydroxyethyl starch (HES) solutions are widely used for volume replacement therapy but are also known to compromise coagulation, impair renal function and increase long-term mortality. To test the hypotheses that HES 130/0.4 has fewer adverse effects than HES 200/0.5 and exerts anti-inflammatory properties, we compared the effects of HES 130/0.4, HES 200/0.5 and saline on in vitro haemostasis and pro-inflammatory platelet function. Whole blood samples from healthy volunteers were mixed with 6% HES 130/0.4, 10% HES 200/0.5, or normal saline to achieve a final haemodilution rate of 10% or 40%. Haemostatic capacity was characterised by thromboelastography (ROTEM) and measurement for FXIIIa activity. Platelet activation and pro-inflammatory platelet functions were characterised by flow cytometry measuring the platelet activation marker CD62P and binding of fibrinogen to platelets as well as the formation of heterotypic platelet-leukocyte conjugates. Compared with saline, HES 130/0.4 dose-dependently impaired formation and firmness of the fibrin clot but did not affect the fibrin crosslinking activity of FXIIIa. At 40% but not at 10% haemodilution rate, HES 200/0.5 also increased platelet fibrinogen binding and both HES solutions increased expression of CD62P, the main receptor for platelet-leukocyte adhesion. HES 130/0.4 but not HES 200/0.5 increased formation of platelet-neutrophil conjugates and, to a lesser degree, platelet-monocyte conjugates. Our data demonstrate that HES 130/0.4 has similar adverse effects as HES 200/0.5. In particular, both types of HES impair coagulation capacity and stimulate, rather than attenuate, pro-inflammatory platelet function.
 
Arterial blood gas variables for all groups during the experiment
Design of the hemorrhagic shock protocol.
Tissue MDA concentrations following HS/R. Data are plotted as the means ± SD. *P <0.05 versus the sham group; &P <0.05 versus the HES 200 group; #P <0.01 versus the GEL group. GEL, gelatin; HES, hydroxyethyl starch; HS/R, hemorrhagic shock/fluid resuscitation; MDA, malondialdehyde.
MPO levels in the liver, lungs, intestine and brain following HS/R in four groups. Data are plotted as the means ± SD. *P <0.05 versus the sham group; &P <0.05 versus the HES 200 group; #P <0.01 versus the GEL group. GEL, gelatin; HES, hydroxyethyl starch; HS/R, hemorrhagic shock/fluid resuscitation; MPO, myeloperoxidase.
Effects of different resuscitation fluids on HS/R-induced intestinal levels of TNF-α and IL-6. Data are plotted as the means ± SD. *P <0.05 versus the sham group; &P <0.05 versus the HES 200 group; #P <0.01 versus the GEL group. GEL, gelatin; HES, hydroxyethyl starch; HS/R, hemorrhagic shock/fluid resuscitation.
This study compared the effect of hydroxyethyl starch 130/0.4, hydroxyethyl starch 200/0.5, and succinylated gelatin on oxidative stress and the inflammatory response in a rodent hemorrhagic shock model. Sodium pentobarbital-anesthetized adult male Wistar rats (200 g - 220 g) were subjected to a severe volume-controlled hemorrhage using arterial blood withdrawal (30 mL/kg - 33 mL/kg), and resuscitated with a colloid solution at the same volume as blood withdrawal (hydroxyethyl starch 130/0.4, hydroxyethyl starch 200/0.5, or succinylated gelatin). Arterial blood gas parameters were monitored. Malondialdehyde (MDA) content, myeloperoxidase (MPO) activity in the liver, lungs, intestine, and brain were measured 2 hours after resuscitation. The levels of tumor necrosis factor (TNF)- alpha and interleukin (IL)-6 in the intestine were also measured. Infusions of hydroxyethyl starch 130/0.4, but not hydroxyethyl starch 200/0.5 or succinylated gelatin, significantly reduced MDA levels and MPO activity in the liver, intestine, lungs, and brain, and it also inhibited the production of TNF-alpha in the intestine 2 hours after resuscitation. However, no significant difference between hydroxyethyl starch 200/0.5 and succinylated gelatin was observed. Hydroxyethyl starch 130/0.4, but not hydroxyethyl starch 200/0.5 and succinylated gelatin, treatment after hemorrhagic shock ameliorated oxidative stress and the inflammatory response in this rat model. No significant differences were observed after hydroxyethyl starch 200/0.5 and succinylated gelatin administration at doses of approximately 33 mL/kg.
 
This prospective randomized clinical study investigated the efficacy and safety of 7.2% hypertonic saline hydroxyethyl starch 200/0.5 (7.2% NaCl/HES 200/0.5) in comparison with 15% mannitol in the treatment of increased intracranial pressure (ICP). Forty neurosurgical patients at risk of increased ICP were randomized to receive either 7.2% NaCl/HES 200/0.5 or 15% mannitol at a defined infusion rate, which was stopped when ICP was < 15 mmHg. Of the 40 patients, 17 patients received 7.2% NaCl/HES 200/0.5 and 15 received mannitol 15%. In eight patients, ICP did not exceed 20 mmHg so treatment was not necessary. Both drugs decreased ICP below 15 mmHg (p < 0.0001); 7.2% NaCl/HES 200/0.5 within 6.0 (1.2-15.0) min (all results are presented as median (minimum-maximum range)) and mannitol within 8.7 (4.2-19.9) min (p < 0.0002). 7.2% NaCl/HES 200/0.5 caused a greater decrease in ICP than mannitol (57% vs 48%; p < 0.01). The cerebral perfusion pressure was increased from 60 (39-78) mmHg to 72 (54-85) mmHg by infusion with 7.2% NaCl/HES 200/0.5 (p < 0.0001) and from 61 (47-71) mmHg to 70 (50-79) mmHg with mannitol (p < 0.0001). The mean arterial pressure was increased by 3.7% during the infusion of 7.2% NaCl/HES 200/0.5 but was not altered by mannitol. There were no clinically relevant effects on electrolyte concentrations and osmolarity in the blood. The mean effective dose to achieve an ICP below 15 mmHg was 1.4 (0.3-3.1) ml/kg for 7.2% NaCl/HES 200/0.5 and 1.8 (0.45-6.5) ml/kg for mannitol (p < 0.05). 7.2% NaCl/HES 200/0.5 is more effective than mannitol 15% in the treatment of increased ICP. A dose of 1.4 ml/kg of 7.2% NaCl/HES 200/0.5 can be recommended as effective and safe. The advantage of 7.2% NaCl/HES 200/0.5 might be explained by local osmotic effects, because there were no clinically relevant differences in hemodynamic clinical chemistry parameters.
 
Kaplan-Meier survival curve and causes of death. Large effect towards a seven-fold higher mortality in the HES group (adjusted hazard ratio 7.12; P = 0.16). ARDS, acute respiratory distress syndrome; HES, hydroxyethyl starch.
Bloodmarkers at days one, two, three and seven. (a) Hematocrit, (b) white blood count, (c) C-reactive protein, (d) glutamat-pyruvat-transaminase (GPT) and (e) glutamat-oxalacetat-transaminase (GOT). Baxter group black bars, hydroxyethyl starch (HES) group grey bars. Plots show mean ± standard deviation, analysis of variance for repeated measurements. There were no differences between the groups.
Despite large experience in the management of severe burn injury, there are still controversies regarding the best type of fluid resuscitation, especially during the first 24 hours after the trauma. Therefore, our study addressed the question whether hyperoncotic hydroxyethyl starch (HES) 200/0.5 (10%) administered in combination with crystalloids within the first 24 hours after injury is as effective as 'crystalloids only' in severe burn injury patients. 30 consecutive patients were enrolled to this prospective interventional open label study and assigned either to a traditional 'crystalloids only' or to a 'HES 200/0.5 (10%)' volume resuscitation protocol. Total amount of fluid administration, complications such as pulmonary failure, abdominal compartment syndrome, sepsis, renal failure and overall mortality were assessed. Cox proportional hazard regression analysis was performed for binary outcomes and adjustment for potential confounders was done in the multivariate regression models. For continuous outcome parameters multiple linear regression analysis was used. Group differences between patients receiving crystalloids only or HES 200/0.5 (10%) were not statistically significant. However, a large effect towards increased overall mortality (adjusted hazard ratio 7.12; P = 0.16) in the HES 200/0.5 (10%) group as compared to the crystalloids only group (43.8% versus 14.3%) was present. Similarly, the incidence of renal failure was 25.0% in the HES 200/0.5 (10%) group versus 7.1% in the crystalloid only group (adjusted hazard ratio 6.16; P = 0.42). This small study indicates that the application of hyperoncotic HES 200/0.5 (10%) within the first 24 hours after severe burn injury may be associated with fatal outcome and should therefore be used with caution. NCT01120730.
 
Hydroxyethylstarch (HES) 200/0.5 is associated with renal failure. Several studies have suggested that renal function is affected but the subsequent arguments leave the clinician in no man's land. A recent study in Critical Care by Simon and colleagues using a two hit animal model of shock demonstrates that the use of a higher molecular weight starch, HES 200/0.5, is associated with impaired renal function when compared with ringers acetate, gelatin or a lower molecular weight starch, HES 130/0.42. The authors conclude that both the lower molecular weight starch and the ringers acetate 'preserve renal function and attenuate tubular damage better than 10% hydroxyethylstarch 200/0.5 in saline'. Added to the previous evidence, the renal effects of HES200/0.5 are probably real. Many clinicians have already moved to the lower molecular weight starches on the basis of doubt rather than certainty, but this study tips the balance. The cause remains elusive and the lack of a mechanism should be seen as a problem.
 
The pulmonary delivery drug system. (a) Clinical in the 'on-vent configuration' and (b) with the hand-held device.
Day 3 serum amikacin concentrations before (0), and at hours 0.5, 1, 3, 6, 9, 12, 13 and 24 after starting the first aerosol. Results are expressed as medians (interquartile range). Black arrows indicate the timing of aerosols.
Day 3 amikacin concentration in the alveolar epithelial lining fluid (ELF) of the 28 assessable patients. The dotted line corresponds to 128 μg/mL, which is 10-fold the critical 90% minimum inhibitory concentration (MIC90) for Pseudomonas aeruginosa. T-bars represent the 10th and 90th percentiles; the horizontal line in the box is the median; the lower and upper limits of the box represent the 25th and 75th percentiles, respectively. Circles represent outliers.
Day 3 amikacin concentration in the tracheal aspirates of the 19 assessable patients. H1 to H6 corresponds to the first six hours following the first aerosol, H7 to H12 to the next six hours (before the second aerosol of the day), H13 to H18 to the six hours following the second nebulization, and H19 to H24 to the last six hours of the day, before next aerosol. T-bars represent the 10th and 90th percentiles; the horizontal line in the box is the median; the lower and upper limits of the box represent the 25th and 75th percentiles, respectively. Circles represent outliers.
Serum amikacin trough concentrations during the study from day 1 (D1) to D10 with the corresponding number of patients. T-bars represent the 10th and 90th percentiles; the horizontal line in the box is the median; the lower and upper limits of the box represent the 25th and 75th percentiles, respectively. Circles represent outliers.
Aminoglycosides aerosolization might achieve better diffusion into the alveolar compartment than intravenous use. The objective of this multicenter study was to evaluate aerosol-delivered amikacin penetration into the alveolar epithelial lining fluid (ELF) using a new vibrating mesh nebulizer (Pulmonary Drug Delivery System (PDDS), Nektar Therapeutics), which delivers high doses to the lungs. Nebulized amikacin (400 mg bid) was delivered to the lungs of 28 mechanically ventilated patients with Gram-negative VAP for 7-14 days, adjunctive to intravenous therapy. On treatment day 3, 30 minutes after completing aerosol delivery, all the patients underwent bronchoalveolar lavage in the infection-involved area and the ELF amikacin concentration was determined. The same day, urine and serum amikacin concentrations were determined at different time points. Median (range) ELF amikacin and maximum serum amikacin concentrations were 976.1 (135.7-16127.6) and 0.9 (0.62-1.73) microg/mL, respectively. The median total amount of amikacin excreted in urine during the first and second 12-hour collection on day 3 were 19 (12.21-28) and 21.2 (14.1-29.98) microg, respectively. During the study period, daily through amikacin measurements were below the level of nephrotoxicity. Sixty-four unexpected adverse events were reported, among which 2 were deemed possibly due to nebulized amikacin: one episode of worsening renal failure, and one episode of bronchospasm. PDDS delivery of aerosolized amikacin achieved very high aminoglycoside concentrations in ELF from radiography-controlled infection-involved zones, while maintaining safe serum amikacin concentrations. The ELF concentrations always exceeded the amikacin minimum inhibitory concentrations for Gram-negative microorganisms usually responsible for these pneumonias. The clinical impact of amikacin delivery with this system remains to be determined. ClinicalTrials.gov Identifier: NCT01021436.
 
Introduction In order to assess the significance of drug levels measured in intensive care medicine, clinical and forensic toxicology, as well as for therapeutic drug monitoring, it is essential that a comprehensive collection of data is readily available. Therefore, it makes sense to offer a carefully referenced compilation of therapeutic and toxic plasma concentration ranges, as well as half-lives, of a large number of drugs and other xenobiotics for quick and comprehensive information. Methods Data have been abstracted from original papers and text books, as well as from previous compilations, and have been completed with data collected in our own forensic and clinical toxicology laboratory. The data presented in the table and corresponding annotations have been developed over the past 20 years and longer. A previous compilation has been completely revised and updated. In addition, more than 170 substances, especially drugs that have been introduced to the market since 2003 as well as illegal drugs, which became known to cause intoxications, were added. All data were carefully referenced and more than 200 new references were included. Moreover, the annotations providing details were completely revised and more than 100 annotations were added. Results For nearly 1,000 drugs and other xenobiotics, therapeutic ("normal") and, if data were available, toxic and comatose-fatal blood-plasma concentrations and elimination half-lives were compiled in a table. Conclusions In case of intoxications, the concentration of the ingested substances and/or metabolites in blood plasma better predicts the clinical severity of the case when compared to the assumed amount and time of ingestion. Comparing and contrasting the clinical case against the data provided, including the half-life, may support the decision for or against further intensive care. In addition, the data provided are useful for the therapeutic monitoring of pharmacotherapies, to facilitate the diagnostic assessment and monitoring of acute and chronic intoxications, and to support forensic and clinical expert opinions.
 
Introduction Morbid obesity and its consequences are considered risk factors for adverse outcome in trauma, although the pathophysiologic mechanisms are incompletely understood. The aim of this study was to compare initial resuscitation, treatment, and short-term outcome of severely injured patients by body mass index (BMI). Methods A total of 1,084 severely injured patients with an injury severity score of 16 or greater were enrolled between 1996 and 2009 and grouped according to BMI. Their course of treatment and in-hospital outcome were analyzed by univariate and multivariate comparison. Results Of these patients, 603 (55.6%) were of normal weight with a BMI between 18.5 and 24.9, 361 (33.3%) had BMI values between 25 and 29.9, and 90 patients (8.3%) were obese (BMI ≥ 30). Thirty patients (2.8%) had BMI levels below 18.5. All groups were comparable with respect to injury severity, initial resuscitation, and time to ICU admission. There was a tendency towards higher mortality in obese patients (mortality 24.4%) and also overweight patients (mortality 18.8%) when compared with patients with a normal BMI (mortality 16.6%). Obese patients showed the highest mortality on day 0 (8.9% vs. 2.8% in the normal-weight group, P = 0.023), mostly due to persistent shock (6.7%). When corrected for BMI, obese patients are provided significantly lower volumes of intravenous fluids during the initial resuscitation period. Conclusion In contrast to the mostly American literature, only a low percentage of trauma patients at a European trauma center are obese. These patients are at risk of higher mortality from persistent hemorrhagic shock in the initial phase after trauma, which may potentially be related to relative hypovolemia during the resuscitation period. In the later course of treatment, no significant differences exist with respect to specific complications, hospital stay, or in-hospital mortality.
 
The demographic and clinical data for the injured patients with or without intubation before hospitalization (600 patients per group)
The intensive-care outcomes of the trauma patients with or without prehospital intubation
Hypoxia and hypoxemia can lead to an unfavorable outcome after severe trauma, by both direct and delayed mechanisms. Prehospital intubation is meant to ensure pulmonary gas exchange. Limited evidence exists regarding indications for intubation after trauma. The aim of this study was to analyze prehospital intubation as an independent risk factor for the posttraumatic course of moderately injured patients. Therefore, only patients who, in retrospect, would not have required intubation were included in the matched-pairs analysis to evaluate the risks related to intubation. The data of 42,248 patients taken from the trauma registry of the German Association for Trauma Surgery (Deutsche Gesellschaft für Unfallchirurgie (DGU)) were analyzed. Patients who met the following criteria were included: primary admission to a hospital; Glasgow Coma Scale (GCS) of 13 to 15; age 16 years or older; maximum injury severity per body region (AIS) ≤ 3; no administration of packed red blood cell units in the emergency trauma room; admission between 2005 and 2008; and documented data regarding intubation. The intubated patients were then matched with not-intubated patients. The study population included 600 matched pairs that met the inclusion criteria. The results indicated that prehospital intubation was associated with a prolonged rescue time (not intubated, 64.8 minutes; intubated, 82.3 minutes; P ≤ 0.001) and a higher volume replacement (not intubated, 911.3 ml; intubated, 1,573.8 ml; P ≤ 0.001). In the intubated patients, coagulation parameters, such as the prothrombin time ratio (PT) and platelet count, declined, as did the hemoglobin value (PT not intubated: 92.3%; intubated, 85.7%; P ≤ 0.001; hemoglobin not intubated, 13.4 mg/dl; intubated, 12.2 mg/dl; P ≤ 0.001). Intubation at the scene resulted in an elevated sepsis rate (not intubated, 1.5%; intubated, 3.7%; P ≤ 0.02) and an elevated prevalence of multiorgan failure (MOF) and organ failure (OF) (OF not intubated, 9.1%; intubated, 23.4%; P ≤ 0.001). Prehospital intubation in trauma patients is associated with a number of risks and should be critically weighed, except in cases with clear indicators, such as posttraumatic apnea.
 
Characteristics of singleinjury and multiple-injury patients.
Diagnoses by body region of 3,177 injury diagnoses registered at People's Hospital of Deyang City after the 2008 Sichuan earthquake, China
Most frequent combinations of injury diagnoses by body region in 856 two-injury patients.
Most frequent combinations of injury diagnosess in 856 two-injury patients.
Flow diagram of the sample obtained at the People's Hospital of Deyang city (PHDC).
Introduction Multiple injuries have been highlighted as an important clinical dimension of the injury profile following earthquakes, but studies are scarce. We investigated the pattern and combination of injuries among patients with two injuries following the 2008 Wenchuan earthquake. We also described the general injury profile, causes of injury and socio-demographic characteristics of the injured patients. Methods A retrospective hospital-based analysis of 1,871 earthquake injured patients, totaling 3,177 injuries, admitted between 12 and 31 May 2008 to the People's Hospital of Deyang city (PHDC). An electronic, webserver-based database with International Classification of Diseases (ICD)-10-based classification of earthquake-related injury diagnoses (IDs), anatomical sites and additional background variables of the inpatients was used. We analyzed this dataset for injury profile and number of injuries per patient. We then included all patients (856) with two injuries for more in-depth analysis. Possible spatial anatomical associations were determined a priori. Cross-tabulation and more complex frequency matrices for combination analyses were used to investigate the injury profile. Results Out of the 1,871 injured patients, 810 (43.3%) presented with a single injury. The rest had multiple injuries; 856 (45.8%) had two, 169 (9.0%) patients had three, 32 (1.7%) presented with four injuries, while only 4 (0.2%) were diagnosed with five injuries. The injury diagnoses of patients presenting with two-injuries showed important anatomical intra-site or neighboring clustering, which explained 49.1% of the combinations. For fractures, the result was even more marked as spatial clustering explained 57.9% of the association pattern. The most frequent combination of IDs was a double-fracture, affecting 20.7% of the two-injury patients (n = 177). Another 108 patients (12.6%) presented with fractures associated with crush injury and organ-soft tissue injury. Of the 3,177 injuries, 1,476 (46.5%) were fractures. Most injuries were located in the head (22.9%) and lower extremities (30.8%). Conclusions Multiple injuries are put forward as an important component of the injury profile after this earthquake. A pattern of injury combinations and spatial aggregation of injuries was also found. Clinical diagnosis and treatment should be adapted to care of these patients. More studies are needed to generalize these findings.
 
Observed versus predicted probability of all-cause CVICU readmission and number needed to identify (NNI) one readmission across risk deciles. The number needed to identify (NNI) one readmission is lower in higher readmission risk deciles. CVIVU: cardiovascular intensive care unit.
Clinical characteristics of cardiac surgical patients with and without a cardiovascular intensive care unit readmission
The APPROACH CVICU readmission risk model nomogram. (A) The CVICU readmission risk score and (B) mean predicted CVICU readmission by APPROACH CVICU readmission risk score. APPROACH: Alberta Provincial Project for Outcomes Assessment in Coronary Heart Disease; CVICU: cardiovascular intensive care unit.
Operative variables in patients with and without a cardiovascular intensive care unit readmission
Postoperative complications in patients with and without a cardiovascular intensive care unit readmission
Introduction: In medical and surgical intensive care units, clinical risk prediction models for readmission have been developed; however, studies reporting the risks for cardiovascular intensive care unit (CVICU) readmission have been methodologically limited by small numbers of outcomes, unreported measures of calibration or discrimination, or a lack of information spanning the entire perioperative period. The purpose of this study was to derive and validate a clinical prediction model for CVICU readmission in cardiac surgical patients. Methods: A total of 10,799 patients more than or equal to 18 years in the Alberta Provincial Project for Outcomes Assessment in Coronary Heart Disease (APPROACH) registry who underwent cardiac surgery (coronary artery bypass or valvular surgery) between 2004 and 2012 and were discharged alive from the first CVICU admission were included. The full cohort was used to derive the clinical prediction model and the model was internally validated with bootstrapping. Discrimination and calibration were assessed using the AUC c index and the Hosmer-Lemeshow tests, respectively. Results: A total of 479 (4.4%) patients required CVICU readmission. The mean CVICU length of stay (19.9 versus 3.3 days, P <0.001) and in-hospital mortality (14.4% versus 2.2%, P <0.001) were higher among patients readmitted to the CVICU. In the derivation cohort, a total of three preoperative (age ≥ 70, ejection fraction, chronic lung disease), two intraoperative (single valve repair or replacement plus non-CABG surgery, multivalve repair or replacement), and seven postoperative variables (cardiac arrest, pneumonia, pleural effusion, deep sternal wound infection, leg graft harvest site infection, gastrointestinal bleed, neurologic complications) were independently associated with CVICU readmission. The clinical prediction model had robust discrimination and calibration in the derivation cohort (AUC c index = 0.799; Hosmer-Lemeshow P = 0.192). The validation point estimates and confidence intervals were similar to derivation model. Conclusions: In a large population-based dataset incorporating a comprehensive set of perioperative variables, we have derived a clinical prediction model with excellent discrimination and calibration. This model identifies opportunities for targeted therapeutic interventions aimed at reducing CVICU readmissions in high-risk patients.
 
Comparison of values for biomarkers between dead and alive 
Comparison of values for biomarkers between remained coma and regained consciousness 
Values of cutoff points and predictive accuracy for remained coma 
Neurological prognostic factors after cardiopulmonary resuscitation (CPR) in patients with cardiac arrest (CA) as early and accurately as possible are urgently needed to determine therapeutic strategies after successful CPR. In particular, serum levels of protein neuron-specific enolase (NSE) and S-100B are considered promising candidates for neurological predictors, and many investigations on the clinical usefulness of these markers have been published. However, the design adopted varied from study to study, making a systematic literature review extremely difficult. The present review focuses on the following three respects for the study design: definitions of outcome, value of specificity and time points of blood sampling. A Medline search of literature published before August 2008 was performed using the following search terms: "NSE vs CA or CPR", "S100 vs CA or CPR". Publications examining the clinical usefulness of NSE or S-100B as a prognostic predictor in two outcome groups were reviewed. All publications met with inclusion criteria were classified into three groups with respect to the definitions of outcome; "dead or alive", "regained consciousness or remained comatose", and "return to independent daily life or not". The significance of differences between two outcome groups, cutoff values and predictive accuracy on each time points of blood sampling were investigated. A total of 54 papers were retrieved by the initial text search, and 24 were finally selected. In the three classified groups, most of the studies showed the significance of differences and concluded these biomarkers were useful for neurological predictor. However, in view of blood sampling points, the significance was not always detected. Nevertheless, only five studies involved uniform application of a blood sampling schedule with sampling intervals specified based on a set starting point. Specificity was not always set to 100%, therefore it is difficult to indiscriminately assess the cut-off values and its predictive accuracy of these biomarkers in this meta analysis. In such circumstances, the findings of the present study should aid future investigators in examining the clinical usefulness of these markers and determination of cut-off values.
 
Details of medical and surgical in-flight emergencies. Percentages are based on 10,189 incidents from two European airlines January 2002 to December 2007.
In-flight medical and surgical emergencies (IMEs) onboard commercial aircrafts occur quite commonly. However, little epidemiological research exists concerning these incidents. Thirty-two European airlines were asked to provide anonymous data on medical flight reports of IMEs for the years 2002 to 2007. The total number of incidents was correlated to revenue passenger kilometers (rpk). Additionally, on-board births and deaths, flight diversions, flight routes (continental/intercontinental) and involvement of a physician or medical professional in providing therapy were analysed. Only four airlines, of which two participated in this study, were able to provide the necessary data. A total of 10,189 cases of IMEs were analysed. Syncope was the most common medical condition reported (5307 cases, 53.5%) followed by gastrointestinal disorders (926 cases, 8.9%) and cardiac conditions (509 cases, 4.9%). The most common surgical conditions were thrombosis (47 cases, 0.5%) and appendicitis (27 cases, 0.25%). In 2.8% of all IMEs, an aircraft diversion was performed. In 86% of cases, a physician or medical professional was involved in providing therapy. A mean (standard deviation) of 14 (+/- 2.3, 10.8 to 16.6 interquartile range) IMEs per billion rpk was calculated. The study demonstrates that although aviation is regulated by a variety of national and international laws, standardised documentation of IMEs is inadequate and needs further development.
 
Initial blood pressure and target ranges at enrollment
Patient entry into CLUE trial.
SBP changes over time in patients randomized to receive either nicardipine or labetalol. Mean percent change and 95% confidence interval (CI), evaluated by Student's T test, relative to presenting blood pressure, during the initial 30 minutes, with the median upper level of target range and median target range indicated by horizontal dotted lines, in patients randomized to receive either nicardipine or labetalol. SBP, systolic blood pressure.
Heart rate changes over time in patients randomized to receive either nicardipine or labetalol. CI, confidence interval; HR, heart rate.
multivariable logistic regression model † for "met target systolic blood pressure within first 30 minutes".
Our purpose was to compare the safety and efficacy of food and drug administration (FDA) recommended dosing of IV nicardipine versus IV labetalol for the management of acute hypertension. Multicenter randomized clinical trial. Eligible patients had 2 systolic blood pressure (SBP) measures ≥180 mmHg and no contraindications to nicardipine or labetalol. Before randomization, the physician specified a target SBP ± 20 mmHg (the target range: TR). The primary endpoint was the percent of subjects meeting TR during the initial 30 minutes of treatment. Of 226 randomized patients, 110 received nicardipine and 116 labetalol. End organ damage preceded treatment in 143 (63.3%); 71 nicardipine and 72 labetalol patients. Median initial SBP was 212.5 (IQR 197, 230) and 212 mmHg (IQR 200,225) for nicardipine and labetalol patients (P = 0.68), respectively. Within 30 minutes, nicardipine patients more often reached TR than labetalol (91.7 vs. 82.5%, P = 0.039). Of 6 BP measures (taken every 5 minutes) during the study period, nicardipine patients had higher rates of five and six instances within TR than labetalol (47.3% vs. 32.8%, P = 0.026). Rescue medication need did not differ between nicardipine and labetalol (15.5 vs. 22.4%, P = 0.183). Labetalol patients had slower heart rates at all time points (P < 0.01). Multivariable modeling showed nicardipine patients were more likely in TR than labetalol patients at 30 minutes (OR 2.73, P = 0.028; C stat for model = 0.72) Patients treated with nicardipine are more likely to reach the physician-specified SBP target range within 30 minutes than those treated with labetalol.
 
Bland and Altman's graph for end-hour intracranial pressure (ICP)
Bland and Altman's graph for mean hourly intracranial pressure (ICP)
Bar graphs depicting the capabilities of the digital and manual systems to capture increases in intracranial pressure (ICP)
Bar graphs depicting the (a) digital and (b) manual percentages of time of intracranial hypertension (HICP) (> 20 mm Hg) for each patient and (c) the difference of digital and manual percentages of time of HICP
The presence of intracranial hypertension (HICP) after traumatic brain injury (TBI) affects patient outcome. Intracranial pressure (ICP) data from electronic monitoring equipment are usually calculated and recorded hourly in the clinical chart by trained nurses. Little is known, however, about how precisely this method reflects the real patterns of ICP after severe TBI. In this study, we compared hourly manual recording with a validated and continuous computerized reference standard. Thirty randomly selected patients with severe TBI and HICP admitted to the neuroscience intensive care unit (Policlinico University Hospital, Milan, Italy) were retrospectively studied. A 24-hour interval with ICP monitoring was randomly selected for each patient. The manually recorded data available for analysis covered 672 hours corresponding to 36,492 digital data points. The two methods were evaluated using the correlation coefficient and the Bland and Altman method. We used the proportion test to analyze differences in the number of episodes of HICP (ICP > 20 mm Hg) detected with the two methods and the paired t test to analyze differences in the percentage of time of HICP. There was good agreement between the digitally collected ICP and the manual recordings of the end-hour values. Bland and Altman analysis confirmed a mean difference between the two methods of 0.05 mm Hg (standard deviation 3.66); 96% of data were within the limits of agreement (+7.37 and -7.28). The average percentages of time of ICP greater than 20 mm Hg were 39% calculated from the digital measurements and 34% from the manual observations. From the continuous digital recording, we identified 351 episodes of ICP greater than 20 mm Hg lasting at least five minutes and 287 similar episodes lasting at least ten minutes. Conversely, end-hour ICP of greater than 20 mm Hg was observed in only 204 cases using manual recording methods. Although manually recorded end-hour ICP accurately reflected the computerized end-hour and mean hour values, the important omission of a number of episodes of high ICP, some of long duration, results in a clinical picture that is not accurate or informative of the true pattern of unstable ICP in patients with TBI.
 
At 07:39 on 11 March 2004, 10 terrorist bomb explosions occurred almost simultaneously in four commuter trains in Madrid, Spain, killing 177 people instantly and injuring more than 2000. There were 14 subsequent in-hospital deaths, bringing the ultimate death toll to 191. This report describes the organization of clinical management and patterns of injuries in casualties who were taken to the closest hospital, with an emphasis on the critically ill. A total of 312 patients were taken to the hospital and 91 patients were hospitalized, of whom 89 (28.5%) remained in hospital for longer than 24 hours. Sixty-two patients had only superficial bruises or emotional shock, but the remaining 250 patients had more severe injuries. Data on 243 of these 250 patients form the basis of this report. Tympanic perforation occurred in 41% of 243 victims with moderate-to-severe trauma, chest injuries in 40%, shrapnel wounds in 36%, fractures in 18%, first-degree or second-degree burns in 18%, eye lesions in 18%, head trauma in 12% and abdominal injuries in 5%. Between 08:00 and 17:00, 34 surgical interventions were performed in 32 patients. Twenty-nine casualties (12% of the total, or 32.5% of those hospitalized) were deemed to be in a critical condition, and two of these died within minutes of arrival. The other 27 survived to admission to intensive care units, and three of them died, bringing the critical mortality rate to 17.2% (5/29). The mean Injury Severity Score and Acute Physiology and Chronic Health Evaluation II scores for critically ill patients were 34 and 23, respectively. Among these critically ill patients, soft tissue and musculoskeletal injuries predominated in 85% of cases, ear blast injury was identified in 67% and blast lung injury was present in 63%. Fifty-two per cent suffered head trauma. Over-triage to the closest hospital probably occurred, and the time of the blasts proved to be crucial to the the adequacy of the medical and surgical response. The number of blast lung injuries seen is probably the largest reported by a single institution, and the critical mortality rate was reasonably low.
 
Predicting whether a critically ill patient will survive intensive care treatment remains difficult. The advantages of a validated strategy to identify those patients who will not benefit from intensive care unit (ICU) treatment are evident. Providing critical care treatment to patients who will ultimately die in the ICU is accompanied by an enormous emotional and physical burden for both patients and their relatives. The purpose of the present study was to examine whether health-related quality of life (HRQOL) before admission to the ICU can be used as a predictor of mortality. We conducted a prospective cohort study in a university-affiliated teaching hospital. Patients admitted to the ICU for longer than 48 hours were included. Close relatives completed the Short-form 36 (SF-36) within the first 48 hours of admission to assess pre-admission HRQOL of the patient. Mortality was evaluated from ICU admittance until 6 months after ICU discharge. Logistic regression and receiver operating characteristic analyses were used to assess the predictive value for mortality using five models: the first question of the SF-36 on general health (model A); HRQOL measured using the physical component score (PCS) and mental component score (MCS) of the SF-36 (model B); the Acute Physiology and Chronic Health Evaluation (APACHE) II score (an accepted mortality prediction model in ICU patients; model C); general health and APACHE II score (model D); and PCS, MCS and APACHE II score (model E). Classification tables were used to assess the sensitivity, specificity, positive and negative predictive values, and likelihood ratios. A total of 451 patients were included within 48 hours of admission to the ICU. At 6 months of follow up, 159 patients had died and 40 patients were lost to follow up. When the general health item was used as an estimate of HRQOL, area under the curve for model A (0.719) was comparable to that of model C (0.721) and slightly better than that of model D (0.760). When PCS and MCS were used, the area under the curve for model B (0.736) was comparable to that of model C (0.721) and slightly better than that of model E (0.768). When using the general health item, the sensitivity and specificity in model D (sensitivity 0.52 and specificity 0.81) were similar to those in model A (0.45 and 0.80). Similar results were found when using the MCS and PCS. This study shows that the pre-admission HRQOL measured with either the one-item general health question or the complete SF-36 is as good at predicting survival/mortality in ICU patients as the APACHE II score. The value of these measures in clinical practice is limited, although it seems sensible to incorporate assessment of HRQOL into the many variables considered when deciding whether a patient should be admitted to the ICU.
 
Number of patients according to genotype and disease outcome 
, the RR can be expressed as follows: 
Outcomes of the study conducted by Rivers and coworkers 
Relative risk and odds ratio have been introduced in earlier reviews (see Statistics reviews 3, 6 and 8). This review describes the calculation and interpretation of their confidence intervals. The different circumstances in which the use of either the relative risk or odds ratio is appropriate and their relative merits are discussed. A method of measuring the impact of exposure to a risk factor is introduced. Measures of the success of a treatment using data from clinical trials are also considered.
 
Inhalation injury in combination with a severe thermal injury increases mortality. Alterations in inflammatory mediators, such as cytokines, contribute to the incidence of multi-organ failure and mortality. The aim of the present study was to determine the effect of inhalation injury on cytokine expression in severely burned children. Thirty severely burned pediatric patients with inhalation injury and 42 severely burned children without inhalation injury were enrolled in the study. Inhalation injury was diagnosed by bronchoscopy during the first operation. Blood was collected within 24 hours of admission and again at five to seven days following admission. Cytokine expression was profiled using multi-plex antibody-coated beads. Significance was accepted at a p value of less than 0.05. The mean percentages of total body surface area burned were 67% +/- 4% (56% +/- 6%, third-degree burns) in the inhalation injury group and 60% +/- 3% (45% +/- 3%, third-degree burns) in the non-inhalation injury group (p value not significant [NS]). Mean age was 9 +/- 1 years in the inhalation injury group and 8 +/- 1 years in the non-inhalation injury group (p value NS). Time from burn to admission in the inhalation injury group was 2 +/- 1 days compared to 3 +/- 1 days in the non-inhalation injury group (p value NS). Mortalities were 40% in the inhalation injury group and 12% in the non-inhalation injury group (p < 0.05). At the time of admission, serum interleukin (IL)-7 was significantly increased in the non-inhalation injury group, whereas IL-12p70 was significantly increased in the inhalation injury group compared to the non-inhalation injury group (p < 0.05). There were no other significant differences between groups. Five to seven days following admission, all cytokines decreased with no differences between the inhalation injury and non-inhalation injury cohorts. In the present study, we show that an inhalation injury causes alterations in IL-7 and IL-12p70. There were no increased levels of pro-inflammatory cytokines, indicating that an inhalation injury in addition to a burn injury does not augment the systemic inflammatory response early after burn.
 
Japan was struck by a magnitude 9.0 earthquake and a tsunami on 11 March 2011. Although this catastrophe has caused the most devastating damage to Japan since World War II, we believe that our systematic preparation for disasters somewhat alleviated the damage. Learning lessons from the magnitude 7.3 Great Hanshin earthquake in 1995, the government organized approximately 700 medical teams specialized in disaster management. In this earthquake of 2011, hundreds of medical teams were successfully deployed and started operations within the first 72 hours. Furthermore, the internet, which was not commonly used in 1995, made significant contributions in communication among clinicians and enabled them to promptly identify the needs of the affected hospitals. In addition, medical professional societies took leadership in the logistics of transferring victims away from the disaster zone. We also observed that the spectrum of causes of death is distinct between the earthquakes of 1995 and 2011. In 1995, many victims died from trauma, including crash injury, and delays in providing hemodialysis contributed to additional deaths. In 2011, in contrast, many victims died from drowning in the tsunami, and most survivors did not have life-threatening injuries.
 
Acute kidney injury (AKI) is a frequent complication of cardiopulmonary bypass (CPB). The lack of early biomarkers has impaired our ability to intervene in a timely manner. We previously showed in a small cohort of patients that plasma neutrophil gelatinase-associated lipocalin (NGAL), measured using a research enzyme-linked immunosorbent assay, is an early predictive biomarker of AKI after CPB. In this study we tested whether a point-of-care NGAL device can predict AKI after CPB in a larger cohort. First, in a cross-sectional pilot study including 40 plasma samples (NGAL range 60 to 730 ng/ml) and 12 calibration standards (NGAL range 0 to 1,925 ng/ml), NGAL measurements by enzyme-linked immunosorbent assay and by Triage NGAL Device (Biosite Inc., San Diego, CA, USA) were highly correlated (r = 0.94). Second, in a subsequent prospective uncontrolled cohort study, 120 children undergoing CPB were enrolled. Plasma was collected at baseline and at frequent intervals for 24 hours after CPB, and analyzed for NGAL using the Triage(R) NGAL device. The primary outcome was AKI, which was defined as a 50% or greater increase in serum creatinine. AKI developed in 45 patients (37%), but the diagnosis using serum creatinine was delayed by 2 to 3 days after CPB. In contrast, mean plasma NGAL levels increased threefold within 2 hours of CPB and remained significantly elevated for the duration of the study. By multivariate analysis, plasma NGAL at 2 hours after CPB was the most powerful independent predictor of AKI (beta = 0.004, P < 0.0001). For the 2-hour plasma NGAL measurement, the area under the curve was 0.96, sensitivity was 0.84, and specificity was 0.94 for prediction of AKI using a cut-off value of 150 ng/ml. The 2 hour postoperative plasma NGAL levels strongly correlated with change in creatinine (r = 0.46, P < 0.001), duration of AKI (r = 0.57, P < 0.001), and length of hospital stay (r = 0.44, P < 0.001). The 12-hour plasma NGAL strongly correlated with mortality (r = 0.48, P = 0.004) and all measures of morbidity mentioned above. Accurate measurements of plasma NGAL are obtained using the point-of-care Triage(R) NGAL device. Plasma NGAL is an early predictive biomarker of AKI, morbidity, and mortality after pediatric CPB.
 
Determination of De Backer's score [3]. Vessel density is calculated as the number of vessels crossing the lines divided by the total length of the lines. Perfusion is then categorized by eye as present (continuous flow for at least 20 s), absent (no flow for at least 20 s) or intermittent (at least 50% of time with no flow). The proportion of perfused vessels (PPV [%]) and perfused vessel density (PVD) are then calculated. A 20 μm cut-off is used to separate small vessels (mostly capillaries) from large vessels (mostly venules).
Determination of mean flow index (MFI) score [15]. The image is divided into four quadrants and the predominant type of flow (absent = 0, intermittent = 1, sluggish = 2, and normal = 3) is assessed in each quadrant. The MFI score represents the averaged values of the four. A 20 μm cut-off is used to separate small vessels (mostly capillaries) from large vessels (mostly venules).
Change in image size during software stabilization. When movements occur, software can re-centre the image using easily recognized structures. However, peripheral parts of the images, not seen on successive images, will be lost so that the final area will be smaller than the original one. The size of the original image is represented by the light grey rectangle, and the final one by the light blue rectangle.
Microvascular alterations may play an important role in the development of organ failure in critically ill patients and especially in sepsis. Recent advances in technology have allowed visualization of the microcirculation, but several scoring systems have been used so it is sometimes difficult to compare studies. This paper reports the results of a round table conference that was organized in Amsterdam in November 2006 in order to achieve consensus on image acquisition and analysis. The participants convened to discuss the various aspects of image acquisition and the different scores, and a consensus statement was drafted using the Delphi methodology. The participants identified the following five key points for optimal image acquisition: five sites per organ, avoidance of pressure artifacts, elimination of secretions, adequate focus and contrast adjustment, and recording quality. The scores that can be used to describe numerically the microcirculatory images consist of the following: a measure of vessel density (total and perfused vessel density; two indices of perfusion of the vessels (proportion of perfused vessels and microcirculatory flow index); and a heterogeneity index. In addition, this information should be provided for all vessels and for small vessels (mostly capillaries) identified as smaller than 20 microm. Venular perfusion should be reported as a quality control index, because venules should always be perfused in the absence of pressure artifact. It is anticipated that although this information is currently obtained manually, it is likely that image analysis software will ease analysis in the future. We proposed that scoring of the microcirculation should include an index of vascular density, assessment of capillary perfusion and a heterogeneity index.
 
There has been controversy regarding the mortality directly attributed to Acinetobacter baumannii infections. Data from six case-control studies have been recently added to the literature regarding the attributable mortality of A. baumannii infections during the past months. The information from these studies, added to the previous knowledge on this issue, provides evidence that A. baumannii infections are indeed associated with increased mortality. In addition, there is relevant evidence from studies examining the effect of inappropriate treatment on mortality; specifically, inappropriate treatment of A. baumannii infections has been associated with excess mortality. We believe that the accumulated data suggest that attributable mortality due to A. baumannii infections should no longer be a controversial issue. The efforts of the scientific community interested in this pathogen should therefore be directed to the development and introduction of new antibiotics effective against multidrug-resistant and pandrug-resistant A. baumannii as well as the implementation of infection control measures that may help us in the control of the increasing problem of A. baumannii infections.
 
Data continue to emerge demonstrating the poor quality of life of ICU survivors in the months and years following critical illness. In this issue of Critical Care, Cuthbertson and colleagues present new data on quality of life from a cohort of ICU survivors who were followed for 5 years. They found that survivors had poor physical quality of life and low quality adjusted life-years in comparison to age-adjusted norms, describing the long-term impact of critical illness as similar to a co-morbidity. Studies are now needed that seek to identify potentially modifiable factors both during and following an ICU admission to allow for eventual improvement in long-term morbidity. Such studies will likely need to incorporate extensive planning for data collection, as well as coordinated linkage with other available datasets that include substantial amounts of patient information from outside of the ICU.
 
Plasma microRNA-124 was already elevated tenfold at the time of admission in patients compared with controls. The horizontal lines represent the medians. miR-124, MicroRNA-124.
A poor outcome was associated with a significant elevation of microRNA-124 at 48 hours after cardiac arrest. The horizontal lines represent the medians. CPC, Cerebral Performance Category scale score; mIR-124, MicroRNA-124. ***P < 0.0001.
Receiver operating characteristic area under the curve of microRNA-124 at 48 hours after cardiac arrest. AUC, Area under the curve; CI, 95% confidence interval.
Plasma microRNA-124 was significantly elevated in patients with poor outcomes at 24 hours after cardiac arrest. The horizontal lines represent the medians. CPC, Cerebral Performance Category scale score; mIR-124, MicroRNA-124. ***P < 0.0001.
Receiver operating characteristic area under the curve of microRNA-124 at 24 hours after cardiac arrest. AUC, Area under the curve; CI, 95% confidence interval.
Early prognostication after successful cardiopulmonary resuscitation is difficult and there is a need for novel methods to estimate the extent of brain injury and predict outcome. In this study, we evaluate the impact of the cardiac arrest syndrome on the plasma levels of selected tissue-specific microRNAs (miRNAs) and assess their ability to prognosticate death and neurological disability. We included 65 patients treated with hypothermia after cardiac arrest in the study. Blood samples were obtained at 24 hours and at 48 hours. For miRNA-screening purposes, custom quantitative polymerase chain reaction (qPCR)-panels were first used. Thereafter, individual miRNAs were assessed at 48 hours with qPCR. miRNAs successful at predicting prognosis at 48 hours were further analyzed at 24 hours. Outcome was measured according to the Cerebral Performance Categories scale (CPC) at 6 months after cardiac arrest and stratified into good (CPC 1-2) or poor (CPC 3-5). At 48 hours, miR-146a, miR-122, miR-208b, miR-21, miR-9 and miR-128 did not differ between the good and poor neurological outcome groups. In contrast, miR-124 was significantly elevated in patients with poor outcome compared to those with a favorable outcome (P <0.0001) at 24 hours and 48 hours after cardiac arrest. Analysis of receiver operating curves showed an area under the curve of 0.87 (95% confidence interval (CI) 0.79 to 0.96) at 24 hours and of 0.89 (95% CI 0.80 to 0.97) at 48 hours after cardiac arrest. The brain-enriched miRNA miR-124 is a promising novel biomarker for prediction of neurological prognosis following cardiac arrest.
 
Intensive-care-unit-acquired weakness is a major complication in critically ill patients. The paper by Hough and coworkers suggests that the current method of manual muscle strength testing with the Medical Research Council sum score is of limited value in the intensive care unit. However, their results raise a number of questions and provide important lessons for implementation of such evaluations in the intensive care unit.
 
Our goal was to assess the effects of titration of a norepinephrine infusion to increasing levels of mean arterial pressure (MAP) on sublingual microcirculation. Twenty septic shock patients were prospectively studied in two teaching intensive care units. The patients were mechanically ventilated and required norepinephrine to maintain a mean arterial pressure (MAP) of 65 mmHg. We measured systemic hemodynamics, oxygen transport and consumption (DO2 and VO2), lactate, albumin-corrected anion gap, and gastric intramucosal-arterial PCO2 difference (DeltaPCO2). Sublingual microcirculation was evaluated by sidestream darkfield (SDF) imaging. After basal measurements at a MAP of 65 mmHg, norepinephrine was titrated to reach a MAP of 75 mmHg, and then to 85 mmHg. Data were analyzed using repeated measurements ANOVA and Dunnett test. Linear trends between the different variables and increasing levels of MAP were calculated. Increasing doses of norepinephrine reached the target values of MAP. The cardiac index, pulmonary pressures, systemic vascular resistance, and left and right ventricular stroke work indexes increased as norepinephrine infusion was augmented. Heart rate, DO2 and VO2, lactate, albumin-corrected anion gap, and DeltaPCO2 remained unchanged. There were no changes in sublingual capillary microvascular flow index (2.1 +/- 0.7, 2.2 +/- 0.7, 2.0 +/- 0.8) and the percent of perfused capillaries (72 +/- 26, 71 +/- 27, 67 +/- 32%) for MAP values of 65, 75, and 85 mmHg, respectively. There was, however, a trend to decreased capillary perfused density (18 +/- 10,17 +/- 10,14 +/- 2 vessels/mm2, respectively, ANOVA P = 0.09, linear trend P = 0.045). In addition, the changes of perfused capillary density at increasing MAP were inversely correlated with the basal perfused capillary density (R2 = 0.95, P < 0.0001). Patients with septic shock showed severe sublingual microcirculatory alterations that failed to improve with the increases in MAP with norepinephrine. Nevertheless, there was a considerable interindividual variation. Our results suggest that the increase in MAP above 65 mmHg is not an adequate approach to improve microcirculatory perfusion and might be harmful in some patients.
 
Number of patients according to level of lactate and mortality
Number of patients according to level of lactate, using a range of cut-off values, and mortality plus sensitivities and specificities
This review introduces some commonly used methods for assessing the performance of a diagnostic test. The sensitivity, specificity and likelihood ratio of a test are discussed. The uses of the receiver operating characteristic curve and the area under the curve are explained.
 
We present a consensus report from the SFAR/SLRF (Société Française d'Anesthésie et de Réanimation/Société de Réanimation de Langue Française) Consensus Conference, held on 13 October 2005 in Paris, France. The consensus report made recommendations on five topics relevant to the treatment of circulatory failure in sepsis and its underlying rationale. These topics are as follows: therapeutic goals of haemodynamic support in sepsis; goals of fluid resuscitation (including transfusion); role of inotropes and vasoactive drugs; role of other treatments; and treatment strategy. This report is reproduced from a translation of the original in Annales Francaises of Anesthésie and Réanimation.
 
The recent bus crash in Switzerland involving many children provides several lessons for the pre-hospital care community. The use of multiple helicopters that are capable of flying at night and that carry advanced medical pre-hospital teams undoubtedly saved lives following the tragedy. We describe the medical response to the incident and the lessons that can be learned for emergency medical services.
 
Recently, it has been shown in several experimental settings that the noble gases xenon and helium have neuroprotective properties. In this study we tested the hypothesis that the noble gas argon has a neuroprotective potential as well. Since traumatic brain injury and stroke are widespread and generate an enormous economic and social burden, we investigated the possible neuroprotective effect in in vitro models of traumatic brain injury and cerebral ischemia. Organotypic hippocampal slice cultures from mice pups were subjected to either oxygen-glucose deprivation or to a focal mechanical trauma and subsequently treated with three different concentrations (25, 50 and 74%) of argon immediately after trauma or with a two-or-three-hour delay. After 72 hours of incubation tissue injury assessment was performed using propidium iodide, a staining agent that becomes fluorescent when it diffuses into damaged cells via disintegrated cell membranes. We could show argon's neuroprotective effects at different concentrations when applied directly after oxygen-glucose deprivation or trauma. Even three hours after application, argon was still neuroprotective. Argon showed a neuroprotective effect in both in vitro models of oxygen-glucose deprivation and traumatic brain injury. Our promising results justify further in vivo animal research.
 
Comparison of biological parameters in patients as a function of the 30% median threshold of ADAMTS13 activity
Severity of disease at presentation and outcome according to the number of hemostatic abnormalities in 72 patients with septic shock
Kaplan-Meier graph for cumulative survival according to the number of hemostatic abnormalities (A) or according to ADAMTS13 activity (B) in 72 patients with septic shock. Group 1: no DIC and ADAMTS13 activity >30%; group 2: either DIC or ADAMTS13 activity <30%; group 3: DIC and ADAMTS13 activity <30%: 30% corresponds to the median of ADAMTS13 activity in the 72 patients with septic shock. ADAMTS13, a disintegrin and metalloprotease with thrombospondin type 1 repeats, member 13; DIC, disseminated intravascular coagulation.
ADAMTS13 (a disintegrin and metalloprotease with thrombospondin type 1 repeats, member 13) deficiency has been reported in patients with sepsis but its clinical relevance and pathophysiology remain unclear. Our objectives were to assess the clinical significance, prognostic value and pathophysiology of ADAMTS13 deficiency in patients with septic shock with and without disseminated intravascular coagulation (DIC). This was a prospective monocenter cohort study of patients with septic shock. Von Willebrand Factor, ADAMTS13-related parameters and plasma IL-6 concentration were measured at inclusion to the study. Patients were categorized into three groups according to the presence of ADAMT13 deficiency (<30%) or DIC. This study included 72 patients with a median age of 59 years (interquartile range (IQR) 50 to 71). Each of the included patients received vasopressors; 55 (76%) were under mechanical ventilation and 22 (33%) underwent renal replacement therapy. Overall, 19 patients (26%) had DIC, and 36 patients had ADMTS13 deficiency (50%). Patients with DIC, ADAMTS13 deficiency or both were more severe at ICU admission. Mortality was higher in septic shock patients from group one. By multivariate analysis, Simplified Acute Physiology Score 2 (SAPS2) score (odds ratio (OR) 1.11/point; 95% CI 1.01 to 1.24) and ADAMTS13 activity <30% (OR 11.86; 95% CI 1.36 to 103.52) were independently associated with hospital mortality. There was no correlation between ADAMTS13 activity and the International Society for Thrombosis and Haemostasis (ISTH) score (rs = -0.97, P = 0.41) suggesting that ADAMTS13 functional deficiency and DIC were independent parameters. IL-6 level was higher in patients with ADAMTS13 activity <30% [895 (IQR 330 to 1843) pg/mL versus 83 (IQR 43 to 118), P = 0.0003). Septic shock was associated with a functional deficiency of ADAMTS13, independently of DIC. ADAMTS13 functional deficiency is then a prognostic factor for mortality in septic shock patients, independently of DIC.
 
Organ dysfunction. Maximum number of organ dysfunctions during the intensive care unit (ICU) stay (upper panel) and hospital mortality according to the number of organ dysfunctions (lower panel) in the three groups of patients. White bars = no cancer; gray bars = haematological cancer; black bars = solid tumours.
Kaplan Meier 60-day survival curves of the three groups of patients. Log Rank score = 20.78; p < 0.01.
Hospital mortality in the three groups of patients overall and in patients with sepsis. White bars = no cancer; gray bars = haematological cancer; black bars = solid tumours. *p < 0.001 versus no-cancer group.
Increasing numbers of cancer patients are being admitted to the intensive care unit (ICU), either for cancer-related complications or treatment-associated side effects, yet there are relatively few data concerning the epidemiology and prognosis of cancer patients admitted to general ICUs. The aim of this study was to assess the characteristics of critically ill cancer patients, and to evaluate their prognosis. This was a substudy of the Sepsis Occurrence in Acutely Ill Patients (SOAP) study, a cohort, multicentre, observational study that included data from all adult patients admitted to one of 198 participating ICUs from 24 European countries during the study period. Patients were followed up until death, hospital discharge or for 60 days. Of the 3147 patients enrolled in the SOAP study, 473 (15%) had a malignancy, 404 (85%) had solid tumours and 69 (15%) had haematological cancer. Patients with solid cancers had the same severity of illness as the non-cancer population, but were older, more likely to be a surgical admission and had a higher frequency of sepsis. Patients with haematological cancer were more severely ill and more commonly had sepsis, acute lung injury/acute respiratory distress syndrome, and renal failure than patients with other malignancies; these patients also had the highest hospital mortality rate (58%). The outcome of all cancer patients was comparable with that in the non-cancer population, with a 27% hospital mortality rate. However, in the subset of patients with more than three failing organs, more than 75% of patients with cancer died compared with about 50% of patients without cancer (p = 0.01). In this large European study, patients with cancer were more often admitted to the ICU for sepsis and respiratory complications than other ICU patients. Overall, the outcome of patients with solid cancer was similar to that of ICU patients without cancer, whereas patients with haematological cancer had a worse outcome.
 
In a meta-analysis of 34 randomized trials evaluating hydroxyethyl starch (HES) 130/0.4 in 9,587 patients and a meta-analysis evaluating HES 130/0.42 in 804 patients, HES 130 was shown to increase mortality and the need for renal replacement therapy (RRT) [1]. In the largest included trial of the meta-analysis, the Crystalloid versus Hydroxyethyl Starch Trial (CHEST) with 7,000 ICU patients [2], HES 130/0.4 increased the need for RRT despite a low average daily dose of only 526 ml. The meta-analysis also included RRT data from the CRYSTMAS trial that had not been reported in the original publication of that trial but were later published in a letter to Critical Care [3] and also incorporated in revised US Prescribing Information for HES 130/0.4 [4]. Some authors have nevertheless recently sought to defend the renal safety of HES 130/0.4 [5], in part by citing the absence of significant signs indicating renal dysfunction in a retrospective study by Boussekey and colleagues [6]. In their study of 363 ICU patients, HES 130/0.4 was administered to 168 patients at the low mean cumulative dose of 763 ml over the first 48 hours [6]. No significant difference in acute kidney injury was detected using the RIFLE criteria. However, Boussekey and colleagues neglected to report their RRT data. Those data should be reported now, so that they may inform the ongoing debate about the renal safety of HES 130/0.4.
 
Neuron specific enolase (NSE) has been proven effective in predicting neurological outcome after cardiac arrest with a current cut off recommendation of 33 microg/l. However, most of the corresponding studies were conducted before the introduction of mild therapeutic hypothermia (MTH). Therefore we conducted a study investigating the association between NSE and neurological outcome in patients treated with MTH. In this prospective observational cohort study the data of patients after cardiac arrest receiving MTH (n = 97) were consecutively collected and compared with a retrospective non-hypothermia (NH) group (n = 133). Serum NSE was measured 72 hours after admission to ICU. Neurological outcome was classified according to the Pittsburgh cerebral performance category (CPC 1 to 5) at ICU discharge. NSE serum levels were significantly lower under MTH compared to NH in univariate analysis. However, in a linear regression model NSE was affected significantly by time to return of spontaneous circulation (ROSC) and ventricular fibrillation rhythm but not by MTH. The model for neurological outcome identified NSE, NSE*MTH (interaction) as well as time to ROSC as significant predictors. Receiver Operating Characteristic (ROC) analysis revealed a higher cutoff value for unfavourable outcome (CPC 3 to 5) with a specificity of 100% in the hypothermia group (78.9 microg/l) compared to the NH group (26.9 microg/l). Recommended cutoff levels for NSE 72 hours after ROSC do not reliably predict poor neurological outcome in cardiac arrest patients treated with MTH. Prospective multicentre trials are required to re-evaluate NSE cutoff values for the prediction of neurological outcome in patients treated with MTH.
 
This is the 20th year of the European Congress. This year there was exceptional diversity and depth of both educational and 'hot-off-the-press' information for both scientists and clinicians. Nurses (often under-rated, but a vital part of the intensive care community) were also accommodated, with a special section of the congress dedicated to them. As we head off into the new century, all those who attended the conference have a good idea of what developments lie ahead in the field of intensive care medicine.
 
Introduction Heparin-induced thrombocytopenia (HIT) is a serious, prothrombotic, immune-mediated adverse reaction triggered by heparin therapy. When HIT is diagnosed or suspected, heparins should be discontinued, and an alternative, fast-acting, parenteral, nonheparin anticoagulation such as argatroban should be initiated. Limited and inconsistent data exist about dosing of argatroban in intensive care unit (ICU) patients with critical illnesses.
 
Patient characteristics grouped according to the highest RIFLE score attained RIFLE classification
Newborns in need of extracorporeal membrane oxygenation (ECMO) support are at high risk of developing acute kidney injury (AKI). AKI may occur as part of multiple organ failure and can be aggravated by exposure to components of the extracorporeal circuit. AKI necessitates adjustment of dosage of renally eliminated drugs and avoidance of nephrotoxic drugs. We aimed to systematically define the incidence and clinical course of AKI in critically ill neonates receiving ECMO support. This study reviewed prospectively collected clinical data (including age, diagnosis, ECMO course, and serum creatinine [SCr]) of all ECMO-treated neonates within our institution spanning a fourteen-year period. AKI was defined using the Risk, Injury, Failure, Loss of renal function, and End-stage renal disease (RIFLE) classification. SCr data were reviewed per ECMO day and compared to age-specific SCr reference values. Accordingly, patients were assigned to RIFLE categories (Risk, Injury, or Failure as 150%, 200% or 300% of median SCr reference values). Data are presented as median and interquartile range (IQR) or number and percentage (%). 242 patients were included of whom 179 survived (74%). Median age at start ECMO was 39 hours [IQR 26 - 63]; median ECMO duration was 5.8 days [IQR 3.9 - 9.4]. 153 patients (64%) had evidence of AKI, with 72 (30%) qualifying as Risk, 55 (23%) as Injury and 26 (11%) as Failure. At the end of the study period only 71 patients (46%) out of all 153 AKI patients improved at least one RIFLE category. Using regression analysis it was found that nitric oxide ventilation (p=0.04) and younger age at start ECMO (p=0.004) were significant predictors of AKI. Survival until intensive care unit discharge was significantly lower for patients in the Failure category (35%) as compared to the Non-AKI (78%), Risk (82%) and Injury category (76%), with all p<0.001, while there were no significant differences between the three latter RIFLE categories. Two thirds of neonates on ECMO suffered from AKI, with a significantly increased mortality risk for patients in the Failure category. As AKI during childhood may predispose for chronic kidney disease in adulthood, long-term monitoring of kidney function after ECMO is warranted.
 
Top-cited authors
Jean-Louis Vincent
  • Université Libre de Bruxelles
Fabio Silvio Taccone
  • Université Libre de Bruxelles
Donat Rudolf Spahn
  • University of Zurich
Louis Riddez
  • Karolinska Institutet
Radko Komadina
  • General and Teaching Hospital Celje