Secondary Logo

Journal Logo

Critical Care and Trauma: Research Reports

Inotropic Support During Experimental Endotoxemic Shock: Part I. The Effects of Levosimendan on Splanchnic Perfusion

Section Editor(s): Takala, JukkaCunha-Goncalves, Doris MD*; Perez-de-Sa, Valeria MD, PhD; Grins, Edgars MD*; Dahm, Peter L. MD, PhD; Thörne, Johan MD, PhD§; Blomquist, Sten MD, PhD*

Author Information
doi: 10.1213/ane.0b013e3181af3fe3

Sepsis-induced disturbances in myocardial function and splanchnic perfusion are a major determinant of multiple organ dysfunction and death in the intensive care unit.1 Early aggressive supportive treatment by means of fluid resuscitation in combination with inotropes and vasopressors can limit the damage caused by sepsis-induced tissue hypoperfusion.2

The pathophysiology of septic cardiac dysfunction is complex, involving β-adrenergic hyporesponsiveness3 and possibly calcium desensitization.4,5 Currently, recommended inotropes and vasopressors for the treatment of septic shock are mainly β- and α-adrenergic agonists. Whereas β-adrenergic effects are mediated by the 3′,5′-cyclic adenosine monophosphate-pathway and have the drawback of increasing intracellular calcium and myocardial oxygen consumption, α-agonists may negatively affect regional blood flow distribution owing to the heterogeneous distribution of splanchnic adrenergic receptors.6

Recently, the calcium sensitizer levosimendan has been tested with the purpose of improving cardiac performance and organ blood flow during experimental and human septic shock.7–10 Unlike α- and β-adrenergic agonists, levosimendan improves myocardial contractility by means of stabilizing the Ca2+-troponin-C complexes without affecting the calcium transients.11 It also causes marked vasodilatation by opening ATP-sensitive potassium channels,12 an effect that was found to be associated with a significantly higher incidence of hypotension in the REVIVE-II study.13 A major concern during early sepsis is that hypotension might lead to myocardial ischemia, especially if the drug is administered before judicious fluid resuscitation has been achieved.14 Even so, pigs with moderate intravascular volume replacement that received levosimendan preemptively were found to show improvements in cardiac output (CO), and in systemic and gut oxygen delivery during the later phase of endotoxemic shock, showing no adverse effects of the drug.7

Although expeditious treatment with reversal of hypoperfusion early during sepsis improves outcome,2 preemptive treatment is hardly feasible. We tested the hypothesis that levosimendan administered after the establishment of severe endotoxemia in animals with moderate intravascular volume substitution would improve systemic and regional blood flow and maintain oxygen homeostasis.


The Local Ethics Committee for Animal Research at Lund University approved the study, and the animals received care in accordance with the Swedish National Guidelines for Care and Use of Animals in Research. Sixteen pigs (31.4 ± 3.4 kg) were fasted overnight with free access to water. After premedication with azaperon (6–8 mg/kg) and atropin (0.03 mg/kg), anesthesia was induced with IV thiopental (10 mg/kg) and maintained with fentanyl (35μg · kg−1 · h−1) and midazolam (0.35 mg · kg−1 · h−1). Vecuronium (0.3 mg · kg−1 · h−1) was used to facilitate abdominal surgery and to avoid shivering during sepsis. A heating mattress was placed on the operating table to avoid temperature loss during surgery. Body temperature was maintained at 36.2°C–38.4°C. Adequate depth of anesthesia was verified in four pigs by means of auditory-evoked potentials (A-line AEP Monitor; Alaris Medical Systems, Danmeter, DK).15 The anesthetic regimen kept the derived A-line Arx Index under 30. The animals were orally intubated, and their lungs mechanically ventilated (ventilator Siemens-Elema 900B™, Solna, Sweden) with a tidal volume of 10 mL · kg−1 · min−1 and an inspired oxygen fraction of 50% in air. Respiratory rate was adjusted to achieve an arterial carbon dioxide partial pressure of 34–41 mm Hg (4.5–5.5 kPa). Arterial and central venous accesses were obtained by surgical insertion of polyurethane catheters into the carotid artery and internal jugular vein. A 6 Fr hemostatic sheath was inserted into the external jugular vein, and a 5.2 Fr multipurpose catheter (MP A-3; SciMed Life Systems, Maple Grove, MN) was guided into a branch of the hepatic vein under fluoroscopy. The cranial caval vein was punctured above the manubrium and a pulmonary artery catheter was inserted through an 8.5 Fr hemostatic sheath. The positions of all catheters were confirmed by fluoroscopy using contrast dye when needed. A median laparotomy was performed, and a short segment of the portal vein, the superior mesenteric artery (SMA), and the hepatic artery distal to the gastro-duodenal artery were dissected free. Precalibrated perivascular ultrasonic transit-time flowprobes (Medi-Stim AS, Oslo, Norway) were placed around these vessels and connected to a flowmeter (CM 4000; Medi-Stim AS). A mesenteric vein was punctured and a 4 Fr catheter was passed into the portal vein for pressure monitoring and blood sampling. A cystostomy catheter was used for urine drainage. The abdominal incision was approximated with a few single sutures and the animals covered with a plastic sheet to avoid heat loss.

Hemodynamic Measurements

Heart rate and vascular pressures (mean arterial blood pressure (MAP), mean pulmonary arterial pressure (MPAP), central venous and portal vein pressures, respectively) were continuously monitored with a Powerlab® monitor (ADInstruments, Hastings, East Sussex, UK) and the data stored in a personal computer. CO was measured with an Oximetrix CO monitor (Abbot Critical Care Systems, Chicago, IL) every 30 min. The mean value of three injections of 5% glucose at room temperature with <10% variation between values was used. Hepatic artery flow (HAF), portal vein flow (PVF), and SMA flow (SMAF) were displayed continuously and the values recorded every 30 min, together with rectal temperature and pulmonary artery occlusion pressure (PAOP). Derived hemodynamic variables were calculated as follows: systemic vascular resistance (SVR) = (MAP − central venous pressure) × CO−1 × 80; pulmonary vascular resistance = (MPAP − PAOP) × CO−1 × 80; gut vascular resistance = (MAP − portal vein pressure) × PVF−1 × 80. Regional flow ratios were calculated as percentage of CO and the transpulmonary gradient as MPAP − PAOP.

Blood Gas, Lactate Concentration, and Hematological Analysis

Arterial, portal vein, and hepatic vein lactate concentrations were measured using an enzymatic colorimetric method (LOX-PAP, Miniphotometer plus LP20; Dr. Lange GmbH, Berlin, Germany) every 60 min. Results are given in mmol/L. Arterial, portal, hepatic vein, and mixed venous blood gases were analyzed every 60 min on an ABL 505 analyzer (Radiometer Medical A/S, Copenhagen, Denmark) at 37°C. Hemoglobin concentration and oxygen saturations were measured with an OSM3 hemoxymeter (Radiometer Medical A/S) with factory-installed correction for the swine hemoglobin absorption characteristics. Blood oxygen content was calculated using standard formulae. Systemic oxygen economy was calculated as: oxygen consumption = (CaO2 − CO2) × CO; oxygen delivery = CO × CaO2; and oxygen extraction ratio = (CaO2 − CO2) × CaO2−1 × 100, where CaO2 and CO2 are arterial and mixed venous oxygen contents. Gut and hepatic oxygen balances were calculated in the same manner, i.e., based on the Fick principle, from the respective flows and oxygen contents. Leukocyte and platelet counts were analyzed by standard laboratory methods. Gut lactate production was calculated as ([arterial lactate-portal vein lactate] × PVF) and hepatic lactate uptake as ([portal vein lactate × PVF] + [arterial lactate × HAF]) − ([HAF+PVF] × hepatic vein lactate), where decreasing values denote a decrease in hepatic extraction.

Preparation of Levosimendan

Immediately before each experiment and according to the manufacturer’s instructions, levosimendan dry powder was dissolved in alcohol and diluted in a bicarbonate buffer vehicle solution to a concentration of 133 μg/mL, as described previously.14

Experimental Protocol

To compensate for fluid losses during surgery and for capillary leakage because of endotoxin infusion, the animals received intravascular fluid administration with 30 mL/kg of dextran 70 (Macrodex® 60 mg/mL in NaCl; Pharmalink, Väsby, Sweden) before baseline measurements. A continuous infusion of Ringer’s acetate solution and 5% glucose (10 mL · kg−1 · h−1) was begun at the start of anesthesia and continued throughout the experiment. The PAOP was kept between 6 and 9 mm Hg. After surgery and stabilization, the animals were randomized to either control or levosimendan groups. Baseline data (t = 0 min) were collected 1 h after surgery. In both groups, an infusion of endotoxin (Lipopolysaccharide B E. coli, 0111: B4; Difco Laboratories, Detroit, MI) was started immediately after baseline measurements and continued until the end of the experiment (300 min). At 100 min, either an infusion of levosimendan (50 μg · kg−1 · h−1) or, in the control group, an infusion of the equivalent amount of vehicle solution was started and kept unchanged for the remaining 200 min.

Statistical Analysis

Statistical analysis was performed using SigmaStat software version 3.5 (Systat Software GmbH, Erkrath, Germany). To evaluate the isolated effects of endotoxemia, all data were pooled into one group (0–90 min for hemodynamics and regional blood flows; 0–60 min for lactate concentrations, and oxygenation indices). Data were analyzed by repeated-measures analysis of variance with Bonferroni post hoc test where indicated (stage versus baseline) except for oxygenation/lactate (paired t-test or signed rank test). Within-group (levosimendan, control) treatment effects (90–300 min, for hemodynamics and regional blood flows, stage versus t = 90 min; 60–300 min, for lactate concentrations and oxygenation indices, stage versus t = 60 min) were analyzed in the same way. Differences between groups for hemodynamic variables at t = 90 min and t = 300 min, and also for oxygenation indices and lactate concentrations at t = 60 min and t = 300 min, were analyzed using the t-test—or the Mann–Whitney rank sum test when normality test failed (Kolmogorov-Smirnov test). P < 0.05 was considered significant. Data are presented as mean ± sem.


Two animals in each group died before the end of the study. Leukocyte sequestration and platelet aggregation, characteristic features of porcine endotoxemia, were indirectly confirmed by a marked decrease in the number of circulating leukocytes and platelets in peripheral blood samples of all animals (not shown).

Effects of Endotoxin Infusion (0–90 min)

Endotoxin infusion induced tachycardia and reduced stroke volume, MAP, and SVR (Tables 1 and 2, Figs. 1A–D, 2A–D, and 3A–F). Although CO decreased at 30 min, it returned to baseline levels as heart rate increased. PAOP decreased (P = 0.007). MPAP and pulmonary vascular resistance increased markedly. Gut vascular resistance increased temporarily and blood flow was diverted away from the splanchnic vascular bed, as evidenced by reduced SMAF and PVF ratios. However, there was a transient compensatory increase in HAF and flow ratio at 90 min. Arterial hemoglobin concentration increased but systemic oxygen delivery remained unchanged, whereas systemic oxygen consumption and extraction ratio (32 ± 2 to 39 ± 2%, P = 0.021) increased. Both gut and hepatic oxygen deliveries decreased. Gut oxygen consumption was unchanged as the extraction ratio increased by 50% (P = 0.002). In contrast, hepatic oxygen extraction ratio remained unchanged (P = 0.134) and hepatic oxygen consumption decreased. Mixed venous and portal vein, but not hepatic vein, saturations decreased. Arterial, portal vein, and hepatic vein lactate increased but gut lactate production and hepatic lactate uptake remained unchanged. Arterial pH and base excess decreased.

Table 1
Table 1:
Systemic and Regional Hemodynamics
Table 2
Table 2:
Oxygenation and Lactate
Figure 1
Figure 1:
Figure 1.
Figure 2
Figure 2:
Figure 2.
Figure 3
Figure 3:
Figure 3.

Effects of Levosimendan/Vehicle Treatment (90–300 min)

There were no statistically significant differences between groups before the start of treatment (t = 90 min) (Tables 1 and 2, Figs. 1A–D, 2A–D, and 3A–F). During treatment, heart rate increased, whereas stroke volume and CO decreased equally in both groups. Although SVR gradually increased in the control group, it remained unchanged in the levosimendan group, and the difference between the groups at 300 min nearly reached statistical significance (P = 0.052). Consequently, MAP decreased only in the latter. PAOP remained unchanged in both groups. Although pulmonary vascular resistance increased similarly in both groups, the transpulmonary gradient increased only in the control group (P = 0.002). MPAP increased and was higher than in the levosimendan group at 300 min. Gut vascular resistance was unchanged but tended to be lower in the levosimendan group at 300 min (P = 0.065). With the exception of SMAF, which only decreased in the levosimendan group, all other regional blood flows decreased in both groups. SMAF and PVF ratios increased in the control group, whereas HAF ratio decreased in both groups.

There were no statistically significant differences in lactate concentrations or oxygen variables between the two groups, either before (t = 60 min) or at the end of the treatment period (t = 300 min). In both groups, hemoglobin levels progressively increased. Thus, systemic oxygen delivery remained unchanged, with the exception of a short-lived increase in the levosimendan group at 120 min, and also a transient decrease in the control group at 240 min. In association with a transitory increase in systemic oxygen consumption in the levosimendan group and a tendency toward increase in the control group, systemic oxygen extraction ratio increased by 50%–60% (P < 0.001) in both groups. At 240 min, hepatic oxygen delivery decreased in the control group, whereas gut oxygen delivery decreased in both groups at t = 300 min. Although gut and hepatic oxygen consumptions remained unchanged, hepatic oxygen extraction ratio doubled in both groups (P = 0.002), and gut extraction ratio increased by 50% in the levosimendan group (P = 0.048). Venous saturations decreased in all vascular beds. The arterial lactate concentration increased transiently in the levosimendan group, and gut lactate production and hepatic lactate uptake remained unchanged in both groups.


In pigs receiving moderate intravascular volume resuscitation, levosimendan administered after the establishment of endotoxemic shock, maintained a low SVR, and prevented further increases in MPAP. However, contrary to our hypothesis, it did not maintain CO and the animals developed hypotension. Likewise, despite a trend toward lower gut vascular resistance, gut blood flow and oxygen supply ultimately decreased.

Our observation of reduced splanchnic blood flow ratio (SMAF and PVF ratios), followed by a transitory increase in HAF ratio during the first 90 min of endotoxemia, is in accordance with the findings of Tenhunen et al.16 (redistribution of blood flow within the splanchnic circulation early during endotoxin infusion) and with a previous report on exhaustion of the hepatic arterial buffer response during systemic hypoperfusion.17 Interestingly, SMAF and PVF ratios recovered to some extent in the control group but not in the levosimendan group. The persistence of reduced fractional blood flows in this group could partly be explained by the fact that during periods of reduced systemic perfusion, vasodilators can induce a vascular steal phenomenon, driving blood away from the splanchnic circulation.18

Contrary to our findings, levosimendan (100–200 μg · kg−1 · h−1) given either before or together with endotoxin infusion fairly maintained MAP, and minimized the decrease in CO and splanchnic blood flow during shock.7,8 Somewhat different perfusion pressures might partially account for the discrepancy between the studies. Besides, more aggressive intravascular volume resuscitation in our study might have improved CO and MAP, as well as systemic and splanchnic blood flows.19 Nevertheless, intravascular volume replacement during experimental sepsis aims at maintaining the PAOP at 5–8 mm Hg,16,20 but despite intravascular fluid administration with dextran 70 and crystalloid solutions enough to maintain the PAOP at the targeted value, our animals never became hyperdynamic. Interestingly, the control groups in both experimental studies cited above developed hypodynamic shock of comparable severity to that observed in our control group; thus, volume status alone cannot explain the conflicting results. Could the preemptive use of levosimendan by itself have improved hemodynamics in severe endotoxemic shock? At least for the heart, the opening of mitochondrial KATP channels protects against reperfusion injury21; thus, it cannot be excluded that pretreatment with levosimendan may have a role in preventing microvascular dysfunction during endotoxemia.

Another important issue that might partly explain the contradictory findings in different experimental studies is levosimendan dosage. At high plasma concentrations, levosimendan causes phosphodiesterase inhibition that leads to an increase in 3′,5′-cyclic adenosine monophosphate.22 Our aim was to study the drug at clinically relevant plasma concentrations. In a dose-response study in healthy pigs, levosimendan at a dose of 80 μg · kg−1 · h−1 for 10 min, followed by 40 μg · kg−1 · h−1 for 30 min, increased CO maximally and resulted in a mean plasma concentration of 67 ng/mL.23 In our previous study14 with endotoxemic pigs, 50 μg · kg−1 · h−1 levosimendan resulted in a plasma concentration of 80.9 ± 10.6 ng/mL.

Despite the decrease in CO after 90 min, systemic oxygen delivery was maintained in both groups at 300 min because of an increase in hemoglobin concentration, which was probably secondary to capillary leakage and to autotransfusion of red blood cells originating from splenic contraction.24 The association of reduced regional oxygen delivery, combined with persistent low hepatic oxygen consumption, along with the fact that the animals never developed a hyperdynamic circulation, might indicate suboptimal volume resuscitation. In both groups, although lactate production was limited and the animals could still maintain high systemic oxygen consumption by increasing the extraction ratio, the portal, hepatic vein, and mixed venous oxygen saturations became markedly reduced, signaling impending oxygen deficit.

MPAP was lower in the levosimendan group. This is in agreement with previous findings in moderately resuscitated endotoxemic pigs,7 but differs from our previous results with relatively hypovolemic pigs in endotoxemic shock.14 Although pulmonary vascular resistance increased in both groups, the transpulmonary pressure gradient was unchanged in the levosimendan group. Slightly lower filling pressures after systemic vasodilatation could account for the lower MPAP in this group, even though the difference in filling pressures between the two groups did not reach statistical significance.

There are some obvious limitations to our study. First, although the porcine cardiovascular physiology is similar to that of humans25 and the porcine endotoxin model to some extent reproduces the inflammatory response during human sepsis,26 it is limited by the lack of a source of infection. Second, inodilators should not be used in incompletely resuscitated subjects; therefore, because end points of fluid resuscitation in sepsis remain unclear27 and there are indications that fluid replacement was suboptimal, our study mainly emphasizes the risks of administering inodilators early in sepsis, i.e., before adequate preload is achieved. Third, we used a short-term model of endotoxemic shock that was not designed for examination of long-term outcomes.

In conclusion, levosimendan given after the establishment of endotoxemic shock to pigs receiving moderate intravascular volume administration prevented further increases in MPAP and maintained gut and SVRs. However, perfusion pressure and all venous oxygen saturations decreased and, contrary to previous studies on the preemptive use of levosimendan, neither systemic nor regional perfusion improved.


The authors thank Orion Pharma, Espoo, Finland, for the supply of levosimendan.


1. Beal AL, Cerra FB. Multiple organ failure syndrome in the 1990s. Systemic inflammatory response and organ dysfunction. JAMA 1994;271:226–33
2. Rivers E, Nguyen B, Havstad S, Ressler J, Muzzin A, Knoblich B, Peterson E, Tomlanovich M. Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med 2001;345:1368–77
3. Silverman HJ, Penaranda R, Orens JB, Lee NH. Impaired beta-adrenergic receptor stimulation of cyclic adenosine monophosphate in human septic shock: association with myocardial hyporesponsiveness to catecholamines. Crit Care Med 1993;21: 31–9
4. Tavernier B, Garrigue D, Boulle C, Vallet B, Adnet P. Myofilament calcium sensitivity is decreased in skinned cardiac fibres of endotoxin-treated rabbits. Cardiovasc Res 1998;38:472–9
5. Yasuda S, Lew WY. Lipopolysaccharide depresses cardiac contractility and beta-adrenergic contractile response by decreasing myofilament response to Ca2+ in cardiac myocytes. Circ Res 1997;81:1011–20
6. Gelman S, Mushlin PS. Catecholamine-induced changes in the splanchnic circulation affecting systemic hemodynamics. Anesthesiology 2004;100:434–9
7. Oldner A, Konrad D, Weitzberg E, Rudehill A, Rossi P, Wanecek M. Effects of levosimendan, a novel inotropic calcium-sensitizing drug, in experimental septic shock. Crit Care Med 2001;29:2185–93
8. Dubin A, Murias G, Sottile JP, Pozo MO, Baran M, Edul VS, Canales HS, Etcheverry G, Maskin B, Estenssoro E. Effects of levosimendan and dobutamine in experimental acute endotoxemia: a preliminary controlled study. Intensive Care Med 2007;33:485–94
9. Matejovic M, Krouzecky A, Radej J, Novak I. Successful reversal of resistent hypodynamic septic shock with levosimendan. Acta Anaesthesiol Scand 2005;49:127–8
10. Morelli A, De Castro S, Teboul JL, Singer M, Rocco M, Conti G, De Luca L, Di Angelantonio E, Orecchioni A, Pandian NG, Pietropaoli P. Effects of levosimendan on systemic and regional hemodynamics in septic myocardial depression. Intensive Care Med 2005;31:638–44
11. Haikala H, Kaivola J, Nissinen E, Wall P, Levijoki J, Linden IB. Cardiac troponin C as a target protein for a novel calcium sensitizing drug, levosimendan. J Mol Cell Cardiol 1995;27: 1859–66
12. Yokoshiki H, Katsube Y, Sunagawa M, Sperelakis N. Levosimendan, a novel Ca2+ sensitizer, activates the glibenclamide-sensitive K+ channel in rat arterial myocytes. Eur J Pharmacol 1997;333:249–59
13. Cleland JG, Freemantle N, Coletta AP, Clark AL. Clinical trials update from the American Heart Association: REPAIR-AMI, ASTAMI, JELIS, MEGA, REVIVE-II, SURVIVE, and PROACTIVE. Eur J Heart Fail 2006;8:105–10
14. Cunha-Goncalves D, Perez-de-Sa V, Dahm P, Grins E, Thorne J, Blomquist S. Cardiovascular effects of levosimendan in the early stages of endotoxemia. Shock 2007;28:71–7
15. Martoft L, Jensen EW, Rodriguez BE, Jorgensen PF, Forslid A, Pedersen HD. Middle-latency auditory evoked potentials during induction of thiopentone anaesthesia in pigs. Lab Anim 2001;35:353–63
16. Tenhunen JJ, Uusaro A, Karja V, Oksala N, Jakob SM, Ruokonen E. Apparent heterogeneity of regional blood flow and metabolic changes within splanchnic tissues during experimental endotoxin shock. Anesth Analg 2003;97:555–63
17. Jakob SM, Tenhunen JJ, Laitinen S, Heino A, Alhava E, Takala J. Effects of systemic arterial hypoperfusion on splanchnic hemodynamics and hepatic arterial buffer response in pigs. Am J Physiol Gastrointest Liver Physiol 2001;280:G819–27
18. Shepherd AP, Riedel GL, Maxwell LC, Kiel JW. Selective vasodilators redistribute intestinal blood flow and depress oxygen uptake. Am J Physiol 1984;247:G377–84
19. Ali SZ, Bracht H, Krejci V, Beck M, Stalder M, Hiltebrand L, Takala J, Brandt S, Jakob SM. The immediate and sustained effects of volume challenge on regional blood flows in pigs. Anesth Analg 2008;106:595–600
20. Hiltebrand LB, Krejci V, Jakob SM, Takala J, Sigurdsson GH. Effects of vasopressin on microcirculatory blood flow in the gastrointestinal tract in anesthetized pigs in septic shock. Anesthesiology 2007;106:1156–67
21. Inoue I, Nagase H, Kishi K, Higuti T. ATP-sensitive K+ channel in the mitochondrial inner membrane. Nature 1991;352:244–7
22. Edes I, Kiss E, Kitada Y, Powers FM, Papp JG, Kranias EG, Solaro RJ. Effects of Levosimendan, a cardiotonic agent targeted to troponin C, on cardiac function and on phosphorylation and Ca2+ sensitivity of cardiac myofibrils and sarcoplasmic reticulum in guinea pig heart. Circ Res 1995;77:107–13
23. Leather HA, Ver Eycken K, Segers P, Herijgers P, Vandermeersch E, Wouters PF. Effects of levosimendan on right ventricular function and ventriculovascular coupling in open chest pigs. Crit Care Med 2003;31:2339–43
24. Takasaki K, Tang LC, Urabe M. The beta-adrenergic responses in isolated splenic capsule area of several species of animals. Jpn J Pharmacol 1979;29:1–7
25. Swindle MM, Smith AC, Hepburn BJ. Swine as models in experimental surgery. J Invest Surg 1988;1:65–79
26. Brix-Christensen V, Andersen SK, Andersen R, Mengel A, Dyhr T, Andersen NT, Larsson A, Schmitz O, Orskov H, Tonnesen E. Acute hyperinsulinemia restrains endotoxin-induced systemic inflammatory response: an experimental study in a porcine model. Anesthesiology 2004;100:861–70
27. Vincent JL, Gerlach H. Fluid resuscitation in severe sepsis and septic shock: an evidence-based review. Crit Care Med 2004; 32:S451–4
© 2009 International Anesthesia Research Society