Effect of mild hypothermia on the vascular actions of phenylephrine in rat aortic rings.

Intraoperative mild hypothermia is common. We have investigated the effects of mild hypothermia (34 vs 38 degrees C) on phenylephrine--(10(-8) to 10(-5) mol litre-1) induced contractions of rat aortic rings mounted for isometric tension recordings. A marked decrease in Emax (maximal tension) (P < 0.05) and significant increase in EC50 (phenylephrine concentration producing 50% of maximal tension) were observed at the lower temperature in endothelium intact rings, but there was no effect of temperature when the endothelium had been removed. The decreased contraction with hypothermia in the endothelium intact vessels was restored to 84% by administration of the nitric oxide synthase inhibitor L-NNA and a small additional amount of tone was restored in the presence of the cyclooxygenase inhibitor, indomethacin. We conclude that mild hypothermia markedly decreased phenylephrine-induced rat aortic contraction in vitro by endothelium dependent mechanisms, largely related to increased nitric oxide production or action.

Paradoxical air embolism during orthoptic liver transplantation: diagnosis by transoesophageal echocardiography.

We describe a case of paradoxical air embolism during orthotopic liver transplantation, early diagnosis, using intra-operative transoesophageal echocardiography after a circulatory failure, allowed early management by hyperbaric oxygen therapy.

Vascular properties of isoflurane: comparison between normal and cirrhotic rats.

Isoflurane is known to dilate blood vessels and to modulate nitric oxide production. Because cirrhosis is characterized by over production of endothelial nitric oxide, isoflurane-induced vasodilatation may be altered in this situation. We have compared the vasodilator effects of isoflurane in normal rats and rats with secondary biliary cirrhosis. Aortic rings (intact or endothelium denuded) from normal and cirrhotic rats were suspended in HEPES solution and preconstricted with KCl 40 mmol litre-1. Isoflurane dose-dependently relaxed vessels in both groups. Maximal relaxation was comparable between normal and cirrhotic rats in intact (mean 80 (SEM 4) vs 81 (6)%; ns) and in denuded (100 (4) vs 95 (5)%; ns) vessels. Intact vessels relaxed more than denuded vessels in both groups (100 (4) vs 80 (4)% (P = 0.0008) in normal rats and 95 (5) vs 80 (6)% (P = 0.0008) in cirrhotic rats). We conclude that cirrhosis did not modify isoflurane-induced vasodilatation and that the modulator effect of endothelium was conserved.

Influence of acute renal failure on FPIA rapid serum assay of midazolam and its main metabolite.

We recently developed a simple and fast assay technique, providing the possibility of monitoring of midazolam (M) during sedation. We compared HPLC vs FPIA for the measurement of the sum M plus alpha 1-hydroxymidazolam (OM), its main and pharmacologically active metabolite, in the serum of sedated ICU patients; this activity referred to as M-like. We identified certain patients in whom M-like activity appeared abnormally high in comparison with HPLC assays. Their common denominators were: long-term sedation with M, and seriously impaired renal function. Further, the conjugates of OM (OMG) accumulated in patients with acute renal failure could contribute to the sedation. We compared the metabolic and analytic behavior of M, OM, and OMG in 2 groups of sedated patients either presenting with normal renal functions (group 1) or with a picture of acute renal failure (group 2). Blood samples were assayed by HPLC and by FPIA and analysis was performed before and after hydrolysis of OMG. Before hydrolysis there was a dramatic accumulation of OMG in the patients of group 2, HPLC vs FPIA results were not different within group 1, while in group 2 the FPIA response exceeded that of HPLC. After hydrolysis, measurement by HPLC was greatly increased in group 2, in each group (vs HPLC) and from one group to another, the FPIA signal (the M-like activity) showed a significant increase. It would be important to take OMG into account as a coprotagonist in sedation whenever circumstances predispose to its accumulation.

Role of endothelium in the modulation of isoflurane-induced vasodilatation in rat thoracic aorta.

The mechanisms by which endothelium attenuates vasodilation caused by isoflurane are not well understood. We examined the role of endothelium-derived substances, nitric oxide (NO), endothelium-derived hyperpolarizing factors (EDHF), prostanoids and endothelins in the response to isoflurane in rat thoracic aorta. Increasing cumulative concentrations of isoflurane were administered to aortic rings suspended in Hepes solution and preconstricted with either phenylephrine 10(-6) mol litre-1 or KCl 40 mmol litre-1 (which inhibit EDHF). Rings were intact, denuded or incubated with an inhibitor of nitric oxide synthesis (N omega-nitro-L-arginine (LNNA 5 x 10(-5) mol litre-1), an inhibitor of prostanoid synthesis (indomethacin 10(-5) mol litre-1) or a blocker of the vascular receptors to endothelins (cyclo (-D-trp-D-Asp-Pro-D-Val-Leu (BQ 123 10(-5) mol litre-1)- Endothelium attenuated isoflurane-induced vasodilation in KCl-constricted rings at concentrations of 4% (mean 95 (SEM 4)% vs 72 (4)%; P = 0.0005) and 5% (100 (4)% vs 80 (4)%; P = 0.0008) and in phenylephrine constricted rings at concentrations of 4% (54 (8)% vs 35 (3)%; P = 0.04) and 5% (78 (10)% vs 49 (5)%; P = 0.03). Relaxation was significantly greater in rings treated with LNNA than in intact rings at concentrations of 4% (85 (4)% vs 72 (4)%; P = 0.0005) and 5% (90 (4)% vs 80 (4)%; P = 0.0008). Indomethacin and BQ 123 did not alter isoflurane-induced vasodilation. We conclude that endothelium attenuated the vasodilator effect of isoflurane by a mechanism which was abolished by inhibition of nitric oxide. We hypothesize that isoflurane inhibits the release of nitric oxide, leading to a relative vasoconstriction counter-balancing its vasodilator effect. In contrast, EDHF, prostanoids and endothelins were not involved in the attenuation of isoflurane-induced vasodilatation.

Plasma concentrations of cyclic 3', 5'-guanosine monophosphate in patients with cirrhosis: relationship with atrial natriuretic peptide and haemodynamics.

Little is known about the plasma concentrations of cyclic 3',5'-guanosine monophosphate (cGMP) in patients with cirrhosis. However, plasma cGMP concentrations provide information on cellular cGMP production by particulate guanylyl cyclases (which are stimulated by natriuretic peptides, such as atrial natriuretic peptide; ANP). In contrast, because intracellular cGMP elicits vasorelaxant mechanisms, plasma cGMP concentrations may be related to haemodynamic alterations in patients with cirrhosis. The aim of the present study was to measure plasma cGMP concentrations in patients with cirrhosis and controls and to examine the relationship between cGMP levels and plasma ANP concentrations and haemodynamic values. Plasma concentrations of cGMP and ANP and splanchnic and systemic haemodynamics were measured in 23 subjects; 13 subjects had cirrhosis and 10 were controls. All subjects had normal glomerular filtration. Plasma cGMP concentrations were significantly higher in patients (6.5 +/- 0.8 pmol/mL) than in controls (2.7 +/- 0.4 pmol/mL), while plasma ANP concentrations did not significantly differ between the two groups (127 +/- 22 and 123 +/- 27 pg/mL, respectively). In patients with cirrhosis, no significant correlation was found between plasma cGMP concentrations and plasma ANP concentrations, hepatic venous pressure gradient, cardiac output or systemic vascular resistance. In conclusion, in patients with cirrhosis, increased plasma cGMP concentrations may be due to an activation of particulate guanylyl cyclases by natriuretic peptides other than ANP. The present study suggest that plasma cGMP concentrations are not related to cirrhosis-induced haemodynamic alterations.

Comparison of high-performance liquid chromatography and polyclonal fluorescence polarization immunoassay for the monitoring of midazolam in the plasma of intensive care unit patients.

Midazolam (M) is used as an induction agent for anesthesia. The main metabolite is alpha-hydroxymidazolam (OM), which is pharmacologically active. Use of M for sedation is a recent application, rapidly gaining favor. Monitoring of the level of sedation is fundamental in that an excessive and prolonged effect is associated with the risk of complications. Thus, it was felt both necessary and useful to measure circulating M levels. We compared a high-performance liquid chromatography (HPLC) assay with fluorescence polarization immunoassay (FPIA) for the measurement of M in the serum of 138 sedated patients in the intensive care unit (i.e., 179 samples). Response of the OM was also assessed. The degree of crossover of the metabolite was between 76.8 and 32.7%. The equation of the regression line for sigma HPLC (i.e., the sum M + OM) versus FPIA was TDx = 1.1585 sigma HPLC + 143.42 (R = 0.966). The 95% confidence interval for the slope was 1.1551, 1.1619. The regression slope differed significantly from 1 (p < 0.001) and shows that FPIA measurements overestimated concentrations obtained by HPLC on the order of 19%. The discrepancy between the two techniques was all the more notable when concentrations were > 1,000 ng/ml. The relative selectivity of Abbott industrial reagent in terms of benzodiazepines leads to the identification of what might be called a midazolam-like (M-like) activity covering both M and OM. The development of a global FPIA method for measurement of this M-like activity in sedated patients provides a satisfactory solution to the question raised.

Haemodynamic and hormonal responses to long-term inhibition of nitric oxide synthesis in rats with portal hypertension.

In portal hypertension, the role of the vasorelaxant nitric oxide (NO) in long-term splanchnic and systemic vascular tone regulation is unclear. This study examined the effects of long-term administration of a NO synthesis inhibitor on haemodynamics in portal hypertensive rats. Rats were randomly assigned to receive either water (placebo) or 100 mg/kg.day of oral N-nitro-L-arginine methylester (L-NAME) for 28 days. At 14 days, the portal vein was ligated in 10 rats from each group. At 28 days, splanchnic and systemic blood flows were measured in 20 normal and 20 portal vein stenosed rats. Plasma atrial natriuretic peptide (ANP) concentrations as well as plasma and urinary cyclic guanosine monophosphate (cGMP) levels were also measured. Porto-systemic shunts were measured in other portal vein stenosed animals that had or had not received L-NAME. Portal vein stenosed rats that received L-NAME had significantly lower portal tributary blood flow and percentages of portal-systemic shunting (7.3 +/- 0.5 versus 3.7 +/- 0.2 ml/min.100 g and 96 +/- 1 versus 68 +/- 5%, respectively) and higher hepatocollateral vascular resistance (147 +/- 10 versus 295 +/- 30 dyn.s.cm-5.100 g.10(3), respectively) than placebo portal vein stenosed rats. Portal pressure, ANP and cGMP levels did not differ between the groups. Arterial pressure was significantly higher and cardiac index lower after L-NAME than after placebo. Normal rats had similar but less marked L-NAME-induced responses than portal hypertensive rats. The presence of a long-term L-NAME-induced vasoconstriction in collateral vessels and splanchnic and systemic arterioles in portal vein stenosed rats indicates that a NO-mediated vasodilator tone may contribute to the development and the maintenance of collateral circulation as well as splanchnic and systemic vasodilation in portal hypertension. Moreover, the NO-mediated vasodilator tone in portal hypertensive animals seems to be increased.

Role of sympathetic cardiovascular tone in control of arterial pressure in rats with cirrhosis.

Although an increase in sympathetic nervous activity has been recognized in cirrhosis, the contribution of this overactivity to the regulation of arterial pressure is unknown. The arterial pressure response to increasing doses of hexamethonium (0.05 to 3.2 mg.kg-1.min-1), a ganglionic blocker that decreases sympathetic cardiovascular tone, was explored in normal rats and in two models of portal hypertension, i.e., rats with cirrhosis and rats with portal vein stenosis. Changes in plasma norepinephrine concentrations were greater in rats with cirrhosis (356 +/- 50 vs 166 +/- 30 pg/ml, p = 0.04) than in normal rats (186 +/- 23 vs 86 +/- 31 pg/ml, p = 0.06) and rats with portal vein stenosis (103 +/- 37 vs 93 +/- 5 pg/ml, p = 0.10). The maximum decrease in arterial pressure was obtained at a dose of 1.6 mg.kg-1.min-1 in each group. However, the decrease in arterial pressure was significantly greater in rats with cirrhosis (-25 +/- 2%) than in normal rats (-11 +/- 1%) and in rats with portal vein stenosis (-13 +/- 2%) (p = 0.04). In conclusion, the results of this study suggest that the sympathetic cardiovascular tone is more important for the maintenance of arterial pressure in rats with cirrhosis than in normal rats and in rats with portal vein stenosis.

Acute haemodynamic responses and inhibition of tumour necrosis factor-alpha by pentoxifylline in rats with cirrhosis.

1. Although pentoxifylline has been shown to reduce portal hypertension, the mechanism for this is unclear. Since pentoxifylline decreases tumour necrosis factor-alpha production and since this cytokine may induce vasodilatation per se, a pentoxifylline-induced decrease in tumour necrosis factor-alpha production may limit arterial vasodilatation and decrease portal pressure. The aim of the present study was to examine the effects of pentoxifylline administration on plasma tumour necrosis factor-alpha concentration and haemodynamics in normal and cirrhotic rats. 2. In both groups, systemic and splanchnic haemodynamics and plasma tumour necrosis factor-alpha concentrations were measured before and 120 min after the administration of saline or pentoxifylline (20 mg/kg intravenous bolus). 3. In cirrhotic rats, pentoxifylline significantly decreased portal pressure (24 +/- 13%) and tributary blood flow (33 +/- 30%). On the other hand, pentoxifylline significantly increased vascular resistance in portal and hepatic arterial territories. Systemic haemodynamics were not altered. In normal rats, pentoxifylline significantly decreased portal pressure but induced no other significant changes in splanchnic or systemic haemodynamics. In cirrhotic rats, plasma tumour necrosis factor-alpha concentrations were significantly reduced after pentoxifylline administration but not after saline administration. No significant correlations were found between pentoxifylline-induced changes in tumour necrosis factor-alpha levels and changes in splanchnic haemodynamics. In normal rats, plasma tumour necrosis factor-alpha concentrations significantly decreased after pentoxifylline or saline administration. 4. This study shows that in rats with cirrhosis, pentoxifylline induces a decrease in both portal pressure and plasma tumour necrosis factor-alpha concentrations. These reductions were not correlated however.

Role of pentobarbitone anaesthesia and sympathetic tone in the haemodynamic effects of isosorbide dinitrate in rats with cirrhosis.

The haemodynamic effects of nitrovasodilators and their mechanisms of action on portal hypertension remain unclear. The splanchnic and systemic haemodynamic response to the infusion of isosorbide dinitrate (100 micrograms/kg per min), a nitrovasodilator, was investigated in cirrhotic rats. The role of the conscious state in the haemodynamic response to isosorbide dinitrate was examined using rats that were anaesthetized with pentobarbitone. The role of sympathetic tone in the haemodynamic response to isosorbide dinitrate was examined using rats pretreated with the ganglion blocker hexamethonium. Isosorbide dinitrate had no haemodynamic effects in conscious, unblocked normal and cirrhotic rats. Isosorbide dinitrate had no haemodynamic effects in normal and cirrhotic rats treated with hexamethonium. In normal anaesthetized rats, isosorbide dinitrate significantly decreased systemic vascular resistance (414 +/- 25 vs 290 +/- 26 dyn.s/cm-5 per 100 g). In cirrhotic anaesthetized rats, isosorbide dinitrate significantly decreased mean arterial pressure (98 +/- 6 vs 79 +/- 7 mmHg), systemic vascular resistance (318 +/- 30 vs 207 +/- 10 dyn.s/cm-5 per 100 g), portal pressure (14.0 +/- 1.0 vs 11.3 +/- 0.9 mmHg) and portal territory vascular resistance (1362 +/- 163 vs 1031 +/- 182 dyn.s/cm5 per 100 g). In conclusion, this study shows that the portal hypotensive effects of isosorbide dinitrate depend upon the alterations of vascular tone by pentobarbitone.

Dose-dependent effects of a nitric oxide biosynthesis inhibitor on hyperdynamic circulation in two models of portal hypertension in conscious rats.

The role of nitric oxide (NO) in the hyperkinetic circulation in portal hypertension has not been clearly elucidated. Different doses of NO inhibitors, haemodynamic values and experimental conditions might explain the discrepant results. The aim of the present study was to investigate the acute effects of a specific biosynthesis inhibitor of NO, Nomega-nitro-L-arginine (L-NNA), on the systemic and splanchnic circulation in normal conscious rats and rats with portal hypertension due to either partial portal vein stenosis or secondary biliary cirrhosis. The administration of L-NNA (15 to 960 micrograms.kg-1.min-1) induced a significant dose dependent increase in arterial pressure which was not different among the three groups of rats. Following an acute and maximal vasopressive dose of L-NNA (1 mg.kg-1.min-1) cardiac index decreased more in portal vein stenosed and cirrhotic rats (-45 +/- 3% and -45 +/- 2%, respectively) than in normal rats (-31 +/- 2%), and systemic vascular resistance increased more in the two groups of portal hypertensive rats than in normals (+ 161 +/- 13% and + 154 +/- 10% vs + 85 +/- 6%, respectively). L-NNA caused a greater decrease in portal tributary blood flow in portal vein stenosed and cirrhotic rats (-63 +/- 4% and -55 +/- 4%, respectively) than in normal rats (-45 +/- 6%). Similarly, the increase in portal territory vascular resistance was significantly more marked in portal vein stenosed and cirrhotic rats (+ 337 +/- 62% and + 214 +/- 24%, respectively) than in normal rats (+ 153 +/- 23%). Portal pressure did not change. Following the acute administration of L-NNA, no significant difference in splanchnic and systemic haemodynamics were noted between portal vein stenosed and normal rats, except for portal pressure. In cirrhotic rats, splanchnic and systemic values remained different from normal rats. This study confirms that NO plays a role in the haemodynamic changes in portal hypertension, and shows that NO inhibitors have a dose-dependent effect in conscious portal hypertensive rats.

Effects of glibenclamide on systemic and splanchnic haemodynamics in conscious rats.

1. The effects of the sulphonylurea, glibenclamide (20 mg kg-1, i.v.), at a dose that blocks vascular potassium channels, on systemic and splanchnic haemodynamics (radioactive microspheres) were studied in conscious rats. 2. Glibenclamide significantly decreased cardiac index and hepatic artery blood flow while it significantly increased vascular resistance in systemic, portal and hepatic arterial territories. 3. In rats with suppressed cardiovascular reflexes, glibenclamide induced vasoconstriction in systemic, portal and hepatic arterial territories. 4. Intracerebroventricular administration of glibenclamide did not alter systemic or regional vascular tone. 5. Glibenclamide blunted the vasodilator effect of the potassium channel opener, diazoxide but not that of the L-type calcium channel blocker, nicardipine. 6. Another sulphonylurea, glipizide (20 mg kg-1, i.v.), induced significant systemic and splanchnic vasoconstriction. 7. Thus, the glibenclamide-induced blockade of vascular potassium channels caused a vasoconstriction in the systemic and splanchnic vascular beds. In these territories, therefore, the opening of glibenclamide-sensitive potassium channels might be responsible for a basal vasodilator tone.

Altered control of vascular tone by adenosine triphosphate-sensitive potassium channels in rats with cirrhosis.

BACKGROUND/AIMS: Because the activation of arterial adenosine triphosphate (ATP)-sensitive potassium (KATP) channels is known to induce vasodilation, these channels may contribute to baseline vasodilator tone in cirrhosis. This study aimed to examine hemodynamic responses to glibenclamide, a KATP channel blocker, and to aprikalim, a vasodilator activating KATP channels, in normal and cirrhotic rats. METHODS: Splanchnic and systemic hemodynamic responses to glibenclamide (2.5, 5, 20, 30 mg/kg, intravenously) were studied. The arterial pressure response to aprikalim (200 mu/kg, intravenously) was studied with and without glibenclamide pretreatment (20 mg/kg). RESULTS: In cirrhotic rats, glibenclamide (5, 20, 30 mg/kg but not 2.5 mg/kg) significantly increased vascular resistance in portal and systemic territories. In normal rats, the latter effects occurred with 20 and 30 mg/kg of glibenclamide only. Aprikalim-induced arterial hypotension was significantly less marked in cirrhotic than in normal rats. Following glibenclamide, aprikalim-induced arterial hypotension was significantly less marked in cirrhotic than in normal animals. CONCLUSIONS: In rats with cirrhosis, the glibenclamide-induced vasoconstriction indicates that a vasodilator tone due to KATP channel opening existed under baseline conditions. Moreover, this study suggests that the control of vascular tone by KATP channels is altered in cirrhosis.

[Relationships between plasma concentrations of endothelin and the severity of liver cirrhosis]. / Relations entre les concentrations plasmatiques d'endothéline et la sévérité de la cirrhose.

OBJECTIVES: Results of studies on plasma endothelin concentrations in patients with cirrhosis are conflicting. Moreover, the relationships between plasma endothelin concentrations and the severity of cirrhosis have not yet been studied. The aim of this study was to measure plasma endothelin concentrations in controls and in patients with cirrhosis. In addition, this study examined the relationships between plasma endothelin concentrations, and the severity of liver disease, splanchnic and systemic hemodynamics. METHODS: Plasma endothelin concentrations (in the hepatic vein and the right atria), hepatic venous pressures, arterial pressure, cardiac output, pulmonary pressures and plasma concentrations of sodium and creatinine were measured in 7 controls and 28 patients with cirrhosis. RESULTS: Plasma endothelin concentrations in the hepatic vein and the right atria were significantly higher in patients with cirrhosis (18.9 +/- 2.9 and 20.2 +/- 3.1 pg/mL, respectively) than in controls (6.1 +/- 1.1 and 7.2 +/- 1.1 pg/mL, respectively). In these patients, hepatic venous plasma endothelin concentrations were significantly correlated with Pugh's score (r = 0.49), hepatic venous pressure gradient (r = -0.44), and plasma sodium concentrations (r = -0.46). No significant correlation was found between plasma endothelin concentrations and systemic hemodynamics. CONCLUSION: Plasma endothelin concentrations are increased in patients with cirrhosis. Moreover, this increase is more marked in patients with severe liver disease than in patients with no or moderate impairment of liver function.

Arterial and mixed venous acid-base status in patients with cirrhosis. Influence of liver failure.

Although it has been established that liver failure is associated with arterial hypocapnia and alkalaemia (i.e., respiratory alkalosis), the influence of liver failure on mixed venous acid-base status has not yet been studied. Thus, arterial and mixed venous acid-base status were simultaneously measured in controls and in a large series of patients with cirrhosis. Grade B patients (n = 28) or Grade C patients (n = 21) had significantly lower arterial and mixed venous carbon dioxide tensions than controls (n = 29). Grade B or Grade C patients also had significantly higher arterial, mixed venous pH, and lower mixed venous bicarbonate concentrations than controls. Among Grade A patients (n = 27), those with the lowest Pugh's score (i.e., equal to five) had significantly lower mixed venous carbon dioxide tension than controls. The other arterial and mixed venous acid-base values did not differ significantly between Grade A patients with the lowest Pugh's score and controls. Grade A patients with a Pugh's score equal to six and Grade B patients had similar acid-base disorders. No significant differences were found between groups concerning the anion gap and plasma chloride concentrations. In conclusion, this study shows that in Grade B or C patients, respiratory alkalosis was responsible for mixed venous hypocapnia, alkalaemia and hypobicarbonataemia. In addition, in Grade A patients with the lowest Pugh's score (equal to five), analysis of arterial and mixed venous blood revealed that mixed venous hypocapnia was the sole anomaly of the acid-base status. This last finding suggests that mixed venous hypocapnia might be an early event preceding the onset of arterial hypocapnia.

Septic shock in patients with cirrhosis: hemodynamic and metabolic characteristics and intensive care unit outcome.

OBJECTIVES: To examine the hemodynamic and metabolic characteristics and ICU outcome of septic shock in patients with cirrhosis. DESIGN: Prospective, comparative study. Measurements performed in the first 24 hrs of septic shock. SETTING: A general hospital ICU. PATIENTS: Twelve patients with cirrhosis and 23 patients without cirrhosis admitted for septic shock. MEASUREMENTS AND MAIN RESULTS: Arterial pressure was measured using an arterial catheter. Pulmonary arterial and right atrial pressures were measured by using a pulmonary artery catheter. Cardiac output was determined by using the thermodilution method. Pulmonary arterial L-lactate plasma concentrations were measured using an automated spectrophotometer, and blood temperature was measured using a cardiac output computer. Arterial and mixed venous PO2, PCO2, and pH values were measured by using specific electrodes. Oxygen saturations and hemoglobin concentrations were measured using a hemoximeter. Patients with cirrhosis had decompensated liver disease (grade C of the Child-Pugh classification). The number of Gram-negative infections and therapeutic interventions were similar in both groups. Patients with cirrhosis had higher cardiac indices (5.14 +/- 0.52 [SE] vs. 3.91 +/- 0.30 L/min/m2, p less than .05), plasma lactate concentrations (9.0 +/- 2.0 vs. 5.2 +/- 0.7 mmol/L, p less than .05) and ICU mortality rates (100% vs. 43%, p less than .05), and lower blood temperatures (35.5 +/- 0.6 vs. 37.6 +/- 0.2 degrees C, p less than .05) than patients without cirrhosis. Systemic vascular resistance, arterial pressure, pulmonary arterial pressure, oxygen delivery and consumption, and arterial and mixed venous acid-base status were not significantly different between the two groups. CONCLUSIONS: In patients with cirrhosis, septic shock was characterized by severe liver dysfunction, low blood temperature, marked increases in cardiac index and lactic acidemia, and a 100% ICU mortality rate. These findings should be taken into account if patients with cirrhosis are to be included in controlled studies on septic shock.