INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Sepsis and septic shock lead to overproduction of nitric oxide (NO) from L-arginine by an inducible NO synthase (NOS) (1). NO itself is regarded as playing an important role in mediating cardiovascular failure associated with sepsis (2). Therefore the pharmacological inhibition of NO synthesis is considered a potential treatment for sepsis-associated hypotension that is refractory to fluid therapy (5, 6).

There is evidence in the literature that NOS inhibition in fact is able to restore macrohemodynamics with equivocal effects on the perfusion of splanchnic organs. In fact, in a long-term hyperdynamic ovine model of septic shock Booke and coworkers demonstrated a redistribution of blood flow away from the pancreas to the colon as well as reduced mesenteric perfusion in animals treated with N-monomethyl-L-arginine (L-NMMA) in order to restore blood pressure (7, 8). Moreover, controversial results exist in the literature concerning the effect of NOS inhibitors on liver function in endotoxic shock: in hypodynamic models NOS inhibitors demonstrated a benefit for mesenteric perfusion owing to hemodynamic stabilization (9), but other authors found detrimental effects of NOS inhibition as well (10), depending on the models used and on the time when NOS inhibitors were administered (see Kilbourn and coworkers [13] for review). Finally, there is evidence of a beneficial effect of NO donors in early stages of sepsis with regard to hepatic perfusion and metabolism (14- 16), suggesting that nonselective NOS inhibition might be detrimental in early sepsis.

The mainstay of the treatment of sepsis-associated hypotension so far is fluid therapy and adrenergic stimulation with norepinephrine (NOR). Only a few reports compare the standard therapy with NOR with a vasopressor treatment with NOS inhibitors (7, 8). Although the effects of NOR and NOS inhibition on macrohemodynamics are well known there is a lack of data comparing the effect of either therapy on hepatic perfusion and liver oxygen transport and, in consequence, on liver metabolic function.

Therefore the objective of the present study was to compare the effects of NOR and nonselective NOS inhibition with L-NMMA on liver blood flow, O₂ kinetics, and metabolic activity in long-term, hyperdynamic, porcine endotoxin (ETX)- induced shock based on a clinical scenario; the purpose was to verify whether there are deleterious side effects of either of the therapeutic approaches when they are used according to a clinical strategy. From the currently available nonselective NOS inhibitors we chose L-NMMA, because it has a relatively short half-life and can be administered continuously with adequate control; moreover, clinical studies with L-NMMA were underway (17). Parameters on hepatic hemodynamics and energy metabolism, however, are mostly not available in human studies. In a previous study we demonstrated that hypermetabolism associated with endotoxic shock causes derangements of hepatocellular energy metabolism despite maintenance of hepatic perfusion and oxygenation (18). Therefore this study was undertaken to investigate whether vasopressor support may affect the sepsis-induced alterations in liver hemodynamics, O₂ transport, and, as a consequence, hepatic energy balance.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animal Preparation

The study protocol was approved by the University Animal Care Committee as well as the federal authorities for animal research of the Regierungspräsidium Tübingen (Baden-Württemberg, Germany). Thirty domestic pigs (Deutsches Landschwein) of either sex with a median body weight of 42 kg (interquartile range, 40-44 kg) were fasted for 24 h, with water ad libitum. The animals were premedicated with intramuscularly administered atropine (Atropinsulfat, 2.5 mg; Braun, Melsungen, Germany) and azaperone (Stresnil, 150-200 mg; Janssen, Neuss, Germany) followed by cannulation of an ear vein. Anesthesia was induced by intravenous administration of sodium pentobarbital (Nembutal, 10 mg kg¹; Sanofi Winthrop, Munich, Germany) and ketamine (Ketavet, 1.5-2.0 mg kg¹; Parke-Davis, Berlin, Germany). The pigs were orally intubated, and their lungs were mechanically ventilated (fraction of inspired oxygen [FI_O2], 0.4; positive end-expiratory pressure [PEEP], 5 cm H₂O) (Servo 900B; Siemens, Erlangen, Germany) with a tidal volume of 15 ml kg¹ at a respiratory rate of 10-12 breaths min¹ adjusted to maintain arterial PCO₂ between 35 and 40 mm Hg. During the surgical preparation the inspired gas mixture consisted of N₂O and O₂, during the observation period the mixture consisted of air and O₂. Anesthesia was maintained with continuous intravenous administration of pentobarbital (200-300 mg h¹), and depth of anesthesia was controlled by continuous electroencephalogram (EEG) monitoring (Neurotrac; Interspec, Conshohocken, PA). The spectral edge frequency was always less than 15 Hz, and the median power frequency was 5-10 Hz. Buprenorphine (Temgesic, 0.3 mg; Boehringer GmbH, Mannheim, Germany) was added every 4 h and before any surgical or noxious stimuli in order to prevent a rise in heart rate and arterial pressure owing to inadequate anesthesia. Muscle paralysis was obtained with alcuronium (Alloferin, 14 mg h¹; Hoffmann-La Roche AG, Basel, Switzerland). The right and left jugular veins and a submandibular vein as well as the right and left femoral arteries were surgically exposed. A central venous catheter for drug, isotope, and fluid infusion was inserted into the superior vena cava, and a balloon-tipped thermodilution pulmonary artery catheter (93A754 7F; Baxter Healthcare, Irvine, CA) was placed for the measurement of central venous pressure (CVP), mean pulmonary artery pressure (MPAP), and pulmonary artery occluded pressure (PAOP) (Medex MX 80 pressure transducers; Medex, Hillard, OH). In one femoral artery a catheter was placed for continuous blood pressure recording and blood sampling, in the other one a 3-Fr thermistor-tipped fiberoptic catheter for thermal-dye double-indicator dilution measurements was placed (FT-Pulsiocath PV 2023; Pulsion, Munich, Germany). Ringer's lactate solution (10 ml kg¹ h¹) was infused intravenously as maintenance fluid. A midline laparotomy was performed, and precalibrated ultrasonic flow probes (Transonic Systems, Ithaca, NY) were placed around the portal vein and the common hepatic artery distal to the takeoff of the gastroduodenal artery. Flows were continuously recorded using a T206 flow meter (Transonic Systems) and flow plots graphically displayed on a plotter. A 4-Fr catheter (CS-16402; Arrow, Reading, PA) was then introduced into the portal vein, and an angiography catheter (7F Multipurpose A-1; Cordis, Roden, The Netherlands) was introduced via the right jugular vein into a hepatic vein under ultrasound guidance. A glass tube was inserted through the abdominal wall onto the liver surface for the placement of the remission spectrophotometry light guide. After instrumentation a stabilization period of 8 h was allowed before baseline measurements were recorded.

Measurements and Calculations

Cardiac output () was determined by thermodilution (66S monitor; Hewlett-Packard, Palo Alto, CA), the data reported being the mean of four or five injections of 10 ml of ice-cold saline randomly spread over the respiratory cycle. The intrathoracic blood volume was measured by arterial thermal-green dye double indicator dilution (COLD Z-021; Pulsion) after injection of 10 ml of cold indocyanine green (2.5 mg ml¹) dissolved in distilled water. The continuously recorded portal venous and hepatic arterial blood flow rates were summed to obtain the total hepatic blood flow. Arterial, portal, and hepatic venous blood samples were analyzed for PO₂, PCO₂, and pH (Nova Stat Profile Ultra; Nova Biomedical, Waltham, MA) as well as total hemoglobin and hemoglobin O₂ saturation (IL 482 CO-oximeter [Instrumentation Laboratories, Lexington, MA], calibrated for pig blood). Systemic O₂ delivery (DO₂) was calculated from the standard formula. Hepatic DO₂ (hDO₂) and O₂ uptake (hVO₂) were calculated as the product of portal venous and hepatic arterial blood flow times the portal venous and the arterial O₂ content, respectively, and the portal-hepatic venous and the arterial-hepatic venous O₂ content differences where appropriate. Arterial blood glucose levels were measured every 2 h using an automatic enzymatic glucose analyzer (Glucometer Elite; Bayer AG, Leverkusen, Germany). Systemic VO₂ was calculated according to the Fick principle, as calorimetric measurement of VO₂ was impossible because of a progressive increase in FI_O2 beyond 0.5 owing to sepsis-associated respiratory failure. Body temperature was kept within 0.5° C of the baseline value by using a heating mattress or external cooling as necessary.

Capillary hemoglobin O₂ saturation (HbSc_O2) on the liver surface was measured by means of remission spectrophotometry using the Erlanger Mikrolightguide spectrophotometer device (EMPHO; Bodenseewerk Gerätetechnik, Überlingen, Germany) as described earlier (18, 19). The means as well as the frequency distributions of the HbSc_O2 were calculated. For this purpose, the Mikrolightguide device was placed on the liver surface through the glass tube, which had been passed through the abdominal wall. The reported values of the HbSc_O2 are the mean of 300 spectra, each recorded at six different measurement sites, i.e., a total of 1,800 spectra was analyzed for each data point in every animal.

Arterial, portal, and hepatic venous lactate concentrations were enzymatically measured in duplicate, using an automatic whole blood lactate analyzer (YSI model 2300 STAT; Schlag Wissenschaftliche Messinstrumente, Bergisch-Gladbach, Germany). Arterial, portal, and hepatic venous alanine levels were assessed in duplicate with the ninhydrin reaction after separation by high-performance liquid chromatography (amino acid analyzer LC 3000; Biotronik, Hamburg, Germany). For the determination of hepatic venous lactate/pyruvate ratios the lactate and pyruvate concentrations were measured in duplicate spectrophotometrically (18). The coefficients of variation for the lactate, pyruvate, and alanine measurements were 1.8, 9.2, and 6.3%. Hepatic lactate and alanine fluxes were subsequently calculated as the product of portal venous and hepatic arterial blood flow times the portal-hepatic venous and the arterial-hepatic venous concentration differences, respectively.

The endogenous glucose production (EGP) rate was determined as described previously (18), using stable, nonradioactive isotope- labeled [6,6-²H₂]glucose (Mass Trace, Woburn, MA). The glucose rate of appearance of unlabeled glucose (Ra) was derived from the arterial plasma isotope enrichment, and the EGP rate was subsequently calculated as the difference between Ra and the infusion rate of unlabeled glucose. The glucose plasma isotope enrichment used for the computation of Ra was the corresponding mean of triplicate blood samples obtained within 10 min each and analyzed by gas chromatography/mass spectrometry (GC 5890, MS 5970; Hewlett-Packard) after derivatization of glucose to 1,5-pentaacetate.

Protocol

The animals were randomly assigned to three groups: endotoxin (ETX, n = 8), norepinephrine (NOR, n = 11), and L-NMMA (n = 11). After recording baseline measurements (preshock) a central venous ETX infusion (Escherichia coli lipopolysaccharide B 0111:B4 [Difco Laboratories, Detroit, MI], 20 mg L¹ in 5% dextrose) was started. The ETX infusion rate was incrementally increased every 20 min until the MPAP reached 50 mm Hg and was then subsequently adjusted to result in moderate pulmonary hypertension with an MPAP of ~ 35-40 mm Hg. This ETX infusion rate was maintained until the end of the experiment. Hydroxyethylstarch (Infukoll HES, 6%; Serum-Werk, Bernburg, Germany) was administered as required to maintain a mean arterial pressure (MAP) of > 60 mm Hg, and a mixture of glucose and xylitol (each 10%) (GX 20%; Pharmacia, Erlangen, Germany) was infused to maintain arterial blood glucose levels between 5 and 7 mmol L¹. Further hemodynamic, gas exchange, and metabolic measurements were recorded 12 h after the start of ETX infusion. In the treatment groups therapy with NOR (0.1 µg kg¹ min¹) and L-NMMA (1 mg kg¹ h¹) was initiated after 12 h of ETX infusion and subsequently titrated to maintain the MAP at preshock levels.

Further acquisition of data was performed 18 and 24 h after the start of ETX infusion (corresponding to 6 and 12 h after the start of vasopressor treatment). After the last set of data had been obtained the animals were killed by KCl injection under deep anesthesia.

Statistical Analysis

All values shown are median and interquartile range unless otherwise stated. Differences between preshock values and those during ETX or vasopressor infusion within each group were tested using a Friedman repeated measures analysis of variance on ranks and a subsequent Student-Newman-Keuls test for multiple comparisons. A value of p < 0.05 was regarded as significant. Differences between the groups were analyzed using a Kruskal-Wallis analysis of variance on ranks with  adjustment according to Bonferroni. A value of p < 0.017 was regarded as significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Table 1 shows that the dosage of the endotoxin infusion did not differ significantly between the three groups. In the vasopressor therapy groups NOR was administered at a median rate of 0.19 and 0.44 µg kg¹ min¹, and L-NMMA at 4.0 and 4.8 mg kg¹ h¹, at 6 and 12 h of treatment, respectively, to achieve the goal of maintaining the MAP at preshock values. Global hemodynamic as well as O₂ exchange parameters in response to the ETX infusion and in the vasopressor treatment groups are summarized in Table 2. After 12 h of ETX infusion a hyperdynamic, normotensive septic shock with decreased systemic vascular resistance (SVR) was established. Adequate volume resuscitation was ensured by monitoring intrathoracic blood volume (ITBV; given as ml kg¹), which significantly increased from preshock values of 26 (26) in the ETX, 26 (25) in the NOR, and 27 (23) in the L-NMMA groups to 38 (33-40), 41 (32-54), and 32 (30-37) at the end of the experiment in the ETX, NOR, and L-NMMA groups, respectively, without significant intergroup differences. Whereas after 24 h ETX had resulted in a significant decrease in MAP despite adequate volume resuscitation, both NOR and L-NMMA maintained blood pressure at preshock levels. NOR significantly increased whereas L-NMMA significantly decreased (but not below the normal baseline value) concomitant with restored SVR. Whereas the time course of global oxygen delivery (DO₂) paralleled that of , global VO₂ was not affected in the three experimental groups. The hepatic blood flow as well as O₂ exchange and metabolic responses are summarized in Table 3. Despite the increased SVR, L-NMMA did not affect hepatic perfusion, demonstrating a redistribution of blood flow to the liver. By contrast, in parallel with the enhanced systemic blood flow NOR increased total hepatic blood flow and thereby hDO₂ without effect on hVO₂. The hepatic alanine uptake rate decreased similarly in the three groups without any intergroup difference.

The unaltered macrocirculatory liver O₂ supply was associated with constant microcirculatory O₂ availability as documented by the unchanged mean HbSc_O2 (Table 4) and HbSc_O2 frequency distribution (Figure 1).

View larger version (35K): [in this window] [in a new window]   Figure 1.   Frequency distribution of capillary hemoglobin O2 saturation (HbScO2) of the liver surface in the ETX (open plot), NOR (gray plot), and L-NMMA (solid plot) groups.

Despite the well-preserved macro- and microcirculatory O₂ availability there was a significant continuous fall in arterial pH from a preshock value of 7.52 (7.50-7.53) to 7.35 (7.32-7.38) in the ETX group and from 7.51 (7.49-7.52) to 7.27 (7.22-7.40) and 7.49 (7.48-7.53) to 7.32 (7.21-7.36) in the NOR group and L-NMMA group, respectively, at the end of the experiment without any intergroup difference. Mixed venous pH significantly decreased in parallel from a preshock level of 7.48 (7.47-7.49) and 7.48 (7.46-7.49) and 7.47 (7.45-7.50) to 7.30 (7.28-7.35) and 7.23 (7.17-7.37) and 7.27 (7.17-7.30) in the ETX, NOR, and L-NMMA groups, respectively, at 24 h. This progressive systemic acidosis was accompanied by a simultaneous significant decline in hepatic venous pH in all groups over the study period ranging from 7.48 (7.47-7.50) to 7.30 (7.25-7.34), from 7.49 (7.47-7.51) to 7.22 (7.17-7.37), and from 7.48 (7.44-7.49) to 7.25 (7.16-7.30) in the ETX, NOR, and L-NMMA groups, respectively, without any intergroup differences (Figure 2). The decreased hepatic venous pH was associated with significantly decreased hepatic lactate uptake rates and the liver lactate balance became negative in all three experimental groups without any effect of the vasopressor treatment. The corresponding values for liver lactate uptake values (given as µmol kg¹ min¹) at preshock compared with after 24 h of septic shock were 13 (11) to 6 (16-5) (ETX), 9 (6) to 5 (10-6) (NOR), and 9 (7) to 17 (24-9) (L-NMMA) in the three groups, respectively; intergroup comparison did not show any significant differences. The decreased lactate uptake rates were accompanied by significantly increased hepatic venous lactate/pyruvate ratios ranging from 25 (15-33) to 61 (32-123), from 31 (22-37) to 105 (42-205), and from 22 (19-29) to 148 (68-187) in the ETX, NOR, and L-NMMA groups, respectively (Figure 3); vasopressor treatment had no further effect.

View larger version (16K): [in this window] [in a new window]   Figure 2.   Arterial (top), mixed venous (middle), and hepatic venous (bottom) pH in the ETX (open plot), NOR (cross-hatched plot), and L-NMMA (hatched plot) groups. Data presented indicate median, interquartile, and 5% and 95% interquantile range. Dagger () designates a significant difference versus preshock.

View larger version (19K): [in this window] [in a new window]   Figure 3.   Liver lactate uptake rates (top) as well as hepatic venous lactate/pyruvate ratio (bottom) in the ETX (open plot), NOR (cross-hatched plot), and L-NMMA (hatched plot) groups. Data presented indicate median, interquartile, and 5% and 95% interquantile range. Dagger () designates a significant difference versus preshock.

EGP (given as µmol kg¹ min¹) significantly increased in the ETX group from a preshock value of 25 (20) to 33 (29- 38) at 12 h and to 32 (23-40) and 37 (29-66) at 18 and 24 h, respectively. NOR transitorily increased the EGP even further (p = 0.037 versus 12 h) from a preshock value of 23 (21) and 35 (31-38) at 12 h of shock to 44 (37-50) at 18 h, and reversed it to 35 (28-50) at 24 h. From a preshock value of 24 (18) and 30 (27) at 12 h L-NMMA resulted in a normalized EGP of 26 (25-44) and 30 (23) at 18 and 24 h, respectively (Figure 4). Nevertheless, the dissociation of EGP and hepatic glucose precursor uptake was not significantly affected by either treatment.

View larger version (19K): [in this window] [in a new window]   Figure 4.   Endogenous glucose production in the ETX (open plot), NOR (cross-hatched plot), and L-NMMA (hatched plot) groups. Data presented indicate median, interquartile, and 5% and 95% interquantile range. Dagger () denotes a significant difference versus preshock; section symbol (§) designates a significant difference NOR versus L-NMMA at the same time.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The aim of this study was to compare the effects of a standard vasopressor therapy with NOR and nonselective NOS inhibition on liver blood flow, O₂ exchange, and metabolism during long-term, hyperdynamic endotoxic shock in pigs.

Vasopressor treatment was started 12 h after the onset of ETX infusion for several reasons: first, because vasodilation associated with septic shock is believed to result from excess NO production induced by increased activity of the inducible isoform of NOS (iNOS), a sufficient time interval to allow for iNOS activation is mandatory. In fact, Villamor and coworkers (20) demonstrated that NOS activity was not enhanced after a 5-h incubation of isolated pig arteries with endotoxin, whereas it had significantly increased (three- to fourfold) after 20 h. Moreover, we have shown that the NO synthesis rate as reflected by the nitrate production rate increased 4-6 h after starting an ETX infusion in a hyperdynamic porcine shock model (21). Finally, in a previous experiment we showed that a continuous 12-h infusion of ETX allowed the establishment of a hyperdynamic shock state with a sustained increase in (18) that fulfills the criteria required for a relevant clinical model of septic shock (22).

Both treatment strategies resulted in an equally effective maintenance of mean blood pressure, the mechanisms, however, being substantially different: NOR maintained the MAP at preshock values as a result of an additional increase in blood flow without significant influence on SVR. This finding confirms previous data reported by Booke and coworkers (7) and probably reflects the intrinsic -adrenoceptor-stimulating properties of NOR (23). In contrast, L-NMMA restored the MAP primarily as a result of increased SVR, a finding that has been reported in animal experiments (7, 8) as well as in human septic shock (5). It is noteworthy that the increased SVR, however, was associated with unchanged portal venous and hepatic arterial and consequently total hepatic blood flow. This finding of a redistribution of toward the splanchnic organs is in sharp contrast to previous studies, in which reduced blood flow to the organs has been reported during NOS inhibition (11). We can only speculate about this difference, but the well preserved sustained increase in resulting from adequate colloid fluid resuscitation, which was reflected by the increased ITBV, may be of particular importance in this context. This hypothesis is confirmed by data from Offner and coworkers (9), who found an increase in mesenteric blood flow in a fluid-resuscitated endotoxic shock model in pigs treated with the nonspecific NOS inhibitor N-nitro-L-arginine methyl ester (L-NAME).

Irrespective of the effect on hepatic macrocirculatory hemodynamics neither of the treatments affected microcirculatory O₂ availability or capillary heterogeneity as evidenced by both the unchanged mean capillary hemoglobin O₂ and the hemoglobin O₂ frequency distributions. This result contrasts with data from Kubes and Granger (24), who reported impaired splanchnic microcirculatory O₂ supply in a feline model of septic shock. It should be noted, however, that this model exhibited depressed and, hence, a hypodynamic shock state without fluid resuscitation. Nevertheless, there are some technical limitations of this method to be taken into account, as there is interference by biological pigments such as melanin and bilirubin. Moreover, user-dependent skills in avoiding movement artifacts as well as in exerting no pressure on the measurement site may influence the remission spectrophotometry results. Finally, the EMPHO system has not yet provided a mathematically rigorous technique for eliminating artifactual spectra, although this problem is attenuated by recording large numbers of spectra at a high frequency. Despite the technical limitations of remission spectrophotometry, this method was demonstrated to detect changes in HbSc_O2 as shown by Temmesfeld-Wollbrück and coworkers (19) in septic patients. It must be noted, however, that in contrast to these data there was no significant change in HbSc_O2 or in its frequency distribution in our model. We can only speculate about this, but the different measurement sites (gastric mucosa versus liver surface) as well as the differences in volume resuscitation and catecholamine regimen might be important in this context. Regardless of the effects on hepatic O₂ availability the hemodynamic stabilization did not affect hepatic oxygen uptake: given the results obtained for hDO₂, hepatic O₂ extraction remained unchanged in the L-NMMA group and even decreased in NOR-treated animals. A possible explanation might be that despite maintained O₂ availability there may be disturbances in cellular O₂ metabolism independent of hemodynamic stability. As a consequence, intracellular O₂ metabolism capacity may determine cellular O₂ uptake rather than circulatory O₂ availability. Thus, our data may add evidence for an existing "cytopathic" hypoxia as described by Fink (25).

NOR treatment led to a transitory increase in EGP beyond the values of the ETX group, probably reflecting the metabolic response of -agonistic properties of this catecholamine in the splanchnic region. This well-known physiologic thermogenic effect of NOR has also been demonstrated in patients with septic shock when NOR was added to the treatment (26) or replaced by the pure -agonist phenylephrine (27). By contrast, NOS inhibition with L-NMMA restored EGP back to normal values in our experiment, a finding that is in sharp contrast to numerous studies, underscoring the inhibitory role of NO in hepatic gluconeogenesis (28). It must be noted, however, that those reports always describe results from studies in vitro or in rodent models of endotoxic shock characterized by severely compromised hemodynamics. Wolfe (29) had already emphasized the different effects of endotoxin on hepatic glucogenic capacity in vivo and in vitro.

A major characteristic of sepsis and septic shock in humans is hypermetabolism, due mainly to pronounced increases in the EGP rate (30) such as we have demonstrated in porcine endotoxic shock (18). At first glance, the L-NMMA-associated decrease in EGP back to normal values may have particular impact with respect to the hepatic O₂ supply/demand relationships. On the basis of the stoichiometry of glucose formation, 6 mol of ATP (corresponding to 1.1 mol of O₂) is required for the de novo synthesis of 1 mol of glucose (30). Assuming that the liver accounted for approximately 60-70% of EGP (33) the median EGP levels of 37 and 44 µmol kg¹ min¹, as found in the ETX and NOR groups, respectively, would correspond to an estimated O₂ demand of about 0.6 ml kg¹ min¹: theoretically the total hVO₂, hence, was required for EGP, and any other metabolic pathway theoretically would therefore have been dependent on anaerobic ATP formation. Consequently, reducing EGP without affecting O₂ availability should have improved the hepatic O₂ and substrate supply/demand balance. It should be underscored, however, that L-NMMA, similarly to NOR, did not alter any parameter of liver energy balance; in particular, it did not influence the progressive fall in lactate balance or the dissociation between EGP and hepatic glucose precursor uptake. Moreover, the continuous increase in hepatic venous lactate/ pyruvate ratios, a marker of cytosolic redox state (34), was not affected either. Several phenomena may explain this finding that neither NOR nor L-NMMA could influence the ETX- induced derangements in hepatic energy metabolism: L-NMMA may not have totally reversed an NO-induced inhibition of mitochondrial respiration, and a potential benefit of increased liver O₂ availability during NOR infusion might have been outweighed by the simultaneous thermogenic effect of this compound resulting in increased O₂ demands. Finally, despite the different effect on hepatic gluconeogenesis neither of the two treatments probably affected the O₂ demand of other liver cells such as those of the reticuloendothelial system, which may account for about 25-50% of hepatic metabolic activity and thus O₂ demands as estimated from glucose uptake studies in parenchymal and nonparenchymal liver cells (35).

In summary, L-NMMA and NOR were equally effective in maintaining MAP during long-term hyperdynamic porcine endotoxic shock. Neither of the treatments compromised hepatic perfusion, or macro- as well as microcirculatory O₂ availability. The ETX-induced dissociation of glucogenic precursor uptake and de novo glucose synthesis could not be reversed by either vasopressor despite the different mechanisms on cardiocirculatory system and key liver energy pathways. In particular, modulating EGP by the two compounds did not result in any difference in hepatic O₂ exchange and lactate clearance. Provided that microcirculatory O₂ supply is preserved, failure of hepatic energy balance is not related to hemodynamic stabilization.