INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Hepato-splanchnic dysoxia is considered by many as an important mechanism in the development of multiple organ failure (1). During sepsis liver metabolism can increase out of proportion to blood flow (4) so that oxygen demand may exceed oxygen supply in the hepato-splanchnic area. Such a phenomenon has been suggested by the common observation of an increased gradient between mixed venous and hepatic venous oxygen saturation (DSO₂) in septic patients (6, 8, 11- 14). Although cellular uncoupling of metabolic processes may also be present, there is experimental evidence that cellular metabolism is preserved if adequate blood flow is provided. In pigs with peritonitis, Hirai and coworkers (10) observed that the ATP content of the liver was closely related to the level of blood flow. In septic rats, Dahn and coworkers (7) reported that liver oxygen consumption (O₂) could increase when more substrate was provided, suggesting that oxidative metabolism was preserved but limited by extrahepatic factors.

An important question is whether liver dysoxia could be revealed by the presence of regional O₂/DO₂ dependency. Dobutamine is a convenient drug to study O₂/DO₂ relationships because it has a rapid onset of action, is well tolerated in septic patients (15), and has limited thermogenic effects in the critically ill (16, 17). Dobutamine has been shown to increase hepato-splanchnic blood flow in endotoxic animals (18, 19) and in septic patients (12). In this study we investigated the effects of dobutamine on hepato-splanchnic O₂/DO₂ relationships in septic patients. To separate the effects of changes in blood flow from the thermogenic effect of the catecholamine, we also compared the effects of the administration of dobutamine to those of the application of a positive end-expiratory pressure (PEEP).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This investigation was approved by the institutional ethics committee and informed consent was obtained from the next of kin. Pregnant women and patients with liver cirrhosis, liver metastases, or acute liver failure were not included.

The study included 42 patients (age, 59 ± 17 yr; males, n = 31 [73%]; weight, 71 ± 13 kg) with severe sepsis as defined by the presence of fever (temperature > 38.5° C) or hypothermia (temperature < 35.5° C), leukocytosis or leukopenia (white blood cell count > 10,000 or < 4,000/mm³), and hypotension (systolic arterial pressure < 90 mm Hg for more than 2 h, requiring fluid challenge and/or vasoactive agents), within the last 24 h, in the presence of a source of infection. Sources of infection included the lungs (n = 29), abdomen (n = 7), bone and soft tissues (n = 3), skin (n = 3), and urinary tract (n = 2). The mean APACHE II score (20) was 17.6 ± 7.9 and the mean SOFA score (21) was 10.4 ± 3.3. Each patient was mechanically ventilated with an assisted/controlled mode or pressure control mode. PEEP was used in 37 patients at a level of 7.4 ± 3.4 cm H₂O.

No patient was treated with dobutamine during the 5 h preceding the study. Other catecholamines were administered in 23 patients, including dopamine in 23 patients at a dose of 14 ± 7 µg/kg · min and norepinephrine in three patients at a dose of 10 ± 9 µg/min. The doses of these vasopressor agents were unchanged 2 h before and during the study. Each patient was sedated with midazolam and morphine which were infused at a constant rate. For the purpose of the study, each patient was paralyzed with pancuronium, given as a bolus of 4 mg 1 h prior to the study, followed by a continuous infusion of 2 mg/h.

Each patient was monitored with an arterial catheter (radial or femoral) and a pulmonary artery catheter (Swan Ganz 7.5F; Baxter Healthcare, Irvine, CA). A venous introducer (cc-350B-8.5F; Baxter) was inserted in the right (n = 40) or left (n = 1) internal jugular vein, or left subclavian vein (n = 1). Through this introducer a radiopaque catheter (multipurpose catheter 5F; Cook, Bjaerskov, Denmark) was inserted under fluoroscopic guidance in the right hepatic vein.

Study Protocol

After baseline measurements, dobutamine was administered as an infusion of 5 µg/kg · min increasing to 10 µg/kg · min after 30 min. The infusion was then discontinued and another set of measurements obtained 30 min later.

In 14 patients, a PEEP level was then applied or increased from baseline to 10, 15, and 20 cm H₂O. PEEP was not manipulated in patients who were treated with a PEEP level higher than 6 cm H₂O (n = 20), had chronic lung disease prohibiting the application of a high PEEP level (n = 6), or had an auto-PEEP level higher than 5 cm H₂O as determined by the occlusion maneuver (n = 2).

Measurements

Hemoglobin level (Hb), hematocrit, aspartate and alanine transaminases, alkaline phosphatase, bilirubin, and prothrombin time were determined at baseline.

Sets of measurement were obtained every 30 min and included core temperature, heart rate, arterial pressure, pulmonary arterial pressure, pulmonary artery balloon occluded pressure, right atrial pressure, hepatic vein pressure, cardiac index, hepato-splanchnic blood flow, arterial, mixed venous and hepatic venous blood gases, and arterial and hepatic venous lactate levels. All pressures were determined at end-expiration, with the zero reference set at midchest. Cardiac index was determined by the thermodilution technique (COM1 or COM2; Baxter), using injections of 10 ml of cold (< 10° C) solution of dextrose 5% in water obtained through a closed system (CO-set; Baxter). Each bolus was injected at end-inspiration. A total of five to seven measurements within 5% of one another was averaged.

Hepato-splanchnic blood flow was determined using the continuous infusion of indocyanine green (ICG) as proposed by Uusaro and coworkers (22). ICG was administered as a bolus of 12 mg (ICG Pulsion; Pulsion, Munich, Germany) followed by a continuous infusion of 1 mg/min. At each point, 3 ml of arterial and hepatic venous blood were withdrawn in triplicate. The samples were centrifuged and the plasma stored at 18° C until analysis within 48 h. Plasma absorbance was determined by spectrophotometry (Uvikon 930 spectrophotometer; Kontron, Basel, Switzerland) at an 800 nm wavelength. In order to limit the effects of background absorbance, the Allen's correction was performed using absorbance determinations at wavelengths of 650 and 950 nm. Plasma ICG levels were then calculated using a reference curve obtained by diluting ICG in blood freshly withdrawn from other patients. A reference curve was performed for each batch of ICG. The hepato-splanchnic blood flow (SPF) was calculated as follows:

SPF (ml/min · M²) = ICG administration rate (mg/min)/([ICG]_a  [ICG]_hv) × (1  hct) × BSA where [ICG]_a and [ICG]_hv are the arterial and hepatic venous ICG concentrations; hct, the hematocrit level; and BSA, the body surface area.

Arterial, mixed venous, and hepatic venous blood gases were determined (analyzer model ABL3; Radiometer, Copenhagen, Denmark) and hemoglobin saturations were measured (Hemoximeter OSM3; Radiometer).

DO₂, hepato-splanchnic oxygen delivery (DO₂spla), O₂, hepato-splanchnic oxygen consumption (O₂spla), whole body and hepato-splanchnic oxygen extraction ratios (EO₂ and EO₂spla) were calculated by the following formulas:

where CI is the cardiac index and Sa_O2, Sv_O2, and Sh_O2 are the arterial, mixed venous, and hepatic venous oxygen saturations.

The arterial and hepatic venous lactate were also determined in triplicate (2300 STAT plus; Yellow Spring Instruments Inc., Yellow Springs, OH) and averaged. Splanchnic lactate consumption (mEq/ min · M²) was calculated as the product of SPF by the arterial hepatic venous lactate difference.

Statistical Analysis

In the first part of the study, a one-way analysis of variance (ANOVA) for repeated measurements was done to assess the effects of incremental doses of dobutamine. When the F value was significant, a Newman-Keuls test was applied. O₂/DO₂ relationships were then analyzed individually with linear regression and the slopes of the hepato-splanchnic O₂/DO₂ relationships were correlated with various physiologic, hemodynamic, biological, and therapeutic parameters (n = 35). A Bonferroni's adjustment was applied for multiple comparisons. The slopes of the O₂/DO₂ relationships determined by dobutamine administration and PEEP application were compared using Student's t test for paired data.

In the second part of the study, patients were separated post hoc in two groups according to their baseline DSO₂ (lower or higher than 10%). Baseline hemodynamic, physiologic, biological, and therapeutic parameters were compared using Student's t test for unpaired data. The effects of incremental doses of dobutamine were assessed using two-way ANOVA (for group and time), followed by a Newman-Keuls test.

Data are presented as mean ± SD, unless stated otherwise. Statistical significance was determined at the 95% confidence interval.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Dobutamine induced a dose-related increase in cardiac index and DO₂ (Table 1). O₂ significantly increased but the O₂/ DO₂ slope was only 7.3 ± 8.8%. Basal hepato-splanchnic blood flow was 877 ± 859 (range, 241 to 4,687) ml/min · M². Hepato-splanchnic blood flow and DO₂spla increased. There was no significant difference in any of the measured parameters before starting, and after discontinuing, the dobutamine infusion.

The increase in hepato-splanchnic blood flow was not dose-related for the whole group since a large interindividual variation was observed. O₂spla increased but with large interindividual variations. Arterial lactate levels remained unchanged. Splanchnic lactate consumption increased in a dose-related fashion (Table 1). ICG clearance increased with dobutamine at the dose of 5 µg/kg · min (Table 1).

The slope of the hepato-splanchnic O₂/DO₂ relationship was significantly correlated with Sh_O2 (y = 66.9  64.8x, r = 0.6, p < 0.05), DSO₂ (y = 3.1 + 55.9x, r = 0.74, p < 0.001) (Figure 1), and EO₂spla (y = 30.6 + 67.5x, r = 0.65, p < 0.05) but not with all the other tested physiological, therapeutic, hemodynamic, and biological parameters. Of note, there was no significant relationship between hepato-splanchnic O₂/DO₂ slopes and the doses of the other adrenergic agents. Splanchnic lactate consumption was significantly correlated with the arterial lactate levels (y = 0.164 + 0.046x, r = 0.34, p < 0.05).

View larger version (15K): [in this window] [in a new window]   Figure 1.   Relationship between the gradient between mixed venous and hepatic venous oxygen saturation (DSO2) and the slope of the hepato-splanchnic O2/DO2 relationship.

In the 14 patients in whom DO₂ was altered successively by dobutamine administration and by PEEP application, the slope of the hepato-splanchnic O₂/DO₂ relationship was similar during these two interventions (21.9 ± 14.2% versus 21.9 ± 13.9%, p = NS) (Table 2).

The patients were separated in two groups according to their gradient between Sv_O2 and Sh_O2 (DSO₂). Thirteen patients had a DSO₂ lower or equal to 10% (Group I) and 29 patients had a DSO₂ higher than 10% (Group II). At baseline, Group II patients were treated with a higher PEEP level (7.5 ± 4.3 versus 4.6 ± 2.6 cm H₂O, p < 0.05) and had a lower ICG clearance (280 ± 159 versus 457 ± 348 mg/min · M², p < 0.05) (Table 3) but all other biological, physiological, therapeutic, and hemodynamic data were similar in both groups (Table 4). DO₂ and O₂ increased similarly in both groups (Table 3) and the slope of the systemic O₂/DO₂ relationship was similar in both groups (8.1 ± 12.5 versus 7.4 ± 6.9%, p = NS). Although DO₂spla increased similarly in both groups, O₂spla increased in Group II but not in Group I (Table 3 and Figure 2). The slope of the hepato-splanchnic O₂/DO₂ relationship was significantly higher in Group II than in Group I (31.2 ± 16.7% versus 10.4 ± 5.1%, p < 0.001). As disclosed in a splanchnic blood flow/EO₂spla diagram (Figure 3), patients in Group I remained on the same isopleth while patients in Group II crossed from one isopleth to another, confirming the increase in O₂. Interestingly, the increase in splanchnic lactate consumption and ICG clearance was similar in both groups. In one patient presenting covariance of hepato-splanchnic O₂ and DO₂, the splanchnic region produced lactate at baseline, consumed lactate during dobutamine administration, and again produced lactate after discontinuation of the dobutamine infusion. There was no statistical difference in survival between the two groups (7 of 13 versus 15 of 29, p = NS).

View larger version (12K): [in this window] [in a new window]   Figure 2.   Relationship between hepato-splanchnic O2 and hepato-splanchnic DO2. Group I: patients with gradient between mixed venous and hepatic venous oxygen saturation lower than or equal to 10%. Group II: patients with gradient between mixed venous and hepatic venous oxygen saturation higher than 10%. Data are presented as mean ± SEM.

View larger version (25K): [in this window] [in a new window]   Figure 3.   Relationship between hepato-splanchnic blood flow and hepato-splanchnic oxygen extraction (see Figure 2 for details).

A difference between Sv_O2 and Sh_O2 greater than 10% represented a valuable cutoff value to predict covariance of hepato-splanchnic O₂ and DO₂ (sensitivity, 0.96; specificity, 0.80; positive predictive value, 89.7%; negative predictive value, 92.3%; proportion of correct classification, 0.91).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Systemic O₂/DO₂ dependency has been reported by some investigators in patients with sepsis (15, 23) and acute respiratory distress syndrome (ARDS) (24, 25), and challenged by others (26, 27). However, the study of whole body O₂/DO₂ relationships is crude since it cannot detect alterations in the distribution of blood flow and oxygen metabolism. There is a consensus among investigators that the study of regional beds is necessary (28, 29). The hepato-splanchnic area is particularly important because hepato-splanchnic dysoxia is considered as a cornerstone in the development of multiple organ failure (1) and yet is hardly recognized (29, 30). The present study demonstrated that covariance of O₂ and DO₂ in the hepato-splanchnic area is common in septic patients, and can be disclosed by dobutamine administration or PEEP application.

Dobutamine increased hepato-splanchnic blood flow in septic patients, even though the effects of dobutamine on hepato-splanchnic blood flow were not clearly dose-related, as a consequence of interindividual variations. This increase was accompanied by an increase in liver metabolism as reflected by increased O₂spla, splanchnic lactate consumption, and (to some extent) ICG clearance. After massive hepatectomy in dogs, Nonami and coworkers (31) observed that dobutamine increased liver DO₂, O₂, and lactate consumption. In a hyperdynamic endotoxic shock dog model, De Backer and coworkers (19) also observed that dobutamine similarly increased liver DO₂ and O₂ but these effects were not greater with 10 than with 5 µg/kg · min. Recently, Reinelt and coworkers (12) studied 12 patients with septic shock and observed that dobutamine administration increased hepato-splanchnic DO₂ from 141 ± 33 to 182 ± 44 ml/min · M² (p < 0.001). In the patients we studied, dobutamine improved the balance between oxygen supply and oxygen demand since Sh_O2 increased in parallel with the increase in hepato-splanchnic blood flow and metabolism.

The range of DSO₂ we observed was wide, from 10.2% to +42.7%. Several investigators have reported a similar range of DSO₂ in critically ill patients with (6, 13) or without (6, 14) sepsis. There are no data in the literature about the range of DSO₂ in normal individuals, but Sv_O2 and Sh_O2 are usually similar in nonseptic patients (6, 8). The DSO₂ values, calculated from individual data of Sv_O2 and Sh_O2 reported by Dahn and coworkers (6), had a standard deviation of 8.9%, so that we thought it was reasonable to choose a 10% cutoff value in baseline DSO₂ to separate the patients into two groups.

We observed that O₂spla increased only in the patients with a widened gradient between Sv_O2 and Sh_O2. We believe that the most likely explanation was the presence of regional oxygen supply dependency. The hepato-splanchnic O₂ extraction was increased at baseline, reflecting an imbalance between oxygen supply and demand in the splanchnic area. Various investigators have shown that liver metabolism increases during sepsis (4). In rats, Breuille and colleagues (5) noted that the contribution of the liver to protein synthesis increased from 15% in control conditions to 30% in sepsis. Dahn and coworkers (9) observed an increase in glucose output of 77% in septic patients as compared with nonseptic patients. In those septic patients, the hepato-splanchnic O₂ was related to glucose output. Albumin synthesis was also increased, reflecting enhanced hepatic function. In another group of septic patients, the increase in hepato-splanchnic O₂ was greater than the increase in DO₂ and was accompanied by a decrease in Sh_O2 (8). In an isolated liver model in which DO₂ was maintained, Dahn and coworkers (7) observed that the hepatic O₂ was increased during sepsis and, more importantly, that the response to a metabolic load was maintained. The latter finding argues against mitochondrial uncoupling, and rather suggests that inadequate blood flow can play an important role in the altered liver function commonly observed in sepsis.

An increase in hepato-splanchnic O₂ has been observed in some septic patients when hepato-splanchnic DO₂ is increased. In 1993, Ruokonen and colleagues (32) first reported an increase in hepato-splanchnic O₂ after the administration of dopamine and norepinephrine in 10 hemodynamically stable patients. More recently Steffes and colleagues (11) observed that hepato-splanchnic DO₂ increased by 26%, and hepato-splanchnic O₂ increased by 9%, in septic patients during red blood cell transfusion. However, splanchnic lactate consumption, which was relatively low at baseline in our study as compared with others (4, 9, 11), decreased suggesting the absence of hypoxia in this area. In our study, lactate consumption increased by 38% but the increase was similar in the patients who presented covariance of hepato-splanchnic O₂ and DO₂ and in the others. It is nevertheless difficult to accept that an increase in O₂ during O₂/DO₂ dependency associated with hypoxia must always be accompanied by an increase in lactate consumption. Indeed, the determinants of hepato-splanchnic lactate consumption are multiple. Lactate consumption is influenced not only by the amount of O₂ available for the conversion from lactate to pyruvate but also by the total amount of lactate delivered to the liver and thus the liver blood flow. Therefore, lactate consumption and O₂ could increase when blood flow is increased, even in the absence of hypoxia. In this study, lactate consumption by the liver was directly related to the arterial lactate levels. More importantly, hepato-splanchnic lactate consumption must equal the difference between lactate consumption by the liver and lactate production by the gut since most of the lactate is delivered to the liver via the portal vein. Dobutamine could have decreased gut lactate production and hence minimized the changes in hepato-splanchnic lactate consumption. In a canine endotoxic shock model, dopexamine, another -adrenergic agent, was found to decrease lactate production in the gut (33). Because portal vein lactate levels were not determined in these studies, it is very difficult to interpret hepato-splanchnic lactate consumption in humans.

The increase in O₂ associated with an increase in DO₂ in the hepato-splanchnic area may be due to the reversal of a reduced cellular metabolism in the presence of decreased oxygen supply, at least in some parts of the liver, by a phenomenon of oxygen conformance. In vitro studies by Schumacker and associates (34) demonstrated that O₂ and cellular metabolism can decrease in parallel to a significant decrease in PO₂. Because some nonessential cellular functions were temporarily withheld, severe hypoxia with PO₂ as low as 10 mm Hg could be tolerated without alterations in cellular integrity, whereas O₂ and the metabolism returned to baseline levels after restoration of normoxic conditions. Although it is quite difficult to demonstrate or exclude this phenomenon in vivo, the similar baseline hepato-splanchnic O₂ in both groups argues against this mechanism.

Another possibility was that the increase in O₂ was related to an increased supply of metabolically active nutrients when blood flow was increased. However, hepato-splanchnic O₂ significantly increased only in patients with a high Sv_O2-Sh_O2 gradient, despite the fact that hepato-splanchnic blood flow increased similarly in the two groups of patients. Furthermore, the increase in lactate consumption, reflecting the increased supply of lactate, was similar in both groups. The three phenomena of regional hypoxia, oxygen conformance, and increased supply of metabolically active nutrients do not necessarily exclude each other. They may even take place in different parts of the liver.

An increase in O₂ could also be due to a thermogenic effect of dobutamine. However, the slopes of the hepato-splanchnic O₂/DO₂ relationship, successively determined in 14 patients during dobutamine administration and after PEEP application, were similar with the two interventions. Also, Reinelt and coworkers (12) recently demonstrated that low-dose dobutamine did not increase neoglucogenesis in septic patients. In these patients hepato-splanchnic O₂ was unchanged despite an increase in blood flow. Finally, administration of a limited dose of dobutamine had similar effects on whole body O₂/DO₂ relationships to those of sodium nitroprusside in healthy volunteers (35) or prostacyclin in septic patients (17). Hence it is unlikely that the thermogenic effect played a significant role in the increase in hepato-splanchnic O₂.

An important question is whether mathematical coupling of data could also be responsible for the increase in O₂ (26). However, mathematical coupling of data could not account for the observation that O₂ increased in one group and not in the other one despite a similar increase in hepato-splanchnic blood flow. Also, mathematical coupling of data could only explain these results if there was a systematic underestimation of the hepato-splanchnic blood flow at baseline and a systematic overestimation of hepato-splanchnic blood flow during dobutamine administration, which seems unlikely. Because hepato-splanchnic DO₂ and O₂ cannot be determined independently, we also analyzed the hepato-splanchnic blood flow/hepato-splanchnic O₂ extraction relationship. This relation avoids the problems of mathematical coupling of data since the two variables are determined independently. Because hemoglobin and arterial saturation remained unchanged, O₂ isopleths could be defined as lines of equal product of hepato-splanchnic blood flow and hepato-splanchnic O₂ extraction. The observation that the increase in hepato-splanchnic blood flow was fully counterbalanced by a decrease in EO₂spla in patients in one group but less completely in patients in the other group supported the validity of our observations. Therefore, mathematical coupling of the data could not account for the altered O₂/DO₂ relationships observed in some septic patients.

Interestingly, the hepato-splanchnic O₂/DO₂ relationships could not be extrapolated from the analysis of systemic O₂/ DO₂ relationships. The evolution of cardiac index, whole body DO₂ and O₂ was similar in both groups. More importantly, the slope of the hepato-splanchnic O₂/DO₂ relationship was not correlated with the slope of the systemic one. On the contrary, the abnormal hepato-splanchnic O₂/DO₂ relationship was related to various indices of hepato-splanchnic dysoxia such as Sh_O2, EO₂spla, and DSO₂. The hepatic ICG clearance was also lower in patients with covariance of hepato-splanchnic O₂ and DO₂, suggesting depressed liver function although other biological markers of liver impairment were not different among groups. Thus, covariance of hepato-splanchnic O₂ and DO₂ could only be recognized by regional studies including hepatic vein catheterization. A DSO₂ cutoff value at 10% could reliably be used to predict covariance of hepato-splanchnic O₂ and DO₂. Of note, the PEEP level was higher in the patients who presented covariance of hepato-splanchnic O₂ and DO₂. The application of PEEP has been shown to decrease splanchnic blood flow in various animal (36) and human (37, 38) studies, and could therefore stress the balance between oxygen supply and demand in the hepato-splanchnic area at baseline.

Inevitably, a number of patients were treated concomitantly with other catecholamines (i.e., dopamine and norepinephrine) and this could have influenced the response to dobutamine. In particular, a different level of receptivity of adrenergic receptors may have influenced the cellular response to dobutamine. However, several findings suggest that the use of other catecholamines did not influence the response to dobutamine. First, there was no significant difference in catecholamine administration between the two groups and there was no significant relationship between the slope of the hepato-splanchnic O₂/ DO₂ relationship and the dose of the adrenergic agents. Next, the response in hepato-splanchnic DO₂ was similar in the patients treated and those not treated with other catecholamines. Finally, the slopes of the O₂/DO₂ relationship were similar with dobutamine administration and PEEP application in 14 patients who had the two interventions.

We were not able to relate the outcome of patients to the presence of covariance of hepato-splanchnic O₂ and DO₂ but this phenomenon could be transient, and outcome is related to a number of factors, including the underlying disease, the source of infection, and the development of complications. Two-thirds of the patients presented covariance of hepato-splanchnic O₂ and DO₂, which was quite a high proportion, but some patients already suffered from multiple organ failure, even though we investigated the patients early in their septic course.

We conclude that covariance of hepato-splanchnic O₂ and DO₂ may be present in many septic patients. Such a phenomenon may be caused by regional O₂/DO₂ dependency owing to tissue hypoxia or cellular O₂ conformance. This phenomenon could be suspected when the gradient between mixed venous and hepatic vein O₂ saturations is increased, and revealed by dobutamine administration or PEEP application. Dobutamine administration is more convenient since generally it does not compromise hepato-splanchnic blood flow.