The influence of thoracic epidural anesthesia on liver hemodynamics in patients under general anesthesia

Correspondence to D. Trepenaitis, Department of Anesthesiology, Kaunas University of Medicine, Eivenių 2, 50028 Kaunas, Lithuania. E-mail: trepenaitis@gmail.com Adresas susirašinėti: D. Trepenaitis, KMU Anesteziologijos klinika, Eivenių 2, 50028 Kaunas. El. paštas: trepenaitis@gmail.com


Summary. Background and objective. Hepatic hypoperfusion is regarded as an important factor in the pathophysiology of perioperative liver injury. Although thoracic epidural anesthesia is a widely used technique, limited data are available about the effects on hepatic blood flow with blockade restricted to thoracic segments in humans. The main objective of the present study was to investigate the effects of thoracic epidural anesthesia on hepatic blood flow under general anesthesia in humans.
Material and methods.In 40 patients under general anesthesia, we assessed hepatic blood flow using plasma disappearance rate of indocyanine green (PDR ICG ) as a simple noninvasive method before and after induction of thoracic epidural anesthesia.The epidural catheter was inserted at the Th 7/8 or Th 8/9 , and 1% lidocaine at a mean (range) dose of 8 (6)(7)(8)(9)(10)

mL was injected. Ephedrine bolus was given to patients who demonstrated a decrease in mean arterial blood pressure below 60 mm Hg after induction of thoracic epidural anesthesia (TEA-E group). Other patients did not receive any catecholamines during the study period (TEA group). Ten patients who did not undergo TEA served as controls (control group).
Results.In 7 patients, administration of ephedrine was necessary to avoid a decrease in mean arterial blood pressure below 60 mm Hg.Thus, the TEA-E group consisted of 7 patients and TEA group of 33.In the TEA group, thoracic epidural anesthesia was associated with a mean 2.3% min -1  decrease in PDR ICG (P<0.05).In the TEA-E group, all seven patients showed a 2.2% min -1 decrease in PDR ICG (P<0.05).Patients in the control group showed a mean 1.1% min -1 increase in PDR ICG (P<0.05).In contrast to hepatic blood flow, cardiac output was not affected by thoracic epidural anesthesia.
Conclusions.In humans, thoracic epidural anesthesia is associated with a decrease in hepatic blood flow.Thoracic epidural anesthesia combined with ephedrine bolus was found to result in further decrease in hepatic blood flow.fects on hepatic blood fl ow.For example, blockade of thoracic efferent sympathetic fi bers causes regional arteriolar dilation and thus increases regional blood fl ow.On the other hand, hypotension as a result of a decreased systemic vascular resistance may reduce hepatic blood fl ow.Lumbar epidural anesthesia has been shown to induce a decrease in hepatic blood fl ow in animals (3,4) and in humans (5,6).In contrast to these studies, Vagts et al. (7), using a pig model, did not observe alterations in total hepatic blood fl ow after induction of thoracic epidural anesthesia despite a signifi cant decrease in systemic arterial blood pressure.Thus, we hypothesized that thoracic epidural anesthesia maintains hepatic blood fl ow also in humans.
The main objective of the present study was to investigate the effects of TEA on hepatic blood fl ow

Introduction
Patients undergoing major abdominal surgery are at risk of developing postoperative complications including multiple organ failure.Hepatic hypoperfusion is an important factor in the pathophysiology of perioperative liver injury (1).Furthermore, it has been hypothesized that hypoperfusion of the liver may initiate or contribute to the development of a systemic infl ammatory response syndrome (2).
For decades, there have been discussions about whether or not thoracic epidural anesthesia (TEA) is able to reduce perioperative risk.Although thoracic epidural anesthesia and analgesia is widely used technique, the effects on hepatic blood fl ow with blockade restricted to thoracic segments have not been investigated extensively.TEA may induce different physiological reactions with confl icting ef-under general anesthesia in humans.Plasma disappearance rate of indocyanine green (PDR ICG ) has been evaluated as a marker of hepatic perfusion and hepatocellular function, and its prognostic value in the critically ill patient has been demonstrated (8)(9).

Material and methods
Participants.The protocol of this prospective study was reviewed and approved by the Kaunas Regional Ethics Committee for Biomedical Research.This study was designed with the aim to enroll 50 patients consecutively: 40 patients who underwent TEA and 10 patients who did not undergo TEA (control group).The control group consisted of patients with contraindications for TEA, in whom the insertion of a epidural catheter was not possible.We aimed at inclusion of 50 ASA physical status II or III patients with a mean age of 64.94 years (range, 33 to 80 years), who underwent elective upper abdominal interventions for carcinoma of the stomach, papilla of Vater, and pancreas.The informed consent was obtained from all the patients.
The abdominal interventions were carried out under combined general-epidural anesthesia.Patients with a history of cardiac or vascular disease, including hypertension, were excluded from the study.Hypertension was defi ned as a mean diastolic blood pressure of more than 100 mm Hg or a systolic blood pressure of more than 160 mm Hg or both at three measurements.Patients taking antihypertensive drugs, including diuretics, as well as antiarrhythmics or α 1 -sympathicolytic medications for benign prostate hypertrophy were excluded.Furthermore, patients who had a history of known hypersensitivity to amide local anesthetics; severe respiratory, renal, or hepatic disease; diabetes mellitus; or neurological, psychiatric, or seizure disorders were excluded.In addition, patients with positive results of Allen's test (no collateral fl ow in the ulnar artery) at both upper extremities were excluded.
Preparation.Patients fasted from midnight.They received 7.5 mg of midazolam orally approximately 45-60 min before the procedure.After the patient was connected to standard equipment to monitor the blood pressure noninvasively and for continuous registration of heart rate and rhythm, an intravenous cannula was inserted into a large vein in the forearm or the antecubital fossa for fl uid or drug administration.
Subsequently, after local infi ltration of the skin with 1.0% lidocaine, the epidural space was accessed under sterile conditions with a 16-gauge Tuohy needle at the Th 7/8 or Th 8/9 interspace with a patient in the sitting position.Using a median or paramedian approach, the epidural space was identifi ed using the loss of resistance to air technique.An epidural catheter was inserted 5 cm into the epidural space.After excluding an intravascular or arachnoid positioning of the catheter by aspiration, a bacterial filter was attached, and the catheter was capped sterilely.Thereafter, the patient was placed again in a supine horizontal position.Standardized general anesthesia was in duced with 3-5 mg/kg of sodium thiopental and 2.5 μg/kg of fentanyl.Rocuronium (0.5 mg/kg) was administered to facilitate tracheal intubation.Anesthesia was maintained with sevofl urane and fentanyl as required.Ventilation was mechanically controlled to maintain end-tidal carbon dioxide tension near 35 mm Hg.After the induction of anesthesia, a 20-gauge arterial cannula was inserted into a radial artery, preferentially in the nondominant arm, for invasive blood pressure monitoring.A central venous line was placed via the right internal jugular vein for central venous pressure monitoring.Patients were actively warmed with forced air on the lower body to maintain intraoperative normothermia (Bair Hugger, Augustine Medical, Eden Prairie, MN).Liver hemodynamics was measured noninvasively using LiMON monitor and fi ngertip sensor (PULSION Medical Systems AG, Munich, Germany), which allows for noninvasive measurement of the blood concentration-time curve of indocyanine green (ICG).ICG was used as an indicator.
Experimental procedure.The experiment consisted of four parts.The fi rst part of the experiment was aimed at obtaining baseline values.Before baseline measurements, all patients received at least 500 mL of crystalloids.Depending on the mean arterial pressure (MAP), further fl uids were administered with the aim to maintain MAP above 60 mm Hg.After complete preparation for the experiment, a 10-min stabilization period was observed.
The second part of the experiment was aimed at measuring hepatic blood fl ow before epidural anesthesia.Patients received a rapid intravenously (≤2 s) injected bolus administration of ICG (0.25 mg/kg body weight), followed by measurement of plasma disappearance (PDR) and retention 15 (R15) rate of ICG using the pulse densitometer.Simultaneously, heart rate, invasive arterial blood pressure, central venous pressure (CVP), cardiac output (CO), and systemic vascular resistance index (SVRI) were continuously recorded.
The third part of the experiment was aimed at assessing the effects of epidural anesthesia on liver hemodynamics.Using the epidural catheter that had been installed before the experiment, 1% lidocaine was epidurally administered at a rate of 1 mL • s -1 .Epidural bolus dose of lidocaine was calculated according to the patient's age: 20-40 years, 1.5 mL/ segment; 40-60 years, 1.3 mL/segment; 60-80 years, 1.0 mL/segment; and more than 80 years, 0.7 mL/segment (10).We aimed to achieve epidural block from Th 5 to Th 12. Fluid administration was continued with approximately 5 mL•kg -1 •h -1 .Hypotension (decrease in SBP >30% of the pre-experimental value or SBP <90 mm Hg) was treated by administering the rapid intravenous infusion of a colloid solution (6% hydroxyethyl starch; Voluven, Fresenius Kabi, Bad Homburg, Germany).After 30 min, the second measurement of PDR ICG and R15 was carried out as described above.A PDR ICG of 18-25% min -1 is considered to be normal.Like all dynamic liver tests, PDR ICG is dependent on sinusoidal perfusion as well as hepatocellular function.While higher values can be observed in hyperdynamic states, lower values are due to acute or chronic hepatocellular dysfunction or due to macrocirculatory or microcirculatory perfusion disorders leading to reduced overall sinusoidal perfusion (11).Central hemodynamic parameters were recorded simultaneously.If MAP decreased below 60 mm Hg within a period of 15 min after injection of lidocaine, 5-mg ephedrine was administered intravenously.Waiting periods of 15 min after a bolus administration of ephedrine were observed before ICG was administered to measure liver hemodynamics.Patients who received ephedrine were separately analyzed as the TEA-E subgroup (TEA-E group).
The fourth part of the experiment was aimed at assessing central hemodynamic parameters (CVP, CO, and SVRI) 1 hour after the epidural bolus.
All measurements were performed before surgery.The function of the epidural catheter was checked by pinprick evaluation of sensory blockade on the fi rst postoperative day.
Hemodynamic assessment.The arterial blood pressure and CVP were monitored throughout the study period using a pressure transducer (MEMSCAP AS, Skoppum, Norway).Further hemodynamic values (CO and SVRI) were assessed using an impedance cardiography monitor (BioZ CardioDynamics International Corp., San Diego, CA, USA).
Statistical analysis.Measured parameters were analyzed using the SPSS for Windows 13.1 (SPSS Inc., Chicago, Illinois, USA).An a priori power analysis was conducted indicating that, given the scatter of the data, a sample size of 40 would be suffi cient to detect a 4% min -1 difference in PDR ICG that is a 20% improvement or deterioration of a normal value of about 20% min -1 with a power of 0.9 and an α error of 5%.All data were expressed as means±SD.When normal distribution of the variables could not be assumed, we compared signifi cance level of Student's and Wilcoxon's tests.Because signifi cance level of these tests was equal, we presented our results like parametric variables.Intragroup changes in hemodynamic variables were analyzed using the Wilcoxon's signed rank test (P 1 value).The statistical signifi cance level was set at 0.05.

Results
Fifty patients were included in this study.No patient was excluded because of epidural catheter dislocation, because on the fi rst postoperative day, a correct catheter position in the epidural space could be confi rmed in all patients.After induction of TEA, seven patients demonstrated a decrease in MAP below 60 mm Hg, so ephedrine bolus was administered (TEA-E group).The remaining 43 patients did not The infl uence of thoracic epidural anesthesia on liver hemodynamics in patients under general anesthesia receive any catecholamines during the study period.There were no signifi cant differences comparing the groups except gender differences (Table 1).
The volume of lidocaine was 8 (6-10) mL both in the TEA and TEA-E groups.There were no signifi cant differences in end-tidal sevofl urane and CO 2 concentrations and amounts of administered fl uids between the two groups.PDR ICG showed a signifi cant (P=0.03)decrease after induction of TEA, but not below the normal value (18% min -1 ).This difference was similar in the TEA and TEA-E groups (Fig. ).In the control group, PDR ICG showed an increase after 30 min following hemodynamic stabilization.TEA was associated with a signifi cant decrease in HR and MAP.TEA was not associated with a signifi cant change in CO and SVRI in the TEA or TEA-E groups (Table 2).

Discussion
In the present study, we have investigated the infl uence of TEA on PDR ICG as a surrogate of liver blood fl ow in patients under general anesthesia.We used PDR ICG as a simple noninvasive method to assess hepatic perfusion.ICG binds avidly to plasma proteins, so that its distribution volume equals the blood plasma volume.ICG is exclusively eliminated from blood by the liver with a half-life of 150-180 s and without enterohepatic circulation (12).The elimination of ICG strongly correlates with hepatic blood fl ow, but is not similar to hepatic fl ow as the extraction ratio may differ among patients and clinical conditions (13).As a matter of principle, both liver blood fl ow and hepatocellular function affect dynamic liver tests such as PDR ICG ; however, rapid changes that was within 30 min as in our study can be explained almost exclusively by changes in sinusoidal blood fl ow (11).
Hepatic blood fl ow in surgical patients can be altered by a variety of factors, including arterial blood pressure, posture changes, carbon dioxide levels, intravascular volume shifts, positive pressure ventilation, and volatile anesthetics (1).In the present study, measurements were performed under general anesthesia and mechanical ventilation; hence, hepatic blood fl ow was presumably already decreased during baseline measurements.Many of the factors affecting hepatic blood fl ow, such as carbon dioxide levels, end-tidal sevofl urane concentration, and mode of ventilation remained constant during this investigation.All measurements were performed before surgery to exclude the infl uence of intraoperative surgical manipulations.Fluid administration was continued during the study period, so there was no decrease in intravascular volume.Measurements were started when patients were hemodynamically stable according to blood pressure and heart rate.In contrast to our hypothesis, we found a consistent decrease in PDR ICG as a marker of regional hepatic perfusion after induction of TEA.The combination of TEA with ephedrine (TEA-E group) did not improve liver blood fl ow.However, cardiac output was not signifi cantly affected by TEA.Presumably, the TEA-related decrease in systemic arterial blood pressure is the main reason for the observed decrease in hepatic blood fl ow.The negative effects of the decrease in systemic arterial blood pressure prevail over the anticipated positive effects of local vasodilation.One reason for the arterial blood pressure effect of hepatic blood fl ow regulation might be that the hepatic artery exhibits almost no autoregulatory capacity, as a nearly linear correlation between systemic arterial blood pressure and hepatic arterial blood fl ow has been found (14,15).
In previous studies, there had been confl icting results on the infl uence of EA on parameters of splanchnic perfusion, most of them using mucosal pH and arterial-mucosal p CO2 difference as surrogate parameters.This might be due to different therapy regimens, patient population, or insertion sites of epidural catheters or simply point to the limitations of gastrointestinal tonometry (16,17).The effects of TEA on hepatic blood fl ow have been investigated in only one animal study and two studies in humans.In contrast to the present study, Vagts et al. (7), using a pig model, observed an unaltered total hepatic blood fl ow after induction of TEA (Th 5 -Th 12 ) despite a 30% decrease in MAP.The reasons for the discrepancy between the data found in pigs and our fi ndings in humans may be species-related and could be due to different physiological reactions on TEA-induced sympathicolysis.Kortgen et al. (11) found that liver perfusion was increased with thoracic but not lumbar EA after major abdominal surgery in most patients.The study by Meierhenrich et al. (18) found a signifi cant decrease in hepatic venous blood fl ow after induction of TEA using multiplane transesophageal echocardiography.The combination of TEA with continuous infusion of norepinephrine seems to induce a further decrease in hepatic blood fl ow.
Three human studies have been performed to evaluate the effects of high lumbar epidural anesthesia (LEA) on hepatic blood fl ow.Kennedy et al. (5) measured hepatic venous fl ow in conscious volunteers by the ICG clearance technique using a hepatic venous catheter and, after induction of high LEA, found a 25% decrease in hepatic blood fl ow despite a constant cardiac output.The infl uence of thoracic epidural anesthesia on liver hemodynamics in patients under general anesthesia fl ow using pulse dye densitometry (PDD), which allows for noninvasive measurement of the blood concentration of ICG.They found a decrease in liver blood fl ow after LEA.We performed TEA and assessed the changes in hepatic blood fl ow by the use of noninvasive ICG clearance test.Nevertheless, we found almost the same reduction in hepatic blood fl ow.Thus, the present data do not support the hypothesis that TEA is superior to LEA with respect to the effects on hepatic perfusion.All 7 patients (TEA-E group) treated with ephedrine to compensate for the TEA-induced decrease in arterial blood pressure showed any improvement in PDR ICG and did not differ from the main TEA group.Neither human nor experimental studies have been performed to investigate the effects of ephedrine in combination with TEA on hepatic blood fl ow.Experimental studies assessed the sole effect of norepinephrine on hepatic blood fl ow and consistently found a decrease in fl ow to be between 20% and 45% (20)(21)(22).The present data do not suggest that TEA attenuates ephe-drine-induced vasoconstriction in the splanchnic bed.
We cannot draw any conclusion as to what extent the observed decrease in hepatic blood fl ow was associated with alterations in hepatic oxygen balance and metabolic activity.However, the decrease might be relevant in specifi c patients' groups, such as those with pre-existing liver disease or those undergoing hepatic surgery.Outcome studies should be performed to evaluate TEA with respect to possible perioperative hepatic damage, especially in these patients' groups.In addition, the intraoperative use of ephedrine in combination with TEA should be investigated in outcome studies and compared with over vasopressor therapies.

Conclusions
In humans, thoracic epidural anesthesia is associated with a decrease in hepatic blood fl ow.Thoracic epidural anesthesia combined with ephedrine bolus was found to result in further decrease in hepatic blood fl ow.
Fig. Effects of thoracic epidural anesthesia on hepatic blood fl owGroup

Table 2 .
Tanaka et al. (6) also performed high LEA.They estimated hepatic blood fl ow by pulse densitometric determination of the PDR ICG and reported a 35% decrease in hepatic blood fl ow.Simon et al. (19) measured liver blood Effects of thoracic epidural anesthesia on global hemodynamic variables 1 1Kauno medicinos universiteto Anesteziologijos klinika,2Kauno medicinos universiteto Chirurgijos klinika