Secondary Logo

Journal Logo

Original Articles – General

Thoracic but not lumbar epidural anaesthesia increases liver blood flow after major abdominal surgery

Kortgen, Andreasa; Silomon, Malteb; Pape-Becker, Christineb; Buchinger, Heikob; Grundmann, Ulrichb; Bauer, Michaela

Author Information
European Journal of Anaesthesiology: February 2009 - Volume 26 - Issue 2 - p 111-116
doi: 10.1097/EJA.0b013e32831c8939



Patients undergoing major abdominal surgery are at risk of developing postoperative complications including multiple organ failure. For decades, there have been discussions about whether or not epidural anaesthesia and analgesia (EAA) is able to reduce perioperative risk. Many investigations have emphasized beneficial effects, for example a decrease in intraoperative blood loss [1], a decrease in the incidence of thromboembolic events [2,3], improved postoperative respiratory [3–5], cardiovascular [3,4] and intestinal function [5,6], and modified immune function [7]. In the intensive care setting, EAA has been proposed to reduce morbidity and mortality in acute pancreatitis, suggesting an increase in splanchnic blood flow as the underlying mechanism [8].

The splanchnic region seems to be of extraordinary importance in patients undergoing major surgery [9] and in the critically ill [10,11]. More specifically, perioperative or peritraumatic splanchnic ischaemia is associated with impaired outcome [12]. Thus, improving blood flow and oxygen supply to the gastrointestinal tract and the liver has the potential to improve patient outcome. Beneficial influence of EAA on bowel function has been demonstrated [13] and has led to its use in fast track concepts [14]. Sympathetic blockade of the splanchnic arteries may increase organ perfusion; however, there is no coherent view of the effects of EAA on splanchnic perfusion. For this reason, lumbar as well as thoracic EAA have been investigated in various studies with conflicting results [15–18]. Most of these studies applied pHi as the surrogate of gastric and intestinal perfusion.

Monitoring of regional organ perfusion is crucial in patients at risk and has been highly recommended by the American College of Critical Care Medicine/Society of Critical Care Medicine task force [19]. Plasma disappearance rate of indocyanine green (PDRICG) has been evaluated as a marker of hepatic perfusion and hepatocellular function and its prognostic value in the critically ill patient has been demonstrated [20,21].

Confounding factors such as insertion site of the epidural catheter (lumbar vs. thoracic), changes in cardiac preload, heart rate and blood pressure are of great importance for the influence of EAA on splanchnic perfusion. Especially the insertion site causing sympathetic blockade in dependent vascular beds might be of extraordinary interest according to their influence on splanchnic perfusion. We, therefore, compared thoracic and lumbar EAA according to their differential influence on PDRICG as a parameter of hepatic perfusion and on blood lactate and central venous oxygen saturation as parameters of global oxygen supply/demand ratio.


After institutional review board approval, the prospective study was conducted on an operative ICU. Investigations were made on 34 ASA physical status II or III patients admitted to the ICU postoperatively after scheduled major abdominal surgery. All patients were still on ventilatory support when transferred to the ICU. After written informed consent, 17 of them received preoperatively a thoracic and 17 a lumbar epidural catheter for postoperative analgesia. Catheters were inserted between interspaces T5/6 and T9/10 (thoracic EAA) and between L1/2 and L4/5 (lumbar EAA), respectively. Catheters were inserted 3 cm into the epidural space. Sites of catheter insertion were chosen in order to obtain sufficient postoperative analgesia for the planned operation site according to routine practice. Operations in the thoracic EAA group were oesophagectomy (six patients), gastrectomy (three), Whipple's operation (four), hemicolectomy (two) and liver resection (two), and in the lumbar EAA group cystectomy (five), rectal resection (four), sigmoidectomy (four), gynaecologic operations (two), melanoma resection of the leg (one) and vascular surgery of the leg (one). A test dose of 3 ml bupivacaine 0.5% followed by 10 μg sufentanil was given before induction of general anaesthesia. There was no further application of local anaesthetics or opioids through the epidural catheters before ICU admission, in order to exclude any effect of local anaesthetics at the first measurements.

After admission to the ICU, local anaesthetics were given through the epidural catheters and the changes in global and regional perfusion parameters were assessed while patients were still sedated with propofol. Propofol infusion rates were not altered throughout the study period to exclude influence on measurements. Patients with a thoracic epidural catheter received a bolus of 10 ml bupivacaine 0.25% immediately followed by an infusion of 8 ml h−1 bupivacaine 0.175%. In patients with a lumbar epidural catheter, the bolus of 12 ml bupivacaine 0.25% was followed by an infusion of 10 ml h−1 bupivacaine 0.125%. The solutions for continuous epidural infusion contained 2 μg ml−1 fentanyl as well, in order to obtain sufficient analgesia. Bolus administration, infusion rates and concentrations were according to standard clinical practice at our institution as higher doses in lumbar EAA are needed to achieve the equivalent expansion of EAA due to the lumbar epidural space being broader than the thoracic epidural space. A lower concentration of bupivacaine for continuous epidural infusion in lumbar EAA was used to minimize motor blockade; however, the studied effect of EAA on liver blood flow is mediated by sympathetic blockade on the vascular bed and is achieved by either concentration, as sympathetic nerves are blocked by low concentrations of local anaesthetics. In addition, the initial bolus concentration of bupivacaine was 0.25% in both groups.

PDRICG was measured noninvasively using the LiMON-Monitor and a fingertip sensor (Pulsion Medical Systems, Munich, Germany) before and 2 h after bolus application of bupivacaine. A dose of 0.25 mg kg−1 ICG was administered intravenously through a central venous catheter. A PDRICG of 18–25 % min−1 is considered to be normal. Like all dynamic liver tests, PDRICG 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. At the same time points, heart rate, blood pressure, central venous oxygen saturation (ScvO2) and blood lactate were documented. For the two latter parameters, blood samples were drawn through a central venous catheter and measured using a standard blood gas analyser (Stat Profile 5, Nova Biomedical, Rödermark, Germany). Volume was administered to keep central venous pressure constant in order to minimize haemodynamic changes as a cause for altered liver perfusion.

Statistical analysis

An a-priori power analysis was conducted indicating that, given the scatter of the data, a sample size of at least 14 would be sufficient to detect a difference of PDRICG of 4 % min−1 – that is, a 20% improvement or deterioration of a normal value of about 20 % min−1 – with a power of 0.8 and an α-error of 5%. Data are expressed as means ± SD. The (paired) t-test, Wilcoxon signed rank test or Mann–Whitney rank sum test was carried out to compare values before and after EAA or thoracic with lumbar EAA and to compare patient, preoperative, intraoperative and postoperative factors, respectively. For categorial variables, the chi-squared test was used. Correlational analyses were performed with Pearson product moment correlation or Spearman rank order correlation. Sigma Stat for Windows 2.03 and SPSS for Windows 13.0 (SPSS Inc., Chicago, Illinois, USA) were used for analysis. P values less than 0.05 were considered significant.


Patients' characteristics and perioperative data are listed in Table 1. Patients with lumbar EAA were older and had a higher BMI than patients receiving thoracic EAA. There were no differences in laboratory parameters, fluid administration, duration of surgery and postoperative blood loss. Three patients in each group received transfusions of packed red blood cells. None of the 34 patients studied was excluded due to altered requirements in noradrenaline or propofol administration during the observation period. Three patients in the thoracic EAA group received noradrenaline (0.01, 0.02 and 0.2 μg kg−1 min−1) and one patient in the lumbar EAA group (0.12 μg kg−1 min−1). Infusion rates in these patients were not altered between the two measurements. Propofol infusion rates were 1.7 ± 0.67 mg kg−1 h−1 (thoracic EAA) and 1.5 ± 0.58 mg kg−1 h−1 (lumbar EAA) and were kept stable during the observation period. Neither noradrenaline nor propofol infusion rates were significantly different between the two groups. No adverse events occurred during the study period and the patients showed no signs of organ dysfunction. In both groups, an average amount of 500–1000 ml crystalloid (Sterofundin, B. Braun Melsungen, Melsungen, Germany) and up to 500 ml colloid solution (HES 130/0,4; Voluven, Fresenius Kabi, Bad Homburg, Germany) was administered intravenously in the first 2 h after epidural application of local anaesthetics. No significant changes in mean arterial pressure, heart rate and central venous pressure were detected in either group (Table 2). Although, after extubation, spreading of EAA could not be reliably tested due to the low anaesthetic concentrations used, analgesic effect of EAA was sufficient to control postoperative pain and was comparable between the two groups as need for additional opioid therapy was similar within the first 24 h (Table 1).

Table 1
Table 1:
Patients' characteristics and perioperative data
Table 2
Table 2:
Clinical parameters at baseline and with epidural anaesthesia and analgesia

PDRICG increased significantly upon thoracic EAA (Fig. 1a). Although the order of magnitude of increases in PDRICG was low, this was a very consistent observation. No significant changes could be detected in blood lactate levels and central venous oxygen saturation (Table 2) as surrogates for global haemodynamics.

Fig. 1
Fig. 1

A different picture was observed for lumbar EAA: PDRICG showed a slight but not significant decrease 2 h after starting lumbar EAA for the group mean; however, this minor change for the overall cohort masked fairly substantial changes in individual patients (Fig. 1b). Blood lactate levels and central venous oxygen saturation did not change after starting lumbar EAA, similar to what was observed in patients receiving thoracic EAA (Table 2).

A significant difference was found in changes of PDRICG between lumbar and thoracic EAA (Fig. 2). Correlations between PDRICG, blood lactate level and ScvO2, as well as between changes in PDRICG and blood lactate level, ScvO2, heart rate, central venous pressure or mean arterial pressure, could not be observed.

Fig. 2
Fig. 2

While patients with a decrease in PDRICG showed no significant differences in basic or measured perioperative and haemodynamic data compared with patients with an increase of PDRICG, more patients with a decrease received lumbar EAA.


In the present study, we have investigated the influence of thoracic and lumbar EAA on PDRICG as a surrogate of liver blood flow in the postoperative course after major abdominal surgery. We chose the postoperative setting to exclude influences of intraoperative surgical manipulations on intestinal and liver blood flow. To further control confounding factors, patients were sedated without changes in propofol administration and were ventilated throughout the study period. In addition, central venous pressure as a marker of cardiac preload was kept stable and noradrenaline infusions were not changed either. Measurements were started when patients were haemodynamically stable according to blood pressure and heart rate. Thoracic EAA induced a consistent rise in PDRICG as a marker of regional hepatic perfusion in this setting, whereas lumbar EAA resulted in an overall decrease in PDRICG with a high interindividual variability.

In former studies, there had been conflicting results on the influence of EAA on parameters of splanchnic perfusion, most of them using mucosal pH and arterial–mucosal pCO2 difference as surrogate parameters. This might be due to different therapy regimens, patient populations or insertion sites of epidural catheters, or simply point to the limitations of gastrointestinal tonometry [22,23]. In one study, EAA did not improve either gastric–arterial or sigmoid–arterial pCO2 differences. It also did not improve arterial–gastric or arterial–sigmoid pH gaps in patients after aortic reconstruction surgery in comparison with a control group without EAA. Hepatic lactate uptake was higher in the EAA group; splanchnic blood flow was not affected. Epidural catheters were inserted between T12 and L1. As a confounding factor, the need for sodium nitroprusside to control blood pressure was higher in the control group [18]. Spackman et al.[16] demonstrated improved gastric–arterial pCO2 difference and small bowel motility in 21 patients with clinical signs of peritonitis and paralytic ileus, when low thoracic or high lumbar EAA was performed in comparison with intravenous morphine. With lumbar EAA in patients undergoing aortic surgery, gastric pHi and the pCO2 gap were not influenced [15], whereas, in another study, thoracic EAA showed a significantly lower decrease in gastric pHi during major abdominal surgery [17]; however, in the latter study, cardiac preload was not controlled and an increase in heart rate was detected that might consecutively cause an increase in cardiac output and thereby in splanchnic perfusion.

In our study, we used PDRICG as a simple noninvasive method to assess splanchnic perfusion that has demonstrated a prognostic value in critically ill patients [20,21]. As a matter of principle, both liver blood flow and hepatocellular function affect dynamic liver tests such as PDRICG; however, rapid changes, that is within 2 h, as in our study, can be explained almost exclusively by changes in sinusoidal blood flow.

Our data lend support to the concept that PDRICG is a simple tool to assess EAA-induced net changes in splanchnic perfusion, a critical vascular bed. Changes in regional perfusion after epidural injection can result from reduced sympathetic activity and vasodilation within the blocked area, whereas, in contrast, a compensatory increased sympathetic activity outside the epidural blockade may lead to reduced regional perfusion through vasoconstriction; however, these changes in regional vascular resistance can be overimposed by alterations in global perfusion, as sympathetic blockade with reduced preload can cause a decrease in cardiac output and variations in heart rate may result equally in similarly directed changes of cardiac output, and they can be affected by administration of vasoactive substances. To reduce preload-induced changes in global perfusion and thereby splanchnic perfusion, central venous pressure was kept stable by volume loading. In none of the patients did vasoactive drugs have to be used or changed as a result of this protocol; however, only a few patients received low-dose noradrenaline during the study period and a correlation between changes in PDRICG and heart rate, central venous pressure, or central venous oxygen saturation were not found. Of course, these are only crude cardiocirculatory parameters that cannot rule out changes in cardiac index, but they represent routinely assessed parameters in stable patients in the perioperative setting, in whom more invasive monitoring is doubtful.

Our data are therefore consistent with a differential influence of lumbar and thoracic EAA on liver blood flow, even if confounding factors such as changes in crude global haemodynamic parameters are excluded. While thoracic EAA increased PDRICG in almost every patient, results in the lumbar EAA group were highly variable. This variability might be due to a different expansion of lumbar EAA, in some cases with additional splanchnic sympathetic blockade and sometimes with a rather compensatory increase in splanchnic sympathetic activity. Unfortunately, owing to the setting of the present investigation with patients still receiving sedatives, expansion of EAA could not be assessed.

In conclusion, thoracic EAA can increase PDRICG as a marker of splanchnic perfusion and liver blood flow postoperatively after major abdominal surgery in most cases, whereas lumbar EAA does not have a consistent net effect. Monitoring PDRICG allows for individual assessment of influence of EAA on hepatic perfusion in this patient population.


The study was supported solely by funding from the Department of Anaesthesiology and Intensive Care Medicine. Michael Bauer is member of the Medical Advisory Board of Pulsion Medical Systems AG, Munich, Germany. Michael Bauer and Andreas Kortgen received honoraria from Pulsion Medical Systems for educational talks.

The study was performed at the Department of Anaesthesiology and Intensive Care Medicine, University of Saarland, Homburg/Saar, Germany.


1 Mauermann WJ, Shilling AM, Zuo ZZ. A comparison of neuroaxial block versus general anesthesia for elective total hip replacement: a meta-analysis. Aneath Analg 2006; 103:1018–1025.
2 Modig J, Maripuu E, Sahlstedt B. Thromboembolism following total hip replacement: a prospective investigation of 94 patients with emphasis on the efficacy of lumbar epidural anesthesia in prophylaxis. Reg Anesth 1986; 11:72–79.
3 Tuman KJ, McCarthy RJ, March RJ, et al. Effects of epidural anesthesia and analgesia on coagulation and outcome after major vascular surgery. Anesth Analg 1991; 73:696–704.
4 Yeager MP, Glass DD, Neff RK, Brinck-Johnsen T. Epidural anesthesia and analgesia in high-risk surgical patients. Anesthesiology 1987; 66:729–736.
5 Liu SS, Wu CL. Effect of postoperative analgesia on major postoperative complications: a systematic update of the evidence. Anesth Analg 2007; 104:689–702.
6 Jansen M, Fass J, Tittel A, et al. Influence of postoperative epidural analgesia with bupivacaine on intestinal motility, transit time, and anastomotic healing. World J Surg 2002; 26:303–306.
7 Beilin B, Shavit Y, Trabekin E, et al. The effects of postoperative pain management on immune response to surgery. Anesth Analg 2003; 97:822–827.
8 Bernhardt A, Kortgen A, Niesel HC, Goertz A. Using epidural anaesthesia in patients with acute pancreatitis – a prospective investigation in 121 patients [in German]. Anaesthesiol Reanimat 2002; 27:16–22.
9 Soong CV, Halliday MI, Barclay GR, et al. Intramucosal acidosis and systemic host responses in abdominal aortic aneurysm surgery. Crit Care Med 1997; 25:1472–1479.
10 Hawker F. Liver dysfunction in critical illness. Anaesth Intens Care 1991; 19:165–181.
11 Kramer L, Jordan B, Druml W, et al, Austrian epidemiologic study on intensive care, ASDI study group. Incidence and prognosis of early hepatic dysfunction in critically ill patients a prospective multicenter study. Crit Care Med 2007; 35:1099–1104.
12 Meregalli A, Oliveira PO, Friedman G. Occult hypoperfusion is associated with increased mortality in hemodynamically stable, high-risk, surgical patients. Crit Care 2004; 8:R60–R65.
13 Carli F, Mayo N, Klubien K, et al. Epidural analgesia enhances functional exercise capacity and health-related quality of life after colonic surgery: results of a randomized trial. Anesthesiology 2002; 97:540–549.
14 Jakobsen DH, Sonne E, Anreasen J, Kehlet H. Convalescence after colonic surgery with fast-track vs conventional care. Colorectal Dis 2006; 8:683–687.
15 Piper SN, Boldt J, Schmidt CC, et al. Hemodynamics, intramucosal pH and regulators of circulation during perioperative epidural analgesia. Can J Anesth 2000; 47:631–637.
16 Spackman DR, McLeod ADM, Prineas SN, et al. Effect of epidural blockade on indicators of splanchnic perfusion and gut function in critically ill patients with peritonitis: a randomised comparison of epidural bupivacaine with systemic morphine. Intensive Care Med 2000; 26:1638–1645.
17 Kapral S, Gollmann G, Bachmann D, et al. The effects of thoracic epidural anesthesia on intraoperative visceral perfusion and metabolism. Anesth Analg 1999; 88:402–406.
18 Väisänen O, Parviainen I, Ruokonen E, et al. Epidural analgesia with bupivacaine does not improve splanchnic tissue perfusion after aortic reconstruction surgery. Br J Anaesth 1998; 81:893–898.
19 Task Force of the American College of Critical Care Medicine, Society of Critical Care Medicine. Practice parameters for hemodynamic support of sepsis in adult patients in sepsis. Crit Care Med 1999; 27:639–660.
20 Sakka SG, Reinhart K, Meier-Hellmann A. Prognostic value of the indocyanine green plasma disappearance rate in the critically ill patients. Chest 2002; 122:1715–1720.
21 Kimura S, Yoshioka T, Shibuya M, et al. Indocyanine green elimination rate detects hepatocellular dysfunction early in septic shock and correlates with survival. Crit Care Med 2001; 29:1159–1163.
22 Bams JL, Mariani MA, Groeneveld ABJ. Predicting outcome after cardiac surgery: comparison of global haemodynamic and tonometric variables. Br J Anaesth 1999; 82:33–37.
23 Hamilton MA, Mythen MG. Gastric tonometry: where do we stand? Curr Opin Crit Care 2001; 7:122–127.

central venous oxygen saturation; epidural anaesthesia; indocyanine green; liver; lumbar; major abdominal; regional perfusion; surgery; thoracic

© 2009 European Society of Anaesthesiology