Epidural Blockade Modifies Perioperative Glucose Production without Affecting Protein Catabolism

With the approval of the Ethics Committee of the Royal Victoria Hospital, 16 patients undergoing elective colorectal surgery for nonmetastatic carcinoma were admitted to the study. Written informed consent was obtained from all patients. All patients had a body mass index (BMI) between 20 and 27 kg/m²and maintained their body weight during the preceding 3 months (< 5% weight loss). Exclusion criteria were any cardiac, hepatic, renal, endocrine, or metabolic disorders, ingestion of any medication known to affect metabolism, and history of severe sciatica or back surgery that contraindicates the use of epidural catheters. We further excluded patients with a plasma albumin concentration less than 35 g/l, with anemia (hemoglobin < 10 g/dl), and patients who received chemotherapy during 6 months before surgery.

Patients were randomly assigned to receive either a combination of epidural blockade with bupivacaine and general anesthesia (EDA group) or general anesthesia alone (control group).

For patients in the EDA group, an epidural catheter was inserted at a thoracic level between T9 and T11 before the operation. Afferent neural blockade was established with 0.5% bupivacaine to achieve a bilateral sensory block from T4 to S5 as judged from perception of pinprick and maintained with intermittent bolus doses of bupivacaine 0.5%. General anesthesia in all patients was induced by 5 mg/kg of thiopentone and 1.5 μg/kg of fentanyl in the EDA group or by 5 μg/kg of fentanyl in the control group, respectively. Tracheal intubation was facilitated with 0.6 mg/kg of rocuronium, and patients’ lungs were ventilated with 30% oxygen in air to maintain normocapnia. No nitrous oxide was used because it has the same molecular weight as ¹³CO²and thus would interfere with the isotope ratio measurement of expired ¹³CO². General anesthesia in the control group was maintained using desflurane at end-tidal concentrations as required to keep heart rate within 20% of preoperative values. In the EDA group, desflurane was administered at end-tidal concentrations of approximately 3 vol% to achieve tolerance of the endotracheal tube and to prevent awareness. The degree of muscle relaxation was monitored using train-of-four ratio, and supplemental doses of rocuronium were applied as needed for complete surgical muscle relaxation. Fluid was given as NaCl 0.9% solution at a rate of 10 ml·kg⁻¹·h⁻¹intraoperatively and 6 ml·kg⁻¹·h⁻¹thereafter. During surgery, the patients were covered with a warming blanket to maintain normothermia. For pain control after surgery, patients in the control group received intravenous morphine. In the epidural group, postoperative analgesia was performed with epidural bupivacaine 0.25% as required to maintain sensory blockade from T8 to L3. Hemodynamic monitoring was performed using a three-lead electrocardiogram monitor and radial artery catheterization for continuous blood pressure measurement.

The rates of appearance of glucose (R^aglucose, endogenous glucose production) and leucine (R^aleucine) were determined before, during, and 2 h after surgery by stable isotope tracer technique using primed continuous infusions of [6,6-²H²]glucose and L-[1-¹³C]leucine (Cambridge Isotope Laboratories, Cambridge, MA). Sterile solutions of the isotopes were prepared by the hospital pharmacy. The bottles were heat sterilized at 121°C for 15 min and kept refrigerated at 4°C until administration. Before the infusion study, each set of solution was confirmed to be free of pyrogens (LAL Pyrogent Test, Whittaker Bioproducts, Walkersville, MD).

All tests were performed on the day of surgery beginning between 7:00 and 8:00 am after fasting for approximately 36 h. Because of the bowel preparation required for colorectal procedures, patients received only clear fluids until midnight the day preceding the operation. A catheter was placed in a superficial vein in the dorsum of the hand and kept patent with saline 0.9% (2 ml·kg⁻¹·h⁻¹). A superficial vein of the contralateral arm was cannulated to pro-vide access for the infusion of [6,6-²H²]glucose and L-[1-¹³C]leucine. Blood and expired air samples were taken to determine baseline enrichments. Thereafter, priming doses of NaH¹³CO³, 1 μmol/kg, L-[1-¹³C]leucine, 4 μmol/kg, and [6,6-²H²]glucose, 22 μmol/kg, were administered fol-lowed immediately by continuous infusions of [6,6-²H²]glucose, 0.22 μmol·kg⁻¹·min⁻¹, and L-[1-¹³C]leucine, 0.06 μmol·kg⁻¹·min⁻¹, respectively. Isotope infusion was uninterrupted throughout the entire study period. Expired breath and blood samples for the determination of isotopic enrichments and for the measurement of metabolic substrates (glucose, lactate, free fatty acids [FFA]) and hormones (insulin, glucagon, cortisol) were collected as indicated in figure 1. Breath samples were collected through a mouthpiece in a 3-l bag and transferred immediately to 10-ml vacutainers until analysis. During controlled ventilation, expired gases were collected by a one-way valve into a 5-l bag. Blood samples were immediately transferred to a heparinized tube, centrifuged at 4°C (3,000 g, 15 min), and the obtained plasma was stored at −70°C until analysis.

Oxygen consumption (Vo²) and carbon dioxide production (Vco²) were measured before and after surgery by indirect calorimetry using the open system indirect calorimetry device Deltatrac Metabolic Monitor (Datex Instrumentarium, Helsinki, Finland). Vco²was also determined during the operation (70 min after skin incision). The values of Vo²and Vco²and the calculated respiratory quotient (RQ) represent an average of the data obtained during a 20-min period on each occasion, with a coefficient of variation less than 10%.

After derivatization of plasma glucose to its pentaacetate compound, the [6,6-²H²]glucose enrichment was quantified by gas chromatography–mass spectrometry using electron-impact ionization. 16Plasma α-[1-¹³C]-ketoisocaproate (α-[1-¹³C]KIC) enrichment was analyzed by electron-impact selected-ion monitoring gas chromatography–mass spectrometry as described earlier, except that t-butyldimethylsylyl rather than trimethylsylyl derivatives were prepared. 17Expired ¹³CO²enrichment for the calculation of leucine oxidation was analyzed by isotope ratio mass spectrometry (Analytical Precision AP2, 003, Manchester, UK).

Plasma concentrations of glucose were quantified using a glucose analyzer 2 (Beckman Instruments, Fullerton, CA) based on a glucose oxidase method. The plasma lactate assay was based on lactate oxidase using the synchron CX 7 system (Beckman Instruments, Fullerton, CA). Circulating concentrations of FFA were measured by an enzymatic assay (Boehringer Mannheim, Laval, Quebec, Canada). Plasma cortisol, insulin, and glucagon concentrations were analyzed by means of a double antibody radioimmunoassay (Amersham International, Amersham, Bucks, UK).

During physiologic and isotopic steady state, the rate of appearance of unlabeled substrate can be derived from the plasma isotope enrichment (APE = atom percent excess) calculated by: R^a= I. (APE^inf/APE^pl− 1), where APE^infis the tracer enrichment in the infusate, APE^plis the tracer enrichment in plasma, and I is the infusion rate of the labeled tracer. The APE values used for the calculation of the rate of appearance were the average of four (pre- and postoperative) and five (intraoperative) APE measurements. Steady state conditions were assumed when the coefficient of variation (CV) of the APE values at isotopic plateau was less than 5%.

During steady state conditions, leucine flux is defined by the equation Q = S + O = B + I, where S is the rate of leucine uptake for protein synthesis, O is the rate of leucine oxidation, B is the rate at which leucine enters the free amino acid pool from endogenous protein breakdown, and I is the rate of leucine intake, including tracer and diet. Therefore, during postabsorptive conditions when there is no exogenous leucine entering the plasma pool, the only source of leucine is that derived from endogenous protein breakdown, and consequently, the rate of leucine breakdown equals leucine flux. 18Plasma α-[1-¹³C]KIC enrichment was used for calculating flux and oxidation of leucine because it has been demonstrated to reflect the intracellular precursor pool enrichment more precisely than leucine itself. 19In the calculation of leucine oxidation, a correction factor of 0.76 was used to account for the fraction of ¹³CO²released from leucine but retained within slow turnover rate pools of the body. 18 

Whole body glucose uptake equals the rate of endogenous glucose production during steady state conditions. However, glucose uptake increases with increasing blood glucose concentration because most glucose uptake occurs in noninsulin-sensitive tissues and consequently depends on the diffusion gradient of glucose. Therefore, whole body glucose uptake is not an accurate measure of the tissue's ability to take up glucose. The fractional glucose clearance rate represents an index of the tissue's capacity to take up glucose. The plasma clearance rate of glucose was calculated as the R^aglucose divided by the corresponding plasma glucose concentration. During surgery, the average of two plasma glucose concentration measurements 60 and 100 min after skin incision was used for the calculation.

The sample size calculation was based on the results of a previous report demonstrating a 50% reduction of splanchnic glucose release by epidural blockade. 7For an expected difference in glucose production of 25% between the groups (power, 80%; α= 5%), 16 patients was calculated to be sufficient. Differences between the groups were analyzed using the Mann–Whitney U test. Within-group comparison of variables was made by analysis of variance for repeated measures with post hoc  analysis by Student–Newman–Keuls test. The relationships between the R^aleucine and R^aglucose were evaluated by the correlation coefficient. A probability of P < 0.05 was considered to be significant. Data are presented as mean ± SD.

The two groups were comparable with regard to patients’ age, height and weight, gender, and American Society of Anesthesiologists (ASA) physical status (table 1). There were no differences between the groups in duration of surgery, estimated blood loss, and amount of crystalloid fluid administered throughout the study period.

During EDA, heart rate and mean arterial blood pressure decreased and remained lower than in the control group (P < 0.05, table 2). The end-tidal desflurane concentration was lower in the EDA group than in the control group at 60 min (EDA, 2.9 ± 0.4 vol%; control, 5.5 ± 1.2 vol%, P < 0.05) and 100 min after skin incision (EDA, 2.9 ± 0.3 vol%; control, 5.5 ± 1.2 vol%, P < 0.05).

In both groups, plasma glucose concentration increased during and after surgery (P < 0.05, table 3), revealing higher values in the control group than in the EDA group (P < 0.05). In all patients, an isotopic plateau of [6,6-²H²]glucose, α-[1-¹³C]KIC, and expired ¹³CO²was achieved (CV < 5%, fig. 2). Intra- and postoperative R^aglucose were lower in the EDA group than in control subjects (P < 0.05). Plasma glucose clearance decreased with both anesthetic techniques (P < 0.05) without showing a difference between the two groups. R^aleucine and protein synthesis decreased during and after surgery (P < 0.05) to a similar extent in both groups. Leucine oxidation intraoperatively decreased in all patients, followed by an increase after surgery (P > 0.05). A correlation between the R^aleucine and R^aglucose was observed in the EDA group (r = 0.504, P < 0.05, fig. 3A), whereas no correlation could be detected in the control group (r =−0.198, P > 0.05, fig. 3B).

Whole body Vo², Vco², and the RQ did not change during the study period (table 4).

Plasma lactate and FFA concentrations remained unaltered with both anesthetic techniques (table 5). There was no perioperative change in plasma insulin concentration. Plasma cortisol concentrations increased throughout the study in both groups (P < 0.05), with lower values during EDA when compared with the control group (P < 0.05). Intraoperative plasma glucagon levels were lower in patients receiving EDA than in the control subjects (P < 0.05).

Considerable attention has been given to the preservation of glucose homeostasis in surgical patients because acute hyperglycemia, a typical feature of the metabolic response to surgery, has been demonstrated to significantly compromise immune function 20and to contribute to poor clinical outcome after cardiopulmonary resuscitation 21and cerebral ischemia 22in humans. It has long been recognized that epidural blockade with local anesthetic, established before and maintained after the operation, prevents or blunts the hyperglycemic response to surgery. 4,7,23It remained unclear, however, if this inhibitory effect of epidural blockade was a consequence of a decrease in glucose production, an increase in glucose utilization, or a combination of both. The major finding of the present study is that patients receiving epidural local anesthetic showed lower intra- and postoperative R^aglucose than subjects in the control group. The absolute rate of glucose production during epidural blockade was identical to values recently reported during hip surgery performed with intrathecal neuraxial blockade. 24Glucose clearance in our study, an indicator of whole body glucose uptake, decreased to a comparable extent with both anesthetics, lending support to the contention that epidural blockade attenuates the hyperglycemic response by modifying glucose production without affecting endogenous tissue glucose utilization. This conclusion corroborates the previous observation of a suppressive effect of epidural local anesthetics on glucose production in parenterally fed patients studied between 1 and 5 days after various surgical interventions. 8The present results are also in accordance with a recent investigation describing an inhibitory influence of epidural blockade on splanchnic glucose release during cholecystectomy. 7Because splanchnic glucose production, calculated from the splanchnic blood flow and the arteriohepatic venous plasma concentration difference of glucose, does not account for the metabolic activity, i.e ., glucose uptake by the gut, this parameter does not represent an accurate measure of whole body glucose production. The fact that all subjects entering the latter study's protocol were continuously given intravenous dextrose further limits the validity of its conclusions.

Whole body glucose production is composed to a varying degree of glycogenolysis and gluconeogenesis. The use of [6,6-²H²]glucose, as in the present protocol, does not allow differentiation between the two biochemical pathways. It seems conceivable, however, that gluconeogenesis is activated in patients undergoing colorectal procedures, as a consequence of the long preoperative fasting period (as a result of bowel preparation) and concurrent depletion of hepatic glycogen stores. 10After 42 h of fasting, gluconeogenesis contributes to more than 90% of the glucose produced in healthy subjects. 9In surgical patients, the rate of this process is further enhanced by the overproduction of counterregulatory hormones, such as catecholamines, cortisol, and glucagon, which all stimulate gluconeogenesis, either directly or indirectly by counteracting the action of insulin. 25Gluconeogenesis in the liver is a highly oxygen-consuming process, accounting for 50% of hepatic oxygen consumption during postabsorptive conditions. 26Thus, the decrease in glucose production as seen during epidural blockade assumes clinical relevance with regard to the energy balance of the liver. Further, gluconeogenesis has been proposed to occupy a central position in catabolic pathways because muscle protein is being hydrolyzed to supply glucose precursors via  the glucoplastic amino acids. 11Hence, it is generally believed that by reducing accelerated gluconeogenesis, amino acids derived from muscle protein breakdown are being directed into anabolic pathways, resulting in less protein loss. 12Confirming this link between perioperative glucose and protein metabolism, we observed a weak, albeit significant, correlation between the R^aleucine and R^aglucose in presence of epidural blockade. 12,27In contrast, no correlation between the two metabolic pathways was detected in the control group, emphasizing the notion that there are more factors than need for gluconeogenic precursors that regulate protein breakdown and gluconeogenesis during surgery. Whole body protein breakdown, synthesis, and amino acid oxidation intraoperatively decreased in all patients, independent of the anesthetic technique used. This depression of protein metabolism is in line with previous studies showing similar decreases in protein breakdown 27–29and decreased rates of intraoperative protein synthesis in the whole body and in specific organ tissues (liver, muscle). 30,31It was also concluded that the effect of anesthetics on protein catabolism during the acute phase of surgical tissue trauma is small. 27 

Because the present protocol was not designed to elucidate underlying mechanisms, we can only speculate about potential endocrine, biochemical, and hemodynamic factors that may have been responsible for the metabolic effects of epidural blockade and surgery in our study.

Much of the catabolic responses to surgical trauma have been ascribed to alterations in the endocrine milieu, in particular to the elevated plasma concentrations of counterregulatory hormones, all of which promote hyperglycemia by stimulating glucose production and decreasing glucose utilization. In accordance with previous reports, epidural blockade attenuated the neuroendocrine response during surgery as reflected by lower plasma cortisol and glucagon concentrations when compared with the control group. 4,7,32Although plasma catecholamine levels were not determined in the present protocol, the hemodynamic pattern in the EDA group characterized by a decreased heart rate and arterial blood pressure strongly suggests efferent blockade of sympathoadrenergic fibers, which has been frequently described during thoracic epidural blockade. 4,23Although the modification of the endocrine responses by epidural local anesthetic fits well with the observed alterations in glucose metabolism, it fails to explain the intraoperative changes in protein metabolism, which occurred independent of the type of anesthesia. Considering the well-known catabolic action of cortisol, the 30% decrease in protein breakdown and amino acid oxidation in presence of increased cortisol levels appear to be paradoxical. It has to be noted, however, that the catabolic effects of corticosteroids are unlikely to take effect within 2–4 h. Cortisol administration has been shown to have only little influence on nitrogen loss, protein breakdown, and amino acid oxidation during the first 24 h in healthy subjects. 33Because insulin represents the key endocrine regulator of protein metabolism, alterations in the perioperative plasma concentration of insulin gain metabolic importance. High spinal anesthesia (T2) after single-shot intrathecal application of local anesthetic has been shown to impair the insulin response to an intravenous dextrose load, whereas basal plasma insulin concentrations were not affected. 34In agreement with previous studies conducted in healthy subjects and surgical patients receiving thoracic–lumbar dermatome blockade, plasma insulin levels did not change in our subjects, neither in the epidural nor in the control group. 35,36Thus, alterations in the insulin system seem an unlikely cause for the intraoperative decrease in protein catabolism. It is interesting to note that hyperglycemia per se , as seen during surgery, has been associated with a decrease in protein breakdown independent of the action of insulin. 37 

Animal studies have shown that local ischemia and hypoxia exert suppressive effects on muscle protein metabolism. 38Even though patients were hemodynamically stable, well hydrated, and oxygenated during surgery, it cannot be ruled out entirely that reduced peripheral muscle perfusion and subsequent regional depression of protein metabolism contributed to the overall effect in the whole body.

We conclude that epidural blockade attenuates the hyperglycemic response to abdominal surgery through the modification of glucose production without affecting glucose utilization. The depression of protein metabolism during and immediately after surgery occurs independent of the anesthetic technique used.

The authors thank Tina Nordolillo, B.Sc. (Biochemist, Department of Anesthesia, McGill University, Montreal, Quebec, Canada), Louise Mazza, B.Sc. (Biochemist, Department of Anesthesia, McGill University, Montreal, Quebec, Canada), and the team of respiratory technicians of the Royal Victoria Hospital (Montreal, Quebec, Canada) for their excellent technical assistance.