Introduction

General anesthesia combined with post operative epidural analgesia for major intra-abdominal surgery results in superior post-operative pain relief, improved mental status, faster return of bowel function, and accelerated patient recovery after surgery, compared with general anesthesia followed by parenteral opioids (1-4). Failure of epidural anesthesia is defined as inadequate pain relief that requires either repositioning or replacement of the catheter, change of medication, or addition of another major pain relief modality (e.g., i.v. PCA). In three major studies with patient populations ranging from 1,000 to 25,000 the failure rate was reported as low as 3%, to as high as 30% (5-7). Epidural failures are attributed to either mal position of catheters not in the epidural space or patient specific anatomy that does not allow equal spread of epidural infusion volumes (5-7). Therefore a potentially large proportion of patients are not experiencing the above mentioned benefits of epidural anesthesia. Although different techniques have been advocated including the use of fluoroscopy, and pre-operative bolusing of catheters, the optimal method for pre or intra-operative determination of a working epidural catheter has yet to be elucidated (8, 9). In this pilot study a novel cardiac output monitor was used to measure hemodynamic changes after bolus administration of local anesthetics into the epidural catheter to evaluate the effectiveness of epidural anesthesia.

The Endotracheal Cardiac Output Monitor (ECOM), (Conmed, Irvine, CA) delivers a stable current (2 mA at 100 kHz) through the tracheal mucosa to the ascending aorta and records impedance changes associated with systolic ejection. From the change of impedance the stroke volume (SV) is determined and cardiac output (CO), systemic vascular resistance (SVR) are calculated real time by using the systolic (SBP), diastolic (DBP), mean arterial blood pressure (MABP) values from arterial line and estimated central venous pressure (CVP) values. Animal studies suggest that the ECOM is safe and accurate in CO determination (10). To date this device has only been studied in the cardiac surgery patient population. Comparison of CO measurements between the ECOM monitor and the more traditional thermodilution (TD) or transesophageal echocardiographic (TEE) techniques has shown poor correlation in absolute values; however, trending analysis yielded 88%, and 100% of CO changes were within 0.5 L/min and 1.0 L/min limits of agreement, respectively (11, 12).

This pilot study was designed to determine the role of hemodynamic changes measured by the ECOM monitor in predicting functional epidural systems. The effectiveness of the epidural systems was compared between patients with and without significant hemodynamic changes.

Methods

The study was accepted by University of Colorado IRB with informed consent, and was funded entirely by an internal departmental research grant. Nineteen patients undergoing intra-abdominal surgery, who accepted an epidural catheter for post-operative pain relief and an arterial line for hemodynamic monitoring, were enrolled into this study. Placement of epidural catheters occurred in the pre-anesthesia area and was performed by either intra-operative providers or members of the acute pain team. All catheters were placed in awake, alert patients lightly sedated with provider choice of intravenous midazolam, fentanyl or combination of the two. The actual placement technique was up to the individual provider. The level of placement of all catheters was between T7- L1. After placement of an epidural catheter and establishment of general anesthesia, systolic, diastolic and mean blood pressures (SBP, DBP, MAP), stroke volume (SV), cardiac output (CO), cardiac index (CI), and systemic vascular resistance (SVR), were monitored and manually recorded in 15 minute intervals using the ECOM monitor. The ECOM monitor was connected to a special endotracheal tube and to a radial arterial line transducer. The SV measurement was derived from the electrodes placed on the endotracheal tube. The electrodes represented a complex electronic system for measuring the changes of the impedance in the ascending aorta, which correlates with the actual SV. For the exact timing of the impedance measurement and for calculating the SVR the system was connected to an arterial line transducer and measured the SBP, SDP and MAP. The patients did not have a central line for this study so a CVP value of 10 mmHg was placed into the system in order to calculate SVR. At the start of abdominal closure the epidural catheters were bolused with 5 ml of 0.25% bupivicaine in ten-minute intervals for a total of 10 ml volume. The hemodynamic parameters where then manually recorded at five-minute intervals until the conclusion of surgery. In the post anesthesia care unit (PACU) the level of dermatomal blockade, with each verterbral level accounting for two dermatomes based on left and right sided blockade, was assessed by using an ice bag to determine temperature discrimination at 20 minutes after arrival. Epidural infusions were started in PACU on all patients with either 0.08% or 0.1% bupivicaine with the addition of 6-10 mcg/ml hydromorphone. Infusion rates were started at between 8 and 10ml/hr based on the extent of incision and the localization of pain. Pain was assessed throughout the study period (every 4 hours, for 24 hours) using a visual analog scale (VAS) with 0 having no pain and 10 having the worst pain imaginable. Patient satisfaction was also evaluated on a 0-10 scale from being unsatisfied (0) to very satisfied (10) with the pain management. All patients were transferred to either the surgical intensive care unit or general surgical floor. Patients were followed by acute pain team personnel until discontinuation of the epidural catheter. The epidural infusion rates were adjusted from 6-14 ml/hr according to the dermatomal level of the block and VAS pain score.

The number of dermatomes blocked at 20 minutes from PACU arrival, the average PACU VAS pain scores, 24-hour epidural medication amounts, average 24-hour VAS pain score, and patient satisfaction score were recorded at 24 hours post-operatively.

Statistical analysis

Results

Twelve males and 7 females were enrolled into the study. Average age of patients was 58.8 ± 9.1 years. Diagnosis included: pancreatic cancer (6), rectal cancer (2), renal cancer (2), colon cancer (2), failed gastric bypass, tubular adenoma, gastrointestinal stromal cell tumor, abdominal liposarcoma, duodenal polyps, pseudomyxoma peritonei, and ulcerative colitis. Dictated procedures included: whipple (6), nephrectomy (2), distal pancreatectomy/splenectomy , right hepatectomy, duodenal resection, antero-posterior resection, abdominoperoneal resection, revision of gastric bypass, small bowel resection, colon resection, radiofrequency hepatic ablation/adrenalectomy, ileo-anal pull-through, and abdominal debulking. Average loss of resistance with epidural placement occurred at 6.6 ± 1.4 cm, with catheters secured to skin at an average of 11.8 ± 1.3 cm. Eleven of 19 catheters were recorded as easy placement, 7/19 required more than one attempt and 1/19 required movement of needle insertion to different vertebral level. All test doses were reported negative with no immediate complications from insertion.

The hemodynamic values were averaged before and after the epidural bolus injection and paired Student’s t- test was used to calculate the significance of the difference in each patient. A significant reduction was observed in SBP, DBP, MAP, SVR, with increases seen in CO, and CI after starting epidural anesthesia. No significant change in SV or HR was observed (Table 1). It was hypothesized that significant change in each of the hemodynamic parameters (SBP, DBP, MAP, SVR, CO, CI) in individual patients might predict a functional epidural catheter and effective postoperative pain relief. The dermatome level, average PACU VAS pain score, average 24-hour VAS pain score and patient satisfaction score were compared in patients with or without significant changes of each of the above hemodynamic values. Patients with significant change in SBP, DBP, MAP, CO or CI did not show any correlation with dermatome level or any of the measured pain scores. Significant SVR reduction was the only hemodynamic parameter, which was associated with higher dermatomal blockade and decreased pain scores. Significant SVR reduction was observed in 12 patients (Group I) and non-significant change was detected in 7 patients (Group II). The number of dermatomes blocked at 20 minutes post PACU arrival was significantly higher and the average 24-hour VAS pain score was significantly lower in Group I (Table 2). The average PACU VAS pain score and patient satisfaction score were not statistically different between the two groups. There was no significant difference between groups I and II in the dose of epidural bupivicaine or epidural hydromorphone received in the first 24 hours post-operatively (Table 2).

Figure 1

Table 1: Hemodynamic changes associated with epidural bolus administration of 10 ml bupivacaine 0.25%

Figure 2

Table 2: Patients divided into two groups based on the presence (Group I) or absence

(Group II) of significant reduction in SVR after bolusing of epidural catheter

Discussion

Epidural catheterization and local anesthetic administration causes a segmental blockade of nervous fibers in the affected dermatomes. Segmental blockade is proportional to the volume of injectant while differential blockade is based on the concentration of local anesthetic. Small unmyelinated sympathetic fibers are blocked first, followed by sensory fibers and finally large myelinated motor fibers being the most resistant to blockade. Sympathetic nerve blockade leads to a decrease in SVR, SBP, DBP, MAP, and if compensatory mechanisms are intact an increase in CO, and CI (13-16). In our study, 10 ml bupivacaine 0.25% as a loading dose induced the above pattern of hemodynamic changes as detected by the ECOM monitor. In previous studies the CO measurements of the ECOM monitor, compared to TD or TEE values in cardiac surgery patients, have shown poor correlation in absolute values but the trending values were found to be acceptable compared with both TD and TEE derived CO measurements (11,12). These results might reflect the fact that the ECOM monitor is more reactive to rapid changes (beat to beat calculation of SV and CO) whereas the values of thermodilution measurement are updated and averaged from 30-second time periods. Despite potential limitations of the ECOM monitor it seemed a reasonable tool with appropriate sensitivity for real-time tracking of the direction and magnitude of hemodynamic changes in the setting of epidural anesthesia. The system can be used only in patients who are intubated, which is a disadvantage especially in cases where the patient is extubated in the operating room but transferred to the intensive care unit setting and potentially needs continued hemodynamic monitoring. The ECOM monitor is sensitive to electronic disturbances generated during surgery but it takes only a short time to get back to displaying the impedance curve and recording the hemodynamic numbers. The current study is the first to our knowledge using this technology in a non-cardiac surgery patient population, and also the first that uses the technology in order to attain hemodynamic indices associated with starting epidural anesthesia.

The ideal confirmation of catheter tip location and effectiveness of the epidural system are yet to be determined. Flouroscopy can be used to accurately define the position of the catheter (8), though this method is expensive, time consuming, and may not be immediately available. Pre-operative bolusing of epidurals can also be used to assess neuraxial blockade prior to onset of general anesthesia. This technique can be used to exclude intravascular or intrathecal placement of the catheter and an anesthetized dermatome may suggest an effective epidural system. A recent study has shown that 80% of patients will have a demonstrated block of at least four dermatomal segments within eight minutes after bolusing the epidural catheter (9). In certain circumstances the pre-operative bolus may induce profound hypotension during anesthetic induction of hemodynamically unstable patients. Another common way of interrogating an epidural system is intra-operative bolusing of local anesthetic. Anesthesiologists often gain confidence of a working epidural if a decrease in blood pressure after local anesthetic bolusing is seen. It has been shown that the onset of epidural anesthesia is accompanied by decrease in SBP, DBP, as well as MAP (16). Our data supports this observation in general; however, significant decrease in SBP, DBP, or MAP did not differentiate between patients in the effectiveness of epidural anesthesia. According to our data, serial measurement of SVR during the surgery and after a bolus injection of an epidural local anesthetic was the only hemodynamic indicator associated with increased dermatomal blockade and better 24-hour pain control.

Patients with optimal epidural anesthesia along with other benefits, usually experience effective pain relief post-operatively. In our study 12/19 (63%) of the patients had a blockade of about 20 dermatomes or 10 vertebral levels, with significant change in SVR and significantly less average 24-hour pain scores (2.2±1.5) after the bolus administration of bupivacaine (Group I, Table 2). At the same time 7/19 (37%) of the patients had significantly less dermatome blockade and significantly less change in SVR (Group II, Table 2). This patient group had a blockade of about 12 dermatomes or 6 vertebral levels which may explain that although these patients had significantly higher average 24-hour VAS pain scores (3.8±1.7), these scores were within the range of 1.12-3.99, which has characterized post laparotomy patients with functioning epidural anesthesia (17-19). This result may also explain why the overall patient satisfaction did not differ significantly between the two groups. There was no difference between the two groups in the amount of bupivacaine or hydromorphone infused through the epidural catheter in the first 24 hrs post-operatively suggesting that epidural anesthesia in group II might be less optimal but still effective to provide adequate pain relief characterized by low pain scores and high level of patient satisfaction. The above data suggests a limited role of hemodynamic measurements by the ECOM monitor in predicting the effectiveness of epidural anesthesia, though it may be able to predict a more extensive dermatome blockade. The cost might not be comparable to the benefits in post-operative 24 hr pain scores, patient satisfaction or the amount of medications infused through the epidural catheter.

Conclusion

This pilot study suggests that significant SVR change after starting epidural anesthesia may identify significantly higher dermatome blockade and reduced pain scores but it is not an absolute requirement for effective post-operative pain relief. Patients without significant SVR change may also have effective dermatome level with relatively low pain scores, high patient satisfaction and requirement of the same amount of pain medications.