Skip Navigation LinksHome > February 2011 - Volume 114 - Issue 2 > Thoracic Epidural Anesthesia Improves Early Outcomes in Pati...
doi: 10.1097/ALN.0b013e318201f571
Pain Medicine

Thoracic Epidural Anesthesia Improves Early Outcomes in Patients Undergoing Off-pump Coronary Artery Bypass Surgery: A Prospective, Randomized, Controlled Trial

Caputo, Massimo M.D.*; Alwair, Hazaim M.R.C.S.†; Rogers, Chris A. Ph.D.‡; Pike, Katie M.Sc.§; Cohen, Alan F.R.C.A.∥; Monk, Christopher F.R.C.A.∥; Tomkins, Sally F.R.C.A.∥; Ryder, Ian F.R.C.A.∥; Moscariello, Cesare M.D.#; Lucchetti, Vincenzo M.D.#; Angelini, Gianni D. F.R.C.S.**

Free Access
Supplemental Author Material
Article Outline
Collapse Box

Author Information

Collapse Box


Background: The aim of this two-center, open, randomized, controlled trial was to evaluate the impact of thoracic epidural anesthesia on early clinical outcomes in patients undergoing off-pump coronary artery bypass surgery.
Methods: Two hundred and twenty-six patients were randomized to either general anesthesia plus epidural (GAE) (n = 109) or general anesthesia only (GA) (n = 117). The primary outcome was length of postoperative hospital stay. Secondary outcomes were: arrhythmia, inotropic support, intubation time, perioperative myocardial infarction, neurologic events, intensive care stay, pain scores, and analgesia requirement.
Results: Baseline characteristics were similar in the two groups. One patient died in the GAE group. Median postoperative stay was significantly reduced in the GAE, compared with the GA, group (5 days, interquartile range [5–6] vs. 6 days, interquartile range [5–7], hazard ratio = 1.39, 95% CI [1.06–1.82]; P = 0.017). The incidence of arrhythmias and the median intubation time were both significantly lower in the GAE, compared with the GA, group (odds ratio = 0.41, 95% CI [0.22–0.78], P = 0.006 and hazard ratio = 1.73, 95% CI [1.31–2.27], P < 0.001, respectively). Patients in the GAE group were more likely to need vasoconstrictors intraoperatively than in the GA group (odds ratio = 2.50, 95% CI [1.22–5.12]; P = 0.012). The GAE, compared with GA, group reported significantly lower levels of impairment for all pain domains and reduced morphine usage (odds ratio = 0.07, 95% CI [0.03–0.17]; P < 0.001).
Conclusion: In patients undergoing off-pump coronary artery bypass surgery, the addition of thoracic epidural to general anesthesia significantly reduces the incidence of postoperative arrhythmias and improves pain control and overall quality of recovery, allowing earlier extubation and hospital discharge.
Back to Top | Article Outline

What We Already Know about This Topic

* Whether thoracic epidural analgesia improves recovery after coronary artery bypass grafting (CABG) surgery is unclear and has not been studied for off-pump CABG
Back to Top | Article Outline

What This Article Tells Us That Is New

* In a randomized trial of more than 200 patients undergoing off-pump CABG, the addition of thoracic epidural analgesia to general anesthesia was associated with reduced hospital stay, improved pain control and quality of recovery, and reduced incidence of dysrhytmias
THE current, greatly improved mortality rate after coronary artery bypass grafting (CABG) is a relatively crude measure of outcome, and data on complications and quality of recovery should be used to assess increasingly important early outcomes for patients and the health care system. A recent report by the Society of Cardiothoracic Surgery of Great Britain and Ireland stressed the fact that “mortality is important but a relatively crude measure, and emphasized the need to look at data on complications and length of stay.”1 Reduction of postoperative morbidity is an obvious goal of clinical care, but is also extremely important in the current era of economic constraint, as this can help to minimize a patient's length of hospital stay and reduce the chances of readmission for complications.1
The search for improvements in quality of recovery for patients undergoing CABG is therefore an important target for doctors and nurses working in this field. In this context, the addition of regional to general anesthesia, with its demonstrated benefits in adult and pediatric patients undergoing major surgery, is receiving increasing attention from anesthetists and cardiac surgeons.2–7 Regional anesthesia may attenuate adverse physiologic stress responses associated with cardiothoracic surgery, including alterations in circulatory (tachycardia, hypertension, vasoconstriction), metabolic (increased catabolism and cortisol release), immunologic (impaired immune response), and hemostatic (platelet activation) systems.8–12
Two recent randomized, controlled trials3,4 have investigated the potential benefits of thoracic epidural anesthesia in patients undergoing conventional CABG with cardiopulmonary bypass (CPB) and cardioplegic arrest. Scott et al.3 demonstrated a reduction in the incidence of supraventricular arrhythmias, postoperative confusion, respiratory tract infection, and renal failure in patients receiving regional anesthesia as well as earlier tracheal extubation on the intensive care unit (ICU), compared with control patients receiving general anesthesia only. Priestley et al.4 also demonstrated better analgesia and earlier extubation with the use of thoracic epidural anesthesia, compared with the general anesthesia control group, with no significant difference in the postoperative length of stay. Others, however, have shown no significant difference in extubation time with or without regional anesthesia.13,14 In general, cardiac anesthetists have been reluctant to adopt this technique given the need for full heparinization during CPB and the potential for bleeding and neurologic complications. The aim of this randomized, controlled trial was to investigate the impact of regional anesthesia on early clinical outcomes in patients undergoing off-pump coronary artery bypass (OPCAB) surgery where the need for systemic heparinization is significantly reduced.
Back to Top | Article Outline

Materials and Methods

Study Design
This was a two-center, open, parallel-group, randomized, controlled trial.
Back to Top | Article Outline
Eligible patients were adults (at least 16 yr) undergoing primary OPCAB surgery without the use of CPB and cardioplegic arrest. Patients requiring salvage CABG, in cardiogenic shock, or who had associated heart valve pathologies were excluded. Patients on intravenous heparin, warfarin, or clopidogrel at the time of surgery or who suffered from bleeding diathesis were also excluded.
Back to Top | Article Outline
Study Settings
The study was conducted at the Bristol Heart Institute (Bristol, United Kingdom) and Clinica Montevergine (Mercogliano, Italy), two specialized regional cardiac surgery centers serving a population of approximately 1.5 million people. The study was approved by the Hospital Research Ethics Committee (Central and South Bristol Research Ethics Committee, Bristol, United Kingdom), and written informed consent was obtained from all patients.
Back to Top | Article Outline
Patients were randomized to receive either general anesthesia plus epidural (GAE) or general anesthesia (GA) only.
Back to Top | Article Outline
Anesthetic Technique
In both groups, the anesthetic technique consisted of premedication with benzodiazepines and induction with propofol at 0.5–1 mg/kg combined with fentanyl (10–20 μg/kg). Neuromuscular blockade was achieved with 0.1–0.15 mg/kg pancuronium bromide or vecuronium, and, following tracheal intubation, the lungs were ventilated to normocapnia with air and oxygen (45–50%). End-tidal carbon dioxide was maintained between 35 and 40 mmHg throughout. Anesthesia was maintained with either isoflurane at 0.8–1.0 minimal anesthetic concentration or intravenous propofol 3–4 mg · kg−1 · h−1, at the discretion of the consultant anesthetist.
In addition, patients in the GAE group had a thoracic epidural catheter sited in the operating theater immediately before surgery at the T2–3 or T3–4 intervertebral space. Bilateral neuraxial block was established from T1 to T10 with an initial bolus of 5 ml bupivacaine, 0.5%, followed by another 5-ml bolus after 10 min. Determination of the spread of block was performed with ethyl chloride spray. If a “bloody tap” was to occur, the operation was postponed for 24 h and commenced only if neurologic examination was completely normal the next morning. Any focal neurologic abnormality resulted in an urgent magnetic resonance imaging scan to exclude epidural hematoma. After induction of GA and when central hemodynamic status was stable, a continuous infusion of 0.125% bupivacaine and 0.0003% clonidine (150 μg in 500 ml) was commenced at an initial rate of 10 ml/h in accordance with the protocol described by Scott et al.3
Back to Top | Article Outline
Surgical Technique
The surgical technique and the method of exposure and stabilization for performing anastomoses in patients undergoing OPCAB surgery have been described previously.15,16 Heparin (150 units/kg) was administered before the start of the first anastomosis to achieve an activated clotting time of 250–350 s. On completion of all anastomoses, protamine sulfate was given to reverse the effect of heparin and return the activated clotting time to less than 120 s.
Back to Top | Article Outline
Postoperative Care
In the GAE group, the epidural infusion continued for 72 h. After surgery, the rate of infusion was titrated by the attending intensivist according to clinical need; the goal was to maintain the neuraxial block between T1 and T10 throughout the infusion. “Top-up” bolus doses up to a maximum of 4 ml bupivacaine, 0.25%, were administered when the patient complained of pain. If more than three increases to the infusion rate or more than three epidural top-up doses were required in any hour, analgesia was considered inadequate.
In the GA group, a patient-controlled analgesia intravenous morphine pump was started in the ICU for 48 h by using 1-mg bolus dosing with a 5-min lockout period. Patient-controlled analgesia was also available, if required, to patients in the GAE group. All patients in both groups received additional oral paracetamol (1 g) every 6 h.
All other aspects of postoperative patient management were according to unit protocol as reported previously.16
Back to Top | Article Outline
Outcome Measures
The primary outcome was the length of postoperative hospital stay, defined as the number of days from surgery to discharge from the hospital. Patients were discharged when apyrexial, with normal blood test results and chest x-ray and full mobility. The period from randomization to surgery (including any delay to the surgery for clinical or operational reasons) and any period of convalescence for rehabilitation or social reasons in another healthcare facility before discharge home were excluded.
Secondary outcomes were (1) new arrhythmia, defined as any duration at any time in the postoperative period on the basis of a rhythm strip or 12-lead electrocardiogram17; (2) requirements for inotropic or vasodilator support (noradrenaline was used as a vasopressor; dopamine was used in bradycardic patients, and adrenaline was mostly for low cardiac output); (3) blood loss in the first 12 h after surgery and transfusion requirement, given if the hematocrit fell to less than or equal to 23%18; (4) intubation time; (5) perioperative myocardial infarction (MI), defined as new Q waves of 0.04 ms and/or a reduction in R waves more than 25% in at least two contiguous leads on electrocardiogram17; (6) chest infection, defined as the presence of purulent sputum associated with fever and requiring antibiotic therapy according to positive sputum culture and wound infection defined according to the National Nosocomial Infections Surveillance Report††; (7) neurologic events (diagnosis of stroke was made if there was evidence of new neurologic deficit with morphological substrate confirmed by computed tomography or nuclear magnetic resonance imaging)17; (8) pain outcomes and analgesia requirement; (9) postoperative ICU stay in days; and (10) quality of life at recruitment, 3 months, and 6 months after surgery, measured using the coronary revascularization outcome questionnaire and short-form health survey.
Pain outcomes were collected by the nursing staff twice daily (am and pm) at days 1–5 postoperatively. Six domains (mobility, sedation, pain, upper limb, lower limb, and vomit) were scored. Ordinal scales (ranging from 0 to 2 or 0 to 5, depending on the item) were used, with higher scores indicating greater severity/impairment. For the United Kingdom patients, information on the use of analgesia and the doses of morphine, tramadol, codeine, and nonsteroidal analgesia was also collected at days 1–5 by the nursing staff.
Quality of life was measured for the United Kingdom patients only. At recruitment, the questionnaires were handed to the patient for completion. Follow-up questionnaires at 3 and 6 months after surgery were sent to the patient's home.
Back to Top | Article Outline
Sample Size
The study size was set at 300 patients (150 per group) to provide 80% power to detect clinically relevant difference of 1 day in postoperative length of stay, assuming a mean postoperative hospital stay of 6.9 days (SD 4.7 days) for the GA group and 5% statistical significance, which were in line with previous studies carried out at the Bristol Heart Institute. Recruitment was slower than expected, and after 120 patients had been recruited from the United Kingdom center, data on length of stay were extracted from the trial database, and the estimate for the SD was recalculated. The data were not unblinded, and no comparison between the groups was made. The distribution of the length of stay was skewed, so for the revised calculations, the data were transformed to the logarithmic scale and a target reduction of 1 day in median length of stay (from 6 to 5 days) was used. The SD (for the two groups combined) was 0.494, giving an effect size of 0.368 and a revised total sample size of 230 patients.
Back to Top | Article Outline
Randomized treatment allocations were generated using Stata version 8 (StataCorp LP, College Station, TX). They were stratified by consultant team with a 1:1 allocation using blocks of varying sizes. Allocation details were concealed in sequentially numbered, opaque sealed envelopes. These were prepared by the clinical trials and evaluation unit. Randomization took place after the research fellow had obtained written informed patient consent. This was usually the night before surgery for a planned morning case or on the morning of surgery for an afternoon case.
Back to Top | Article Outline
Statistical Methods
Continuous variables were summarized using the mean ± SD (or median and interquartile range [IQR] if the distribution was skewed), and categorical data were summarized as a number and percentage. “Time-to-event” outcomes (e.g., length of hospital stay and time to extubation) were compared using hazard ratios (HRs) from Cox proportional hazards models, censoring at death for any deaths before the event occurred. The time-to-event curves were constructed using the Kaplan–Meier method. Binary outcomes were compared using odds ratios (ORs). Continuous outcomes were modeled using linear regression, with logarithmic transformations if distributions were skewed. For transformed data, the results were transformed back to the original scale after analysis and the results presented as geometric means, with the difference between the treatment groups expressed as a ratio of geometric means. All effect estimates are reported with 95% CIs.
Postoperative pain scores were normalized, and the resulting scores analyzed using mixed models with patient and patient by time fitted as random effects. Again, changes in treatment effect over time were assessed using the F test. In addition, the potential inclusion of nonlinear time effects (and relevant interaction terms) was assessed using F tests. Pain and analgesia data collected on day 5 were excluded because of differential missing data patterns between treatment groups (at all other times missing data were minimal).
Analyses were carried out on the basis of intention to treat. All analyses were adjusted for consultant team, as the randomization was stratified by team. Center-by-treatment and consultant team-by-treatment interactions were examined, and no evidence of differing treatment effects between centers and teams was found. The validity of the assumptions underpinning the models fitted was checked. The primary outcome was reanalyzed (1) excluding protocol violations, (2) adjusting for consultant team and the additive euroSCORE, and (3) redefining length of stay to include the time from randomization to surgery as sensitivity analyses.
No correction was made for multiple comparisons, but our interpretation of the findings takes into account the consistency of the associations observed and their magnitude as well as their statistical significance. All mixed models were fitted in SAS version 9.1 (SAS Institute, Inc., Cary, NC). All other analyses were performed using Stata version 10.1 (StataCorp LP).
Back to Top | Article Outline


Patient Recruitment
Fig. 1
Fig. 1
Image Tools
Patient recruitment took place between August 2003 and November 2007. Two hundred and twenty-six patients were enrolled in the study, with 109 randomly assigned to GAE and 117 to GA (see fig. 1). One hundred and fifty-seven of these patients were recruited by the United Kingdom center; the remaining 69 patients were recruited by the Italian center. Data on patients screened but not enrolled were not collected.
Back to Top | Article Outline
Protocol Violations
There were 18 protocol violations in patients allocated to GAE who received GA because the epidural could not be inserted. Of these, bloody taps occurred in seven patients, and severe hypotension requiring a bolus of adrenaline intraoperatively without any clinical consequences occurred in two patients. The failure rate was distributed across the study period; it was not concentrated at the beginning of the study (see Supplemental Digital Content 1, which is a table showing epidural failures, Six operations were carried out on-pump; three were classified as on-pump at the first anastomosis, and three were converted from off- to on-pump sometime after the first anastomosis (fig. 1).
No patient withdrew from the trial.
Back to Top | Article Outline
Baseline Characteristics and Operative Details
Table 1
Table 1
Image Tools
Table 1
Table 1
Image Tools
Patient demographics are presented in table 1. The patients' mean age was 65.7 yr (SD, 8.7 yr), and 90% were male. The median euroSCORE was 3 (IQR, 1,4). Sixteen percent were smokers, and 23% were diabetic. The majority of the patients (70%) had triple-vessel coronary disease, and 28% had more than 50% stenosis in the left main stem. Patient characteristics were similar in the two groups, although lung disease/chronic obstructive airways disease was more common in the GAE group (23% vs. 12%). Also, the percentage of insulin-dependent diabetics was higher in the GA group (12% vs. 4%), although the overall proportion of diabetics was similar in the two groups (20% GAE, 26% GA).
The mean number of grafts was 2.7 (SD, 0.7), and mean operative duration 193 min (SD, 70).
Back to Top | Article Outline
Numbers Analyzed
The primary analysis was intention to treat and involved all patients who were randomly assigned. Data are complete, except where indicated by footnotes to the relevant table.
Back to Top | Article Outline
Primary Outcome
Table 2
Table 2
Image Tools
Fig. 2
Fig. 2
Image Tools
The median time from surgery to discharge was 1 day shorter for patients in the GAE group compared with those in the GA group (5 days, IQR [5–6] vs. 6 days, IQR [5–7]; see table 2). The estimated HR (GAE/GA) of 1.39 (95% CI [1.06, 1.82]; P = 0.017), indicates that, at any given time, the “hazard” of discharge from hospital for a patient in the GAE group was approximately 40% higher than that of a patient in the GA group (i.e., GAE patients were significantly more likely to be discharged earlier than those in the GA group). Figure 2 shows the time to discharge by treatment group.
The majority of patients at the United Kingdom center were discharged home. Eight patients in the GAE group and five in the GA group were discharged to another healthcare facility, and one patient in the GAE group was discharged to a convalescence facility. Healthcare systems in Italy are such that the majority of patients are discharged to a convalescence/rehabilitation facility before returning home.
Back to Top | Article Outline
Secondary Outcomes
Clinical Outcomes.
Secondary clinical outcomes are reported in table 2. The need for intraoperative vasoconstrictors differed significantly between the two treatment groups; patients in the GAE group were 2.5 times more likely to need vasoconstrictors intraoperatively (OR = 2.50, 95% CI [1.22, 5.12]; P = 0.012). However, there was no significant difference between the groups with regard to the need for postoperative inotropes (P = 0.30). Of the 42 patients who needed vasoconstrictors interoperatively, 21 received dopamine (10 in the GAE group, 11 in the GA group), 2 received adrenaline (both in the GA group), and 20 were given noradrenaline (18 in the GAE group, 2 in the GA group). One patient (in the GA group) received both dopamine and adrenaline. Of the 121 patients given inotropes postoperatively, the dose was minimal (3–5 μg/kg per min) in 61 cases (40 in the GAE group, 21 in the GA group), moderate (more than 5 μg/kg per min) in 57 cases (19 in the GAE group, 38 in the GA group), and maximal (adrenaline/enoximone) in 3 cases (all in the GAE group). Doses less than 3 μg/kg per min were classified as “none.” In total, 66/109 (61%) patients in the GAE group and 59/117 (50%) patients in the GA group required inotropic support intra- and/or postoperatively (OR = 1.59, 95% CI [0.89, 2.82]; P = 0.12), and 38 patients (24 patients in the GAE group and 14 in the GA group) required inotropic support both intra- and postoperatively (OR = 2.06, 95% CI [0.99, 4.29]; P = 0.053).
The incidence of new arrhythmias and median intubation time were both significantly lower in the GAE group compared with the GA group (OR = 0.41, 95% CI [0.22–0.78], P = 0.006 and HR = 1.73, 95% CI [1.31–2.27], P < 0.001, respectively).
There was no evidence of statistically significant differences in the incidence of vasodilator support (either intra- or postoperatively) or postoperative complications (MI, infections, and neurologic events) between the two groups (see table 2). However, the incidence of postoperative complications was generally low (MI 5%, stroke 2%, infection 17%), and therefore comparisons between treatment groups have low power. Ten patients were reoperated on for reasons of bleeding (5), low cardiac output (2), mediastinitis (1), and reason not recorded (2). Thirty-four patients had a pneumothorax or lung collapse (12 in the GAE group and 22 in the GE group; P = 0.11). Lengths of ICU stay (see fig. 2), median blood loss, and transfusion requirements were also similar between the two groups.
There was one death in the ICU after 13 days. This patient (from the GAE group) suffered an MI postoperatively and died of multiorgan failure.
Back to Top | Article Outline
Pain and Mobility Outcomes.
Pain and mobility data were analyzed for 193 patients (92 in the GAE group, 101 in the GA group); the first 33 Italian patients recruited were excluded, as the data were not collected for five of the six domains. Scores for the pain domain were collected, but only for days 1–3.
Fig. 3
Fig. 3
Image Tools
The effect of GAE versus GA on pain scores is illustrated in figure 3. The GAE group reported significantly lower levels of impairment/pain for all six domains of mobility (mobility, sedation, pain, upper- and lower-limb motor, and vomit), compared with the GA group. The effect of treatment differed over time, with the greatest differences observed in the immediate postoperative period. Although the differences in average scores between the groups reduced over time, all but the mobility score remained significantly lower for the GAE group at the end of postoperative day 3, with scores on the pain domain significantly lower up to postoperative day 4, compared with those for the GA group. Full details of all scores at each time point can be found in Supplemental Digital Content 2,
Table 3
Table 3
Image Tools
Analgesia results are presented in table 3. The proportion of patients receiving any analgesia was similar in the two groups (P = 0.54). However, there was less morphine usage in the GAE group compared with the GA group (OR 0.07, 95% CI [0.03, 0.17]; P < 0.001), and for the subgroup of patients prescribed morphine, the median total dosage was lower in the GAE group compared with the GA group (12 mg, IQR [3, 29] vs. 21 mg, IQR [11, 33.5]; P = 0.004). Usage of other analgesics (tramadol, codeine, and nonsteroidal analgesic drugs) was similar in the two groups. Details of analgesic use at each day can be found in Supplemental Digital Content 3,
Back to Top | Article Outline
Quality-of-life Scores.
No statistically significant differences were found between the two groups on either the coronary revascularization outcome questionnaire or the short-form health survey measures. Full details are reported in Supplemental Digital Content 4,
Back to Top | Article Outline
Sensitivity Analyses
The sensitivity analyses for the primary outcome did not alter any conclusions: (1) excluding protocol violations (HR = 1.51, 95% CI [1.12, 2.02]; P = 0.006), (2) including adjustment for the additive euroSCORE (HR = 1.34, 95% CI [1.02, 1.77]; P = 0.035), and (3) redefining the length of stay to include the period from randomization to surgery (HR = 1.36, 95% CI [1.04, 1.79]; P = 0.025). The distribution and median time from randomization to surgery were similar in both groups (median, 1 day; P = 0.34). There were differences in overall prevalence of inotropic and vasodilator support between the United Kingdom and Italian centers, but there was no evidence of differing treatment effects between the centers.
Back to Top | Article Outline


Our study has clearly demonstrated that adding perioperative regional anesthesia to general anesthesia with “conventional” postoperative analgesia reduces the length of postoperative hospital stay by an average of 1 day. Its use also reduces postoperative arrhythmias and intubation time, improves postoperative pain control, mobility, and physiotherapy cooperation, and reduces nausea in patients undergoing OPCAB surgery.
Superior analgesia, and the avoidance of parental opioids and their side effects, led to significantly quicker extubation times in the GAE group, confirming the results of other studies.3–12 This superior analgesia was also associated with improved mobility and physiotherapy cooperation and less nausea, probably related to the significant decreased use of opioids. These factors, together with a significant decrease in the incidence of postoperative arrhythmias, explain the early hospital discharge in patients receiving GAE. Such a reduction in hospital stay could have a significant implication, in terms of hospital resources, in times of steadily increasing healthcare costs.
The overall reduced risk (40%) of new postoperative arrhythmias in the GAE group compared with patients in the GA-only group is one of the most compelling findings of the trial. In the anesthetized patient, the major determinant of heart rate is the balance of the sympathetic and parasympathetic activity. High regional anesthesia, including the upper five thoracic segments, blocks the cardiac afferent and efferent sympathetic fibers, resulting in a loss of the chronotropic and inotropic drive to the myocardium. This results in significantly improved intraoperative stability, in terms of heart rate and hemodynamics, and fewer postoperative supraventricular arrhythmias. In the GAE group, we observed remarkably stable heart rates throughout the surgical procedure; β-adrenergic blockers were not needed, even during periods of intense surgical stimulation, such as sternotomy, and while positioning the heart to perform distal coronary anastomosis. Other investigators have reported lower catecholamine levels and slower heart rates after CABG in thoracic epidural-treated patients,19–23 and similar reduction in new supraventricular arrhythmias.22,23 Our trial is also in accordance with previous reports that showed larger intraoperative vasopressor requirements in GAE-treated CABG patients, compared with controls, related to the vasodilator effect of epidural anesthesia. The resulting hypotension has a potential impact on the incidence of myocardial ischemia in patients with critical coronary stenosis; nevertheless, we did not find any evidence of increased MI between the two groups.
The main argument against the use of thoracic epidural anesthesia for conventional CABG surgery is the fear of an increased risk of epidural hematoma caused by the need to administer a large dose of heparin immediately before CPB. This perceived increased risk cannot be quantified, but the incidence of epidural hematoma after catheter insertion without heparinization is approximately 1 in 10,000.21 Such a risk must be balanced by important clinical advantages if the technique is to be justified. In OPCAB surgery, the need for heparinization is reduced to half the dose as that used for conventional CABG, making the use of epidural anesthesia a more attractive approach. There were six conversions to CPB in the study, three of whom had received an epidural. One of the three patients randomized to receive an epidural had a postoperative MI. There is no evidence in the literature to suggest any increased risk or potential medical-legal concerns with epidural and full heparinization on CPB.3
When assessing the efficacy of GAE, compared with GA, no adjustment was made for the number of statistical comparisons, but our interpretation of the findings takes into account the consistency and clinical plausibility of the associations observed and their magnitude, as well as their statistical significance. If we were to apply a Bonferroni correction (the most conservative adjustment) in order to maintain an overall 5% significance level across all secondary outcomes compared, two parameters would be considered statistically significant: intubation time and use of opioids.
Back to Top | Article Outline
The main limitation of the study was that it was not blinded, and there was no independent adjudication of clinical endpoints. Although it is recognized that lack of masking can affect the assessment of clinical parameters, all patients were managed according to strict protocols, and data were collected in a consistent manner throughout.
The exclusion of the first 33 patients recruited from the Italian center from the analyses of pain outcomes was a further limitation, insofar as the reduced sample size reduced the statistical power of the comparisons. However, we do not believe it biased our findings, as the patients were a consecutive series; their exclusion was not influenced in any way by their treatment allocation or clinical outcome.
The study was slow in recruiting, partially because not all of the anesthetists in the two units were agreeable to the use of epidural anesthesia. This reflects the controversy already mentioned. A large number of patients were also excluded because they were receiving intravenous heparin or clopidogrel. Given the increasing use of clopidogrel in routine clinical practice, this may represent a significant limitation to the application of the epidural anesthesia technique, and a further randomized, controlled trial including patients on clopidogrel may be warranted.
In this study, there were 11 failures of epidural catheter insertion and 7 bloody taps that delayed the surgery for at least 1 day. The failure to establish a working epidural was not related to the learning curve and is consistent with previous reports.4 Bloody taps are a recognized complication that is entirely unpredictable. The failure rate also reflects the cautious approach taken by our anesthetists. If one or two attempts failed, they did not persist for fear of causing a hematoma and risking patient injury.
Back to Top | Article Outline


This is the first randomized, controlled clinical trial of the use of epidural anesthesia in patients undergoing OPCAB surgery. We have demonstrated that the addition of thoracic epidural to conventional general anesthesia accounts for a significant reduction in the incidence of postoperative arrhythmias and improvement in overall quality of recovery, allowing an earlier tracheal extubation and hospital discharge.
The Perfusion Department and the nursing staff in both the Bristol Heart Insitute, Bristol, United Kingdom, and the Clinica Montevergine, Mercogliano, Italy, are thanked for their support.
Back to Top | Article Outline


1. Bridgewater B, Keogh B: Surgical “league tables”: Ischaemic heart disease. Heart 2008; 94:936–42

2. Goldstein S, Dean D, Kim SJ, Cocozello K, Grofsik J, Silver P, Cody RP: A survey of spinal and epidural techniques in adult cardiac surgery. J Cardiothorac Vasc Anesth 2001; 15:158–68

3. Scott NB, Turfrey DJ, Ray DA, Nzewi O, Sutcliffe NP, Lal AB, Norrie J, Nagels WJ, Ramayya GP: A prospective randomised study of the potential benefits of thoracic epidural anesthesia and analgesia in patients undergoing coronary artery bypass grafting. Anesth Analg 2001; 93:528–35

4. Priestley MC, Cope L, Halliwell R, Gibson P, Chard RB, Skinner M, Klineberg PL: Thoracic epidural anesthesia for cardiac surgery: The effects on tracheal intubation time and length of hospital stay. Anesth Analg 2002; 94:275–82

5. Buggy DJ, Smith G: Epidural anaesthesia and analgesia: Better outcome after major surgery? Growing evidence suggests so. BMJ 1999; 319:530–1

6. Humphreys N, Bays SM, Parry AJ, Pawade A, Heyderman RS, Wolf AR: Spinal anesthesia with an indwelling catheter reduces the stress response in pediatric open heart surgery. Anesthesiology 2005; 103:1113–20

7. Lattermann R, Carli F, Schricker T: Epidural blockade suppresses lipolysis during major abdominal surgery. Reg Anesth Pain Med 2002; 27:469–75

8. Ganapathy S, Murkin JM, Dobkowski W, Boyd D: Stress and inflammatory response after beating heart surgery versus conventional bypass surgery: The role of thoracic epidural anesthesia. Heart Surg Forum 2001; 4:323–7

9. Kirnö K, Friberg P, Grzegorczyk A, Milocco I, Ricksten SE, Lundin S: Thoracic epidural anesthesia during coronary artery bypass surgery: Effects on cardiac sympathetic activity, myocardial blood flow and metabolism, and central hemodynamics. Anesth Analg 1994; 79:1075–81

10. Loick HM, Schmidt C, Van Aken H, Junker R, Erren M, Berendes E, Rolf N, Meissner A, Schmid C, Scheld HH, Möllhoff T: High thoracic epidural anesthesia, but not clonidine, attenuates the perioperative stress response via sympatholysis and reduces the release of troponin T in patients undergoing coronary artery bypass grafting. Anesth Analg 1999; 88:701–9

11. Liem TH, Hasenbos MA, Booij LH, Gielen MJ: Coronary artery bypass grafting using two different anesthetic techniques: Part 2: Postoperative outcome. J Cardiothorac Vasc Anesth 1992; 6:156–61

12. Naesh O, Haljamäe H, Hindberg I, Holm J, Jivegård L, Wennmalm A: Epidural anaesthesia prolonged into the postoperative period prevents stress response and platelet hyperaggregability after peripheral vascular surgery. Eur J Vasc Surg 1994; 8:395–400

13. Tenling A, Joachimsson P-O, Tydén H, Hedenstierna G: Thoracic epidural analgesia as an adjunct to general anaesthesia for cardiac surgery. Effects on pulmonary mechanics. Acta Anaesthesiol Scand 2000; 44:1071–6

14. Hedenstierna G, Tenling A: The lung during and after thoracic anaesthesia. Curr Opin Anaesthesiol 2005; 18:23–8

15. Pitsis AA, Angelini GD: Off pump coronary bypass grafting of the circumflex artery. Eur J Cardiothorac Surg 1999; 16:478–9

16. Angelini GD, Taylor FC, Reeves BC, Ascione R: Early and midterm outcome after off-pump and on-pump surgery in Beating Heart Against Cardioplegic Arrest Studies (BHACAS 1 and 2): A pooled analysis of two randomised controlled trials. Lancet 2002; 359:1194–9

17. Miceli A, Capoun R, Fino C, Narayan P, Bryan AJ, Angelini GD, Caputo M: Effects of angiotensin-converting enzyme inhibitor therapy on clinical outcome in patients undergoing coronary artery bypass grafting. J Am Coll Cardiol 2009; 54:1778–84

18. Murphy GJ, Reeves BC, Rogers CA, Rizvi SI, Culliford L, Angelini GD: Increased mortality, postoperative morbidity, and cost after red blood cell transfusion in patients having cardiac surgery. Circulation 2007; 116:2544–52

19. Stenseth R, Bjella L, Berg EM, Christensen O, Levang OW, Gisvold SE: Thoracic epidural analgesia in aortocoronary bypass surgery. I: Haemodynamic effects. Acta Anaesthesiol Scan 1994; 38:826–33

20. Moore CM, Cross MH, Desborough JP, Burrin JM, Macdonald IA, Hall GM: Hormonal effects of thoracic extradural analgesia for cardiac surgery. Br J Anaesth 1995; 75:387–93

21. Chakravarthy M, Nadiminti S, Krishnamurthy J, Thimmannagowda P, Jawali V, Royse CF, Minzter BH: Temporary neurologic deficits in patients undergoing cardiac surgery with thoracic epidural supplementation. J Cardiothorac Vasc Anesth 2004; 18:512–20

22. Gonca S, Kilickan L, Dalcik C, Dalcik H, Bayindir O: The cardioprotective effects of thoracal epidural anaesthesia are induced by the expression of vascular endothelial growth factor and inducible nitric oxide synthase in cardiopulmonary bypass surgery. J Cardiovasc Surg 2007; 48:93–102

23. Royse CF: High thoracic epidural anaesthesia for cardiac surgery. Curr Opin Anaesthesiol 2009; 22:84–7

†† Available at: Accessed August 17, 2010. Cited Here...

Cited By:

This article has been cited 3 time(s).

Cochrane Database of Systematic Reviews
Epidural analgesia for cardiac surgery
Svircevic, V; Passier, MM; Nierich, AP; van Dijk, D; Kalkman, CJ; van der Heijden, GJ
Cochrane Database of Systematic Reviews, (6): -.
ARTN CD006715
Interactive Cardiovascular and Thoracic Surgery
Superior haemodynamic stability during off-pump coronary surgery with thoracic epidural anaesthesia: results from a prospective randomized controlled trial
Rajakaruna, C; Rogers, C; Pike, K; Alwair, H; Cohen, A; Tomkins, S; Angelini, GD; Caputo, M
Interactive Cardiovascular and Thoracic Surgery, 16(5): 602-607.
Epidurals for Cardiac Surgery: Can We Substantially Reduce Surgical Morbidity or Should We Focus on Quality of Recovery?
Royse, C
Anesthesiology, 114(2): 232-233.
PDF (111) | CrossRef
Back to Top | Article Outline

Supplemental Digital Content

Back to Top | Article Outline

© 2011 American Society of Anesthesiologists, Inc.

Publication of an advertisement in Anesthesiology Online does not constitute endorsement by the American Society of Anesthesiologists, Inc. or Lippincott Williams & Wilkins, Inc. of the product or service being advertised.

Article Tools