Best Evidence Topics
  • Send this BET as an Email
  • Make a Comment on this BET

In patients undergoing cardiac surgery, thoracic epidural analgesia combined with general anesthesia results in faster recovery and fewer complications but does not affect length of hospital stay

Three Part Question

In [adult patients undergoing cardiac surgery], is the use of [Thoracic epidural anaesthesia] associated with [faster recovery, fewer complications and shorter length of stay]

Clinical Scenario

You are the anaesthetist assessing a 75-year-old current smoker with an exercise tolerance of only 200 yds, who is due to have coronary grafting tomorrow. You approach the surgeon to suggest that you place a thoracic epidural prior to induction. He is reluctant for you to do this as he says that epidurals can be dangerous with full heparinisation and anyway he isn't aware of any evidence that it speeds recovery. You decide to summarize the evidence for him.

Search Strategy

Medline 1966 to October Week 3 2005 using OVID interface
EMBASE 1980 to 2005 Week 43
[exp Cardiopulmonary Bypass/ OR OR exp Thoracic Surgery/ OR exp Cardiac surgical procedures/ OR Coronary art$ OR Cardiopulmonary OR exp Cardiopulmonary Bypass/ OR exp Cardiovascular Surgical Procedures/ OR exp Thoracic Surgical Procedures/ OR exp Coronary Artery Bypass/ OR cardiac OR exp Heart Transplantation/ ] AND [exp Anesthesia, Epidural/ OR Epidural OR exp Analgesia, Epidural/ OR epidural OR exp Anesthesia, Conduction/ae, cl, st, td, mo OR exp Anesthesia, Local/ad, ae, ct, st, td, mo OR exp Anesthesia, Spinal/ae, ct, st, td, mo OR exp Nerve Block/ae, ct, st, td, mo] AND [exp ANESTHESIA RECOVERY PERIOD/ OR exp "RECOVERY OF FUNCTION"/ OR "Length of Stay"/ OR exp Intraoperative Complications/ OR exp Postoperative Complications/ OR surgical OR *"Postoperative Complications"/] LIMIT [humans AND english language]

Search Outcome

Using the reported search, 313 papers were identified on Medline and 368 on Embase. A further paper was found by hand-searching of reference lists. Fifteen papers represented the best evidence on the subject and are summarised below. A further 6 papers were not summarised due to small study size, non-randomised design and our own space constraints (Kessler, Hemmerling, Tenling, Senseth, Liem, Liem).

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
Liu et al,
Meta-analysis of 15 trials enrolling 1178 patients over period 1996 to 1st Jan 2004 obtained from searching Medline, American College of Physicians Journal Club, Cochrane database of Systematic reviews, and Database of Abstracts of Reviews of EffectsMeta-analysis (level 1a)Odds of deathTEA 0.7% vs. GA 0.3% OR 1.56 (0.35–6.91) P=0.56. 1178 ptsAlmost all the RCT's were small including 3 studies with <20 participants Pooled numbers still too few to determine effects on myocardial ischaemia and mortality 7 studies reported in the table below not part of the meta-analysis
Odds of MITEA 2.3% vs. GA 3.4% OR 0.74 (0.34–1.59); P=0.44. 1026 pts
Odds of suffering dysrhythmias (AF or tachycardia)TEA 17.8% vs. GA 30% OR 0.52 (0.29–0.93) P=0.03. 913 pts
Odds of pulmonary complicationsTEA 17.2% vs. GA 30.3% OR 0.41 (0.27–0.60) P<0.00001. 644 pts
Odds of time to tracheal extubation (hours)TEA 6.9 h vs. GA 10.4 h Mean Difference –4.5 (–7 to –2) P=0.0005. 905 pts
Odds of postoperative pain (Visual analogue scale)TEA 12.4 mm vs. GA 19.6 mm Mean Difference –7.8 mm; (–15 to –0.6); P=0.03. 392 pts
Incidence of neurological injuryNo cases of spinal/epidural haematoma
Scott et al,
420 patients undergoing elective CABG surgery Standard GA with 72hrs narcotic analgesia v standard GA with 96hrs TEA Elective CABG pts with normal preoperative coagulation status Epidurals inserted at T2/3 or T3/4 in OR immediately preoperatively – operation postponed if "bloody tap" obtained Bilateral T1-T10 block instituted with Bupivacaine 0.5% and maintained with Bupivacaine / Clonidine infusion perioperatively +/- additional Bupivacaine top-ups postop.PRCT (level 1b)Time to extubation (to achieve standard unit extubation criteria)51/206 (25%) TEA vs.11/202 (5.5%) GA extubated within 4 h (P<0.0001)Avoidance of Beta-blockade during study period unless clinically indicated might be confounding factor in development of dysrhythmias 73/202 (36%) GA v 100/202 (49%) TEA had no preoperative complications p=0.012
Incidence of new supraventricular arrhythmias21/206 (10.2 %) TEA vs. 45/202 (22.3%) GA P=0.001
Incidence of proven respiratory tract infection31/206 (15.3%) TEA vs. 59/202 (29.2%) GA group developed LRTIs P=0.0007
Maximal inspiratory lung volume (ILV)-subset analysis of 47 TEA & 46 GA ptsMaximal ILV better in TEA vs. GA group – 985± 326 ml vs. 733±208 ml; P<0.0001
Overall any complication rate in 1st 5 days postoperatively84/206 (40.8%) TEA vs. 108/202 (53.5%) GA P=0.011
Neurological sequelaeNo permanent neurological complications associated with use of TEA
Turfey et al,
218 consecutive patients undergoing CABG surgery 118/218 Standard GA+target-controlled Alfentanil for 1st 24hrs post-op, then IV PCA Morphine for 48hrs 100/218 Standard GA + TEA. TEA instituted with 0.5% Bupivacaine and maintained with a Bupivacaine / Clonidine infusion Epidurals inserted immediately before induction of anaesthesia and run for 5 days postoperatively TEA not used in presence of abnormal coagulation (Aspirin only not a contraindication)Retrospective Cohort (level 2b)Time to tracheal extubation21% TEA vs. 2% GA extubated within 1 h of surgery (P<0.002)Not a randomised study Decision to use TEA based on Anaesthetist's preference Study from same unit as that of Scott et al
Incidence of postoperative arrhythmias18/100 (18%) TEA vs. 38/118 (32%) GA (P=0.02)
Incidence of respiratory complication16% TEA vs. 21% GA pts P=NS
Other complicationsNo significant differences in incidence of other renal, GI or haematological complication. No significant difference in cardiac complications, difference in IABP use/ no VAD use. No difference in inotrope use
Incidence of neurolgical complication6/100 (6%) TEA vs. 9/118 (7.6%). GA pts had neurological complication– not significant. TEA group CNS complications – one leg weakness (no epidural haematoma on MRI)
Barrington et al,
120 pts undergoing elective CABG surgery Comparison of high thoracic epidural anesthesia / analgesia + general anaesthetic (HTEA) to general anaesthesia only (GA) Pts with abnormal coagulation, recent Aspirin ingestion or evidence of neurological dysfunction excluded from study HTEA group – epidural inserted at T1/2 or T2/3 the day before surgery and established with Ropivacaine / Fentanyl GA group – standard GA with postoperative morphine infusionPRCT (level 1b)Time to tracheal extubationMean time to extubation 15 min HTEA vs. 430 min GA (P<0.0001)

60% HTEA pts vs. 5% GA pts extubated within 30 min of completion of surgery (P<0.0001)

Time to extubation beyond 30 min not significantly different HTEA vs. GA
Significant higher incidence of peripheral vascular & cerebrovasc disease in HTEA v GA group (p=0.016 & 0.002 respectively) Group allocations not masked – might bias towards early extubation
VAS assessment of postoperative analgesiaHTEA pain scores significantly less than GA scores at all time points with/without coughing
Cardiac troponin (cTnI) & CK-MB levels preop and 12 and 24 h post-cross-clamp releasecTnI levels HTEA vs. GA – both raised at 12 & 24 h but no significant difference between groups
12 lead ECGs preop and on postop days 1 & 5 (2 blinded observers)3/60 (5%) HTEA vs. 5/60 (8.3%) GA pts developed new ECG changes 1/60 (1.6%) HTEA vs. 2/60 (3.3%) GA pts had evidence of transmural MI on basis of ECG but no difference in incidence of myocardial ischaemia on basis of ECG or biochemical criteria
Neurological sequelaeNo neurological complications related to HTEA
Priestley et al,
100 pts randomised to either standard GA + postop IV Morphine PCA or standard GA + TEA for 48hrs postop, then epidural PCA Meperidine for further 24hrs TEA inserted on night before surgery at T1-T4 and induced with Ropivacaine / Fentanyl. Surgery NOT postponed in presence of "bloody tap" but attempt made to resite elsewhere Exclusions - pre-existing coagulopathy - anticoagulation - potent anti-platelet drugs - systemic / local infection - arthritis of spine with history of associated neurological deficit - cognitive impairment Aspirin & other NSAID's usually discontinued 1 week before surgery but not a contraindication to TEAPRCT (level 1b)Time to tracheal extubationTEA vs GA medians 3.2 vs. 6.7 h (P<0.0001)Study powered on basis of length of hospital stay Male preponderance in both groups (5-7:1) Decision to discharge made by unblended surgeons Whilst study did not demonstrate earlier discharge in TEA group it did demonstrate earlier discharge of both study TEA & control groups in comparison to non-study pts
Time to achieve standard mobilisation goal of ability to climb 2 flights of stairs unaidedNo significant difference TEA vs GA in time to achieve standard mobilisation goals, criteria for discharge, and actual time to discharge
Time to achieve standard unit discharge criteria24% of total cohort discharged on Day 4 & 33% on Day 5 compared to 2% & 18 of pts outwith study population
Actual time to dischargeNo significant differences
Respiratory functionNo significant differences
Atrial fibrillation10/50 (20%)GA vs. 11/50 (22%) TEA developed AF
Postoperative pain VAS scoresAnalgesia better in TEA group at rest and during coughing but statistically significant during 1st postop 24 h only - 16/50 (32%) GA vs 6/50 (12%) TEA had VAS >40 mm (P=0.015)
TEA complicationsNo major complications related to epidural technique 4/50 (8%) epidurals 'failed' 3/50 (6%) TEA vs 2/50 (4%) GA pts developed postop urinary catheterisation
Royse et al,
80 CABG pts randomised to GA+high TEA (HTEA) or GA+PCA Morphine HTEA inserted at T1/2 or T2/3 on night before surgery. Induced & maintained with Ropivacaine / Fentanyl until 3rd postop dayPRCT (level 1b)Pain on 1st and 2nd postop days assessed by 10cm VAS at rest and with coughingSignificantly better pain scores with HTEA vs GA (mean cm scores): Day 1 Rest 0.02 vs. 0.8 P=0.032. Day 1 Cough 1.2 vs. 4.4 P<0.001 Day 2 Rest 0.1 vs. 1.2 P=0.044 Day 2 Cough 1.5 vs. 3.6 P=0.005
Post operative ventilation/hospital stayHTEA vs. GA 2.6± 2.5 h vs. 5.4±3.1 h (P<0.001) Length of ITU/hospital stay – no significant difference HTEA vs. GA Postop creatinine/atrial fibrillation – no significant difference HTEA vs. GA
Assessment of lung function - preop and daily spirometry and SpO2 on air - arterial blood gases on postop day 1Lung function lower both groups - – O2 higher in HTEA than GA group; P=0.041 (P=0.041 corrected)1 – PEFR higher in HTEA than GA group on all days incl Day 3 when epidural discontinued P=0.001 (0.003 corrected). 1. P values corrected using technique to test 6 hypotheses
Nygard et al,
163 pts scheduled for CABG surgery 2x2 factorial study to compare high TEA, Amiodarone therapy or both against control in preventing postop AF Exclusions - recent use of Amiodarone - Amiodarone toxicity - Warfarin therapy or coagulopathy - use of antidysrhythmics (other than Beta blockers, digoxin & calcium-channel blockers) Group E - high TEA Group A - Amiodarone Group E+A - high TEA + Amiodarone Group C - Control Epidural groups - epidural inserted at T1-3 the day before surgery and maintained for 96hrs. Initiated with Bupivacaine & maintained with Bupivacaine / Morphine Amiodarone groups - 1800mg orally, then IV infusion commencing after induction of anaesthesia for 3 days - 900mg/24hrs All patients received standard GA and non-TEA pts IV morphine analgesia 5 days of continuous Holter monitoring - automated & blinded manual reportingPRCT (level 1b)Incidence of new onset AF - 5 minGroup E - 22/44 (50%). Group A - 10/36 (27.8%). Group E+A - 10/35 (28.6%). Group C - 20/48 (41.7%). No signif difference in AF incidence between TEA (E & E +A) groups and non-TEA (A & C) groups P=0.538

No significant difference in incidence of AF requiring treatment in TEA E & E+A) vs. non-TEA (A & C) groups - 28/79 (35%; 95% CI 25 - 47%) vs. 25/84 (30% 95% CI 20 - 41%) P=0.66
Jideus et al,
141 patients undergoing elective CABG 45/141 randomly selected for standardised GA with intraoperative & postoperative TEA; 96/141 control group TEA inserted at T2/3, T3/4 or T4/5 the day before surgery and established with Bupivacaine. TEA postoperatively maintained with Bupivacaine / Sufentanil for 96hrs or until onset of sustained AF 96/141 control patients ‘V standardised GA only with postop Ketobemidone analgesia. Both groups monitored with Holters for 24hrs preop & 96hrs postopPRCT (level 1b)Incidence of sustained AF >30 s on Holter13/41 (31.7%) TEA vs. 29/80 (36.3%) GA (P=0.77)

Onset time of AF 1.9 days (TEA) vs. 2.2 days (GA) post-op P=0.23
Lack of information about randomisation process Control group twice the size of treatment group Exclusions in control group may have affected AF outcomes Beta-blockers more likely to be withdrawn in TEA group ‘V might affect AF outcomes
Pre and postop levels of Neuropeptide Y (NPY), Chromogranin A (CgA), Chromogranin B (CgB) Pancreatic polypeptide (PP) Epinephrine NorepinephrinePostop NPY levels significantly lower in TEA vs. GA group(P=0.002 vs. P=0.60). CgA & CgB levels unchanged postop in both groups PP significantly lower in both TEA & GA groups postop (P=0.04 & P=0.0001). Raised Epinephrine in both groups in day 1 though less marked in TEA group (TEA P=0.002 vs. GA P<0.001). Pre & postop Norepinephrine levels unchanged in TEA group (P=0.73) but significant increase in GA group (P=0.002)
Salvi et al,
106 consecutive pts undergoing off-pump surgery under combined TEA & general anaesthesia TEA performed immediately prior to anaesthesia at T1/2 or T2/3 and induced and maintained with Ropivacaine / Sufentanil to block C7/T1 to T6/T7. Catheter removed on 3rd postop day Exclusion criteria - abnormal coag studies - abnormal platelet counts - recent anticoagulation / thrombolysis - coagulation disorders - neuroaxial pathology - recent acute infarctRetrospective cohort (level 2b)Time to extubationMean time to extubation 4.6± 2.9 h (median 4; range 0.5–12 h)Uncontrolled cohort group – no comparison group TEA unsuccessful in 2/106 (1.8%) pts Bloody tap in 1/106 (0.9%) pts No dural taps
Length of ITU stayMean ITU stay 1.5± 0.8 days (median 1.0; range 0.5–6 days)
Incidence of MI4/106 (3.8%) pts presented with signs of myocardial ischaemia – attenuated/abolished by institution of TEA
Incidence of AF11/206 (10.6%) developed AF during ITU stay
Assessment of pain using Visual analogue scoreGood analgesia at rest and with coughing
Neurological complications1/106 (0.9%) pts developed postop paraplegia - emergency MRI excluded 2° to spinal/epidural haematoma
Chakravarthy et al,
2113 adults undergoing cardiac surgery (CABG, valve surgery, congenital surgery) under general + epidural anaesthesia over 13 year period Epidural catheter inserted at C7 to T3 level day evening before surgery in patients with normal coagulation studies and platelet number, no anti-platelet therapy for 1 week and heparin discontinuation for 6-12 hours pre-insertion. Epidural avoided in patients on anti-platelet therapy, with coagulation problems, with previous surgery to cervical or thoracic spine or local infection. Evolving epidural regime over time using combinations of Lidocaine+/-Epinephrine / Bupivacaine / Morphine / Buprenorphine / FentanylCohort (level 2b)Dural puncture19/2224 (0.85%) – unable to ID epidural space/insert catheter –18/19 – dural puncture –10/18 resited successfully

22/2224 (0.98%) –haemorrhagic tap –12/22 resited successfully

2113 epidurals performed
No outcome parameters Audit designed to look at complications, not outcomes No standard Epidural or GA regime – regime evolved over study period
Complications4/2113 – develpoed temporary neurological deficits (transient monoplegias) of up to 24 h which recovered on withdrawal/removal of epidural catheter

No other neurological sequelae
Sanchez et al,
571 consecutive pts scheduled for CABG over 2 year period Pts not given epidural if on Warfarin; with a prothrombin-proconvertin<50%; or on Aspirin within last 7 days Epidural inserted day before surgery at T1-T3 level. Induced with Bupivacaine & maintained with Bupivacaine / Fentanyl. Standard GACohort (level 2b)ExtubationMedian time to extubation 345 min (range 75–1305 min)13/571 (2%) epidurals abandoned - 12 uncertain blocks + one unsuccessful catheter placement 558/571 (98%) epidurals successful 1/558 dural puncture – surgery postponed. After 1 day without complication, epidural resited and surgery proceeded 403/558 (72%) on antiplatelet medication
ITU stayMedian duration of ITU stay 1 day (range 1–13 days)

445/558 (80%) epidurals left for 4–5 days
ComplicationsNo documented spinal haematomas causing neurolgoical symptoms. No severe radiating back pains. No unexplained/progressive neurological deficits
Pastor et al,
714 pts undergoing CABG under combined TEA + standard GA over 7yr period TEA inserted between T1-T3 immediately preop & induced and maintained with Bupivacaine or Ropivacaine periop & Ropivacaine / Fentanyl postop. Inclusion criteria - APTT<45 secs - PT (INR) < 1.5 - Platelets > 80,000/μl - adequate (7 days ) preop suspension of aspirin & clopidigrol Exclusion criteria - pre-existing neurological disorder - failed block - infection at puncture site Technical criteria - max 3 attempts at TEA - midline approach - TEA at least 60 mins before heparinisation - MRI / neurosurgery availability TEA usually terminated at 4 days - normal coagulation - before therapeutic anticoagulation - >12hrs after last but >4hrs before next dose of LMW heparin if usedProspective cohort (level 2b)Technical complications of TEA6/714 (0.84%) dural punctures, no complaints of pain/ paraesthesia; no 'bloody taps'

6/6 (100%) dural punctures successfully resited elsewhere. No postop complications in this group
Observational study designed to confirm that adherence to standard safety measures adverts risk of epidural haematoma No comparison of TEA group to control group
Time to extubation (standard unit criteria)Median time to extubation 0 h (0–168) 496/714 (75.3%) extubated in the OR
Length of ITU stayMedian length of ITU stay 48 h (22–456 h)
Length of hospital stayMedian length of hospital stay 8 days (3–121)
General morbidity and mortality26/714 (3.6%) pts required, prolonged IPPV, reintubation or tracheostomy
Complications of TEA3/714 (0.42%) awoke with suspicious motor findings and underwent MRI scanning

No evidence of epidural haematoma or sub-acute/chronic space occupying processes
Ho et al,
Hong Kong
4583 patients undergoing cardiac surgery utilising TEA in period up to 1999 reported in previously published trials Assessment of risk of epidural haematoma / spinal injury using Hanley & Lippman-Hand mathematical modelling techniqueCohort (level 2b)Mathematical calculation of statistical risk of epidural haematoma / spinal injuryRisk of spinal injury from epidural haematoma during TEA between 1:150,000 and 1:1,500 (95% confidence) and up to 1:1,000 (99% confidence)Remains a mathematical estimation
Berendes et al,
73 pts scheduled for CABG surgery randomised to GA only (control) or GA+TEA TEA inserted day before surgery at C7/T1 level & initiated with Bupivacaine 0.5% & Sufentanil and maintained with Bupivacaine Exclusions - abnormal coagulation - renal insufficiency - ventricular dysfunction (EF<50%) - valvular surgery - heart failurePRCT (level 1b)Regional LV function by transoesophageal echo (TOE)37/73 GA only vs. 36/73 GA+ Postop LV function scores significantly better in TEA vs. GA pts: 0.78 vs. 0.38; P<0.05Pts with pre-existing ventricular dysfunction excluded Presentation of TOE, ANP & BNP graphical with little absolute numerical data Although 2 year mortality a 2° outcome study not powered to detect significant difference
Pre & postop cardiac troponin IPeak Cardiac Troponin I significantly less in TEA vs. GA pts 1.6 vs. 5.7 ng/ml; P<0.05
Pre, intra & postop Atrial & Brain natriuretic peptideANP maximally raised during reperfusion after cross-clamping – TEA raised significantly less than GA increase (98 vs. 211 ng/ml; P=0.03)

BNP increase 6 h post-ITU admission & peaked 24 h postop – TEA significantly attenuated BNP peak – mean BNP peak TEA vs. GA 108 vs. 189 ng/ml (P=0.01)
Extubation timeMean time to extubation 3.4 vs. 9.2 h
ComplicationsNo complications related to TEA; No other significant complications
Filinger et al,
60 pts undergoing cardiac surgery Excluded if any contraindication to TEA (heparin, warfarin, coagulopathy, local/systemic infection) 30/60 GA+TEA Epidural inserted immediately before induction of anaesthesia and induced with morphine/bupivacaine 30/60 GA+postop IVI MorphinePRCT (level 1b)Time to tracheal extubationTime to tracheal extubation GA vs. TEA 9.5 vs. 10.7 h (P=0.49)Study powered using total hospital costs as proxy for hospital outcome
Duration of ITU stayITU stay GA vs. TEA 30 vs. 31.7 h (P=0.84) Hospital stay GA vs. TEA 160 vs. 156 h (P=0.84)
Duration of hospital stayHospital stay GA vs. TEA 160 vs. 156 h (P=0.84)
Pain control using 100mm VASsPain control on 1st postop day - no significant difference between GA and TEA groups
Urinary free cortisolUrinary free cortisol GA vs. TEA 120 vs. 150 ΅g/dl (P=0.25)
Total hospital chargesTotal charges GA vs. TEA $36,955 vs. $40,026 (P=0.20
Cardiopulmonary complication rate20/30 (67%) GA vs. 30/30 (100%) TEA group required catecholamines post-CPB (P=0.0005)

Incidence of AF 23% in both groups

No difference in incidence of ischaemia between groups

9/30 (30%) GA developed 12 'complications'; 12/30 (40%) TEA developed 17 'complications'

No neurological sequelae related to TEA


Various benefits have been reported in combining thoracic epidural anaesthesia (TEA) with general anaesthesia (GA) during cardiac surgery. Liu et al performed a meta-analysis in 2004 which studied outcomes of 1178 patients from 15 studies including eight identified in our search. They found that the use of TEA led to earlier extubation (OR -4.5; 95%CI -7 to -2; p=0.0005) and improved pain scores at rest (OR -7.8; 95%CI -15 to -0.6; p=0.03) and with activity (OR -11.6; 95%CI -19.7 to -3.5; p=0.005) compared to GA alone. Scott et al, Turfrey et al, Barrington et al, Priestley et al, and Royse et al have all confirmed these findings. The meta-analysis also identified a decrease in pulmonary complications (OR 0.41; 95%CI 0.27-0.60; p<0.00001) with Scott reporting a lower incidence of chest infection and better maximal inspiratory lung volume, and Royse identifying better postoperative lung function and cooperation with physiotherapy in their TEA groups. However Priestly et al reported no difference in blood gases, spirometry or chest X-ray appearances in their postoperative TEA group. Liu also reported decreased dysrhythmias with TEA (OR 0.52; 95%CI 0.29-0.93; p=0.03) although there were no significant differences in myocardial infarction and mortality. But whilst Scott and Turfrey both identified lower incidences of atrial fibrillation (AF) with TEA, Barrington, Priestley and Royse did not. However Scott discontinued Beta-blockade postoperatively which may be confounding factor and it has been suggested that his use of Clonidine as an epidural adjunct might enhance TEA anti-dysrhythmic effect. Jideus et al recognised that whilst TEA suppressed sympathetic activity and protected against postoperative tachycardia and supraventricular premature beats, there was no difference in the proportion of patients developing AF. Nygard's 2x2 factorial studied the effects of TEA, Amiodarone or both as prophylaxis against AF and found Amiodarone but not TEA to be protective. A number of studies have looked at biochemical markers of myocardial injury in patients undergoing cardiac surgery utilising TEA. Barrington and Priestly measured Cardiac Troponin and CK-MB levels postoperatively and Fillinger et al looked at urinary free cortisol but there were no significant differences between TEA and control groups. Jideus measured Neuropeptide Y (NPY) levels and identified significant falls in their TEA group postoperatively versus control. Berendes et al combined Troponin I, ANP and BNP measurement with transoesophageal echo (TOE) assessment of LV function and identified differences in ANP and BNP responses to surgery in their TEA group. They also recognised that postoperative LV function was better in the TEA group and identified a correlation between peak BNP, Troponin and TOE-derived LV function motion score implying a possible prolonged postoperative benefit. Other effects have been noted. Scott's group identified significant decreases in acute confusional states and renal failure in their TEA group but whilst there was a lower incidence of stroke this did not reach statistical significance . Royse noted lower rates of post-traumatic stress disorder and postoperative depression in those patients receiving TEA. Salvi et al identified a small subgroup of patients whose developing myocardial ischaemia was attenuated / abolished by institution of TEA. Not surprisingly Fillinger reported that TEA patients required more vasoconstrictor and catecholamine but less vasodilator therapy than control patients. There is currently no evidence to suggest that use of TEA is associated with earlier hospital discharge although this in part may be confounded by the populations studied. For example 60% of Scott's study population were in receipt of allowances if they stayed in hospital! In Priestley's study, both their TEA and control groups were discharged earlier than non-study patients. Fillinger et al found no significant cost benefits in using TEA. There is no agreement on the optimal TEA drug regime. Bupivacaine and Ropivacaine are widely used and whilst most authors used opioid supplementation, the two studies reporting greatest anti-dysrhythmic effect used Clonidine. The major disadvantage of TEA remains the fear of epidural haematoma in the soon-to-be anticoagulated patient. Whilst a small number of individual cases have been reported, increasing number of audits and cohort studies including those by Salvi, Chakravarthy, Sanchez, and Pastor which specifically looked for neurological complications seem to confirm the apparent safety of the technique when basic sensible precautions are taken. Chakravarthy et al reported only 4 temporary neurological deficits in a large series of 2113 epidurals in cardiac surgery patients over a 13 year period . A small proportion of their patients experienced dural or bloody taps but in most cases epidurals were resited and surgery proceeded. Likewise Pastor et al reported a low incidence of dural puncture although three of his series and one each of Scott's, Turfrey's and Salvi's patients underwent negative MRI scanning after developing suspicious lower limb signs. There were no other unexplained neurological deficits or signs of epidural haematoma in any study. In 2000 Ho et al used mathematical modelling to calculate the statistical risk of epidural haematoma following TEA as between 1:150,000 and 1:1,500 (95% confidence) and up to 1:1,000 (99% confidence) based on published data available at that time. They also acknowledged the importance of factors including presence of a normal coagulation system prior to insertion and removal of the catheter, avoidance of repeated attempts, postponing surgery in the presence of a bloody tap and close neurological surveillance in minimising risk. These precautions are repeated in most studies. It has also been suggested that timing of catheter insertion is important with Barrington, Priestly, Royse, Nygard, Jideus, Berendes, Chakravarthy and Sanchez advocating insertion the day before surgery and Scott, Turfrey, Fillinger, Salvi and Pastor on the day of surgery. Finally Scott et al recommended that TEA should only be used in an institution with immediate MRI availability to facilitate rapid diagnosis of epidural haematoma and also recognised the importance of using anaesthetic drugs with rapid offsets to facilitate early neurological assessment.

Clinical Bottom Line

TEA in combination with general anaesthesia results in more rapid extubation and significantly better pain relief in patients undergoing cardiac surgery. There may also be a lower incidence of respiratory, cardiovascular, renal and psychological complications although the length of hospital stay is not affected. The optimal drug regime has probably not been established. Whilst there is a theoretical risk of epidural haematoma and a few have been reported the risks appear small on the basis of the evidence available and are minimised by the use of appropriate well-documented basic sensible precautions to ensure normal haemostasis prior to catheter insertion and removal.


  1. Liu SS, Block BM, Wu CL. Effects of perioperative central neuroaxial analgesia on outcome after coronary artery bypass surgery: a meta-analysis. Anesthesiology 2004; 101: 153-61
  2. Scott NB, Turfrey DJ, Ray DA, Nzewi O, Sutcliffe NP, Lal AB, Norrie J, Nagels WJ, Ramayya GP. A prospective randomized study of the potential benefits of thoracic epidural anesthesia and analgesia in patients undergoing coronary artery bypass grafting. Anesthesia & Analgesia. 2001; 93: 528-35.
  3. Turfrey DJ, Ray DA, Sutcliffe NP, Ramayya P, Kenny GN, Scott NB. Thoracic epidural anaesthesia for coronary artery bypass graft surgery. Effects on postoperative complications. Anaesthesia 1997; 52: 1090-5.
  4. Barrington MJ, Kluger R, Watson R, Scott DA, Harris KJ. Epidural anesthesia for coronary artery bypass surgery compared with general anesthesia alone does not reduce biochemical markers of myocardial damage. Anesthesia & Analgesia 2005; 100: 921-8.
  5. 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. Anesthesia & Analgesia 2002; 94: 275-82.
  6. Royse C, Royse A, Soeding P, Blake D, Pang J. Prospective randomized trial of high thoracic epidural analgesia for coronary artery bypass surgery. Annals of Thoracic Surgery 2003; 75: 93-100.
  7. Nygard E, Sorensen LH, Hviid LB, Pedersen FM, Ravn J, Thomassen L, Svendsen JH, Eliasen K, Krogsgaard K, Aldershvile J. Effects of amiodarone and thoracic epidural analgesia on atrial fibrillation after coronary artery bypass grafting. Journal of Cardiothoracic & Vascular Anesthesia 2004; 18: 709-14.
  8. Jideus L, Joachimsson PO, Stridsberg M, Ericson M, Tyden H, Nilsson L, Blomstrom P, Blomstrom-Lundqvist C. Thoracic epidural anesthesia does not influence the occurrence of postoperative sustained atrial fibrillation. Annals of Thoracic Surgery 2001; 72: 65-71.
  9. Salvi L, Sisillo E, Brambillasca C, Juliano G, Salis S, Marino MR. High thoracic epidural anesthesia for off-pump coronary artery bypass surgery. Journal of Cardiothoracic & Vascular Anesthesia 2004; 18: 256-62.
  10. Chakravarthy M, Thimmangowda P, Krishnamurthy J, Nadiminti S, Jawali V. Thoracic epidural anesthesia in cardiac surgical patients: a prospective audit of 2,113 cases. Journal of Cardiothoracic & Vascular Anesthesia 2005; 19: 44-8.
  11. Sanchez R, Nygard E. Epidural anesthesia in cardiac surgery: is there an increased risk Journal of Cardiothoracic & Vascular Anesthesia 1998; 12:170-3.
  12. Pastor MC, Sanchez MJ, Casas MA, Mateu J, Bataller ML. Thoracic epidural analgesia in coronary artery bypass graft surgery: seven years' experience. Cardiothoracic & Vascular Anesthesia 2003; 17: 154-9.
  13. Ho AM, Chung DC, Joynt GM. Neuroaxial blockade and hematoma in cardiac surgery: estimating the risk of a rare event that has not (yet) occurred. Chest 2000; 117 : 551-550.
  14. Berendes E, Schmidt C, Van Aken, H. Hartlage MG, Wirtz S, Reinecke H, Rothenburger M, Scheld HH, Schluter B, Brodner G, Walter M. Reversible Cardiac Sympathectomy by High Thoracic Epidural Anesthesia Improves Regional Left Ventricular Function in Patients Undergoing Coronary Artery Bypass Grafting: A Randomized Trial. Archives of Surgery 2003; 138: 1283-1290.
  15. Kessler P, Aybek T, Neidhart G, Dogan S, Lischke V, Bremerich DH, Byhahn C. Comparison of three anesthetic techniques for off-pump coronary artery bypass grafting: general anesthesia, combined general and high thoracic epidural anesthesia, or high thoracic epidural anesthesia J Cardiothorac Vasc Anesth 2005; 19:132–39.
  16. Hemmerling TM, Prieto I, Choiniere JL, Basile F, Fortier JD. Ultra-fast-track anesthesia in off-pump coronary artery bypass grafting: a prospective audit comparing opioid-based anesthesia vs. thoracic epidural-based anesthesia. Can J Anaesth 2004; 51:2163–168.
  17. Tenling A, Joachimsson PO, Tyden H, Wegenius G, Hedenstierna G. Thoracic epidural anesthesia as an adjunct to general anesthesia for cardiac surgery: effects on ventilation-perfusion relationships. J Cardio-thorac Vasc Anesth 1999; 13:3258–264.
  18. Stenseth R, Bjella L, Berg EM, Christensen O, Levang OW, Gisvold SE. Effects of thoracic epidural analgesia on pulmonary function after coronary artery bypass surgery. Eur J Cardio-thorac Surg 1999; 10:10859–865.
  19. 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; 6:2156–161.
  20. Liem TH, Booij LH, Hasenbos MA, Gielen MJ. Coronary artery bypass grafting using two different anesthetic techniques: Part I: Hemodynamic results. J Cardiothorac Vasc Anesth 1992; 6:2148–155.
  21. Fillinger MP, Yeager MP, Dodds TM, Fillinger MF, Whalen PK, Glass DD. Epidural anesthesia and analgesia: effects on recovery from cardiac surgery. J Cardiothorac Vasc Anesth 2002; 16:15–20.