Editorial II

Does epidural analgesia improve surgical outcome?

J. C. Ballantyne1

1 Harvard Medical School, Massachusetts General Hospital Pain Center, 15 Parkman Street, WACC 333, Boston, MA 02114, USA E-mail: jballantyne{at}partners.org

{dagger}LMA® is the property of Intavent Limited.

When the safety and efficacy of neuraxial opioid administration was established in humans in the late 1970s,1 the utility of postoperative epidural analgesia increased markedly. The opioid provided good analgesia without the high doses of local anaesthetic that were previously used, thus allowing patients to retain their mobility while receiving effective epidural analgesia. As the use of epidurals for postoperative pain became more widespread, clinicians began to assess whether epidural analgesia would improve surgical outcome. The superior analgesia of epidural vs systemic analgesia during the postoperative period, at least after intra-abdominal and intrathoracic procedures, was quickly confirmed by multiple randomized trials.2 However, it was much more difficult to show benefit in terms of improvement in surgical outcome, and early trials had mixed results.

 Efforts to determine the true benefits of postoperative epidural analgesia are often confounded when epidural anaesthesia is used during surgery. It is common in clinical practice, and in trials, that epidurals placed for postoperative use are also used to produce dense neuraxial block or epidural anaesthesia during surgery. Postoperatively, low-dose local anaesthetics are used with opioids, and epidural analgesia takes over. I shall use the term epidural anaesthesia/analgesia when referring to intraoperative epidural anaesthesia combined with postoperative epidural analgesia. Epidural anaesthesia, with its accompanying dense sympathetic block, might logically improve cardiac function by reducing cardiac work, reduce thrombosis by improving lower extremity blood flow, and reduce stress responses, depending on the degree and level of both the sensory and sympathetic block. Postoperative epidural analgesia, on the other hand, could have different benefits, largely related to superior analgesia, continuous low-dose local anaesthetic effects, and avoidance of systemic opioids. They might include improved bowel mobility, improved coughing and breathing, earlier ambulation, and consequently a lower incidence of thrombosis. The distinction between the intra- and postoperative effects of epidurals becomes relevant when conducting and interpreting trials and meta-analyses.

When, in 1987, Yeager and colleagues3 reported significant improvement in morbidity and mortality in high-risk patients receiving epidural anaesthesia/analgesia, the anaesthesia community felt that it finally had strong support for using epidurals. This study was truncated at 55 patients by the monitoring committee because the early results favoured the epidural treatment so strongly that the committee felt it would be unethical to continue the trial. Subsequent questions about the validity of the findings from such a small study prompted other investigators to plan large, multi-centre trials assessing the role of epidural anaesthesia/analgesia in high-risk patients undergoing major surgery.46 The first was conducted in United States Veterans Affairs (VA) hospitals, incorporated 1021 high-risk patients, and found few differences in mortality or major (life-threatening) morbidity. The only significant differences were in a subgroup of 374 patients undergoing abdominal aortic surgery, who displayed a significant reduction in mortality, myocardial infarction, stroke, and respiratory failure.4 The second was conducted in Australia (the MASTER trial), incorporated 915 high-risk patients, and found no significant difference in mortality or major morbidity other than a reduction in the incidence of respiratory failure.5 The second phase of the MASTER trial analysis, in which the trial patients were stratified by risk category, was published in February 2003. Again, no difference in mortality was found, and the only benefit in terms of major morbidity was a reduction in respiratory failure, despite linking specific risks (e.g. cardiac) with specific outcomes.6 On the strength of these two large trials, can we assume we are mistaken in thinking that epidural anaesthesia/analgesia improves surgical outcome?

As discussed by Peyton and colleagues6 in their recent publication of the MASTER trial subset analyses, the results of previously published meta-analyses incorporating multiple small trials assessing benefits of epidurals for surgery, differed markedly from the results of larger trials. Targeted meta-analyses of epidural anaesthesia/analgesia had shown selected benefit—improved pulmonary function,7 and reduced myocardial infarction8—that was not confirmed by the large trials. A broader, comprehensive meta-analysis by Rodgers and colleagues,9 published in 2000, found that overall, spinals and epidurals reduce surgical mortality by 30% when compared with general anaesthesia without neuraxial block. While it is obvious that a meta-analysis that includes trials of mixed neuraxial interventions cannot be applied to the question of whether epidural anaesthesia/analgesia is effective in improving surgical outcome, the significance and size of the effect on mortality, and the fact that the trials of epidural anaesthesia/analgesia are included in this analysis, certainly gives the impression that epidural anaesthesia/analgesia might be beneficial in terms of mortality.

Meta-analysis is a useful tool for combining data from similar studies. However, there are limitations to the technique, and meta-analysis cannot reasonably be used when there are differences between studies that have a significant impact on the outcomes being assessed. If excessive patient or treatment heterogeneity is ignored, meta-analysis tends to oversimplify complex issues, and its conclusions are misleading. To some extent, all meta-analyses of epidural anaesthesia/analgesia suffer from excessive treatment heterogeneity because of important differences in epidural level and medications used that are not always clearly defined in the contributing studies. The Rodgers meta-analysis suffers particularly because, at least in its overall analysis, it combines truly non-homogeneous interventions such as single shot spinals without general anaesthesia, and epidural anaesthesia/analgesia with general anaesthesia.

Another important reason for discrepancies between meta-analyses and subsequent large multi-centre trials is thought to be the inclusion of older, outdated trials in meta-analyses.1012 Recent changes in anaesthetic and surgical practice have radically altered perioperative morbidity, and made older studies less relevant to today’s practice. Short-acting drugs, the LMA,{dagger} new standards of monitoring and vigilance, better optimization of preoperative medical status, less invasive surgical techniques, accelerated recovery procedures, and modern thromboprophylaxis, have all contributed.13 14 The large trials of epidural anaesthesia/analgesia conducted recently in the USA4 and Australia,5 6 may have failed to show differences simply because they sought only differences in major (life-threatening) morbidity and mortality. If these events are rare enough, they may not be revealed by trials, even large trials, which do not have enough power to detect differences in rare events.15 16 Although the authors of the VA and MASTER trials used power analysis to estimate the size of study they would need to identify differences, recently instituted perioperative treatment procedures used in advanced medical centres have markedly reduced surgical morbidity and mortality,14 17 and may have obscured differences attributable to epidural anaesthesia/analgesia alone in these settings. The meta-analyses, on the other hand, show differences because they incorporate older studies, conducted before improvements in morbidity and mortality, when the influence of neuraxial block on thromboembolism, in particular, was important. Careful reading of the Rodgers meta-analysis,9 using the authors’ own well-constructed subgroup analyses, makes it clear that benefit pertains mainly to neuraxial anaesthesia used for fractured hip and vascular surgery, especially in older studies, where thromboembolism was a high risk, and thromboprophylaxis was not used. Just as the favourable effect of neuraxial anaesthesia on thromboembolism before the routine use of thromboprophylaxis may have skewed the overall result of the Rodgers meta-analysis,9 this specific benefit may no longer be important. However, this and other benefits of neuraxial anaesthesia may still be worthwhile where modern technology and drugs are not available.

There are, of course, many other reasons that the results of meta-analyses may differ from those of subsequent large trials. Both meta-analyses and large trials suffer from some loss of control over inclusion and exclusion criteria. Many of the weaknesses of small trials are compounded in meta-analysis. For example, meta-analyses of small trials may overestimate treatment effects because positive trials are published preferentially, and only large differences can be detected (publication bias).18 A weakness of large trials, on the other hand, is that because they must have procedures that are compatible with a broad range of practice, the investigators have less control over factors that influence outcome, including the skill and experience of the anaesthetists, and many other aspects of perioperative care.5 11 12 15 19 Questions remain about the influence of epidural level, dosing regimens, and the measuring of dynamic pain vs pain at rest, that may be important in terms of explaining negative results.46 20 21

If we believe that epidural anaesthesia/analgesia does not influence catastrophic outcome in present-day practice, are there any benefits at all? First, it is important to remember that, for many surgical procedures, particularly upper abdominal and thoracic, trials of epidural analgesia consistently demonstrate superior analgesia.2 22 In many geographical areas, patients consider epidurals standard analgesic care after major surgery, and patient satisfaction has become an indicator of quality of medical care.23 Second, several benefits of epidural anaesthesia/analgesia in terms of reducing non-life-threatening morbidity have been demonstrated in selected patients. These effects are not always directly related to the analgesic efficacy of epidurals, but nevertheless play a key role in integrated approaches to improving surgical outcome.24 25 They include reduced blood loss, improved bowel mobility, fewer cardiac ischaemic events, improved cough and reduced atelectasis, improved short- and long-term activity levels, and improved quality of life. 22 2427 The effects of continuous postoperative local anaesthetics on the bowel may be particularly effective in terms of reducing ileus and hospital stay.2 24 25

The current literature suggests that epidural anaesthesia/analgesia does have benefits that are important in terms of patient satisfaction, and important in integrated perioperative care. Yet it may no longer be possible, in advanced medical centres, to show the benefit of any anaesthetic or analgesic option in isolation in terms of its effect on major morbidity and mortality, given the safety of modern anaesthesia, the rarity of a catastrophic outcome, and the role of integrated approaches to improving surgical outcome. We must continue to question the role of epidurals in perioperative care, not forgetting that there are also well-established risks, some serious, that are rarely considered in trials of surgical outcome, but which remain important every time we offer an epidural to a surgical patient.19 28 29

References

1 Bromage PR, Camporesi E, Leslie J. Epidural narcotics in volunteers: sensitivity to pain and to carbon dioxide. Pain 1980; 9: 145–60[CrossRef][ISI][Medline]

2 Jorgensen H, Wetterslev J, Moiniche S, Dahl JB. Epidural local anaesthetics versus opioid-based analgesic regimens for postoperative gastrointestinal paralysis, PONV and pain after abdominal surgery. Cochrane Database Systemat Rev 2003; 1

3 Yeager MP, Glass DD, Neff RK, Brinck-Johnsen T. Epidural anesthesia and analgesia in high-risk surgical patients. Anesthesiology 1987; 66: 729–36[ISI][Medline]

4 Park WY, Thompson JS, Lee K. Effect of epidural anesthesia and analgesia on perioperative outcome. A randomized, controlled Veterans Affairs Cooperative Study. Ann Surg 2001; 234: 560–71[CrossRef][ISI][Medline]

5 Rigg JRA, Jamrozik K, Myles PS, et al. Epidural anaesthesia and analgesia and outcome of major surgery: a randomised trial. Lancet 2002; 359: 1276–82[CrossRef][ISI][Medline]

6 Peyton PJ, Myles PS, Silbert BS, et al. Perioperative epidural analgesia and outcome after major abdominal surgery in high-risk patients. Anesth Analg 2003; 96: 548–54[Abstract/Free Full Text]

7 Ballantyne JC, Carr DB, DeFerranti S, et al. The comparative effects of postoperative analgesic therapies on pulmonary outcome: cumulative meta-analyses of randomized, controlled trials. Anesth Analg 1998; 86: 598–612[Abstract]

8 Beattie W, Badner N, Choi P. Epidural analgesia reduced postoperative myocardial infarction: a meta-analysis. Anesth Analg 2001; 93: 853–8[Abstract/Free Full Text]

9 Rodgers A, Walker WS, McKee A, et al. Reduction of postoperative mortality and morbidity with epidural or spinal anaesthesia: results from overview of randomised trials. Br Med J 2000; 321: 1493–7[Abstract/Free Full Text]

10 Cappelleri JC, Ioannidis JPA, de Ferranti SD, et al. Large trials versus meta-analyses of smaller trials: how do their results compare? JAMA 1996; 276: 1332–8[Abstract]

11 LeLorier J, Gregoire G, Benhaddad A, Lapierre J, Derderian F. Discrepancies between meta-analyses and subsequent large randomised controlled trials. N Engl J Med 1997; 337: 536–42[Abstract/Free Full Text]

12 Myles P. Why we need large randomized studies in anaesthesia. Br J Anaesth 1999; 83: 833–4[Free Full Text]

13 Sentinel events: approaches to error reduction and prevention. Jt Comm J Qual Improv 1998; 24: 175–86[Medline]

14 Callum KG, Gray AJG, Hoile RW, et al. The 2000 report of the National Confidential Enquiry into Perioperative Deaths: data collection period 1 April 1998 to 31 March 1999. London: National Confidential Enquiry into Perioperative Deaths, November 2000

15 McPeek B. Inference, generalizability and a major change in anesthetic practice. Anesthesiology 1987; 66: 723–4[ISI][Medline]

16 Benson K, Hartz AJ. A comparison of observational studies and randomized, controlled trials. New Engl J Med 2000; 342: 1878–86[Abstract/Free Full Text]

17 Birkmeyer JD, Siewers AE, Finlayson EVQ, et al. Hospital volume and surgical mortality in the United States. N Engl J Med 2002; 346: 1128–37[Abstract/Free Full Text]

18 Pogue J, Yusuf S. Overcoming the limitations of current meta-analysis of randomised controlled trials. Lancet 1998; 351: 47–52[CrossRef][ISI][Medline]

19 Wildsmith JAW. No sceptic me, but the long day’s task is not yet done: the 2002 Gaston Labat lecture. Reg Anesth Pain Med 2002; 27: 503–8[CrossRef][ISI][Medline]

20 Norris EJ, Beattie C, Perler BA, et al. Double-masked randomized trial comparing alternate combinations of intraoperative anesthesia and postoperative analgesia in abdominal aortic surgery. Anesthesiology 2001; 95: 1054–67[CrossRef][ISI][Medline]

21 Andrease M. Underdosing the epidural invalidates a good clinical trial. Anesthesiology 2002; 97: 1026–7

22 Liu S, Carpenter R, Neal JM. Epidural anesthesia and analgesia. Their role in postoperative outcome. Anesthesiology 1995; 82: 1474–506[ISI][Medline]

23 Wu C, Maqibuddin M, Fleisher LA. Measurement of patient satisfaction as an outcome of regional anesthesia and analgesia: a systematic review. Reg Anesth Pain Med 2001; 26: 196–208[CrossRef][ISI][Medline]

24 Kehlet H. A multi-modal approach to control postoperative pathophysiology and rehabilitation. Br J Anaesth 1997; 78: 606–17[Abstract/Free Full Text]

25 Basse L, Raskov HH, Jakobsen D, et al. Accelerated postoperative recovery program after colonic resection improves physical performance, pulmonary function and body composition. Br J Surg 2002; 89: 446–53[CrossRef][ISI][Medline]

26 Atanassoff PG. Effects of regional anesthesia on perioperative outcome. J Clin Anesth 1996; 8: 446–55[CrossRef][ISI][Medline]

27 Carli F, Mayo N, Klubien K, Schricker T, Trudel J, Belliveau P. Epidural analgesia enhances functional exercise capacity and health-related quality of life after colonic surgery. Anesthesiology 2002; 97: 540–9[ISI][Medline]

28 Bergqvist D, Wu CL, Neal JM. Anticoagulation and neuraxial regional anesthesia: perspectives. Reg Anesth Pain Med 2003; 28: 163–6[CrossRef][ISI][Medline]

29 Horlocker TT, Wedel DJ, Benzon H, et al. Regional anesthesia in the anticoagulated patient: defining the risks (The second ASRA consensus conference on neuraxial anesthesia and anticoagulation). Reg Anesth Pain Med 2003; 28: 172–97[CrossRef][ISI][Medline]