Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Which anaesthetic technique - revisited


Bandolier 86 was unhappy with a meta-analysis [1] examining the use of neuraxial blockade on mortality after surgery. Although in many respects an excellent review, the conclusion that neuraxial blockade reduced mortality depended on five (out of 141) trials. These trials had death rates of over 10%, had fewer than 100 patients, and had only 6% of the total number of patients.

They were not patients like ours. In the other 131 trials the death rate was about 2% and there was no difference between neuraxial blockade and control. The authors of the review were themselves cautious about the result. A new, large, randomised trial [2] designed specifically to test the hypothesis shows just how sensible they were.


The trial, conducted in Australia, East Asia and the Middle East, randomised highest-risk patients undergoing high-risk major abdominal procedures to intraoperative epidural anaesthesia or general anaesthesia with balanced technique with intraoperative and postoperative opioids. Randomisation was from a central office, but the study was not, nor could be, blind.

The intention was to select a patient population where about half the patients were expected to have a major postoperative complication within 30 days of operation. The outcome was a combined endpoint of death or at least one morbid endpoint (renal failure, cardiovascular event etc).


Over six years, 888 patients were randomised. There were no differences in the proportion of risk factors: 45% had diabetes, 27% myocardial ischaemia, 15% acute myocardial infarction and 12% cardiac failure, as the most common risk factors.

The combined endpoint occurred in 60% of patients. There were no differences between the two procedures for the combined endpoint, death, and all specified endpoints with the exception of the need for prolonged ventilation or re-intubation, which occurred less frequently with epidurals. There was only one intraoperative death, and most deaths occurred at least four days after surgery (Figure 1).

Figure 1: Postoperative mortality and anaesthetic method


This randomised trial confirms the conservative view of the meta-analysis, that anaesthetic technique makes no difference to operative mortality. Another recent large (1,021 patients) randomised trial [3] in a different patient group also found no difference.

The 131 trials in the meta-analysis with death rates of below 10% in controls had 173 deaths with a relative risk of 0.8 (95% 0.6 to 1.1). Adding in the 42 deaths in this study changes the relative risk to 0.9 (0.7 to 1.2). The lesson is that meta-analyses will be wrong if they include trials that are invalid, either because they may be biased, or because their patients are not like ours.

That's the main reason why meta-analysis and randomised trials sometimes give different results. It's not that either method is wrong, just that people use them wrongly.


  1. A Rodgers et al. Reduction in postoperative mortality and morbidity with epidural or spinal anaesthesia: results from overview of randomised trials. BMJ 2000 321: 1-12.
  2. JR Rigg et al. Epidural anaesthesia and analgesia and outcome of major surgery: a randomised trial. Lancet 2002 359: 1276-1282.
  3. WY Park et al. Effect of epidural anesthesia and analgesia on perioperative outcome: a randomized, controlled Veteran Affairs cooperative study. Annals of Surgery 2001 234: 560-569.
previous or next story in this issue