Eight RCTs (232 patients) were included in the review, of which one (17 patients) had a cross-over design. Jadad scores ranged from 1 to 3 and four RCTs were considered to be high quality. Only one trial reported using an intention-to-treat analysis. One trial used blinded outcome evaluation. Four reported allocation concealment, with two reporting appropriate methods.
There were no statistically significant differences between treatment groups in any of the main analyses for the primary outcomes of mortality (four studies), intensive care unit stay (five studies), hospital stay (five studies) or duration of mechanical ventilation (three studies). Statistically significant heterogeneity (I2>90%) was detected for all analyses except mortality where I2=0%.
When the two studies that compared thoracic epidural analgesia with local anaesthetics versus parenteral opioids were pooled for the outcome of mechanical ventilation, the duration was found to be significantly lower in the epidural groups (weighted mean difference was -4.17 days, 95% CI: -5.45 to -2.88; p<0.01; 73 patients), with no evidence of statistical heterogeneity (I2=0%). Subgroup analyses of this comparison for the outcome of intensive care unit and hospital stay were not statistically significant, although evidence of heterogeneity was not found (I2=0% in both cases).
Meta-analyses for secondary outcomes were not possible due to missing data. A brief narrative synthesis was provided.