One hundred RCTs were included in the review (the number of patients was not provided).
All surgical sites.
Epidural analgesia was associated with significantly better post- operative analgesia than parenteral opioids (WMD 10, 95% CI: 9.5, 10.5, P<0.001), based on all studies and observation points. Analysis by post-operative day also found that epidural analgesia was significantly better than parenteral opioids for each time point after surgery (P<0.01). There was no significant difference between intravenous patient-controlled analgesia and intramuscular or subcutaneous opioid administration (data not provided).
Thoracic surgery (TEA 18 RCTs; LEA 9 RCTs).
TEA was associated with significantly better post-operative pain control compared with parenteral opioids when using local anaesthetic with or without opioid (WMD 11.2, 95% CI: 9.9, 12.5, P<0.001), or opioids alone (WMD 2.9, 95% CI: 0.4, 5.3, P=0.002).
LEA was associated with significantly better post-operative pain control compared with parenteral opioids when using opioids alone (WMD 4.2, 95% CI: 2.2, 6.2, P<0.001).
When categorised by method of pain assessment, TEA and LEA were generally associated with significantly better post-operative rest and incident pain control than parenteral opioids when using local anaesthetic with or without opioid (P<0.001), or opioids alone (P<0.001). The exception was TEA with opioids alone for rest pain (WMD 0.6, 95% CI: -3, 1.5, P=0.12).
Abdominal surgery (TEA 21 RCTs; LEA 8 RCTs).
TEA was associated with significantly better post-operative pain control compared with parenteral opioids when using local anaesthetic with or without opioid (WMD 10.9, 95% CI: 10.1, 11.6, P<0.001), or opioids alone (WMD 6.7, 95% CI: 3.4, 9.6, P<0.001).
LEA was associated with significantly better post-operative pain control compared with parenteral opioids when using local anaesthetic with or without opioid (WMD 17.8, 95% CI: 15.8, 19.9, P<0.001), or opioids alone (WMD 8.5, 95% CI: 6.2, 10.8, P<0.001).
When categorised by method of pain assessment, TEA and LEA were associated with significantly better post-operative rest and incident pain control compared with parenteral opioids when using local anaesthetic with or without opioid (P<0.001), or opioids alone (P<0.001).
Pelvic and Caesarean delivery (TEA 2 RCTs; LEA 23 RCTs).
TEA was associated with significantly better post-operative pain control compared with parenteral opioids when using local anaesthetic with or without opioid (WMD 10.5, 95% CI: 7.1, 14.0, P<0.001).
LEA was associated with significantly better post-operative pain control compared with parenteral opioids when using local anaesthetic with or without opioid (WMD 7.2, 95% CI: 5.8, 8.6, P<0.001), or opioids alone (WMD 8.6, 95% CI: 7.2, 10.1, P<0.001).
When categorised by method of pain assessment, TEA and LEA were associated with significantly better post-operative rest and incident pain control compared with parenteral opioids when using local anaesthetic with or without opioid (P<0.001), or opioids alone (P<0.001).
Lower extremity surgery (LEA 13 RCTs).
LEA was associated with significantly better post-operative pain control compared with parenteral opioids when using local anaesthetic with or without opioid (WMD 12.6, 95% CI: 10.1, 15.0, P<0.001), or opioids alone (WMD 9.4, 95% CI: 6.8, 11.9, P<0.001).
When categorised by method of pain assessment, LEA was associated with significantly better post-operative rest and incident pain control compared with parenteral opioids when using local anaesthetic with or without opioid (P<0.01), or opioids alone (P<0.01).
Complications. The rates for all complications were relatively low across the included studies. The incidence rate of nausea or vomiting was significantly lower with local anaesthetic LEA than with parenteral opioids. No difference was found for TEA with local anaesthetic with or without opioid, or LEA with opioid alone.
Epidural analgesia was associated with a significantly higher incidence rate of pruritus compared with parenteral analgesia. LEA with opioid alone was associated with a 6-fold increased incidence from parenteral opioid (38% versus 6%, P<0.001).
Epidural analgesia was associated with higher incidence of motor block or numbness compared with epidural. The rate of numbness was higher in those given LEA with opioid alone.
TEA with local anaesthetic, with or without opioid, was associated with an increased rate of hypotension compared with parenteral analgesia. However, the overall incidence of hypotension was low.
The authors stated that there were insufficient data to evaluate return of function, major complications, and several other minor complications including confusion-delirium, sedation, constipation, urinary retention, headache and backache.
Publication bias.
It was estimated that additional trials with at least 94,273 participants demonstrating no statistical difference between treatment regimens would be required to invalidate the results of this review. The funnel plot did not suggest the presence of publication bias.