Four relevant RCTs were identified (n=129, range 30 to 36); all had a modified Jadad score of 3. Withdrawals ranged from 0% to 14%.
Postoperative pain scores were significantly lower in the epidural group than in the intravenous morphine group after 24 hours (WMD -12.59, 95% CI -24.40 to -0.78, I2=0%; four studies), after 48 hours (WMD -10.13, 95% CI -19.11 to -1.16, I2=0%; four studies) and after 72 hours (WMD -11.53, 95% CI -20.85 to -2.21, I2=0%; three studies). Subgroup analysis for the two RCTs that used one epidural catheter after 24 hours found no significant benefit for epidural analgesia, but they also used lower anaesthetic doses. Subgroup analysis for the two RCTs that used two epidural catheters after 24 hours and higher anaesthetic doses (both studies originated from the same centre and author) found a significant benefit for epidural anaesthesia (WMD -15.02, 95% CI -28.34 to -1.70, I2=0%).
Nausea rates (three studies) were significantly lower in epidural groups (0% versus 38%) than intravenous morphine groups (0% versus 20%) in two studies after 24 hours; there was no significant difference between groups in a third study.
Pruritus rates (three studies) were significantly lower in the epidural groups (0% versus 38%) than the intravenous morphine groups (7% versus 33%) in two studies after 24 hours; there was no significant difference between groups in a third study.
Rescue analgesic use (four studies) was significantly lower in epidural groups than intravenous morphine groups in two studies (WMD -7.5mg morphine per patient, 95% CI -11.26 to -3.69); there was no significant difference between groups in the two other studies.
For return of bowel function (four studies), two studies found no significant differences between groups for resumption of liquid intake (one RCT), time to liquid or solid intake, return of bowel sounds and nasogastric tube removal (one RCT). Two studies found the epidural group had more rapid return to bowel sounds, first flatus, first tolerated meal and first bowel movement.
Two studies assessed patient satisfaction using a zero to 10 scale and found a higher satisfaction rate in the epidural group than the intravenous morphine group (WMD 1.62, 95% CI 1.26 to 1.97).