Skip navigation
Acute Pain | Chronic Pain | General

Patient controlled epidural analgesia for labour

Clinical bottom line

Patient controlled epidural analgesia were less likely to need unscheduled anaesthetic interventions, receive a slightly lower dose of local anaesthetic, and have less motor block than those who have continuous epidural infusion.


M van der Vyver et al. Patient-controlled epidural analgesia versus continuous infusion for labour analgesia: a meta-analysis. British Journal of Anaesthesia 2002 89: 459-465.

Systematic review

Randomised trials comparing patient controlled epidural analgesia with continuous epidural infusion were sought in a variety of electronic databases, including the Cochrane Library. The same local anaesthetic had to be used. A variety of outcome measures were sought.


There were nine randomised trials with 641 patients, most of whom received bupivacaine. Five studies had quality scores of 3/5 or more.

In five studies, the number of women who needed a clinician top up of the epidural for analgesia was reported (Figure 1). In women with continuous epidural infusion, 41% did not need a top up. In women with patient controlled epidural analgesia only 69% did not need a top up. The relative risk was 1.7 (1.4 to 2.0) and the number of women needed to be treated with patient controlled epidural analgesia rather than continuous epidural infusion to prevent one needing an unscheduled top up was 3.6 (2.8 to 5.2).

Figure 1: Women not requiring unscheduled epidural top up

There was some evidence of decreased motor weakness in four trials with 300 patients. Here no motor weakness was seen in 90% of women receiving patient controlled epidural analgesia compared with 80% in those receiving continuous epidural infusion. The NNT for one women to be free of motor weakness was 9 (5.3 to 43).

Evidence on pain scores and maternal satisfaction was not available.


The interesting issue here is again the paucity of clinical data. On the face of it, patient controlled epidural analgesia has benefits, though much more time would need to be given to patient education. The evidence we have addresses a research an agenda, and is inadequate for implementation.