Skip navigation
Acute Pain | Chronic Pain | General

Patient-controlled analgesia update

Clinical bottom line

Patient-controlled analgesia with opioid produces modest improvement in pain relief compared to the same opioid given conventionally. Patients preferred it, and there were no more or fewer adverse events reported.


B Walder et al. Efficacy and safety of patient-controlled opioid analgesia for acute postoperative pain. Acta Anaesthesiologica Scandinavica 2001 45: 795-804.

Systematic review

This review was an update of a previous review of patient controlled analgesia. Several electronic databases, including the Cochrane Library, were searched. Studies had to be randomised comparisons of an opioid given by intravenous patient-controlled device and the same opioid given subcutaneously, intravenously or intramuscularly. Comparisons of different opioids were excluded.


Thirty-two randomised trials with some 2,100 patients were included. Most (27) had quality scores of 2/5 or less. The median number of patients per trial was 44, and trials varied in drug, bolus dose, and lockout times of the devices. Morphine was the most used opioid, in 22 trials.

The review examined a wide variety of endpoints. Morphine consumption was not significantly different between patient-controlled devices and conventional administration. Visual analogue scales for pain at rest at various times from about 12 to 120 hours after surgery was reported, the tendency was for lower pain intensity scores for patient-controlled devices.

For dichotomous efficacy information, patients using patient-controlled devices had more satisfaction (82%) than those having conventional analgesia (68%) , with a number needed to treat for one more patient to be satisfied of 8 (5 to 15).

Adverse events (nausea, vomiting, bradypnoea, hypoxia, sedation, pruritus) occurred no more frequently with either method of opioid delivery.


Here we have an extremely good review of, on the whole, studies lacking in quality. Methodological quality was low, with most trials possibly showing some bias. Size was also a problem, with a median trial size of 44 patients. So even the few conclusions that have been drawn can only be treated with a degree of caution.

The best conclusion si that patient-controlled analgesia probably produces analgesia as good as, or a little better than, that provided by conventional opioid delivery. The question is whether conventional opioid delivery in the real world is always as good as it is in these trials. There is a huge research agenda here.