Our hypothesis was that deep anaesthesia, as estimated by a low target bispectral index (BIS) of 30-40, would result in less postoperative pain than that achieved at a conventional depth of anaesthesia. We undertook a randomized double-blind controlled study at two tertiary teaching hospitals in New Zealand (2010-1) recruiting 135 adult patients ASA I-II presenting for non-emergent surgery under general anaesthesia requiring tracheal intubation. Anaesthesia was maintained with desflurane and a multimodal analgesia regimen comprising fentanyl infusion, i.v. paracetamol, and parecoxib. Patients were randomly assigned to either a low BIS (30-40) group or a high BIS (45-60) group. Desflurane concentrations were titrated to achieve these targets. Postoperative pain was assessed by: the pain on awakening (0-10, verbal rating scale, VRS(awake)) in the post-anaesthetic care unit; pain on activity at 20-24 h after operation (VRS(d1A)); and the rate of morphine patient-controlled analgesia (PCA) usage over the first 24 h. There was no statistically significant difference between the two groups for any of the pain scores. The median [inter-quartile range (IQR)] VRS(awake) was 4.0 (0-8) for the low and 4.0 (0-8) for the high BIS groups (P=0.56). The median (IQR) VRS(d1A) was 3.0 (1-5) for the low and 3.0 (1.5-4.5) for the high BIS groups (P=0.83). The median PCA morphine consumption in the low BIS group was 0.61 mg h(-1) (0.04-1.5) vs 0.43 mg h(-1) (0-1.59) in the high BIS group (P=0.98). We conclude that there is no clinically useful analgesic effect of a deep anaesthesia regimen.