We have evaluated prilocaine as a topical anaesthetic agent for fibreoptic bronchoscopy in comparison with lidocaine in terms of efficacy and safety. Forty patients were included in a randomised double-blind parallel-group study. Efficacy was assessed using visual analogue scales, a patient ranking scale and the number of doses of local anaesthetic and intravenous sedative required. Measures of toxicity included peak plasma concentration of local anaesthetic, whether supplementary oxygen was needed and change in methaemoglobin concentration. For most of the outcome variables, the medians and quartiles were similar for the two local anaesthetics. However, the median peak plasma concentration of prilocaine (0.5 micrograms.ml-1) was less than one-third that of lidocaine (1.76 micrograms.ml-1). The merits and hazards of using multiple-regression modelling to improve the precision of the analysis of the results are considered. We conclude that prilocaine can be used successfully as a topical anaesthetic agent for fibreoptic bronchoscopy and is associated with a lower risk of toxicity.