It is inappropriate to subject children to distressing procedures if this can be avoided. Moreover, many procedures may be difficult or unsafe in the uncooperative child. The number of valuable but invasive techniques in use has increased steadily over the years, and to facilitate these, sedative agents are often given. Such use of sedation is generally supervised by non-anaesthetists. Despite the obvious benefits associated with this practice, there are unresolved concerns about its efficacy and safety. The ideal sedative regimen would act predictably and rapidly and would induce a level and duration of sedation appropriate to the procedure being performed. In practice, few regimens are truly satisfactory in these terms. General anaesthesia may therefore be preferred, but for many procedures (for example, liver biopsy) this might seem inappropriate. Widespread use of general anaesthesia for these purposes would also have substantial resource implications for paediatric departments. The use of sedation in our hospital is probably comparable with that in many similar institutions.1-3 Many children undergo cardiac catheterisation and other radiological procedures under sedation. Protocols vary, but often include high dose chloral hydrate, or a combination of temazepam and droperidol, and in many cases it is necessary to induce quite deep levels of sedation. As an indication of the volume and importance of this activity, magnetic resonance imaging alone accounts for more than 1100 procedures in our radiology department each year. We also perform about 800 gastrointestinal endoscopic procedures annually in infants and children, and for most of these we employ intravenous sedation using a combination of midazolam and pethidine. Gastrointestinal endoscopy became part of paediatric practice about 20 years ago, and from the beginning this invasive technology posed great challenges with regard to sedation.3 In this article I shall discuss the difficulties and controversies surrounding endoscopic sedation in order …