The local analgesic efficacy of a cream formulation of lidocaine and prilocaine (EMLA) in reducing pain at venous cannulation was investigated in children scheduled for elective surgery. Forty children participated in a double-blind, randomized comparison between EMLA and inactive placebo cream. Another group of 18 children without any local treatment was studied as an additional control material. Subjective pain scores, expressed with a visual analogue scale, were significantly lower in the EMLA group compared with both the group treated with placebo cream (P less than 0.001) and the open control group (no cream; P less than 0.01). Local pallor and slight oedema were the only side-effects, registered in both cream-treated groups. A preliminary study was also carried out with 10 children (five with EMLA and five without) in order to determine whether catecholamine and vasopressin levels in venous blood are affected by the stress and anxiety associated with venepuncture in children premedicated with oral flunitrazepam. No significant hormone responses were, however, detected. The lidocaine concentrations measured in venous blood taken from the application site of EMLA cream were low, and there were no measurable levels of lidocaine in simultaneous blood samples from the opposite extremity. In our opinion EMLA cream is safe and alleviates effectively the pain associated with venepuncture, and thus deserves a place in the routine premedication of children.