In order to evaluate the sensitivity of an adrenaline test dose for detecting intravascular injection and the effect of atropine pretreatment, 90 ASA physical status I and II patients were allocated randomly to two groups, to receive i.v. saline 1 ml (n = 46) or i.v. atropine 0.5 mg (n = 44). Five minutes later, all patients received an i.v. test dose of 2% lignocaine 3 ml with adrenaline 15 micrograms at a rate of 1 ml s-1. The groups were similar with respect to basal heart rate (HR). HR remained unchanged after saline injection, but increased slightly 5 min after atropine injection (mean 78 (SD 15) beat min-1 vs 87 (20) beat min-1 (P less than 0.05). After the test dose of lignocaine with adrenaline, HR increased significantly in both groups at 30 s, 1 and 2 min, and remained increased at 3 min in the atropine group. The maximum increase in HR was greater in the atropine group than in the saline group (31 (4) beat min-1 vs 26 (11) beat min-1 (P less than 0.05). However, when individual maximum HR changes are considered, five patients in the saline group and four in the atropine group had an increase less than or equal to 10 beat min-1, and three patients in the saline group had no change or a decrease in HR. Defining a positive result to a test dose as an increase in HR of greater than 10 beat min-1, the sensitivity of the adrenaline test dose was 83 (5.5)% in the saline group and 91 (3.5)% in the atropine group (ns).(ABSTRACT TRUNCATED AT 250 WORDS)