EDITOR: Toxic reactions to local anaesthetics usually occur as a result of accidental overdosage or inadvertent intravenous (i.v.) injections. We report such a case; the reasons for this occurrence and suggestions for prevention are discussed. A 53-yr-old, 60 kg, ASA I female patient was scheduled for elective total knee replacement. There was nothing of note in her medical history. After institution of monitoring and establishment of venous access, the patient was premedicated with midazolam 3 mg. The baseline heart rate was 78 beats min−1 and blood pressure was 140/80 mmHg. An epidural catheter was inserted at the L3-4 interspace. Lidocaine 2% 3 mL was injected through the epidural catheter as a test dose. After 5 min, as the patient did not feel any change in her legs, a bolus of bupivacaine 0.5% 12 mL was injected slowly. Heart rate was 74 beats min−1 and blood pressure was 110/65 mmHg. The level of the regional blockade was T10. General anaesthesia was then induced with fentanyl 0.05 mg, propofol 2 mg kg−1 and atracurium 0.5 mg kg−1 and maintained with sevoflurane 1% in a mixture of oxygen and air. The operation was uneventful and lasted 110 min. The patient was extubated at the end of the operation and remained for 15 min in the recovery room before she was transferred to her bed. Postoperative analgesia was planned with bolus injections of a mixture of bupivacaine 0.25% and morphine 2 mg in 6 mL solution, via the epidural catheter. Two hours later, the patient requested epidural analgesia and a nurse injected the solution through the catheter. Six hours later another injection of same dose of bupivacaine was given i.v. by mistake by the same nurse. The patient was in some distress and complained of ringing in her ears, palpitations and dizziness. She could co-operate for examination of muscle tone and strength, which were normal as were the deep tendon reflexes. There were no lateralizing findings. The patient was afebrile. The i.v. bupivacaine produced a tachycardia of 120 beats min−1 but in sinus rhythm. The tachycardia reduced slightly over the next 20 min and normal blood pressure was always sustained. Investigations performed were arterial blood gases, serum glucose and electrolyte concentrations: all were normal other than PaCO2 4.0 kPa, and serum glucose 9.6 mmol L−1. The patient was immediately transferred to the intensive care unit where, fortunately, she passed the night uneventfully and no treatment was required. She was discharged from the hospital on the seventh postoperative day. Our normal practice for total knee replacement surgery is to site an epidural catheter and administer intermittent boluses of local anaesthetic combined with opioid at intervals of 6-8 h postoperatively. This epidural catheter had been sited uneventfully and had been used to give one bolus dose of bupivacaine. The second dose was an explicit i.v. injection despite clear written orders that it was to be given through the epidural catheter. Complications arise from the use of drugs and equipment, and from human error [1]. Local anaesthetics and opioids are commonly used for epidural analgesia in the postoperative period. There are reports of accidental injections of drugs with consequences ranging from no clinical effect to irreversible damage [2,3]. In our patient, we believe that the low concentration and volume of the bupivacaine protected the patient from an untoward reaction. Eldor and Frankel [4] reported a case of inadvertent i.v. injection using bupivacaine 0.5% where tinnitus was an early sign; indeed, our patient's first complaint was of tinnitus. Dysrrhythmias may occur with i.v. injection of bupivacaine. The striking feature of accidental i.v. bupivacaine is the great difficulty in resuscitating the patient [5]. In convulsant dosages, local anaesthetics cause increases in blood pressure, heart rate and cardiac output by stimulating the autonomic control centres in the brainstem [6]; however, in our patient, the low concentration of the bupivacaine produced only a sinus tachycardia. Because the nurses take part in postoperative analgesia management rather than the anaesthesiologists, they need to be trained and regulated. This case history demonstrates an inherent safety advantage of the 'bolus top-up' system over the 'continuous infusion' analgesia method when patients are cared for by a busy nursing team. Our nurse was still with the patient when the adverse symptoms and signs developed. Therefore, the effects of the inadvertent i.v. administration were observed and the nurse was able to call the anaesthesiologist immediately. If an intended epidural infusion had been connected, incorrectly, to an i.v. line, the nurse would have been occupied elsewhere by the time that toxic effects began to occur and a dangerous delay in diagnosis and management could have ensued. S. Karaca E. Ö. Ünlüsoy Department of Anesthesiology; Faculty of Medicine Cerrahpasa; University of Istanbul; Istanbul, Turkey