SIR: The patient was a 73-year-old woman who developed major depression triggered by domestic stress in June 2002. She was prescribed 40 mg of paroxetine and 0.8 mg of alprazolam at bedtime and 3 mg of etizolam during the day. Full remission resulted, and treatment was maintained.

The patient remained well until August 2002, when she was admitted to the urology department of a hospital as a kidney donor for her daughter. She received a last dosage of medication at bedtime on the day before surgery. At 07:00 the following day, diazepam (2 mg) and ranitidine hydrochloride (150 mg) were administered preoperatively. Left-side donor nephrectomy was performed following intravenous injection of thiopental (200 mg) at 10:00 and inhalation with sevoflurane for maintenance of general anesthesia. The operation was completed without complications at 14:00. By 14:30, the patient had completely regained consciousness. Post-operative laboratory data revealed no abnormalities. However, symptoms of thirst and dizziness were suddenly reported from 16:00. Thereafter, disorderly behavior gradually developed. At 21:20, a urologist administered etizolam (1 mg) and alprazolam (0.8 mg), but the patient remained agitated and complained of visual hallucinations such as insects. Since psychotic symptoms persisted despite intravenous injection of haloperidol (5 mg) at 22:25, diazepam (5 mg) at 23:40, and hydroxyzine hydrochloride (50 mg) at 01:57 the following day, the urologist referred the patient to a psychiatrist (the author) at 10:00 the next morning. Although visual hallucinations had disappeared, a diagnosis of delirium was made since the patient was assessed to be not fully oriented and was unable to follow verbal commands (Japan Coma Scale score = 1-3). Tiapride hydrochloride (75 mg) during the day and chlorpromazine (15 mg) at bedtime were prescribed and the patient slept well. Psychotic manifestations had resolved completely by 10:30 on postoperative day 2.

We attribute delirium in this case to the abrupt withdrawal of paroxetine, because: 1) the patient had been taking a relatively high dose of paroxetine for a prolonged period; 2) delirium appeared soon after cessation of paroxetine; 3) delirium developed despite absence of intraoperative complications; 4) delirium was preceded by thirst and dizziness, which are common symptoms of SSRI withdrawal;⁺¹ 5) no drugs known to interact with paroxetine⁺² were administered intraoperatively; and 6) full recovery from delirium occurred within a short period. Another possible candidate for withdrawal delirium in this case was benzodiazepine. However, as psychotic manifestations were exacerbated despite administration of benzodiazepine before surgery and in the period when delirium initially appeared, discontinuation of benzodiazepine seems unlikely to have caused the delirium.

Paroxetine is reportedly more likely to induce withdrawal symptoms than other SSRIs, due to the following pharmacokinetics^+3,+4,+5: (1) short half-life; (2) absence of active metabolites; and (3) non-linear kinetics. Physicians should pay close attention to delirium as well as other symptoms of SSRIs, particularly in cases in whom high-dose paroxetine has been maintained for long periods.