Obstructive urinary symptoms due to chronic non-bacterial prostatitis are due to incomplete relaxation of the bladder neck or inappropriate contraction of the external urethral sphincter during voiding. Patients may respond to adrenergic blocking drugs; if effective, treatment may need to be continued indefinitely. Botulinum toxin has been used to weaken striated and smooth muscles. Toxin injections into the external urethral sphincter to relieve urethral obstruction were described by Dykstra and Sidi. Apart from treatment of detrusor-sphincter dyssynergia, we have found no reports in the literature on the treatment of voiding dysfunction in non-bacterial prostatitis. Four consecutive men (mean age 30·7 [SD 5·7] years) with chronic non-bacterial prostatitis and poor bladder emptying because of spastic external urethral sphincter (mean duration of symptoms 18 [3] months), who failed to respond to tamsulosin 0·4 mg once daily for more than 4 months were enrolled. All the patients were examined, had uroflowmetric studies to assess times of urinary flow (TQ) and maximum urinary flow (TQmax), maximum flow (Qmax), average flow (Qave), and total urinary volume (Vcomp), and had anorectal manometry at rest (RT) and after maximum contraction (MC). An increased value of TQ and TQmax with a normal value of Qmax was taken to be indicative of incomplete relaxation of bladder neck. 1, 4, and 8 weeks after treatment, patients underwent the same assessments. With the patient lying on his left side, a 26-gauge monopolar needle electrode was inserted in the perineum in the anterior midline, about 1·5–2·0 cm from the anus and directed toward the prostatic apex, without sedation or local anaesthesia. 30 U of type A botulinum toxin were injected. No local complications or systemic side-effects were seen. Within 1 week of injection all patients had a striking improvement in their voiding; none complained of urinary incontinence. At 4 weeks, three patients showed a continuing improvement. At 8 weeks, the same three patients were satisfied with the therapy and none of them complained of urinary incontinence. The patients were followed up for a mean of 12 months. No relapse occurred in the three patients who improved. 6 weeks after treatment, the fourth patient reported worsening voiding. He was also depressed and required fluoxetine. He was re-treated with 50 U of botulinum toxin and the urinary symptoms improved. Uroflowmetric study showed a decrease in TQ and TQmax values at 1, 4, and 8 weeks compared with baseline values (table). Other variables were not changed.