Sixteen of nineteen patients completed a randomised double-blind crossover trial to assess the effect of oral mexiletine (10 mg/kg bodyweight daily) on the symptoms and signs of chronic painful diabetic neuropathy. The median age of the sixteen patients was 50 years (range 30-64). Assessment with a five-item clinical symptom scale showed significant improvement during the mexiletine phase compared with the placebo phase. Pain was reduced during mexiletine but not during placebo, as assessed by a visual analogue rating scale. Mexiletine treatment had no effect on tendon reflexes, vibration threshold levels, beat-to-beat variation in heart rate during deep breathing, and postural blood pressure response. Mild side-effects were seen in three of the sixteen patients during mexiletine treatment.