A single-blind dose-escalation study with the selective serotonin reuptake inhibitor paroxetine was conducted in 19 diabetic patients with neuropathy symptoms. The effect of treatment was evaluated by self-rating using visual analog scales. After an initial placebo period, paroxetine doses were increased from 10 mg/day in 10 mg steps, until the dose was 30-70 mg/day. In all except four patients, there was a marked relief of symptoms. Plasma concentrations of paroxetine above 300-400 nM were required to insure maximal relief in the majority of patients responding on paroxetine, but a considerable interindividual variation was observed (10-800 nM, median of 195 nM). The therapeutic effect appeared to increase gradually as the plasma concentration increased. The great interindividual variation in the pharmacokinetics of paroxetine was confirmed, but as the effect is maximal within approximately 1 week, and the drug is nontoxic, it may be clinically feasible simply to titrate the dose from 20 mg/day until the maximal effect is achieved. However, it is advised that titration to an effect, in diabetic neuropathy using doses above 50 mg/day, be undertaken with care as there is limited experience with doses above this level in any population. The beneficial effect of paroxetine appeared to be maintained unaltered during an additional 1 month open-label treatment on optimal paroxetine doses.