In a randomized, double-blind crossover study, 40 patients with postherpetic neuralgia were given single oral doses of clonidine, 0.2 mg, codeine, 120 mg, ibuprofen, 800 mg, or inert placebo. Pain relief and side effects were recorded for 6 hours. Patients reported significantly more relief after clonidine than after the other three treatments. Codeine and ibuprofen were ineffective. Sedation, dizziness, and other side effects were more frequent after clonidine (74%) or codeine (69%) than after placebo (36%) or ibuprofen (28%). Reported pain relief was greater during trials in which side effects were present. A single, mild side effect was associated with as much additional pain relief as multiple, severe side effects. Clonidine's superiority to codeine, which had a similar incidence of side effects, argues for a specific analgesic effect. In addition, side effects may have contributed to clonidine analgesia, perhaps by suggesting to patients that they had received a potent drug.