Carbamazepine therapy in restless legs. Discrimination between responders and non-responders.

One hundred and seventy-four patients suffering from restless legs were included in a double-blind trial. Eighty-four patients were treated with carbamazepine (CBZ) and 90 with placebo. Discrimination analysis was carried out in order to characterize the patients who did not benefit from the treatment. Patients treated with CBZ were divided into responders and non-responders. A discriminant function classified 15 out of 19 actual non-responders as non-responders and 60 out of 65 actual responders as responders. By using the "leaving-one-out" technique, 14 of the non-responders and 57 of the responders were still correctly classified. The probability of erroneously classifying a patient increased from about 10 to 15% by this correction. The discriminant function classified approximately 80% of the patients in the placebo group as responders to CBZ.