Of 30 patients who completed a study of progabide (PGB) as an add-on to both phenytoin (PHT) and carbamazepine (CBZ), 11 volunteered for a double-blind withdrawal protocol in which the PHT and CBZ were to be withdrawn. All patients were receiving 24-32 mg/kg/day PGB in combination with PHT and CBZ. Each patient was randomly assigned to withdrawal of either CBZ or PHT in the first block, and then withdrawal from the other in the second block in an attempt to achieve PGB monotherapy. Seizure occurrence was monitored by sequential analysis, and if a significant increase over baseline seizure frequency occurred, the dose of PGB was increased to a maximum of 45 mg/kg/day. If seizure frequency remained above baseline, the drug being withdrawn was added back and an attempt made to withdraw the other. The study was terminated if these adjustments were not successful in decreasing seizure frequency to baseline. At the conclusion of the study, three patients were being treated with PGB and PHT, two with CBZ and PGB, and six with all three. This study demonstrated the applicability of sequential analysis to antiepileptic drug trials.