Current pharmacotherapy for panic disorder generally consists of a trial of antidepressants or benzodiazepines for at least 3 months, with continuation for 6 to 12 months when patients respond favorably. However, the most appropriate time to discontinue medication has not yet been determined. The author discusses reasons for discontinuation, both appropriate and inappropriate; optimal timing for discontinuation; outcomes of discontinuation (i.e., remaining well, relapse, rebound, and withdrawal); strategies, including dosage taper, to avoid or minimize unfavorable outcomes; and long-term outcomes in patients treated for panic disorder. The author concludes from the data available on long-term outcomes that, although there is preliminary understanding of the course of acutely treated panic disorder patients, additional data from long-term follow-up studies are needed.