We conducted a meta-analysis to update the knowledge of long-term efficacy of different treatments in panic disorder with or without agoraphobia. Included were 68 studies pertaining to 106 treatment conditions and 1346 patients. Effect sizes Cohen's d were calculated within the treatment conditions at posttest and at follow-up for panic and agoraphobia. A comparison was made between six treatments: high-potency benzodiazepines, antidepressants, psychological panic management, exposure in vivo, antidepressants combined with exposure, and psychological panic management combined with exposure in vivo. The mean (± SD) duration of the follow-up period was 62 ± 89 weeks. In the majority of the studies (84%), follow-up had a naturalistic character. The lack of information about treatments received between posttest and follow-up limits the interpretation of the results. For all conditions, the treatment gains at posttest were maintained during the follow-up period. The mean (± SD) d for panic was 1.11 ± 0.70 at posttest and 1.28 ± 0.61 at follow-up; for agoraphobia, the mean d at posttest was 1.36 ± 1.10 and at follow-up it was 1.41 ± 0.82. Significant differences were found in efficacy on agoraphobic measures at follow-up between the combination of antidepressants and exposure in vivo versus psychological panic management, exposure in vivo, and the combination of psychological panic management and exposure. Overall, the data suggest that different treatment options for panic disorder with or without agoraphobia are effective at both posttest and follow-up. Research on long-term treatment, discontinuation of therapies, and interventions between posttest and follow-up need more attention, for pharmacotherapy as well as psychotherapy.