Background: It is often suggested in the literature that alexithymic patients are less responsive to psychotherapy than nonalexithymic patients. However, few empirical studies have examined this issue. Furthermore, it is unclear whether or not alexithymia itself may improve during psychotherapy. Methods: Fifty-five consecutive outpatients with panic disorder received short-term cognitive-behavioral group therapy (CBGT) and were followed up 6 months later. Nineteen patients (35%) were on concomitant antidepressant medication. Alexithymia was measured using the 20-item Toronto Alexithymia Scale (TAS-20). Both completers and intention-to-treat analyses were calculated, taking into consideration the potentially confounding effect of comorbid conditions. Results: Baseline alexithymia did not predict outcome of CBGT, neither at posttreatment nor at follow-up. The presence of comorbid axis I disorders predicted nonresponse at posttreatment but not at follow-up. TAS-20 total scores decreased over time, with the TAS-20 factors 1 (difficulty identifying feelings) and 2 (difficulty describing feelings) decreasing significantly, while factor 3 (externally oriented thinking) remained largely stable. Conclusions: These findings are encouraging for cognitive-behavioral therapists working with patients with alexithymia who suffer from panic disorder: CBGT outcome does not appear to be negatively affected by alexithymia, and some alexithymic characteristics may even be reduced following CBGT. Assessing alexithymia at treatment onset may be useful for individually tailoring therapeutic interventions.