Objective: To examine the effectiveness of paroxetine and cognitive-behavioural therapy (CBT) in elderly patients suffering from panic disorder with or without agoraphobia (PD(A)). Method: Forty-nine patients aged 60+ years with confirmed PD(A) were randomly assigned to 40 mg paroxetine, individual CBT, or to a 14-week waiting list. Outcomes, with avoidance behaviour and agoraphobic cognitions being the primary measures, were assessed at baseline and at weeks 8, 14 (conclusion CBT/waiting list), and at week 26 (treated patients only) and analysed using mixed models. Results: All outcome measures showed that the patients having received CBT and those treated with paroxetine had significantly better improvement compared with those in the waiting-list condition. With one patient (1/20, 5%) in the CBT and three (3/14, 17.6%) in the paroxetine condition dropping out, attrition rates were low. Conclusion: Patients with late-life panic disorder respond well to both paroxetine and CBT. Although promising, the outcomes warrant replication in larger study groups.