Background: The value of internal (interoceptive) cues for exposure is under debate and so was tested in a pilot controlled study. Methods: Outpatients with panic disorder and severe agoraphobia were randomised to 10 weeks of self-exposure to either (1) both internal (interoceptive) and external cues (n = 12) or (2) external cues only (n = 14). Both groups were trained in slow deep breathing and asked to carry out daily self-exposure homework. Neither group had cognitive restructuring. Results: By post-treatment and follow-up all outcome measures improved significantly in both treatment groups. The two groups did not differ significantly in outcome, though slightly more patients who had exposure to both internal and external cues improved 50% or more on phobic avoidance and fear. Conclusions: A larger controlled study is now worthwhile to tell if such small differences can be significant.