It is widely believed that comorbid personality disorders (PDs) have a negative influence on the effects of treatment for axis-1 anxiety disorders. However, many studies reporting negative influences suffer from methodological problems, such as inference of PD after treatment by clinical judgement. The aim of our studies was to investigate the influence of comorbid PDs on outcome of CBT for anxiety disorders in a double blind prospective design.

Axis-1 and axis-2 disorders were assessed with SCID interviews. Therapists and patients were blind for outcome of SCID-II interview. Patients received state of the art CBT for their main anxiety disorder. We controlled for baseline levels. Outcome was assessed with Fear Questionnaire and SCL-90.

In a very large sample of more than 1800 patients we didn't find evidence for a negative influence of comorbid PDs. The only effect we found was that comorbid borderline PD was associated with drop-out. In a smaller sample (N = 398) we found evidence that PDs predicted higher avoidance levels after treatment, but no other psychopathology. Interestingly, specific beliefs related to PDs, notably mistrust and dependency beliefs, were related to higher symptom levels after treatment.

The influence of PDs on CBT of anxiety disorders is not strong. When effects were found, they were very small. Two cognitive beliefs seem to be central in interference of PDs with CBT: mistrust and dependency related beliefs.