Background: Persistence is said to be a feature of personality disorder, but there are few long-term prospective studies of the condition. A total of 200 patients with anxiety and depressive disorders involved in a randomised controlled trial initiated in 1983 had full personality status assessed at baseline. We repeated assessment of personality status on three subsequent occasions over 30 years. Methods: Personality status was recorded using methods derived from the Personality Assessment Schedule, which has algorithms for allocating Diagnostic and Statistical Manual of Mental Disorders (DSM) and the 11th International Classification of Diseases (ICD-11) categories. The category and severity of personality diagnosis were recorded at baseline in the randomised patients with DSM-III anxiety and depressive diagnoses. The same methods of assessing personality status was repeated at 2, 12 and 30 years after baseline. Results: Using the ICD-11 system, 47% of patients, mainly those with no personality disturbance at baseline, retained their personality status; of the others 16.8% improved and 20.4% worsened to more severe disorder. In DSM-III diagnosed patients, those diagnosed as Cluster A and Cluster C increased in frequency (from 14% to 40%, p < 0.001, and 21.5% to 36%, p < 0.001, respectively) over follow-up, while those with Cluster B showed little change in frequency (22% to 18%, p = 0.197). Conclusion: In this population of patients with common mental disorders, personality status showed many changes over time, inconsistent with the view that personality disorder is a persistent or stable condition. The increase in diagnoses within the Cluster A and C groups suggests personality disorder generally increases in frequency as people age.