Objectives: Our aim was to examine patterns of Axis II co-morbidity using data from the national survey of psychiatric morbidity among prisoners in England and Wales. Methods: A one-in-five sub-sample of participants in a survey of psychiatric morbidity among prisoners in England and Wales was interviewed using the Schedule for Clinical Assessment in Neuropsychiatry and the Structured Clinical Interview for the fourth edition of the Diagnostic and Statistical Manual of Mental Health Disorders Axis II personality disorders (PD). Logistic regression analysis was conducted adjusting for confounders of associations with co-morbid psychopathology disorders, age and gender. Results: The most prevalent Axis II disorders in the sample were anti-social, paranoid and borderline PD. Following logistic regression, anti-social and borderline PD demonstrated high levels of co-morbidity with both Axis I and other Axis II disorders, narcissistic PD with other Axis II, and paranoid and avoidant PD with Axis I disorders. Conclusions: Certain Axis II disorders may increase the risk for lifetime Axis I disorders. Although appropriate statistical procedures reduce the level of Axis II co-morbidity, some patterns may be artefacts of a diagnostic system encouraging multiple diagnostic categories.