OBJECTIVE: The purpose of this study was to examine patterns of co- occurrence of axis II disorders in a group of consecutive patients evaluated with two contrasting structured interviews. METHOD: One hundred of 106 consecutive applicants for long-term, inpatient treatment of severe personality psychopathology were assessed, face-to- face, by psychiatrists using the Structured Clinical Interview for DSM- III-R Personality Disorders (SCID-II) and the Personality Disorder Examination (PDE). The percent of co-occurrence of pairs of disorders diagnosed by each structured interview separately was calculated, and significance levels were determined by using chi-square tests of independence. Finally, odds ratios were computed for the odds of each pair of disorders occurring together compared with the odds for the occurrence of each disorder alone. RESULTS: The two interview methods revealed different comorbidity patterns. Significant covariation was found for 29 pairs of disorders diagnosed with the PDE, compared with 12 pairs diagnosed with the SCID-II. Six pairs of disorders covaried significantly and were associated with odds ratios greater than 4: histrionic with borderline, histrionic with narcissistic, narcissistic with antisocial, narcissistic with passive-aggressive, avoidant with schizotypal, and avoidant with dependent. CONCLUSIONS: Substantial overlap occurred among personality disorders. In this group of patients, consistent patterns of comorbidity involving narcissistic, avoidant, and histrionic personality disorders suggest that categorical distinctions between them and certain other DSM-III-R personality disorders may be illusory. The question of which of two overlapping disorders is more valid, however, is left unanswered. For clinical purposes, a two-level diagnostic convention is proposed.Abstract Teaser