BACKGROUND 50 % of the patients seen in the mental health clinics have personality disorders and their diagnosis is very important to establish the patient's prognosis and treatment. The clinician usually only uses one nosology, the ICD-10 or the DSM-IV and is aware of the differences between them. However, he/she gives preference to his/her favorite nosology. METHODS A total of 248 patients seen in a mental health clinic were included in the study. They were diagnosed of anxiety (n = 121) and depressive (n = 127) disorders, with a mean age of 32.3 years for anxiety disorders and 32.4 years depressive disorders. The International Personality Disorders Examination (IPDE) Screening Questionnaire in its DSM-IV and ICD-10 versions was administered to observe the degree of concordance between the two nosologies. CONCLUSIONS The female population predominated in our sample, these being 35.5 % men with anxiety disorders and 65,5 % women, and 29.10 % men and 70.9 % women with depressive disorders. An observational study was performed by studying the percentage with 3 or more positive items and also percentage with 6 or more positive items. It was observed that the DSM-IV version produced more positive diagnoses and more comorbidity. Thus, it is necessary to consider whether the ICD-10 version subdiagnoses personality disorders or the DSM-IV version gives false positives. The qualitative differences between the ICD-10 and the DSM-IV must also be considered. The ICD-10 does not include narcissistic and schizotypal personality disorders and subdivides the borderline personality disorder into two subcategories, so that a given patient can receive different diagnoses according to the nosology used. CONCLUSIONS In the patient sample studied, diagnostic discrepancies in personality disorders between DSM-IV and ICD-10 are so important that the World Health Organization and the American Psychiatry Association should consider unifying categories and diagnostic criteria of personality disorders in the next nosologies.