Consumption rates of anxiolytic drugs, and especially of benzodiazepines, remain very high in France compared to other Western countries, whereas clinical guidelines limit their indications to short term treatments and only for some precise anxiety disorders. Recent epidemiologic surveys in the community indicated that more than 15% of people used once or more an anxiolytic drug in the past year. The issue of chronic treatments is particularly crucial because of their poor benefit/risk ratio in most anxiety disorders (limited efficacy, cognitive side effects, withdrawal and dependence problems). To address this important public health issue, and knowing that, in France, benzodiazepines are prescribed mainly by general physicians, our aims were to explore psychiatric diagnoses in GP's patients with chronic use of anxiolytic benzodiazepines. We included 4 425 patients consuming such drugs regularly for six months or more, and assessed their anxiety and depression symptoms through various clinical scales (Hospital Anxiety and Depressive scale - HAD, Clinical Global Impression scale - CGI, Sheehan Disability Scale - SDS, Cognitive Dependence to Benzodiazepines scale - CDB) and with the Mini International Neuropsychiatric Interview for DSM IV criteria. Only 2.2% of the subjects had neither anxious nor depressive symptoms as indicated by low scores on both subscores (less than 8) of the HAD scale, used as a screener. Nearly three quarters of the 4,257 subjects (73.2%), had CGI scores of at least 5 (markedly ill to extremely ill). Social and familial disability was also high in more than 40% of the sample (marked to extreme disruption according to SDS scores). About half of the sample had CDB scores suggesting a benzodiazepine dependence. According to the MINI, 85.1% of the patients had at least one current DSM IV diagnosis of affective disorder. The most frequent diagnoses were major depressive episode (60%), generalized anxiety disorder (61.2%), and panic disorder (22.5%). An anxiety and depressive comorbidity wad found in 41.9% of the subjects. Some methodological limitations must be taken into account in the discussion of our results, and especially the fact that the included patients were not supposed to be totally representative of all patients consuming anxiolytic benzodiazepines in general practice. However, the size of our sample is sufficiently large to limit possible biases in patient selection. The main result of this study is that a great majority of the patients had significant symptomatology, in particular major depressive episodes and generalized anxiety disorder, often with marked severity and disability. These data are in line with the knowledge of a lack of efficacy of benzodiazepines in depressive and most anxiety disorders, despite long term treatment. They also confirm the current guidelines which recommend prescribing serotoninergic antidepressants, and not benzodiazepines, when long term treatments are needed for severe and chronic affective disorders. This epidemiologic study leads to the conclusion that a specific and attentive diagnostic assessment should be done in all patients receiving benzodiazepines for more than three months, in order to purpose in many cases other long term therapeutic strategies.