The association of major depressive disorders with personality disorders is relevant in terms of clinical, therapeutic and prognostic aspects. However, the prevalence of this association remains unclear. This may be due to methodological considerations. Nonetheless, it could be estimated from this review that 20% to 50% of inpatients and 50% to 85% of out-patients with a current major depressive disorder have an associated personality disorder. Cluster B personality disorders, in particular borderline (10-30%), histrionic (2-20%) and antisocial (0-10%), seem to be overrepresented, even if the narcissistic one is rare (less than 5%). The main characteristic of Cluster C personality disorders is the great variability of results across studies, except for the obsessive compulsive personality disorder, whose prevalence is consistent and rather high (0-20%). Cluster A personality disorders are an heterogeneous group, since the prevalence of schizotypal personality disorder is rather high (0-20%), the prevalence of paranoid personality disorder is low (less than 5%) and the prevalence of schizoid personality disorder is quite variable from one study to another. The prevalence of personality disorders among patients with a lifetime major depression has been insufficiently studied, although it may concern half of these patients. The prevalence of current of lifetime major depression among patients with a personality disorder has not been sufficiently studied and results are very scattered. Also, the coexistence of personality disorder and major depression is frequent, and this review emphasizes the heterogeneity of the personality styles associated with major depression. Finally, an optimization of methods and the adjunction of a dimensional point of view to the categorical approach may help to study the comorbidity of major depression and personality disorders and its consequences.