The relationship between anxiety and impulsivity is controversial and not well explored. In a previous study we compared impulsivity, measured by different rating tools, in patients with anxiety disorders vs. healthy controls. In the same sample we now explore the influence of comorbid soft bipolar spectrum disorders on the relationship between anxiety disorders and impulsivity. A sample including 47 subjects with anxiety disorder(s) and 45 control subjects matched for demographic, educational and work characteristics underwent a diagnostic evaluation by the Mini Neuropsychiatric Interview (MINI); a symptomatological evaluation by the Bech-Rafaelsen Depression and Mania Scale (BRDMS), the State-Trait Anxiety Inventory (STAI), the Hypomania Check List (HCL-32) and the Clinical Global Impression (CGI); a temperamental and personological evaluation by the Questionnaire for the Affective and Anxious Temperament Evaluation of Memphis, Pisa, Paris and San Diego-Modified (TEMPS-M), the Separation Anxiety Symptoms Inventory (SASI), the Interpersonal Sensitivity Symptoms Inventory (ISSI); and, finally, a psychometric and a neuro-cognitive evaluation of impulsivity by the Barratt Impulsiveness Scale (BIS) and the Immediate and Delayed Memory Task (IMT/DMT). The initial sample of patients with anxiety disorders was then subdivided into two subgroups depending on the presence of comorbid cyclothymia (Cyclo+, n=26 and Cyclo-, n=21). For the diagnosis of cyclothymic disorder, we used both the DSM-IV-TR criteria and also a modified threshold for hypomania with a duration of 2 days. We compared symptomatological, temperamental, personological and impulsivity measures in Cyclo+, Cyclo- and controls. The comparison between Cyclo+, Cyclo- and controls showed that Cyclo+ are the most impulsive subjects in all the investigated measures and are characterized by greatest symptomatological impairment, highest scores in temperamental scales, and highest levels of interpersonal sensitivity and separation anxiety. Cyclo- subjects resulted to be more impulsive compared to controls concerning the retrospective trait measures, but not in the neuro-cognitive test. Correlational cross-sectional study. In our patients with anxiety disorders, without lifetime comorbidity with major mood episodes, trait and state impulsivity resulted to be greater than in controls. In particular impulsivity was highest in patients with both anxiety disorders and cyclothymia. In anxious-cyclothymic patients also separation anxiety and interpersonal sensitivity were more severe than in anxious patients without cyclothymia and controls. Our findings suggest that impulsivity, rather than being directly related to the presence of the anxiety disorder, could be associated with comorbidity with cyclothymia.