Dysfunctions of dopaminergic transmission have been implicated in the pathogenesis of mood disorders and variants at the catechol-O-methyltransferase (COMT) gene have been recently reported to be associated with bipolar disorder. In this study, we tried to replicate this finding in an independent sample of participants affected by mood disorders. Eight hundred and five psychiatric in-patients affected by bipolar (n=446) and major depressive (n=359) disorder were included. All patients and 116 controls were typed for the COMT variants using polymerase chain reaction techniques. A subsample of 458 patients was assessed for their lifetime symptomatology. The sample was analyzed alone and pooled with 111 participants and 127 controls collected in the center of Pisa and previously reported. COMT variants were not associated with mood disorders in our sample, a trend was observed toward an excess of COMT*A allele in the pooled sample. The analysis of symptomatology factors did not show any significant difference as well. Possible stratification factors such as sex, current age and age of onset did not affect the observed results. Control subjects from the two centers showed a similar genotype distribution. In conclusion, COMT variants were not a major liability factor for mood disorder and for mood disorder symptomatology in the present sample; however, we could not subdivide our sample on the basis of panic comorbidity as in the original paper and this could be considered a possible reason for the discrepant finding.