Several studies suggest that OCD is probably a heterogeneous disorder with several putative subtypes. Differences in treatment response, disease course, and clinical features are suggestive of the existence of subtypes with specific underlying pathophysiological substrates. Whether studying OCD from a phenomenological or genetic standpoint phenotypic characterization becomes crucial. This manuscript summarizes data that support the existence of a subgroup of OCD related to tics, a group with an early age onset; a possible subgroup related to rheumatic fever; and a group with poor insight. Validation of these putative subtypes may be possible with genetic family studies and will contribute to the development of more specific treatments.