OBJECTIVE: The authors performed an analysis of their published chromosome 18 linkage data on 28 families in which there was bipolar disorder to test the potential of comorbid panic disorder to define a genetic subtype of bipolar disorder. METHOD: Families ascertained through probands with bipolar I disorder were stratified into three groups based on a history of panic disorder, panic attacks, or no panic attacks in the probands. Multipoint nonparametric linkage analysis was performed on data from bipolar I and II family members in each group. RESULTS: Linkage scores for five consecutive 18q marker loci were highest in the families of the probands with panic disorder and lowest for the families of the probands without panic attacks. CONCLUSIONS: This study supports the authors' previously reported clinical hypothesis of a genetic subtype of bipolar disorder identified by comorbid panic disorder. The hypothesis merits prospective testing.