This controlled family study explores (1) whether panic disorder and unipolar depression share familial factors, and (2) whether the co-occurrence of lifetime diagnoses of panic disorder and unipolar depression in individuals defines a distinct diagnostic subtype in terms of familial aggregation. To be most informative, the familial lifetime prevalence rates for panic disorder and unipolar depression have to be determined in a set of four proband groups: 78 patients with unipolar depression and panic disorder. 121 patients with unipolar depression alone (no panic disorder), 81 patients with panic disorder alone (no unipolar depression), and 109 control probands sampled in the general population were compared by lifetime prevalence rates for panic disorder and unipolar depression in their first-degree relatives. Altogether 1046 relatives were interviewed directly; family history information was available on another 346 subjects. Both disorders were aggregating in families. We found modest overlap of familial components; the relative risk of panic disorder only in relatives of patients with unipolar depression only was 2.3, and the relative risk of unipolar depression only in relatives of patients with panic disorder only was 1.8. The comorbid condition did not represent a distinct subtype in terms of familial aggregation. Excess comorbidity was observed in affected relatives independent of the diagnostic status of the index case. Thus, a sharing of familial factors of aetiological relevance between panic disorder and unipolar depression might explain a limited proportion of comorbid cases. However, the major proportion of comorbidity between panic disorder and unipolar depression may still be due to non-familial factors.