DISCUSSION
The analyses reported here demonstrate that methodological approaches are critical to our eventual understanding of a possible relationship between MDD and PD. The two approaches taken by Crowe et al. (6), Noyes et al. (15), and Leckman et al. (12) differ in their use of the primary/secondary distinction. From the work of Crowe, and Noyes and colleagues, it is clear that there is not an elevated risk of chronologically primary MDD in families of probands with chronologically primary PD. However, from the work of Leckman et al. and others (12,19) and from the results reported here it is clear that the rate of MDD (without regard to whether it is primary or secondary) is significantly elevated in families of both MDD and PD probands. What is not clear from the data presented here is whether that increased rate of MDD is among all relatives of PD or agoraphobic probands or whether it occurs only in those families where the proband had a diagnosis of MDD. Leckman et al. (12) reported that there was a significant difference between relatives of MDD versus MDD + PD probands. Crowe et al. (6) and Noyes et al. (15) also reported that the rate of secondary depression was elevated among relatives of PD probands with secondary depression, but the data do not allow a direct comparison to the results of Leckman et al. (12). Nevertheless, it appears that the co-occurrence of MDD and PD within the same individual is an important predictive factor for MDD among the relatives regardless of whether PD or MDD is the chronologically primary disorder. Clinical studies have examined patients with both MDD and PD and have shown differences when compared to patients with MDD alone. Van Valkenburg et al. (20) showed such patients to have the poorest outcome on follow-up. They were also the patients who responded least favorably to tricyclic antidepressant intervention. Nutzinger and Zapotoczky (16) also demonstrated that individuals with "cardiac phobia" who had a secondary depression were more likely to be ill at follow-up. Phenomenological studies have suggested that the episodes of depression that occur together with panic disorder may be somewhat unique. Vitone and Uhde (21) found that patients with both panic and depression were more likely to have episodes of severe depression with melancholic features. However, the episodes of depression tended to be shorter, with only 24% of patients having depressive episodes lasting more than 2 weeks. Given these findings and the data reported in this study, it appears that individuals with both PD and MDD may have a unique syndrome when compared to individuals with only MDD or PD. In order to determine if this is the case, additional phenomenological, biological, and pharmacological studies are needed in which patients with MDD alone, PD alone, PD + MDD, and MDD + PD are studied together with their families. It must be emphasized that in future studies particular attention needs to be paid to how diagnostic assessments are made and reported. Reporting only primary diagnoses can obscure important information. On the other hand, reporting only rates of diagnosis can also be misleading. All relevant diagnoses of each individual need to be reported in such a way so that it is possible to determine whether relatives have the same diagnostic profile as the probands.