Over 9% of the subjects in the NIMH Epidemiologic Catchment Area sample reported a history of at least one panic attack, yet only 1.5% of the sample met DSM-III criteria for panic disorder. In light of the revisions of these criteria for DSM-IV, two of those diagnostic criteria, frequency of panic attacks and number of symptoms during a panic attack, are examined in this paper. We evaluated the relationship between those criteria and each of three measures of social morbidity:psychiatric hospitalization, emergency room (ER) use for emotional problems, and suicide attempts. Although only symptom count was significantly associated with either ER use or suicide attempts, both diagnostic criteria were associated with psychiatric hospitalization. The threshold for symptom count was also considered and while three appeared to have slightly better sensitivity and specificity, the continued use of four symptoms seemed acceptable. The diagnostic and clinical implications of the findings are considered.