To the Editor.—
Weissman et al1 and Johnson et al2 reported cross-sectional survey results from the National Institute of Mental Health Epidemiologic Catchment Area Program to show that a history of panic attacks signaled a markedly high lifetime prevalence of suicide attempts, even in the absence of other diagnosed psychiatric disorders. On this basis, the authors argued that physicians should be attentive to panic symptoms shown in their patients. We were concerned that reliance on cross-sectional lifetime prevalence data might create a spurious artifact in the form of an association between a history of panic attacks and a history of suicide attempts. For this reason, we extended a prior casecontrol analysis based on 40 community survey respondents who recently attempted suicide and 160 matched controls.3 The recent suicide attempters were detected at cross-section and at follow-up after 1 year in the prospective component of the Epidemiologic Catchment