We compared smoking prevalence in 217 patients with panic disorder with that in 217 age- and sex-matched control subjects who were obtained by telephone survey from the same neighborhoods. Data were obtained for current smoking habits and smoking status at either the onset of illness (patients) or 10 years previously (control subjects). Patients had been ill for 10.6 (SD = 10.0) years. Female patients with panic disorder had a significantly higher smoking prevalence at the onset of their illness than did control subjects 10 years previously (54% vs. 35%). The current smoking prevalence for female patients was also significantly higher than that of control subjects (40% vs. 25%). Male smoking rates did not differ between patients and control subjects. Caffeine use did not appear to explain these findings. These data suggest a link between smoking behavior and panic disorder in women.