We appreciate the comments of Hiyama and colleagues on our article in CHEST (August 2000).1They stated that the percentage of cryptogenic hemoptysis (42%) was particularly high in our study compared with other previously published studies and with their own results (12%). The distribution of causes of hemoptysis differed greatly in different studies,2depending on geographic setting, time, and design of the study, and patient inclusion (or exclusion) criteria. Hiyama and colleagues studied a clinical series of 51 patients hospitalized for hemoptysis during a 6-year period, with detailed information and prospective follow-up. In contrast, we performed a retrospective epidemiologic study, covering a broad population (6,349 subjects who had, during a 3-year period, received a discharge diagnosis of spontaneous hemoptysis, with less detailed information collected from a large electronic discharge data file. This also explains why we had no information on smoking history or on the amount of expectorated blood. Furthermore, in our study, patients were not selected, and our data file included all kinds of patients, regardless of their age, the completeness of the etiologic investigation, and the amount and/or duration of bleeding. It is possible that a small number of the cases could have been erroneously classified as cryptogenic hemoptysis, but it is noteworthy that, in the study by Adelman et al,3 cryptogenic hemoptysis was identified in approximately 30% of all patients.