Study objectives: Epistaxis is a common emergency, but no emergency department (ED)–based epidemiology has been published. There is also significant controversy ns, such as the influence of age and weather. We study the demographics and trends in epistaxis visits to all US EDs for the 10-year period 1992 to 2001. Methods: ED visit data from the National Hospital Ambulatory Medical Care Survey (NHAMCS) were used to estimate the nationwide number of epistaxis visits for 1992 to 2001. Inclusion criteria were International Classification of Diseases, Ninth Revision, Clinical Modification diagnosis code 784.7 or chief complaint code 1405.1. Traumatic visits were identified as those with a supplementary classification of external injury and poisoning code (E-codes) or a reason-for-visit code related to an injury, poisoning, or adverse effect (codes 5001 to 5999). National estimates were obtained using NHAMCS patient visit weights. Rates were calculated using midyear population estimates from the US Census Bureau, by age, sex, race, and region (Northeast, Midwest, South, West). The standard error of the estimate was used to calculate 95% confidence intervals (CIs). Stata 7.0 software (StataCorp, College Station, TX) was used for analyses. Results: From 1992 to 2001, 4,503,000 visits for epistaxis occurred in US EDs (95% CI 3,993,000 to 5,013,000), accounting for 4.6 of every 1,000 ED visits (95% CI 4.1 to 5.1). On a population basis, there were 1.7 visits (95% CI 1.5 to 1.9) per 1,000 US residents. The sex-specific rates per 1,000 ED visits were 5.3 (95% CI 4.5 to 6.0) for men and 4.0 (95% CI 3.4 to 4.7) for women. Rates were similar among blacks and whites, Hispanics and non-Hispanics, and in metropolitan and nonmetropolitan areas. Rates did not change during the 10-year period of study. Arrival by ambulance was reported in 15% of visits (95% CI 12% to 18%). Most patients (94%) were discharged (95% CI 91% to 97%). Use of a medication for epistaxis was reported in 57% of visits (95% CI 52% to 62%). Cocaine was most common, followed by phenylephrine, silver nitrate, and oxymetazoline. Overall frequency of visits for epistaxis was strongly related to age, which was not because of a general trend toward more visits with increasing age: of 1,000 ED visits by those aged 70 to 79 years, epistaxis was present in 12.0 (95% CI 9.1 to 14.9) versus 2.3 (95% CI 1.2 to 3.3) for visits by those aged 20 to 39 years. The rate per 1,000 ED visits for children younger than 10 years was 4.0 (95% CI 3.0 to 5.1). Eighty-three percent of epistaxis cases were nontraumatic (95% CI 80% to 86%). Patients with nontraumatic epistaxis were older than those with traumatic epistaxis, with mean ages of 49.4 years (95% CI 47.2 to 51.6 years) and 30.8 years (95% CI 26.2 to 35.5 years), respectively. Nontraumatic epistaxis varied with the seasons, accounting for 50 of every 1,000 ED visits during the winter months of December to February (95% CI 40 to 60) and 34 per 1,000 ED visits during the rest of the year (95% CI 30 to 39). The seasonal variation of nontraumatic epistaxis was observed in all regions except the Western region. Traumatic epistaxis did not vary seasonally. Conclusion: Epistaxis is present in about one half of 1% of all ED visits. Among ED patients with epistaxis, the elderly predominate. The proportion of ED visits among those aged 70 to 79 years was nearly 6 times greater than among those aged 20 to 39 years. Even small children younger than 10 years had one third as high a proportion of all ED visits because of epistaxis compared with those aged 70 to 79 years. Nontraumatic epistaxis is significantly more common during the winter, possibly because of cold or low humidity. Though there was no seasonal variation in the Western region, the seasonal variation in the other regions was strong enough to cause significant variation in the nationwide sample. Our study was limited by the fact that NHAMCS records information about visits only, not individual patients. We were also unable to answer the much-debated question of the relationship between epistaxis and hypertension, because hypertension may not be recorded reliably in NHAMCS. Finally, the predominance of cocaine as a therapeutic choice could be due to reporting bias, given the close tracking of this controlled substance.