A presentation of findings from a large population of anaphylaxis cases. Retrospective chart review and follow-up questionnaire provided data on 266 subjects (113 males and 153 females) aged 12 to 75 years (mean age, 38 years) who were referred to a university-affiliated private allergy-immunology practice in Memphis, Tenn, for evaluation and management of anaphylaxis from January 1978 through March 1992. Of 266 subjects, 162 (61%) had three or more anaphylactic episodes, 41 (15%) had two episodes, and 63 (24%) had one episode. Atopy was present in 98 individuals (37%). Physicians thought foods, spices, and food additives caused anaphylaxis in 89 individuals (34%); crustaceans and peanut accounted for about half of these cases. Medications were thought to have caused the anaphylactic episodes in 52 individuals (20%); nonsteroidal anti-inflammatory drugs in about half of these cases. Other probable causes included exercise (n = 19), latex (n = 2), hormonal changes (n = 2), and insect bites (n = 4). A suspected cause could not be determined in 98 individuals (37%). These subjects were diagnosed as having idiopathic anaphylaxis. Of the 266 subjects, 102 responded to a follow-up survey; 68 (67%) of the 102 were thought to have identifiable causes of anaphylaxis (32 of whom [47%] failed to carry epinephrine syringes for self-administration despite instructions to do so). In contrast, of 34 subjects with idiopathic anaphylaxis who responded to the survey, only three (9%) did not carry epinephrine. (1) Atopy is common in subjects who experience anaphylaxis, regardless of its origin; (2) crustaceans and nonsteroidal anti-inflammatory drugs are the most common food and medication groups, respectively, thought to cause anaphylaxis; (3) causative agents can be identified for two thirds of the subjects, and recurrent attacks are the rule; and (4) subjects with idiopathic anaphylaxis are more likely to carry epinephrine for self-administration than those with identifiable causes.