Despite the importance of anaphylaxis, little information is available on its clinical features. To evaluate the clinical and allergologic features of anaphylaxis in children referred to the allergology and immunology unit of A. Meyer Children's Hospital (Florence, Italy) from 1994 to 1996. Ninety-five episodes of anaphylaxis occurred in 76 children (50 boys and 26 girls). Sixty-six children (87%) had only one episode of anaphylaxis, while 10 (13%) had two or more episodes. Sixty-two (82%) of the 76 patients had a personal history of atopic symptoms, although 14 (18%) did not. Sixty (79%) of the 76 children studied had at least one positive skin prick test to one or more of the common inhalant and/or food allergens. Children with venom-induced anaphylaxis usually had negative skin tests to the allergens tested. A younger age and eczema were more frequent among children with food-dependent anaphylaxis, whereas an older age together with urticaria-angioedema were common among those with exercise-induced anaphylaxis. The mean latent period (+/-SD) of the anaphylaxis episodes was 15.4 +/- 27.5 minutes. Skin and respiratory manifestations had an earlier onset and were more common than the gastrointestinal and cardiovascular ones. The most frequent clinical manifestation in children with food anaphylaxis was gastrointestinal symptoms, whereas cardiovascular symptoms were rare. The most probable causative agents in the 95 episodes described were foods (57%), drugs (11%), hymenoptera venom (12%), exercise (9%), additives (1%), specific immunotherapy (1%), latex (1%), and vaccines (2%), but in 6 cases (6%) the agent was never determined. Among the foods, seafood and milk were the most frequently involved. As for location, 57% of the anaphylactic events occurred in the home (54/95), 12% outdoors (11/95); 5% in restaurants (5/95); 3% in the doctor's office (3/95); 3% in hospitals (3/95); 3% on football fields (3/95); 2% on the beach (2/95); 1% in the gym (1/95); 1% at school (1/95); and 1% in the operating room (1/95). In the remaining 12% of cases (11/95) the site remained unknown. Sixty-two percent of the patients (59/95) were treated in an emergency room or hospital, while 32% (30/95) were not (this information is lacking in 6% of the cases [6/95]). Patients were treated with corticosteroids in 72% of the cases (68/95), with antihistamines in 20% (19/95), with epinephrine in 18% (17/95), with beta2-agonists in 5% (5/95), and with oxygen in 4% (4/95). In our area, foods, particularly seafood and milk, seem to be the most important etiologic factors triggering anaphylaxis. Food-induced anaphylaxis often occurs in younger children with a severe food allergy, whereas exercise-induced anaphylaxis occurs more often in older children with a history of urticaria-angioedema. The venom-induced variant usually presents itself in nonatopic subjects. Given the fact that most of the children had only one anaphylactic reaction, prevention is almost impossible. Epinephrine, although it is the first-choice treatment of anaphylaxis, often goes unused, even in hospitals and doctors' offices.