[Double-blind food challenges in a general hospital: useful and safe, but not without pitfalls]

To describe the results of double-blind placebo-controlled food challenges (DBPCFC) with cow's milk, hen's eggs, soy, peanuts and hazelnuts in a general hospital.

Retrospective chart review.

Food challenges were performed between January 2006 and June 2011, in children 0-18 years of age, on two half-day hospital admissions with a 1-week interval. Tests were performed double-blind following a standardized protocol with validated recipes.

Overall, 234 food challenges were performed, with cow's milk (160), peanuts (35), hen's eggs (21), hazelnuts (11), and soy (7). In two thirds of the cases, the DBPCFC was negative (cow's milk: 57.5%; peanuts: 40.0%; hen's eggs: 66.7%, hazelnuts: 90.9%, soy: 100%). There were no significant associations between patient characteristics and a positive DBPCFC, except for symptoms in the history from 3 different organ systems (p = 0.007). In 2 children (0.9%), serious systemic allergic reactions occurred. Symptoms were recorded on 29.3% of challenges with placebo. In a fifth of the children with a negative test (30/137), symptoms returned when reintroducing the allergen into the diet, mostly (66.7%) transient. Of the 85 tests regarded as positive by the attending physician, 19 (22.4%) were false-positive, i.e. they were not positive according to predefined criteria; a false-positive test result occurred in particular with non-specific symptoms.

The DBPCFC can be safely performed in a general hospital for a range of food allergens. The test result is negative in most cases except for peanuts. Non-specific symptoms may hinder the interpretation of the DBPCFC and can result in false-positive results.