Objective. There is a growing body of clinical and laboratory evidence
to support the notion that food allergy plays a role in the pathogene
sis of atopic dermatitis (AD). However, the incidence of IgE-mediated
food allergy in children with AD is not well established. Design. A pr
ospective study to determine the prevalence of IgE-mediated food hyper
sensitivity among patients referred to a university-based dermatologis
t for evaluation of AD. Setting. University hospital pediatric dermato
logy clinic. Patients. A total of 63 patients with AD were recruited (
35 male; 32 white, 24 African-American, 7 Asian). Methods. Patients we
re assigned an AD symptom score (SCORAD) and were screened for food-sp
ecific serum IgE antibodies to six foods (milk, egg, wheat, soy, peanu
t, fish) known to be the most allergenic in children. The levels of fo
od-specific serum IgE were determined by the CAP System fluoroscein-en
zyme immunoassay (CAP); patients with a value greater than or equal to
0.7 kIUa/L were invited for an additional allergy evaluation. Those w
ith CAP values below the cutoff were considered not food allergic. Pat
ients were considered to be allergic if they met one of the following
criteria for at least one food: 1) reaction on food challenge; 2) CAP
value more than the 95% confidence interval predictive for a reaction;
3) convincing history of an acute significant (hives, respiratory sym
ptoms) reaction after the isolated ingestion of a food to which there
was a positive CAP or prick skin test. Results. A total of 63 patients
(median age, 2.8 years; median SCORAD, 41.1) were recruited; 22 had n
egative CAP values (without a significant difference in age or SCORAD
score, compared with the 41 with positive specific IgE values). Furthe
r allergy evaluation was offered to the 41 remaining patients; 10 were
lost to follow-up and 31 were evaluated further. Of these, 19 underwe
nt a total of 50 food challenges (36 double-blind, placebo-controlled,
and 14 open), with 11 patients experiencing 18 positive challenges (9
4% with skin reactions). Additionally, 6 patients had a convincing his
tory with a predictive level of IgE; 5 had a convincing history with p
ositive, indeterminate levels of IgE; and 1 had predictive levels of I
gE (to egg and peanut) without a history of an acute reaction. Overall
, 23/63 (37%; 95% confidence interval, 25% to 50%) had clinically sign
ificant IgE-mediated food hypersensitivity without a significant diffe
rence in age or symptom score between those with or without food aller
gy.Conclusions. Approximately one third of children with refractory, m
oderate-severe AD have IgE-mediated clinical reactivity to food protei
ns. The prevalence of food allergy in this population is significantly
higher than that in the general population, and an evaluation for foo
d allergy should be considered in these patients.