To determine why a significant proportion of birch pollen-sensitive patients do not have the oral allergy syndrome (OAS), possible predictive in vivo or in vitro tests for OAS were sought in a large retrospective and prospective follow-up study performed in 283 patients with clinical evidence of birch pollen hypersensitivity. OAS was associated with more severe respiratory symptoms and with higher birch-specific and total IgE levels; moreover, its onset was clearly related to duration of birch pollinosis. The prospective part of this study, performed in 63 patients without OAS, confirmed these findings and highlighted the very high negative predictive value of both skin prick tests with fresh foods and RAST with food allergens. This work suggests that about 15% of patients with birch pollen hypersensitivity are not prone to OAS and that their anti-birch IgE might be directed against determinants that do not crossreact with food allergens.