Food extracts for diagnostic purposes often lack sufficient activity and consistency. Biologically standardized food extracts are not available on the market. Using extracts from plant-derived foods as examples, we investigated factors which may be important for the quality of such extracts. Divergent allergenic activities were found between strains of apples, but not within varieties of celery tuber (celeriac), hazelnut, and peanut, respectively. Heating of the food remarkably reduced the activity of apple, hazelnut, and celeriac, but had little effect on peanut. By contrast, heating of semipurified protein extracts from celery tuber and apple for 30 min at 100 degrees C did not deplete the immunoreactivity of the major allergens, indicating that this is an inappropriate test for identifying labile food allergens. Due to their high endogenous enzyme activities, apples and other fruits require special extraction procedures applying either low temperature or enzyme inhibitors. Variation of extraction conditions had little effect on the composition and activity of extracts from hazelnut. The storage stability of skin test solutions from plant foods can be improved by avoiding phenol as an additive and by including 50% of glycerol. For model studies considering neoallergens, IgE was raised in mice against native and heated celery tuber, respectively. When extracts from nonthermally and thermally processed celeriac were subjected to an RBL-cell mediator release assay with these sera, an inverse ranking was obtained with anti-heated celeriac IgE and anti-native celeriac IgE, respectively. These data indicated that new epitopes had been formed by the heating process. Since all parameters were tested in model experiments with either human or murine IgE, their relevance has to be proven in further clinical investigations.