Insight into the dynamic time courses of patch test reactions may be useful for patch test evaluation. To put this to the proof, we have analysed the time courses of allergic patch test reactions that were filed between 1990 and 1994 by the Information Network of Departments of Dermatology in Germany (IVDK). Based on reaction strength at days 1 to 3, each allergic patch test was assigned to 1 of 5 distinct dynamic reaction patterns: crescendo, plateau, decrescendo, crescendo/decrescendo, or decrescendo/ crescendo. The frequencies of these patterns were determined for those 10 standard allergens yielding the highest numbers of allergic reactions (n between 180 and 1240), and dynamic patterns were classified as type 1 (crescendo or plateau patterns) or type 2 (other patterns). Influences of population characteristics upon the classification of patterns were evaluated by logistic regression analyses. Different frequencies of reaction patterns were found for distinct allergens. For all 10 allergens, type-1 patterns were significantly more often observed for allergic reactions assessed as clinically relevant. For 5 allergens, female sex of patients was associated with a lower frequency of type-2 patterns. A history of atopic dermatitis and the age of patients were only weakly associated with a certain classification of reaction patterns in some of the allergens. In a comparison of reaction patterns obtained with different allergens in the same population, we found that nickel sulfate elicited type 1 reaction patterns more often than fragrance mix. We conclude that the dynamic reaction patterns should be taken into account in the evaluation of patch tests. The relation of type-1 patterns to clinical relevance of the reaction needs further attention.