BACKGROUND: In 1984, the German, Austrian, and Swiss Photopatch Test Group was founded to standardize the photopatch test procedure and to investigate photoallergic reactions, as well as the epidemiology of photoallergy, in central Europe. Therefore in a first test period from 1985-1990, 32 test substances were applied on the backs of patients suspected to be photosensitive. After evaluation of these data, some substances were dismissed, and others were additionally integrated into the test tray. Thus a modified test tray comprising 26 test substances was used for the second test period (1991-1997). OBJECTIVE: According to the standard photopatch test procedure defined in the first test period, the aim of this multicenter study was to apply compounds from the modified second test tray to a large group of photosensitive patients. After evaluation of the second test period (1991-1997), the outcome was compared with the results of the first period. On the basis of these results and influenced by concurrently published case reports, a third modified photopatch test tray has been established. METHODS: After the application of a duplicate test tray for 24 hours, one test site was irradiated with 10 J/cm(2) UVA, and the other test site served as the control area. Readings were performed immediately and 24, 48, and 72 hours after irradiation. Test reactions were qualitatively graded according to a 4-point scale and classified by the investigators of the participating centers. In 49 participating clinics 1261 photopatch tests were performed. All data were subjected to computer-assisted analysis by using a specially developed software to classify all positive test reactions as plain contact or photoinduced reactions (nonspecific, toxic, or allergic photoreactions) and to define substance-specific reaction patterns. RESULTS: In test period 1 data of 1129 patients were evaluated. From 2859 positive test reactions, 28.6% were excluded as plain contact reactions, 71.4% were found to be photoinduced reactions, and 3.8% were classified as photoallergic. In test period 2 data of 1261 patients were evaluated. One thousand four hundred fifteen positive test reactions were observed, and of these, 28.7% were excluded as plain contact reactions, 71.3% were classified as photoinduced reactions, and 8.1% were classified as photoallergic reactions. In both test periods nonsteroidal anti-inflammatory drugs, disinfectants, and phenothiazines represented the leading photoallergens in the evaluated central European region. By using computer-assisted reaction pattern analysis, substance-specific reaction patterns could be distinguished. These substance-specific reaction patterns comprised 4 main categories: the well-known decrescendo (phototoxic) and crescendo (photoallergic) reaction patterns, as well as a combined and a plateau pattern. CONCLUSION: The test modification after the first test period led to a notably reduced number of positive (mainly nonphotoallergic and thus nonrelevant) test reactions per patient in the second test period (from 2.6 to 1.1). In contrast, the percentage of photoallergic reactions increased significantly from 3.8% to 8.1% of all positive test reactions. Test modifications after the first test period led to a remarkably improved specificity of the photopatch test. Furthermore, substance-specific reaction patterns observed in test period 1 were confirmed in test period 2.