Macrolide-resistant Helicobacter (H.) pylori represent an increasing therapeutic problem. Macrolide resistance is usually determined phenotypically in vitro with methods such as E-test or agar dilution test. A prerequisite for those tests, however, is bacterial culture that is not routinely set up in the course of gastroscopy. In contrast, formalin-fixed, paraffin-embedded biopsies are regularly available from patients who have undergone gastroscopy. In such biopsies macrolide-resistant H. pylori can be detected by the genotype-based technique of fluorescence in situ hybridization (FISH). Experience gained by this new method, however, is still extremely limited, especially in formalin-fixed tissue. Therefore, we retrospectively investigated formalin-fixed, paraffin-embedded biopsy specimens by FISH in 104 patients suffering from therapy-resistant H. pylori gastritis. To test the accuracy of FISH, we initially examined specimens from 53 patients for whom results of the E-test were available. Next we analyzed biopsies from another 51 patients that had been selected since phenotypical resistance testing had failed despite documented culturing attempts. In all 104 patients, H. pylori was detected by FISH and could thus be investigated for macrolide resistance. Overall, macrolide-resistant bacteria were found in 71 patients (68.3%). In 49 of 53 patients (92.4%), FISH and E-test returned identical results (no significant discordance according to McNemar's chi(2)-test). Taken together, our study demonstrates that FISH is a highly sensitive and reliable method for detecting macrolide-resistant H. pylori in formalin-fixed, paraffin-embedded biopsy specimens, which represents the routine method of processing tissue obtained upon gastroscopy.