We evaluated the performance of three enzyme-linked immunosorbent assays (ELISAs) in detecting serum immunoglobulin G (IgG) and IgA antibodies to Helicobacter pylori; two were new ones from Pyloriset (Pyloriset EIA-G update and Pyloriset EIA-A update; Orion Diagnostica, Espoo, Finland), and the third was the Malakit EIA-G (Biolab, Limal, Belgium). Serum samples from 154 dyspeptic patients were collected. As a reference method, multiple biopsy specimens from different anatomical areas of the stomach were obtained by endoscopy and were analyzed by culture and/or histology and direct urease testing. Accordingly, 126 patients (82%) were found to be H. pylori positive and 28 patients (18%) were found to be H. pylori negative. To validate serology as a predictor of H. pylori infection, sensitivity, specificity, positive and negative predictive values, and accuracy of the assays were calculated against the H. pylori status as determined by the reference method. The corresponding data for the different ELISAs were 100%, 79%, 95%, 100%, and 96% for the Pyloriset ELA-G update, 81%, 89%, 97%, 52%, and 82% for the Pyloriset EIA-A update, and 87%, 86%, 96%, 60%, and 87% for the Malakit EIA-G, respectively. We conclude that the Pyloriset EIA-G update is a reliable and accurate test and that because of its 100% sensitivity, conjunctional IgA testing is not necessary. Its 100% negative predictive value makes it a very useful screening test. For purposes of excluding infection with H. pylori, the performance of the Malakit EIA-G is moderate but can be improved by conjunctional IgA testing. The Pyloriset EIA-A update can be useful as such a conjunctional test.