BACKGROUND It is unclear whether near-patient whole-blood diagnostic tests for Helicobacter pylori are of comparable accuracy to laboratory based ELISA for screening of dyspeptic patients prior to endoscopy. OBJECTIVE To compare two ELISA and two whole-blood tests in order to determine whether near-patient H. pylori diagnostic tests are an acceptable alternative to laboratory based ELISA tests for screening of dyspeptic patients prior to endoscopy. METHODS One hundred and seven consecutive patients with dyspepsia (median age, 32 years; range, 16-45 years) were evaluated with Helico-G ELISA, Hmcap ELISA and Helisal whole-blood tests. A further 111 dyspeptic patients (median age, 51 years; range, 16-96 years) were evaluated with the Immunocard whole-blood test only. The 'gold standard' for infection was based on histology and the rapid urease test (CLO). RESULTS Compared to the Helico-G test, both near-patient tests had a higher false negative rate (23-37% vs 5%, P< 0.003), and lower sensitivity and negative predictive value. The Immunocard had a higher specificity than did the Helisal (87% vs 63%, P=0.006); otherwise both near-patient whole-blood tests had similar performance. At a sensitivity of 95%, the Hmcap ELISA was more specific than the Helico-G ELISA (75% vs 67%) and had fewer false positives (25% vs 32%). The near-patient tests would wrongly classify up to 40% H. pylori positive dyspeptic patients and exclude them from endoscopy, compared to 5-6% for ELISA. CONCLUSIONS Near-patient whole-blood H. pylori diagnostic tests are less accurate and thus not an acceptable alternative to laboratory based ELISA tests.