Several direct/invasive and indirect/non-invasive diagnostic tests are available for the diagnosis of H. pylori infection. Invasive tests require biopsy sampling of the gastric mucosa and include rapid urease test, histology, bacterial culture and polymerase chain reaction technique. Non-invasive tests include the urea breath test and serological assays. This review gives a critical comparative analysis of accuracy, advantages and limitations of the different diagnostic tests including current cost and availability in Belgium. Rapid urease testing (RUT) of gastric biopsy specimens is probably the initial test of choice in patients undergoing endoscopy because of its low cost, rapid availability of results, simplicity and accuracy. Histological examination of gastric biopsy samples should be mandatory at the initial presentation of the patient because it also gives insight on the status of the gastric mucosa (inflammation & premalignant changes). Although not mandatory for primary diagnosis, a biopsy for culture and sensitivity testing should always be obtained when it is available and when endoscopy is undertaken as part of the patient's management. Among the non-invasive tests, the place of serology remains questionable for other than epidemiological purposes. How is H. pylori infection best diagnosed? How many tests are needed in routine clinical practice? The answer will depends on the clinical setting and local availability of the tests. For primary diagnosis in dyspeptic patients--where endoscopy is an important tool--a biopsy-based detection system is appropriate an we recommend the use of at least two diagnostic tests based on different principles, like RUT (with 1 or 2 biopsy specimen/test) and histology (including antrum & corpus biopsies) which are widely available. Alternatively a urea breath test may also be recommended when endoscopy is not required. Post-treatment monitoring seems to be justified in most cases and must always be performed at least 4-6 weeks after completion of therapy. The urea breath test is probably the method of choice for non-invasive testing in this clinical setting. When endoscopy is required, multiple biopsy specimens both from the antrum and the corpus and the use of at least two different diagnostic methods must be performed. Whenever possible, culture should always be done as it is very specific and allows testing of antimicrobial susceptibility which is mandatory in case of treatment failures. Neither the "Test and Treat" nor the "Test and Scope" strategies have been investigated in terms of effectiveness of symptoms relief and cost in Belgium and cannot therefore be recommended at this time.