Pathologists are frequently asked to evaluate gastric mucosal biopsy specimens for the presence of Helicobacter pylori infection and for potentially important changes such as intestinal metaplasia. No agreed-on system is both available and prospectively shown to provide reliable estimates of the underlying pathological condition. The Sydney System combined topographical, morphological, and causative information for evaluation of gastric biopsy specimens and provided recommendations regarding biopsy site and number. Both the biopsy sites and number were changed in 1994. Gastric biopsy specimens from patients who had multiple biopsies performed on predetermined sites were examined to compare the original and the revised Sydney Systems for the detection of intestinal metaplasia and H pylori. The diagnosis based on both versions of the Sydney System was then compared with that obtained by evaluating all specimens. Forty-six patients were studied, 20 with H pylori infection and 36 with intestinal metaplasia. Using either version of the Sydney System correctly categorized H pylori infection status in 100%. Both the original and the revised Sydney recommendations seriously underestimated the prevalence of intestinal metaplasia. Intestinal metaplasia was missed in more than 50% of those with confirmed intestinal metaplasia. No set or site of biopsy specimens was found that could reliably exclude the presence of intestinal metaplasia. Current and future studies that use the Sydney System as basis for detecting intestinal metaplasia are not likely to be reliable. Likewise, using the Sydney System to test posttherapy or with time will not accurately reflect the true status of intestinal metaplasia.