Between 1967 and 1976, 1,525 Slovenian patients with a histological diagnosis of intestinal metaplasia (IM) were classified according to subtype of IM based on morphology and mucin staining; 518 cases were diagnosed with type I, 197 with type II and 275 with type III, but in 291 the diagnosis of IM was not confirmed. Patients who had developed cancer or died up to 1986 were identified by record linkage at the Slovenia Cancer Registry and the Central Population Registry in Slovenia. A total of 34 incident cases of gastric cancer occurring at least 6 months after the diagnosis of IM were identified. The standardised incidence ratio (SIR) for stomach cancer was 2.23 in the whole cohort. It was highest for IM type III, followed by type II and IM-unconfirmed, but not increased for type I. The relative risk (RR) of developing gastric cancer based on Cox's proportional hazards model was 2.14 for type II and 4.58 for type III, compared with type I. The RR was especially increased for a subgroup of type III secreting sulphomucins in their goblet cells in comparison with types I-II negative to sulphomucins. Our results confirm that subtyping of IM is useful for identifying individuals at high risk for gastric cancer.

Interobserver agreement in the application of the Sydney system to reversibility of gastritis after H. pylori was good. More strict criteria should be used for atrophy and to differentiate normal and mild chronic inflammation.

No abstract