Gastric intestinal metaplasia

We read with interest the article by Dinis-Ribeiro et al concerning a follow up model for patients with atrophic gastritis and intestinal metaplasia (IM).¹ The model proposed is based on “a minimum of two endoscopic biopsies”, although 40% of patients had at least three biopsies, and more than 15% of patients had more than four biopsies performed. Within that setting, the authors investigated the significance of type I, II, and III IM, and concluded that those with type I IM should have a less intensive follow up protocol, whereas in those with type III IM, “a hunt for high grade neoplasia should be performed”.

In previous publications, several authors have reported that IM may be detected in biopsies taken exclusively from endoscopically abnormal areas,² whereas others recommended that gastric biopsies should be harvested from pre-established mucosal sites.³ In this respect, the Sydney system³ for grading of gastritis has provided practical guidelines for optimal biopsy sampling of the gastric mucosa. Notwithstanding, using the Sydney system’s recommendations, El-Zimaity and Graham⁴ found that IM was missed in more than 50% of the biopsies from “Sydney” sites in patients with confirmed gastric IM on multiple site sampling. These authors concluded that the minimum number of biopsies needed to identify IM should probably be eight,⁴ and emphasised that current and future studies that use the Sydney system as a basis for detecting gastric IM are not likely to be reliable. Thus, it appears that sampling gastric biopsies from pre-established mucosal sites,² or from endoscopically abnormal areas³ may be both insufficient to calculate IM prevalence figures, and inadequate to estimate the possible risk of gastric IM in long follow up studies.⁴

By assessing IM at low power (×4) examination, we found in histological sections from gastrectomy specimens that IM could be either spotty or extended (encompassing one or more entire low power microscopic field/section).⁵ When extended IM was present in ⩾ 5 histological sections, IM was considered to be widespread. Thus, IM may be spotty, or extended and widely distributed. Against that background, the proposal of Dinis-Ribeiro et al of obtaining two or three biopsies (in most of their series) as representative for the state of the gastric mucosa appears highly unsatisfactory. Moreover, several authors have been unable to corroborate the prognostic significance of incomplete IM type III.^6,7 Kato et al found that incomplete IM usually occurs in the antrum, whereas complete IM is usually seen in the fundus.⁷

The article by Dinis-Ribeiro et al points out that two pathologists reviewed all the slides, and that “agreement was achieved in 85% of the cases. In case of disagreement, a consensus was obtained…”. Because no mention is made regarding double blind examination with coated slides, we assume that the authors did not carry out that procedure. The possibility of intraobserver and of interobserver variations was not explored.

According to Dinis-Ribeiro “…endoscopic examination throughout the entire gastric cavity may still fail to diagnosis dysplasia and cancer”. The authors forgot to include areas with IM, particularly when chromography of the gastric mucosa is not performed. That procedure was not mentioned in the list of “methods” used in their follow up studies.