• atrophic gastritis
• intestinal metaplasia
• stomach

We read with interest the paper by Walker (Gut 2003;52:1–4). We agree that histology remains the most suitable test for both detecting and assessing reversion of atrophic gastritis. Such a view elicits two basic questions, however: (1) how consistent are pathologists in recognising gastric atrophy? and (2) in Walker’s words, “where to biopsy?” and, we might add, “how extensively to biopsy”, to correctly evaluate any presence/regression of gastric atrophy?

Concerning the first point on the classification of atrophic gastritis, the current literature is largely biased by the inconsistency of the histological criteria used to categorise atrophy.^1,2 To amalgamate the different viewpoints and also test interobserver agreement in atrophy classification/scoring, an international group of pathologists recently published an extensive description of the different phenotypes of gastric atrophy.³ By merging Western and Eastern experiences, the new proposal extensively describes the diagnostic categories that should be adopted to enable acceptable comparisons between clinicopathological studies involving gastric atrophy (both non-metaplastic and metaplastic). The proposed classification also introduces a new diagnostic category (that is, indefinite for atrophy) which suspends any evaluation of atrophy when high grade inflammation—mostly related to Helicobacter pylori infection—interferes with a reliable assessment of the “loss of appropriate glands”.

As for the number and location of biopsies for atrophy assessment, the recommendations of the updated Sydney system⁴ seem a suitable compromise between the excessive pathologists’ demands and the operating limits of routine practice.

The question of “where to biopsy” is more intriguing. No doubt both the oxyntic and antral mucosa need to be tested, but endoscopists too often neglect the recommendation to take an additional angular sample.⁵

We studied 504 consecutive H pylori positive patients who underwent gastroscopy for untreated non-ulcer dyspepsia. In all patients, biopsies were obtained (Pentax, Japan: KW2415S) from: (i) oxyntic mucosa (one biopsy from the lesser curvature 4–6 cm proximal to the angulus and one from the greater curvature 4–8 cm distal from cardia); (ii) antral mucosa (one biopsy each from the greater and lesser curvatures, 3–5 cm proximal to the pyloric ring), and (iii) only one additional biopsy from the incisura angularis. Histological categories included the basic distinction between non-atrophic and atrophic gastritis.³ Two pathologists independently assessed the biopsies with a 93% consistency (Fleiss’ K value=0.91). Table 1 shows the atrophy prevalence according to biopsy location.

In this series, the importance of sampling the incisura angularis is emphasised by the percentage of atrophic gastritis (46%) that would have been missed if sampling had not included the angular mucosa. The NND (number needed to diagnose⁶) values in detecting atrophy calculated for incisura, antrum, and corpus sampling were 2.17, 4.85, and 63.29, respectively.

Because of different pricing policies in different countries, it is difficult to estimate the additional cost of processing and interpreting the extra biopsy sample taken from the incisura angularis. In most countries, however, the cost of the endoscopy procedure is far higher than the cost of histological examination, and the price of an angular biopsy seems amply balanced by the benefit of an appropriate assessment of gastric disease and a suitable evaluation of the patient’s cancer risk.