We read with great interest the systematic review by Vannella et al. addressing the incidence of gastric cancer (GC) in pernicious anaemia (PA). The meta-analysis concluded that PA is associated with a nearly sevenfold relative risk of GC. While we agree that PA, as a reflection of atrophic gastritis, is a precancerous condition, it is important to separate PA resulting from autoimmune gastritis, from the PA due to other causes, including Helicobacter pylori.
Pooling may obscure risks, especially when a particular outcome arises from different pathogeneses. Most cases of PA associated with gastric atrophy are now considered as an outcome of chronic H. pylori infection, and thus PA may simply reflect the most severe injury (advanced atrophic gastritis spreading from antral to oxyntic mucosa; i.e. OLGA stage III-IV). Histologically, the ‘pure diseases’ are theoretically easy to separate, in that autoimmune gastritis spares the antrum and H. pylori does not.[4, 5]
We recently published a study addressing autoimmune gastritis. Among those with pure autoimmune phenotype (i.e. non-atrophic gastric antrum with corpus-restricted atrophic gastritis), neoplastic epithelial lesions (both intraepithelial or invasive) were never detected. All four cases of incidental epithelial neoplasia arose in extensive gastric atrophy (involving both antral and oxyntic mucosa; i.e. OLGA stages III–IV), and three had active H. pylori infection.
Among the 562 patients considered, 116 had a repeat endoscopy/biopsy at least 24 months after the index endoscopy (mean interval between initial and last procedure: 54 months; in total 522 person-years). Among these 116 patients, 10 were H. pylori positive at the initial endoscopy and one still had H. pylori infection at the last check-up. No cases of neoplastic lesions (intraepithelial or invasive) were detected at the end of follow-up.
By keeping together the available information, we would like to put forward few considerations. PA associated with atrophic gastritis is a constellation of clinico-biological situations, with the most common causes being H. pylori (pure environmental aetiology), autoimmune (primary autoimmunity), combined H. pylori -autoimmune aetiology and H. pylori-driven autoimmune corpus atrophy. The full range of causes is still only partially explored. The paucity of the literature included in Vannella's meta-analysis emphasises the need for prospective studies to examine the relationship and assign responsibility of host-related vs. environmental (H. pylori) oncogenetic factors.
The availability of H. pylori-free generations (as well as the contemporary H. pylori-eradicated populations) will allow epidemiological panorama to clarify the (probably low) oncogenic relation of GC (in contrast with gastric carcinoids) related to autoimmune gastric diseases. Currently, it is thought that the greater the extent of atrophy (antrum and corpus), the higher the atrophy-related GC risk. As such, ‘pure autoimmune’ atrophy should theoretically be restricted only to a fraction of the gastric mucosa (the oxyntic target of the autoimmune attack) and should theoretically reduce the atrophy-associated cancer risk. In addition, H. pylori associated with ongoing inflammation and eradication is expected to result in a reduction in GC risk. The inflammation in ‘pure autoimmune’ gastritis likely burns out as the native (i.e. oxyntic) target is destroyed.
Irrespective of the gastritis aetiology, any staging system of gastric atrophy (including scoring of pseudopyloric metaplasia, as proposed by OLGA) reflects the overall atrophy-associated GC risk. Therefore, gastritis staging could be usefully applied in supporting the clinical strategies for GC secondary prevention.[6, 7]