Epidemiologic findings on serologically defined chronic atrophic gastritis strongly depend on the choice of the cutoff-value

Chronic atrophic gastritis (CAG), a precursor of intestinal gastric cancer, is mostly ascertained noninvasively by serum pepsinogens in epidemiologic studies. However, serological definitions vary widely. We aimed to investigate the impact of this variation on estimated prevalence of CAG and its association with its main risk factors, age and Helicobacter pylori infection. Serum pepsinogen I and II and antibodies against H. pylori were measured by ELISA among 9,444 women and men aged 50-74 years in a population-based cohort study in Saarland/Germany. Application of the various definitions resulted in a wide range of prevalence estimates of CAG prevalence (2.1%-8.2%, with an outlier of 18.8% for one particular definition) and its associations with age and H. pylori infection (age adjusted odds ratios, OR, for CagA positive H. pylori infection: 0.98-4.48). Definitions of CAG based on both pepsinogen I and the pepsinogen I/II ratio or on the pepsinogen I/II ratio only revealed much clearer associations with both age and H. pylori infection than definitions of CAG based on pepsinogen I only (ORs for H. pylori infection: 1.45-4.48 and 0.86-1.30, respectively). Epidemiologic findings on CAG lack comparability due to the heterogeneity in serologic definitions of CAG. The association of age and H. pylori infection with CAG may be strongly underestimated in studies in which CAG is defined by pepsinogen I only.