OBJECTIVE Serum pepsinogens and gastrin have been proposed as markers of gastritis, but have seldom been studied in children. In this study the aim was to identify host- and Helicobacter pylori-related factors linked to variations in serum gastrin, PGI, PGII, and to evaluate the potential of these biomarkers for diagnosing gastritis, whether H. pylori-associated or not. METHODS Ninety-two dyspeptic children referred for endoscopy (peptic ulcer exclusion) were included in the study. H. pylori status (urease, culture, histology) was assessed, and genotype determined (PCR) in H. pylori-positive subjects. Serum gastrin, PGI and PGII levels were measured by standard radioimmunoassay (RIA). RESULTS PGI and PGII levels were significantly higher in H. pylori-positive subjects (p=0.007; p=0.012, respectively). Gastrin levels were significantly higher in H. pylori-negative subjects (p=0.035). PGI and PGII were associated significantly with higher antrum inflammation scores (p=0.002; p=0.016, respectively); only PGI was associated with age, after controlling for inflammation (p=0.033) and for activity (p=0.037). The contribution of virulence factors could not be assessed owing to the low number of virulent strains. After multivariate analysis, only antrum inflammation was independently associated with PGI level (p=0.012). Receiver operating characteristic (ROC) analysis showed a low PGI and PGII discriminant power for predicting antrum inflammation. CONCLUSIONS Pepsinogen levels as measured in this study seem predominantly to reflect antral inflammation, but they are not an effective screening test for gastritis (H. pylori-positive or -negative) in dyspeptic children.