Evaluating the Predictive Accuracy of the Kyoto Classification for Helicobacter pylori Infection

Aims Helicobacter pylori (HP) is a gram-negative bacterium infecting over 50% of the global population and is associated with several gastrointestinal diseases, including chronic gastritis, gastric/duodenal ulcers, gastric carcinoma, and MALT lymphoma. Despite advances in endoscopy, endoscopic findings are not routinely used to assess infection risk. The Kyoto Classification has emerged as a promising tool for diagnosing HP infection/gastric cancer risk, based on five endoscopic findings: mucosal atrophy, intestinal metaplasia, enlarged gastric folds, nodularity, and diffuse redness. This study assesses the effectiveness of the Kyoto Classification in identifying active HP infection in a Portuguese population, with a focus on its applicability in non-Asian cohorts, where data remains limited.

Methods This retrospective observational study included patients who underwent upper gastrointestinal endoscopy (EGD) with gastric biopsies for HP detection between January and March 2023. Patients who had received antibiotics or proton pump inhibitors within 15 days before the procedure were excluded. Endoscopic characteristics were recorded according to the Kyoto Classification, and HP detection was confirmed through histological examination. Univariate analysis using Pearson's correlation coefficients assessed the association between endoscopic findings and active infection. ROC curve analysis was also performed to evaluate the diagnostic accuracy of the Kyoto Classification [1] [2] [3] [4].

Results A total of 154 patients (mean age 59.96 years, 55.2% male) were included. Active HP infection was identified in 39 patients (25.3%). Primary EGD indications were pre-bariatric surgery assessment (33.1%), anemia (13.6%), dyspepsia (10.4%), and abdominal pain (9.7%). Univariate analysis revealed that nodularity (r=0.439) and gastric fold hypertrophy (r=0.332) were the variables most strongly associated with active HP infection. Mucosal atrophy (r=0.066) and intestinal metaplasia (r=–0.029) had minimal impact in predicting infection. More than half of the patients with a Kyoto score≥2 presented with active HP infection, with scores ranging from 0 to 5. A threshold of 2 points was considered indicative of increased infection probability. However, ROC curve analysis, with an AUC of 0.673 (95% CI: 0.556-0.790), suggested that despite reasonable performance, the Kyoto Classification was not robust enough to be used alone as a diagnostic tool and was more valuable for ruling out infection in cases with low scores.

Conclusions The Kyoto Classification demonstrates moderate predictive accuracy for HP infection in the studied Portuguese population, particularly in excluding infection in cases with low scores. Although a score≥2 correlates with an increased likelihood of infection, the scoring system alone lacks sufficient reliability for definitive diagnostic use. These findings highlight the need for further multicenter studies to validate the Kyoto Classification’s applicability in non-Asian populations, ensuring its broader clinical utility.

Conflicts of Interest

Authors do not have any conflict of interest to disclose.

• References

• 1 Kikuchi S., Kusano C.. et al. 'Usefulness of the Kyoto Classification Score for Prediction of Current Helicobacter pylori Infection.' ResearchGate. 2022
• 2 Toyoshima O.. et al. 'Endoscopic Findings in the Kyoto Classification of Gastritis and Gastric Cancer Risk: A Retrospective Study.'. Gastrointestinal Endoscopy 2022; 97 (5): 957-967
• 3 Yamaguchi Y., Nagata N.. et al. 'The Kyoto Classification of Gastritis and its Association with Gastric Cancer Risk.'. Journal of Gastroenterology 2023; 58: 35-45
• 4 Kanemasa K., Saito Y.. et al. 'Usefulness of the Kyoto Classification Score for Prediction of Current Helicobacter pylori Infection.'. World Journal of Gastroenterology 2023; 28 (43): 6078-6090

Publication History

Article published online:
27 March 2025