Incidence of Gastric Cancer and Neuroendocrine Tumors in Autoimmune Gastritis: A Systematic Review and Meta-Analysis of Follow-Up Studies.

INTRODUCTION

A central controversy in autoimmune gastritis (AIG) concerns the true incidence of gastric cancer (GC) and gastric neuroendocrine tumors (gNETs). Reported risks vary widely, potentially influenced by Helicobacter pylori (HP) status, AIG subtypes, and age-factors not addressed in prior meta-analyses. This study aimed to determine the pooled incidence of GC and gNETs in AIG and to conduct a comprehensive stratified evaluation.

METHODS

We searched PubMed, Embase, and Cochrane Library through May 2025 for observational studies reporting incident GC or gNETs in AIG. Subgroup analyses, sensitivity analyses, and meta-regression were conducted.

RESULTS

Twenty-three studies were included: 19 hospital-based cohorts (n = 4309) and 4 registries (n = 36,015). The pooled incidence of GC was 0.12/100 person-years (PY; 95% CI, 0.06-0.17; I2 = 45.7%) and of gNETs 1.09/100 PY (95% CI, 0.61-1.58; I2 = 81.3%). By subtype, GC incidence was 0.29/100 PY in pernicious anemia (PA), 0.12 in histopathology-defined AIG, and 0.09 in parietal cell antibody (PCA)-positive patients; gNETs were highest in the hypergastrinemia subgroup (1.74/100 PY). By HP status, GC incidence was 0.12/100 PY in HP-naïve and 0.07 in HP-exposed patients. Age-stratified analyses showed a consistent upward trend in patients ≥ 60 years for both GC (0.27 vs. 0.10/100 PY) and gNETs (1.48 vs. 0.87/100 PY).

CONCLUSION

This study underscores the neoplastic risk associated with AIG. GC incidence is highest in PA and lower in PCA-positive patients, likely reflecting different AIG stages. HP-naïve status does not confer protection, and long-term GC risk persists. Age is a key modifier, providing a useful reference for risk stratification and individualized surveillance.

TRIAL REGISTRATION: PROSPERO database (CRD420251047873).