Helicobacter pylori infection and insulin resistance in a screening population: frequentist and Bayesian analyses

Le résumé fourni par la source

BACKGROUND: Helicobacter pylori (H. pylori) infection has been associated with insulin resistance and metabolic syndrome, particularly in East Asian populations. Whether these associations are independent or explained by socioeconomic and lifestyle confounding remains unclear. We examined this relationship in a large European cohort using rigorous confounder adjustment and Bayesian sensitivity analyses. METHODS: This cross-sectional study included 4,681 asymptomatic adults (2007-2020), with H. pylori status determined by gastric biopsy. The primary outcome was insulin resistance (HOMA-IR > 2.5); secondary outcomes included HOMA2 model parameters and metabolic syndrome (ATP III criteria). Poisson regression estimated incidence rate ratios (IRR) with sequential adjustment for age, sex, education, alcohol, smoking, and diet. Bayesian analyses quantified the probability of clinically meaningful effects. RESULTS: Of 4,681 participants (median age 57; 52% male), 868 (18.5%) were H. pylori-positive. Infected individuals had lower education, higher BMI, and higher alcohol consumption. Crude insulin resistance prevalence was higher in H. pylori-positive participants (31% vs. 27%; P = 0.015). In unadjusted regression, H. pylori was associated with insulin resistance (IRR 1.15, 95% CI 1.01-1.32), but this association disappeared after full adjustment (IRR 1.12, 95% CI 0.95-1.31; P = 0.176). No association with metabolic syndrome was found (IRR 1.03; P = 0.681). Bayesian analysis showed an 87.9% probability of practical equivalence, arguing strongly against a clinically meaningful independent effect. CONCLUSIONS: The association between H. pylori and insulin resistance is fully explained by socioeconomic and lifestyle confounding. H. pylori acts as a marker of adverse metabolic risk rather than a causal factor, and eradication cannot be recommended for metabolic syndrome prevention.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.