The association between red blood cell distribution width-to-albumin ratio and Helicobacter pylori seropositivity in U.S. adults: an observational study from NHANES 1999–2000 with external validation

ObjectivesHelicobacter pylori (Hp) seropositivity, linked to inflammatory morbidity, may reflect host inflammation-nutrition imbalance. The red blood cell distribution width-to-albumin ratio (RAR) has been proposed as a composite marker reflecting this axis. Using NHANES data and an external validation cohort of 250 hospital patients, we aimed to examine the association between RAR and H. pylori seropositivity.MethodsThis study included U.S. adults from NHANES 1999–2000 and an external validation cohort of 250 hospital patients. Survey-weighted logistic regression assessed the association between RAR and H. pylori seropositivity. Restricted cubic splines (RCS) evaluated non-linearity. In addition, we performed a series of analyses, including restricting the range of RAR to limit inter-queue overlap, excluding subjects with clinical anemia, conducting sensitivity analyses by adjusting spline function parameters and handling extreme values, as well as subgroup analyses.ResultsIn the main adult dataset of NHANES 1999–2000, RCS regression analysis revealed a significant nonlinear association between RAR and Hp seropositivity rate (overall test p < 0.001; nonlinear test p < 0.001), with the inflection point at RAR = 2.85. Below this point, higher RAR was associated with greater odds of seropositivity (OR = 3.59; 95% CI: 1.76–7.31); above it, higher RAR related to lower odds (OR = 0.73; 95% CI: 0.55–0.97). RCS analysis of the external validation cohort showed a linear RAR-Hp seropositivity association (overall p = 0.457; nonlinear p = 0.937). Most RAR values clustered above the inflection point (2.85), potentially reducing statistical power. Restricting both cohorts to RAR ≥ 2.85 yielded consistent linear downward trends, with no significant nonlinearity in either cohort, indicating the validation cohort reflects post-inflection local relationships. In categorical analyses, the third RAR quartile remained positively associated after full adjustment, whereas the highest quartile did not, consistent with nonlinearity. Subgroup analyses showed no significant associations overall, though potential heterogeneity was observed by hypertension (p for interaction = 0.040) and age (p for interaction = 0.046).ConclusionA statistically significant non-linear association pattern was observed between RAR and the seropositivity rate of Helicobacter pylori in NHANES; however, this association was sensitive to exposure distribution and could be partially explained by hematological factors. The findings were not replicated in the external cohort, and the overall evidence does not support a robust or clinically applicable association. RAR may reflect the inflammation-nutrition microenvironment associated with Helicobacter pylori exposure, rather than functioning as an independent predictive biomarker. Prospective studies are warranted to elucidate the causal relationship and clinical implications.