Dietary live microbe intake and asthma risk and mediator analysis: a cross-sectional study of NHANES 2003-2018.

OBJECTIVE: This study investigated the association between dietary live microbe intake and asthma risk, exploring potential mediating roles of exhaled nitric oxide (NO), C-reactive protein (CRP), and forced vital capacity (FVC). METHODS: Cross-sectional data from NHANES (2003–2018) were analyzed, including 28,025 participants (14.2% asthma prevalence). Dietary intake was categorized into low, medium, and high live microbe levels using 24-hour recalls. Weighted logistic regression, restricted cubic spline analysis, and mediation analyses were performed to assess associations and nonlinear relationships. RESULTS: In the fully adjusted model (Model 3), dietary live microbe intake was not significantly associated with asthma risk (OR = 0.90, 95% CI: 0.78–1.04, p = 0.161). Although inverse and nonlinear associations appeared in crude and partially adjusted models, these patterns did not persist following full covariate adjustment, indicating that the early associations were likely attributable to confounding rather than a true protective effect. A nonlinear relationship emerged: asthma risk initially decreased with medium-high microbe intake but increased at higher levels (nonlinearity p = 0.003). Mediation analysis revealed NO positively mediated the association (11.7% mediation effect), while FVC exerted a negative mediation effect (-11.1%). CRP showed no significant mediation. These findings indicate that although moderate dietary live microbe intake appeared protective in crude analyses, this effect was not significant after full adjustment, and NO and FVC may partially mediate the observed relationships. CONCLUSION: Although restricted cubic spline analysis suggested a U-shaped pattern in unadjusted and partially adjusted models, this nonlinearity did not persist after full covariate adjustment, indicating that the initial inverse associations were largely driven by confounding rather than a biological effect of live microbial intake. The mediation pathways involving NO and FVC are exploratory and not indicative of causal mechanisms.