Gut microbiota-based prediction of clinical response to sublingual immunotherapy in Artemisia pollen-induced allergic rhinitis: A prospective cohort study.

The clinical response rate to sublingual immunotherapy (SLIT) for Artemisia pollen-induced allergic rhinitis was 54.41%.

• 111 out of 204 adults achieved clinical response.
• Clinical response was defined as ≥30% reduction in combined symptom-medication score (CSMS) during the peak pollen season.
• This was a single-center prospective cohort study.
• All participants underwent baseline stool collection before initiating standardized SLIT.

All three prespecified gut microbiota features independently predicted SLIT treatment response in multivariable analysis.

• Butyrate-producing bacteria (Faecalibacterium, Roseburia, Eubacterium rectale group) composite abundance: OR = 1.59, q = 0.006.
• Prevotella-to-Bacteroides (P/B) ratio: OR = 1.43, q = 0.020.
• Shannon diversity index: OR = 1.33, q = 0.046.
• Gut microbiota was analyzed using 16S rRNA sequencing of the V3-V4 region.
• All q-values indicate significance after multiple testing correction.

Adding gut microbiota features to clinical variables (Model B) significantly improved discrimination compared to clinical variables alone (Model A).

• Model B AUC = 0.79 vs. Model A AUC = 0.71.
• ΔAUC = 0.08, P = 0.021.
• Model B showed improved calibration with intercept α = -0.03 and slope β = 0.98.
• Net reclassification improvement for Model B was NRI = 0.36, P = 0.002.
• Decision curve analysis confirmed greater net benefit for Model B across clinically relevant threshold probabilities.

A parsimonious model (Model C) derived via L1 regularization maintained good predictive performance with high sensitivity and specificity.

• Optimism-corrected AUC for Model C = 0.78.
• Sensitivity = 77.48%.
• Specificity = 72.04%.
• Model C was developed using L1 (LASSO) regularization to select a reduced set of predictors.

The study used a prespecified set of microbiota features including Shannon diversity index, butyrate-producing bacteria composite abundance, and the Prevotella-to-Bacteroides ratio.

• Sample size was 204 adults with Artemisia pollen-induced AR.
• Gut microbiota profiling was performed on baseline stool samples collected before SLIT initiation.
• 16S rRNA sequencing targeted the V3-V4 hypervariable region.
• Three prediction models were developed and compared: Model A (clinical only), Model B (clinical plus microbiota), and Model C (parsimonious).

The gut microbiota's role in modulating mucosal immunity and allergic responses had not been previously well-explored as a predictor of SLIT outcomes in Artemisia pollen-induced allergic rhinitis.

• The authors describe predictive value of gut microbiota for SLIT outcomes as 'underexplored' in this condition.
• The study prospectively enrolled adults with Artemisia pollen-induced AR.
• The findings support integration of gut microbiota assessment into pretreatment stratification algorithms for allergen immunotherapy.