{"paper_id":"378ce3e2-9f0b-4f61-8d03-33cf83301b62","body_text":"Predictive Value of Baseline Gut Microbiome Characteristics for the Response in the Next Pollen Season After Sublingual Immunotherapy in Artemisia Pollen–Induced Allergic Rhinitis: A Single-Center Prospective Cohort Study. | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Predictive Value of Baseline Gut Microbiome Characteristics for the Response in the Next Pollen Season After Sublingual Immunotherapy in Artemisia Pollen–Induced Allergic Rhinitis: A Single-Center Prospective Cohort Study. Fei Wang1, Jinjin Yang, Liming Bao, Ya nan, Bat Jin This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8234895/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Artemisia pollen allergic rhinitis is a major health burden, with sublingual immunotherapy showing variable effectiveness. This study explores the potential of gut microbiota as a biomarker to better predict treatment outcomes. Current predictive methods, such as IgE levels, skin prick tests, and symptom scales, often fail to accurately predict treatment outcomes. Objective To evaluate the predictive value of baseline gut microbiome features for sublingual immunotherapy response and develop a practical clinical score for patient stratification. Methods A single-center prospective cohort study enrolled 204 participants. Pretreatment stool samples were analyzed using 16S rRNA V3–V4 sequencing to assess Shannon diversity, the proportion of butyrate-producing bacteria, and the Prevotella/Bacteroides ratio. Three models were developed, with Model A based on clinical variables, Model B incorporating microbiome features, and Model C using L1 regularization for feature selection. Model performance was evaluated through AUC (DeLong), calibration intercept α and slope β (Bootstrap), NRI/IDI, and decision curve analysis, with Model C validated internally. Results The median improvement in CSMS over the peak 6-week pollen season was 32.83% (95% CI 28.87–36.61), with a response rate of 54.41% (95% CI 47.56–61.10). In Model B, microbiome features significantly predicted response, with ORs of 1.59 for butyrate-producing bacteria, 1.43 for the Prevotella/Bacteroides ratio, and 1.33 for Shannon diversity. Model B increased AUC from 0.71 to 0.79 (P = 0.021) and showed improved calibration (α=−0.03; β = 0.98). Model C, with a threshold of p ≥ 0.62, had a sensitivity of 77.48%, specificity of 72.04%, and AUC of 0.78. Conclusions Baseline gut microbiome features enhance the prediction of sublingual immunotherapy outcomes. The interpretable, low-dimensional score offers a practical tool for patient stratification and decision-making, with potential for further validation and clinical application. gut microbiota sublingual immunotherapy allergic rhinitis predictive model Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 1 Introduction Allergic rhinitis is a significant health burden, with Artemisia pollen being the primary allergen in northern China. Symptoms worsen during the pollen season, leading to increased airway reactivity( 1 ). Sublingual immunotherapy is regarded as a disease-modifying approach, but shows marked interindividual heterogeneity in response, and reliable determination of the probability of treatment efficacy before treatment is lacking( 2 ). Current decision-making mainly relies on specific IgE, skin prick test intensity and symptom scales, with limited predictive power and inconsistent thresholds( 3 , 4 ). Mucosal tolerance is influenced by the gut microbiome, with short-chain fatty acids like butyrate, microbial diversity, and the Prevotella/Bacteroides ratio playing a key role in suppressing Th2 inflammation( 5 ). Previous studies mostly suggested associations, but were mainly cross-sectional or post-treatment analyses, allergens and regimens were heterogeneous, and statistical evaluations focused on associations rather than predictive evidence with clinical utility( 6 ). In clinical practice, prospective data targeting sublingual immunotherapy with Artemisia pollen are still lacking to demonstrate that a single baseline stool measurement can predict response in the next pollen season before treatment and provide quantifiable increment and actual net benefit on top of conventional clinical variables ( 7 ). Interpretable and low-cost predictive tools are also scarce, and existing models rarely evaluate discrimination, calibration, reclassification and decision utility simultaneously within a unified outcome framework, with insufficient robustness testing against antibiotic exposure, technical batches and threshold setting( 8 ). How to incorporate pre-specified, reproducible microbiome features into clinical models under conditions with controllable confounding, and to use them in an operational way to serve treatment initiation and persistence, allocation of follow-up intensity and doctor–patient communication, still requires systematic validation. This study, based on a single-center prospective cohort and focusing on patients with Artemisia pollen–induced allergic rhinitis receiving sublingual immunotherapy, used a single baseline stool 16S rRNA sequencing and standardized clinical data to verify the independent predictive value of three prespecified gut features for next-year clinical response and to examine their incremental value in clinical models, and then developed and internally validated a low-dimensional, interpretable GM-SLIT score, evaluating clinical feasibility in terms of discrimination, calibration, reclassification and decision curves. The results can be summarized into two aspects of novelty: baseline gut features can achieve usable probability stratification before treatment; and they provide meaningful decision benefits at clinically relevant thresholds. 2 Materials and methods 2.1 Study design and participants This study employs a single-center, prospective cohort design conducted at a specialized outpatient clinic offering standardized allergen immunotherapy. Participants complete baseline assessments and stool collection between November 1, 2023, and March 31, 2024, and initiate Artemisia pollen sublingual immunotherapy (SLIT) within two weeks of completing the baseline assessment. Outcomes are assessed during the Artemisia pollen season, defined as occurring between August 1 and October 15, 2024. The study protocol receives approval from the institutional ethics committee (approval No. 2021-033), and all participants provide written informed consent before any study procedures. 2.1.1 Inclusion and exclusion criteria Inclusion criteria were: aged 18–60 years; seasonal allergic rhinitis meeting the ARIA diagnostic criteria; serum Artemisia pollen–specific IgE ≥ 0.35 kUA/L or positive skin prick test to Artemisia pollen (wheal diameter ≥ 3 mm with valid controls); planned to receive single Artemisia pollen SLIT and able to provide a baseline stool sample before initiation; having completed electronic diary recording in the previous natural pollen season with diary completeness ≥ 80% during the entire peak 6 weeks. Exclusion criteria were: previous Artemisia pollen–specific immunotherapy; systemic glucocorticoid therapy or use of immunosuppressive drugs within the past 4 weeks; pregnancy or lactation; moderate-to-severe persistent asthma requiring maintenance with inhaled or oral corticosteroids; inflammatory bowel disease, celiac disease, short bowel syndrome or major gastrointestinal surgery within the past 6 months; other conditions judged by the investigators to affect adherence or safety. 2.1.2 Participant flow and target sample size Participants were consecutively recruited from the outpatient clinic in the International Mongolian Hospital of Inner Mongolia. Following eligibility screening and informed consent, baseline data and stool samples were collected, and participants then began SLIT with Artemisia pollen. The target sample size was calculated based on the number of events needed for predictive model development. The joint model, which included five clinical variables and three microbiome features, aimed to estimate eight parameters. With an expected response rate of approximately 50%, at least 12.5 events per parameter were required, resulting in a minimum of 100 events. Consequently, the total sample size was set to ≥ 200 participants. To account for a 10% loss to follow-up or non-evaluable outcomes, 220 participants were ultimately enrolled. 2.2 Intervention protocol and follow-up All participants received standardized Artemisia pollen sublingual drops from the same manufacturer. The treatment began with a 14-day build-up phase, during which the dose was gradually increased to reach the maintenance dose. Participants continued the maintenance phase with a fixed once-daily dose until the end of the pollen season( 9 ). In case of local or systemic reactions, allergists adjusted the dose based on severity. Adherence was monitored through electronic diary check-ins, bottle weight measurements, and medication refill records, with adherence defined as the proportion of actual medication days ≥ 80% of planned maintenance days. Follow-up visits included: baseline, end of the build-up phase, pre-pollen season, mid-pollen season (via phone or outpatient visit), and end of the pollen season. Any medication interruptions or switches were recorded, along with the corresponding dates and reasons. 2.2.1 Definition of the pollen season and observation window The pollen season interval was defined using daily Artemisia pollen counts from the local aerobiological monitoring station. The season began when the 5-day moving average of pollen count first reached 20 grains/m³ and ended when the 5-day moving average remained below 20 grains/m³ for seven consecutive days ( 10 ). The consecutive 6 weeks with the highest moving average were designated as the peak observation period, during which all primary outcomes were calculated based on diary data. 2.3 Outcome measures and definition of response The primary outcome was the binary variable of “clinical response in the next pollen season”. Seasonal-level indices were calculated using the daily combined symptom-medication score (CSMS) recommended by EAACI ( 11 ): the daily Rhinoconjunctivitis Total Symptom Score (RTSS) recorded 6 symptoms, each scored 0–3, with a total score of 0–18; the daily rescue medication score (RMS) was graded 0–3 (0 for no medication; 1 for use of oral H1 antihistamines or antiallergic eye drops ( 12 ); 2 for use of nasal corticosteroids without systemic corticosteroids; 3 for use of systemic corticosteroids on that day); the daily CSMS = RTSS/6 + RMS, with a range of 0–6. The mean CSMS during the peak 6 weeks of the previous natural pollen season was used as the baseline seasonal value, and the mean CSMS during the peak 6 weeks of the next pollen season was used as the follow-up value ( 13 ). The relative reduction proportion was defined as (baseline-season CSMS − follow-up-season CSMS)/baseline-season CSMS. A relative reduction ≥ 30% was prespecified as response. Secondary outcomes included the continuous value of CSMS in the follow-up season, change in RQLQ total score and change in the number of days of medication use, for descriptive supplementation. 2.3.1 Collection and grading of adverse events All SLIT-related adverse events were recorded in real time during the build-up and maintenance phases through outpatient visits and electronic diaries. Local and systemic reactions were classified using the World Allergy Organization (WAO) grading criteria, with details on the most severe grade, start and end times, and management measures. Serious adverse events were reported to the ethics committee and safety supervisor within 24 hours. 2.4 Stool sample collection and 16S rRNA sequencing Baseline stool samples were collected at home within 14 days before SLIT initiation, using sampling tubes with stabilizer. Samples were thoroughly mixed on-site, kept at room temperature for no more than 24 hours, and stored at − 80°C within 4 hours of arrival at the laboratory. DNA was extracted using a standardized kit that included a mechanical disruption step, with blank sampling tubes and extraction negative controls included in each batch. Microbial quality control materials with known composition were added to monitor batch stability. The V3–V4 region of the 16S rRNA gene was amplified using the 341F (5′-CCTACGGGNGGCWGCAG-3′) and 806R (5′-GACTACHVGGGTATCTAATCC-3′) primers for paired-end PCR. Libraries were sequenced using 2×250 bp paired-end sequencing on the Illumina MiSeq platform, with each batch including negative controls and quality control bacterial communities. The target sequencing depth was ≥ 25,000 reads/sample; samples not meeting this standard underwent repeat extraction and amplification. 2.5 Microbial data processing and prespecified features Raw sequences were processed in QIIME 2 (version 2023.5) for adapter trimming, quality control, and denoising. Amplicon sequence variants (ASVs) were generated using DADA2, with chimera removal. Taxonomic annotation was performed using the SILVA 138 reference database, with a confidence threshold of 0.8 for annotation from phylum to genus level. Contamination assessment was conducted using the prevalence method in decontam to differentiate negative controls and low-biomass samples, and suspected contaminant ASVs were removed. The rarefaction depth for diversity analysis was set to 10,000 reads/sample. For regression modeling, community composition data were transformed using centered log-ratio (CLR) after adding a pseudocount of 0.5 to the relative abundance matrix to mitigate compositional bias. The prespecified microbiome features were limited to three and locked in the analysis plan: the Shannon index; the composite proportion of butyrate-producing bacteria, defined as the sum of the relative abundances of Faecalibacterium, Roseburia and the Eubacterium rectale group; and the Prevotella/Bacteroides ratio, calculated as log10 (Prevotella relative abundance + 0.5) − log10 (Bacteroides relative abundance + 0.5)( 14 ). All features were standardized as z-scores before modeling. 2.6 Clinical variables and potential confounders At baseline, demographic information (age, sex, body mass index, duration of rhinitis) and lifestyle (smoking status) were recorded. Allergic indices included Artemisia-specific IgE concentration (entered into the model after log10 transformation) and whether there was concomitant sensitization to other pollens (specific IgE to grass or tree pollens ≥ 0.35 kUA/L was defined as the binary variable of polysensitization). Baseline disease severity was described using the RQLQ total score and the previous season CSMS as continuous variables; only the RQLQ total score was retained in the modeling as the representative of severity. Use of antibiotics, probiotics and proton pump inhibitors in the past 3 months was recorded by questionnaires and cross-checked with the prescription system for sensitivity analyses. Technical batch variables (DNA extraction batch and sequencing run) were used for quality control and sensitivity analyses. 2.7 Statistical analysis All analyses were conducted using R 4.3.2. Continuous variables were reported as means with standard deviations or medians with interquartile ranges, while categorical variables were presented as frequencies and percentages. A two-sided significance level of α = 0.05 was used. Before modeling, independent variables were centered and standardized, and outliers were assessed using Cook’s distance and leverage, with Winsorization applied when necessary. Missing covariates were addressed using multiple imputation (MICE, 20 imputed datasets), with chained equations including all predictors and the outcome variable; missing outcomes were not imputed, and the main analysis was performed on the evaluable population. All regression coefficients, predictive performance measures, and reclassification indices were estimated based on pooled results after imputation, combined according to Rubin’s rules. 2.7.1 Main analysis: assessment of independent predictive value A multivariable Logistic regression model was constructed with “clinical response in the next pollen season” as the dependent variable. Clinical Model A included five prespecified clinical variables: age, sex, RQLQ total score, Artemisia-specific IgE (log10) and polysensitization (binary). Combined Model B added three microbiome features to Model A (Shannon index, composite proportion of butyrate-producing bacteria, Prevotella/Bacteroides ratio). Adjusted odds ratios and 95% confidence intervals were reported for each predictor, and variance inflation factors were used to assess collinearity in Combined Model B. 2.7.2 Incremental predictive value and model performance Discrimination was evaluated using the area under the receiver operating characteristic curve (AUC), and differences in AUC between Model A and Model B were compared using the DeLong method; the Brier score, calibration intercept and calibration slope were also reported, and a spline-smoothed calibration curve was plotted. Risk reclassification was assessed using the category-free net reclassification improvement (NRI) and the integrated discrimination improvement (IDI), with the threshold probability range prespecified as 0.2–0.8. Decision curve analysis was used to compare net benefit at different threshold probabilities. 2.7.3 Construction of the predictive tool and internal validation On the basis of Model B, L1 regularization was used for feature selection, retaining at most 5 variables to form the parsimonious Model C. A point-based score and a nomogram were constructed according to the relative magnitudes of the regression coefficients, and a threshold was provided to map predicted probabilities to risk strata (high probability and low probability). Internal validation used 1,000-bootstrap resampling to estimate optimism and perform correction, and the optimism-corrected AUC, Brier score and calibration indices were reported. For the final score, the sensitivity, specificity, positive predictive value and negative predictive value at the classification threshold were provided. 2.7.4 Sensitivity analyses and control of multiple comparisons Three prespecified sensitivity analyses were performed: excluding participants who had used antibiotics within 30 days before baseline; resetting the response threshold to 25% and 35% to recompute model stability; and adding technical batch indicator variables into the regression to examine the impact of batch effects. For multiple testing of the three microbiome features, the Benjamini–Hochberg method was used to control the false discovery rate at 5%, and no additional correction was applied for overall model comparisons. 2.8 Data quality control and bias management Electronic data were collected using REDCap, with field validity checks and logical verification rules specified. Microbiological experiments were performed in a batch-wise standardized workflow, and negative and positive quality control materials were used to monitor contamination and drift; within- and between-batch sequencing consistency was assessed using Pearson correlation and Bray–Curtis distance. Stool samples and sequence identifiers were transferred in a double-blind manner between the statistical and clinical teams, and statisticians were unaware of clinical outcomes before database lock. To reduce information bias, symptom and medication diaries used standardized templates and daily push reminders; missing records were verified by telephone and the reasons were documented. Medication use records were verified by comparison with outpatient pharmacy data to improve the accuracy of medication scores. 3 Results 3.1 Study participants and baseline characteristics A total of 356 patients were screened, and 204 (57.3%) met the inclusion criteria and were included in the analysis set. The main reasons for exclusion were not meeting inclusion criteria (31.5%), not completing stool collection or quality control (4.9%) and not completing treatment follow-up (11.5%) (Fig. 1 ). Continuous variables were compared using independent-samples t tests, and categorical variables were compared using Pearson χ² tests. Standardized mean differences (SMD) were ≤ 0.21 (considered mild), and the tests did not yield statistically significant results (all P > 0.05) (Table 1 ). Table 1 Baseline clinical characteristics of participants (stratified by response status in the next pollen season) Variable Overall (n = 204) Responders (n = 111) Non-responders (n = 93) |SMD| Test statistic P Age (years) 32.80 ± 9.63 31.94 ± 9.36 33.82 ± 9.90 0.20 t = − 1.385 0.168 Sex (male/female) 96 (47.06%); 108 (52.94%) 49 (44.14%); 62 (55.86%) Male 47 (50.54%); Female 46 (49.46%) 0.13 χ² = 0.830 0.362 Body mass index (kg/m²) 23.84 ± 3.16 23.56 ± 3.10 24.18 ± 3.22 0.20 t = − 1.393 0.165 Duration of rhinitis (years) 7.41 ± 4.15 7.02 ± 3.97 7.88 ± 4.33 0.21 t = − 1.467 0.144 Baseline-season CSMS (daily mean) 2.36 ± 0.62 2.31 ± 0.60 2.41 ± 0.65 0.16 t = − 1.133 0.259 RQLQ total score 3.27 ± 1.08 3.19 ± 1.04 3.37 ± 1.12 0.17 t = − 1.181 0.239 Artemisia pollen–specific IgE (log10, kUA/L) 0.96 ± 0.48 0.93 ± 0.46 1.00 ± 0.50 0.15 t = − 1.033 0.303 Polysensitization (yes/no) 91 (44.61%); 113 (55.39%) 46 (41.44%); 65 (58.56%) 45 (48.39%); 48 (51.61%) 0.14 χ² = 0.988 0.32 Comorbid asthma (yes/no) 36 (17.65%); 168 (82.35%) 17 (15.32%); 94 (84.68%) 19 (20.43%); 74 (79.57%) 0.13 χ² = 0.911 0.34 Smoking status (never/former/current) 137 (67.16%); 37 (18.14%); 30 (14.71%) 77 (69.37%); 18 (16.22%); 16 (14.41%) 60 (64.52%); 19 (20.43%); 14 (15.05%) 0.11 χ² = 0.687 0.709 Interval from baseline stool collection to SLIT initiation (days) 6.23 ± 2.75 6.07 ± 2.81 6.42 ± 2.69 0.13 t = − 0.907 0.366 Notes: Responder definition: during the next pollen season, the 6-week peak-period mean CSMS decreases by ≥ 30% vs. the baseline season; CSMS calculation: CSMS = RTSS/6 + RMS; reported as the daily mean within the season; Specific IgE values are presented after log10 transformation. 3.2 Data completeness and SLIT exposure/adherence Continuous variables were compared using independent-samples t tests, and categorical variables were compared using Pearson χ² tests. The results showed that diary completeness and medication adherence were high in each group and the differences were not significant (all P > 0.05), suggesting good adherence to SLIT treatment and sufficient data completeness (Table 2 ). Table 2 Overview of electronic diary completeness during the pollen season and SLIT exposure/adherence Measure Overall (n = 204) Responders (n = 111) Non-responders (n = 93) Test statistic P value Valid diary days (days) 39.08 ± 3.12 39.34 ± 2.96 38.73 ± 3.31 t = 1.375 0.079 Diary completeness (%) 93.01 ± 7.43 93.67 ± 7.06 92.22 ± 7.88 t = 1.372 0.103 Planned maintenance medication days (days) 243.83 ± 30.37 245.09 ± 29.08 242.32 ± 31.79 t = 0.644 0.486 Actual days on medication (days) 228.93 ± 38.41 231.42 ± 36.87 225.95 ± 39.89 t = 1.010 0.274 Medication adherence (%) 93.90 ± 8.34 94.47 ± 7.94 93.31 ± 8.57 t = 0.996 0.323 Total SLIT duration (days) 257.84 ± 31.72 259.71 ± 30.06 255.60 ± 33.51 t = 0.914 0.318 AE-related interruption or dose reduction (yes/no, n, %) 29 (14.22%); 175 (85.78%) 14 (12.61%); 97 (87.39%) 15 (16.13%); 78 (83.87%) χ² = 0.513 0.448 Note: Adherence was defined as the proportion of actual medication days ≥ 80% of planned maintenance medication days. 3.3 16S sequencing and quality control The method adopted 16S rRNA V3–V4 sequencing, with a rarefaction depth of 10,000 reads. Panel A used the Hodges–Lehmann method to evaluate between-group differences in sequencing depth, and Panel B plotted LOESS rarefaction curves. The results showed that sequencing depths were similar between the two groups (HL = 779 reads, 95% CI − 428 to 1,962), with no obvious systematic bias (Fig. 2 ). 3.4 Primary outcome: CSMS change and response rate Using the peak 6-week diary, the relative change in CSMS was right-skewed; the median improvement was 32.83% (95% CI 28.87–36.61). With the ≥ 30% threshold, the response rate was 54.41% (95% CI 47.56–61.10). The kernel density peak was close to the 30% threshold (Fig. 3 ). 3.5 Independent predictive value of baseline gut microbiota In Model B, the three prespecified microbiome features were independently associated with response (all corrected by BH-FDR), with OR = 1.59 per + 1 SD for the composite proportion of butyrate-producing bacteria, OR = 1.43 for the Prevotella/Bacteroides ratio and OR = 1.33 for the Shannon index (all q < 0.05); the other clinical covariates did not reach statistical significance (Fig. 4 ). 3.6 Incremental predictive value of microbiome features Model discrimination was evaluated using ROC curves (DeLong test to compare AUC between Model A and Model B), and calibration was assessed by Bootstrap analysis. After adding microbiome features, the AUC of Model B increased from 0.71 for Model A to 0.79, and the difference in AUC between the two models was statistically significant (ΔAUC = 0.08, P = 0.021). Model B showed better calibration (α=−0.03; β = 0.98), closer to the ideal values (Fig. 5 A, B). The category-free NRI and IDI were used to evaluate the incremental predictive ability of Model B (Bootstrap test). After adding microbiome features, both NRI_total and IDI increased significantly (both P < 0.01), indicating that the predictive performance of Model B was clearly improved compared with Model A (Table 3 ). Decision curve analysis evaluated net benefit (Bootstrap 1000 times). Within the range of pₜ=0.10–0.80, the overall net benefit of Model B was higher; at pₜ=0.30 and 0.50, ΔNB was 0.04 and 0.03, respectively, both superior to Model A (Fig. 5 C). Table 3 Category-free NRI and IDI of Model B compared with Model A Metric Value 95%CI P_Boot NRI_events 0.27 0.10–0.46 0.006 NRI_non-events 0.09 −0.03–0.22 0.146 NRI_total(=events + non-events) 0.36 0.18–0.54 0.002 IDI 0.05 0.02–0.09 0.004 Note: Category-free NRI and IDI were calculated according to the Pencina method. 3.7 Parsimonious predictive tool and internal validation Model C retained 5 variables after L1 regularization. A point-based score was constructed by linearly transforming the regression coefficients (Butyrate-producer composite = 100 points as the reference; P/B ratio = 71.29 points; Shannon index = 58.37 points; RQLQ total and Artemisia sIgE log10 contributed negative points), and the total score T was mapped to the predicted probability p; p ≥ 0.62 was defined as high response probability. For the example individual, the total score (T = 158.21) corresponded to a predicted probability of about p = 69.83%, indicating that this subject belonged to the high-probability response group (Fig. 6 A, B). Based on the 2×2 table, Se, Sp, PPV, NPV and F1 were calculated, with 95% CI obtained by Bootstrap, and the threshold was determined by the Youden J. With stratification at p ≥ 0.62, the response rate in the high-probability response group was 76.79%, which was markedly higher than 27.17% in the low-probability response group; the classification performance of Model C was: Se 77.48%, Sp 72.04%, PPV 76.79%, NPV 72.83%, F1 77.13% (Table 4 ). Harrell method Bootstrap (1000) was used for optimism correction. Internal validation showed AUC_optim = 0.78; Brier = 0.187; calibration intercept α=−0.01 and calibration slope β = 0.97 (Table 5 ). Table 4 Risk stratification and classification performance of the GM-SLIT score Metric Result (95% CI) Threshold (predicted probability) p ≥ 0.62 (Youden J = 0.495) High-probability response group (n, %) 112 (54.90%); observed response rate 76.79% (68.16–83.64) Low-probability response group (n, %) 92 (45.10%); observed response rate 27.17% (19.14–37.04) Confusion matrix (TP, FP, FN, TN) [TP = 86, FP = 26; FN = 25, TN = 67] Sensitivity (Se) 77.48% (68.86–84.25) Specificity (Sp) 72.04% (62.19–80.15) Positive predictive value (PPV) 76.79% (68.16–83.64) Negative predictive value (NPV) 72.83% (62.96–80.86) F1-score 77.13% (69.74–83.02) Note: The threshold was determined by maximizing the Youden index (J = 0.495 in this table). Sample size N = 204, responders = 111; stratification and performance all refer to Model C. Table 5 Performance of Model C after internal validation (optimism-corrected) Metric Value 95% CI Optimism-corrected AUC (AUC_optim) 0.78 0.72–0.84 Brier score 0.187 0.172–0.204 Calibration intercept α −0.01 −0.17–0.14 Calibration slope β 0.97 0.83–1.12 3.8 Prespecified sensitivity analyses In the sensitivity analyses, ΔAUC varied only slightly across scenarios (− 0.01 to + 0.01), and Δβ ranged from − 0.06 to + 0.04; at pₜ=0.30 and 0.50, the differences in net benefit were 0.03–0.06 and 0.02–0.04, respectively. The effect directions of the three microbiome features remained consistently stable (≥ 90%) (Table 6 ). Table 6 Summary of robustness of Model B under sensitivity scenarios (95% CI) Metric Exclude antibiotics within 30 days pre-baseline Response threshold = 25% Response threshold = 35% Include batch indicators in regression ΔAUC + 0.01 (− 0.01–0.03) −0.01 (− 0.03–0.01) + 0.01 (− 0.02–0.03) + 0.001 (− 0.018–0.02) Δ calibration slope + 0.03 (− 0.02–0.07) −0.06 (− 0.11–−0.01) + 0.04 (− 0.01–0.08) + 0.01 (− 0.03–0.05) ΔNB_0.30 0.05 (0.02–0.08) 0.03 (0.01–0.05) 0.06 (0.03–0.09) 0.04 (0.01–0.07) ΔNB_0.50 0.03 (0.01–0.06) 0.02 (0.01–0.04) 0.04 (0.02–0.06) 0.03 (0.01–0.05) Direction consistency—Shannon 92.18% (85.64–96.02) 90.84% (83.98–94.98) 92.63% (86.52–96.25) 91.42% (84.83–95.43) Direction consistency—butyrate-producer composite 96.27% (91.65–98.56) 95.53% (90.41–98.20) 96.88% (92.74–98.85) 96.02% (91.18–98.39) Direction consistency—P/B ratio 94.11% (88.32–97.27) 93.02% (86.84–96.65) 94.37% (88.79–97.33) 93.58% (87.92–96.92) Note: NRI and IDI are not repeated in this table; reclassification results are shown in Table 3 ; all indices are based on Model B (clinical + microbiome features). 3.9 Safety Oral local reactions were the most common adverse events (mainly grade 2), and most participants did not require dose adjustment or interruption of treatment; visits or emergency visits due to adverse events were rare, and only 3 cases (1.47%) discontinued treatment because of adverse events (Table 7 ). Table 7 Adverse events related to sublingual immunotherapy (n, %) Event category Cases Most severe WAO grade Any dose adjustment/interruption Emergency/urgent care Oral local reactions 82 (40.20%) 2 22 (26.83%) 0 (0.00%) Gastrointestinal discomfort 21 (10.29%) 2 4 (19.05%) 0 (0.00%) Skin reactions 18 (8.82%) 2 3 (16.67%) 0 (0.00%) Respiratory symptoms 12 (5.88%) 2 4 (33.33%) 1 (8.33%) Systemic reaction (non-shock) 5 (2.45%) 3 4 (80.00%) 1 (20.00%) Anaphylactic reaction 1 (0.49%) 4 1 (100.00%) 1 (100.00%) Discontinuation due to AE 3 (1.47%) — 3 (100.00%) 1 (33.33%) 4 Discussion After adjusting for age, sex, symptom burden, Artemisia-specific IgE, and polysensitization, the three gut features remained independently associated with clinical response. Effect directions were stable after error control, and sequencing quality, diary completeness, and adherence met analysis requirements. Baseline clinical differences were mild, supporting the attribution of associations to microbiome exposure rather than measurement bias. The composite index of butyrate-producing bacteria contributed most significantly. By increasing short-chain fatty acid levels, it may inhibit histone deacetylase activity, induce differentiation of regulatory T cells, and increase IL-10 production. This process may enhance tight junctions of the intestinal mucosal barrier, thereby reducing allergic inflammation along the gut–lung axis( 15 ); an increase in Shannon diversity represents increased metabolic redundancy of the microbial community, making it easier to generate tolerance-related metabolites and limit pro-inflammatory colonization( 16 ); a higher Prevotella/Bacteroides ratio usually means that fermentation pathways of complex carbohydrate substrates predominate, or that the sensitizing drive to allergens is reduced via remodeling of dendritic cell phenotypes( 17 ), together explaining the differences in response probability under similar clinical backgrounds. Existing studies have repeatedly suggested that butyrate-producing bacteria and high diversity are positively associated in allergic diseases and in the remission of immunotherapy( 18 ), but most are cross-sectional or post-treatment analyses, making it difficult to clarify causality versus concomitant changes; other studies have also suggested that the effect of Prevotella is context-dependent, and diet and host background can modulate the direction and magnitude of the effect( 19 ). The present prospective design, with a single allergen and prespecified features as the core, under the premise of controllable confounding and quality assurance, advances the signal from the level of association to the attribute of prediction, providing empirical evidence for using baseline stool measurement for pretreatment precise stratification, and laying the foundation for subsequent incorporation of microecological indices into clinical decision-making. On the basis of the model containing only clinical variables, after introducing the three baseline gut features, the overall predictive performance was markedly improved, with discrimination increasing from moderate to better and probability calibration becoming more ideal; reclassification of individual probabilities was improved, with true responders more easily identified and fewer misclassified as responders. Across clinically acceptable thresholds, net benefit was consistently higher than that of reference strategies, suggesting the practical value of incorporating microbiome data into decision-making( 20 ). Sensitivity analyses showed stable results across scenarios, unaffected by antibiotic exposure, threshold changes, or experimental batches. The incremental value brought by microbiome indices may arise from the gut community providing additional information for mucosal immune homeostasis that is not covered by traditional demographic, allergic and symptom variables( 21 ). Butyrate-producing bacteria, diversity, and Prevotella/Bacteroides-dominated communities impact short-chain fatty acid production, metabolic redundancy, and substrate utilization. These factors affect dendritic cell phenotypes, Treg induction, and mucosal barrier integrity, influencing tolerance to sublingual allergens( 22 ). These pieces of information are related to IgE titers or symptoms but are not redundant, providing complementary signals statistically, making individualized probability estimates closer to the true risk, and achieving a net improvement in the trade-off between overtreatment and undertreatment at intervention thresholds. Existing predictive frameworks usually rely on IgE, skin prick test intensity and symptom scales ( 23 ), with only limited improvement and lacking quantitative demonstration of clinical utility; previous studies of gut microbiota and immunotherapy outcomes have mostly been correlation analyses, with few systematically examined from the perspectives of discrimination, calibration, reclassification and net benefit. This evidence, after control of confounding and resampling-based correction, confirms that microbiome indices can achieve quantifiable predictive improvement without substantially increasing complexity, and present clinical decision value through net benefit evaluation, providing an implementation pathway for screening candidate populations, promoting adherence and optimizing follow-up intensity. The GM-SLIT score, based on L1 regularization, uses five baseline variables to form a nomogram with a single threshold. It effectively stratifies patients before treatment and shows stable classification performance after internal optimism correction. Follow-up records were complete and medication adherence was relatively high, and safety events were mainly mild to moderate, meeting the implementation conditions for deployment in outpatient settings. The weighting of the score is consistent with immunologic directions, with butyrate-producing bacteria and community diversity corresponding to a higher probability of tolerance, specific IgE and symptom burden corresponding to a lower probability of response, and the Prevotella/Bacteroides structural ratio reflecting differences in fermentation substrate utilization and short-chain fatty acid pathways( 24 ), making the model output interpretable and facilitating communication. Implementation of the tool requires only a single baseline stool sample and routine 16S rRNA sequencing, with quality control, compositional transformation and standardization automatically completed in the background, and the interface presenting individualized probabilities and stratification prompts, on which basis decisions can be made about whether to prioritize initiation or persistence of immunotherapy, arrange more intensive follow-up and adherence interventions, and link with pollen-season management. The low-dimensional microbiome-integrated score complements IgE, skin prick tests, and symptom scales, offering better discrimination and calibration( 25 ); compared with models with multiple features and complex algorithms but insufficient interpretability, the present tool achieves transferable clinical utility with lower complexity and uses decision curves to quantify the net improvement in the trade-off between overtreatment and undertreatment, reflecting the application value of microbiome indices in individualized stratification of immunotherapy. This single-center study had a moderate sample size, and external validation is needed. The models were corrected by internal resampling, so their generalizability remains uncertain. Microbiological testing used 16S rather than metagenomics and metabolomics, making it difficult to resolve strains and functional pathways and to directly quantify short-chain fatty acids. Dietary structure, PPIs and antibiotics, and environmental exposures may still produce residual confounding, although sensitivity analyses were performed. Outcomes depended on electronic diaries, with risks of self-report bias and missingness. The next step will be to carry out prospective validation in multicenter populations with different dietary and pollen exposure backgrounds, to unify sampling and bioinformatics pipelines, perform external calibration and model updating, and evaluate re-calibration of thresholds in different medical settings; to conduct metagenomics and targeted metabolomics in representative subsamples, pair measurements of stool and serum short-chain fatty acids, and construct functional pathway mediation models; to improve prospective recording of diet and medication exposures and verify them with prescription and dispensing data; and to carry out decision impact studies, embedding the score into outpatient workflows to examine real-world improvements in adherence, allocation of follow-up resources and patient benefit. 5 Conclusions Baseline gut microbiota characteristics can independently predict clinical response to sublingual immunotherapy in Artemisia pollen-induced allergic rhinitis, with butyrate-producing bacteria, diversity, and the Prevotella/Bacteroides ratio as key indicators. Combining microbiome indices with conventional clinical variables improves predictive model discrimination and calibration, offering net benefits at clinically acceptable thresholds. The low-dimensional, interpretable score based on prespecified features enables patient stratification before treatment and is feasible for outpatient use. It can guide treatment initiation and persistence, optimize follow-up intensity, and improve resource allocation for individualized immunotherapy. External validation across different regions and exposure backgrounds, as well as integration with functional omics, is needed to support clinical translation and identify modifiable pathways. Declarations Conflict of Interest The authors declare no conflicts of interest in relation to this study. Ethical Statement This study was approved by the institutional ethics committee of International Mongolian Hospital of Inner Mongolia (Approval No. 2021-033), and all participants provided written informed consent prior to any study procedures. The research adhered to the ethical principles of the Declaration of Helsinki, ensuring confidentiality and voluntary participation. Consent for Publication Written informed consent was obtained from all participants for the publication of their data. Competing Interests The authors declare that they have no competing interests. Funding This study received no funding. Author Contribution F.W. and J.Y. conceptualized and designed the study. F.W. and L.B. were responsible for the methodology and statistical analysis. Y.N. and B. conducted patient recruitment, clinical assessments, and data collection. F.W. and J.Y. drafted the manuscript and performed critical revisions. All authors contributed to the data interpretation and approved the final manuscript. Data Availability The datasets supporting the conclusions of this article are included within the paper. Further data are available upon reasonable request from the corresponding author. References Zhang Y, Zhang L. Increasing Prevalence of Allergic Rhinitis in China. Allergy Asthma Immunol Res. 2019;11(2):156–69. López JF, Bel Imam M, Satitsuksanoa P, Lems S, Yang M, Hwang YK, et al. Mechanisms and biomarkers of successful allergen-specific immunotherapy. Asia Pac Allergy. 2022;12(4):e45. Pitsios C. Allergen Immunotherapy: Biomarkers and Clinical Outcome Measures. J Asthma Allergy. 2021;14:141–8. Boonpiyathad T, Lao-Araya M, Chiewchalermsri C, Sangkanjanavanich S, Morita H. Allergic Rhinitis: What Do We Know About Allergen-Specific Immunotherapy? Front Allergy. 2021;2:747323. Amoroso C, Perillo F, Strati F, Fantini MC, Caprioli F, Facciotti F. The Role of Gut Microbiota Biomodulators on Mucosal Immunity and Intestinal Inflammation. Cells. 2020;9(5). He Z, Vadali VG, Szabady RL, Zhang W, Norman JM, Roberts B, et al. Increased diversity of gut microbiota during active oral immunotherapy in peanut-allergic adults. Allergy. 2021;76(3):927–30. Zissler UM, Schmidt-Weber CB. Predicting Success of Allergen-Specific Immunotherapy. Front Immunol. 2020;11:1826. Steyerberg EW, Vickers AJ, Cook NR, Gerds T, Gonen M, Obuchowski N, et al. Assessing the performance of prediction models: a framework for traditional and novel measures. Epidemiology. 2010;21(1):128–38. Lou H, Wang X, Wei Q, Zhao C, Xing Z, Zhang Q, et al. Artemisia Annua sublingual immunotherapy for seasonal allergic rhinitis: A multicenter, randomized trial. World Allergy Organ J. 2020;13(9):100458. Pfaar O, Bastl K, Berger U, Buters J, Calderon MA, Clot B, et al. Defining pollen exposure times for clinical trials of allergen immunotherapy for pollen-induced rhinoconjunctivitis - an EAACI position paper. Allergy. 2017;72(5):713–22. Pfaar O, Demoly P, van Gerth R, Bonini S, Bousquet J, Canonica GW, et al. Recommendations for the standardization of clinical outcomes used in allergen immunotherapy trials for allergic rhinoconjunctivitis: an EAACI Position Paper. Allergy. 2014;69(7):854–67. Palathumpattu B, Pieper-Fürst U, Acikel C, Sahin H, Allekotte S, Singh J, et al. Correlation of the combined symptom and medication score with quality of life, symptom severity and symptom control in allergic rhinoconjunctivitis. Clin Transl Allergy. 2022;12(10):e12191. Worm M, Rak S, de Blay F, Malling HJ, Melac M, Cadic V, et al. Sustained efficacy and safety of a 300IR daily dose of a sublingual solution of birch pollen allergen extract in adults with allergic rhinoconjunctivitis: results of a double-blind, placebo-controlled study. Clin Transl Allergy. 2014;4(1):7. Hjorth MF, Christensen L, Larsen TM, Roager HM, Krych L, Kot W, et al. Pretreatment Prevotella-to-Bacteroides ratio and salivary amylase gene copy number as prognostic markers for dietary weight loss. Am J Clin Nutr. 2020;111(5):1079–86. Corrêa RO, Castro PR, Moser R, Ferreira CM, Quesniaux VFJ, Vinolo MAR, et al. Butyrate: Connecting the gut-lung axis to the management of pulmonary disorders. Front Nutr. 2022;9:1011732. Menni C, Zhu J, Le Roy CI, Mompeo O, Young K, Rebholz CM, et al. Serum metabolites reflecting gut microbiome alpha diversity predict type 2 diabetes. Gut Microbes. 2020;11(6):1632–42. Alemao CA, Budden KF, Gomez HM, Rehman SF, Marshall JE, Shukla SD, et al. Impact of diet and the bacterial microbiome on the mucous barrier and immune disorders. Allergy. 2021;76(3):714–34. Di Costanzo M, Carucci L, Berni Canani R, Biasucci G. Gut Microbiome Modulation for Preventing and Treating Pediatric Food Allergies. Int J Mol Sci. 2020;21:15. Chung WSF, Walker AW, Bosscher D, Garcia-Campayo V, Wagner J, Parkhill J, et al. Relative abundance of the Prevotella genus within the human gut microbiota of elderly volunteers determines the inter-individual responses to dietary supplementation with wheat bran arabinoxylan-oligosaccharides. BMC Microbiol. 2020;20(1):283. Vickers AJ, van Calster B, Steyerberg EW. A simple, step-by-step guide to interpreting decision curve analysis. Diagn Progn Res. 2019;3:18. Ding G, Yang X, Li Y, Wang Y, Du Y, Wang M, et al. Gut microbiota regulates gut homeostasis, mucosal immunity and influences immune-related diseases. Mol Cell Biochem. 2025;480(4):1969–81. Luu M, Monning H, Visekruna A. Exploring the Molecular Mechanisms Underlying the Protective Effects of Microbial SCFAs on Intestinal Tolerance and Food Allergy. Front Immunol. 2020;11:1225. Mormile I, Granata F, Detoraki A, Pacella D, Della Casa F, De Rosa F et al. Predictive Response to Immunotherapy Score: A Useful Tool for Identifying Eligible Patients for Allergen Immunotherapy. Biomedicines. 2022;10(5). Di Costanzo M, De Paulis N, Biasucci G, Butyrate. A Link between Early Life Nutrition and Gut Microbiome in the Development of Food Allergy. Life (Basel). 2021;11(5). Xie S, Jiang S, Zhang H, Wang F, Liu Y, She Y, et al. Prediction of sublingual immunotherapy efficacy in allergic rhinitis by serum metabolomics analysis. Int Immunopharmacol. 2021;90:107211. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {\"props\":{\"pageProps\":{\"initialData\":{\"identity\":\"rs-8234895\",\"acceptedTermsAndConditions\":true,\"allowDirectSubmit\":true,\"archivedVersions\":[],\"articleType\":\"Research Article\",\"associatedPublications\":[],\"authors\":[{\"id\":573402076,\"identity\":\"3b583802-48f9-44fa-bcd2-1b851ab74020\",\"order_by\":0,\"name\":\"Fei Wang1\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"International Mongolian Hospital of Inner Mongolia\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Fei\",\"middleName\":\"\",\"lastName\":\"Wang1\",\"suffix\":\"\"},{\"id\":573402077,\"identity\":\"22e44cd0-a76c-42f9-a6dd-b344eb441ea0\",\"order_by\":1,\"name\":\"Jinjin Yang\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"International Mongolian Hospital of Inner Mongolia\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Jinjin\",\"middleName\":\"\",\"lastName\":\"Yang\",\"suffix\":\"\"},{\"id\":573402078,\"identity\":\"cbdb1e23-bdbd-4f7c-87da-512fffe0617c\",\"order_by\":2,\"name\":\"Liming Bao\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"International Mongolian Hospital of Inner Mongolia\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Liming\",\"middleName\":\"\",\"lastName\":\"Bao\",\"suffix\":\"\"},{\"id\":573402079,\"identity\":\"7fdb9d76-3079-4017-8c85-d9ec59c891af\",\"order_by\":3,\"name\":\"Ya nan\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"Inner Mongolia Medical University\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Ya\",\"middleName\":\"\",\"lastName\":\"nan\",\"suffix\":\"\"},{\"id\":573402080,\"identity\":\"a1ae837f-ef2d-413b-aac9-d9a39418ad86\",\"order_by\":4,\"name\":\"Bat Jin\",\"email\":\"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAzUlEQVRIiWNgGAWjYBACAxBOYLCRYTgAZPEQp4UZpCWNhzQtQHCYBC3m7OcPFDzccZ6H70YC44O3bQz58g0EtFj2JDMYJJ65zSN5I4HZcG4bg+WGA4QcdgCkpe02j8GNBDZp3jYGAwOCfjn/GKTlHEgL+2+QFoIOM7gBtuUA2BZmkBYGQg6znPHYAKglmUfyzMNmyTnnJAwMCGkx5098ZvizzU6O73jywQ9vymwIOwwI2CAeFkgEqZUgrB4ImB+AKX5CDhoFo2AUjIIRCwD9wD7GdNKkVwAAAABJRU5ErkJggg==\",\"orcid\":\"\",\"institution\":\"International Mongolian Hospital of Inner Mongolia\",\"correspondingAuthor\":true,\"prefix\":\"\",\"firstName\":\"Bat\",\"middleName\":\"\",\"lastName\":\"Jin\",\"suffix\":\"\"}],\"badges\":[],\"createdAt\":\"2025-11-29 06:53:12\",\"currentVersionCode\":1,\"declarations\":\"\",\"doi\":\"10.21203/rs.3.rs-8234895/v1\",\"doiUrl\":\"https://doi.org/10.21203/rs.3.rs-8234895/v1\",\"draftVersion\":[],\"editorialEvents\":[],\"editorialNote\":\"\",\"failedWorkflow\":false,\"files\":[{\"id\":100400987,\"identity\":\"1613f8bf-9b72-48de-8c79-3b7f613e985d\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 11:58:37\",\"extension\":\"docx\",\"order_by\":0,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":5428809,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"prespective...docx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/ba2fcee5d8b74388ab5466b5.docx\"},{\"id\":100401001,\"identity\":\"5056259e-2db0-4821-8711-077396008315\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 11:58:38\",\"extension\":\"json\",\"order_by\":1,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":7338,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"75806bca2e854f56995ece8071225ed3.json\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/58fb6a0bdebc1a98ea445707.json\"},{\"id\":100401264,\"identity\":\"a58aac88-5417-4194-9a22-500074ee1a9e\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 11:58:50\",\"extension\":\"xml\",\"order_by\":2,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":114811,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"75806bca2e854f56995ece8071225ed31enriched.xml\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/109194f112be80c34d640162.xml\"},{\"id\":100400552,\"identity\":\"1fcc7ae3-b63e-4214-869a-bf057a58aac7\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 11:58:17\",\"extension\":\"jpeg\",\"order_by\":3,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":889276,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"floatimage1.jpeg\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/a0e394879070c641ab1c5ea9.jpeg\"},{\"id\":100401733,\"identity\":\"8ca30bd1-fc4c-46ae-8055-a16c903a8606\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 11:59:17\",\"extension\":\"jpeg\",\"order_by\":4,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":735760,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"floatimage2.jpeg\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/42f04ba6349154ac581af194.jpeg\"},{\"id\":100401703,\"identity\":\"be0507df-1049-4449-8f0b-3c382f414797\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 11:59:11\",\"extension\":\"jpeg\",\"order_by\":5,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":330742,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"floatimage3.jpeg\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/d8ceab6c082e4a1ac0f9fc05.jpeg\"},{\"id\":100402373,\"identity\":\"7407c60b-60eb-461c-97a9-f88ab407402f\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 12:00:02\",\"extension\":\"jpeg\",\"order_by\":6,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":392938,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"floatimage4.jpeg\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/0f2f2000faa9b2f62cd3a833.jpeg\"},{\"id\":100421491,\"identity\":\"e2ab9ec5-f8a0-465f-9fe5-9ea75b331514\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 13:33:07\",\"extension\":\"jpeg\",\"order_by\":7,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":4012112,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"floatimage5.jpeg\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/635e2469d14f770a989c4933.jpeg\"},{\"id\":100401751,\"identity\":\"125465ed-62a1-4fe5-8c98-6a8d349081f4\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 11:59:17\",\"extension\":\"jpeg\",\"order_by\":8,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":205147,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"floatimage6.jpeg\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/2b83a9cf5ce9ce49c2a8893a.jpeg\"},{\"id\":100401031,\"identity\":\"0ac9f591-39e3-42f9-a1f2-8034e96a9798\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 11:58:39\",\"extension\":\"png\",\"order_by\":9,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":466878,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"Onlinefloatimage1.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/2220539f72feba334dc3f69a.png\"},{\"id\":100402221,\"identity\":\"bf42c6d0-03ea-4397-97d9-e9c54ca7e583\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 11:59:49\",\"extension\":\"png\",\"order_by\":10,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":167707,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"Onlinefloatimage2.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/86a438e2a92be8e3f9b0c3f2.png\"},{\"id\":100401877,\"identity\":\"ab83e26b-77d0-44ad-9f8d-15c2859528e7\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 11:59:24\",\"extension\":\"png\",\"order_by\":11,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":135275,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"Onlinefloatimage3.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/23419a1f489f5c160eabad55.png\"},{\"id\":100402291,\"identity\":\"aa7765cf-e325-41b6-ab3e-7ce9f13f9a12\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 11:59:52\",\"extension\":\"png\",\"order_by\":12,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":108764,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"Onlinefloatimage4.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/8ec91747069f671a29e441a1.png\"},{\"id\":100402396,\"identity\":\"8df834b4-99fe-4600-9c1e-7caba51194bd\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 12:00:04\",\"extension\":\"png\",\"order_by\":13,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":958358,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"Onlinefloatimage5.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/4cdd51ab493eabf152203c78.png\"},{\"id\":100401242,\"identity\":\"4b8d84dc-9fd6-4506-9f06-6518ff2b4245\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 11:58:47\",\"extension\":\"png\",\"order_by\":14,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":44608,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"Onlinefloatimage6.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/421985d1e385ef8296249312.png\"},{\"id\":100401742,\"identity\":\"ebc347ee-c37a-486a-8c0e-73d13e4956a6\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 11:59:17\",\"extension\":\"xml\",\"order_by\":15,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":116208,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"75806bca2e854f56995ece8071225ed31structuring.xml\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/2ca6acec9bf655983e7621a8.xml\"},{\"id\":100402143,\"identity\":\"48ebe305-6cee-43bc-a13f-f0a01bbb380b\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 11:59:45\",\"extension\":\"html\",\"order_by\":16,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":125823,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"earlyproof.html\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/1d20528c50a18c5355bfa980.html\"},{\"id\":100402147,\"identity\":\"7a932e50-fc9b-43ba-acaf-ee4dce492788\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 11:59:45\",\"extension\":\"jpeg\",\"order_by\":1,\"title\":\"Figure 1\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":889276,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003eFlow chart of participant enrollment, follow-up and analysis set (STROBE flow)\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"image1.jpeg\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/83b5283f9109d09f17c1fb4c.jpeg\"},{\"id\":100401791,\"identity\":\"b11f3f75-4907-45e3-bf6c-8f0f7f27eb44\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 11:59:19\",\"extension\":\"jpeg\",\"order_by\":2,\"title\":\"Figure 2\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":735760,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003e16S rRNA Sequencing Quality Overview\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"image2.jpeg\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/af75788b531730e5287c79bd.jpeg\"},{\"id\":100401723,\"identity\":\"e174ec96-16c0-4f81-9895-5581a4deea5b\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 11:59:12\",\"extension\":\"jpeg\",\"order_by\":3,\"title\":\"Figure 3\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":330742,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003eDistribution of relative CSMS change with response threshold\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"image3.jpeg\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/2f12a2199b25f0b2cf7a13b2.jpeg\"},{\"id\":100402199,\"identity\":\"741b8fdb-3417-4470-b439-8d995d249be8\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 11:59:48\",\"extension\":\"jpeg\",\"order_by\":4,\"title\":\"Figure 4\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":392938,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003eMultivariable adjusted odds ratios for pre-specified microbiome features and clinical response (Model B)\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"image4.jpeg\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/ce54b65e5e306bc8eb8242e9.jpeg\"},{\"id\":100402300,\"identity\":\"84f64c06-0c69-497d-a30d-0874944dc605\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 11:59:53\",\"extension\":\"jpeg\",\"order_by\":5,\"title\":\"Figure 5\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":4012112,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003eEvaluation of model predictive performance and clinical benefit. A: ROC curves. B: Calibration. C: Decision Curve Analysis (DCA)\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"image5.jpeg\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/bdc7c8daff8e3a7e3463886c.jpeg\"},{\"id\":100401765,\"identity\":\"d948cedd-d42c-46ff-8d5c-0ab9e74e45b9\",\"added_by\":\"auto\",\"created_at\":\"2026-01-16 11:59:18\",\"extension\":\"jpeg\",\"order_by\":6,\"title\":\"Figure 6\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":205147,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003eGM-SLIT parsimonious predictive score nomogram and application example. A: Nomogram (Model C). B: Example case (standardized inputs)\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"image6.jpeg\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/a606a76589c39854970631c4.jpeg\"},{\"id\":100755294,\"identity\":\"dca5b09e-1b6c-499f-b8bb-25a815c7b51c\",\"added_by\":\"auto\",\"created_at\":\"2026-01-21 06:18:49\",\"extension\":\"pdf\",\"order_by\":0,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"manuscript-pdf\",\"size\":8000061,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"manuscript.pdf\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8234895/v1/84237177-00b5-44dd-ab7e-fc1689033bf9.pdf\"}],\"financialInterests\":\"No competing interests reported.\",\"formattedTitle\":\"Predictive Value of Baseline Gut Microbiome Characteristics for the Response in the Next Pollen Season After Sublingual Immunotherapy in Artemisia Pollen–Induced Allergic Rhinitis: A Single-Center Prospective Cohort Study.\",\"fulltext\":[{\"header\":\"1 Introduction\",\"content\":\"\\u003cp\\u003eAllergic rhinitis is a significant health burden, with Artemisia pollen being the primary allergen in northern China. Symptoms worsen during the pollen season, leading to increased airway reactivity(\\u003cspan citationid=\\\"CR1\\\" class=\\\"CitationRef\\\"\\u003e1\\u003c/span\\u003e). Sublingual immunotherapy is regarded as a disease-modifying approach, but shows marked interindividual heterogeneity in response, and reliable determination of the probability of treatment efficacy before treatment is lacking(\\u003cspan citationid=\\\"CR2\\\" class=\\\"CitationRef\\\"\\u003e2\\u003c/span\\u003e). Current decision-making mainly relies on specific IgE, skin prick test intensity and symptom scales, with limited predictive power and inconsistent thresholds(\\u003cspan citationid=\\\"CR3\\\" class=\\\"CitationRef\\\"\\u003e3\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR4\\\" class=\\\"CitationRef\\\"\\u003e4\\u003c/span\\u003e). Mucosal tolerance is influenced by the gut microbiome, with short-chain fatty acids like butyrate, microbial diversity, and the Prevotella/Bacteroides ratio playing a key role in suppressing Th2 inflammation(\\u003cspan citationid=\\\"CR5\\\" class=\\\"CitationRef\\\"\\u003e5\\u003c/span\\u003e). Previous studies mostly suggested associations, but were mainly cross-sectional or post-treatment analyses, allergens and regimens were heterogeneous, and statistical evaluations focused on associations rather than predictive evidence with clinical utility(\\u003cspan citationid=\\\"CR6\\\" class=\\\"CitationRef\\\"\\u003e6\\u003c/span\\u003e). In clinical practice, prospective data targeting sublingual immunotherapy with Artemisia pollen are still lacking to demonstrate that a single baseline stool measurement can predict response in the next pollen season before treatment and provide quantifiable increment and actual net benefit on top of conventional clinical variables (\\u003cspan citationid=\\\"CR7\\\" class=\\\"CitationRef\\\"\\u003e7\\u003c/span\\u003e). Interpretable and low-cost predictive tools are also scarce, and existing models rarely evaluate discrimination, calibration, reclassification and decision utility simultaneously within a unified outcome framework, with insufficient robustness testing against antibiotic exposure, technical batches and threshold setting(\\u003cspan citationid=\\\"CR8\\\" class=\\\"CitationRef\\\"\\u003e8\\u003c/span\\u003e). How to incorporate pre-specified, reproducible microbiome features into clinical models under conditions with controllable confounding, and to use them in an operational way to serve treatment initiation and persistence, allocation of follow-up intensity and doctor\\u0026ndash;patient communication, still requires systematic validation. This study, based on a single-center prospective cohort and focusing on patients with Artemisia pollen\\u0026ndash;induced allergic rhinitis receiving sublingual immunotherapy, used a single baseline stool 16S rRNA sequencing and standardized clinical data to verify the independent predictive value of three prespecified gut features for next-year clinical response and to examine their incremental value in clinical models, and then developed and internally validated a low-dimensional, interpretable GM-SLIT score, evaluating clinical feasibility in terms of discrimination, calibration, reclassification and decision curves. The results can be summarized into two aspects of novelty: baseline gut features can achieve usable probability stratification before treatment; and they provide meaningful decision benefits at clinically relevant thresholds.\\u003c/p\\u003e\"},{\"header\":\"2 Materials and methods\",\"content\":\"\\u003cdiv id=\\\"Sec3\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e2.1 Study design and participants\\u003c/h2\\u003e \\u003cp\\u003eThis study employs a single-center, prospective cohort design conducted at a specialized outpatient clinic offering standardized allergen immunotherapy. Participants complete baseline assessments and stool collection between November 1, 2023, and March 31, 2024, and initiate Artemisia pollen sublingual immunotherapy (SLIT) within two weeks of completing the baseline assessment. Outcomes are assessed during the Artemisia pollen season, defined as occurring between August 1 and October 15, 2024. The study protocol receives approval from the institutional ethics committee (approval No. 2021-033), and all participants provide written informed consent before any study procedures.\\u003c/p\\u003e \\u003cdiv id=\\\"Sec4\\\" class=\\\"Section3\\\"\\u003e \\u003ch2\\u003e2.1.1 Inclusion and exclusion criteria\\u003c/h2\\u003e \\u003cp\\u003eInclusion criteria were: aged 18\\u0026ndash;60 years; seasonal allergic rhinitis meeting the ARIA diagnostic criteria; serum Artemisia pollen\\u0026ndash;specific IgE\\u0026thinsp;\\u0026ge;\\u0026thinsp;0.35 kUA/L or positive skin prick test to Artemisia pollen (wheal diameter\\u0026thinsp;\\u0026ge;\\u0026thinsp;3 mm with valid controls); planned to receive single Artemisia pollen SLIT and able to provide a baseline stool sample before initiation; having completed electronic diary recording in the previous natural pollen season with diary completeness\\u0026thinsp;\\u0026ge;\\u0026thinsp;80% during the entire peak 6 weeks. Exclusion criteria were: previous Artemisia pollen\\u0026ndash;specific immunotherapy; systemic glucocorticoid therapy or use of immunosuppressive drugs within the past 4 weeks; pregnancy or lactation; moderate-to-severe persistent asthma requiring maintenance with inhaled or oral corticosteroids; inflammatory bowel disease, celiac disease, short bowel syndrome or major gastrointestinal surgery within the past 6 months; other conditions judged by the investigators to affect adherence or safety.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec5\\\" class=\\\"Section3\\\"\\u003e \\u003ch2\\u003e2.1.2 Participant flow and target sample size\\u003c/h2\\u003e \\u003cp\\u003eParticipants were consecutively recruited from the outpatient clinic in the International Mongolian Hospital of Inner Mongolia. Following eligibility screening and informed consent, baseline data and stool samples were collected, and participants then began SLIT with Artemisia pollen. The target sample size was calculated based on the number of events needed for predictive model development. The joint model, which included five clinical variables and three microbiome features, aimed to estimate eight parameters. With an expected response rate of approximately 50%, at least 12.5 events per parameter were required, resulting in a minimum of 100 events. Consequently, the total sample size was set to \\u0026ge;\\u0026thinsp;200 participants. To account for a 10% loss to follow-up or non-evaluable outcomes, 220 participants were ultimately enrolled.\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec6\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e2.2 Intervention protocol and follow-up\\u003c/h2\\u003e \\u003cp\\u003eAll participants received standardized Artemisia pollen sublingual drops from the same manufacturer. The treatment began with a 14-day build-up phase, during which the dose was gradually increased to reach the maintenance dose. Participants continued the maintenance phase with a fixed once-daily dose until the end of the pollen season(\\u003cspan citationid=\\\"CR9\\\" class=\\\"CitationRef\\\"\\u003e9\\u003c/span\\u003e). In case of local or systemic reactions, allergists adjusted the dose based on severity. Adherence was monitored through electronic diary check-ins, bottle weight measurements, and medication refill records, with adherence defined as the proportion of actual medication days\\u0026thinsp;\\u0026ge;\\u0026thinsp;80% of planned maintenance days. Follow-up visits included: baseline, end of the build-up phase, pre-pollen season, mid-pollen season (via phone or outpatient visit), and end of the pollen season. Any medication interruptions or switches were recorded, along with the corresponding dates and reasons.\\u003c/p\\u003e \\u003cdiv id=\\\"Sec7\\\" class=\\\"Section3\\\"\\u003e \\u003ch2\\u003e2.2.1 Definition of the pollen season and observation window\\u003c/h2\\u003e \\u003cp\\u003eThe pollen season interval was defined using daily Artemisia pollen counts from the local aerobiological monitoring station. The season began when the 5-day moving average of pollen count first reached 20 grains/m\\u0026sup3; and ended when the 5-day moving average remained below 20 grains/m\\u0026sup3; for seven consecutive days (\\u003cspan citationid=\\\"CR10\\\" class=\\\"CitationRef\\\"\\u003e10\\u003c/span\\u003e). The consecutive 6 weeks with the highest moving average were designated as the peak observation period, during which all primary outcomes were calculated based on diary data.\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec8\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e2.3 Outcome measures and definition of response\\u003c/h2\\u003e \\u003cp\\u003eThe primary outcome was the binary variable of \\u0026ldquo;clinical response in the next pollen season\\u0026rdquo;. Seasonal-level indices were calculated using the daily combined symptom-medication score (CSMS) recommended by EAACI (\\u003cspan citationid=\\\"CR11\\\" class=\\\"CitationRef\\\"\\u003e11\\u003c/span\\u003e): the daily Rhinoconjunctivitis Total Symptom Score (RTSS) recorded 6 symptoms, each scored 0\\u0026ndash;3, with a total score of 0\\u0026ndash;18; the daily rescue medication score (RMS) was graded 0\\u0026ndash;3 (0 for no medication; 1 for use of oral H1 antihistamines or antiallergic eye drops (\\u003cspan citationid=\\\"CR12\\\" class=\\\"CitationRef\\\"\\u003e12\\u003c/span\\u003e); 2 for use of nasal corticosteroids without systemic corticosteroids; 3 for use of systemic corticosteroids on that day); the daily CSMS\\u0026thinsp;=\\u0026thinsp;RTSS/6\\u0026thinsp;+\\u0026thinsp;RMS, with a range of 0\\u0026ndash;6. The mean CSMS during the peak 6 weeks of the previous natural pollen season was used as the baseline seasonal value, and the mean CSMS during the peak 6 weeks of the next pollen season was used as the follow-up value (\\u003cspan citationid=\\\"CR13\\\" class=\\\"CitationRef\\\"\\u003e13\\u003c/span\\u003e). The relative reduction proportion was defined as (baseline-season CSMS\\u0026thinsp;\\u0026minus;\\u0026thinsp;follow-up-season CSMS)/baseline-season CSMS. A relative reduction\\u0026thinsp;\\u0026ge;\\u0026thinsp;30% was prespecified as response. Secondary outcomes included the continuous value of CSMS in the follow-up season, change in RQLQ total score and change in the number of days of medication use, for descriptive supplementation.\\u003c/p\\u003e \\u003cdiv id=\\\"Sec9\\\" class=\\\"Section3\\\"\\u003e \\u003ch2\\u003e2.3.1 Collection and grading of adverse events\\u003c/h2\\u003e \\u003cp\\u003eAll SLIT-related adverse events were recorded in real time during the build-up and maintenance phases through outpatient visits and electronic diaries. Local and systemic reactions were classified using the World Allergy Organization (WAO) grading criteria, with details on the most severe grade, start and end times, and management measures. Serious adverse events were reported to the ethics committee and safety supervisor within 24 hours.\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec10\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e2.4 Stool sample collection and 16S rRNA sequencing\\u003c/h2\\u003e \\u003cp\\u003eBaseline stool samples were collected at home within 14 days before SLIT initiation, using sampling tubes with stabilizer. Samples were thoroughly mixed on-site, kept at room temperature for no more than 24 hours, and stored at \\u0026minus;\\u0026thinsp;80\\u0026deg;C within 4 hours of arrival at the laboratory. DNA was extracted using a standardized kit that included a mechanical disruption step, with blank sampling tubes and extraction negative controls included in each batch. Microbial quality control materials with known composition were added to monitor batch stability. The V3\\u0026ndash;V4 region of the 16S rRNA gene was amplified using the 341F (5\\u0026prime;-CCTACGGGNGGCWGCAG-3\\u0026prime;) and 806R (5\\u0026prime;-GACTACHVGGGTATCTAATCC-3\\u0026prime;) primers for paired-end PCR. Libraries were sequenced using 2\\u0026times;250 bp paired-end sequencing on the Illumina MiSeq platform, with each batch including negative controls and quality control bacterial communities. The target sequencing depth was \\u0026ge;\\u0026thinsp;25,000 reads/sample; samples not meeting this standard underwent repeat extraction and amplification.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec11\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e2.5 Microbial data processing and prespecified features\\u003c/h2\\u003e \\u003cp\\u003eRaw sequences were processed in QIIME 2 (version 2023.5) for adapter trimming, quality control, and denoising. Amplicon sequence variants (ASVs) were generated using DADA2, with chimera removal. Taxonomic annotation was performed using the SILVA 138 reference database, with a confidence threshold of 0.8 for annotation from phylum to genus level. Contamination assessment was conducted using the prevalence method in decontam to differentiate negative controls and low-biomass samples, and suspected contaminant ASVs were removed. The rarefaction depth for diversity analysis was set to 10,000 reads/sample. For regression modeling, community composition data were transformed using centered log-ratio (CLR) after adding a pseudocount of 0.5 to the relative abundance matrix to mitigate compositional bias.\\u003c/p\\u003e \\u003cp\\u003eThe prespecified microbiome features were limited to three and locked in the analysis plan: the Shannon index; the composite proportion of butyrate-producing bacteria, defined as the sum of the relative abundances of Faecalibacterium, Roseburia and the Eubacterium rectale group; and the Prevotella/Bacteroides ratio, calculated as log10 (Prevotella relative abundance\\u0026thinsp;+\\u0026thinsp;0.5)\\u0026thinsp;\\u0026minus;\\u0026thinsp;log10 (Bacteroides relative abundance\\u0026thinsp;+\\u0026thinsp;0.5)(\\u003cspan citationid=\\\"CR14\\\" class=\\\"CitationRef\\\"\\u003e14\\u003c/span\\u003e). All features were standardized as z-scores before modeling.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec12\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e2.6 Clinical variables and potential confounders\\u003c/h2\\u003e \\u003cp\\u003eAt baseline, demographic information (age, sex, body mass index, duration of rhinitis) and lifestyle (smoking status) were recorded. Allergic indices included Artemisia-specific IgE concentration (entered into the model after log10 transformation) and whether there was concomitant sensitization to other pollens (specific IgE to grass or tree pollens\\u0026thinsp;\\u0026ge;\\u0026thinsp;0.35 kUA/L was defined as the binary variable of polysensitization). Baseline disease severity was described using the RQLQ total score and the previous season CSMS as continuous variables; only the RQLQ total score was retained in the modeling as the representative of severity. Use of antibiotics, probiotics and proton pump inhibitors in the past 3 months was recorded by questionnaires and cross-checked with the prescription system for sensitivity analyses. Technical batch variables (DNA extraction batch and sequencing run) were used for quality control and sensitivity analyses.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec13\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e2.7 Statistical analysis\\u003c/h2\\u003e \\u003cp\\u003eAll analyses were conducted using R 4.3.2. Continuous variables were reported as means with standard deviations or medians with interquartile ranges, while categorical variables were presented as frequencies and percentages. A two-sided significance level of α\\u0026thinsp;=\\u0026thinsp;0.05 was used. Before modeling, independent variables were centered and standardized, and outliers were assessed using Cook\\u0026rsquo;s distance and leverage, with Winsorization applied when necessary. Missing covariates were addressed using multiple imputation (MICE, 20 imputed datasets), with chained equations including all predictors and the outcome variable; missing outcomes were not imputed, and the main analysis was performed on the evaluable population. All regression coefficients, predictive performance measures, and reclassification indices were estimated based on pooled results after imputation, combined according to Rubin\\u0026rsquo;s rules.\\u003c/p\\u003e \\u003cdiv id=\\\"Sec14\\\" class=\\\"Section3\\\"\\u003e \\u003ch2\\u003e2.7.1 Main analysis: assessment of independent predictive value\\u003c/h2\\u003e \\u003cp\\u003eA multivariable Logistic regression model was constructed with \\u0026ldquo;clinical response in the next pollen season\\u0026rdquo; as the dependent variable. Clinical Model A included five prespecified clinical variables: age, sex, RQLQ total score, Artemisia-specific IgE (log10) and polysensitization (binary). Combined Model B added three microbiome features to Model A (Shannon index, composite proportion of butyrate-producing bacteria, Prevotella/Bacteroides ratio). Adjusted odds ratios and 95% confidence intervals were reported for each predictor, and variance inflation factors were used to assess collinearity in Combined Model B.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec15\\\" class=\\\"Section3\\\"\\u003e \\u003ch2\\u003e2.7.2 Incremental predictive value and model performance\\u003c/h2\\u003e \\u003cp\\u003eDiscrimination was evaluated using the area under the receiver operating characteristic curve (AUC), and differences in AUC between Model A and Model B were compared using the DeLong method; the Brier score, calibration intercept and calibration slope were also reported, and a spline-smoothed calibration curve was plotted. Risk reclassification was assessed using the category-free net reclassification improvement (NRI) and the integrated discrimination improvement (IDI), with the threshold probability range prespecified as 0.2\\u0026ndash;0.8. Decision curve analysis was used to compare net benefit at different threshold probabilities.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec16\\\" class=\\\"Section3\\\"\\u003e \\u003ch2\\u003e2.7.3 Construction of the predictive tool and internal validation\\u003c/h2\\u003e \\u003cp\\u003eOn the basis of Model B, L1 regularization was used for feature selection, retaining at most 5 variables to form the parsimonious Model C. A point-based score and a nomogram were constructed according to the relative magnitudes of the regression coefficients, and a threshold was provided to map predicted probabilities to risk strata (high probability and low probability). Internal validation used 1,000-bootstrap resampling to estimate optimism and perform correction, and the optimism-corrected AUC, Brier score and calibration indices were reported. For the final score, the sensitivity, specificity, positive predictive value and negative predictive value at the classification threshold were provided.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec17\\\" class=\\\"Section3\\\"\\u003e \\u003ch2\\u003e2.7.4 Sensitivity analyses and control of multiple comparisons\\u003c/h2\\u003e \\u003cp\\u003eThree prespecified sensitivity analyses were performed: excluding participants who had used antibiotics within 30 days before baseline; resetting the response threshold to 25% and 35% to recompute model stability; and adding technical batch indicator variables into the regression to examine the impact of batch effects. For multiple testing of the three microbiome features, the Benjamini\\u0026ndash;Hochberg method was used to control the false discovery rate at 5%, and no additional correction was applied for overall model comparisons.\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec18\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e2.8 Data quality control and bias management\\u003c/h2\\u003e \\u003cp\\u003eElectronic data were collected using REDCap, with field validity checks and logical verification rules specified. Microbiological experiments were performed in a batch-wise standardized workflow, and negative and positive quality control materials were used to monitor contamination and drift; within- and between-batch sequencing consistency was assessed using Pearson correlation and Bray\\u0026ndash;Curtis distance. Stool samples and sequence identifiers were transferred in a double-blind manner between the statistical and clinical teams, and statisticians were unaware of clinical outcomes before database lock. To reduce information bias, symptom and medication diaries used standardized templates and daily push reminders; missing records were verified by telephone and the reasons were documented. Medication use records were verified by comparison with outpatient pharmacy data to improve the accuracy of medication scores.\\u003c/p\\u003e \\u003c/div\\u003e\"},{\"header\":\"3 Results\",\"content\":\"\\u003cdiv id=\\\"Sec20\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e3.1 Study participants and baseline characteristics\\u003c/h2\\u003e \\u003cp\\u003e A total of 356 patients were screened, and 204 (57.3%) met the inclusion criteria and were included in the analysis set. The main reasons for exclusion were not meeting inclusion criteria (31.5%), not completing stool collection or quality control (4.9%) and not completing treatment follow-up (11.5%) (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003e). Continuous variables were compared using independent-samples t tests, and categorical variables were compared using Pearson χ\\u0026sup2; tests. Standardized mean differences (SMD) were \\u0026le;\\u0026thinsp;0.21 (considered mild), and the tests did not yield statistically significant results (all P\\u0026thinsp;\\u0026gt;\\u0026thinsp;0.05) (Table\\u0026nbsp;\\u003cspan refid=\\\"Tab1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003e).\\u003c/p\\u003e \\u003cp\\u003e \\u003c/p\\u003e \\u003cp\\u003e \\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab1\\\" border=\\\"1\\\"\\u003e \\u003ccaption language=\\\"En\\\"\\u003e \\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 1\\u003c/div\\u003e \\u003cdiv class=\\\"CaptionContent\\\"\\u003e \\u003cp\\u003eBaseline clinical characteristics of participants (stratified by response status in the next pollen season)\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/caption\\u003e \\u003ccolgroup cols=\\\"7\\\"\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c3\\\" colnum=\\\"3\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c4\\\" colnum=\\\"4\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c5\\\" colnum=\\\"5\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c6\\\" colnum=\\\"6\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c7\\\" colnum=\\\"7\\\"\\u003e\\u003c/div\\u003e \\u003cthead\\u003e \\u003ctr\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eVariable\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eOverall (n\\u0026thinsp;=\\u0026thinsp;204)\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eResponders (n\\u0026thinsp;=\\u0026thinsp;111)\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003eNon-responders (n\\u0026thinsp;=\\u0026thinsp;93)\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e|SMD|\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003eTest statistic\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003eP\\u003c/p\\u003e \\u003c/th\\u003e \\u003c/tr\\u003e \\u003c/thead\\u003e \\u003ctbody\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eAge (years)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e32.80\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;9.63\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e31.94\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;9.36\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e33.82\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;9.90\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.20\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003et\\u0026thinsp;=\\u0026thinsp;\\u0026minus;\\u0026thinsp;1.385\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e0.168\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eSex (male/female)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e96 (47.06%); 108 (52.94%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e49 (44.14%); 62 (55.86%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003eMale 47 (50.54%); Female 46 (49.46%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.13\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003eχ\\u0026sup2; = 0.830\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e0.362\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eBody mass index (kg/m\\u0026sup2;)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e23.84\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;3.16\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e23.56\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;3.10\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e24.18\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;3.22\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.20\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003et\\u0026thinsp;=\\u0026thinsp;\\u0026minus;\\u0026thinsp;1.393\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e0.165\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eDuration of rhinitis (years)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e7.41\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;4.15\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e7.02\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;3.97\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e7.88\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;4.33\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.21\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003et\\u0026thinsp;=\\u0026thinsp;\\u0026minus;\\u0026thinsp;1.467\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e0.144\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eBaseline-season CSMS (daily mean)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e2.36\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;0.62\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e2.31\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;0.60\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e2.41\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;0.65\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.16\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003et\\u0026thinsp;=\\u0026thinsp;\\u0026minus;\\u0026thinsp;1.133\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e0.259\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eRQLQ total score\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e3.27\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;1.08\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e3.19\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;1.04\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e3.37\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;1.12\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.17\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003et\\u0026thinsp;=\\u0026thinsp;\\u0026minus;\\u0026thinsp;1.181\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e0.239\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eArtemisia pollen\\u0026ndash;specific IgE (log10, kUA/L)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.96\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;0.48\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.93\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;0.46\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e1.00\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;0.50\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.15\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003et\\u0026thinsp;=\\u0026thinsp;\\u0026minus;\\u0026thinsp;1.033\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e0.303\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003ePolysensitization (yes/no)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e91 (44.61%); 113 (55.39%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e46 (41.44%); 65 (58.56%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e45 (48.39%); 48 (51.61%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.14\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003eχ\\u0026sup2; = 0.988\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e0.32\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eComorbid asthma (yes/no)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e36 (17.65%); 168 (82.35%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e17 (15.32%); 94 (84.68%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e19 (20.43%); 74 (79.57%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.13\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003eχ\\u0026sup2; = 0.911\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e0.34\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eSmoking status (never/former/current)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e137 (67.16%); 37 (18.14%); 30 (14.71%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e77 (69.37%); 18 (16.22%); 16 (14.41%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e60 (64.52%); 19 (20.43%); 14 (15.05%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.11\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003eχ\\u0026sup2; = 0.687\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e0.709\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eInterval from baseline stool collection to SLIT initiation (days)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e6.23\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;2.75\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e6.07\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;2.81\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e6.42\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;2.69\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.13\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003et\\u0026thinsp;=\\u0026thinsp;\\u0026minus;\\u0026thinsp;0.907\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e0.366\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003c/tbody\\u003e \\u003c/colgroup\\u003e \\u003ctfoot\\u003e \\u003ctr\\u003e\\u003ctd colspan=\\\"7\\\"\\u003e\\u003cem\\u003eNotes: Responder definition: during the next pollen season, the 6-week peak-period mean CSMS decreases by \\u0026ge;\\u0026thinsp;30% vs. the baseline season; CSMS calculation: CSMS\\u0026thinsp;=\\u0026thinsp;RTSS/6\\u0026thinsp;+\\u0026thinsp;RMS; reported as the daily mean within the season; Specific IgE values are presented after log10 transformation.\\u003c/em\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e \\u003c/tfoot\\u003e \\u003c/table\\u003e\\u003c/div\\u003e \\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec21\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e3.2 Data completeness and SLIT exposure/adherence\\u003c/h2\\u003e \\u003cp\\u003eContinuous variables were compared using independent-samples t tests, and categorical variables were compared using Pearson χ\\u0026sup2; tests. The results showed that diary completeness and medication adherence were high in each group and the differences were not significant (all P\\u0026thinsp;\\u0026gt;\\u0026thinsp;0.05), suggesting good adherence to SLIT treatment and sufficient data completeness (Table\\u0026nbsp;\\u003cspan refid=\\\"Tab2\\\" class=\\\"InternalRef\\\"\\u003e2\\u003c/span\\u003e).\\u003c/p\\u003e \\u003cp\\u003e \\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab2\\\" border=\\\"1\\\"\\u003e \\u003ccaption language=\\\"En\\\"\\u003e \\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 2\\u003c/div\\u003e \\u003cdiv class=\\\"CaptionContent\\\"\\u003e \\u003cp\\u003eOverview of electronic diary completeness during the pollen season and SLIT exposure/adherence\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/caption\\u003e \\u003ccolgroup cols=\\\"6\\\"\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c3\\\" colnum=\\\"3\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c4\\\" colnum=\\\"4\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c5\\\" colnum=\\\"5\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c6\\\" colnum=\\\"6\\\"\\u003e\\u003c/div\\u003e \\u003cthead\\u003e \\u003ctr\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eMeasure\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eOverall (n\\u0026thinsp;=\\u0026thinsp;204)\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eResponders (n\\u0026thinsp;=\\u0026thinsp;111)\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003eNon-responders (n\\u0026thinsp;=\\u0026thinsp;93)\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003eTest statistic\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003eP value\\u003c/p\\u003e \\u003c/th\\u003e \\u003c/tr\\u003e \\u003c/thead\\u003e \\u003ctbody\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eValid diary days (days)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e39.08\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;3.12\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e39.34\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;2.96\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e38.73\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;3.31\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003et\\u0026thinsp;=\\u0026thinsp;1.375\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e0.079\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eDiary completeness (%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e93.01\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;7.43\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e93.67\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;7.06\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e92.22\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;7.88\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003et\\u0026thinsp;=\\u0026thinsp;1.372\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e0.103\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003ePlanned maintenance medication days (days)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e243.83\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;30.37\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e245.09\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;29.08\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e242.32\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;31.79\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003et\\u0026thinsp;=\\u0026thinsp;0.644\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e0.486\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eActual days on medication (days)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e228.93\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;38.41\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e231.42\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;36.87\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e225.95\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;39.89\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003et\\u0026thinsp;=\\u0026thinsp;1.010\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e0.274\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eMedication adherence (%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e93.90\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;8.34\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e94.47\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;7.94\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e93.31\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;8.57\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003et\\u0026thinsp;=\\u0026thinsp;0.996\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e0.323\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eTotal SLIT duration (days)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e257.84\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;31.72\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e259.71\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;30.06\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e255.60\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;33.51\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003et\\u0026thinsp;=\\u0026thinsp;0.914\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e0.318\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eAE-related interruption or dose reduction (yes/no, n, %)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e29 (14.22%); 175 (85.78%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e14 (12.61%); 97 (87.39%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e15 (16.13%); 78 (83.87%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003eχ\\u0026sup2; = 0.513\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e0.448\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003c/tbody\\u003e \\u003c/colgroup\\u003e \\u003ctfoot\\u003e \\u003ctr\\u003e\\u003ctd colspan=\\\"6\\\"\\u003e\\u003cem\\u003eNote: Adherence was defined as the proportion of actual medication days\\u0026thinsp;\\u0026ge;\\u0026thinsp;80% of planned maintenance medication days.\\u003c/em\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e \\u003c/tfoot\\u003e \\u003c/table\\u003e\\u003c/div\\u003e \\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec22\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e3.3 16S sequencing and quality control\\u003c/h2\\u003e \\u003cp\\u003eThe method adopted 16S rRNA V3\\u0026ndash;V4 sequencing, with a rarefaction depth of 10,000 reads. Panel A used the Hodges\\u0026ndash;Lehmann method to evaluate between-group differences in sequencing depth, and Panel B plotted LOESS rarefaction curves. The results showed that sequencing depths were similar between the two groups (HL\\u0026thinsp;=\\u0026thinsp;779 reads, 95% CI \\u0026minus;\\u0026thinsp;428 to 1,962), with no obvious systematic bias (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig2\\\" class=\\\"InternalRef\\\"\\u003e2\\u003c/span\\u003e).\\u003c/p\\u003e \\u003cp\\u003e \\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec23\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e3.4 Primary outcome: CSMS change and response rate\\u003c/h2\\u003e \\u003cp\\u003eUsing the peak 6-week diary, the relative change in CSMS was right-skewed; the median improvement was 32.83% (95% CI 28.87\\u0026ndash;36.61). With the \\u0026ge;\\u0026thinsp;30% threshold, the response rate was 54.41% (95% CI 47.56\\u0026ndash;61.10). The kernel density peak was close to the 30% threshold (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig3\\\" class=\\\"InternalRef\\\"\\u003e3\\u003c/span\\u003e).\\u003c/p\\u003e \\u003cp\\u003e \\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec24\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e3.5 Independent predictive value of baseline gut microbiota\\u003c/h2\\u003e \\u003cp\\u003eIn Model B, the three prespecified microbiome features were independently associated with response (all corrected by BH-FDR), with OR\\u0026thinsp;=\\u0026thinsp;1.59 per +\\u0026thinsp;1 SD for the composite proportion of butyrate-producing bacteria, OR\\u0026thinsp;=\\u0026thinsp;1.43 for the Prevotella/Bacteroides ratio and OR\\u0026thinsp;=\\u0026thinsp;1.33 for the Shannon index (all q\\u0026thinsp;\\u0026lt;\\u0026thinsp;0.05); the other clinical covariates did not reach statistical significance (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig4\\\" class=\\\"InternalRef\\\"\\u003e4\\u003c/span\\u003e).\\u003c/p\\u003e \\u003cp\\u003e \\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec25\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e3.6 Incremental predictive value of microbiome features\\u003c/h2\\u003e \\u003cp\\u003eModel discrimination was evaluated using ROC curves (DeLong test to compare AUC between Model A and Model B), and calibration was assessed by Bootstrap analysis. After adding microbiome features, the AUC of Model B increased from 0.71 for Model A to 0.79, and the difference in AUC between the two models was statistically significant (ΔAUC\\u0026thinsp;=\\u0026thinsp;0.08, P\\u0026thinsp;=\\u0026thinsp;0.021). Model B showed better calibration (α=\\u0026minus;0.03; β\\u0026thinsp;=\\u0026thinsp;0.98), closer to the ideal values (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig5\\\" class=\\\"InternalRef\\\"\\u003e5\\u003c/span\\u003eA, B). The category-free NRI and IDI were used to evaluate the incremental predictive ability of Model B (Bootstrap test). After adding microbiome features, both NRI_total and IDI increased significantly (both P\\u0026thinsp;\\u0026lt;\\u0026thinsp;0.01), indicating that the predictive performance of Model B was clearly improved compared with Model A (Table\\u0026nbsp;\\u003cspan refid=\\\"Tab3\\\" class=\\\"InternalRef\\\"\\u003e3\\u003c/span\\u003e). Decision curve analysis evaluated net benefit (Bootstrap 1000 times). Within the range of pₜ=0.10\\u0026ndash;0.80, the overall net benefit of Model B was higher; at pₜ=0.30 and 0.50, ΔNB was 0.04 and 0.03, respectively, both superior to Model A (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig5\\\" class=\\\"InternalRef\\\"\\u003e5\\u003c/span\\u003eC).\\u003c/p\\u003e \\u003cp\\u003e \\u003c/p\\u003e \\u003cp\\u003e \\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab3\\\" border=\\\"1\\\"\\u003e \\u003ccaption language=\\\"En\\\"\\u003e \\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 3\\u003c/div\\u003e \\u003cdiv class=\\\"CaptionContent\\\"\\u003e \\u003cp\\u003eCategory-free NRI and IDI of Model B compared with Model A\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/caption\\u003e \\u003ccolgroup cols=\\\"4\\\"\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c3\\\" colnum=\\\"3\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c4\\\" colnum=\\\"4\\\"\\u003e\\u003c/div\\u003e \\u003cthead\\u003e \\u003ctr\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eMetric\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eValue\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e95%CI\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003eP_Boot\\u003c/p\\u003e \\u003c/th\\u003e \\u003c/tr\\u003e \\u003c/thead\\u003e \\u003ctbody\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eNRI_events\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.27\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.10\\u0026ndash;0.46\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.006\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eNRI_non-events\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.09\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e\\u0026minus;0.03\\u0026ndash;0.22\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.146\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eNRI_total(=events\\u0026thinsp;+\\u0026thinsp;non-events)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.36\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.18\\u0026ndash;0.54\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.002\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eIDI\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.05\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.02\\u0026ndash;0.09\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.004\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003c/tbody\\u003e \\u003c/colgroup\\u003e \\u003ctfoot\\u003e \\u003ctr\\u003e\\u003ctd colspan=\\\"4\\\"\\u003e\\u003cem\\u003eNote: Category-free NRI and IDI were calculated according to the Pencina method.\\u003c/em\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e \\u003c/tfoot\\u003e \\u003c/table\\u003e\\u003c/div\\u003e \\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec26\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e3.7 Parsimonious predictive tool and internal validation\\u003c/h2\\u003e \\u003cp\\u003eModel C retained 5 variables after L1 regularization. A point-based score was constructed by linearly transforming the regression coefficients (Butyrate-producer composite\\u0026thinsp;=\\u0026thinsp;100 points as the reference; P/B ratio\\u0026thinsp;=\\u0026thinsp;71.29 points; Shannon index\\u0026thinsp;=\\u0026thinsp;58.37 points; RQLQ total and Artemisia sIgE log10 contributed negative points), and the total score T was mapped to the predicted probability p; p\\u0026thinsp;\\u0026ge;\\u0026thinsp;0.62 was defined as high response probability. For the example individual, the total score (T\\u0026thinsp;=\\u0026thinsp;158.21) corresponded to a predicted probability of about p\\u0026thinsp;=\\u0026thinsp;69.83%, indicating that this subject belonged to the high-probability response group (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig6\\\" class=\\\"InternalRef\\\"\\u003e6\\u003c/span\\u003eA, B). Based on the 2\\u0026times;2 table, Se, Sp, PPV, NPV and F1 were calculated, with 95% CI obtained by Bootstrap, and the threshold was determined by the Youden J. With stratification at p\\u0026thinsp;\\u0026ge;\\u0026thinsp;0.62, the response rate in the high-probability response group was 76.79%, which was markedly higher than 27.17% in the low-probability response group; the classification performance of Model C was: Se 77.48%, Sp 72.04%, PPV 76.79%, NPV 72.83%, F1 77.13% (Table\\u0026nbsp;\\u003cspan refid=\\\"Tab4\\\" class=\\\"InternalRef\\\"\\u003e4\\u003c/span\\u003e). Harrell method Bootstrap (1000) was used for optimism correction. Internal validation showed AUC_optim\\u0026thinsp;=\\u0026thinsp;0.78; Brier\\u0026thinsp;=\\u0026thinsp;0.187; calibration intercept α=\\u0026minus;0.01 and calibration slope β\\u0026thinsp;=\\u0026thinsp;0.97 (Table\\u0026nbsp;\\u003cspan refid=\\\"Tab5\\\" class=\\\"InternalRef\\\"\\u003e5\\u003c/span\\u003e).\\u003c/p\\u003e \\u003cp\\u003e \\u003c/p\\u003e \\u003cp\\u003e \\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab4\\\" border=\\\"1\\\"\\u003e \\u003ccaption language=\\\"En\\\"\\u003e \\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 4\\u003c/div\\u003e \\u003cdiv class=\\\"CaptionContent\\\"\\u003e \\u003cp\\u003eRisk stratification and classification performance of the GM-SLIT score\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/caption\\u003e \\u003ccolgroup cols=\\\"2\\\"\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e \\u003cthead\\u003e \\u003ctr\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eMetric\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eResult (95% CI)\\u003c/p\\u003e \\u003c/th\\u003e \\u003c/tr\\u003e \\u003c/thead\\u003e \\u003ctbody\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eThreshold (predicted probability)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003ep\\u0026thinsp;\\u0026ge;\\u0026thinsp;0.62 (Youden J\\u0026thinsp;=\\u0026thinsp;0.495)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eHigh-probability response group (n, %)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e112 (54.90%); observed response rate 76.79% (68.16\\u0026ndash;83.64)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eLow-probability response group (n, %)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e92 (45.10%); observed response rate 27.17% (19.14\\u0026ndash;37.04)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eConfusion matrix (TP, FP, FN, TN)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e[TP\\u0026thinsp;=\\u0026thinsp;86, FP\\u0026thinsp;=\\u0026thinsp;26; FN\\u0026thinsp;=\\u0026thinsp;25, TN\\u0026thinsp;=\\u0026thinsp;67]\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eSensitivity (Se)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e77.48% (68.86\\u0026ndash;84.25)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eSpecificity (Sp)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e72.04% (62.19\\u0026ndash;80.15)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003ePositive predictive value (PPV)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e76.79% (68.16\\u0026ndash;83.64)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eNegative predictive value (NPV)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e72.83% (62.96\\u0026ndash;80.86)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eF1-score\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e77.13% (69.74\\u0026ndash;83.02)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003c/tbody\\u003e \\u003c/colgroup\\u003e \\u003ctfoot\\u003e \\u003ctr\\u003e\\u003ctd colspan=\\\"2\\\"\\u003e\\u003cem\\u003eNote: The threshold was determined by maximizing the Youden index (J\\u0026thinsp;=\\u0026thinsp;0.495 in this table). Sample size N\\u0026thinsp;=\\u0026thinsp;204, responders\\u0026thinsp;=\\u0026thinsp;111; stratification and performance all refer to Model C.\\u003c/em\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e \\u003c/tfoot\\u003e \\u003c/table\\u003e\\u003c/div\\u003e \\u003c/p\\u003e \\u003cp\\u003e \\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab5\\\" border=\\\"1\\\"\\u003e \\u003ccaption language=\\\"En\\\"\\u003e \\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 5\\u003c/div\\u003e \\u003cdiv class=\\\"CaptionContent\\\"\\u003e \\u003cp\\u003ePerformance of Model C after internal validation (optimism-corrected)\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/caption\\u003e \\u003ccolgroup cols=\\\"3\\\"\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c3\\\" colnum=\\\"3\\\"\\u003e\\u003c/div\\u003e \\u003cthead\\u003e \\u003ctr\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eMetric\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eValue\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e95% CI\\u003c/p\\u003e \\u003c/th\\u003e \\u003c/tr\\u003e \\u003c/thead\\u003e \\u003ctbody\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eOptimism-corrected AUC (AUC_optim)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.78\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.72\\u0026ndash;0.84\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eBrier score\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.187\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.172\\u0026ndash;0.204\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eCalibration intercept α\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e\\u0026minus;0.01\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e\\u0026minus;0.17\\u0026ndash;0.14\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eCalibration slope β\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.97\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.83\\u0026ndash;1.12\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003c/tbody\\u003e \\u003c/colgroup\\u003e \\u003c/table\\u003e\\u003c/div\\u003e \\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec27\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e3.8 Prespecified sensitivity analyses\\u003c/h2\\u003e \\u003cp\\u003eIn the sensitivity analyses, ΔAUC varied only slightly across scenarios (\\u0026minus;\\u0026thinsp;0.01 to +\\u0026thinsp;0.01), and Δβ ranged from \\u0026minus;\\u0026thinsp;0.06 to +\\u0026thinsp;0.04; at pₜ=0.30 and 0.50, the differences in net benefit were 0.03\\u0026ndash;0.06 and 0.02\\u0026ndash;0.04, respectively. The effect directions of the three microbiome features remained consistently stable (\\u0026ge;\\u0026thinsp;90%) (Table\\u0026nbsp;\\u003cspan refid=\\\"Tab6\\\" class=\\\"InternalRef\\\"\\u003e6\\u003c/span\\u003e).\\u003c/p\\u003e \\u003cp\\u003e \\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab6\\\" border=\\\"1\\\"\\u003e \\u003ccaption language=\\\"En\\\"\\u003e \\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 6\\u003c/div\\u003e \\u003cdiv class=\\\"CaptionContent\\\"\\u003e \\u003cp\\u003eSummary of robustness of Model B under sensitivity scenarios (95% CI)\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/caption\\u003e \\u003ccolgroup cols=\\\"5\\\"\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c3\\\" colnum=\\\"3\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c4\\\" colnum=\\\"4\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c5\\\" colnum=\\\"5\\\"\\u003e\\u003c/div\\u003e \\u003cthead\\u003e \\u003ctr\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eMetric\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eExclude antibiotics within 30 days pre-baseline\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eResponse threshold\\u0026thinsp;=\\u0026thinsp;25%\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003eResponse threshold\\u0026thinsp;=\\u0026thinsp;35%\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003eInclude batch indicators in regression\\u003c/p\\u003e \\u003c/th\\u003e \\u003c/tr\\u003e \\u003c/thead\\u003e \\u003ctbody\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eΔAUC\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e+\\u0026thinsp;0.01 (\\u0026minus;\\u0026thinsp;0.01\\u0026ndash;0.03)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e\\u0026minus;0.01 (\\u0026minus;\\u0026thinsp;0.03\\u0026ndash;0.01)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e+\\u0026thinsp;0.01 (\\u0026minus;\\u0026thinsp;0.02\\u0026ndash;0.03)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e+\\u0026thinsp;0.001 (\\u0026minus;\\u0026thinsp;0.018\\u0026ndash;0.02)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eΔ calibration slope\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e+\\u0026thinsp;0.03 (\\u0026minus;\\u0026thinsp;0.02\\u0026ndash;0.07)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e\\u0026minus;0.06 (\\u0026minus;\\u0026thinsp;0.11\\u0026ndash;\\u0026minus;0.01)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e+\\u0026thinsp;0.04 (\\u0026minus;\\u0026thinsp;0.01\\u0026ndash;0.08)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e+\\u0026thinsp;0.01 (\\u0026minus;\\u0026thinsp;0.03\\u0026ndash;0.05)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eΔNB_0.30\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.05 (0.02\\u0026ndash;0.08)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.03 (0.01\\u0026ndash;0.05)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.06 (0.03\\u0026ndash;0.09)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.04 (0.01\\u0026ndash;0.07)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eΔNB_0.50\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.03 (0.01\\u0026ndash;0.06)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.02 (0.01\\u0026ndash;0.04)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.04 (0.02\\u0026ndash;0.06)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.03 (0.01\\u0026ndash;0.05)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eDirection consistency\\u0026mdash;Shannon\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e92.18% (85.64\\u0026ndash;96.02)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e90.84% (83.98\\u0026ndash;94.98)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e92.63% (86.52\\u0026ndash;96.25)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e91.42% (84.83\\u0026ndash;95.43)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eDirection consistency\\u0026mdash;butyrate-producer composite\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e96.27% (91.65\\u0026ndash;98.56)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e95.53% (90.41\\u0026ndash;98.20)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e96.88% (92.74\\u0026ndash;98.85)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e96.02% (91.18\\u0026ndash;98.39)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eDirection consistency\\u0026mdash;P/B ratio\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e94.11% (88.32\\u0026ndash;97.27)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e93.02% (86.84\\u0026ndash;96.65)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e94.37% (88.79\\u0026ndash;97.33)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e93.58% (87.92\\u0026ndash;96.92)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003c/tbody\\u003e \\u003c/colgroup\\u003e \\u003ctfoot\\u003e \\u003ctr\\u003e\\u003ctd colspan=\\\"5\\\"\\u003e\\u003cem\\u003eNote: NRI and IDI are not repeated in this table; reclassification results are shown in\\u003c/em\\u003e Table\\u0026nbsp;\\u003cspan refid=\\\"Tab3\\\" class=\\\"InternalRef\\\"\\u003e3\\u003c/span\\u003e; \\u003cem\\u003eall indices are based on Model B (clinical\\u0026thinsp;+\\u0026thinsp;microbiome features).\\u003c/em\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e \\u003c/tfoot\\u003e \\u003c/table\\u003e\\u003c/div\\u003e \\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec28\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e3.9 Safety\\u003c/h2\\u003e \\u003cp\\u003eOral local reactions were the most common adverse events (mainly grade 2), and most participants did not require dose adjustment or interruption of treatment; visits or emergency visits due to adverse events were rare, and only 3 cases (1.47%) discontinued treatment because of adverse events (Table\\u0026nbsp;\\u003cspan refid=\\\"Tab7\\\" class=\\\"InternalRef\\\"\\u003e7\\u003c/span\\u003e).\\u003c/p\\u003e \\u003cp\\u003e \\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab7\\\" border=\\\"1\\\"\\u003e \\u003ccaption language=\\\"En\\\"\\u003e \\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 7\\u003c/div\\u003e \\u003cdiv class=\\\"CaptionContent\\\"\\u003e \\u003cp\\u003eAdverse events related to sublingual immunotherapy (n, %)\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/caption\\u003e \\u003ccolgroup cols=\\\"5\\\"\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c3\\\" colnum=\\\"3\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c4\\\" colnum=\\\"4\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c5\\\" colnum=\\\"5\\\"\\u003e\\u003c/div\\u003e \\u003cthead\\u003e \\u003ctr\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eEvent category\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eCases\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eMost severe WAO grade\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003eAny dose adjustment/interruption\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003eEmergency/urgent care\\u003c/p\\u003e \\u003c/th\\u003e \\u003c/tr\\u003e \\u003c/thead\\u003e \\u003ctbody\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eOral local reactions\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e82 (40.20%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e2\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e22 (26.83%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0 (0.00%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eGastrointestinal discomfort\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e21 (10.29%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e2\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e4 (19.05%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0 (0.00%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eSkin reactions\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e18 (8.82%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e2\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e3 (16.67%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0 (0.00%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eRespiratory symptoms\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e12 (5.88%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e2\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e4 (33.33%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e1 (8.33%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eSystemic reaction (non-shock)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e5 (2.45%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e3\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e4 (80.00%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e1 (20.00%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eAnaphylactic reaction\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e1 (0.49%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e4\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e1 (100.00%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e1 (100.00%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eDiscontinuation due to AE\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e3 (1.47%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e\\u0026mdash;\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e3 (100.00%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e1 (33.33%)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003c/tbody\\u003e \\u003c/colgroup\\u003e \\u003c/table\\u003e\\u003c/div\\u003e \\u003c/p\\u003e \\u003c/div\\u003e\"},{\"header\":\"4 Discussion\",\"content\":\"\\u003cp\\u003eAfter adjusting for age, sex, symptom burden, Artemisia-specific IgE, and polysensitization, the three gut features remained independently associated with clinical response. Effect directions were stable after error control, and sequencing quality, diary completeness, and adherence met analysis requirements. Baseline clinical differences were mild, supporting the attribution of associations to microbiome exposure rather than measurement bias. The composite index of butyrate-producing bacteria contributed most significantly. By increasing short-chain fatty acid levels, it may inhibit histone deacetylase activity, induce differentiation of regulatory T cells, and increase IL-10 production. This process may enhance tight junctions of the intestinal mucosal barrier, thereby reducing allergic inflammation along the gut\\u0026ndash;lung axis(\\u003cspan citationid=\\\"CR15\\\" class=\\\"CitationRef\\\"\\u003e15\\u003c/span\\u003e); an increase in Shannon diversity represents increased metabolic redundancy of the microbial community, making it easier to generate tolerance-related metabolites and limit pro-inflammatory colonization(\\u003cspan citationid=\\\"CR16\\\" class=\\\"CitationRef\\\"\\u003e16\\u003c/span\\u003e); a higher Prevotella/Bacteroides ratio usually means that fermentation pathways of complex carbohydrate substrates predominate, or that the sensitizing drive to allergens is reduced via remodeling of dendritic cell phenotypes(\\u003cspan citationid=\\\"CR17\\\" class=\\\"CitationRef\\\"\\u003e17\\u003c/span\\u003e), together explaining the differences in response probability under similar clinical backgrounds. Existing studies have repeatedly suggested that butyrate-producing bacteria and high diversity are positively associated in allergic diseases and in the remission of immunotherapy(\\u003cspan citationid=\\\"CR18\\\" class=\\\"CitationRef\\\"\\u003e18\\u003c/span\\u003e), but most are cross-sectional or post-treatment analyses, making it difficult to clarify causality versus concomitant changes; other studies have also suggested that the effect of Prevotella is context-dependent, and diet and host background can modulate the direction and magnitude of the effect(\\u003cspan citationid=\\\"CR19\\\" class=\\\"CitationRef\\\"\\u003e19\\u003c/span\\u003e). The present prospective design, with a single allergen and prespecified features as the core, under the premise of controllable confounding and quality assurance, advances the signal from the level of association to the attribute of prediction, providing empirical evidence for using baseline stool measurement for pretreatment precise stratification, and laying the foundation for subsequent incorporation of microecological indices into clinical decision-making.\\u003c/p\\u003e \\u003cp\\u003eOn the basis of the model containing only clinical variables, after introducing the three baseline gut features, the overall predictive performance was markedly improved, with discrimination increasing from moderate to better and probability calibration becoming more ideal; reclassification of individual probabilities was improved, with true responders more easily identified and fewer misclassified as responders. Across clinically acceptable thresholds, net benefit was consistently higher than that of reference strategies, suggesting the practical value of incorporating microbiome data into decision-making(\\u003cspan citationid=\\\"CR20\\\" class=\\\"CitationRef\\\"\\u003e20\\u003c/span\\u003e). Sensitivity analyses showed stable results across scenarios, unaffected by antibiotic exposure, threshold changes, or experimental batches. The incremental value brought by microbiome indices may arise from the gut community providing additional information for mucosal immune homeostasis that is not covered by traditional demographic, allergic and symptom variables(\\u003cspan citationid=\\\"CR21\\\" class=\\\"CitationRef\\\"\\u003e21\\u003c/span\\u003e). Butyrate-producing bacteria, diversity, and Prevotella/Bacteroides-dominated communities impact short-chain fatty acid production, metabolic redundancy, and substrate utilization. These factors affect dendritic cell phenotypes, Treg induction, and mucosal barrier integrity, influencing tolerance to sublingual allergens(\\u003cspan citationid=\\\"CR22\\\" class=\\\"CitationRef\\\"\\u003e22\\u003c/span\\u003e). These pieces of information are related to IgE titers or symptoms but are not redundant, providing complementary signals statistically, making individualized probability estimates closer to the true risk, and achieving a net improvement in the trade-off between overtreatment and undertreatment at intervention thresholds. Existing predictive frameworks usually rely on IgE, skin prick test intensity and symptom scales (\\u003cspan citationid=\\\"CR23\\\" class=\\\"CitationRef\\\"\\u003e23\\u003c/span\\u003e), with only limited improvement and lacking quantitative demonstration of clinical utility; previous studies of gut microbiota and immunotherapy outcomes have mostly been correlation analyses, with few systematically examined from the perspectives of discrimination, calibration, reclassification and net benefit. This evidence, after control of confounding and resampling-based correction, confirms that microbiome indices can achieve quantifiable predictive improvement without substantially increasing complexity, and present clinical decision value through net benefit evaluation, providing an implementation pathway for screening candidate populations, promoting adherence and optimizing follow-up intensity.\\u003c/p\\u003e \\u003cp\\u003eThe GM-SLIT score, based on L1 regularization, uses five baseline variables to form a nomogram with a single threshold. It effectively stratifies patients before treatment and shows stable classification performance after internal optimism correction. Follow-up records were complete and medication adherence was relatively high, and safety events were mainly mild to moderate, meeting the implementation conditions for deployment in outpatient settings. The weighting of the score is consistent with immunologic directions, with butyrate-producing bacteria and community diversity corresponding to a higher probability of tolerance, specific IgE and symptom burden corresponding to a lower probability of response, and the Prevotella/Bacteroides structural ratio reflecting differences in fermentation substrate utilization and short-chain fatty acid pathways(\\u003cspan citationid=\\\"CR24\\\" class=\\\"CitationRef\\\"\\u003e24\\u003c/span\\u003e), making the model output interpretable and facilitating communication. Implementation of the tool requires only a single baseline stool sample and routine 16S rRNA sequencing, with quality control, compositional transformation and standardization automatically completed in the background, and the interface presenting individualized probabilities and stratification prompts, on which basis decisions can be made about whether to prioritize initiation or persistence of immunotherapy, arrange more intensive follow-up and adherence interventions, and link with pollen-season management. The low-dimensional microbiome-integrated score complements IgE, skin prick tests, and symptom scales, offering better discrimination and calibration(\\u003cspan citationid=\\\"CR25\\\" class=\\\"CitationRef\\\"\\u003e25\\u003c/span\\u003e); compared with models with multiple features and complex algorithms but insufficient interpretability, the present tool achieves transferable clinical utility with lower complexity and uses decision curves to quantify the net improvement in the trade-off between overtreatment and undertreatment, reflecting the application value of microbiome indices in individualized stratification of immunotherapy.\\u003c/p\\u003e \\u003cp\\u003eThis single-center study had a moderate sample size, and external validation is needed. The models were corrected by internal resampling, so their generalizability remains uncertain. Microbiological testing used 16S rather than metagenomics and metabolomics, making it difficult to resolve strains and functional pathways and to directly quantify short-chain fatty acids. Dietary structure, PPIs and antibiotics, and environmental exposures may still produce residual confounding, although sensitivity analyses were performed. Outcomes depended on electronic diaries, with risks of self-report bias and missingness. The next step will be to carry out prospective validation in multicenter populations with different dietary and pollen exposure backgrounds, to unify sampling and bioinformatics pipelines, perform external calibration and model updating, and evaluate re-calibration of thresholds in different medical settings; to conduct metagenomics and targeted metabolomics in representative subsamples, pair measurements of stool and serum short-chain fatty acids, and construct functional pathway mediation models; to improve prospective recording of diet and medication exposures and verify them with prescription and dispensing data; and to carry out decision impact studies, embedding the score into outpatient workflows to examine real-world improvements in adherence, allocation of follow-up resources and patient benefit.\\u003c/p\\u003e\"},{\"header\":\"5 Conclusions\",\"content\":\"\\u003cp\\u003eBaseline gut microbiota characteristics can independently predict clinical response to sublingual immunotherapy in Artemisia pollen-induced allergic rhinitis, with butyrate-producing bacteria, diversity, and the Prevotella/Bacteroides ratio as key indicators. Combining microbiome indices with conventional clinical variables improves predictive model discrimination and calibration, offering net benefits at clinically acceptable thresholds. The low-dimensional, interpretable score based on prespecified features enables patient stratification before treatment and is feasible for outpatient use. It can guide treatment initiation and persistence, optimize follow-up intensity, and improve resource allocation for individualized immunotherapy. External validation across different regions and exposure backgrounds, as well as integration with functional omics, is needed to support clinical translation and identify modifiable pathways.\\u003c/p\\u003e\"},{\"header\":\"Declarations\",\"content\":\"\\u003cp\\u003e \\u003ch2\\u003eConflict of Interest\\u003c/h2\\u003e \\u003cp\\u003eThe authors declare no conflicts of interest in relation to this study.\\u003c/p\\u003e \\u003c/p\\u003e\\u003cp\\u003e \\u003ch2\\u003eEthical Statement\\u003c/h2\\u003e \\u003cp\\u003eThis study was approved by the institutional ethics committee of International Mongolian Hospital of Inner Mongolia (Approval No. 2021-033), and all participants provided written informed consent prior to any study procedures. The research adhered to the ethical principles of the Declaration of Helsinki, ensuring confidentiality and voluntary participation.\\u003c/p\\u003e \\u003c/p\\u003e\\u003cp\\u003e \\u003ch2\\u003eConsent for Publication\\u003c/h2\\u003e \\u003cp\\u003e Written informed consent was obtained from all participants for the publication of their data.\\u003c/p\\u003e \\u003c/p\\u003e\\u003cp\\u003e \\u003ch2\\u003eCompeting Interests\\u003c/h2\\u003e \\u003cp\\u003eThe authors declare that they have no competing interests.\\u003c/p\\u003e \\u003c/p\\u003e\\u003ch2\\u003eFunding\\u003c/h2\\u003e \\u003cp\\u003eThis study received no funding.\\u003c/p\\u003e\\u003ch2\\u003eAuthor Contribution\\u003c/h2\\u003e\\u003cp\\u003eF.W. and J.Y. conceptualized and designed the study. F.W. and L.B. were responsible for the methodology and statistical analysis. Y.N. and B. conducted patient recruitment, clinical assessments, and data collection. F.W. and J.Y. drafted the manuscript and performed critical revisions. All authors contributed to the data interpretation and approved the final manuscript.\\u003c/p\\u003e\\u003ch2\\u003eData Availability\\u003c/h2\\u003e\\u003cp\\u003eThe datasets supporting the conclusions of this article are included within the paper. Further data are available upon reasonable request from the corresponding author.\\u003c/p\\u003e\"},{\"header\":\"References\",\"content\":\"\\u003col\\u003e\\u003cli\\u003e\\u003cspan\\u003eZhang Y, Zhang L. Increasing Prevalence of Allergic Rhinitis in China. Allergy Asthma Immunol Res. 2019;11(2):156\\u0026ndash;69.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eL\\u0026oacute;pez JF, Bel Imam M, Satitsuksanoa P, Lems S, Yang M, Hwang YK, et al. Mechanisms and biomarkers of successful allergen-specific immunotherapy. Asia Pac Allergy. 2022;12(4):e45.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003ePitsios C. Allergen Immunotherapy: Biomarkers and Clinical Outcome Measures. J Asthma Allergy. 2021;14:141\\u0026ndash;8.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eBoonpiyathad T, Lao-Araya M, Chiewchalermsri C, Sangkanjanavanich S, Morita H. Allergic Rhinitis: What Do We Know About Allergen-Specific Immunotherapy? Front Allergy. 2021;2:747323.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eAmoroso C, Perillo F, Strati F, Fantini MC, Caprioli F, Facciotti F. The Role of Gut Microbiota Biomodulators on Mucosal Immunity and Intestinal Inflammation. Cells. 2020;9(5).\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eHe Z, Vadali VG, Szabady RL, Zhang W, Norman JM, Roberts B, et al. Increased diversity of gut microbiota during active oral immunotherapy in peanut-allergic adults. Allergy. 2021;76(3):927\\u0026ndash;30.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eZissler UM, Schmidt-Weber CB. Predicting Success of Allergen-Specific Immunotherapy. Front Immunol. 2020;11:1826.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eSteyerberg EW, Vickers AJ, Cook NR, Gerds T, Gonen M, Obuchowski N, et al. Assessing the performance of prediction models: a framework for traditional and novel measures. Epidemiology. 2010;21(1):128\\u0026ndash;38.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eLou H, Wang X, Wei Q, Zhao C, Xing Z, Zhang Q, et al. Artemisia Annua sublingual immunotherapy for seasonal allergic rhinitis: A multicenter, randomized trial. World Allergy Organ J. 2020;13(9):100458.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003ePfaar O, Bastl K, Berger U, Buters J, Calderon MA, Clot B, et al. Defining pollen exposure times for clinical trials of allergen immunotherapy for pollen-induced rhinoconjunctivitis - an EAACI position paper. Allergy. 2017;72(5):713\\u0026ndash;22.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003ePfaar O, Demoly P, van Gerth R, Bonini S, Bousquet J, Canonica GW, et al. Recommendations for the standardization of clinical outcomes used in allergen immunotherapy trials for allergic rhinoconjunctivitis: an EAACI Position Paper. Allergy. 2014;69(7):854\\u0026ndash;67.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003ePalathumpattu B, Pieper-F\\u0026uuml;rst U, Acikel C, Sahin H, Allekotte S, Singh J, et al. Correlation of the combined symptom and medication score with quality of life, symptom severity and symptom control in allergic rhinoconjunctivitis. Clin Transl Allergy. 2022;12(10):e12191.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eWorm M, Rak S, de Blay F, Malling HJ, Melac M, Cadic V, et al. Sustained efficacy and safety of a 300IR daily dose of a sublingual solution of birch pollen allergen extract in adults with allergic rhinoconjunctivitis: results of a double-blind, placebo-controlled study. Clin Transl Allergy. 2014;4(1):7.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eHjorth MF, Christensen L, Larsen TM, Roager HM, Krych L, Kot W, et al. Pretreatment Prevotella-to-Bacteroides ratio and salivary amylase gene copy number as prognostic markers for dietary weight loss. Am J Clin Nutr. 2020;111(5):1079\\u0026ndash;86.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eCorr\\u0026ecirc;a RO, Castro PR, Moser R, Ferreira CM, Quesniaux VFJ, Vinolo MAR, et al. Butyrate: Connecting the gut-lung axis to the management of pulmonary disorders. Front Nutr. 2022;9:1011732.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eMenni C, Zhu J, Le Roy CI, Mompeo O, Young K, Rebholz CM, et al. Serum metabolites reflecting gut microbiome alpha diversity predict type 2 diabetes. Gut Microbes. 2020;11(6):1632\\u0026ndash;42.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eAlemao CA, Budden KF, Gomez HM, Rehman SF, Marshall JE, Shukla SD, et al. Impact of diet and the bacterial microbiome on the mucous barrier and immune disorders. Allergy. 2021;76(3):714\\u0026ndash;34.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eDi Costanzo M, Carucci L, Berni Canani R, Biasucci G. Gut Microbiome Modulation for Preventing and Treating Pediatric Food Allergies. Int J Mol Sci. 2020;21:15.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eChung WSF, Walker AW, Bosscher D, Garcia-Campayo V, Wagner J, Parkhill J, et al. Relative abundance of the Prevotella genus within the human gut microbiota of elderly volunteers determines the inter-individual responses to dietary supplementation with wheat bran arabinoxylan-oligosaccharides. BMC Microbiol. 2020;20(1):283.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eVickers AJ, van Calster B, Steyerberg EW. A simple, step-by-step guide to interpreting decision curve analysis. Diagn Progn Res. 2019;3:18.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eDing G, Yang X, Li Y, Wang Y, Du Y, Wang M, et al. Gut microbiota regulates gut homeostasis, mucosal immunity and influences immune-related diseases. Mol Cell Biochem. 2025;480(4):1969\\u0026ndash;81.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eLuu M, Monning H, Visekruna A. Exploring the Molecular Mechanisms Underlying the Protective Effects of Microbial SCFAs on Intestinal Tolerance and Food Allergy. Front Immunol. 2020;11:1225.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eMormile I, Granata F, Detoraki A, Pacella D, Della Casa F, De Rosa F et al. Predictive Response to Immunotherapy Score: A Useful Tool for Identifying Eligible Patients for Allergen Immunotherapy. Biomedicines. 2022;10(5).\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eDi Costanzo M, De Paulis N, Biasucci G, Butyrate. A Link between Early Life Nutrition and Gut Microbiome in the Development of Food Allergy. Life (Basel). 2021;11(5).\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eXie S, Jiang S, Zhang H, Wang F, Liu Y, She Y, et al. Prediction of sublingual immunotherapy efficacy in allergic rhinitis by serum metabolomics analysis. Int Immunopharmacol. 2021;90:107211.\\u003c/span\\u003e\\u003c/li\\u003e\\u003c/ol\\u003e\"}],\"fulltextSource\":\"\",\"fullText\":\"\",\"funders\":[],\"hasAdminPriorityOnWorkflow\":false,\"hasManuscriptDocX\":true,\"hasOptedInToPreprint\":true,\"hasPassedJournalQc\":\"\",\"hasAnyPriority\":false,\"hideJournal\":true,\"highlight\":\"\",\"institution\":\"\",\"isAcceptedByJournal\":false,\"isAuthorSuppliedPdf\":false,\"isDeskRejected\":\"\",\"isHiddenFromSearch\":false,\"isInQc\":false,\"isInWorkflow\":false,\"isPdf\":false,\"isPdfUpToDate\":true,\"isWithdrawnOrRetracted\":false,\"journal\":{\"display\":true,\"email\":\"info@researchsquare.com\",\"identity\":\"researchsquare\",\"isNatureJournal\":false,\"hasQc\":true,\"allowDirectSubmit\":true,\"externalIdentity\":\"\",\"sideBox\":\"\",\"snPcode\":\"\",\"submissionUrl\":\"/submission\",\"title\":\"Research Square\",\"twitterHandle\":\"researchsquare\",\"acdcEnabled\":true,\"dfaEnabled\":false,\"editorialSystem\":\"\",\"reportingPortfolio\":\"\",\"inReviewEnabled\":false,\"inReviewRevisionsEnabled\":true},\"keywords\":\"gut microbiota, sublingual immunotherapy, allergic rhinitis, predictive model\",\"lastPublishedDoi\":\"10.21203/rs.3.rs-8234895/v1\",\"lastPublishedDoiUrl\":\"https://doi.org/10.21203/rs.3.rs-8234895/v1\",\"license\":{\"name\":\"CC BY 4.0\",\"url\":\"https://creativecommons.org/licenses/by/4.0/\"},\"manuscriptAbstract\":\"\\u003cp\\u003e\\u003cb\\u003eBackground\\u003c/b\\u003e\\u003c/p\\u003e \\u003cp\\u003eArtemisia pollen allergic rhinitis is a major health burden, with sublingual immunotherapy showing variable effectiveness. This study explores the potential of gut microbiota as a biomarker to better predict treatment outcomes. Current predictive methods, such as IgE levels, skin prick tests, and symptom scales, often fail to accurately predict treatment outcomes.\\u003c/p\\u003e\\u003cp\\u003e\\u003cb\\u003eObjective\\u003c/b\\u003e\\u003c/p\\u003e \\u003cp\\u003eTo evaluate the predictive value of baseline gut microbiome features for sublingual immunotherapy response and develop a practical clinical score for patient stratification.\\u003c/p\\u003e\\u003cp\\u003e\\u003cb\\u003eMethods\\u003c/b\\u003e\\u003c/p\\u003e \\u003cp\\u003eA single-center prospective cohort study enrolled 204 participants. Pretreatment stool samples were analyzed using 16S rRNA V3\\u0026ndash;V4 sequencing to assess Shannon diversity, the proportion of butyrate-producing bacteria, and the Prevotella/Bacteroides ratio. Three models were developed, with Model A based on clinical variables, Model B incorporating microbiome features, and Model C using L1 regularization for feature selection. Model performance was evaluated through AUC (DeLong), calibration intercept α and slope β (Bootstrap), NRI/IDI, and decision curve analysis, with Model C validated internally.\\u003c/p\\u003e\\u003cp\\u003e\\u003cb\\u003eResults\\u003c/b\\u003e\\u003c/p\\u003e \\u003cp\\u003eThe median improvement in CSMS over the peak 6-week pollen season was 32.83% (95% CI 28.87\\u0026ndash;36.61), with a response rate of 54.41% (95% CI 47.56\\u0026ndash;61.10). In Model B, microbiome features significantly predicted response, with ORs of 1.59 for butyrate-producing bacteria, 1.43 for the Prevotella/Bacteroides ratio, and 1.33 for Shannon diversity. Model B increased AUC from 0.71 to 0.79 (P\\u0026thinsp;=\\u0026thinsp;0.021) and showed improved calibration (α=\\u0026minus;0.03; β\\u0026thinsp;=\\u0026thinsp;0.98). Model C, with a threshold of p\\u0026thinsp;\\u0026ge;\\u0026thinsp;0.62, had a sensitivity of 77.48%, specificity of 72.04%, and AUC of 0.78.\\u003c/p\\u003e\\u003cp\\u003e\\u003cb\\u003eConclusions\\u003c/b\\u003e\\u003c/p\\u003e \\u003cp\\u003eBaseline gut microbiome features enhance the prediction of sublingual immunotherapy outcomes. The interpretable, low-dimensional score offers a practical tool for patient stratification and decision-making, with potential for further validation and clinical application.\\u003c/p\\u003e\",\"manuscriptTitle\":\"Predictive Value of Baseline Gut Microbiome Characteristics for the Response in the Next Pollen Season After Sublingual Immunotherapy in Artemisia Pollen–Induced Allergic Rhinitis: A Single-Center Prospective Cohort Study.\",\"msid\":\"\",\"msnumber\":\"\",\"nonDraftVersions\":[{\"code\":1,\"date\":\"2026-01-16 09:16:46\",\"doi\":\"10.21203/rs.3.rs-8234895/v1\",\"editorialEvents\":[{\"type\":\"communityComments\",\"content\":0}],\"status\":\"published\",\"journal\":{\"display\":true,\"email\":\"info@researchsquare.com\",\"identity\":\"researchsquare\",\"isNatureJournal\":false,\"hasQc\":true,\"allowDirectSubmit\":true,\"externalIdentity\":\"\",\"sideBox\":\"\",\"snPcode\":\"\",\"submissionUrl\":\"/submission\",\"title\":\"Research Square\",\"twitterHandle\":\"researchsquare\",\"acdcEnabled\":true,\"dfaEnabled\":false,\"editorialSystem\":\"\",\"reportingPortfolio\":\"\",\"inReviewEnabled\":false,\"inReviewRevisionsEnabled\":true}}],\"origin\":\"\",\"ownerIdentity\":\"adf7979e-00ac-4d39-9433-ccf3326a90b0\",\"owner\":[],\"postedDate\":\"January 16th, 2026\",\"published\":true,\"recentEditorialEvents\":[],\"rejectedJournal\":[],\"revision\":\"\",\"amendment\":\"\",\"status\":\"posted\",\"subjectAreas\":[],\"tags\":[],\"updatedAt\":\"2026-01-21T06:15:24+00:00\",\"versionOfRecord\":[],\"versionCreatedAt\":\"2026-01-16 09:16:46\",\"video\":\"\",\"vorDoi\":\"\",\"vorDoiUrl\":\"\",\"workflowStages\":[]},\"version\":\"v1\",\"identity\":\"rs-8234895\",\"journalConfig\":\"researchsquare\"},\"__N_SSP\":true},\"page\":\"/article/[identity]/[[...version]]\",\"query\":{\"redirect\":\"/article/rs-8234895\",\"identity\":\"rs-8234895\",\"version\":[\"v1\"]},\"buildId\":\"XKTyCvWXoU3ODBz1xrDgd\",\"isFallback\":false,\"isExperimentalCompile\":false,\"dynamicIds\":[84888],\"gssp\":true,\"scriptLoader\":[]}","source_license":"CC-BY-4.0","license_restricted":false}