Early-onset wheeze trajectories in infants: the Phenotyping of Wheezing Infants (P-WIN) study

Background: Recurrent wheeze in infancy is common; although symptoms often resolve, some children develop persistent disease. To better capture clinical heterogeneity, we analyzed wheeze trajectories based on symptom frequency rather than simply recording presence or absence. Objective: To identify distinct wheeze phenotypes in 1-year-old children with recurrent wheeze based on longitudinal wheezing frequency trajectories. Methods: We conducted a two-year, multicenter prospective cohort study involving children aged 12–23 months with recurrent wheeze at 27 sites in Japan. Monthly caregiver-reported wheeze frequency was collected, and trajectories were classified using latent class growth analysis. Clinical characteristics, environmental exposures, and biomarkers were assessed at enrollment and age 3. Ordinal and binary logistic regression analyses were performed to identify risk and protective factors. Results: Among 253 enrolled children, 219 completed follow-up. Four trajectories were identified: Early-Resolving (24.2%), Low-Frequency with Mid-Peak (57.1%), Persistent High-Frequency (12.8%), and Late- Peaking High-Frequency (5.9%). The latter two groups showed greater symptom burden, including more frequent corticosteroid use and interference with daily activities. Ordinal logistic regression showed that parental allergic rhinitis and pet ownership were associated with lower odds of more severe trajectories. Binary logistic regression comparing high- (Clusters 3–4) versus low- frequency (Clusters 1–2) groups revealed parental smoking as a strong risk factor (OR 5.49), while allergic rhinitis (OR 0.12) and pet ownership (OR 0.11) remained protective. Conclusions: High-frequency wheeze trajectories were linked to greater clinical burden. Early identification of at-risk children and targeted environmental interventions—particularly avoidance of passive smoking—may reduce morbidity in early-onset recurrent wheeze. Early-onset wheeze trajectories in infants: the Phenotyping of Wheezing Infants (P-WIN) study Rei Kanai1)2)3), Mizuho Nagao1)2)3), Yasunori Sato4), Jun Atsuta5), Chiho Tatsumoto6), Tadashi Matsuda7), Yohei Watanabe 8), Yoko Miura 9), Noriyuki Yanagida 9), Shigeru Suga 2)3), Kiyosu Taniguchi 2)3), Takao Fujisawa1)2) 1. Allergy Center, NHO Mie National Hospital, Tsu, Japan 2. Department of Pediatrics, NHO Mie National Hospital, Tsu, Japan 1 Posted on 3 Jun 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.174895576.62817881/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. 3. Division of Child Medical Health and Development Department of Molecular and Experimental Medicine Mie University Graduate School of Medicine, Tsu, Japan 4. Department of Biostatistics, Keio University Graduate School of Medicine, Tokyo, Japan 5. Atsuta Pediatric Clinic Allergy Clinic, Tsu, Japan 6. Aozora children’s hospital, Kagoshima, Japan 7. Matsuda Pediatric Clinic, Kuwana, Japan 8. Department of Pediatrics, Sendai Medical Center, Sendai, Japan 9. NHO Sagamihara National Hospital, Kanagawa, Japan Running title: Wheeze Trajectories in One-Year-Olds Corresponding author: Takao Fujisawa, MD, PhD NHO Mie National Hospital 357 Osato-kubota, Tsu, Mie 514-0125, Japan [email protected] Word count:3,462, Number of Table;4, Number of figures; 2 Supporting Information Tables; 5, Supporting Information Figures; 2 Conflict of Interest: All authors declare that they have no conflicts of interest related to this work. Financial Support: This study was supported by a research grant from the National Hospital Organization of Japan.

Preschool wheeze, Trajectory analysis, Latent class growth analysis, Asthma phenotypes, Envi- ronmental risk factors

Background: Recurrent wheeze in infancy is common; although symptoms often resolve, some children develop persistent disease. To better capture clinical heterogeneity, we analyzed wheeze trajectories based on symptom frequency rather than simply recording presence or absence.

To identify distinct wheeze phenotypes in 1-year-old children with recurrent wheeze based on longitudinal wheezing frequency trajectories.

We conducted a two-year, multicenter prospective cohort study involving children aged 12–23 months with recurrent wheeze at 27 sites in Japan. Monthly caregiver-reported wheeze frequency was collected, and trajectories were classified using latent class growth analysis. Clinical characteristics, envi- ronmental exposures, and biomarkers were assessed at enrollment and age 3. Ordinal and binary logistic regression analyses were performed to identify risk and protective factors.

Among 253 enrolled children, 219 completed follow-up. Four trajectories were identified: Early- Resolving (24.2%), Low-Frequency with Mid-Peak (57.1%), Persistent High-Frequency (12.8%), and Late- Peaking High-Frequency (5.9%). The latter two groups showed greater symptom burden, including more frequent corticosteroid use and interference with daily activities. Ordinal logistic regression showed that parental allergic rhinitis and pet ownership were associated with lower odds of more severe trajectories. Binary logistic regression comparing high- (Clusters 3–4) versus low-frequency (Clusters 1–2) groups revealed parental smoking as a strong risk factor (OR 5.49), while allergic rhinitis (OR 0.12) and pet ownership (OR 0.11) remained protective.

High-frequency wheeze trajectories were linked to greater clinical burden. Early identification of at-risk children and targeted environmental interventions—particularly avoidance of passive smoking—may reduce morbidity in early-onset recurrent wheeze.

2 Posted on 3 Jun 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.174895576.62817881/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. Recurrent wheezing in infancy and early childhood is a common clinical concern, affecting up to 30% of children under the age of three. It places a substantial burden on healthcare systems and is a frequent source of anxiety for caregivers.1 Although many children experience spontaneous resolution of symptoms, a considerable subset develop persistent wheezing or progress to asthma, with potential long-term respiratory and systemic consequences. Longitudinal studies have demonstrated that impaired lung function in early life is associated not only with asthma, but also with adverse long-term outcomes, including chronic obstructive pulmonary disease (COPD), cardiovascular and metabolic comorbidities, and increased all-cause mortality. 2-5 Children with recurrent wheeze are more likely to follow suboptimal lung function trajectories, characterized by failure to achieve peak lung function in early adulthood and accelerated decline thereafter.. 6 These findings highlight the critical need to identify high-risk phenotypes during the preschool years, when timely intervention may still modify the course of disease. Several large-scale birth cohort studies—including the Tucson Children’s Respiratory Study, 7,8 the Avon Longitudinal Study of Parents and Children (ALSPAC), 9 the PIAMA birth cohort, 10 and the Tokyo Chil- dren’s Health, Illness and Development (T-CHILD) Study 11 have classified early wheeze phenotypes using trajectory- or cluster-based approaches. These efforts have consistently identified subtypes such as transient early wheeze, persistent wheeze, and intermediate-onset wheeze. Importantly, all of these studies enrolled children from the general population and followed them longitudinally from birth through school age and beyond, while focusing on early wheezing patterns primarily during the preschool years. These efforts have provided valuable insights into the natural history and phenotypic heterogeneity of early-life wheezing at the population level. In parallel, recent advances in data-driven analytical methods—such as latent class and trajectory modeling— have enabled more refined phenotypic classifications of preschool wheeze. 12These techniques have revealed a broader spectrum of symptom trajectories than was previously recognized. However, many prior studies have relied on binary data indicating the presence or absence of wheeze at selected time points, potentially overlooking important variations in symptom burden and timing. In contrast, trajectory modeling based on wheeze frequency could offer greater resolution in characterizing disease expression and severity. To address these gaps, we conducted the Phenotyping of Wheezing Infants (P-WIN) study, a prospective cohort study targeting children who had already developed recurrent wheeze by age one and followed them over a two-year period. We applied latent class growth analysis (LCGA) to characterize longitudinal patterns of wheeze frequency at ages 1, 2, and 3 years. This trajectory-based, data-driven approach was designed to capture early heterogeneity in symptom patterns within this high-risk group. Our aim was to identify distinct early wheeze trajectories and to provide insights that could ultimately inform prediction of later respiratory outcomes based on early clinical manifestations.

Study Design and Participants This study is part of the Phenotyping of Wheezing Infants (P-WIN) study, a prospective, multicenter cohort conducted at 27 clinical sites across Japan, including both primary care clinics and allergy specialty centers (Supporting Information Figure E1 and Table E5). Participant enrollment was carried out between July 2016 and January 2020. Eligible participants were children aged 12 to 23 months with a documented history of at least one wheezing episode occurring more than one month before enrollment. At registration, the presence of expiratory wheeze lasting over 24 hours was confirmed by a study physician. Children were excluded if wheezing was attributable to underlying conditions such as congenital tracheal stenosis, tracheomalacia, or cardiac anomalies; a history of perinatal respiratory disease requiring mechanical ventilation; or diagnosed immunodeficiency. Data Collection At enrollment and at the final evaluation at age 3, clinical data were collected through physical examinations, physician-completed questionnaires, and blood sampling. Monthly follow-up data were obtained using elec- 3 Posted on 3 Jun 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.174895576.62817881/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. tronic patient-reported outcomes (ePRO) submitted by caregivers. At age 2, an interim clinical assessment was conducted; due to the COVID-19 pandemic, this was performed via telephone interviews when in-person visits were not feasible. Caregivers provided detailed information on demographics, perinatal history, infant feeding practices, eczema, respiratory symptoms, family history of allergic disease, environmental exposures (e.g., parental smoking, pet ownership), and childcare settings. Throughout the follow-up period, frequency of wheezing, wheezing- related symptoms, treatments, and episodes of respiratory infection were documented. For respiratory infec- tions, caregiver-reported episodes described as “bronchiolitis” were classified as indicative of lower respiratory tract infection. Wheezing severity during follow-up was assessed based on caregiver reports of interference with daily ac- tivities, use of oral corticosteroids, and hospitalizations due to wheezing. Diagnoses of food allergy, allergic rhinitis, and atopic dermatitis were based on clinical evaluations and physician-completed questionnaires. Study physicians at each site performed blood tests for complete blood count, white blood cell differential, total IgE, and allergen-specific IgE antibodies to house dust mites (Dermatophagoides pteronyssinus and D. farinae), Japanese cedar, dog, cat, egg white, and milk, measured using the ImmunoCAP system (Thermo Fisher Scientific, Uppsala, Sweden). Statistical Analysis Latent class growth analysis (LCGA) was employed to identify distinct wheeze trajectory clusters over the two-year follow-up period based on monthly symptom reports. The optimal number of classes was determined using the Bayesian Information Criterion (BIC) and Bayes factors. Age-related changes in wheeze frequency within each trajectory group were evaluated using a linear mixed-effects model, followed by Tukey’s multiple comparison test for post hoc pairwise comparisons. Categorical variables were compared using Fisher’s exact test, and continuous variables using the Kruskal– Wallis test. After identifying trajectory clusters by LCGA, ordinal logistic regression was conducted to assess associations between cluster membership and potential risk or protective factors, treating the clusters as ordered categories. In addition, binary logistic regression was performed by grouping clusters into higher- and lower-frequency wheeze trajectories, to further examine factors associated with more severe patterns. All statistical tests were two-sided, with a significance threshold set at p < 0.05. Analyses were exploratory in nature, aiming to identify potential predictors of wheeze trajectory phenotypes without prespecified hy- potheses. LCGA was performed using SAS version 9.4 (SAS Institute, Cary, NC). All other statistical analyses were conducted using JMP version 17.0 (SAS Institute) and GraphPad Prism version 9.3.1 (Graph- Pad Software, San Diego, CA). Written informed consent was obtained from all parents or legal guardians prior to participation. Ethics Statement This study was approved by the Ethics Committee of NHO Mie National Hospital (approval number: H30- 0122001).

Characteristics of the study participants Among the 253 children enrolled in the study, 34 discontinued follow-up during the two-year observation period, yielding a final analytic sample of 219 participants. Baseline characteristics of the cohort are sum- marized in Table 1. The majority of participants were male. Preterm birth and low birth weight were not notably overrepresented. More than half of the participants had a history of infantile eczema, while approx- imately 30% had current allergic conditions such as atopic dermatitis or food allergy. A similar proportion had a family history of asthma or other allergic diseases. 4 Posted on 3 Jun 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.174895576.62817881/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. The prevalence of parental smoking in this cohort exceeded national estimates reported by the Ministry of Health, Labour and Welfare of Japan, which found smoking rates among individuals in their twenties to be 20.6% for men and 5.2% for women. In our cohort, the rates were 37.9% for fathers and 6.4% for mother. (https://www.mhlw.go.jp/content/10900000/001338334.pdf) Pet ownership was reported in 35.6% of participants—substantially higher than the rates reported in the Japan Environment and Children’s Study (JECS), which found ownership rates of 15.1% for dogs and 8.1% for cats. 13 Although specific pet types were not recorded in our cohort, the total ownership rate suggests a greater level of exposure compared to national estimates. Identification of wheeze trajectory classes Using LCGA Wheezing episode frequencies during the year prior to enrollment, the first year, and the second year of follow-up were analyzed using latent class growth analysis (LCGA). To determine the optimal number of trajectory classes, models with increasing numbers of latent classes were sequentially evaluated, beginning with a one-class model. A four-class solution was selected based on the lowest Bayesian Information Criterion (BIC) value and the highest Bayes factor (Supporting Information Table E1). The four identified trajectory clusters are illustrated in Figure 1. • Cluster 1: Early-Resolving Wheeze (ER) — This group exhibited a marked decline in wheezing fre- quency over the two-year period. • Cluster 2: Low-Frequency with Mid-Peak Wheeze (LFMP) — Participants showed low-frequency wheezing at age 1, a peak at age 2, and a subsequent decline by age 3. • Cluster 3: Persistent High-Frequency Wheeze (PHF) — Children in this group experienced consistently high wheezing frequency from enrollment through age 3. • Cluster 4: Late-Peaking High-Frequency Wheeze (LPHF) — Wheezing frequency remained low at ages 1 and 2, followed by a sharp increase at age 3, ultimately exceeding that of all other clusters at that time point. To further characterize these patterns, wheezing frequencies were compared across clusters at baseline (age 1) and at age 3 (Figure 2). At baseline, Cluster 3 (PHF) exhibited the highest wheeze frequency, which remained significantly elevated at age 3 compared to the other clusters ( p < .001). In contrast, Cluster 4 (LPHF) began with relatively low wheeze frequency but showed a significant increase by age 3, resulting in the greatest wheeze burden at that time point ( p < .0001 vs. Clusters 1–3). Cluster 1 (ER) showed moderately low wheeze frequency at baseline, followed by a marked reduction by age 3, ultimately exhibiting the lowest frequency among all clusters ( p < .0001) Clinical Relevance of Wheeze Trajectory Clusters To evaluate the clinical relevance of the identified wheeze trajectories, we analyzed indicators of wheeze severity and potential lower respiratory tract involvement during the second year of follow-up (Table 2). Significant differences in severity-related features were observed across the four clusters. Interference with daily activities was most frequently reported in Cluster 3 (50.0%) and Cluster 4 (46.2%), indicating a greater disease burden compared to Cluster 1 (17.0%) and Cluster 2 (28.8%) (p = .0091). The use of oral corticosteroids was highest in Cluster 4 (23.1%), while no children in Cluster 1 required corticosteroid treatment (p = .0099). In addition, the prevalence of caregiver-reported episodes of “bronchiolitis”—used here as a proxy for pos- sible lower respiratory tract infection—was significantly higher in Cluster 4 (23.1%) and Cluster 3 (17.9%) compared to Clusters 1 and 2 ( p = .0032). Factors associated with Wheeze Trajectory Clusters To investigate background factors associated with each wheeze trajectory cluster, we examined differences in perinatal, familial, and environmental characteristics across the four clusters (Supporting Information Table 5 Posted on 3 Jun 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.174895576.62817881/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. E2). Several variables demonstrated significant variation among clusters. For example, maternal antibiotic use during pregnancy was more frequently reported in Cluster 4 (23.1%) than in the other clusters (p = .039). The prevalence of parental allergic rhinitis and pet ownership was significantly lower in Clusters 3 and 4, whereas parental smoking, particularly in Cluster 4 (76.9%), was markedly more common (p < .01 for all comparisons). These findings suggest that specific perinatal exposures and household environmental factors may contribute to the development of distinct early wheeze phenotypes.. Sensitization and other biomarkers At age 3, Japanese cedar pollen sensitization was more prevalent in Clusters 3 and 4 compared to Clusters 1 and 2 (p = .015). Sensitization to other allergens, total IgE, and eosinophil counts did not significantly differ across clusters (Supporting Information Table E3). Use of Asthma Controller Medications Across Clusters The use of leukotriene receptor antagonists (LTRAs) was high at enrollment (up to 78.6%) and during the first year (up to 100%), with no significant differences among clusters. In the second year, LTRA use remained high overall but differed significantly among clusters (p = .027). Although usage appeared lower in Cluster 1 (69.8%), no post hoc comparisons were performed to determine which clusters contributed to the observed difference. Inhaled corticosteroid (ICS) use ranged from 17.7% to 25.0% at enrollment, 38.5% to 56.8% in the first year, and 30.2% to 49.6% in the second year, with no significant differences across clusters (Supporting Information Table E4). Risk Factors Associated with Severe Wheeze Trajectories To identify factors associated with more severe wheeze phenotypes, ordinal logistic regression was performed, treating the trajectory clusters as an ordinal outcome from least to most severe (Cluster 1 to Cluster 4; Table 3). In addition to variables found to be statistically significant in univariate analyses, several factors were included in the multivariate model based on clinical relevance and exploratory rationale. These included use of iron supplements, which exhibited noticeable differences across clusters despite not reaching statistical significance, as well as sex and parental history of asthma, both of which have been previously reported as relevant predictors in the literature. The analysis revealed that parental allergic rhinitis (either parent) and pet ownership were significantly associated with lower odds of assignment to more severe clusters. Both variables had odds ratios <1 with statistical significance, indicating a protective association with milder wheeze trajectories. To further clarify these associations, we conducted a binary logistic regression, grouping Clusters 1 and 2 as the low-frequency wheeze group and Clusters 3 and 4 as the high-frequency group (Table 4). The same set of covariates was used in this model. Consistent with the ordinal model, both parental allergic rhinitis (OR: 0.12, 95% CI: 0.05–0.28, p < .0001) and pet ownership (OR: 0.11, 95% CI: 0.03–0.32, p < .0001) were associated with significantly reduced odds of belonging to the high-frequency group. In contrast, parental smoking was strongly associated with increased odds of high-frequency wheeze (OR: 5.49, 95% CI: 2.30– 13.12, p = .0001). A trend toward increased risk was also observed for antibiotic use during pregnancy (OR: 4.22, p = .069), although this did not reach statistical significance.

In this study, we aimed to identify wheeze trajectories in infants with early-onset recurrent wheezing, fo- cusing specifically on one-year-old children with a history of two or more documented episodes. Previous birth cohort studies have demonstrated that early-onset persistent wheeze is associated with subsequent asthma development and impaired lung growth. 7-11 To further characterize early wheeze phenotypes— beyond the binary classification of symptom presence or absence used in earlier studies—we conducted a detailed two-year follow-up based on wheeze frequency. This approach identified four distinct trajectories: Early-Resolving Wheeze (24.2%), Low-Frequency with Mid-Peak Wheeze (57.1%), Persistent High-Frequency 6 Posted on 3 Jun 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.174895576.62817881/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. Wheeze (12.8%), and Late-Peaking High-Frequency Wheeze (5.9%). Analysis of symptom severity during the second year revealed that the latter two clusters were associated with more severe clinical manifestations, suggesting a higher risk of unfavorable long-term outcomes. A key strength of this study is the integration of prospectively collected, monthly caregiver-reported symptom data with a trajectory-based analytical approach. Although wheeze trajectories were modeled using annual summary measures, the underlying data were derived from monthly electronic patient-reported outcomes (ePROs). This design minimized recall bias and ensured consistent, timely documentation of wheezing symptoms throughout the follow-up period. We intentionally chose not to model the monthly data directly due to seasonal fluctuations in wheeze frequency. 14 Instead, wheeze frequencies were aggregated over 12- month periods to generate robust and comparable estimates of symptom burden across participants. To leverage this structured longitudinal dataset, we applied Latent Class Growth Analysis (LCGA), a model-based clustering method suitable for identifying distinct developmental trajectories. Unlike previ- ous studies—such as those using Longitudinal Latent Class Analysis (LLCA) 10 or Group-Based Trajectory Modeling (GBTM), 11 which typically relied on binary symptom reports collected approximately once per year—our approach utilized monthly wheeze frequency data, offering greater phenotypic resolution. This more granular symptom tracking enabled the identification of nuanced patterns, including mid-course exac- erbations and late-onset increases in wheezing, which may not be readily detectable using less frequent or dichotomous measures. Although long-term outcome data are not yet available for our cohort, children classified into high-frequency wheeze trajectories already demonstrated greater clinical severity, as indicated by more frequent use of oral corticosteroids and greater interference with daily activities. These findings suggest that such children may represent a high-risk subgroup with underlying pathophysiological alterations. Previous research has shown that severe or persistent preschool wheezing may be accompanied by early signs of airway remodeling. 15-17 Airway remodeling, in turn, has been linked to impaired lung function, 18 placing affected children on a trajectory toward long-term respiratory morbidity that may extend through adolescence and into adulthood,6 and potentially persist into later life. 4 Our classification—based on wheeze frequency rather than binary symptom presence—may thus be useful not only for guiding short-term management, but also for identifying children who could benefit from intensified long-term surveillance and early preventive strategies. Despite efforts to identify high-risk wheeze phenotypes early in life, few interventions have demonstrated lasting effects. Inhaled corticosteroids (ICS) are commonly used to alleviate symptoms in preschool children, but their benefits generally do not persist after discontinuation. 19,20 In our cohort, ICS was prescribed in approximately 30–50% of cases, primarily in response to symptom exacerbation rather than as a preventive measure. Leukotriene receptor antagonists (LTRAs) were more frequently used, reflecting prevailing clinical practices in Japan. However, neither treatment appeared to influence wheeze trajectory. Biologic agents such as dupilumab, which target upstream pathways of type 2 inflammation, have emerged as promising disease-modifying therapies. 21-23These agents improve symptom control in severe cases and may potentially prevent airway remodeling and progression of the atopic march. However, supporting evidence in preschool populations remains limited. Further research is needed to explore whether biologics could modify the course of early-life wheezing disorders. Numerous studies have demonstrated that early-life exposure to tobacco smoke—including passive smoking— is associated with adverse respiratory outcomes.24,25 A systematic review of trajectory-specific risk factors for childhood wheeze reported that tobacco exposure increases the risk across multiple phenotypes, including early-transient, early-persistent, and late-onset trajectories. 26 Our findings reinforce this well-established evidence by demonstrating that even among children who had already developed recurrent wheeze by age one, passive smoking remained significantly associated with greater symptom severity. Moreover, parental smoking rates in our cohort exceeded national averages for adults in Japan across most clusters, with the exception of maternal smoking in Cluster 2. Although our study did not include a control group and statistical testing was not performed, the higher rates of parental smoking compared to national estimates raise the possibility that early-life exposure to tobacco smoke may have contributed to the development of recurrent 7 Posted on 3 Jun 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.174895576.62817881/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. wheeze in some children who might not otherwise have met the inclusion criteria for this cohort. This remains a hypothesis-generating observation that warrants further investigation in appropriately controlled studies.. One proposed mechanism by which tobacco exposure contributes to wheezing and asthma is through epi- genetic modification. Both maternal 27,28 and paternal29 tobacco smoke exposure have been associated with increased DNA methylation in immune-related genes, potentially leading to immune dysregulation and air- way inflammation. Importantly, tobacco exposure represents a modifiable risk factor. Several studies have demonstrated that family-based smoking cessation interventions can reduce asthma-related morbidity in children.30,31 Accordingly, pediatricians should routinely assess parental smoking during clinical encounters and provide targeted referrals to evidence-based cessation programs when indicated. 32 In contrast, parental asthma was not significantly associated with trajectory classification, while parental allergic rhinitis appeared to be a protective factor against high-frequency wheeze. These findings differ from previous reports in general populations, where both parental asthma and allergic rhinitis are typically considered risk factors for childhood wheezing and asthma development. 25 One possible explanation for this discrepancy lies in the specific characteristics of our cohort, which included only children who had already experienced recurrent wheezing by age one. This clinical homogeneity may have unmasked differential influences of asthma- versus rhinitis-related genetic or immunologic pathways—effects that may be obscured in more heterogeneous general population studies. Similarly, pet ownership was associated with a lower risk of persistent high-frequency wheeze. Although previous studies have reported conflicting results regarding the impact of pet exposure,25 the protective effect observed in our cohort may reflect immune-modulatory or microbiome-mediated mechanisms, 33 particularly within a population already exhibiting early-onset wheezing. These findings underscore the need for further investigation into host–environment interactions in high-risk pediatric subgroups. This study has several limitations. First, although the multicenter design contributed to a certain degree of population diversity, the moderate sample size may limit the generalizability of our findings. Second, while the use of monthly ePROs helped reduce recall bias, wheeze frequency was analyzed using annual aggregates, which may have masked finer-grained temporal variations in symptom patterns. Third, long- term outcomes such as lung function and asthma diagnosis beyond age 3 were not available at the time of analysis. Nonetheless, as this cohort is being followed prospectively, future longitudinal data will be critical to validating the clinical significance of the wheeze trajectories identified in this early period. In conclusion, this study identified four distinct symptom trajectories among young children with recurrent wheeze using latent class growth analysis, revealing meaningful differences in clinical severity and associated environmental factors. High-frequency trajectories were linked to greater disease burden, highlighting the importance of early identification and sustained clinical monitoring. These findings provide a clinically useful framework for phenotypic stratification in preschool wheeze and lay the groundwork for future research focused on prognosis and early intervention strategies.

We gratefully acknowledge Mr. Shinobu Tanimura of the Department of Clinical Research, Mie National Hospital, for his exceptional contribution to data collection and database management. We also thank Ms. Kumiko Ohta for her valuable administrative assistance throughout the study. During manuscript preparation, the authors used ChatGPT (GPT-4, OpenAI) to assist with English language refinement and structural editing. All content was reviewed and revised by the authors to ensure scientific accuracy. The authors take full responsibility for the integrity and final version of the manuscript. Key Message: Using longitudinal wheeze frequency, we identified four distinct trajectories among one-year-old children with recurrent wheeze. High-frequency trajectories were associated with greater clinical severity and treatment 8 Posted on 3 Jun 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.174895576.62817881/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. burden. Parental smoking increased the risk of severe trajectories, whereas parental allergic rhinitis and pet ownership were protective. Author Contributions Rei Kanai : Writing-Original Draft Preparation (equal), Data Curation (equal), Formal Analysis (support- ing), Visualization (supporting) Mizuho Nagao : Conceptualization (lead), Funding Acquisition (lead), Project Administration (Lead), Data curation (equal), Formal Analysis (supporting) Yasunori Sato : Methodology (lead), Formal Analysis (lead) Jun Atsuta: Investigation (equal) Chiho Tatsumoto: Investigation (equal) Tadashi Matsuda: Investigation (equal) Yohei Watanabe: Investigation (equal) Yoko Miura : Investigation (equal) Noriyuki Yanagida: Investigation (equal), Writing-Review & Editing (supporting) Kiyosu Taniguchi : Supervision (equal) Shigeru Suga : Supervision (equal) Takao Fujisawa: Concep- tualization (supporting), Methodology (supporting), Visualization (Lead), Formal Analysis (supporting), Writing – Original Draft Preparation (Lead), Writing- Review & Editing (Lead), .

1. Bloom CI, Franklin C, Bush A, Saglani S, Quint JK. Burden of preschool wheeze and progression to asthma in the UK: Population-based cohort 2007 to 2017. J Allergy Clin Immunol. 2021;147(5):1949-1958.2. Bui DS, Agusti A, Walters H, et al. Lung function trajectory and biomarkers in the Tasmanian Longitudinal Health Study. ERJ Open Res. 2021;7(3).3. Bui DS, Burgess JA, Lowe AJ, et al. Childhood lung function predicts adult chronic obstructive pulmonary disease and asthma-chronic obstructive pulmonary disease overlapsyndrome. Am J Respir Crit Care Med. 2017;196(1):39-46.4. Bui DS, Lodge CJ, Burgess JA, et al. Childhood predictors of lung function trajectories and future COPD risk: a prospective cohort study from the first to the sixth decade of life. Lancet Respir Med. 2018;6(7):535-544.5. Marott JL, Ingebrigtsen TS, Colak Y, Vestbo J, Lange P. Lung Function Trajectories Leading to Chronic Obstructive Pulmonary Disease as Predictors of Exacerbations and Mortality. Am J Respir Crit Care Med. 2020;202(2):210-218.6. Belgrave DCM, Granell R, Turner SW, et al. Lung function trajectories from pre-school age to adulthood and their associations with early life factors: a retrospective analysis of three population-based birth cohort studies.Lancet Respir Med. 2018;6(7):526-534.7. Martinez FD, Wright AL, Taussig LM, Holberg CJ, Halonen M, Morgan WJ. Asthma and wheezing in the first six years of life. N Engl J Med. 1995;332(3):133- 138.8. Stein RT, Holberg CJ, Morgan WJ, et al. Peak flow variability, methacholine responsiveness and atopy as markers for detecting different wheezing phenotypes in childhood. Thorax. 1997;52(11):946-952.9. Henderson J, Granell R, Heron J, et al. Associations of wheezing phenotypes in the first 6 years of life with atopy, lung function and airway responsiveness in mid-childhood. Thorax. 2008;63(11):974-980.10. Savenije OE, Granell R, Caudri D, et al. Comparison of childhood wheezing phenotypes in 2 birth cohorts: ALSPAC and PIAMA. J Allergy Clin Immunol. 2011;127(6):1505-1512 e1514.11. Yang L, Narita M, Yamamoto- Hanada K, Sakamoto N, Saito H, Ohya Y. Phenotypes of childhood wheeze in Japanese children: A group- based trajectory analysis. Pediatr Allergy Immunol. 2018;29(6):606-611.12. Custovic D, Fontanella S, Custovic A. Understanding progression from pre-school wheezing to school-age asthma: Can modern data approaches help? Pediatr Allergy Immunol. 2023;34(12):e14062.13. Minatoya M, Araki A, Miyashita C, et al. Cat and Dog Ownership in Early Life and Infant Development: A Prospective Birth Cohort Study of Japan Environment and Children’s Study. Int J Environ Res Public Health. 2019;17(1).14. Wu TJ, Pan SC, Chen BY, Chin WS, Guo YL. Different seasonal effect on asthma trajectories: A population-based birth cohort study. Pediatr Allergy Immunol. 2018;29(8):873-877.15. Saglani S, Payne DN, Zhu J, et al. Early 9 Posted on 3 Jun 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.174895576.62817881/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. detection of airway wall remodeling and eosinophilic inflammation in preschool wheezers. Am J Respir Crit Care Med. 2007;176(9):858-864.16. Lezmi G, Gosset P, Deschildre A, et al. Airway Remodeling in Preschool Children with Severe Recurrent Wheeze. Am J Respir Crit Care Med. 2015;192(2):164-171.17. Beaufils F, Mondenx M, Fayon M, et al. Bronchial Remodeling Identifies New Endotypes in Severe Preschool Wheezers. Am J Respir Crit Care Med. 2024;210(7):941-944.18. Malmstr ¨om K, Malmberg LP, O’Reilly R, et al. Lung function, airway remodeling, and inflammation in infants: outcome at 8 years. Ann Allergy Asthma Immunol. 2015;114(2):90-96.19. Guilbert TW, Morgan WJ, Zeiger RS, et al. Long-term inhaled corticosteroids in preschool children at high risk for asthma. N Engl J Med. 2006;354(19):1985-1997.20. Ermers MJ, Rovers MM, van Woensel JB, Kimpen JL, Bont LJ, Group RSVCS. The effect of high dose inhaled corticosteroids on wheeze in infants after respiratory syncytial virus infection: randomised double blind placebo controlled trial. BMJ. 2009;338:b897.21. Spergel JM, Du Toit G, Davis CM. Might biologics serve to interrupt the atopic march?J Allergy Clin Immunol. 2023;151(3):590-594.22. Caminiti L, Galletta F, Foti Randazzese S, et al. Early Assessment of Efficacy and Safety of Biologics in Pediatric Allergic Diseases: Preliminary Results from a Prospective Real-World Study. Children (Basel). 2024;11(2).23. Bacharier LB, Guilbert TW, Katelaris CH, et al. Dupilumab Improves Lung Function Parameters in Pediatric Type 2 Asthma: VOYAGE Study. J Allergy Clin Immunol Pract. 2024;12(4):948-959.24. Strachan DP, Cook DG. Health effects of passive smoking. 6. Parental smoking and childhood asthma: longitudinal and case-control studies. Thorax. 1998;53(3):204-212.25. Hossenbaccus L, Linton S, Ramchandani R, Gallant MJ, Ellis AK. Insights into allergic risk factors from birth cohort studies. Ann Allergy Asthma Immunol. 2021;127(3):312- 317.26. Owora AH, Zhang Y. Childhood wheeze trajectory-specific risk factors: A systematic review and meta-analysis. Pediatr Allergy Immunol. 2021;32(1):34-50.27. Rotroff DM, Joubert BR, Marvel SW, et al. Maternal smoking impacts key biological pathways in newborns through epigenetic modification in Utero. BMC Genomics. 2016;17(1):976.28. Breton CV, Byun HM, Wenten M, Pan F, Yang A, Gilliland FD. Prenatal tobacco smoke exposure affects global and gene-specific DNA methylation. Am J Respir Crit Care Med. 2009;180(5):462-467.29. Wu CC, Hsu TY, Chang JC, et al. Paternal Tobacco Smoke Correlated to Offspring Asthma and Prenatal Epigenetic Programming. Front Genet. 2019;10:471.30. Nabi-Burza E, Drehmer JE, Hipple Walters B, et al. Treating Parents for Tobacco Use in the Pediatric Setting: The Clinical Effort Against Secondhand Smoke Exposure Cluster Randomized Clinical Trial. JAMA Pediatr. 2019;173(10):931- 939.31. Wilson SR, Yamada EG, Sudhakar R, et al. A controlled trial of an environmental tobacco smoke reduction intervention in low-income children with asthma. Chest. 2001;120(5):1709-1722.32. Farber HJ, Walley SC, Groner JA, Nelson KE, Section on Tobacco C. Clinical Practice Policy to Protect Children From Tobacco, Nicotine, and Tobacco Smoke. Pediatrics. 2015;136(5):1008-1017.33. Kim YJ, Bunyavanich S. Microbial influencers: the airway microbiome’s role in asthma. J Clin Invest. 2025;135(4). Table 1 Participant Demographics and Clinical Background at Enrollment Baseline characteristics n % Sex Male 147 67.1 Perinatal Iron supplements during pregnancy 28 12.7 Antibiotics during pregnancy 11 5.0 Cesarean section 36 16.4 Preterm (<37weeks) 19 8.7 Low birth weight 23 10.5 Nutrition Breast feeding or mixed feeding 196 89.5 Comorbid allergic diseases Infantile eczema 123 56.2 Atopic dermatitis 67 30.6 Food allergy 57 26.0 Allergic rhinitis 35 16.0 Family history Paternal asthma 59 26.9 Maternal asthma 56 25.6 Paternal allergic rhinitis 99 45.2 Maternal allergic rhinitis 98 44.7 10 Posted on 3 Jun 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.174895576.62817881/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. Environmental factors Paternal smoking 83 37.9 Maternal smoking 14 6.4 Pet ownership 78 35.6 Nursery attendance 156 71.2 Presence of siblings 142 64.8 Table 2 Clinical Indicators of Wheezing Severity and Lower Respiratory Involvement During the Second Year of Observation Cluster 1 Cluster 2 Cluster 3 Cluster 4 p-value Severity Indicators of Wheezing Interference with daily activities 17.0% 28.8% 50.0% 46.2% .0090 Use of oral corticosteroids 0.0% 8.8% 10.7% 23.1% .0099 Possible lower respiratory tract infection Reported episode of “bronchiolitis” 0.0% 9.6% 17.9% 23.1% .0032 “Bronchiolitis” episodes were defined as caregiver-reported respiratory illnesses suggestive of lower airway involvement, without requiring physician-confirmed diagnosis. p -values were calculated using the Fisher’s exact test. Table 3 Ordinal Logistic Regression Analysis of Factors Associated with Wheeze Severity Cluster Membership OR 95%CI p-value Sex (male) 0.82 0.47-1.43 .47 Use of antibiotics during pregnancy 2.30 0.70-7.54 .17 Use of iron supplement 0.46 0.21-1.01 .053 Parental asthma 1.38 0.80-2.38 .25 Parental allergic rhinitis 0.38 0.21-0.68 .0012 Parental smoking 1.45 0.84-2.50 .19 Pet ownership 0.47 0.27-0.82 .0085 The analysis modeled cluster membership as an ordinal outcome, ordered from Cluster 1 (least severe) to Cluster 4 (most severe). Odds ratios (ORs) and 95% confidence intervals (CIs) reflect the likelihood of assignment to more severe clusters. Table 4 Binary Logistic Regression Analysis Identifying Risk Factors for Membership in High-Frequency Wheeze Clusters (Clusters 3 and 4) OR 95%CI p-value Sex (male) 1.74 0.75-4.00 0.20 Use of antibiotics during pregnancy 4.22 0.89-20.0 0.069 Use of iron supplement 0.33 0.08-1.42 0.14 Parental asthma 1.10 0.47-2.59 0.83 Parental allergic rhinitis 0.12 0.05-0.28 <0.001 Parental smoking 5.49 2.30-13.12 <0.001 11 Posted on 3 Jun 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.174895576.62817881/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. OR 95%CI p-value Pet ownership 0.11 0.03-0.32 <0.001 Clusters 1 and 2 were grouped as the low-frequency wheeze group, and Clusters 3 and 4 as the high- frequency wheeze group. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated to estimate the likelihood of belonging to the high-frequency group.. Figure legends Figure 1 Wheeze frequency trajectories over a two-year period in four distinct latent classes identified by latent class growth analysis. The vertical axis indicates the annual number of wheezing episodes, and the horizontal axis shows child age (1, 2, and 3 years). Data are presented as mean ± standard deviation. The four trajectory clusters are labeled as follows: Cluster 1: Early-Resolving Wheeze (ER) Cluster 2: Low-Frequency with Mid-Peak Wheeze (LFMP) Cluster 3: Persistent High-Frequency Wheeze (PHF) Cluster 4: Late-Peaking High-Frequency Wheeze (LPHF) Changes in wheeze frequency over time within each cluster were analyzed using a linear mixed-effects model, followed by Tukey’s post hoc test for pairwise comparisons. Statistical significance is indicated as follows: * p < 0.05, *** p < 0.001, **** p < 0.0001. Figure 2 Comparison of wheezing frequency among the four trajectory clusters at (A) baseline (age 1) and (B) age 3. Wheezing frequency is presented as box-and-whisker plots showing the median, interquartile range (IQR), and full range. The vertical axis represents the number of wheezing episodes reported over the preceding 12 months. Group comparisons were conducted using the Kruskal–Wallis test, followed by Dunn’s post hoc test for multiple comparisons. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001. Figure 1 Hosted file image2.emf available at https://authorea.com/users/930765/articles/1301889-early-onset- wheeze-trajectories-in-infants-the-phenotyping-of-wheezing-infants-p-win-study Figure 2 12