==== Front Clin Transl Allergy Clin Transl Allergy 10.1002/(ISSN)2045-7022 CLT2 Clinical and Translational Allergy 2045-7022 John Wiley and Sons Inc. Hoboken 10.1002/clt2.12271 CLT212271 Original Article Original Article Patterns of aeroallergen sensitization in asthma patients identified by latent class analysis: A cross‐sectional study in China Zhang Jiale https://orcid.org/0000-0002-1214-5659 1 Luo Wenting https://orcid.org/0000-0001-9356-7126 1 Li Guoping 2 Ren Huali 3 Su Jie 4 Sun Jianxin 5 Zhong Ruifen 6 Wang Siqin 7 Li Zhen'an 8 Zhao Yan 9 Ke Huashou 10 Chen Ting 11 Xv Chun 12 Chang Zhenglin 1 Wu Liting 1 Zheng Xianhui 1 Xv Miaoyuan 1 Ye Qingyuan 1 Hao Chuangli 13 hcl_md@163.com Sun Baoqing 1 sunbaoqing@vip.163.com 1 Department of Clinical Laboratory National Center for Respiratory Medicine, National Clinical Research Center for Respiratory Disease, State Key Laboratory of Respiratory Disease, Guangzhou Institute of Respiratory Health, The First Affiliated Hospital of Guangzhou Medical University Guangzhou China 2 Laboratory of Allergy and Precision Medicine Department of Pulmonary and Critical Care Medicine Chengdu Institute of Respiratory Health Chengdu Third People's Hospital Branch of National Clinical Research Center for Respiratory Disease Chengdu China 3 Department of Allergy State Grid Beijing Electric Power Hospital Capital Medical University Electric Power Teaching Hospital Beijing China 4 The Second People's Hospital of Foshan Foshan China 5 The Second People's Hospital of Zhaoqing Zhaoqing China 6 Dongguan Eighth People's Hospital Dongguan China 7 Henan Provincial People's Hospital Zhengzhou China 8 Foshan Maternal Child Health Hospital Foshan China 9 Department of Allergy The First Affiliated Hospital Harbin Medical University Harbin China 10 Maoming Maternal and Child Health Hospital Maoming China 11 Shengli Clinical Medical College of Fujian Medical University, Department of Otorhinolaryngology Head and Neck Surgery, Fujian Provincial Hospital Fuzhou China 12 Jiangxi Medical College Shangrao China 13 Department of Respiratory Medicine Children's Hospital of Soochow University Suzhou China * Correspondence Chuangli Hao, Department of Respiratory Medicine, Children’s Hospital of Soochow University, Suzhou, China. Email: hcl_md@163.com Baoqing Sun, Department of Allergy and Clinical Immunology, Department of Laboratory, National Center for Respiratory Medicine, National Clinical Research Center for Respiratory Disease, State Key Laboratory of Respiratory Disease, Guangzhou Institute of Respiratory Health First Affiliated Hospital of Guangzhou Medical University, 151 Yanjiangxi Road, Guangzhou, Guangdong 510120, China. Email: sunbaoqing@vip.163.com 01 7 2023 7 2023 13 7 10.1002/clt2.v13.7 e1227113 5 2023 01 2 2023 02 6 2023 © 2023 The Authors. Clinical and Translational Allergy published by John Wiley & Sons Ltd on behalf of European Academy of Allergy and Clinical Immunology. https://creativecommons.org/licenses/by/4.0/ This is an open access article under the terms of the http://creativecommons.org/licenses/by/4.0/ License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. Abstract Background This cross‐sectional study aimed to identify latent sensitization profiles of asthma patients in mainland China, unveiling the association between regional differences and sensitization patterns. Methods 1056 asthma participants from 10 medical centers divided into eastern and western cohorts were clustered into four individual sensitization patterns, respectively, by using an unsupervised statistical modeling method, latent class analysis (LCA), based on the levels of 12 aeroallergens specific IgE reactivities. Moreover, differences in clinical characteristics and environmental exposures were compared in different sensitization patterns. Results Four distinct sensitization patterns in the two cohorts were defined as follows, respectively. Eastern cohort: Class 1: “High weed pollen and house dust mites (HDMs) sensitization” (8.87%), Class 2: “HDMs dominated sensitization” (38.38%), Class 3: “High HDMs and animal dander sensitization” (6.95%), Class 4: “Low/no aeroallergen sensitization” (45.80%). Western cohort: Class 1: “High weed pollen sensitization” (26.14%), Class 2: “High multi‐pollen sensitization” (15.02%), Class 3: “HDMs‐dominated sensitization” (10.33%), Class 4: “Low/no aeroallergen sensitization” (48.51%). Of note, the significant statistical difference in age, asthma control test score (ACT) and comorbidities were observed within or between different sensitization patterns. Exposure factors in different sensitization patterns were pointed out. Conclusions Asthmatic patients with distinct sensitization patterns were clustered and identified through the LCA method, disclosing the relationship between sensitization profiles of multiple aeroallergens and geographical differences, providing novel insights and potential strategies for atopic disease monitoring, management and prevention in clinical practice. airborne allergen asthma latent class analysis (LCA) sensitization pattern sIgE Foundation of the State Key Laboratory of Respiratory DiseaseZ‐2022‐09 Guangdong Zhong Nanshan Medical Foundation20220042 Guangzhou Science and Technology Foundation2023A03J0365 Suzhou "clinical medicine expert team" introduction projectSZYJTD201806 Jiangsu Key Research and Development Program (Social Development) ProjectBE2021656 Medical Scientific Research Foundation of Guangdong Province of ChinaB2021375 source-schema-version-number2.0 cover-dateJuly 2023 details-of-publishers-convertorConverter:WILEY_ML3GV2_TO_JATSPMC version:6.3.0 mode:remove_FC converted:01.07.2023 Zhang J , Luo W , Li G , et al. Patterns of aeroallergen sensitization in asthma patients identified by latent class analysis: A cross‐sectional study in China. Clin Transl Allergy. 2023;e12271. 10.1002/clt2.12271 Jiale Zhang, Wenting Luo, Guoping Li, Huali Ren, Jie Su, Jianxin Sun, Ruifen Zhong, Siqin Wang, Zhen'an Li, Yan Zhao, Huashou Ke, Ting Chen, Chun Xv, Zhenglin Chang, Liting Wu, Xianhui Zheng, Miaoyuan Xu and Qingyuan Ye should be considered co‐first authors. Chuangli Hao and Baoqing Sun should be considered joint co‐corresponding authors. ==== Body pmc1 INTRODUCTION Globally, about 300 million patients were affected by asthma currently, 1 which was one of the most common inflammatory disorders of the airways worldwide. Roughly estimated, the prevalence of asthma in people over 20 years was 4.26% in China, reaching 45.7 million in total, posing a tremendous social and economic burden. 2 As a primary causative factor, exposure and sensitization to aeroallergens are responsible for most of the onset and persistence of asthma to a large extent. 3 Alternations have been observed in patterns of aeroallergens sensitization in China in the last decade, demonstrating that geographical differences affect asthmatic patients sensitized. 4 Besides, previous studies focused more on the northern or southern coastal areas and school‐age children, 5 while little is known about the differences in aeroallergens sensitization patterns in the western and southwestern regions in mainland China and lack large‐scale epidemiology research in different age periods. Furthermore, quite a part of asthma patients is ignorant of themselves having co‐sensitization to multiple allergens, thus making it more difficult to control and manage in clinical practice. 6 Conventionally, sensitization patterns of asthmatic patients were defined based on the presence or absence of sensitization to a panel of allergens or the overall number of sensitized allergens in a screening combo, leading to ignorance of the individual heterogeneity and variability of the polysensitization to multi‐allergens. More recently, latent class analysis (LCA) as an unsupervised statistical modeling method has been widely and successfully used in the identification of phenotypes in multiple chronic respiratory diseases. Latent class analysis is to explain the correlation among different observed variables with the least latent variables, satisfying the requirement of local independence between observed variables within each latent category meanwhile. 7 Then, all subjects would be clustered and divided into different latent classes, by calculating and comparing the biggest posterior conditional probability. Concerning a latent variable, the larger the conditional probability, the greater the weight, indicating the latent variable showed a strong influence on the observed variable in the latent category. Compared with traditional hierarchical cluster analysis (HCA), the essence of LCA is to discover potential variables, regarded as a combination of principal component analysis (PCA) and cluster analysis, providing a more integral view of detailed information reflected by explicit variables without eliminating variables and reducing people's intervention. Therefore, the aim of this multicenter study was first to establish a flexible statistical model for clustering asthma patients from eastern and western regions of China mainland into distinct sensitization patterns through the LCA method respectively. Secondly, identify and verify differences in clinical characteristics among different latent categories in two cohorts based on levels of 12 aeroallergens specific IgE. Finally, explore possible risk factors in different cluster populations by using multivariate logistic regression, providing new insights and strategies in the clinical management and precision diagnosis of atopic diseases. 2 MATERIALS AND METHODS 2.1 Patients and study design This cross‐sectional study was conducted in 10 medical centers in China between September 2021 and October 2022, including 5 eastern regions (Jiangsu, Shandong, Guangdong, Henan, and Hebei) and 5 western regions (Gansu, Sichuan, Ningxia, Inner Mongolia, and Shaanxi). Patients can be diagnosed as asthma if he/she meets the following symptoms and meets any of the objective tests for airflow limitation, except wheezing, shortness of breath, chest tightness and cough caused by other diseases: (a) recurrent wheezing, shortness of breath, with or without chest tightness or cough, frequent at night and in the morning; (b) sporadic or diffuse wheezing sounds could be heard in both lungs; (c) objective inspection of variable airflow limitation: bronchodilation test was positive, positive bronchial excitation test, peak expiratory flow (PEF) was >10%, or the weekly variation rate of PEF was >20%. Patients who at least sensitized to one of the allergens in Skin prick test (SPT) (Grade 1–4) or sIgE detection (>0.35 kU/L) will be diagnosed as allergic asthma (AA). After excluding the missing data, 1056 asthmatic patients aged 0–86 years were recruited, according to the guidelines of the Global Initiative for Asthma (GINA) 8 and Chinese. 9 After signing informed consent, patients underwent an SPT with 16 aeroallergens and collected 5 mL venous blood for serum‐sIgE detection of 12 aeroallergens. Furthermore, patients and/or their legal guardians filled in a questionnaire. The study design was presented in Figure 1 and created with BioRender.com. FIGURE 1 Graphical abstract of this cross‐sectional study. 2.2 Standardized questionnaire The questionnaire consisted of basic information, clinical history, ACT score and environmental exposures. Demographic characteristics included age, gender, ethnic group and delivery mode. Clinical history showed patients' comorbidity situations. Additionally, environmental exposures consisted of living surroundings and the history of exposures. Lastly, the ACT score was calculated after answering five questions about asthma control as listed below. 10 (i) How often has asthma influenced your daily activities (work/study/rest) in the past four weeks? (ii) How many times have you had difficulty breathing (anhelation, shortness of breath or poor breathing) in the past four weeks? (iii) How many times have you woken up at night or earlier than usual in the morning due to asthma symptoms (wheezing, coughing, dyspnea, chest tightness or pain), in the past four weeks? (iv) How many times have you used emergency medicine (such as salbutamol) in the past four weeks? (v) How did you assess your asthma control over the past four weeks? According to GINA criteria, the ACT score was classified into three different groups: (i) ACT score<20, uncontrolled; (ii) 20 = ACT score<25, partially controlled; and (iii) ACT score≥25, controlled. All completed questionnaires were verified and double‐checked by two well‐trained investigators. 2.3 Skin prick test (SPT) All candidates underwent SPT by using commercial extracts (ALK‐Abell´o Lab‐oratory, Madrid, Spain) of 16 common aeroallergens, including 4 indoor allergens (Dermatophagoides pteronyssinus, Blattella germanica, Aspergillus fumigatus, Penicillium notatum) and 12 outdoor allergens (Ambrosia elatior, Artemisia vulgaris, Chenopodium album, Betula verrucosa, Phleum pratense, Ulmus campestris, Salix fragilis, Cynodon dactylon, Populus alba, Mediterranean cypress, Platanus hispanica, Phragmites communis). Histamine and normal saline serve as positive and negative controls respectively. All performances followed the standard operation procedure. After 15 min, the wheal reaction was measured as the mean of the maximum diameter and the length of the perpendicular line through its middle. Any allergen showing the size of a wheal ≥3 mm than the negative control should be considered a positive reaction. The result was presented as skin index (SI = mean size of allergen wheal/mean size of histamine wheal). SI accounts for 25%, 50%, 100% and 200% of the histamine‐induced wheal area was defined as Grade 1, 2, 3 and 4 respectively. Grades 1–4 were considered as positive skin reactions, whereas Grade 0 was suggested as a negative reaction. All patients enrolled in the study had discontinued antiallergic drugs for at least 14 days before the SPT. 2.4 Detection of the level of sera‐specific IgE (sIgE) Asthma patients were offered blood draw for the detection of 12 aeroallergen IgE, including 6 indoor allergens, D1 (Dermatophagoides pteronyssinus), D2 (Dermatophagoides farinae), E1 (cat), E3 (horse), E5 (dog) I6 (Blattella germanica), and 6 outdoor allergens, W6 (mugwort), W7 (marguerite), W8 (dandelion), W9 (plantain), G6 (timothy), T3 (birch). Serum samples were centrifuged for 10 min at 3000 rpm and maintained at −80°C for long‐term storage. Samples were analyzed with the ALLEOS 2000TM allergen detection system (Hycor Biomedical, USA). Based on RAST classification, levels of sIgE were quantitatively categorized into six classes: Class 0, <0.35 kU/L; Class 1, 0.35–0.70 kU/L; Class 2: 0.70–3.50 kU/L; Class 3: 3.50–17.50 kU/L; Class 4: 17.50–50.00 kU/L; Class 5: 50.00–100.00 kU/L and Class 6: ≥100.00 kU/L. 2.5 Latent class analysis (LCA) Latent class analysis is a classification method of various latent variables by establishing a statistical model to describe the correlation between explicit variables and latent variables. First, the classification models were repeatedly fitted in a stepwise fashion and selected through model comparison based on the optimal model fit parameters (Akaike Information Criterion, AIC; Bayesian Information Criterion, BIC; entropy; Vuong‐Lo‐Mendell‐Rubin Likelihood Ratio, LMR p‐value; Parametric Bootstrapped Likelihood Ratio Test, BLRT p‐value; Additional File 1 Tables S1‐S2). Second, all subjects were divided into different categories based on conditional probability with post hoc tests. Lastly, summarize and define the characteristics of different sensitization profiles. Each cohort has identified one cluster with low aeroallergen‐sensitized populations for subsequent analysis. 2.6 Statistical analysis Quantitative variables were described as mean and standard deviation (mean ± S.D.), while qualitative variables were expressed as frequency or percentage. Different data were analyzed by one‐way ANOVA, chi‐square test or multiple logistic regression with IBM SPSS v. 22 software (IBM Corp., Armonk, NY, USA), respectively. Figures were plotted by GraphPad Prism v. 7 (San Diego, CA, USA) and R v. 4.2.1(Core Team 2022). Latent class analysis was conducted by Mplus software v.8.3 (Los Angeles, CA). A p‐value <0.05 was regarded as statistically significant. Significance is displayed in figures as follows: *, p < 0.05; **, p < 0.01; and ***, p < 0.001. 3 RESULTS 3.1 Characteristics of study participants In this multicenter study, 417 and 639 asthmatic patients were recruited from the eastern and western regions of China, respectively. The median age of the patients was 20.89 years 461 (43.66%) and 595 (56.34%) participants were female and male respectively. More specific details were shown in Table 1 and Additional File 1 Table S3. TABLE 1 Baseline characteristics and demographics of asthmatic patients. Characteristics Total (N = 1056) Eastern cohort (N = 417) Western cohort (N = 639) p‐value Clinical information Age, (mean ± SD) 20.89 ± 19.59 22.93 ± 21.23 19.57 ± 18.36 0.006 Gender, No. (%) 0.238 Male 595(56.34) 226(54.20) 369(57.75) Female 461(43.66) 191(45.80) 270(42.25) Ethnic group <0.001 Han Chinese 981(92.90) 412(98.80) 569(89.05) China's ethnic minorities 75(7.10) 5(1.20) 70(10.95) Delivery mode, No. (%) 0.519 Eutocia 757(71.69) 299(71.70) 458(71.67) Cesarean 299(28.31) 118(28.30) 181(28.33) Asthma control test score, (mean ± SD), No. (%) 23.17 ± 0.42 23.52 ± 0.72 22.95 ± 0.51 0.505 <20 398(37.69) 164(39.33) 234(36.62) 20≤ ACT score <25 453(42.90) 153(36.69) 300(47.26) ≥25 205(19.41) 100(23.98) 105(16.12) Allergic diseases Asthma, No. (%) N.A. Yes 1056(100.00) 417(39.49) 639(60.51) Allergic rhinitis (AR), No. (%) 0.133 Yes 867(82.10) 356(85.37) 511(80.22) No 189(17.90) 61(14.63) 128(19.78) Allergic conjunctivitis (AC), No. (%) 0.917 Yes 505(47.82) 199(47.72) 306(47.89) No 551(52.18) 218(52.28) 333(52.11) Skin allergy, No. (%) 0.565 Yes 495(46.79) 192(46.04) 303(47.27) No 448(42.34) 182(43.65) 266(41.50) Unknow 113(10.87) 43(10.31) 70(11.23) Food allergy, No. (%) 0.003 Yes 87(8.22) 27(6.47) 60(9.36) No 725(68.53) 346(82.97) 379(59.13) Unknow 244(23.25) 44(10.56) 200(31.51) Environmental exposures Trees, grass or flowers exposure, No. (%) <0.001 Symptoms appear after exposure 287(27.17) 75(17.99) 212(33.18) Without symptoms after exposure 591(55.97) 247(59.23) 344(53.83) No exposure 178(16.86) 95(22.78) 83(12.99) Furry animal exposure, No. (%) <0.001 Symptoms appear after exposure 110(10.42) 23(5.52) 87(13.62) Without symptoms after exposure 338(32.01) 138(33.09) 200(31.30) No exposure 608(57.57) 256(61.39) 352(55.08) Location of residence, No. (%) 0.006 Urban 932(88.26) 354(84.89) 578(90.45) Rural 124(11.74) 63(15.11) 61(9.55) Domestic storey, No. (%) 0.691 <9 373(35.32) 144(34.53) 229(35.84) ≥9 683(64.68) 273(65.47) 410(64.16) Tobacco smoke exposure, No. (%) 0.291 Yes 547(51.80) 208(49.88) 339(53.05) No 509(48.20) 209(50.12) 300(46.95) Using mattress, No. (%) 0.383 Yes 945(89.49) 369(88.49) 576(90.14) No 111(10.51) 48(11.51) 63(9.86) Type of quilts, No. (%) 0.001 cotton 835(78.92) 319(76.50) 516(80.50) others 330(31.19) 162(38.85) 168(26.21) Using air conditioners, No. (%) <0.001 Yes 696(65.91) 390(93.53) 306(47.89) No 360(34.09) 27(6.47) 333(52.11) 3.2 Sensitization rates and distributions of 16 representative aeroallergens on SPT Remarkably, a high sensitization rate to house dust mites (HDMs) of 50.0% was noted in patients in the eastern regions, followed by cockroaches (13.0%), while sensitization rates of other allergens were below 10.0%. On the contrary, SPT results in western areas showed that 45.0% of patients were sensitization to mugwort, followed by ragweed (30.0%), goosefoot (30.0%), elm (29.0%), poplar (23.0%) and Bermuda grass (23.0%), demonstrating that the western cohort featured in sensitization to multi‐pollen with an increasing sensitization rate of HDMs (20.0%) (Figure 2A). Specifically, over half of the cases developed grade 2 reactions on SPT in 13 allergens, except Der. p and mugwort were mainly developed grade 3 reactions (Figure 2B). The difference in sensitization rates of different aeroallergens between the two cohorts was statistically significant (***, p < 0.001). FIGURE 2 Sensitization rates (A) and distributions (B) of 16 aeroallergens on Skin prick test (SPT) in two regions. The differences in sensitization rates of different aeroallergens were compared between the two cohorts (*, p < 0.05; **, p < 0.01; and ***, p < 0.001). Der. p (Dermatophagoides pteronyssinus), Bla. g (Blattella germanica), Asp. f (Aspergillus fumigatus). 3.3 Classification and characteristics of sensitization patterns in different cohorts In the eastern cohort, 417 participants were clustered into four distinct sensitization categories using the LCA model. Four classes were defined as follows. Class 1: “High weed pollen and HDMs sensitization” (n = 37 [8.87%]), Class 2: “HDMs dominated sensitization” (n = 160 [38.38%]), Class 3: “High HDMs and animal dander sensitization” (n = 29 [6.95%]), Class 4: “Low/no aeroallergen sensitization” (n = 191 [45.80%]) (Figure 3A). Besides, the top 5 sIgE levels of allergen‐specific shown in Figure 3B–E verified the accuracy of sensitization patterns in clustering and classification. Patients in Class 4 were characterized with close to zero probability of sensitization to all 12 aeroallergens. Except for the Class 4 population, half of the patients in the eastern cohort were commonly sensitization to HDMs (Der. p [52.0%], Der. f [51.0%]). Patients in Class 1 exhibited high sensitization to both weed pollen (marguerite [94.59%], dandelion [81.08%], mugwort [75.68%]) and HDMs (Der. f [64.86%], Der. p [59.46%]). Moreover, both Class 2 and Class 3 populations showed a higher probability of sensitization to HDMs (Class 2 Der. p [100.0%], Der. f [99.38%] vs. Class 3 Der. p [100.0%], Der. f [100.0%]), and Class 3 has a high sensitized rate of animal dander (cat [100.0%], horse [62.07%], dog [48.28%]). FIGURE 3 Differences in 12 aeroallergens sensitization patterns. (A) based on sIgE reactivity (B‐E) in the eastern cohort. In the eastern cohort, the radar plots showed the sensitization rate of each aeroallergen in different sensitization patterns with the posterior probability (Figure 3A). While the bar charts demonstrated the top 5 sensitization rates of 12 aeroallergens in different sensitization patterns (Figure 3B–E). 12 aeroallergens included W6 (mugwort), W7 (marguerite), W8 (dandelion), W9 (plantain), G6 (timothy), T3 (birch), D1 (Dermatophagoides pteronyssinus), D2 (Dermatophagoides farinae), E1 (cat), E3 (horse), E5 (dog) and I6 (Blattella germanica). In the western cohort, 167 patients of Class 1 (26.14%) were labeled as having high sensitization to weed pollen. While 96 patients of Class 2 (15.02%) showed a high probability of sensitization to weed, tree, and grass pollens. High HDMs‐dominated sensitization was noted in 66 patients of Class 3 (10.33%) population, while low aeroallergen sensitization was observed in 310 patients of Class 4 (48.51%) (Figure 4A). Additionally, as shown in the Figure 4B–E, patients in Class 1 and Class 2 were characterized by high sensitization to pollens; the top 3 sensitization rates were marguerite, mugwort, and plantain, followed by dandelion and cat (Class 1)/birch (Class 2). Patients in Class 3 demonstrated a significantly high sensitized rate to HDMs primarily, while Class 4 populations presented a low probability of sensitization to all 12 aeroallergens. FIGURE 4 Differences in 12 aeroallergens sensitization patterns. (A) based on sIgE reactivity (B‐E) in the western cohort. In the western cohort, the radar plots showed the sensitization rate of each aeroallergen in different sensitization patterns with the posterior probability (Figure 4A). While the bar charts demonstrated the top 5 sensitization rates of 12 aeroallergens in different sensitization patterns (Figure 4B–E). 12 aeroallergens included W6 (mugwort), W7 (marguerite), W8 (dandelion), W9 (plantain), G6 (timothy), T3 (birch), D1 (Dermatophagoides pteronyssinus), D2 (Dermatophagoides farinae), E1 (cat), E3 (horse), E5 (dog) and I6 (Blattella germanica). 3.4 Differences of age among distinct sensitization patterns Age played a critical role in different sensitization patterns. In the eastern cohort, based on the reference category (Class 4), results showed that asthma patients with high HDMs‐dominated sensitization, Class 2 (41.88% vs. 19.37%, p < 0.001) and Class 3 (44.83% vs. 19.37%, p < 0.001), were mainly aged 0–6 years, suggesting that infancy was susceptible to AA induced by HDMs (p < 0.001). Besides, HDMs‐sensitized asthma patients combined with pollen and animal dander sensitization (Class 1) showed a trend of sensitization during the school‐age children period, indicating the probability of poly‐sensitization was increased with rising age and exposure to various allergens (32.43% vs. 17.80%, p < 0.05) (Figure 5A). FIGURE 5 Comparisons of age among different sensitization patterns in eastern (A) and western (B) cohorts. In the eastern cohort, compared with the reference category (Class 4), results showed that asthma patients with HDMs‐dominated high sensitization, Class 2 (41.88% vs. 19.37%, p < 0.001) and Class 3 (44.83% vs. 19.37%, p < 0.001), were mainly infancy (aged 0–6 years). While the Class 1 population (32.43% vs. 17.80%, p < 0.05) showed a trend of sensitization during the school‐age children period (aged 7–14 years) (Figure 5A). In the western cohort, results revealed that the patients of Class 1 (35.33% vs. 30.97, p < 0.05) and Class 2 (42.17% vs. 30.97%, p < 0.001) groups tend to have high proportions of school‐age children (7–14 years) than other clusters (Figure 5B) (*, p < 0.05; **, p < 0.01; and ***, p < 0.001.). In the western cohort, results revealed that the patients of Class 1 (35.33% vs. 30.97, p < 0.05) and Class 2 (42.17% vs. 30.97%, p < 0.001) groups tend to have high proportions of school‐age children (7–14 years) than other clusters. In contrast with different cohorts with similar sensitization characteristics, patients predominantly with high sensitization to HDMs focused more on young and middle‐aged adults in the western cohort population (Class 3, 42.42%, p < 0.05), while patients in the eastern cohort mainly aged 0–6 years (Class 2, 41.88%, p < 0.01). Differences in age proportion between the western and eastern cohort cohorts showed statistical significance (≤6 years, 22.73% vs. 41.88%, p = 0.006; 14