Bidirectional association between asthma and chronic rhinosinusitis: Two longitudinal follow-up studies using a national sample cohort

The aim of this study was to evaluate an epidemiologic association of asthma and chronic rhinosinusitis (CRS) using a national sample cohort of the Korean population. We collected data from the Korean Health Insurance Review and Assessment Service-National Sample Cohort between 2002 and 2013, and two different case-control cohorts were designed (1st: asthmatic patients matched in a 1:1 ratio with 204,119 non-asthmatics as control I, 2nd: CRS patients matched in a 1:4 ratio with 124,020 non-CRS patients as control II). Bidirectional association was examined using Cox proportional hazard models stratified by age, sex, income, and region of residence. Patients with asthma had an increased risk of developing CRS [adjusted hazard ratio (95% confidence interval) = 1.74 (1.67–1.80)], both with nasal polyps [1.55 (1.36–1.78)], without nasal polyps [1.74 (1.67–1.81)]. In the second cohort, patients with CRS had increased risk of developing asthma [1.85 (1.80–1.91)] with similar results for those with and without nasal polyps. The strongest association for risk of CRS was in 20–39 years old men with asthma [2.41 (1.97–2.96)], while the strongest association for increased risk of asthma in those with CRS group was also seen in this same subgroup [2.40 (2.18–2.63)]. CRS and asthma had a bidirectional influence on each other. CRS increased the risk of asthma, and asthma increased the risk of CRS, especially in young men.

Study ii: cRS increased the risk of developing asthma. CRS patients were 1:4 matched with control I group who had same general characteristics (Fig. 1B). The time from index date, the date of the diagnosis of CRS, to asthma diagnosis was 32.9 months (SD = 32.1) in the CRS group and 34.1 months (SD = 32.5) in the control II group (Fig. 2B). The rate of asthma was higher in the CRS group [20.2% (6,270/31,005)) than in the control II group [10.3% (12,731/124,020), P < 0.001] ( Table 1). The general characteristics of participants were exactly same (P = 1.000), while the rates of atopic dermatitis, COPD, and CCI score ≥ 2 were higher in the CRS group than in the control II group (P < 0.001, each comparison).
Patients with CRS had increased risk of developing asthma [1.85 (1.80-1.91)]. The risk of developing asthma among patients with in the CRSwNP and CRSsNP groups was 1.96 (1.79-2.15) and 1.84 (1.78-1.90), respectively, compared to the control II group (Table 4). In study II, asthma was more prevalent in CRSwNP and CRSsNP patients compared to each control group (Table S3). Between two CRS phenotypes, some characteristics were different, including age, prevalence of atopic dermatitis, COPD, and asthma (Table S4).
In subgroup analyses, the risk of developing asthma was higher in the CRS group (P < 0.05, each comparison) ( Table 5). The strongest associations for risk of CRS was highest in men and women aged 20-39 years old [3.04 (2.67-3.47) and 2.40 (2.18-2.63), respectively].

Discussion
Many epidemiologic studies have identified the risks of CRS among asthma patients and vice versa. However, this is the first study to examine the bidirectional association of CRS and asthma using nationwide population-based data. This study demonstrated that the risk of CRS in asthma patients and that of asthma in CRS patients was higher than that of each respective control group. In the subgroup analysis, both diseases showed the highest hazard ratio among 20-39 years old men.
Comorbid asthma is highly prevalent in CRS patients and is associated with severity and refractoriness of CRS 24 . A multinational survey study (n = 52,000) found that asthma was highly associated with CRS, especially in young people, which is consistent with our results 13 . A Taiwanese population-based study with the same setting as our Study I found an increased risk of CRSsNP in asthma patients, but the association with CRSwNP was not significant 25 . The HR for CRSsNP in our study was relatively higher than that for CRSwNP, although statistical significance was reached in both CRS phenotypes. This discrepancy may be explained by the smaller number of patients with CRSwNP compared to those with CRSsNP in both Korean and Taiwanese populations. Sinonasal tissues from Asian CRS patients were less eosinophilic in both CRSwNP and CRSsNP 26 . Furthermore, CRSwNP were also less eosinophilic in Chinese patients compared to European patients 27 . The regional difference of inflammatory patterns of CRS and asthma comorbidity might affect the low incidence of CRSwNP in both Asian national cohort studies. A study from the Canadian national cohort found that the presence of CRS resulted in an increased risk of developing asthma 28 . Another population-based study from Canada found CRS to be more frequent in asthma patients 29 .
Upper airway symptoms caused by infection or inflammation of the paranasal sinuses can hinder disease control or cause acute exacerbation of asthma 30 . Also, patient-reported severity of CRS is negatively associated with that of asthma control 31  www.nature.com/scientificreports www.nature.com/scientificreports/ the severity of asthma, the frequency of attacks, and the number of medication use 32,33 . Regab et al. conducted a randomized study in CRS patients with comorbid asthma that assessed medical and surgical therapy of CRS, and found that both types of treatment had a positive effect on asthma control 34 . Among patients who underwent FESS due to recalcitrance to medical treatment, those who underwent operation earlier in the disease continuum had a decreased risk of new asthma diagnosis 10 .  www.nature.com/scientificreports www.nature.com/scientificreports/ Asthmatic CRS patients have terrible symptoms, including hyposmia and nasal blockage, compared to non-asthmatic CRS patients 35 . In addition, asthmatic CRS patients are known to be highly comorbid with nasal polyposis 5 . However, in this study, the prevalence of CRSsNP was higher than CRSwNP and the HR of CRSsNP was higher than that of CRSwNP. One of the possible reasons of the low incidence of CRSwNP is regional difference (Western versus Asia) as we mentioned above, and the other is possibility of the missing of J33 code in patients with CRSwNP in clinical setting. Usually clinicians give the patients a diagnostic code of J32 (chronic rhinosinusitis) and add J33 (nasal polyposis), if a patient has nasal polyps. There may be a chance that the J33 code is missing during the process. On the contrary, Marino et al. discovered that comorbid asthma did not impair sinus pneumatization and was not associated with the Lund-Mackay computed tomography (CT) score reflecting CRS disease severity 36 . In CRSwNP patients, comorbid asthma had no impact on tissue eosinophilia and CT score 37 . Further study is needed to determine whether asthma affects the development and progression of CRS.
Treatment with biologics is already popular in asthma patients and is increasingly being used for CRS, particularly CRSwNP patients 38,39 . The effectiveness of biologics treatment in asthmatic CRSwNP patients was proved by www.nature.com/scientificreports www.nature.com/scientificreports/ randomized controlled trial 40 . The findings of a bidirectional association between asthma and CRS could be used as evidence for biologics treatment in patients with CRS.
Although this study is the first to use a nationwide cohort to reveal the bidirectional association between asthma and CRS, it is not without limitations. First, the well-known risk factors of both diseases, such as smoking and obesity, could not analyzed due to the characteristics of the data. Second, although this result included a temporal relationship between the two diseases, the causal mechanism could not be elucidated. Third, CRSwNP might be underestimated due to missing J33 code. Lastly, a few cases who may have CRS history prior to 2012 could not be excluded in study 1, because there was no data before 2012.
In summary, our results show that CRS and asthma have a bidirectional influence on each other regardless of the presence of CRSwNP or CRSsNP. CRS increases the risk of asthma, and asthma increases the risk of CRS, especially in young men.

Study population and data collection. The ethics committee of the Institutional Review Board of Hallym
University approved the use of these data (2017-I102). Written informed consent was waived by the Institutional Review Board of Hallym University. All methods of data management and analysis performed in accordance with the guidelines and regulations of the institutional ethic committee. We collected data from the Korean Health Insurance Review and Assessment Service-National Sample Cohort, as previously described in our studies 41-43 . participant selection. Of 1,125,691 cases with 114,369,638 medical claim codes from 2002 to 2013, we included subjects diagnosed as asthma (ICD-10: J45) or status asthmaticus (J46). Among them, we included subjects who were diagnosed with asthma by a physician more than 2 times, and treated with asthma-related medications, including inhaled corticosteroid (ICS), ICSs combined with long-acting β2-agonists (LABAs), oral leukotriene antagonists (LTRAs), short-acting β2-agonists (SABAs), systemic LABAs, xanthine derivatives, and systemic corticosteroids (n = 230,764). The follow-up period was 12 years. The detailed method is described in the previous study 42,44 .
CRS was diagnosed using ICD-10 codes (J32). Among these, we selected the participants who had treatment more than 2 times due to CRS and had undergone head and neck CT (Claim codes: HA401-HA416, HA441-HA443, HA451-HA453, HA461-HA463, or HA471-HA473), as in our previous study 41 . Out of 1,125,691 cases, 5,177 participants were diagnosed as CRSwNP (J32 plus J33 for nasal polyposis), and 34,572 participants were diagnosed as CRSsNP (J32 only). Study i. The asthma patients were matched 1:1 with participants among this cohort who were not diagnosed with asthma (control I, n = 894,927). We matched variables between two groups, such as age, group, sex, income group, and region of residence. We sorted the control I group using another random number order and then selected from top to bottom to prevent selection bias. We set the index date as the date of the diagnosis of asthma, and we assumed that control I participants were involved at the same time as matched asthma participant 43 . Subjects who diagnosed CRS (J32) before the index date were excluded from the two groups. We excluded 5,963 participants from the asthma group because of the previous history of CRS and 20,682 participants due to insufficient matching. The average follow-up period from index date to last date (31st December 2013) or death date was almost similar in the asthma (86.1 months, SD = 42.9) and the control I group (85.6 months, SD = 43.2). After 1:1 matching, 204,119 patients with asthma and 204,119 control I participants were allocated (Fig. 1A). The occurrence of CRS was analyzed in the asthma and control I groups.
Study ii. We matched 1:4 of the CRS and control II group. The control II group (n = 1,085,942) was selected from the cohort who were not diagnosed with CRS from 2002 through 2013. Matching the variables (age, group, sex, income group, and region of residence) and setting the index date were performed same as Study I. In the CRS group, we excluded 8,744 participants who had previous history of asthma. There were no patients with CRS who could not sufficiently matched to the Control II. The mean follow-up time from index date to last date (31st December 2013) or death date was almost similar in the CRS group (87.5 months, SD = 41.2) and control II group (86.8 months, SD = 41.4). Therefore, 31,005 of CRS patients and 124,020 control II participants were included (Fig. 1B). We analyzed the occurrence of the asthma in the CRS and control II group. Also, we analyzed the occurrence of asthma in CRSwNP and their control, and CRSsNP and their control.
COPD was defined as who were diagnosed J43 (emphysema) through J44 (Other chronic obstructive pulmonary disease), and treated with SABA, LABA, LAMA, and corticosteroid. Atopic dermatitis (L20) was defined that who was treated more than two time as a previous study 45 Table 5. Subgroup analysis of crude and adjusted hazard ratios (95% confidence interval) of chronic rhinosinusitis for asthma according to age and sex in study II. *Cox-proportional hazard regression model, Significance at P < 0.05. † Stratified model for age, sex income, and region of residence. ‡ Adjusted model for atopic dermatitis, chronic obstructive pulmonary disease, and Charlson Comorbidity Index. CRS, chronic rhinosinusitis.