Association between diabetes and asthma

Accepted Manuscript

Association between Diabetes and Asthma: Evidence from a Nationwide Korean Study Ji Yeon Baek , Seung Eun Lee , Kyungdo Han , Eun Hee Koh PII: DOI: Reference:

S1081-1206(18)30668-9 https://doi.org/10.1016/j.anai.2018.08.008 ANAI 2666

To appear in:

Annals of Allergy, Asthma Immunology

Received date: Revised date: Accepted date:

12 June 2018 2 August 2018 13 August 2018

Please cite this article as: Ji Yeon Baek , Seung Eun Lee , Kyungdo Han , Eun Hee Koh , Association between Diabetes and Asthma: Evidence from a Nationwide Korean Study, Annals of Allergy, Asthma Immunology (2018), doi: https://doi.org/10.1016/j.anai.2018.08.008

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Association between Diabetes and Asthma: Evidence from a Nationwide Korean Study Short titles: Association between Diabetes and Asthma

Ji Yeon Baek, MDa, #, Seung Eun Lee, MD, MSb, #, Kyungdo Han, PhDc,

CR IP T

Eun Hee Koh, MD, PhDa* On Behalf of the Taskforce Team of Diabetes Fact Sheet of the Korean Diabetes Association

These authors contributed equally to this work.

AN US

#

Author affiliations a

Department of Internal Medicine, Asan Medical Center, University of Ulsan College of

b

M

Medicine, Seoul, Korea.

Department of Internal Medicine, Dongguk University Ilsan Hospital, Dongguk University

Department of Biostatistics, The Catholic University of Korea, Seoul, Korea.

CE

PT

c

ED

College of Medicine, Goyang, Korea.

AC

*Corresponding author contact information Address for correspondence: Eun Hee Koh Department of Internal Medicine, Asan Medical Center, University of Ulsan College of Medicine, 88, Olympic-ro 43-gil, Songpa-gu, Seoul 05505, Korea. Phone: +82-2-3010-3248, Fax: +82-2-2045-4034, E-mail: [email protected] Conflicts of interest: none

ACCEPTED MANUSCRIPT Funding source: none Keywords: Asthma; Diabetes Mellitus; Diabetic retinopathy Abbreviations: DM, diabetes mellitus; DR, diabetic retinopathy; DLco, diffusing capacity of the lung for carbon monoxide; ICS, inhaled corticosteroid; LABA, long-acting beta-2 agonist;

CR IP T

LAMA, long-acting muscarinic antagonist; SABA, short-acting beta-2 agonist; SAMA, shortacting muscarinic antagonist; LTRA, leukotriene receptor antagonist; NHIS, the Korean National Health Insurance Service; DB, database; ICD, International Statistical Classification of Diseases and Related Health Problems; HR, hazard ratio; CI, confidence interval.

AN US

Word count: 1914 (excluding abstract, acknowledgments, e-Supplement material, figure legends, references, and tables). Figures: one

CE

Introduction

PT

ED

M

Tables: three

Asthma

is

a

chronic

inflammatory

airway

disease

characterized

by

airway

AC

hyperresponsiveness and airway inflammation, resulting in bronchoconstriction, airway remodeling, and airway obstruction.1,2 It is a serious health issue world-wide, affecting more than 300 million individuals. The global prevalence of clinical asthma has increased during recent decades.3 Although the absolute prevalence of asthma in Korea is reported to be lower than that of other countries,4,5 the prevalence of asthma has been steadily increasing from 4.94% to 5.7% between 2006 and 2010.6

ACCEPTED MANUSCRIPT Another pervasive disease, type 2 diabetes, is defined as a progressive insulin secretory defect on a background of insulin resistance.7 It is widely accepted that obesity or insulin resistance reflects a chronic inflammatory state in adipose tissue and liver.8,9 Subjects with high levels of interleukin-6 and C-reactive protein are more likely to develop type 2

diseases such as rheumatoid arthritis or psoriasis.11,12

CR IP T

diabetes.10 Furthermore, patients with type 2 diabetes are at increased risk of inflammatory

Since these two disease entities share a common pathophysiology (‘chronic inflammation or meta-inflammation’),13,14 many epidemiologic studies have examined the association between diabetes and asthma. An observational Nurses’ Health Study revealed a null

AN US

association between asthma and type 2 diabetes.15 By contrast, several studies report a positive association between the two diseases.16-18 Furthermore, there is an inverse association between diabetes severity (as defined by disease duration and type of antidiabetic

M

medication) and lung function, suggesting that the lung is a target organ of diabetic injury.19 The aim of this study was to investigate the effect of diabetes on asthma development using

ED

data from a large population-based study in Korea. In addition, we divided diabetes into

CE

development.

PT

diabetes with or without retinopathy to evaluate the impact of disease severity on asthma

AC

Methods

Data source and study population The Korean National Health Insurance Service (NHIS), a single-payer system, is mandatory

for all residents in Korea.20 The NHIS provides free annual or biannual regular health checkups to all applicable examinees, including i) employee subscribers and regional insurance subscribers who are regional householders; ii) dependents and household members of

ACCEPTED MANUSCRIPT employee subscribers (40 years or older); and iii) medical aid beneficiaries (householders aged 19–64 years and household members aged 41–64 years).20 The results from regular health check-ups are held in the Health Check-up database (DB) of the Korean NHI. Along with the Health Check-up DB, the Korean NHIS DB comprises a Qualification DB and a Health Claim DB through which hospitals claim healthcare costs from the NHIS. 21

CR IP T

In this study, we enrolled subjects who underwent regular health check-ups between 2005 and 2008. The data from the regular health check-ups were combined with those of Claim DB (2005 -2013), which included information on patients’ diagnoses, treatments, procedures, surgical histories and drugs prescribed.22 Subjects were excluded if 1) they had a history of

AN US

asthma, 2) there was any missing information about laboratory results or demographic factors, or 3) the examinee was younger than 30 years of age. In addition, if subjects underwent two or more examinations between 2005 and 2008, only the first check-up was considered. The

M

final sample size was 13,154,348, which was classified according to the status of diabetes or diabetic retinopathy (DR) at the time of enrollment; subjects were followed-up until the date

ED

of death or until 31 December 2013, whichever came first (Fig. 1). The primary endpoint was

PT

newly diagnosed asthma. The study protocol was approved by the Korean National Institute for Bioethics Policy (P01-201504-21-005). Anonymized and de-identified information were

CE

used for all analyses.

AC

Definitions of type 2 diabetes and diabetic retinopathy Using claim data, individuals were defined as having type 2 diabetes if they had a

diagnostic code for this disease and a prescription for antidiabetic drugs.20 A diagnosis was made according to the International Statistical Classification of Diseases and Related Health Problems, 10th revision (ICD-10), codes E11–E14. Sulfonylureas, metformin, meglitinides, thiazolidinediones, dipeptidyl peptidase-4 inhibitors, α-glucosidase inhibitors, and insulin were regarded as relevant prescriptions. Also, subjects with fasting plasma glucose levels

ACCEPTED MANUSCRIPT ≥126 mg/dL at a regular health check-up were considered to have type 2 diabetes. Subjects with diabetes were further sub-classified according to presence of DR. DR was defined according to ICD-10 code H36.0 for patients with diabetes.

Diagnosis of asthma

CR IP T

Asthma was defined using a billing code for asthma, relevant prescriptions, and performance of designated asthma-related tests from the claim DB. Subjects with diagnostic codes (ICD-10) for asthma (J45-J46) were defined as having asthma if they also had 1) related prescriptions (e.g., inhaled corticosteroid (ICS), ICS + long-acting beta-2 agonist

AN US

(LABA), LABA, systemic LABA, long-acting muscarinic antagonist (LAMA), short-acting beta-2 agonist (SABA), short-acting muscarinic antagonist (SAMA), leukotriene receptor antagonist (LTRA), systemic corticosteroid, xanthine derivatives, or systemic SABA) or 2)

M

they performed relevant tests (e.g., pulmonary function test or air way provocation test)

PT

Statistical analysis

ED

according to previous described criteria.6,7

Baseline characteristics of the participants were compared using one-way analysis of

CE

variance (continuous variables) and the chi-square test (categorical variables). When significant, data were subjected to Bonferroni post-hoc analysis to identify inter-group

AC

differences. The rate of incidence was expressed as the number of events per 1,000 personyears. Cox proportional hazards regression analysis was used to calculate the hazard ratio (HR) for asthma depending on the presence of diabetes or DR, with adjustments for covariates believed to be important confounders (BMI, smoking, heavy drinking, exercise, hypertension and dyslipidemia). A two-sided P value < 0.05 was considered statistically significant. The SAS survey procedure (version 9.2; SAS Institute, Cary, NC) was used for

ACCEPTED MANUSCRIPT data management and analysis.

Results Baseline characteristics of the study subjects As shown in Table 1, subjects with incident asthma were older and had more comorbidities,

CR IP T

such as hypertension and dyslipidemia, than subjects without. However, patients with incident asthma smoked and drank less and were more physically active than patients without asthma. The mean follow-up duration for subjects with incident asthma was 3.5 years, and

AN US

that for subjects without incident asthma was 7.1 years.

Incidence and risk of asthma according to diabetic status

By the end of the follow-up period, the incidence rates of asthma among non-DM, DM

M

without DR, and DR patients were 27.1, 30.1, and 38.4 per 1,000 person-years, respectively (Table 2). After adjusting for age and sex (model 2), those with DM without DR had a lower

ED

incidence of asthma than those without DM (HR, 0.95; 95% confidence interval [CI], 0.946–

PT

0.954). Even after adjusting for other clinical variables (BMI, smoking, alcohol, exercise, hypertension, and dyslipidemia), DM patients without DR had a significantly lower risk of

CE

asthma (HR, 0.943; 95% CI, 0.939–0.948). By contrast, DM patients with DR had a higher risk of developing asthma than subjects without DM (HR, 1.067; 95% CI, 1.053–1.081).

AC

Next, we compared the incidence of asthma according to age group (Table 3). For all age

groups, having DM without DR was significantly associated with a lower risk of asthma development (age 30-39: HR, 0.916; 95% CI, 0.898–0.934; age 40-59: HR, 0.939; 95% CI, 0.932–0.964; age ≥60: HR, 0.934; 95% CI, 0.928–0.94). Regarding the association between DR and incident asthma, DR patients aged <60 years had a higher risk of developing asthma; this association was not significant in subjects aged ≥60 years.

ACCEPTED MANUSCRIPT

Discussion This prospective large population-based study showed that type 2 diabetes itself does not increase the risk of asthma. However, the risk of incident asthma increased significantly if diabetic patients also had retinopathy (Table 2). When analyzed according to age group, the

CR IP T

data revealed that the positive association between DR and incident asthma was abolished in subjects aged ≥60 years (Table 3).

The null association between diabetes and incident asthma is consistent with findings reported by previous studies.15,23 As Rana et al.15 mentioned, this result can be explained by a

AN US

discrepancy in inflammatory mechanisms. Macrophage activation can be classified as two separate polarization states: M1 and M2.24 M1 or ‘classically activated’ macrophages show increased production of proinflammatory cytokines such as TNF-α and IL-6. By contrast, M2

M

or ‘alternatively activated’ macrophages produce anti-inflammatory cytokines such as IL-4 and IL-10.24 In an obese or insulin-resistant state, inflammation of adipose tissue and liver is

ED

associated mainly with activation of M1 macrophages, which is triggered by multiple stimuli such as lipopolysaccharides, free fatty acids or cytokines (TNF-α and IL-1β)25,26. By contrast,

PT

alternatively activated M2 macrophages are abundant in the airways of asthmatic patients27

CE

with eosinophilic inflammation caused by increased levels of Th2 cytokines (IL-4, IL-5, IL-9 and IL-13).1,2,28 Although type 2 diabetes and asthma seem to share inflammatory

AC

mechanisms, the different functions of macrophages may have caused the null association observed in this study. Interestingly, we found that diabetic patients with accompanying retinopathy were at

significantly higher risk of developing asthma. Previous studies reveal that diabetic severity, as indicated by the intensity of antidiabetic treatment or fasting glucose levels, shows a graded relationship with pulmonary dysfunction.19,29 These results suggest that the lung is a

ACCEPTED MANUSCRIPT target organ for diabetic injury. Accordingly, the diffusing capacity of the lung for carbon monoxide (DLco) is related directly to hemoglobin A1c levels and diabetic complications.30 The authors postulate that advanced glycation end-products, which predict long-term progression of retinopathy and nephropathy, also predispose to dysfunction of the alveolar microvasculature. In addition to microangiopathy of the lung, several other mechanisms

CR IP T

including autonomic neuropathy in respiratory muscle,31 abnormal lung elasticity,32 and comorbidities such as ischemic heart disease, can worsen respiratory function in patients with advanced diabetic complications.33 In line with the results of previous studies, the higher incidence of asthma in patients with DR in this study supports the possibility that the lung is

AN US

one of the target organs of diabetes.

We found that subjects with DR aged <60 years had a higher incidence of asthma than those without DR after adjusting for confounding variables (Table 3); this association was not

M

significant in subjects aged ≥60 years. Although the reason for the conflicting results between age groups is not clear, we suspect that the mitigated risk of incident asthma in subjects aged

ED

≥60 can be explained by the presence of pulmonary microangiopathy in advanced diabetes. Because aging itself can lead to derangement of connective tissue metabolism in conjunction

PT

with reduced DLco and lung function,34,35 the effect of DR on asthma development may be

CE

ameliorated in older age groups. Future studies should examine whether proper glycemic control can mitigate the risk of asthma development, especially in younger diabetic patients.

AC

This study has some limitations. First, we did not take sub-phenotypes of asthma, such as

atopic or non-atopic asthma, into account because the ICD-10 codes used in this study could not discriminate between these sub-phenotypes. Second, the data did not include information about diabetes duration or lung function. Third, because we defined retinopathy using claim data from the Korean NHIS, any inaccuracies in the claims codes may lead to misclassification of disease. Despite these limitations, the large sample size is a major

ACCEPTED MANUSCRIPT strength of our study. In addition, we used a relatively strict definition of type 2 diabetes or asthma based on combining diagnostic codes and prescription information; this should reduce possible inaccuracies in the claims data. In conclusion, the results of this large population-based study suggest that type 2 diabetes without retinopathy is not a significant risk factor for asthma development. However, patients

CR IP T

with DR are at a greater risk of incident asthma. Further studies are required to investigate whether strict glycemic control can mitigate the higher risk of incident asthma, especially in younger patients. Furthermore, physicians should be alert to patients’ respiratory symptoms and lung function, particularly in those with complications associated with advanced diabetes.

AN US

Acknowledgments

The authors thank the Korean National Health Insurance Service for providing access to their data (database no. NHIS-2015-4-008).

Renauld JC. New insights into the role of cytokines in asthma. J Clin Pathol.

ED

[1]

M

References

2001;54:577-589.

Sumi Y, Hamid Q. Airway remodeling in asthma. Allergol Int. 2007;56:341-348.

[3]

To T, Stanojevic S, Moores G, et al. Global asthma prevalence in adults: findings from

CE

PT

[2]

the cross-sectional world health survey. BMC Public Health. 2012;12:204. Song WJ, Kang MG, Chang YS, Cho SH. Epidemiology of adult asthma in Asia:

AC

[4]

toward a better understanding. Asia Pac Allergy. 2014;4:75-85.

[5]

Akinbami LJ, Moorman JE, Bailey C, et al. Trends in asthma prevalence, health care use, and mortality in the United States, 2001-2010. NCHS Data Brief. 2012:1-8.

[6]

Kim S, Kim J, Kim K, et al. Healthcare use and prescription patterns associated with adult asthma in Korea: analysis of the NHI claims database. Allergy. 2013;68:1435-

ACCEPTED MANUSCRIPT 1442. [7]

Park SY, Kim JH, Kim HJ, et al. High Prevalence of Asthma in Elderly Women: Findings From a Korean National Health Database and Adult Asthma Cohort. Allergy Asthma Immunol Res. 2018;10:387-396.

[8]

Shoelson SE, Lee J, Goldfine AB. Inflammation and insulin resistance. J Clin Invest.

[9]

CR IP T

2006;116:1793-1801.

Tateya S, Kim F, Tamori Y. Recent advances in obesity-induced inflammation and insulin resistance. Front Endocrinol (Lausanne). 2013;4:93.

[10]

Wang X, Bao W, Liu J, et al. Inflammatory markers and risk of type 2 diabetes: a

[11]

AN US

systematic review and meta-analysis. Diabetes Care. 2013;36:166-175.

Lu MC, Yan ST, Yin WY, Koo M, Lai NS. Risk of rheumatoid arthritis in patients with type 2 diabetes: a nationwide population-based case-control study. PLoS One.

[12]

M

2014;9:e101528.

Wu CY, Shieh JJ, Shen JL, Liu YY, Chang YT, Chen YJ. Association between

ED

antidiabetic drugs and psoriasis risk in diabetic patients: results from a nationwide

[13]

PT

nested case-control study in Taiwan. J Am Acad Dermatol. 2015;72:123-130. Perez MK, Piedimonte G. Metabolic asthma: is there a link between obesity, diabetes,

[14]

CE

and asthma? Immunol Allergy Clin North Am. 2014;34:777-784. Forno E. Asthma and diabetes: Does treatment with metformin improve asthma?

AC

Respirology. 2016;21:1144-1145.

[15]

Rana JS, Mittleman MA, Sheikh J, et al. Chronic obstructive pulmonary disease, asthma, and risk of type 2 diabetes in women. Diabetes Care. 2004;27:2478-2484.

[16]

Thomsen SF, Duffy DL, Kyvik KO, Skytthe A, Backer V. Risk of asthma in adult twins with type 2 diabetes and increased body mass index. Allergy. 2011;66:562-568.

[17]

Mueller NT, Koh WP, Odegaard AO, Gross MD, Yuan JM, Pereira MA. Asthma and

ACCEPTED MANUSCRIPT the risk of type 2 diabetes in the Singapore Chinese Health Study. Diabetes Res Clin Pract. 2013;99:192-199. [18]

Ehrlich SF, Quesenberry CP, Jr., Van Den Eeden SK, Shan J, Ferrara A. Patients diagnosed with diabetes are at increased risk for asthma, chronic obstructive

Care. 2010;33:55-60. [19]

CR IP T

pulmonary disease, pulmonary fibrosis, and pneumonia but not lung cancer. Diabetes

Yeh HC, Punjabi NM, Wang NY, et al. Cross-sectional and prospective study of lung function in adults with type 2 diabetes: the Atherosclerosis Risk in Communities (ARIC) study. Diabetes Care. 2008;31:741-746.

Lee YH, Han K, Ko SH, Ko KS, Lee KU. Data Analytic Process of a Nationwide

AN US

[20]

Population-Based Study Using National Health Information Database Established by National Health Insurance Service. Diabetes Metab J. 2016;40:79-82. Song SO, Jung CH, Song YD, et al. Background and data configuration process of a

M

[21]

nationwide population-based study using the korean national health insurance system.

Kim L, Kim JA, Kim S. A guide for the utilization of Health Insurance Review and

PT

[22]

ED

Diabetes Metab J. 2014;38:395-403.

Assessment Service National Patient Samples. Epidemiol Health. 2014;36:e2014008. Ma J, Xiao L, Knowles SB. Obesity, insulin resistance and the prevalence of atopy

CE

[23]

and asthma in US adults. Allergy. 2010;65:1455-1463. Liu YC, Zou XB, Chai YF, Yao YM. Macrophage polarization in inflammatory

AC

[24]

diseases. Int J Biol Sci. 2014;10:520-529.

[25]

Lumeng CN, Bodzin JL, Saltiel AR. Obesity induces a phenotypic switch in adipose tissue macrophage polarization. J Clin Invest. 2007;117:175-184.

[26]

Fujisaka S, Usui I, Bukhari A, et al. Regulatory mechanisms for adipose tissue M1 and M2 macrophages in diet-induced obese mice. Diabetes. 2009;58:2574-2582.

ACCEPTED MANUSCRIPT [27]

Melgert BN, ten Hacken NH, Rutgers B, Timens W, Postma DS, Hylkema MN. More alternative activation of macrophages in lungs of asthmatic patients. J Allergy Clin Immunol. 2011;127:831-833.

[28]

Kudo M, Ishigatsubo Y, Aoki I. Pathology of asthma. Front Microbiol. 2013;4:263.

[29]

Walter RE, Beiser A, Givelber RJ, O'Connor GT, Gottlieb DJ. Association between

CR IP T

glycemic state and lung function: the Framingham Heart Study. Am J Respir Crit Care Med. 2003;167:911-916. [30]

Chance WW, Rhee C, Yilmaz C, et al. Diminished alveolar microvascular reserves in type 2 diabetes reflect systemic microangiopathy. Diabetes Care. 2008;31:1596-1601. Williams JG, Morris AI, Hayter RC, Ogilvie CM. Respiratory responses of diabetics

AN US

[31]

to hypoxia, hypercapnia, and exercise. Thorax. 1984;39:529-534. [32]

Schuyler MR, Niewoehner DE, Inkley SR, Kohn R. Abnormal lung elasticity in

[33]

M

juvenile diabetes mellitus. Am Rev Respir Dis. 1976;113:37-41. Klein OL, Krishnan JA, Glick S, Smith LJ. Systematic review of the association

Ruivo S, Viana P, Martins C, Baeta C. Effects of aging on lung function. A

PT

[34]

ED

between lung function and Type 2 diabetes mellitus. Diabet Med. 2010;27:977-987.

comparison of lung function in healthy adults and the elderly. Rev Port Pneumol.

[35]

CE

2009;15:629-653.

Stam H, Hrachovina V, Stijnen T, Versprille A. Diffusing capacity dependent on lung

AC

volume and age in normal subjects. J Appl Physiol (1985). 1994;76:2356-2363.

Figure 1. Among 17,391,531 subjects, 4,237,183 subjects were excluded. Finally, 13,154,348 subjects were enrolled and divided into three groups according to diabetes and retinal

ACCEPTED MANUSCRIPT complications. Subjects were followed-up until primary endpoint occurred. Table 1. Baseline characteristics of subjects according to asthma status Asthma

P value

10,819,045

2,335,303

<.001

48±12.3

52±13.1

<.001

30-39, n (%)

3,033,004 (28.03)

458,187 (19.62)

<.001

40-49, n (%)

3,287,602 (30.39)

573,101 (24.54)

<.001

50-59, n (%)

2,432,933 (22.49)

581,212 (24.89)

<.001

60-69, n (%)

1,377,518 (12.73)

465,774 (19.94)

<.001

70-79, n (%)

593,531 (5.49)

225,432 (9.65)

<.001

80-, n (%)

94,457 (0.87)

31,597 (1.35)

<.001

Age

AN US

n

CR IP T

No asthma

6,254,108 (57.81)

1,107,115 (47.41)

<.001

BMI (kg/m2)

23.8±3.1

23.9±3.6

<.001

Current smoker (%)

26.35

19.76

<.001

Heavy drinker (%)

10.26

8.75

<.001

Physically active subjects (%)

8.38

9.52

<.001

3,016,835 (27.88)

760,801 (32.58)

<.001

1,642,107 (15.18)

415,173 (17.78)

<.001

928,882 (8.59)

220,388 (0.98)

<.001

Hypertension Dyslipidemia

PT

Type 2 diabetes

ED

Comorbidities, n (%)

M

Men, n (%)

CE

Values are expressed as mean ± standard deviation or as %.

AC

Abbreviations: BMI, body mass index

ACCEPTED MANUSCRIPT 14

1

Table 2. Incidence and risk of asthma according to diabetic status

Initial diabetic status Non-DM Without DR

DR

2,114,915

197,407

22,981

27.1

30.1 1.11 (1.105-1.115) 0.95 (0.946-0.954) 0.943 (0.939-0.948)

1 (reference)

38.4

1.407 (1.389-1.425) 1.079 (1.065-1.093) 1.067 (1.053-1.081)

AN US

1 (reference)

CR IP T

Variables Total Asthma cases (n) Asthma incidence rate (per 1,000 person-years) Model 1a HR (95% CI) Model 2b HR (95% CI) Model 3c HR (95% CI)

P for trend

DM

1 (reference)

<.001 <.001 <.001

Abbreviations: DM, diabetes mellitus; DR, diabetic retinopathy; HR, hazard ratio; CI,

3

confidence interval.

4

a

Model 1: unadjusted.

5

b

Model 2: adjusted for age and gender.

6

c

Model 3: adjusted for age, gender, BMI, smoking, alcohol, exercise, hypertension, and

7

dyslipidemia.

10 11 12 13 14 15 16

ED

PT

CE

9

AC

8

M

2

ACCEPTED MANUSCRIPT 15

17 18

Table 3. Incidence and risk of asthma according to diabetic status stratified by age group Initial diabetic status

P for trend

DM Without DR

447,845

10,068

19.8

16.6 0.839 (0.822–0.855) 0.917 (0.899–0.935) 0.916 (0.898–0.934)

1 (reference)

1 (reference) 1,063,947

1.143 (1.016-1.286) 1.21 (1.075-1.362) 1.203 (1.068-1.354) 8,163

24.1

31.8

0.94 (0.933–0.947) 0.957 (0.95–0.964) 0.939 (0.932–0.946)

1.248 (1.222-1.276) 1.145 (1.12-1.17) 1.129 (1.105-1.154)

603,123

105,136

14,544

43.7

41.5

44.2

0.951 (0.944–0.957) 0.945 (0.939–0.951) 0.934 (0.928–0.94)

1.019 (1.003-1.036) 1.017 (1.001-1.034) 1.012 (0.996-1.029)

M

ED

1 (reference) 1 (reference)

PT

1 (reference)

CE

22.3

82,203

25.6

AC

274

AN US

1 (reference)

DR

CR IP T

Non-DM Variable 30–39 Asthma cases (n) Asthma incidence rate (per 1,000person-years) Model 1a HR (95% CI) Model 2b HR (95% CI) Model 3c HR (95% CI) 40–59 Asthma cases (n) Asthma incidence rate (per 1,000person-years) Model 1a HR (95% CI) Model 2b HR (95% CI) Model 3c HR (95% CI) ≥60 Asthma cases (n) Asthma incidence rate (per 1,000person-years) Model 1a HR (95% CI) Model 2b HR (95% CI) Model 3c HR (95% CI)

1 (reference) 1 (reference) 1 (reference)

0.027 0.001 0.002

<.001 <.001 <.001

0.023 0.042 0.147

19 20

Abbreviations: DM, diabetes mellitus; DR, diabetic retinopathy; HR, hazard ratio; CI,

21

confidence interval.

ACCEPTED MANUSCRIPT 16

a

Model 1: unadjusted.

23

b

Model 2: adjusted for age and gender.

24

c

Model 3: adjusted for age, gender, BMI, smoking, alcohol, exercise, hypertension, and

25

dyslipidemia.

CR IP T

22

26 27 28

AN US

29 30 31 32

M

33 34

ED

35 36

PT

37

40

AC

39

CE

38