Inhaled bronchodilators and the risk of tachyarrhythmias

International Journal of Cardiology 190 (2015) 133–139

Contents lists available at ScienceDirect

International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard

Inhaled bronchodilators and the risk of tachyarrhythmias Chang-Hoon Lee a,b, Seongmi Choi a,c, Eun Jin Jang a,d, Han-Mo Yang e, Ho Il Yoon a,f, Yun Jung Kim a, Jimin Kim a, Jae-Joon Yim a,b, Deog Kyeom Kim a,g,⁎ a

National Evidence-based Healthcare Collaborating Agency, Seoul, Republic of Korea Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine, Seoul National University College of Medicine, Seoul National University Hospital, Seoul, Republic of Korea Real Estate R&D Institute, Korea Appraisal Board, Daegu, Republic of Korea d Department of Information Statistics, College of Natural Science, Andong National University, Andong, Republic of Korea e Division of Cardiology, Department of Internal Medicine, Seoul National University College of Medicine, Seoul National University Hospital, Seoul, Republic of Korea f Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine, Seoul National University College of Medicine, Seoul National University Bundang Hospital, Seongnam-Si, Republic of Korea g Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine, Seoul National University College of Medicine, Seoul Metropolitan Government–Seoul National University Boramae Medical Center, Seoul, Republic of Korea b c

a r t i c l e

i n f o

Article history: Received 10 February 2015 Received in revised form 14 April 2015 Accepted 15 April 2015 Available online 17 April 2015 Keywords: Tachyarrhythmias Inhaled β2 agonists Inhaled muscarinic antagonists

a b s t r a c t Background/objectives: There have been controversies about whether inhaled bronchodilators could increase the risk of clinically important tachyarrhythmias. We investigated the association between inhaled bronchodilators and the development of tachyarrhythmias, including atrial fibrillation and other paroxysmal tachyarrhythmias in real practice. Methods: We conducted a nested case–control study with the use of the nationwide insurance claims database of the Health Insurance Review and Assessment Service (Seoul, Republic of Korea). Overall, 3312 cases with newly developed tachyarrhythmias including atrial fibrillation and other paroxysmal tachyarrhythmias and 9732 matched (up to 1:5) controls were identified from 545,508 subjects without acute major cardiovascular events in the past year between January 1, 2011 and December 31, 2011. Conditional logistic regression analysis adjusted by comorbidities, cardiovascular drugs and healthcare utilization was performed. Results: In various multivariate models, the use of inhaled long-acting muscarinic antagonists (LAMAs) or longacting inhaled β2 agonists (LABAs) was significantly associated with tachyarrhythmias. Statistically significant effects of LAMAs on tachyarrhythmias were found only in the non-users of β-blockers. We did not find a statistically significant difference in the impact of a LABA without a LAMA vs a LAMA without a LABA (aOR, 0.93; 95% CI, 0.74–1.18), or a multiplicative or additive interaction between a LABA and a LAMA. Conclusions: Inhaled LAMAs and LABAs were significantly and comparably associated with an increased risk of tachyarrhythmias. © 2015 Published by Elsevier Ireland Ltd.

1. Introduction Inhaled bronchodilators, including inhaled β2-adrenergic receptor agonists and anti-muscarinic antagonists, are the cornerstone therapies for obstructive lung diseases [1,2]. Because β2-agonist use is associated with an increased risk of cardiovascular events [3–6] and antimuscarinic antagonists are known to reduce the parasympathetic activation of heart rate which leads to an increase in tachyarrhythmias and myocardial ischemia [7], studies have raised concerns that inhaled bronchodilator use could also increase the risk of cardiovascular events, including tachyarrhythmia [8–12]. However, there are controversies ⁎ Corresponding author at: Department of Internal Medicine, Seoul National University College of Medicine, Seoul Metropolitan Government Seoul National University Boramae Medical Center, 395 Shindaebang-2-Dong, Dongjak-Gu, Seoul 156-707, Republic of Korea. E-mail address: [email protected] (D.K. Kim).

http://dx.doi.org/10.1016/j.ijcard.2015.04.129 0167-5273/© 2015 Published by Elsevier Ireland Ltd.

concerning the use inhaled bronchodilators and the risk of tachyarrhythmias [13,14]. Additionally, it is uncertain how the drugcontaining device impacts the clinical outcomes [15,16]. We investigated the association between inhaled bronchodilators, focusing on longacting muscarinic antagonists (LAMAs) and the development of clinically important tachyarrhythmias including atrial fibrillation and other paroxysmal tachyarrhythmias in the clinic. 2. Methods 2.1. Data source We used the Health Insurance Review and Assessment Service (HIRA; Seoul, South Korea) database, which includes 50.9 million South Koreans from the National Health Insurance (NHI) and National Medical Aid (NMA) databases. The HIRA database contains information

134

C.-H. Lee et al. / International Journal of Cardiology 190 (2015) 133–139

on demographics and all of the medical services rendered, along with diagnostic codes (International Statistical Classification of Diseases and Related Health Problems, 10th edition code, ICD-10 code) and all of the medications prescribed. Key fields, such as drug name, quantity, date dispensed, and duration, are missing or out of range in only b0.5% of the records. This study was approved by the ethics review committee of the National Evidence-based Healthcare Collaborating Agency, Seoul, Republic of Korea. Informed consent was waived by ethics review committee of the National Evidence-based Healthcare Collaborating Agency, Seoul, Republic of Korea. 2.2. Study design and study population A nested case–control study was conducted based on the HIRA database. The source population consisted of all of the individuals who were dispensed inhaled respiratory drugs for 30 days or longer between 1 January 2011 and 31 December 2011. The initiation date was defined as the date of the first use of the inhaled respiratory drugs during either hospitalization or an outpatient visit. We excluded the following patients from this cohort: those who had prescriptions for inhaled respiratory drugs for 30 days or longer during the year prior to the initiation date, those who were diagnosed as having acute major cardiovascular events including acute myocardial infarction, stroke and tachyarrhythmias during the year prior to the initiation date, and those who were under 20 years of age or over 100 years of age. The detailed patient selection flow is presented in Fig. 1, and the final eligible cohort included 226,588 new users of inhaled respiratory drugs. 2.3. Definition of cases with tachyarrhythmia Within the eligible cohort, we identified individuals based on the ICD-10 diagnoses of tachyarrhythmia (ICD-10 code I47–I48) that occurred after the initiation date of the inhaled respiratory drugs. The date of the first assignment of the tachyarrhythmia ICD-10 code was called the index date. 2.4. Definition of controls We performed individual matching to select control patients per case. The control patients were selected from patients without ICD-10 codes for tachyarrhythmia. Each case was matched with up to five

controls based on matching variables such as age (±5 years old), sex, initiation date of inhalers (±15 days), diagnosis of hypertension (ICD10 code I10–I15), diabetes mellitus (DM; ICD-10 code E10–E14), COPD (ICD-10 code J41), ischemic heart disease (IHD; ICD-10 code I20, I25), diagnosis of other heart diseases during one year before the index date and the Charlson Comorbidity Index (CCI) score during one year before the index date. Other heart diseases were defined as rheumatic disease (ICD-10 code I00–I09) and cardiomyopathies, arrhythmias, valvular diseases and pericardial diseases (ICD-10 code I30–I52). The CCI variable was categorized into the following 3 groups: 0–1, 2– 3, and ≥ 4. The index date for the controls was defined as the index date of the matched case. 2.5. Exposure to inhaled medications Inhaled drugs included inhaled corticosteroids (ICSs; beclomethasone, budesonide, triamcinolone, ciclesonide, fluticasone, or flunisolide), a short-acting inhaled β2-agonist (SABA; salbutamol, fenoterol, procaterol, or terbutaline), a long-acting inhaled β2-agonist (LABA; salmeterol or formoterol), a LAMA (tiotropium), or a combination of an ICS and a LABA (budesonide/formoterol or fluticasone/salmeterol). We classified the inhaler users as using inhaled drugs for 30 days or longer during one year, while respiratory drugs requiring a nebulizer were excluded in this study. When we assess the risk of tachyarrhythmia by each inhaler, each inhaler user was defined as having an inhaler prescription for 30 days or longer during the 90-day period before the index date. If each inhaler prescription was for less than 30 days during the 90-day period before the index date, the patients were considered non-users. 2.6. Covariates We considered the covariates for tachyarrhythmia risk adjustment as follows: other chronic respiratory disease, comorbidities, health care utilization, and concomitant medications. Other chronic respiratory diseases were classified as tuberculosis-lung (ICD-10 code B90), bronchiectasis (ICD-10 code J47), asthma (ICD-10 code J45–46), and others. Comorbidities included chronic kidney disease or dialysis (ICD-10 code N17–N19), and dyslipidemia (ICD-10 code E780, E789). We used health care utilization such as number of hospitalization (0, 1, ≥2), outpatient visit (b15, 15–30, 31–50, N50), and emergency room (ER) visit (0, ≥1)

Fig. 1. Flowchart for patient selection.

C.-H. Lee et al. / International Journal of Cardiology 190 (2015) 133–139 Table 1 Baseline characteristics of cases with tachyarrhythmias and controls. Tachyarrhythmia Control (N = 3312) (N = 9732)

Sex Men Women Agea Mean ± SD 20–49 50–59 60–69 70–79 ≥80 Hypertension Diabetes mellitus Ischemic heart disease Dyslipidemiab Chronic kidney diseaseb COPD Respiratory diagnosis other than COPDb,c TB-lung Bronchiectasis Asthma others Current concomitant medicationd ACEI/ARB β-blocker Statin Aspirin Thiazide CCB β-Agonists Methylxanthines Anti-arrhythmic drugs Digitalis or other anti-HF drugs Concomitant medicatione ACEI/ARB β-Blocker Statin Aspirin Thiazide CCB β-Agonists Methylxanthines Anti-arrhythmic drugs Digitalis or other anti-HF drugs MPR of concomitant medicationf ACEI/ARB Mean ± SD Median (Q1, Q3) 0 0 b ≤0.3 0.3 b ≤0.7 0.7 b ≤1 β-Blocker Mean ± SD Median (Q1, Q3) 0 0 b ≤0.3 0.3 b ≤0.7 0.7 b ≤1 Statin Mean ± SD Median (Q1, Q3) 0 0 b ≤0.3 0.3 b ≤0.7 0.7 b ≤1 Aspirin Mean ± SD Median (Q1, Q3) 0 0 b ≤0.3 0.3 b ≤0.7 0.7 b ≤1

135

Table 1 (continued) Tachyarrhythmia Control (N = 3312) (N = 9732)

p-Value

n

n

(%)

n

(%)

1570 1742

(47.4%) (52.6%)

4266 (43.8%) Matched 5466 (56.2%)

68.8 ± 12.5 264 (8.0%) 390 (11.8%) 767 (23.2%) 1316 (39.7%) 575 (17.4%) 2292 (69.2%) 1206 (36.4%) 937 (28.3%) 633 (19.1%) 304 (9.2%) 1860 (56.2%)

69.8 ± 11.5 566 (5.8%) 978 (10.0%) 2372 (24.4%) 4081 (41.9%) 1735 (17.8%) 7274 (74.7%) 3763 (38.7%) 2845 (29.2%) 1931 (19.8%) 887 (9.1%) 5787 (59.5%)

Matched

161 205 2515 431

(4.9%) (6.2%) (75.9%) (13.0%)

336 669 7274 1453

(3.5%) b.001 (6.9%) (74.7%) (14.9%)

474 276 246 314 266 447 80 362 25 129

(14.3%) (8.3%) (7.4%) (9.5%) (8.0%) (13.5%) (2.4%) (10.9%) (0.8%) (3.9%)

1654 889 896 1052 879 1232 254 947 27 320

(17.0%) (9.1%) (9.2%) (10.8%) (9.0%) (12.7%) (2.6%) (9.7%) (0.3%) (3.3%)

b.001 0.162 0.002 0.031 0.079 0.214 0.541 0.047 b.001 0.098

1053 623 592 755 613 992 205 906 45 255

(31.8%) (18.8%) (17.9%) (22.8%) (18.5%) (30.0%) (6.2%) (27.4%) (1.4%) (7.7%)

3662 2146 2137 2423 2087 2766 513 2455 88 691

(37.6%) (22.1%) (22.0%) (24.9%) (21.4%) (28.4%) (5.3%) (25.2%) (0.9%) (7.1%)

b.001 b.001 b.001 0.015 0.000 0.093 0.045 0.016 0.025 0.251

Matched Matched Matched 0.362 0.911 Matched

0.2 ± 0.3 0 (0, 0.54) 2099 (63.4%) 193 (5.8%) 490 (14.8%) 530 (16.0%)

0.3 ± 0.4 b.001 0 (0,0.62) 5453 (56.0%) b.001 726 (7.5%) 1701 (17.5%) 1852 (19.0%)

0.1 ± 0.3 0 (0, 0.011) 2470 (74.6%) 258 (7.8%) 337 (10.2%) 247 (7.5%)

0.2 ± 0.3 b.001 0 (0,0.078) 6961 (71.5%) b.001 762 (7.8%) 1017 (10.5%) 992 (10.2%)

0.1 ± 0.3 0 (0, 0) 2627 (79.3%) 114 (3.4%) 282 (8.5%) 289 (8.7%)

0.2 ± 0.3 b.001 0 (0, 0.078) 7200 (74.0%) b.001 439 (4.5%) 1052 (10.8%) 1041 (10.7%)

0.2 ± 0.3 0 (0, 0.011) 2380 (71.9%) 201 (6.1%) 382 (11.5%) 349 (10.5%)

0.2 ± 0.3 0.013 0 (0, 0.23) 6848 (70.4%) 0.130 562 (5.8%) 1160 (11.9%) 1162 (11.9%)

Thiazide Mean ± SD Median (Q1, Q3) 0 0 b ≤0.3 0.3 b ≤0.7 0.7 b ≤1 CCB Mean ± SD Median (Q1, Q3) 0 0 b ≤0.3 0.3 b ≤0.7 0.7 b ≤1 β-Agonists Mean ± SD Median (Q1, Q3) 0 0 b ≤0.3 0.3 b ≤0.7 0.7 b ≤1 Methylxanthines Mean ± SD Median (Q1, Q3) 0 0 b ≤0.3 0.3 b ≤0.7 0.7 b ≤1 Anti-arrhythmic drugs Mean ± SD Median (Q1, Q3) 0 0 b ≤0.3 0.3 b ≤0.7 0.7 b ≤1 Digitalis or other anti-HF drugs Mean ± SD Median (Q1, Q3) 0 0 b ≤0.3 0.3 b ≤0.7 0.7 b ≤1 Health care utilizationg Number of hospitalization Mean ± SD Median (Q1, Q3) 0 1 ≥2 Number of outpatient visit Mean ± SD Median (Q1, Q3) b15 15–30 31–50 N50 Number of ER visit Mean ± SD Median (Q1, Q3) 0 ≥1

(%)

n

p-Value

(%)

0.1 ± 0.3 0 (0, 0) 2554 (77.1%) 180 (5.4%) 313 (9.5%) 265 (8.0%)

0.1 ± 0.3 0.001 0 (0, 0.033) 7207 (74.1%) 0.003 594 (6.1%) 984 (10.1%) 947 (9.7%)

0.2 ± 0.3 0 (0, 0.42) 2108 (63.6%) 256 (7.7%) 466 (14.1%) 482 (14.6%)

0.2 ± 0.3 0.152 0 (0, 0.37) 6344 (65.2%) 0.268 737 (7.6%) 1248 (12.8%) 1403 (14.4%)

1.1 ± 0.4 1 (1, 1) 2993 (90.4%) 224 (6.8%) 83 (2.5%) 12 (0.4%)

1.1 ± 0.4 0.513 1 (1, 1) 8892 (91.4%) 0.035 532 (5.5%) 260 (2.7%) 48 (0.5%)

1.6 ± 0.9 1 (1, 2) 2085 (63.0%) 606 (18.3%) 379 (11.4%) 242 (7.3%)

1.6 ± 1.0 0.298 1 (1,2) 6334 (65.1%) 0.011 1634 (16.8%) 980 (10.1%) 784 (8.1%)

1.0 ± 0.3 1 (1, 1) 3217 (97.1%) 56 (1.7%) 31 (0.9%) 8 (0.2%)

1.0 ± 0.2 0.006 1 (1, 1) 9578 (98.4%) b.001 69 (0.7%) 50 (0.5%) 35 (0.4%)

1.2 ± 0.6 1 (1, 1) 2926 (88.3%) 157 (4.7%) 144 (4.3%) 85 (2.6%)

1.2 ± 0.6 0.227 1 (1, 1) 8708 (89.5%) 0.003 433 (4.4%) 299 (3.1%) 292 (3.0%)

1.4 ± 2.4 1 (0,2) 1645 (49.7%) 731 (22.1%) 936 (28.3%)

1.6 ± 2.4 b.001 1 (0,2) 3892 (40.0%) b.001 2391 (24.6%) 3449 (35.4%)

44.9 ± 38.3 34 (20, 57) 518 (15.6%) 945 (28.5%) 819 (24.7%) 1030 (31.1%)

44.6 ± 38.4 0.750 34 (20, 57) 1529 (15.7%) 0.800 2779 (28.6%) 2475 (25.4%) 2949 (30.3%)

0.8 ± 2.3 0 (0, 1) 2123 (64.1%) 1189 (35.9%)

0.9 ± 3.5 0.008 0 (0, 1) 5590 (57.4%) b.001 4142 (42.6%)

p-values were derived from independent t-test for continuous variables and χ2-test for categorical variables, respectively. a Age at initiation date. b During 1-year period before index date until index date. c Respiratory disease priority: TB-lung N Bronchiectasis N Asthma N Others. d 14 days or longer within 30 days prior to index date. e Either more than 30 days or more than twice on prescription within 90 days prior to index date. f Medication possession ratio (MPR) within 90 days prior to index date. g Within 1-year prior to index date.

136

C.-H. Lee et al. / International Journal of Cardiology 190 (2015) 133–139

to adjust patient severity. Concomitant medications included angiotensin converting enzyme inhibitors (ACEIs)/angiotensin receptor blockers (ARBs), β-blockers, statins, aspirin, thiazide, calcium-channel blockers (CCBs), β-agonists, methylxanthines, anti-arrhythmic drugs, and inotropics (digitalis and anti-HF drugs). 2.7. Statistical analysis Baseline characteristics of cases and controls were summarized by descriptive statistics such as proportion, mean, standard deviation (SD), median, first quartile (Q1), and third quartile (Q3). We also

summarized the continuous variables into the appropriated categorical variables based on their distributions. An independent t-test for continuous variables and a χ2-test for categorical variables were used to determine statistical significance. The association between the use of inhaled respiratory medication and tachyarrhythmia was investigated by conditional logistic regression analysis. We adjusted the following covariates: age, other chronic respiratory disease, chronic kidney disease or dialysis, dyslipidemia, use of concomitant medications, number of hospitalization, outpatient visit, and ER visit. The unadjusted odds ratios (ORs) and adjusted odds ratios (aORs) are presented with 95% confidence interval (CI).

Table 2 Risk of tachyarrhythmias according to inhaled drug use.

Model 1 ICS LABA Neither ICS nor LABA ICS without LABA ICS with LABA LABA LAMA SABA Model 2 ICS LABA Neither ICS nor LABA ICS without LABA ICS with LABA LABA LAMA No LAMA LAMA SMI LAMA DPI Both LAMA SMI and DPI SABA Model 3 ICS LABA Neither ICS nor LABA ICS without LABA ICS with LABA LABA LAMA MPR 0 0 b ≤0.25 0.25 b ≤0.5 0.5 b ≤0.75 0.75 b ≤1 SABA Model 4 ICS LABA Neither ICS nor LABA ICS without LABA ICS with LABA LABA LAMA prescription days 0 0 b ≤30 30 b ≤90 SABA Model 5 Neither LABA nor LAMA LABA without LAMA LAMA without LABA Both LABA and LAMA LABA vs LAMA a

Adjusteda

Adjustedb

Tachyarrhythmia (N = 3312)

Control (N = 9732)

Unadjusted

n

(%)

n

(%)

OR (95% CI)

p-Value

OR (95% CI)

p-Value

OR (95% CI)

p-Value

2647 59 562 44 241 315

(79.9%) (1.8%) (17.0%) (1.3%) (7.3%) (9.5%)

8017 144 1475 96 629 822

(82.4%) (1.5%) (15.2%) (1.0%) (6.5%) (8.4%)

– 1.2 (0.87, 1.66) 1.2 (1.06, 1.35) 1.34 (0.92, 1.96) 1.3 (1.09, 1.54) 1.09 (0.94, 1.27)

– 0.269 0.003 0.131 0.003 0.262

– 1.21 (0.87, 1.67) 1.18 (1.05, 1.34) 1.35 (0.93, 1.98) 1.27 (1.07, 1.51) 1.1 (0.95, 1.28)

– 0.253 0.006 0.117 0.006 0.208

– 1.19 (0.85, 1.66) 1.17 (1.03, 1.32) 1.27 (0.86, 1.87) 1.27 (1.06, 1.51) 1.09 (0.94, 1.28)

– 0.308 0.017 0.221 0.008 0.264

2647 59 562 44

(79.9%) (1.8%) (17.0%) (1.3%)

8017 144 1475 96

(82.4%) (1.5%) (15.2%) (1.0%)

– 1.2 (0.87, 1.66) 1.2 (1.06, 1.35) 1.34 (0.92, 1.96)

– 0.269 0.003 0.131

– 1.21 (0.87, 1.68) 1.18 (1.05, 1.34) 1.36 (0.93, 1.98)

– 0.249 0.006 0.116

– 1.19 (0.85, 1.66) 1.17 (1.03, 1.32) 1.28 (0.87, 1.88)

– 0.302 0.016 0.217

2990 27 211 84 315

(90.3%) (0.8%) (6.4%) (2.5%) (9.5%)

8876 67 555 234 822

(91.2%) (0.7%) (5.7%) (2.4%) (8.4%)

– 1.25 (0.78, 1.99) 1.3 (1.08, 1.56) 1 (0.77, 1.31) 1.09 (0.94, 1.27)

– 0.347 0.005 0.987 0.262

– 1.21 (0.76, 1.93) 1.28 (1.06, 1.53) 1 (0.76, 1.31) 1.1 (0.95, 1.28)

– 0.419 0.009 1.000 0.203

– 1.18 (0.73, 1.9) 1.28 (1.06, 1.54) 1.02 (0.77, 1.34) 1.09 (0.94, 1.28)

– 0.497 0.01 0.901 0.258

2647 59 562 44

(79.9%) (1.8%) (17.0%) (1.3%)

8017 144 1475 96

(82.4%) (1.5%) (15.2%) (1.0%)

– 1.2 (0.87, 1.66) 1.2 (1.06, 1.35) 1.34 (0.92, 1.96)

– 0.269 0.003 0.131

– 1.21 (0.87, 1.67) 1.18 (1.05, 1.33) 1.36 (0.93, 1.99)

– 0.252 0.007 0.112

– 1.19 (0.85, 1.66) 1.16 (1.03, 1.32) 1.28 (0.87, 1.88)

– 0.305 0.018 0.216

2994 62 131 52 73

(90.4%) (1.9%) (4.0%) (1.6%) (2.2%)

8891 189 311 151 190

(91.4%) (1.9%) (3.2%) (1.6%) (2.0%)

– 0.85 (0.62, 1.15) 1.42 (1.13, 1.78) 1.18 (0.85, 1.65) 1.38 (1.03, 1.85)

– 0.287 0.003 0.320 0.031

– 0.84 (0.62, 1.15) 1.4 (1.12, 1.76) 1.16 (0.83, 1.62) 1.34 (1.003, 1.8)

– 0.279 0.004 0.390 0.048

315

(9.5%)

822

(8.4%)

1.09 (0.94, 1.27)

0.262

1.1 (0.95, 1.29)

0.202

– 0.85 (0.62, 1.16) 1.42 (1.13, 1.79) 1.15 (0.82, 1.61) 1.32 (0.98, 1.78) p for trend 1.09 (0.94, 1.28)

– 0.304 0.003 0.417 0.064 0.009 0.262

2647 59 562 44

(79.9%) (1.8%) (17.0%) (1.3%)

8017 144 1475 96

(82.4%) (1.5%) (15.2%) (1.0%)

– 1.2 (0.87, 1.66) 1.2 (1.06, 1.35) 1.34 (0.92, 1.96)

– 0.269 0.003 0.131

– 1.21 (0.88, 1.68) 1.19 (1.05, 1.34) 1.36 (0.93, 1.99)

– 0.247 0.005 0.112

– 1.19 (0.86, 1.66) 1.17 (1.03, 1.33) 1.28 (0.87, 1.88)

– 0.298 0.014 0.214

2990 150 172

(90.3%) (4.5%) (5.2%)

8876 376 480

(91.2%) (3.9%) (4.9%)

– 1.18 (0.96, 1.45) 1.23 (1.01, 1.49)

– 0.124 0.039

– 1.17 (0.95, 1.44) 1.2 (0.99, 1.46)

– 0.135 0.070

315

(9.5%)

822

(8.4%)

1.09 (0.94, 1.27)

0.262

1.1 (0.95, 1.28)

0.204

– 1.19 (0.97, 1.47) 1.18 (0.97, 1.44) p for trend 1.09 (0.94, 1.28)

– 0.1 0.099 0.051 0.258

2556 515 150 91

(77.2%) (15.5%) (4.5%) (2.7%)

7750 1353 411 218

(79.6%) (13.9%) (4.2%) (2.2%)

– 1.18 (1.04, 1.34) 1.27 (1.03, 1.56) 1.5 (1.15, 1.96)

– 0.011 0.024 0.002

– 1.19 (1.05, 1.35) 1.27 (1.03, 1.57) 1.5 (1.15, 1.96) 0.93 (0.74, 1.17)

– 0.007 0.022 0.003 0.553

– 1.16 (1.02, 1.32) 1.24 (1.005, 1.54) 1.51 (1.15, 1.98) 0.93 (0.74, 1.18)

– 0.028 0.045 0.003 0.553

Adjusted by other inhaled medication. Adjusted by other inhaled medication, age, other chronic respiratory disease, chronic kidney disease or dialysis, dyslipidemia, number of hospitalization, number of outpatient visit, number of ER visit, concomitant medication of ACEI/ARB, beta-blocker, statin, aspirin, thiazide, CCB, β-agonists, methylxanthines, anti-arrhythmic drugs, digitalis or other anti-HF drugs. b

C.-H. Lee et al. / International Journal of Cardiology 190 (2015) 133–139

Subgroup analyses for LAMAs and LABAs were conducted according to beta-blocker use, IHD, DM, hypertension, and COPD. A p-value of less than 0.05 was regarded as significant, and all of the statistical analyses were performed using SAS V.9.2 (SAS Institute, Cary, NC, USA). 3. Results In total, 545,508 individuals with prescriptions of inhaled respiratory drugs for 30 days or longer between January 1, 2011 and December 31, 2011 were identified from the database. Of these individuals, the following were excluded: 248,344 individuals with previous prescriptions for inhaled respiratory drugs for 30 days or longer during the year before the current initiation of inhaled respiratory medication; 21,360 individuals diagnosed as having any tachyarrhythmias during the 1-year period before the index date; and 49,216 individuals who were b20 years old, N100 years old or had an unknown age. Finally, a cohort of 226,588 new users of inhaled respiratory drugs was identified. During the study period, 4918 individuals in this cohort were diagnosed with tachyarrhythmia. After excluding 1606 (32.7%) cases without matched controls, 3312 cases with tachyarrhythmia and 9732 matched controls were included in the analysis (Fig. 1). Because of the large sample size, there were significant differences in several covariates, including chronic respiratory diseases other than COPD. However, the concomitant use of several drugs and the number of hospitalization or ER visits within the previous year, majority of covariates were well balanced between cases with tachyarrhythmia and controls because of extensive matching (Table 1). We used five statistical models to evaluate the association between inhaled LAMAs and tachyarrhythmia. In model 1 (main model), LAMAs significantly increased the risk of tachyarrhythmia (aOR, 1.27; 95% CI, 1.06–1.51). In model 2, we evaluated the difference in the impact on the risk of tachyarrhythmias between LAMAs in a DPI device and LAMAs in a SMI device. Although only the LAMAs in a DPI were significantly associated with tachyarrhythmia (aOR, 1.28; 95% CI, 1.06–1.54), LAMAs in a SMI also showed aOR higher than 1. Model 3 was the same as model 1, except we evaluated LAMAs in a MPR instead of the general use of LAMAs (answered as ‘yes’ or ‘no’), which showed partly the dose response between LAMA in a MPR and the risk of tachyarrhythmias (p for trend = 0.009). Model 4 used model 1 but replaced LAMA use with LAMA prescription days. Using this comparison, we found marginal statistical significance in the association between LAMA prescription days and the risk of tachyarrhythmia (p for trend = 0.051). In models 1 and 2, ICSs combined with LABAs were also significantly associated with tachyarrhythmias (model 1; aOR, 1.17; 95% CI, 1.03–1.32, model 2; aOR, 1.17; 95% CI, 1.03–1.32). Although LABAs without ICSs showed

137

no statistical significance, the aORs of LABAs were higher than those of the ICSs combined with LABAs. Therefore, in model 5, we analyzed the impact of either LAMAs or LABAs and both on the risk of tachyarrhythmias. LAMAs without LABAs and LABAs without LAMAs showed a similar impact on the development of tachyarrhythmia (LABAs without LAMAs vs LAMAs without LABAs, aOR, 0.93; 95% CI, 0.74–1.18). There were no significant multiplicative (P for interaction = 0.775) or additive (RERI = 0.113; 95% CI, −0.363 to 0.589) interactions between LABAs and LAMAs (Table 2). In the subgroup analyses, the statistically significant effects of LAMAs on tachyarrhythmias were found in the β-blocker non-users (Fig. 2). These impacts of LAMAs on tachyarrhythmias remained statistically significant even in the sensitivity analyses, where the drug prescriptions within 1 or 4 weeks before the event were excluded (Table 3). 4. Discussion 4.1. Comments on this study In these various covariates (age, sex, hypertension, diabetes mellitus, ischemic heart disease and COPD)-matched case–control study with an adjustment for other covariates to reduce selection bias, LAMAs significantly increased the risk of tachyarrhythmias by 20–30% compared with no LAMA use. There were some dose–response relationships, although definite linear associations were not observed. Recently, there have been concerns that LAMAs in a SMI could increase the cardiovascular risk [17]. Our study found that the use of LAMAs in a SMI does not have a higher risk of tachyarrhythmias than does the use of LAMAs in a DPI. Interestingly, the sensitivity analyses showed that the impacts of LAMAs on tachyarrhythmias remained unchanged after excluding the prescriptions of drugs within 1 or 4 weeks before the event. This suggests that not only was there no bias but also the effects of LAMAs on the heart could persist for a longer time than expected. In our study, LABAs also increased the risk of tachyarrhythmias. In the main model, LABAs combined with ICSs significantly increased the risk of tachyarrhythmias. However, the effects of the LABAs combined with ICSs on the occurrence of tachyarrhythmia are thought to be mainly due to the effects of LABAs, although there are some papers reporting the association between corticosteroids and tachyarrhythmias [18–20]. The aORs of LABAs were at least similar to those of ICSs with LABAs, and the small number of users is possibly responsible for the failure to show statistical significance. In model 5, both LAMA and LABA use showed the highest aORs compared with either LAMAs without LABAs or LABAs without LAMAs (LABAs without LAMAs vs neither LAMAs nor LABAs, aOR, 1.16; 95% CI, 1.02–1.32, LAMAs without LABAs vs neither LAMAs

Fig. 2. Subgroup analysis for the association between inhaled LAMA use and the risk of tachyarrhythmias.

138

C.-H. Lee et al. / International Journal of Cardiology 190 (2015) 133–139

nor LABAs, aOR, 1.24; 95% CI, 1.005–1.54, both vs neither, aOR 1.51, 95% CI, 1.15–1.98). However, we did not find a statistically significant additive or synergistic interaction between LABAs and LAMAs on tachyarrhythmias. There was no significant difference in the tachyarrhythmia risk between LABAs and LAMAs (LABAs without LAMAs vs LAMAs without LABAs, aOR, 0.93; 95% CI, 0.74–1.18), which corresponds to recent cohort studies [12,21]. Our study showed that inhaled LABAs and LAMAs, which are commonly used in patients with respiratory diseases, can increase the risk of tachyarrhythmias. Given that we excluded those who were diagnosed as having acute major cardiovascular events during the year prior to the initiation date, matching variables included ischemic and other heart diseases and comorbidity index, and adjusted the number of hospitalization and other concomitant drug prescription, our results could reveal that tachyarrhythmias can be induced even in a structurally and functionally normal heart given sufficient hypoxic stress and intake of drug. Sympathetic system activation is most likely the mechanism that is responsible for the occurrence of tachyarrhythmia with inhaled LAMAs and LABAs. In fact, in our study, there were no statistically significant increases in the risk of tachyarrhythmias in either the LAMA patients (Fig. 2) or the ICSs with LABA users (data not shown) who concomitantly used β-blockers, which also suggests the importance of an increase in heart rate. β-Adrenergic receptors are subdivided into three groups: β1, β2 and β3. Although they are classically distributed in the heart, airway smooth muscle and adipose tissue, respectively, the β1- and β2-adrenergic receptors are nearly 50% identical in the amino acid sequences. In addition, β2-receptors are also observed in

cardiac muscle. Therefore, β2-agonists, including new selective β2agonists (salmeterol and formoterol), increase the heart rate [22–24]. Muscarinic (M3) receptor antagonists also inhibit the parasympathetic control of the heart rate [7,10]. The principle that a receptor agonist activity is concentration dependent and certainly not absolute could apply to inhaled β-agonists and muscarinic antagonists. The proinflammatory effects of β2-agonists [25] and LAMAs [26] might also contribute to the development of tachyarrhythmias. Our study has several strengths. Most of all, many observable variables including comorbidities, drugs and health care utilization were well-balanced between the cases in whom tachyarrhythmias developed and the controls as shown in Table 1. We also performed sensitivity analyses in which the use of drugs just before (1 week or 4 weeks) the events was excluded to rule out a protopathic bias [27]. The LAMAs remained statistically significant in all of the models (Table 3). Furthermore, our nationwide database-based study can more reflect a more accurate clinical picture. The majority of RCTs regarding LABA safety used LABA-single drugs, but in actual practice most of the LABAs are used as a combination drug with ICSs, as represented in our study. In addition, RCTs have narrow eligible criteria. Although the post-hoc study of the TORCH trial showed that salmeterol alone or in combination with fluticasone did not increase the risk of cardiovascular events among COPD patients [28], it is unknown how many participants at risk for arrhythmias were included in the trial, despite defining one of the exclusion criteria as “other conditions likely to interfere with the study or cause death within 3 years” [29]. Additionally, our datasets included nearly all of the Korean population, leading to a large number of tachyarrhythmic cases (n = 3312) being included.

Table 3 Sensitivity analysis after exclusion of recent prescriptions.

Model 1 1) Excluding prescriptions within 1 week before the event ICS LABA Neither ICS nor LABA ICS without LABA ICS with LABA LABA LAMA SABA 2) Excluding prescriptions within 4 weeks before the event ICS LABA Neither ICS nor LABA ICS without LABA ICS with LABA LABA LAMA SABA Model 5 1) Excluding prescriptions within 1 week before the event Neither LABA nor LAMA LABA without LAMA LAMA without LABA Both LABA and LAMA 2) Excluding prescriptions within 4 weeks before the event Neither LABA nor LAMA LABA without LAMA LAMA without LABA Both LABA and LAMA a

Adjusteda

Adjustedb

Tachyarrhythmia (N = 3312)

Control (N = 9732)

Unadjusted

n

(%)

n

(%)

OR (95% CI)

p-Value

OR (95% CI)

p-Value

OR (95% CI)

p-Value

2732 54 490 36 218 263

(82.5%) (1.6%) (14.8%) (1.1%) (6.6%) (7.9%)

8209 129 1307 87 569 703

(84.4%) (1.3%) (13.4%) (0.9%) (5.8%) (7.2%)

– 1.22 (0.87, 1.71) 1.2 (1.05, 1.36) 1.24 (0.82, 1.87) 1.29 (1.08, 1.54) 1.09 (0.92, 1.28)

– 0.256 0.005 0.308 0.005 0.307

– 1.22 (0.87, 1.72) 1.18 (1.04, 1.34) 1.25 (0.83, 1.88) 1.27 (1.06, 1.52) 1.1 (0.93, 1.29)

– 0.244 0.010 0.293 0.009 0.273

– 1.19 (0.84, 1.69) 1.16 (1.02, 1.32) 1.17 (0.77, 1.78) 1.26 (1.05, 1.51) 1.08 (0.91, 1.27)

– 0.319 0.024 0.462 0.014 0.384

2908 31 353 20 168 169

(87.8%) (0.9%) (10.7%) (0.6%) (5.1%) (5.1%)

8641 100 934 57 444 441

(88.8%) (1.0%) (9.6%) (0.6%) (4.6%) (4.5%)

– 0.85 (0.55, 1.3) 1.19 (1.03, 1.38) 1.04 (0.61, 1.77) 1.28 (1.05, 1.56) 1.14 (0.93, 1.39)

– 0.448 0.016 0.896 0.014 0.202

– 0.85 (0.55, 1.3) 1.17 (1.01, 1.35) 1.02 (0.6, 1.75) 1.25 (1.03, 1.53) 1.12 (0.92, 1.37)

– 0.444 0.033 0.944 0.026 0.248

– 0.82 (0.53, 1.27) 1.15 (0.99, 1.33) 0.97 (0.56, 1.67) 1.24 (1.01, 1.52) 1.1 (0.89, 1.34)

– 0.369 0.069 0.913 0.037 0.377

2642 452 144 74

(79.8%) (13.6%) (4.3%) (2.2%)

7954 1209 384 185

(81.7%) (12.4%) (3.9%) (1.9%)

– 1.18 (1.03, 1.35) 1.29 (1.04, 1.59) 1.45 (1.09, 1.93)

– 0.014 0.200 0.012

– 1.19 (1.04, 1.36) 1.29 (1.05, 1.6) 1.44 (1.08, 1.93)

– 0.011 0.180 0.120

– 1.16 (1.01, 1.32) 1.25 (1.005, 1.55) 1.47 (1.09, 1.97)

– 0.038 0.045 0.011

2820 324 119 49

(85.1%) (9.8%) (3.6%) (1.5%)

8433 855 308 136

(86.7%) (8.8%) (3.2%) (1.4%)

– 1.21 (1.04, 1.41) 1.34 (1.06, 1.69) 1.28 (0.91, 1.81)

– 0.014 0.013 0.161

– 1.2 (1.03, 1.4) 1.35 (1.07, 1.7) 1.26 (0.89, 1.78)

– 0.019 0.012 0.188

– 1.17 (0.997, 1.36) 1.31 (1.03, 1.66) 1.26 (0.89, 1.8)

– 0.054 0.025 0.193

Adjusted by other inhaled medication. Adjusted by other inhaled medication, age, other chronic respiratory disease, chronic kidney disease or dialysis, dyslipidemia, number of hospitalization, number of outpatient visit, number of ER visit, concomitant medication of ACEI/ARB, beta-blocker, statin, aspirin, thiazide, CCB, β-agonists, methylxanthines, anti-arrhythmic drugs, digitalis or other anti-HF drugs. b

C.-H. Lee et al. / International Journal of Cardiology 190 (2015) 133–139

4.2. Limitations There are also several limitations. Our study did not identify serious or fatal cases, although some have questioned the true clinical impact of inhaled drug-associated tachyarrhythmias [30]. As tachyarrhythmia could be promoted in concert with hypoxic lung disease and/or variable myocardial or coronary perfusion abnormalities, unidentified myocardial ischemia or dysfunction, and severity of hypoxia could not be covered in our matching procedures because of the limitations in the analysis for the claims database, which might be a bias. This study did not include any new LABAs and LAMAs. The new LABAs, including indacaterol, vilanterol and olodaterol, and the new LAMAs, including glycopyrronium, umeclidinium and aclidinium, were not introduced to South Korea until 2011. Although there are studies reporting the association between beta adrenergic receptor polymorphisms and tachyarrhythmia [31,32], we could not get the genomic information of our participants. Further studies are needed. 4.3. Conclusions In conclusion, inhaled LAMAs and LABAs were significantly and comparably associated with an increased risk of tachyarrhythmias. Risk–benefit balance of drugs should be considered before selecting bronchodilators. Grant support This study was supported by the National Evidence-based Healthcare Collaborating Agency (NA13-004), which contributed to the study's design, conduct, and reporting. Conflict of interest The authors report no relationships that could be construed as a conflict of interest. References [1] Global Initiative for Asthma, Global Strategy for Asthma Management and Prevention, 2014. [2] Global Initiative for Chronic Obstructive Lung Disease, Global Strategy for the Diagnosis, Management, and Prevention of Chronic Obstructive Pulmonary Disease, 2014. [3] S. Suissa, B. Hemmelgarn, L. Blais, P. Ernst, Bronchodilators and acute cardiac death, Am. J. Respir. Crit. Care Med. 154 (1996) 1598–1602. [4] R.M. Martin, N.R. Dunn, S.N. Freemantle, R.D. Mann, Risk of non-fatal cardiac failure and ischaemic heart disease with long acting beta 2 agonists, Thorax 53 (1998) 558–562. [5] D.H. Au, R.N. Lemaitre, J.R. Curtis, N.L. Smith, B.M. Psaty, The risk of myocardial infarction associated with inhaled beta-adrenoceptor agonists, Am. J. Respir. Crit. Care Med. 161 (2000) 827–830. [6] D.H. Au, J.R. Curtis, N.R. Every, M.B. McDonell, S.D. Fihn, Association between inhaled beta-agonists and the risk of unstable angina and myocardial infarction, Chest 121 (2002) 846–851. [7] J.M. van Vlymen, J.L. Parlow, The effects of reversal of neuromuscular blockade on autonomic control in the perioperative period, Anesth. Analg. 84 (1997) 148–154. [8] H. Worth, K.F. Chung, J.M. Felser, H. Hu, P. Rueegg, Cardio- and cerebrovascular safety of indacaterol vs formoterol, salmeterol, tiotropium and placebo in COPD, Respir. Med. 105 (2011) 571–579.

139

[9] S.R. Salpeter, T.M. Ormiston, E.E. Salpeter, Cardiovascular effects of beta-agonists in patients with asthma and COPD: a meta-analysis, Chest 125 (2004) 2309–2321. [10] S. Singh, Y.K. Loke, P. Enright, C.D. Furberg, Pro-arrhythmic and pro-ischaemic effects of inhaled anticholinergic medications, Thorax 68 (2013) 114–116. [11] M. Cazzola, F. Imperatore, A. Salzillo, F. Di Perna, F. Calderaro, A. Imperatore, et al., Cardiac effects of formoterol and salmeterol in patients suffering from COPD with preexisting cardiac arrhythmias and hypoxemia, Chest 114 (1998) 411–415. [12] A. Gershon, R. Croxford, A. Calzavara, T. To, M.B. Stanbrook, R. Upshur, et al., Cardiovascular safety of inhaled long-acting bronchodilators in individuals with chronic obstructive pulmonary disease, JAMA Intern. Med. 173 (2013) 1175–1185. [13] M.L. Decramer, N.A. Hanania, J.O. Lotvall, B.P. Yawn, The safety of long-acting beta2agonists in the treatment of stable chronic obstructive pulmonary disease, Int. J. Chron. Obstruct. Pulmon. Dis. 8 (2013) 53–64. [14] S.C. Campbell, G.J. Criner, B.E. Levine, S.J. Simon, J.S. Smith, C.J. Orevillo, et al., Cardiac safety of formoterol 12 microg twice daily in patients with chronic obstructive pulmonary disease, Pulm. Pharmacol. Ther. 20 (2007) 571–579. [15] S. Singh, Y.K. Loke, P.L. Enright, C.D. Furberg, Mortality associated with tiotropium mist inhaler in patients with chronic obstructive pulmonary disease: systematic review and meta-analysis of randomised controlled trials, BMJ 342 (2011) d3215. [16] E.D. Bateman, Tiotropium Respimat increases the risk of mortality: con, Eur. Respir. J. 42 (2013) 590–593. [17] A.G. Mathioudakis, V. Chatzimavridou-Grigoriadou, E. Evangelopoulou, G.A. Mathioudakis, N.M. Siafakas, Comparative mortality risk of tiotropium administered via handihaler or respimat in COPD patients: are they equivalent? Pulm. Pharmacol. Ther. 28 (2014) 91–97. [18] E. Cholongitas, Paroxysmal atrial tachycardia associated with prednisolone administration, J. Postgrad. Med. 55 (2009) 316. [19] A. Vasheghani-Farahani, M.A. Sahraian, L. Darabi, A. Aghsaie, A. Minagar, Incidence of various cardiac arrhythmias and conduction disturbances due to high dose intravenous methylprednisolone in patients with multiple sclerosis, J. Neurol. Sci. 309 (2011) 75–78. [20] V. Shaygannejad, F. Ashtari, M. Alinaghian, R. Norouzi, M. Salari, F. Fatehi, Short-term safety of pulse steroid therapy in multiple sclerosis relapses, Clin. Neuropharmacol. 36 (2013) 1–3. [21] M. Jara, S.F. Lanes, C. Wentworth III, C. May, S. Kesten, Comparative safety of longacting inhaled bronchodilators: a cohort study using the UK THIN primary care database, Drug Saf. 30 (2007) 1151–1160. [22] T. Frielle, K.W. Daniel, M.G. Caron, R.J. Lefkowitz, Structural basis of beta-adrenergic receptor subtype specificity studied with chimeric beta 1/beta 2-adrenergic receptors, Proc. Natl. Acad. Sci. U. S. A. 85 (1988) 9494–9498. [23] P. Gupta, M.S. O'Mahony, Potential adverse effects of bronchodilators in the treatment of airways obstruction in older people: recommendations for prescribing, Drugs Aging 25 (2008) 415–443. [24] M. Johnson, The beta-adrenoceptor, Am. J. Respir. Crit. Care Med. 158 (1998) S146–S153. [25] O.E. Brodde, Beta 1- and beta 2-adrenoceptors in the human heart: properties, function, and alterations in chronic heart failure, Pharmacol. Rev. 43 (1991) 203–242. [26] D.D. Sin, S.F. Man, Why are patients with chronic obstructive pulmonary disease at increased risk of cardiovascular diseases? The potential role of systemic inflammation in chronic obstructive pulmonary disease, Circulation 107 (2003) 1514–1519. [27] R.I. Horwitz, A.R. Feinstein, The problem of “protopathic bias” in case–control studies, Am. J. Med. 68 (1980) 255–258. [28] P.M. Calverley, J.A. Anderson, B. Celli, G.T. Ferguson, C. Jenkins, P.W. Jones, et al., Cardiovascular events in patients with COPD: TORCH study results, Thorax 65 (2010) 719–725. [29] B.J. Barnes, J.M. Hollands, Drug-induced arrhythmias, Crit. Care Med. 38 (2010) S188–S197. [30] J.A. Falk, S. Kadiev, G.J. Criner, S.M. Scharf, O.A. Minai, P. Diaz, Cardiac disease in chronic obstructive pulmonary disease, Proc. Am. Thorac. Soc. 5 (2008) 543–548. [31] A. Biolo, N. Clausell, K.G. Santos, R. Salvaro, P. Ashton-Prolla, A. Borges, et al., Impact of beta1-adrenergic receptor polymorphisms on susceptibility to heart failure, arrhythmogenesis, prognosis, and response to beta-blocker therapy, Am. J. Cardiol. 102 (2008) 726–732. [32] R.G. Aleong, W.H. Sauer, A.D. Robertson, S.B. Liggett, M.R. Bristow, Adrenergic receptor polymorphisms and prevention of ventricular arrhythmias with bucindolol in patients with chronic heart failure, Circ. Arrhythm. Electrophysiol. 6 (2013) 137–143.