Abstract
Non-steroidal anti-inflammatory drugs (NSAIDs) are associated with increased risk of cardiovascular disease. Yet, the risk of atrial fibrillation (AF) associated with NSAIDs among patients with prior myocardial infarction (MI) has not been examined, and such data could contribute considerably to the risk–benefit assessment of NSAID use in this clinical context.
Using nationwide administrative registries in Denmark, we studied patients aged ≥30 years admitted with first-time MI and without prior AF in the period of 1997–2011. Risk of AF associated with NSAID use vs. no NSAID use was analysed by multivariable time-dependent Cox proportional hazard models. Of the 86 496 patients [mean age 66 (SD 13) years; 64% men] included in this study, 44.1% filled at least one NSAID prescription after discharge from MI. During a mean follow-up of 5.3 years, 7831 (8.9%) developed AF. The confidence intervals rate (95% CI) of AF per 100 person-years with NSAID treatment was 2.2 (2.0–2.4) compared with 1.7 (1.6–1.7) without NSAIDs. In the adjusted model, the risk of AF after NSAID treatment increased [Hazard ratio (HR) 1.27 (1.14–1.40)]. An increased risk of AF was seen regardless of the type of NSAID or with short-term (0–14 days) treatment [HR 1.45 (1.24–1.69)]. When the risk of death in patients exposed [crude rate 23.3 (19.7–27.5)] vs. not exposed [crude rate 17.4 (95% CI 16.8–18.1)] to NSAIDs at the time of AF was compared, NSAID use was associated with a poorer prognosis [HR 1.35 (1.14–1.60)].
Our study suggests that the use of NSAIDs might be associated with the increased risk of AF in post-MI patients.
Introduction
Atrial fibrillation (AF) is the most common arrhythmia, and it is frequently encountered in the post-myocardial infarction (MI) period.1–3 Post-MI patients with AF have increased risk of thrombo-embolic complications, in particular ischaemic strokes, and death.4–7 Hence, AF constitutes a common complication in post-MI patients that should be avoided, if possible.8
Non-steroidal anti-inflammatory drugs (NSAIDs) are commonly used in the general population, and two recent studies linked the use of NSAIDs to AF.9,10 Current guidelines discourage any use of NSAIDs in patients with cardiovascular disease, despite this many patients with MI still receive NSAIDs.11–13 Yet, the risk of developing AF associated with NSAIDs among patients with prior MI has not been examined, and such data could contribute considerably to the risk–benefit assessment of NSAIDs in this clinical context. Accordingly, increased risk of AF in post-MI patients exposed to NSAIDs would have important clinical consequences. In this study, we therefore assessed the risk of AF associated with NSAID treatment in real-life, first-time MI patients.
Method
Study design
This study was a nationwide cohort study of AF incidence in Denmark in the period of 1997–2011 among patients with first-time MI without a history of AF.
Data sources
In Denmark, all residents have a unique and permanent identification number, which enable us to combine data from four Danish nationwide administrative registries: (i) Information regarding morbidity was obtained from the Danish National Patient Registry, which since 1978 has kept records of all hospital admissions in Denmark.14 Each hospital admission is registered with one main discharge diagnosis, and if appropriate one or more supplementary diagnoses, according to the International Classification of Diseases (ICD) codes, until 1994 the 8th revision (ICD-8) and from 1994 the 10th revision (ICD-10). (ii) Vital and migration status was obtained from the Central Person Registry. (iii) The cause of death was collected from the National Causes of Death Register, in which immediate, contributory, and underlying causes of death are recorded using the ICD-10 classification system. (iv) Information on concomitant pharmacotherapy was obtained from the Danish Register of Medicinal Product Statistics (national prescription registry), where all prescriptions dispensed from Danish pharmacies are available from 1995 onwards. Due to partial reimbursement of drug expenses by the Danish health-care system, all pharmacies in Denmark are required to register each drug dispensing in the national prescription registry, ensuring complete registration.15 Each drug dispensed is registered according to an international classification of drugs, the Anatomical Therapeutical Chemical (ATC) system, as well as the date of dispensing, quantity dispensed, strength, and formulation. All ATC and ICD codes are available in Supplementary material online, Tables S1 and S2.
Study population
We identified a cohort that included all patients admitted with a first-time MI from 1997 to 2011 who had no prior diagnosis of AF. As done previously, the database was systematically screened to ensure that any transfer of patients between hospitals was registered as one admission, since our analyses used the date of discharge to select patients and to calculate the dispensed medicine.12,16 To avoid selection bias in the exposure allocation due to the high mortality in relation to the MI, the cohort was restricted to individuals alive 30 days post discharge. We used a new-user design as suggested by Ray and excluded NSAID users who collected a prescription 30 days prior to inclusion (n = 6576).17 Patients were followed from 30 days post-MI until one of the following events (whichever came first): AF, emigration, death, or end of study period (31 December 2011).
Exposure and concomitant pharmacotherapy
We identified all claimed prescriptions for NSAIDs (ATC M01A) from the national prescription registry following discharge from index hospitalization (MI). The selective cyclooxygenase (COX)-2 inhibitors, rofecoxib and celecoxib, and the most commonly used non-selective NSAIDs, ibuprofen, diclofenac, and naproxen, were examined separately. All other NSAIDs, excluding glucosamine (M01AX05), were combined in a common group called ‘other NSAIDs’.
The national prescription registry does not include information on prescribed daily dosage of the medication. Therefore, by calculating average dosages of consecutive prescriptions, the daily dosage was estimated at each new prescription dispensing. This method allowed dosages to change at the dispensing of a new prescription. If only one prescription was available for an individual, a standard dose was used to estimate the daily dose. High dosage was defined as being above the upper limit of the recommended minimal dosage for each drug: ibuprofen >1200 mg, diclofenac ≥100 mg, naproxen >500 mg, rofecoxib >25 mg, and celecoxib >200 mg. The method used to determine the dose and treatment duration has been described previously.16,18 Ibuprofen is the only NSAID available in Denmark without a prescription (since 2001) but only in low doses (200 mg) and in limited quantity. The sale constitutes ∼15–20% of all NSAIDs after 2001.
Comorbidity and concomitant medication
Comorbidity was defined by using the Ontario acute MI mortality prediction rule, modified for the ICD-10.19 To further enhance the comorbidity score, we identified discharge diagnoses up to 1 year before the index hospitalization.20 Concomitant use of β-, angiotensin-converting enzyme (ACE) inhibitors/angiotensin 2 receptor blockers, statins, loop diuretics, spironolactone, and anti-diabetic drugs, the latter a proxy for prevalent diabetes,21 was also identified.
Outcome
The outcome of interest was first-time diagnosis of AF after discharge from the index MI. First-time AF admission was indicated by the fact that the National Patient Registry had not registered any prior admissions for AF in the previous 19 years. We selected 19 years since that was the maximal time we were able to look back in the registry for patients admitted in 1997, and this length of history was applied to all patients in our cohort. We used primary and/or secondary diagnoses. The diagnose of AF has been validated with a positive predictive value of 97% in the Danish National Patients Registry.22
Statistics
Unadjusted incidence rates of events per 100 person-years for AF were calculated for all NSAIDs as a group and for the individual NSAIDs separately. We used Cox proportional hazard models to estimate the risk of developing AF associated with NSAID use. Exposure to NSAIDs was included as time-dependent covariates in the models, i.e. patients were only considered at risk, when they were exposed to the drug. Each individual could have multiple independent treatment courses with the same drug but also with different drugs. Hence, current NSAID use was specifically compared with non-current NSAID use in our analyses. To assess the relation of duration of NSAID use on AF risk, we categorized the duration of NSAID exposure into short (0–14 days), medium (15–30 days), and long (31–90 days) terms for each individual period of NSAID treatment. Furthermore, we investigated the prognostic impact of AF after NSAID use. All of the models were adjusted for age, gender, index year, concomitant medication, and comorbidity. The validity of the proportional hazard assumption, linearity of continuous variables, and lack of interaction were found to be valid unless otherwise indicated.
All statistical analyses and data management were performed with version 11.0 (Stata Corp LP, College Station, TX, USA) and the SAS statistical software package, ver. 9.1 (SAS Institute, Inc., Cary, NC, USA).
Ethics
In Denmark, retrospective register studies do not require approval from the ethics committees. The Danish Data Protection Agency approved this study (No. 2007-58-0015, international reference: GEH-2014-014) and made the individual-level data available to us in an anonymized format, so specific individuals could not be identified.
Results
From 1997 to 2011, a total of 128 751 patients aged >30 years were admitted with first-time MI; of these, 86 496 (67.2%) were alive 30 days after discharge and had no prior diagnosis of AF and were therefore included in the study. Descriptive data are shown in Table 1 and Figure 1. The mean age was 66.6 (SD 13.3) years, and men comprised 64.3% of the study cohort. Of the 86 496 patients included in the study, 44.1% filled at least one prescription for an NSAID during follow-up. The median time was 4.5 years (interquartile range 1–14.9). In general, the patients prescribed with non-selective NSAIDs were younger, and a higher proportion of the patients were male compared with patients prescribed selective COX-2 inhibitors. Otherwise, no major differences between the treatments groups were found. The distribution of NSAIDs prescribed after discharge from MI was 2.8% for rofecoxib, 3.0% for celecoxib, 10.1% for other NSAIDs, 14.7% for diclofenac, and 29.9% for ibuprofen. More than 20% of patients received more than one NSAID. The patients who did not receive NSAIDs had a mean age of 68.5 years and more often were male.
Baseline characteristics of the total study population and individual treatment groups
| Characteristic | Total population | No NSAID | Overall NSAID | Exposure group | |||||
|---|---|---|---|---|---|---|---|---|---|
| Rofecoxib | Celecoxib | Ibuprofen | Diclofenac | Naproxen | Other NSAIDs | ||||
| N (%) | N (%) | N (%) | N (%) | N (%) | N (%) | N (%) | N (%) | N (%) | |
| Total patients | 86.496 (100.0) | 48 305 (55.9) | 38 191 (44.1) | 2423 (2.8) | 2584 (3.0) | 25 825 (29.9) | 12 699 (14.7) | 2243 (2.6) | 8692 (10.1) |
| Mean age (SD) (years) | 66.6 (13.3) | 68.5 (13.3) | 64.2 (12.9) | 68.9 (12.5) | 69.1 (12.3) | 62.7 (12.8) | 63.1 (12.5) | 62.7 (12.4) | 65.6 (12.4) |
| Women | 30 829 (35.6) | 17 828 (36.9) | 13 001 (34.0) | 1193 (49.2) | 1306 (50.5) | 8200 (31.8) | 4067 (32.0) | 697 (31.1) | 3380 (38.9) |
| Men | 55 667 (64.4) | 30 477 (63.1) | 25 190 (66.0) | 1230 (51.8) | 1278 (49.5) | 17 625 (68.2) | 8632 (68.0) | 1546 (68.9) | 5312 (61.1) |
| Comorbidity | |||||||||
| Cardiac arrhythmias | 1885 (2.2) | 1.132 (2.3) | 753 (2.0) | 61 (2.5) | 55 (2.1) | 507 (2.0) | 239 (1.9) | 40 (1.8) | 175 (2.0) |
| Peripheral vascular disease | 3210 (3.7) | 2060 (4.3) | 1150 (3.0) | 98 (4.0) | 122 (4.7) | 722 (2.8) | 347 (2.7) | 52 (2.3) | 269 (3.1) |
| Cerebral vascular disease | 3648 (4.2) | 2458 (5.1) | 1190 (3.1) | 102 (4.2) | 103 (4.0) | 724 (2.8) | 377 (3.0) | 55 (2.5) | 281 (3.2) |
| Diabetes with complications | 3478 (4.0) | 2174 (4.5) | 1304 (3.4) | 101 (4.2) | 96 (3.7) | 896 (3.5) | 440 (3.5) | 71 (3.2) | 271 (3.1) |
| Acute renal failure | 680 (0.8) | 519 (1.1) | 161 (0.4) | 20 (0.8) | 9 (0.4) | 98 (0.4) | 35 (0.3) | 6 (0.3) | 31 (0.4) |
| Chronic renal failure | 1062 (1.2) | 832 (1.7) | 230 (0.6) | 10 (0.4) | 14 (0.5) | 135 (0.5) | 65 (0.5) | 8 (0.4) | 45 (0.5) |
| Malignancy | 2026 (2.3) | 1371 (2.8) | 655 (1.7) | 48 (2.0) | 63 (2.4) | 373 (1.4) | 181 (1.4) | 25 (1.1) | 148 (1.7) |
| Shock | 213 (0.3) | 142 (0.3) | 71 (0.2) | 2 (0.1) | 4 (0.2) | 45 (0.2) | 22 (0.2) | 5 (0.2) | 13 (0.2) |
| COPD | 734 (0.9) | 509 (1.1) | 225 (0.6) | 20 (0.8) | 21 (0.8) | 126 (0.5) | 72 (0.6) | 11 (0.5) | 61 (0.7) |
| Gastric ulcer | 3609 (4.2) | 2371 (4.9) | 1238 (3.2) | 145 (6.0) | 143 (5.5) | 713 (2.8) | 411 (3.2) | 49 (2.2) | 311 (3.6) |
| PCI | 31 142 (36.0) | 18 278 (37.8) | 12 864 (33.7) | 371 (15.3) | 393 (15.2) | 9192 (35.6) | 4008 (31.6) | 752 (33.5) | 2585 (29.7) |
| Chronic heart failure | 7471 (8.6) | 5020 (10.4) | 2451 (6.4) | 246 (10.2) | 231 (8.9) | 1479 (5.7) | 691 (5.4) | 127 (5.7) | 558 (6.4) |
| Rheumatic disease | 4339 (5.0) | 2398 (5.0) | 1941 (5.1) | 143 (5.9) | 170 (6.6) | 1202 (4.7) | 628 (5.0) | 116 (5.2) | 552 (6.4) |
| Hypertension | 28 610 (33.1) | 16 766 (34.7) | 11 844 (31.0) | 742 (30.6) | 862 (33.4) | 7767 (30.1) | 3747 (29.5) | 672 (30.0) | 2952 (34.0) |
| Concomitant medical treatment | |||||||||
| β-Blockers | 66 055 (76.4) | 36 120 (74.8) | 29 935 (78.4) | 1679 (69.0) | 1835 (71.0) | 20 496 (79.4) | 10 095 (79.5) | 1786 (79.6) | 6785 (78.1) |
| ACE inhibitors | 38 459 (44.5) | 22 491 (46.6) | 15 968 (41.8) | 906 (37.4) | 972 (37.6) | 10 715 (41.5) | 5138 (40.5) | 917 (40.9.) | 3547 (40.8) |
| Statins | 54 291 (62.8) | 30 679 (62.9) | 23 612 (61.8) | 8310 (34.3) | 951 (36.8) | 16 713 (64.7) | 7690 (60.6) | 1339 (59.7) | 4913 (56.5) |
| ASA | 68 937 (79.7) | 38 536 (79.8) | 30 401 (79.6) | 1757 (72.5) | 1884 (72.9) | 20 756 (80.4) | 9992 (78.7) | 1788 (79.7) | 6768 (77.9) |
| Clopidogrel | 44 371 (51.3) | 26 179 (54.2) | 18 192 (47.6) | 434 (17.9) | 570 (22.1) | 12 901 (50.0) | 5582 (44.0) | 1018 (45.4) | 3633 (41.8) |
| Spironolactone | 5870 (6.8) | 3751 (7.8) | 2119 (5.6) | 161 (6.4) | 187 (7.2) | 1325 (5.1) | 602 (4.7) | 118 (5.3) | 488 (5.6) |
| Loop diuretics | 25 949 (30.0) | 16 075 (33.3) | 9874 (25.9) | 916 (37.8) | 977 (37.8) | 6012 (23.3) | 3053 (24.0) | 553 (24.7) | 2459 (28.3) |
| Glucose-lowering drugs | 9273 (10.7) | 5493 (12.8) | 4780 (9.9) | 248 (10.2) | 265 (10.3) | 2533 (9.8) | 1247 (9.8) | 206 (9.2) | 847 (9.7) |
| PPI | 17 736 (20.5) | 10 764 (22.3) | 6972 (18.3) | 608 (25.1) | 645 (25.0) | 4406 (17.1) | 2266 (17.8) | 381 (17.0) | 1740 (20.0) |
| Antiarrhythmic drugs | 72 959 (84.4) | 40 076 (83.0) | 32 883 (86.1) | 1970 (81.3) | 2142 (82.9) | 22 373 (86.6) | 11 005 (86.7) | 1953 (87.1) | 7527 (86.6) |
| Warfarin | 1101 (1.3) | 749 (1.6) | 352 (0.9) | 13 (0.5) | 18 (0.7) | 232 (0.9) | 97 (0.8) | 19 (0.9) | 73 (0.8) |
| Characteristic | Total population | No NSAID | Overall NSAID | Exposure group | |||||
|---|---|---|---|---|---|---|---|---|---|
| Rofecoxib | Celecoxib | Ibuprofen | Diclofenac | Naproxen | Other NSAIDs | ||||
| N (%) | N (%) | N (%) | N (%) | N (%) | N (%) | N (%) | N (%) | N (%) | |
| Total patients | 86.496 (100.0) | 48 305 (55.9) | 38 191 (44.1) | 2423 (2.8) | 2584 (3.0) | 25 825 (29.9) | 12 699 (14.7) | 2243 (2.6) | 8692 (10.1) |
| Mean age (SD) (years) | 66.6 (13.3) | 68.5 (13.3) | 64.2 (12.9) | 68.9 (12.5) | 69.1 (12.3) | 62.7 (12.8) | 63.1 (12.5) | 62.7 (12.4) | 65.6 (12.4) |
| Women | 30 829 (35.6) | 17 828 (36.9) | 13 001 (34.0) | 1193 (49.2) | 1306 (50.5) | 8200 (31.8) | 4067 (32.0) | 697 (31.1) | 3380 (38.9) |
| Men | 55 667 (64.4) | 30 477 (63.1) | 25 190 (66.0) | 1230 (51.8) | 1278 (49.5) | 17 625 (68.2) | 8632 (68.0) | 1546 (68.9) | 5312 (61.1) |
| Comorbidity | |||||||||
| Cardiac arrhythmias | 1885 (2.2) | 1.132 (2.3) | 753 (2.0) | 61 (2.5) | 55 (2.1) | 507 (2.0) | 239 (1.9) | 40 (1.8) | 175 (2.0) |
| Peripheral vascular disease | 3210 (3.7) | 2060 (4.3) | 1150 (3.0) | 98 (4.0) | 122 (4.7) | 722 (2.8) | 347 (2.7) | 52 (2.3) | 269 (3.1) |
| Cerebral vascular disease | 3648 (4.2) | 2458 (5.1) | 1190 (3.1) | 102 (4.2) | 103 (4.0) | 724 (2.8) | 377 (3.0) | 55 (2.5) | 281 (3.2) |
| Diabetes with complications | 3478 (4.0) | 2174 (4.5) | 1304 (3.4) | 101 (4.2) | 96 (3.7) | 896 (3.5) | 440 (3.5) | 71 (3.2) | 271 (3.1) |
| Acute renal failure | 680 (0.8) | 519 (1.1) | 161 (0.4) | 20 (0.8) | 9 (0.4) | 98 (0.4) | 35 (0.3) | 6 (0.3) | 31 (0.4) |
| Chronic renal failure | 1062 (1.2) | 832 (1.7) | 230 (0.6) | 10 (0.4) | 14 (0.5) | 135 (0.5) | 65 (0.5) | 8 (0.4) | 45 (0.5) |
| Malignancy | 2026 (2.3) | 1371 (2.8) | 655 (1.7) | 48 (2.0) | 63 (2.4) | 373 (1.4) | 181 (1.4) | 25 (1.1) | 148 (1.7) |
| Shock | 213 (0.3) | 142 (0.3) | 71 (0.2) | 2 (0.1) | 4 (0.2) | 45 (0.2) | 22 (0.2) | 5 (0.2) | 13 (0.2) |
| COPD | 734 (0.9) | 509 (1.1) | 225 (0.6) | 20 (0.8) | 21 (0.8) | 126 (0.5) | 72 (0.6) | 11 (0.5) | 61 (0.7) |
| Gastric ulcer | 3609 (4.2) | 2371 (4.9) | 1238 (3.2) | 145 (6.0) | 143 (5.5) | 713 (2.8) | 411 (3.2) | 49 (2.2) | 311 (3.6) |
| PCI | 31 142 (36.0) | 18 278 (37.8) | 12 864 (33.7) | 371 (15.3) | 393 (15.2) | 9192 (35.6) | 4008 (31.6) | 752 (33.5) | 2585 (29.7) |
| Chronic heart failure | 7471 (8.6) | 5020 (10.4) | 2451 (6.4) | 246 (10.2) | 231 (8.9) | 1479 (5.7) | 691 (5.4) | 127 (5.7) | 558 (6.4) |
| Rheumatic disease | 4339 (5.0) | 2398 (5.0) | 1941 (5.1) | 143 (5.9) | 170 (6.6) | 1202 (4.7) | 628 (5.0) | 116 (5.2) | 552 (6.4) |
| Hypertension | 28 610 (33.1) | 16 766 (34.7) | 11 844 (31.0) | 742 (30.6) | 862 (33.4) | 7767 (30.1) | 3747 (29.5) | 672 (30.0) | 2952 (34.0) |
| Concomitant medical treatment | |||||||||
| β-Blockers | 66 055 (76.4) | 36 120 (74.8) | 29 935 (78.4) | 1679 (69.0) | 1835 (71.0) | 20 496 (79.4) | 10 095 (79.5) | 1786 (79.6) | 6785 (78.1) |
| ACE inhibitors | 38 459 (44.5) | 22 491 (46.6) | 15 968 (41.8) | 906 (37.4) | 972 (37.6) | 10 715 (41.5) | 5138 (40.5) | 917 (40.9.) | 3547 (40.8) |
| Statins | 54 291 (62.8) | 30 679 (62.9) | 23 612 (61.8) | 8310 (34.3) | 951 (36.8) | 16 713 (64.7) | 7690 (60.6) | 1339 (59.7) | 4913 (56.5) |
| ASA | 68 937 (79.7) | 38 536 (79.8) | 30 401 (79.6) | 1757 (72.5) | 1884 (72.9) | 20 756 (80.4) | 9992 (78.7) | 1788 (79.7) | 6768 (77.9) |
| Clopidogrel | 44 371 (51.3) | 26 179 (54.2) | 18 192 (47.6) | 434 (17.9) | 570 (22.1) | 12 901 (50.0) | 5582 (44.0) | 1018 (45.4) | 3633 (41.8) |
| Spironolactone | 5870 (6.8) | 3751 (7.8) | 2119 (5.6) | 161 (6.4) | 187 (7.2) | 1325 (5.1) | 602 (4.7) | 118 (5.3) | 488 (5.6) |
| Loop diuretics | 25 949 (30.0) | 16 075 (33.3) | 9874 (25.9) | 916 (37.8) | 977 (37.8) | 6012 (23.3) | 3053 (24.0) | 553 (24.7) | 2459 (28.3) |
| Glucose-lowering drugs | 9273 (10.7) | 5493 (12.8) | 4780 (9.9) | 248 (10.2) | 265 (10.3) | 2533 (9.8) | 1247 (9.8) | 206 (9.2) | 847 (9.7) |
| PPI | 17 736 (20.5) | 10 764 (22.3) | 6972 (18.3) | 608 (25.1) | 645 (25.0) | 4406 (17.1) | 2266 (17.8) | 381 (17.0) | 1740 (20.0) |
| Antiarrhythmic drugs | 72 959 (84.4) | 40 076 (83.0) | 32 883 (86.1) | 1970 (81.3) | 2142 (82.9) | 22 373 (86.6) | 11 005 (86.7) | 1953 (87.1) | 7527 (86.6) |
| Warfarin | 1101 (1.3) | 749 (1.6) | 352 (0.9) | 13 (0.5) | 18 (0.7) | 232 (0.9) | 97 (0.8) | 19 (0.9) | 73 (0.8) |
ACE inhibitors, angiotensin-converting enzyme inhibitors; antiarrhythmic drugs, digoxin, β-blockers, calcium antagonists, class1C, sotalol; ASA, acetylsalicylic acid; COPD, chronic obstructive pulmonary disease; PCI, percutaneous coronary intervention; PPI, proton pump inhibitors.
Baseline characteristics of the total study population and individual treatment groups
| Characteristic | Total population | No NSAID | Overall NSAID | Exposure group | |||||
|---|---|---|---|---|---|---|---|---|---|
| Rofecoxib | Celecoxib | Ibuprofen | Diclofenac | Naproxen | Other NSAIDs | ||||
| N (%) | N (%) | N (%) | N (%) | N (%) | N (%) | N (%) | N (%) | N (%) | |
| Total patients | 86.496 (100.0) | 48 305 (55.9) | 38 191 (44.1) | 2423 (2.8) | 2584 (3.0) | 25 825 (29.9) | 12 699 (14.7) | 2243 (2.6) | 8692 (10.1) |
| Mean age (SD) (years) | 66.6 (13.3) | 68.5 (13.3) | 64.2 (12.9) | 68.9 (12.5) | 69.1 (12.3) | 62.7 (12.8) | 63.1 (12.5) | 62.7 (12.4) | 65.6 (12.4) |
| Women | 30 829 (35.6) | 17 828 (36.9) | 13 001 (34.0) | 1193 (49.2) | 1306 (50.5) | 8200 (31.8) | 4067 (32.0) | 697 (31.1) | 3380 (38.9) |
| Men | 55 667 (64.4) | 30 477 (63.1) | 25 190 (66.0) | 1230 (51.8) | 1278 (49.5) | 17 625 (68.2) | 8632 (68.0) | 1546 (68.9) | 5312 (61.1) |
| Comorbidity | |||||||||
| Cardiac arrhythmias | 1885 (2.2) | 1.132 (2.3) | 753 (2.0) | 61 (2.5) | 55 (2.1) | 507 (2.0) | 239 (1.9) | 40 (1.8) | 175 (2.0) |
| Peripheral vascular disease | 3210 (3.7) | 2060 (4.3) | 1150 (3.0) | 98 (4.0) | 122 (4.7) | 722 (2.8) | 347 (2.7) | 52 (2.3) | 269 (3.1) |
| Cerebral vascular disease | 3648 (4.2) | 2458 (5.1) | 1190 (3.1) | 102 (4.2) | 103 (4.0) | 724 (2.8) | 377 (3.0) | 55 (2.5) | 281 (3.2) |
| Diabetes with complications | 3478 (4.0) | 2174 (4.5) | 1304 (3.4) | 101 (4.2) | 96 (3.7) | 896 (3.5) | 440 (3.5) | 71 (3.2) | 271 (3.1) |
| Acute renal failure | 680 (0.8) | 519 (1.1) | 161 (0.4) | 20 (0.8) | 9 (0.4) | 98 (0.4) | 35 (0.3) | 6 (0.3) | 31 (0.4) |
| Chronic renal failure | 1062 (1.2) | 832 (1.7) | 230 (0.6) | 10 (0.4) | 14 (0.5) | 135 (0.5) | 65 (0.5) | 8 (0.4) | 45 (0.5) |
| Malignancy | 2026 (2.3) | 1371 (2.8) | 655 (1.7) | 48 (2.0) | 63 (2.4) | 373 (1.4) | 181 (1.4) | 25 (1.1) | 148 (1.7) |
| Shock | 213 (0.3) | 142 (0.3) | 71 (0.2) | 2 (0.1) | 4 (0.2) | 45 (0.2) | 22 (0.2) | 5 (0.2) | 13 (0.2) |
| COPD | 734 (0.9) | 509 (1.1) | 225 (0.6) | 20 (0.8) | 21 (0.8) | 126 (0.5) | 72 (0.6) | 11 (0.5) | 61 (0.7) |
| Gastric ulcer | 3609 (4.2) | 2371 (4.9) | 1238 (3.2) | 145 (6.0) | 143 (5.5) | 713 (2.8) | 411 (3.2) | 49 (2.2) | 311 (3.6) |
| PCI | 31 142 (36.0) | 18 278 (37.8) | 12 864 (33.7) | 371 (15.3) | 393 (15.2) | 9192 (35.6) | 4008 (31.6) | 752 (33.5) | 2585 (29.7) |
| Chronic heart failure | 7471 (8.6) | 5020 (10.4) | 2451 (6.4) | 246 (10.2) | 231 (8.9) | 1479 (5.7) | 691 (5.4) | 127 (5.7) | 558 (6.4) |
| Rheumatic disease | 4339 (5.0) | 2398 (5.0) | 1941 (5.1) | 143 (5.9) | 170 (6.6) | 1202 (4.7) | 628 (5.0) | 116 (5.2) | 552 (6.4) |
| Hypertension | 28 610 (33.1) | 16 766 (34.7) | 11 844 (31.0) | 742 (30.6) | 862 (33.4) | 7767 (30.1) | 3747 (29.5) | 672 (30.0) | 2952 (34.0) |
| Concomitant medical treatment | |||||||||
| β-Blockers | 66 055 (76.4) | 36 120 (74.8) | 29 935 (78.4) | 1679 (69.0) | 1835 (71.0) | 20 496 (79.4) | 10 095 (79.5) | 1786 (79.6) | 6785 (78.1) |
| ACE inhibitors | 38 459 (44.5) | 22 491 (46.6) | 15 968 (41.8) | 906 (37.4) | 972 (37.6) | 10 715 (41.5) | 5138 (40.5) | 917 (40.9.) | 3547 (40.8) |
| Statins | 54 291 (62.8) | 30 679 (62.9) | 23 612 (61.8) | 8310 (34.3) | 951 (36.8) | 16 713 (64.7) | 7690 (60.6) | 1339 (59.7) | 4913 (56.5) |
| ASA | 68 937 (79.7) | 38 536 (79.8) | 30 401 (79.6) | 1757 (72.5) | 1884 (72.9) | 20 756 (80.4) | 9992 (78.7) | 1788 (79.7) | 6768 (77.9) |
| Clopidogrel | 44 371 (51.3) | 26 179 (54.2) | 18 192 (47.6) | 434 (17.9) | 570 (22.1) | 12 901 (50.0) | 5582 (44.0) | 1018 (45.4) | 3633 (41.8) |
| Spironolactone | 5870 (6.8) | 3751 (7.8) | 2119 (5.6) | 161 (6.4) | 187 (7.2) | 1325 (5.1) | 602 (4.7) | 118 (5.3) | 488 (5.6) |
| Loop diuretics | 25 949 (30.0) | 16 075 (33.3) | 9874 (25.9) | 916 (37.8) | 977 (37.8) | 6012 (23.3) | 3053 (24.0) | 553 (24.7) | 2459 (28.3) |
| Glucose-lowering drugs | 9273 (10.7) | 5493 (12.8) | 4780 (9.9) | 248 (10.2) | 265 (10.3) | 2533 (9.8) | 1247 (9.8) | 206 (9.2) | 847 (9.7) |
| PPI | 17 736 (20.5) | 10 764 (22.3) | 6972 (18.3) | 608 (25.1) | 645 (25.0) | 4406 (17.1) | 2266 (17.8) | 381 (17.0) | 1740 (20.0) |
| Antiarrhythmic drugs | 72 959 (84.4) | 40 076 (83.0) | 32 883 (86.1) | 1970 (81.3) | 2142 (82.9) | 22 373 (86.6) | 11 005 (86.7) | 1953 (87.1) | 7527 (86.6) |
| Warfarin | 1101 (1.3) | 749 (1.6) | 352 (0.9) | 13 (0.5) | 18 (0.7) | 232 (0.9) | 97 (0.8) | 19 (0.9) | 73 (0.8) |
| Characteristic | Total population | No NSAID | Overall NSAID | Exposure group | |||||
|---|---|---|---|---|---|---|---|---|---|
| Rofecoxib | Celecoxib | Ibuprofen | Diclofenac | Naproxen | Other NSAIDs | ||||
| N (%) | N (%) | N (%) | N (%) | N (%) | N (%) | N (%) | N (%) | N (%) | |
| Total patients | 86.496 (100.0) | 48 305 (55.9) | 38 191 (44.1) | 2423 (2.8) | 2584 (3.0) | 25 825 (29.9) | 12 699 (14.7) | 2243 (2.6) | 8692 (10.1) |
| Mean age (SD) (years) | 66.6 (13.3) | 68.5 (13.3) | 64.2 (12.9) | 68.9 (12.5) | 69.1 (12.3) | 62.7 (12.8) | 63.1 (12.5) | 62.7 (12.4) | 65.6 (12.4) |
| Women | 30 829 (35.6) | 17 828 (36.9) | 13 001 (34.0) | 1193 (49.2) | 1306 (50.5) | 8200 (31.8) | 4067 (32.0) | 697 (31.1) | 3380 (38.9) |
| Men | 55 667 (64.4) | 30 477 (63.1) | 25 190 (66.0) | 1230 (51.8) | 1278 (49.5) | 17 625 (68.2) | 8632 (68.0) | 1546 (68.9) | 5312 (61.1) |
| Comorbidity | |||||||||
| Cardiac arrhythmias | 1885 (2.2) | 1.132 (2.3) | 753 (2.0) | 61 (2.5) | 55 (2.1) | 507 (2.0) | 239 (1.9) | 40 (1.8) | 175 (2.0) |
| Peripheral vascular disease | 3210 (3.7) | 2060 (4.3) | 1150 (3.0) | 98 (4.0) | 122 (4.7) | 722 (2.8) | 347 (2.7) | 52 (2.3) | 269 (3.1) |
| Cerebral vascular disease | 3648 (4.2) | 2458 (5.1) | 1190 (3.1) | 102 (4.2) | 103 (4.0) | 724 (2.8) | 377 (3.0) | 55 (2.5) | 281 (3.2) |
| Diabetes with complications | 3478 (4.0) | 2174 (4.5) | 1304 (3.4) | 101 (4.2) | 96 (3.7) | 896 (3.5) | 440 (3.5) | 71 (3.2) | 271 (3.1) |
| Acute renal failure | 680 (0.8) | 519 (1.1) | 161 (0.4) | 20 (0.8) | 9 (0.4) | 98 (0.4) | 35 (0.3) | 6 (0.3) | 31 (0.4) |
| Chronic renal failure | 1062 (1.2) | 832 (1.7) | 230 (0.6) | 10 (0.4) | 14 (0.5) | 135 (0.5) | 65 (0.5) | 8 (0.4) | 45 (0.5) |
| Malignancy | 2026 (2.3) | 1371 (2.8) | 655 (1.7) | 48 (2.0) | 63 (2.4) | 373 (1.4) | 181 (1.4) | 25 (1.1) | 148 (1.7) |
| Shock | 213 (0.3) | 142 (0.3) | 71 (0.2) | 2 (0.1) | 4 (0.2) | 45 (0.2) | 22 (0.2) | 5 (0.2) | 13 (0.2) |
| COPD | 734 (0.9) | 509 (1.1) | 225 (0.6) | 20 (0.8) | 21 (0.8) | 126 (0.5) | 72 (0.6) | 11 (0.5) | 61 (0.7) |
| Gastric ulcer | 3609 (4.2) | 2371 (4.9) | 1238 (3.2) | 145 (6.0) | 143 (5.5) | 713 (2.8) | 411 (3.2) | 49 (2.2) | 311 (3.6) |
| PCI | 31 142 (36.0) | 18 278 (37.8) | 12 864 (33.7) | 371 (15.3) | 393 (15.2) | 9192 (35.6) | 4008 (31.6) | 752 (33.5) | 2585 (29.7) |
| Chronic heart failure | 7471 (8.6) | 5020 (10.4) | 2451 (6.4) | 246 (10.2) | 231 (8.9) | 1479 (5.7) | 691 (5.4) | 127 (5.7) | 558 (6.4) |
| Rheumatic disease | 4339 (5.0) | 2398 (5.0) | 1941 (5.1) | 143 (5.9) | 170 (6.6) | 1202 (4.7) | 628 (5.0) | 116 (5.2) | 552 (6.4) |
| Hypertension | 28 610 (33.1) | 16 766 (34.7) | 11 844 (31.0) | 742 (30.6) | 862 (33.4) | 7767 (30.1) | 3747 (29.5) | 672 (30.0) | 2952 (34.0) |
| Concomitant medical treatment | |||||||||
| β-Blockers | 66 055 (76.4) | 36 120 (74.8) | 29 935 (78.4) | 1679 (69.0) | 1835 (71.0) | 20 496 (79.4) | 10 095 (79.5) | 1786 (79.6) | 6785 (78.1) |
| ACE inhibitors | 38 459 (44.5) | 22 491 (46.6) | 15 968 (41.8) | 906 (37.4) | 972 (37.6) | 10 715 (41.5) | 5138 (40.5) | 917 (40.9.) | 3547 (40.8) |
| Statins | 54 291 (62.8) | 30 679 (62.9) | 23 612 (61.8) | 8310 (34.3) | 951 (36.8) | 16 713 (64.7) | 7690 (60.6) | 1339 (59.7) | 4913 (56.5) |
| ASA | 68 937 (79.7) | 38 536 (79.8) | 30 401 (79.6) | 1757 (72.5) | 1884 (72.9) | 20 756 (80.4) | 9992 (78.7) | 1788 (79.7) | 6768 (77.9) |
| Clopidogrel | 44 371 (51.3) | 26 179 (54.2) | 18 192 (47.6) | 434 (17.9) | 570 (22.1) | 12 901 (50.0) | 5582 (44.0) | 1018 (45.4) | 3633 (41.8) |
| Spironolactone | 5870 (6.8) | 3751 (7.8) | 2119 (5.6) | 161 (6.4) | 187 (7.2) | 1325 (5.1) | 602 (4.7) | 118 (5.3) | 488 (5.6) |
| Loop diuretics | 25 949 (30.0) | 16 075 (33.3) | 9874 (25.9) | 916 (37.8) | 977 (37.8) | 6012 (23.3) | 3053 (24.0) | 553 (24.7) | 2459 (28.3) |
| Glucose-lowering drugs | 9273 (10.7) | 5493 (12.8) | 4780 (9.9) | 248 (10.2) | 265 (10.3) | 2533 (9.8) | 1247 (9.8) | 206 (9.2) | 847 (9.7) |
| PPI | 17 736 (20.5) | 10 764 (22.3) | 6972 (18.3) | 608 (25.1) | 645 (25.0) | 4406 (17.1) | 2266 (17.8) | 381 (17.0) | 1740 (20.0) |
| Antiarrhythmic drugs | 72 959 (84.4) | 40 076 (83.0) | 32 883 (86.1) | 1970 (81.3) | 2142 (82.9) | 22 373 (86.6) | 11 005 (86.7) | 1953 (87.1) | 7527 (86.6) |
| Warfarin | 1101 (1.3) | 749 (1.6) | 352 (0.9) | 13 (0.5) | 18 (0.7) | 232 (0.9) | 97 (0.8) | 19 (0.9) | 73 (0.8) |
ACE inhibitors, angiotensin-converting enzyme inhibitors; antiarrhythmic drugs, digoxin, β-blockers, calcium antagonists, class1C, sotalol; ASA, acetylsalicylic acid; COPD, chronic obstructive pulmonary disease; PCI, percutaneous coronary intervention; PPI, proton pump inhibitors.
Flow chart of population.
Risk of atrial fibrillation associated with use of non-steroidal anti-inflammatory drug
During the observation period, 7831 subjects were hospitalized with AF. Confidence intervals rate (95% CI) of AF in events per 100 person-years with ongoing NSAID treatment was 2.2 (2.0–2.4) compared with 1.7 (1.6–1.7) without NSAID treatment (Figure 1). In the adjusted model, the risk of AF with NSAID treatment increased [Hazard ratio (HR) 1.27 (1.14–1.40)] compared with no NSAID treatment. Table 2 shows comparison of the individual NSAIDs. When comparing the effect of the different types of NSAIDs, COX-2 inhibitors (rofecoxib, celecoxib), non-selective NSAIDs (ibuprofen, diclofenac, naproxen), and other NSAIDs (e.g. ketoprofen, etodolac), all were associated with both increased confidence intervals rates and increased adjusted risk compared with no NSAID use, Figure 2. With use of the earlier recommended dosage of NSAID, further increase in AF was present suggesting a dose–response relationship (Table 2). The insignificant result may be due to a low power.
Dose–response of the individual non-steroidal anti-inflammatory drug treatment regimes and risk of atrial fibrillation
| Treatment | Number of events of AF | Crude rate of AF | Hazard ratio of AF (95% CI) |
|---|---|---|---|
| No NSAID | 7433 | 1.7 (1.6–1.7) | Reference |
| Any NSAID | 398 | 2.2 (2.0–2.4) | 1.27 (1.14–1.40) |
| Rofecoxib | 18 | 2.4 (1.5–3.8) | 1.22 (0.77–1.93) |
| ≤25 mg | 14 | 2.0 (1.2–3.5) | 1.04 (0.62–1.76) |
| ≥25 mg | 4 | 6.2 (2.3–16.6) | 2.84 (1.06–7.57) |
| Celecoxib | 22 | 2.3 (1.5–3.5) | 1.12 (0.73–1.69) |
| ≤200 mg | 16 | 2.2 (1.3–3.5) | 1.06 (0.65–1.73) |
| ≥200 mg | 6 | 2.7 (1.2–6.0) | 1.28 (0.57–2.84) |
| Diclofenac | 83 | 2.1 (1.7–2.6) | 1.28 (1.03–1.58) |
| ≤100 mg | 69 | 2.0 (1.6–2.5) | 1.23 (0.97–1.56) |
| ≥100 mg | 14 | 2.5 (1.5–4.2) | 1.55 (0.92–2.62) |
| Ibuprofen | 180 | 2.1 (1.9–2.5) | 1.31 (1.13–1.53) |
| ≤1200 mg | 145 | 2.1 (1.8–2.5) | 1.33 (1.13–1.57) |
| ≥1200 mg | 35 | 2.4 (1.7–3.3) | 1.26 (0.90–1.76) |
| Naproxen | 14 | 1.9 (1.1–3.2) | 1.09 (0.65–1.85) |
| ≤500 mg | 4 | 1.3 (0.5–3.3) | 0.65 (0.24–1.74) |
| ≥500 mg | 10 | 2.3 (1.3–4.4) | 1.51 (0.90–2.37) |
| Treatment | Number of events of AF | Crude rate of AF | Hazard ratio of AF (95% CI) |
|---|---|---|---|
| No NSAID | 7433 | 1.7 (1.6–1.7) | Reference |
| Any NSAID | 398 | 2.2 (2.0–2.4) | 1.27 (1.14–1.40) |
| Rofecoxib | 18 | 2.4 (1.5–3.8) | 1.22 (0.77–1.93) |
| ≤25 mg | 14 | 2.0 (1.2–3.5) | 1.04 (0.62–1.76) |
| ≥25 mg | 4 | 6.2 (2.3–16.6) | 2.84 (1.06–7.57) |
| Celecoxib | 22 | 2.3 (1.5–3.5) | 1.12 (0.73–1.69) |
| ≤200 mg | 16 | 2.2 (1.3–3.5) | 1.06 (0.65–1.73) |
| ≥200 mg | 6 | 2.7 (1.2–6.0) | 1.28 (0.57–2.84) |
| Diclofenac | 83 | 2.1 (1.7–2.6) | 1.28 (1.03–1.58) |
| ≤100 mg | 69 | 2.0 (1.6–2.5) | 1.23 (0.97–1.56) |
| ≥100 mg | 14 | 2.5 (1.5–4.2) | 1.55 (0.92–2.62) |
| Ibuprofen | 180 | 2.1 (1.9–2.5) | 1.31 (1.13–1.53) |
| ≤1200 mg | 145 | 2.1 (1.8–2.5) | 1.33 (1.13–1.57) |
| ≥1200 mg | 35 | 2.4 (1.7–3.3) | 1.26 (0.90–1.76) |
| Naproxen | 14 | 1.9 (1.1–3.2) | 1.09 (0.65–1.85) |
| ≤500 mg | 4 | 1.3 (0.5–3.3) | 0.65 (0.24–1.74) |
| ≥500 mg | 10 | 2.3 (1.3–4.4) | 1.51 (0.90–2.37) |
AF, atrial fibrillation; CI, confidence interval; NSAID, non-steroidal anti-inflammatory drug.
Dose–response of the individual non-steroidal anti-inflammatory drug treatment regimes and risk of atrial fibrillation
| Treatment | Number of events of AF | Crude rate of AF | Hazard ratio of AF (95% CI) |
|---|---|---|---|
| No NSAID | 7433 | 1.7 (1.6–1.7) | Reference |
| Any NSAID | 398 | 2.2 (2.0–2.4) | 1.27 (1.14–1.40) |
| Rofecoxib | 18 | 2.4 (1.5–3.8) | 1.22 (0.77–1.93) |
| ≤25 mg | 14 | 2.0 (1.2–3.5) | 1.04 (0.62–1.76) |
| ≥25 mg | 4 | 6.2 (2.3–16.6) | 2.84 (1.06–7.57) |
| Celecoxib | 22 | 2.3 (1.5–3.5) | 1.12 (0.73–1.69) |
| ≤200 mg | 16 | 2.2 (1.3–3.5) | 1.06 (0.65–1.73) |
| ≥200 mg | 6 | 2.7 (1.2–6.0) | 1.28 (0.57–2.84) |
| Diclofenac | 83 | 2.1 (1.7–2.6) | 1.28 (1.03–1.58) |
| ≤100 mg | 69 | 2.0 (1.6–2.5) | 1.23 (0.97–1.56) |
| ≥100 mg | 14 | 2.5 (1.5–4.2) | 1.55 (0.92–2.62) |
| Ibuprofen | 180 | 2.1 (1.9–2.5) | 1.31 (1.13–1.53) |
| ≤1200 mg | 145 | 2.1 (1.8–2.5) | 1.33 (1.13–1.57) |
| ≥1200 mg | 35 | 2.4 (1.7–3.3) | 1.26 (0.90–1.76) |
| Naproxen | 14 | 1.9 (1.1–3.2) | 1.09 (0.65–1.85) |
| ≤500 mg | 4 | 1.3 (0.5–3.3) | 0.65 (0.24–1.74) |
| ≥500 mg | 10 | 2.3 (1.3–4.4) | 1.51 (0.90–2.37) |
| Treatment | Number of events of AF | Crude rate of AF | Hazard ratio of AF (95% CI) |
|---|---|---|---|
| No NSAID | 7433 | 1.7 (1.6–1.7) | Reference |
| Any NSAID | 398 | 2.2 (2.0–2.4) | 1.27 (1.14–1.40) |
| Rofecoxib | 18 | 2.4 (1.5–3.8) | 1.22 (0.77–1.93) |
| ≤25 mg | 14 | 2.0 (1.2–3.5) | 1.04 (0.62–1.76) |
| ≥25 mg | 4 | 6.2 (2.3–16.6) | 2.84 (1.06–7.57) |
| Celecoxib | 22 | 2.3 (1.5–3.5) | 1.12 (0.73–1.69) |
| ≤200 mg | 16 | 2.2 (1.3–3.5) | 1.06 (0.65–1.73) |
| ≥200 mg | 6 | 2.7 (1.2–6.0) | 1.28 (0.57–2.84) |
| Diclofenac | 83 | 2.1 (1.7–2.6) | 1.28 (1.03–1.58) |
| ≤100 mg | 69 | 2.0 (1.6–2.5) | 1.23 (0.97–1.56) |
| ≥100 mg | 14 | 2.5 (1.5–4.2) | 1.55 (0.92–2.62) |
| Ibuprofen | 180 | 2.1 (1.9–2.5) | 1.31 (1.13–1.53) |
| ≤1200 mg | 145 | 2.1 (1.8–2.5) | 1.33 (1.13–1.57) |
| ≥1200 mg | 35 | 2.4 (1.7–3.3) | 1.26 (0.90–1.76) |
| Naproxen | 14 | 1.9 (1.1–3.2) | 1.09 (0.65–1.85) |
| ≤500 mg | 4 | 1.3 (0.5–3.3) | 0.65 (0.24–1.74) |
| ≥500 mg | 10 | 2.3 (1.3–4.4) | 1.51 (0.90–2.37) |
AF, atrial fibrillation; CI, confidence interval; NSAID, non-steroidal anti-inflammatory drug.
Incidence rates per 100 person-years and adjusted Cox proportional hazard analysis of atrial fibrillation associated with the use of non-steroidal anti-inflammatory drug treatment in patients with prior myocardial infarction. No, number of events.
Duration of non-steroidal anti-inflammatory drug use and non-steroidal anti-inflammatory drug use at the time of atrial fibrillation
Rates of AF decreased according to duration of NSAID treatment with confidence intervals rates of 2.5 (2.2–2.9), 2.4 (2.0–3.0), and 2.2 (1.8–2.6) for 0–14 days (short-term use), 15–30 days (medium-term use), and 31–90 days (long-term use), respectively (Figure 3). Relative to no NSAID treatment, corresponding HRs of AF with short-, moderate-, and long-term NSAID treatments were 1.45 (1.24–1.70), 1.41 (1.15–1.72), and 1.27 (1.06–1.55). Comparing the risk of death in patients exposed to NSAID at the time of AF [confidence intervals rate 23.3 (19.7–27.5)] vs. not exposed to NSAID at the time of AF [confidence intervals rate 17.4 (16.8–18.1)], a poorer prognosis was found in patients exposed to NSAID at the time of AF [HR 1.35 (1.14–1.60)].
Incidence rates per 100 person-years and adjusted Cox proportional hazard analysis of atrial fibrillation according to the duration of non-steroidal anti-inflammatory drug treatment associated with the use of NSAID treatment in patients with prior myocardial infarction. No, number of events.
Sensitivity analysis
Using the Wald test, we examined interactions between the use of NSAID and available covariates. No clinically important interactions were found. When the study period was restricted to years without over-the-counter non-selective NSAID availability (1997–2001), our results did not change (data not shown). Heart failure, hypertension, and osteoarthritis might be associated with AF. We have therefore made a sensitivity analysis adjusting for these variables in the COX models. The result remains the same (data not shown). We estimated that an unmeasured confounder would have to elevate the risk by 2.8–3.5 to fully explain the increased risk for AF observed with overall NSAID treatment. To investigate the effect of heart failure (HF), we calculated the incidence of coincident HF and AF, 1157 (1.8%) patients had HF together with AF. To analyse the difference between the two groups by time, we made a cumulative incidence graph, which showed a small increased risk of developing AF taking any NSAID.
Discussion
This nationwide study examined the risk of AF with NSAID use compared with no NSAID use in post-MI patients. The main result was that utilization of NSAIDs was associated with an increased risk of AF in this large patient population.
Atrial fibrillation is one of the most common arrhythmias and often coincides with MI.1–3 Previous studies have reported that the presence of AF in MI patients was a powerful adverse prognostic factor for in-hospital and long-term mortality even after adjustment of age and gender.4–7,23,24 Indeed, the Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Arteries trial reported significantly higher 1-year mortality in MI patients with AF compared with those without AF.6 Furthermore, patients developing AF after an MI had an increased risk for thrombo-embolic complications, in particular stroke, both during hospitalization and during the follow-up period.5–7 Recently, the use of NSAIDs has been found to be associated with development of AF in a population-based study in Denmark, and these investigators reported a 40–70% increased relative risk of AF in patients using NSAIDs compared with those not using NSAIDs, with the most pronounced risk being associated with the selective COX-2 inhibitors and the lowest risk with the non-selective NSAIDs.9,10 Comparable results were found in a meta-analysis that summarized data from 114 clinical trials involving 116 094 patients using COX-2 inhibitors.25 Rofecoxib was found to be associated with the highest risk of any kind of arrhythmias (relative risk 2.90; 95% CI 1.07–7.88), but the risks of different subtypes of arrhythmias such as AF were not examined. Another study examined the risk of chronic AF associated with long-term use of NSAIDs and found that both non-selective NSAIDs and COX-2 inhibitors were associated with increased risk of chronic AF.26 To the best of our knowledge, our study is the first to report the association between the use of NSAID and the risk of AF in post-MI patients. Notably, we found that the use of both the non-selective NSAIDs and the selective NSAIDs was associated with relative high risk of developing AF, but with an absolute risk of 0.4% per year, i.e. 4 per 1000 patients per year. Time-to-event analyses performed in some of the clinical trials and in register studies of the selective COX-2 inhibitors have shown similar results as our data.12,27,28 The use of NSAIDs may be associated with increased risk of AF through adverse renal effects such as fluid retention, electrolyte disturbances, and hypertension, but evidence for such effects is limited. NSAIDs inhibit the COX enzymes, which are involved in the production of prostaglandins. The two major COX isoenzymes are COX-1 and COX-2, which both form prostaglandin H2 from arachidonic acid. Platelets express only COX-1, but endothelial cells express both COX-1 and COX-2. Platelets play an important role in the cardiovascular haemostasis. COX-1 produces thromboxane A2, which stimulates platelet aggregation and vasoconstriction, and increases vascular and cardiac remodelling. COX-2 mediates synthesis of prostaglandin I2, which is a potent vasodilator, inhibits platelet function, and promotes renal sodium excretion. A proposed mechanism for the cardiovascular risk of NSAIDs has been the observed shift in the prothrombotic/antithrombotic balance on endothelial surfaces towards thrombosis after NSAID exposure, i.e. because of a predominant inhibition of COX-2 in face of unopposed TXA” production from platelets. However, this theory of balanced vs. unbalanced COX inhibition is now debatable, because the non-selective NSAIDs have also been associated with increased cardiovascular risk. However, other mechanisms may explain the harmful effects of NSAIDs. Prostaglandin I2 has been found to act as a restraint on many prothrombotic stimuli (e.g. collagen, thrombin, adenosine diphosphate, adrenaline, serotonin, and TxA2).29–37 Non-steroidal anti-inflammatory drugs are often used in the general population, and despite the fact that guidelines specifically discourage the use of NSAID in patients with cardiovascular diseases, no decline in the use of NSAIDs has been seen in Denmark (http://www.medstat.dk/da/viewDataTables/). An association between the use of NSAIDs and the development of AF in post-IM patients could therefore have major public health impact. The HAS-BLED [uncontrolled hypertension >160 systolic, abnormal renal/liver function, prior haemorrhagic stroke, history of major bleeding, labile international normalized ratio, elderly (age > 65), drugs therapy with aspirin, NSAIDs, and alcohol intake (>7 alcoholic drinks per week)] score is used for stratification of bleeding risk in patients with AF, and one of the risk factors included in this score is the use of NSAIDs. Therefore, the use of NSAIDs would appear to represent a double-edged sword in that it increases the risk of AF and at the same time adds to bleeding risk, not least in the advent of anticoagulation therapy.
Limitations
The main limitation of our study is its observational nature. There is lack of information about important clinical parameters such as blood pressure, body mass index (BMI), smoking habits, lipid levels, and left ventricular ejection fraction, and therefore, the effect of unmeasured confounders cannot be excluded. Our calculations showed that if an unmeasured confounder or a combination of confounders was present in 20% of the cohort treated with NSAIDs, the confounder would have to elevate the risk by a factor of 2.8–3.5 to explain the increased risk observed in our study. Existence of such a confounder or a combination of confounders is highly unlikely, but not entirely impossible, since we had no information on other important risk factors such as smoking, lipid levels, or BMI. Information about the precise indication for initiation of NSAID treatment was not available, but NSAIDs are not recommended for the treatment of patients with ischaemic heart disease and it is unlikely that AF would be treated with NSAIDs. However, the clear relation between degree of COX-2 inhibition and risk of developing AF indicates the importance of the drugs rather than the indications. Having users treated with NSAID prior to the inclusion may result in confounding (healthy-user effect), e.g. because subjects not experiencing AF while taking NSAIDs remain on treatment, while AF events in others results consequently ends in discontinuation of treatment. However, our estimates when excluding prevalent NSAID users remained the same. Another possible bias when using prescription data is the uncertainty about adherence to treatment. Indeed, in observational studies, there is always a possibility that the patients do not take their prescribed medications. On the other hand, non-adherence would dilute the association between the exposure and outcome.
This nationwide study of patients with prior MI demonstrated that the treatment with most NSAIDs is associated with increased risk of developing AF compared with no NSAID treatment.
This study supports previous studies and adds important information to the current knowledge of the cardiovascular safety of NSAIDs in patients with prior MI. Further studies and preferably randomized clinical trials are warranted to establish the cardiovascular safety of NSAIDs.
Supplementary material
Supplementary material is available at European Heart Journal online.
Funding
Dr Gislason is supported by an unrestricted clinical research scholarship from the Novo Nordisk Foundation. The funding source had no influence on study design, interpretation of the results, or the decision to submit the article.
Conflict of interest: Dr Gislason reports grants and personal fees from AstraZeneca, grants from Bristol Meyers Squibb, and grants and personal fees from Pfizer, outside the submitted work.
References
