Atrial fibrillation (AF) is the most common arrhythmia in clinical practice. The goal of the present study was to quantify the association between use of nonsteroidal anti-inflammatory drugs (NSAIDs) and risk of AF incidence. MEDLINE and EMBASE were searched for studies that reported risk of AF associated with nonaspirin NSAID use. Combined relative risk (RR) estimates and 95% confidence intervals (CIs) were calculated using the random-effects model. Stratified meta-analyses were used to discern which patients were at the highest risk of AF due to NSAID use. Five studies were identified that met the inclusion criteria, 3 of which additionally reported specifically on the association between selective NSAIDs and risk of AF. Overall, NSAID use was associated with a 12% increased risk for AF incidence (RR 1.12, 95% CI 1.06 to 1.18). The association was found to be apparent among new users (RR 1.53, 95% CI 1.37 to 1.70). The increased risk of AF might be explained by the occurrence of chronic heart failure and kidney disease. In addition, use of selective NSAIDs was still related to an increased risk of AF (RR 1.24, 95% CI 1.18 to 1.30). Sensitivity analyses found results to be robust. In conclusion, use of nonaspirin NSAIDs was associated with an increased risk of incident AF. The association was found to be apparent for new users, with a 53% increase in risk. These findings suggest that AF needs to be added to the cardiovascular risks to be considered when prescribing NSAIDs.
Highlights
- •
Nonsteroidal anti-inflammatory drug use was associated with a 12% increased risk for atrial fibrillation incidence (relative risk 1.12, 95% confidence interval 1.06 to 1.18).
- •
The association was found to be greater among new users, with a 53% increase in the risk.
- •
These findings suggest that atrial fibrillation needs to be added to the cardiovascular risks to be considered when prescribing nonsteroidal anti-inflammatory drugs.
Atrial fibrillation (AF) is the most common rhythm disorder observed in clinical practice, which is associated with increased mortality and morbidity and a threefold to fourfold increased risk of thromboembolic stroke. Although the precise mechanisms of AF are not fully understood, inflammation may be involved. If so, anti-inflammatory medications, including nonsteroidal anti-inflammatory drugs (NSAIDs), may confer protection against AF. However, the epidemiologic literature regarding the effect of NSAID use on AF is mixed. Any confirmed association between use of NSAIDs and AF would have major clinical and public health implications because NSAIDs are commonly prescribed in daily practice to treat inflammatory conditions and pain. Therefore, the purpose of this study was to quantify the association between use of nonaspirin NSAIDs and AF incidence by meta-analyzing existing studies and to discern which patients were at a greater risk of AF due to NSAID use.
Methods
Our systematic review was conducted according to the Meta-analysis of Observational Studies in Epidemiology guidelines. Each author independently conducted a systematic search of MEDLINE and EMBASE from their inception to June 8, 2014, using the following key words: anti-inflammatory agents, nonsteroidal anti-inflammatory drugs, NSAID(s), and AF. Initial searches were restricted to English language publications and studies conducted in humans. A secondary search with no language restriction did not identify any additional relevant articles. Reference lists of articles identified for inclusion in the meta-analysis were examined to identify additional potentially relevant studies.
Abstracts of identified articles were screened to exclude studies that clearly did not meet the eligibility criteria. The full text of those selected for further review was obtained and evaluated. Studies were included if they met the following criteria: (1) presented original data from an epidemiologic study; (2) defined the outcome of interest as AF, based on clearly stated diagnostic criteria or identified through diagnostic codes with additional confirmation; (3) defined exposure as use of nonaspirin NSAIDs; (5) described adjustment for potential confounding; and (6) reported effect estimates with confidence intervals (CIs), standard errors, or sufficient information to calculate these.
Both authors independently extracted data from each study. Discrepancies were discussed and resolved by agreement. The following data were extracted from each study: study name, year of publication, setting, study design, number of participants, mean age, outcome definition, exposure definitions, duration of exposure, methods for confounding adjustment and variables adjusted for, effect estimates, and CIs or standard errors (or information required to compute these). When multiple effect estimates were reported, maximally adjusted estimates were extracted.
The quality of each study was assessed independently by 2 authors (GL and Y-PY) using the Newcastle-Ottawa Scale (NOS). The NOS consists of 3 parameters of quality: selection, comparability, and outcome. The NOS assigns a maximum of 4 points for selection, 2 points for comparability, and 3 points for exposure and/or outcome. Therefore, a score of 9 points indicate the highest quality, 6 to 8 points indicate medium quality, and <6 points indicate low quality. Any discrepancies were resolved by consensus.
Our meta-analysis and statistical analyses were performed by Stata 12.0 (Stata Corp. LP, College Station, Texas). A p value <0.05 was considered statistically significant, unless otherwise specified. Publication bias was assessed with funnel plots and Egger regression asymmetry test. Heterogeneity was measured using Cochran Q and the I 2 statistic. For the Q statistic, a p value <0.1 was considered statistically significant for heterogeneity; whereas for I 2 , a value >50% was considered significant heterogeneity. Measures of association were combined under the assumption that odds ratios were accurate approximations of relative risks (RRs). The pooled analyses were calculated using the random-effects model (DerSimonian and Laird method), which accounts for heterogeneity among studies.
Stratified analyses were conducted to examine differences by status of NSAID exposure (current vs recent use), duration of exposure for current users (new vs long-term use), and whether having co-morbidities such as congestive heart failure (HF) and chronic kidney disease in current users. Exposure to NSAIDs was categorized as current when the supply of the most recent prescription lasted until the index date (date of their initial diagnosis of AF) or ended in the several months (1 to 2 months) before it and as recent when it ended between 1 to 2 months and 1 year before the index date. For current users, new use was defined as those having their first ever NSAID prescription within 1 to 2 months before the index date, and long-term use was defined as those having their first ever NSAID >2 months before the index date.
To test the robustness of the results, we performed a 1-way sensitivity analysis. The scope of this analysis was to evaluate the influence of individual studies by estimating the average RR in the absence of each study. An additional subgroup analysis stratified by study design (cohort vs case-control) was also performed to estimate the robustness of the results.
Results
We retrieved 105 citations from database searches. After title and abstract screening, 96 were found not to be relevant to this meta-analysis and were excluded. After detailed evaluation of the remaining 9 full-text articles, 5 were excluded for reasons described in Figure 1 . Thus, 5 studies met our inclusion criteria and were included in the primary analyses.
Three of the 5 studies included in the analysis were case-control studies, and the remaining 2 were cohort studies ( Table 1 ). All 5 studies reported on the association between nonaspirin NSAIDs and AF, with 3 additionally reporting specifically on the association between selective NSAIDs and risk of AF. Most of the studies were controlled for major confounders (namely age, gender, smoking, cardiovascular disease, and diabetes). Of the 5 studies, 4 were conducted in Europe and 1 in Asia. Based on the NOS, 2 studies were of high quality and 3 of medium quality.
| Reference | Setting (Country) | n (Cases/Controls for Case-Control Studies) | Study Design | AF Diagnosis | Exposure Assessment | Potential Confounders Adjusted | Study quality |
|---|---|---|---|---|---|---|---|
| De Caterina, et al. (2010) | The UK General Practice Research Database | 11,560 (1,560/10,000) | Case-control | Electronic record and based on diagnostic code and validated through a questionnaire sent to primary care physicians | Automated prescription data | age, sex, smoking, BMI, alcohol use, co-morbidities such as ischemic heart disease, hypertension, and diabetes | High |
| Schmidt, et al. (2011) | the Danish National Registry of Patients | 358,520 (32,602/325,918) | Case-control | Electronic record and based on diagnostic code and diagnosed during hospitalization, outpatient clinic and emergency department | Aarhus University Prescription Database | age, sex, alcohol use, co-morbidities such as diabetes, hypertension, COPD, rheumatoid arthritis, and hyperthyroidism | High |
| Bäck, et al. (2012) | the Swedish Prescribed Drug Register, Patient Register, Cause of Death Register | 6792 ,112 | Cohort | Followed for incident events, and diagnosed based on ICD | Prescribed Drug Register record | age, sex, income, educational level, hypertension, rheumatoid arthritis, | Medium |
| Chao, et al. (2013) | the National Health Insurance Research Database (Taiwan) | 80,080 (7,280/72,800) | Case-control | Electronic record and based on diagnostic code | Automated prescription data | age, sex, co-morbidities such as hypertension, and diabetes, ischemic heart disease, hyperthyroidism | Medium |
| Krijthe, et al. (2014) | the Rotterdam Study (The Netherlands) | 8,423 | Cohort | Followed for incident events and diagnosed base on 12-lead ECG as well as medical records | Automated prescription data | age, sex, BMI, blood pressure, total cholesterol, smoking status, ischemic heart disease, diabetes | Medium |
We found that NSAID use was associated with a 12% increased risk of AF (RR 1.12, 95% CI 1.06 to 1.18; Figure 2 ). There was a moderate heterogeneity for the outcome (I 2 = 65%, p <0.001). Sensitivity analysis showed that the RRs for AF were similar without great fluctuation, ranging from 1.10 (95% CI 1.04 to 1.15) to 1.14 (95% CI 1.08 to 1.19). In addition, subgroup analysis showed that both cohort and case-control studies exhibit similar findings (data are not shown).
When the exposure to NSAIDs was categorized as current use and recent use, the RR associated with the current use of NSAIDs was 1.18 (95% CI 1.13 to 1.23) and the RR for recent users was 1.05 (95% CI 1.01 to 1.08). No heterogeneity was observed in these results (I 2 = 0%; Figure 2 ). When stratified by duration of NSAID exposure in current users, a greater risk of AF was noted for new users (RR 1.53, 95% CI 1.37 to 1.70; Figure 3 ).
Among patients with HF, there was an increased risk of AF in current NSAID users (RR 1.82, 95% CI 1.42 to 2.32; Figure 4 ). In addition, among patients with chronic kidney disease, there was also a greater risk of AF in current NSAID users than patients without this co-morbidity (RR 1.58, 95% CI 1.34 to 1.85).
