Abstract
Background
Atrial fibrillation (AF) is common in elderly patients and is associated with ischemic stroke. We sought to explore the current status of antithrombotic therapy in elderly patients with nonvalvular persistent AF.
Methods
This is a retrospective study and data were collected from the First Hospital of China Medical University. A total of 300 patients were enrolled from January 2015 to June 2017. Patients were divided into two groups: Group 1 (from 65‐74) and Group 2 (older than 75). The status of antithrombotic treatment was recorded. Follow‐ups were done at 7, 90 , 180, and 360 days after discharge. The occurrence of stroke was recorded.
Results
For 287 patients with a CHA2DS2‐VASc score ≥2, 41.10% received oral anticoagulants (OAC), 27.20% received new oral anticoagulants (NOAC), 42.20% received antiplatelet agents, 16.70% received no antithrombotic treatment. From 2015 to the first half 2017, the ratio of OAC was 25.90%, 51.89%, and 49.30%, respectively; ratio of NOAC were 16.90%, 30.19%, and 39.10%, respectively. During the four times follow‐up, percentage of patients who had good treatment compliance was 65%, 49.2%, 38.5%, and 25% stroke rate was 6.7% in total 300 patients. Logistic regression analysis showed age older than 75 (odds ratio [OR] 4.812), prior stroke (OR 4.109) were risk factors of stroke, and OAC treatment (OR 0.021) could prevent stroke.
Conclusion
Ratio of antithrombotic therapy in elderly patients with nonvalvular persistent AF was low and drug compliance was poor. Age, prior history of stroke, and OAC treatment are the important predictive factors of stoke in elderly patients.
Keywords: antithrombotic treatment, atrial fibrillation, elderly, stroke
1. INTRODUCTION
Atrial fibrillation (AF) is one of the most common cardiac arrhythmias. With the development of global population aging, its prevalence is expected to increase 3‐fold in the next three decades.1 AF was associated with an increasing risk of mortality and morbidity from stroke, which seriously affected the quality of life of patients.2 The data from the 1980s to 1990s estimated that 16% of ischemic strokes were associated with AF. And AF‐associated stroke increased to 28% in published global data from 2013 to 20146. CHA2DS2VASc score was recommended to assess risk of thromboembolic events: 1 point for age between 65 and 74 years, 2 point for age ≥75 years.3 Using antiplatelet agents alone to prevent thromboembolism in AF was not recommended by the 2016 ESC guideline, oral anticoagulant (OAC) was recommended in patients with different profiles for stroke prevention.4 A 2012 to 2015 data from China revealed the proportions of anticoagulants use in patients 65 to 74 years of age and 75 years or older were 28.5% and 20.7%, respectively.5 To further improve the standardized treatment and provide reference for narrowing the gap between domestic and foreign guidelines, we analyzed the antithrombotic treatment and stroke in elderly patients with nonvalvular persistent AF in our center.
2. METHODS
2.1. Participants
A retrospective single‐center survey was conducted in patients with nonvalvular persistent AF who were admitted to the First Affiliated Hospital of China Medical University between January 2015 and May 2017. Inclusion criteria: age older than 65 years, Chronic AF, defined as previously documented in the medical record detected by a 12‐lead electrocardiogram or 24‐hour Holter monitor. Patients were excluded if they died during hospitalization, or had a mechanical heart valve mitral stenosis or malignant tumor. Patients were divided into two groups by age: Group 1 (from 65‐74) and Group 2 (older than 75).
2.2. Data collection
Data collected from medical records included age, sex, body mass index (BMI), medical history and comorbid conditions, biomarkers, echocardiographic result, CHA2DS2‐VASc scores, HAS‐BLED scores, and antithrombotic treatment (no antithrombotic treatment, antiplatelet agents, oral anticoagulants (OAC), new oral anticoagulants (NOAC) + antiplatelet agents, vitamin K antagonist, VKA + antiplatelet agents). The CHA2DS2‐VASc score (cardiac failure or dysfunction [1 point], hypertension [1 point], age ≥75 years [2 points], diabetes [1 point], prior stroke/transient ischemic attack (TIA) [2 points], vascular disease other than cerebrovascular disease [1 point], age 65 to 74 years [1 point], female sex [1 point]) was used for stroke risk stratification.
2.3. Follow‐up
Patients discharged from hospital received follow‐up after 7, 90, 180, and 360 days to know the situation of the usage of antithrombotic drug and the occurrence of stroke.
2.4. Statistics
All statistical analyses were performed using spss Statistics 21.0 (SPSS Inc., Chicago, IL). The mean ± SD or median and interquartile range were used to describe continuous variables. Characteristics between two independent samples were compared using the chi‐square test or t test for categorical variables. The nonparametric test (Mann‐Whitney U test) was used for abnormal distribution data. Logistic regression analysis was performed to identify independent predictors of stroke. P value less than 0.05 on both sides was considered statistically significant.
3. RESULTS
3.1. General information
A total of 300 patients were admitted in this study. There were 134 patients in Group 1, 166 patients in Group 2. Comparing to Group 1, BMI of Group 2 was lower, the number of dilated cardiomyopathy (DCM) in Group 2 was less, D‐dimer in Group 2 was higher, estimated glomerular filtration rate (eGFR) in Group 2 was lower, the average left atrial diameter (LAD) in Group 2 was lower (P < 0.05). The most common three diseases related to AF were hypertension, coronary heart disease, and heart failure (Table 1).
Table 1.
Baseline characteristics of patients
| Variables | Group 1 (n = 134) | Group 2 (n = 166) | P |
|---|---|---|---|
| Age (years), mean (SD) | 69.31 ± 2.94 | 80.12 ± 3.81 | 0.000 |
| Sex, female, n (%) | 51 (38.1) | 86 (51.8) | 0.017 |
| BMI (kg/m2 ) | 25.79 ± 4.30 | 23.91 ± 4.21 | 0.375 |
| Smoking, n (%) | 37 (27.6) | 31 (18.17) | 0.071 |
| Alcohol, n (%) | 18 (13.4) | 13 (7.8) | 0.123 |
| Hypertension, n (%) | 87 (64.9) | 121 (72.9) | 0.141 |
| Coronary heart disease, n (%) | 64 (48.5) | 83 (53.6) | 0.701 |
| Myocardial infarction, n (%) | 12 (9.0) | 39 (13.0) | 0.237 |
| Heart failure, n (%) | 53 (39.6) | 67 (40.4) | 0.906 |
| DCM, n (%) | 9 (6.7) | 1 (0.6) | 0.007 |
| Diabetes mellitus, n (%) | 39 (29.1) | 43 (25.9) | 0.538 |
| Prior history of stroke, n (%) | 25 (18.7) | 44 (26.5) | 0.104 |
| Thromboembolism, n (%) | 2 (1.5) | 5 (3.0) | 0.388 |
| Hyperthyroidism, n (%) | 2 (1.5) | 3 (1.8) | 0.833 |
| CKD, n (%) | 5 (3.7) | 9 (5.4) | 0.492 |
| Stent placement, n (%) | 20 (14.9) | 24 (14.5) | 0.364 |
| BNP (pg/mL) | 263 (129 504) | 321 (164 611) | 0.334 |
| cTNI (ng/mL) | 0.009 (0.003,0.026) | 0.012 (0.004,0.038) | 0.080 |
| D‐D (μg/mL) | 0.490 (0.328,1.136) | 0.725 (0.400,1.396) | 0.873 |
| eGFR (mL/min) | 87.44 ± 28.78 | 79.48 ± 25.23 | 0.012 |
| SCR (μmol/L) | 79.83 ± 27.49 | 79.38 ± 36.25 | 0.906 |
| UA (μmol/L) | 376.92 ± 125.89 | 354.92 ± 115.44 | 0.126 |
| TSH (mIU/L) | 1.690 (0.996, 2.529) | 1.439 (0.886, 2.550) | 0.851 |
| LAD (mm) | 48.7 ± 6.3 | 46.4 ± 6.4 | 0.002 |
| IVS (mm) | 9.5 ± 1.8 | 9.9 ± 2.4 | 0.112 |
Abbreviations: BMI, body mass index; BNP, brain natriuretic peptide; CKD, chronic kidney disease; cTNI, cardiac troponin I; DCM, dilated Cardiomyopathy; D‐D, D‐DIMER; eGFR, estimated glomerular filtration rate; IVS, interventricular septal thickness; LAD, left atrial diameter; SCR, serum creatine; TSH, thyroid stimulating hormone; UA, uric acid.
3.2. Stroke risk stratification and hemorrhage risk stratification
The average CHA2DS2‐VASc score of all 300 patients was 3.70 ± 1.47 points, and 287 patients (95.7%) had score ≥2 points. When CHA2DS2‐VASc score was 2, 4, and 5 points, the number of patients in Group 2 was more than that in Group 1 (P < 0.05). No significant difference was observed between the two groups when score ≥6. The average HASBLED score was 1.95 ± 0.75 points, and 237 patients (79.0%) had HASBLED score <3 points. HASBLED score in Group 2 was higher than Group 1 (P < 0.05) (Table 2).
Table 2.
(a) CHA2DS2VASc score and (b) HASBLED score
| Score | Overall (n = 300) | Group 1 (n = 134) | Group 2 (n = 166) |
|---|---|---|---|
| (a) | |||
| Average, mean (SD) | 3.70 ± 1.47 | 3.04 ± 1.39 | 4.23 ± 1.32* |
| ≥2 score, n (%) | 287 (95.7) | 121 (42.2) | 166 (57.8) |
| 1 score, n (%) | 13 (4.3) | 13 (9.7) | 0 |
| 2 score, n (%) | 58 (19.3) | 44 (38.2) | 14 (8.4) |
| 3 score, n (%) | 63 (21.0) | 33 (24.6) | 30 (18.1) |
| 4 score, n (%) | 83 (27.7) | 21 (15.7) | 62 (37.3) |
| 5 score, n (%) | 51 (17.0) | 16 (11.9) | 35 (21.1) |
| 6 score, n (%) | 21 (7.0) | 5(3.7) | 16(9.6) |
| 7 score, n (%) | 9 (3.0) | 2 (1.5) | 7 (4.2) |
| 8 score, n (%) | 1 (0.3) | 0 | 1 (0.6) |
| 9 score, n (%) | 1 (0.3) | 0 | 1 (0.6) |
| Score | Overall (n = 300) | Group 1 (n = 134) | Group 2 (n = 166) |
|---|---|---|---|
| (b) | |||
| Average, mean (SD) | 1.95 ± 0.75 | 1.83 ± 0.74 | 2.04 ± 0.74** |
| <3 score, n (%) | 237 (79.0) | 109 (81.3) | 128 (77.1) |
| 1 score, n (%) | 85 (28.3) | 49 (36.6) | 36 (21.7) |
| 2 score, n (%) | 152 (50.7) | 60 (44.8) | 92 (55.8) |
| 3 score, n (%) | 58 (19.3) | 24 (17.9) | 34 (20.5) |
| 4 score, n (%) | 4 (1.3) | 1 (0.7) | 3 (1.8) |
| 5 score, n (%) | 1 (0.3) | 0 | 1(0.6) |
P = 0.000 (P value of Group 1 vs Group 2).
P = 0.011 (P value of Group 1 vs Group 2).
3.3. Drug analysis of antithrombotic therapy in patients with CHA2DS2‐VASc score ≥2
CHA2DS2‐VASc score≥2. The comparison of OAC and antiplatelet use between Groups 1 and 2 are listed in Table 3. Comparing antithrombotic treatment in Groups 1 and 2, there is no significant difference between the two groups (all P > 0.05).
Table 3.
Antithrombotic situation in patients with CHA2DS2VASc score ≥2
| Antithrombotic treatment | Overall (n = 300) | Group 1 (n = 134) | Group 2 (n = 166) |
|---|---|---|---|
| No antithrombosis, n (%) | 48 (16.7) | 19 (15.7) | 29 (17.5) |
| Antiplatelet alone, n (%) | 121 (42.2) | 51 (42.1) | 70 (42.2) |
| OAC alone, n (%) | 118 (41.1) | 51 (42.1) | 67 (40.4) |
| Warfarin alone, n (%) | 22 (7.7) | 12 (9.9) | 10 (6.0) |
| NOAC alone, n (%) | 78 (27.2) | 34 (28.1) | 44 (26.5) |
| Warfarin + antiplatelet, n (%) | 4 (1.4) | 1 (0.8) | 3 (1.8) |
| NOAC + antiplatelet, n (%) | 14 (4.9) | 4 (3.3) | 10 (6.0) |
Abbreviations: OAC, oral anticoagulant; NOAC, new oral anticoagulant.
All P value >0.05.
3.4. Antithrombotic therapy for patients CHA2DS2VASc score ≥2 during hospitalization
Among 287 patients with CHA2DS2VASc score ≥2 points, 41.10% received OAC, 27.20% received NOAC, 42.20% received antiplatelet agents, and 16.70% received no antithrombotic treatment. From 2015 to 2017, calculated per half year, the ratio of anticoagulant treatment were 23.70%, 28.35%, 48.10%, 55.60%, and 49.30% respectively; ratio of using NOAC were 16.90%, 17.00%, 23.10%, 37.00%, 49.30%, and 39.10%, respectively (Figure 1A).
Figure 1.
A, Trends of antithrombotic treatment rates among patients CHA2DS2‐VASc scores ≥2. B, Antithrombotic therapy for patients CHA2DS2‐VASc score ≥2 after discharge. Ratio of patients insisting on antithrombotic treatment as doctors' advice decreased with time. NOAC, new oral anticoagulant; OAC, oral anticoagulant
3.5. Antithrombotic therapy for patients CHA2DS2VASc score ≥2 after discharge
After leaving hospital, follow‐ups were done for patients with CHA2DS2VASc score ≥2. About 118 patients received anticoagulant therapy during hospitalization. The follow‐up ratio was 100% in 7 and 90 days, 79.4% for 180 days, and 60% for 360 days. Ratio of patients insisting on antithrombotic treatment as doctors' advice decreased with time (Figure 1B).
3.6. Factors affecting the stroke associated with elderly patients with nonvalvular persistent atrial fibrillation
During the entire follow‐up, the stroke rate of 300 patients was 6.7%. Logistic analysis showed age older than 75 years old (odds ratio [OR]: 4.284, 95% confidence interval [CI]: 1.315‐18.585), prior stoke/TIA (OR: 4.943, 95% CI: 1.391‐13.190) were risk factors for stroke, and anticoagulation (OR: 0.023, 95% CI: 0.083‐0.858) was an important protective factor in preventing stroke (Table 4).
Table 4.
Factors affecting the stroke
| Factors | OR value | CI (95%) | P |
|---|---|---|---|
| Age ≥75 y | 4.812 | 1.315‐17.144 | 0.011 |
| Anticoagulants | 0.210 | 0.083‐0.859 | 0.034 |
| Previous stroke/TIA | 4.109 | 1.562‐10.8 | 0.023 |
Abbreviations: CI, confidence interval; OR, odds ratio; TIA, transient ischemic attack.
4. DISCUSSION
Patients with AF suffered a 5‐fold higher stroke risk than the general population7. AF‐associated ischemic stroke often causes disability and high mortality. Stroke rates ranged from 7% to 9% per year. Three hundred subjects in this study were followed up for 3 to 12 months and the ischemic stroke rate was 6.7%. However, in our study only 21.4% patients who had ischemic strokes received OAC, which is lower than International general level.8 Recent studies have shown that biomarkers (eg, BNP, brain natriuretic peptide, cTnI, cardiac troponin I, D‐dimer, and creatine) can also improve stroke risk stratification for patients with AF. Some studies validate ABC (age, biomarkers, clinical history)‐stroke score as a novel risk stratification system for predicting stroke.9, 10, 11 In this study, we found older age, prior stroke, or TIA were closely related to stroke risk, and anticoagulant agents could significantly reduce stroke rate.
In our study, comorbid conditions of elderly patients with nonvalvular persistent AF were hypertension (69.3%), coronary heart disease (49.0%), heart failure (40.0%), and diabetes (27.3%). A 1999 to 2001 Chinese survey showed comorbidities of AF were hypertension (40.3%), coronary heart disease (34.8%), heart failure (33.1%), and diabetes (4.1%).12 In both Western and Asian studies, hypertension (21%‐73%) and diabetes (8%‐45%) were the most common comorbidities.13 Our result coincides with those studies.
DCM is one of the causes of AF. The prevalence of AF in DCM ranged from 22% to 25%.14 Our study showed the number of DCM in Group 1 was more than that in Group 2. More attention should be paid on the patients with DCM and 65 to 74 years of age for stroke prevention. In our study, only 10 patients had DCM, which is not enough to make a significant comparison. More researches with large sample size are needed to confirm the result. Previous studies had shown that LAD was associated with incidence of AF and age.15, 16 However, our study was contrary to those studies, which may result from more patients with DCM in Group 1.
High D‐dimer level is associated with thrombosis. Both RELY and ARISTOTLE study verified that the level of D‐dimer was associated with the incidence of stoke. In addition, for patients with an eGFR lower than 45 mL/min, the relative risk of stroke was 1.39 more than patients with eGFR higher than 60 mL/min.10 In our study, D‐dimer in Group 2 was higher and eGFR in Group 2 was lower, which showed risk of thromboembolism increased with age. Thus, strict anticoagulantion therapy should be taken for elderly patients with AF.
In our study, 41.10% patients with nonvalvular persistent AF received anticoagulant treatment (among whom 27.20% received NOACs), 42.20% patients received pure antiplatelet agents. The anticoagulation situation is superior to that of data in our country from 2011 to 2014.17 In that report, only 36.5% patients received anticoagulant treatment. The baseline data from GLORIA‐AF phase 2 demonstrates that 79.9% received OAC (among whom 47.6% received NOACs), 12.1% received antiplatelet treatment.8 There was a big gap in the antithrombotic situation between our center and the international level. However, the good thing was the anticoagulation ratio increased from 2015 to 2017, especially the usage of NOACs. There was no significant difference in anticoagulation between the two groups after leaving hospital, and both groups had poor compliance. During the entire study period, the application of OAC significantly reduced after discharge, and three quarters of patients discontinued anticoagulants after 1 year. Compared with good compliance in recent studies in the United States which showed 47.5% patients using NOACs and 40.2% in patients using Warfarin, anticoagulation therapy in our country is quite insufficient.18 Some studies found that low prescription rate was partly responsible for underuse of anticoagulants. However, less effort had been focused on increasing drugs compliance. The true medication compliance may be lower than what we found. Another interesting result is many patients use pure antiplatelet agents. In fact, only 21 patients had percutaneous coronary intervention (PCI) within 1 year, all of them had dual antiplatelet therapy (aspirin + clopidogrel, or aspirin + ticagrelor); 23 patients who had PCI more than 1 year used single antiplatelet without OAC. We cannot attribute the great amount of antiplatelet use to PCI. During our follow‐up, patients were more willingly to accept antiplatelet treatment than OAC. Patients need to test International Normalized Ratio (INR) when using Warfarin, which is inconvenient; and NOACs are too expensive.
In all, our study showed poor OAC rate in elderly patients with AF and using OAC could decrease stroke risk. In clinical practice, we should take a more active attitude to prescribe OAC in elderly patients with AF. And encourage patients to follow doctors' advice, to have good drug compliance is an important step to reduce stroke risk.
5. LIMITATION
There were several limitations in this study. First, this was a single‐center study, the sample size was small, and the results may not be generalizable to all patients. Second, it is a retrospective study, not all patients followed‐up at 7, 90, 180, and 360 days after leaving hospital because of loss to follow‐up and time limit. Third, the reasons of the poor compliance were not registered. Lastly, only one biochemical index was recorded resulting in partial bias in the results.
Conflicts of interest
The authors declare no potential conflict of interests.
Yu L‐J, Chen S, Xu Y, Zhang Z‐X. Clinical analysis of antithrombotic treatment and occurrence of stroke in elderly patients with nonvalvular persistent atrial fibrillation. Clin Cardiol. 2018;41:1353–1357. 10.1002/clc.23057
Lu‐Jiao Yu and Shuai Chen contributed equally to this study.
REFERENCES
- 1. Morin DP, Bernard ML, Madias C, Rogers PA, Thihalolipavan S, Estes NAM III. The state of the art: atrial fibrillation epidemiology, prevention, and treatment. Mayo Clin Proc. 2016;91(12):1778‐1810. [DOI] [PubMed] [Google Scholar]
- 2. Son MK, Lim NK, Kim HW, Park HY. Risk of ischemic stroke after atrial fibrillation diagnosis: a national sample cohort. PLoS One. 2017;12(6):e0179687. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Zhu WG, Xiong QM, Hong K. Meta‐analysis of CHADS2 versus CHA2DS2‐VASc for predicting stroke and thromboembolism in atrial fibrillation patients independent of anticoagulation. Tex Heart Inst J. 2015;42(1):6‐15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Kirchhof P, Benussi S, Kotecha D, et al. 2016 ESC Guidelines for the management of atrial fibrillation developed in collaboration with EACTS. Europace. 2016;18(11):1609‐1678. [DOI] [PubMed] [Google Scholar]
- 5. Zhang J, Yang XA, Zhang Y, et al. Oral anticoagulant use in atrial fibrillation‐associated ischemic stroke: a retrospective, multicenter survey in northwestern China. J Stroke Cerebrovasc Dis. 2017;26(1):125‐131. [DOI] [PubMed] [Google Scholar]
- 6. Perera KS, Vanassche T, Bosch J, et al. Global survey of the frequency of atrial fibrillation‐associated stroke: embolic stroke of undetermined source global registry. Stroke. 2016;47(9):2197‐2202. [DOI] [PubMed] [Google Scholar]
- 7. Stroke Risk in Atrial Fibrillation Working Group . Independent predictors of stroke in patients with atrial fibrillation: a systematic review. Neurology. 2007;69(6):546‐554.17679673 [Google Scholar]
- 8. Huisman MV, Rothman KJ, Paquette M, et al. The changing landscape for stroke prevention in AF: findings from the GLORIA‐AF registry phase 2. J Am Coll Cardiol. 2017;69(7):777‐785. [DOI] [PubMed] [Google Scholar]
- 9. Hijazi Z, Lindback J, Alexander JH, et al. The ABC (age, biomarkers, clinical history) stroke risk score: a biomarker‐based risk score for predicting stroke in atrial fibrillation. Eur Heart J. 2016;37(20):1582‐1590. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Hijazi Z, Oldgren J, Siegbahn A, Wallentin L. Application of biomarkers for risk stratification in patients with atrial fibrillation. Clin Chem. 2017;63(1):152‐164. [DOI] [PubMed] [Google Scholar]
- 11. Connolly SJ, Ezekowitz MD, Yusuf S, et al. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med. 2009;361(12):1139‐1151. [DOI] [PubMed] [Google Scholar]
- 12. Wen‐Hang QI, Society of Cardiology, Chinese Medical Association . Retrospective investigation of hospitalised patients with atrial fibrillation in mainland China. Int J Cardiol. 2005;105(3):283‐287. [DOI] [PubMed] [Google Scholar]
- 13. Tse HF, Wang YJ, Ahmed Ai‐Abdullah M, et al. Stroke prevention in atrial fibrillation—an Asian stroke perspective. Heart Rhythm. 2013;10(7):1082‐1088. [DOI] [PubMed] [Google Scholar]
- 14. Tuomainen PO, Magga J, Fedacko J, et al. Idiopathic dilated cardiomyopathy and chronic atrial fibrillation. Clin Physiol Funct Imaging. 2014;34(2):133‐137. [DOI] [PubMed] [Google Scholar]
- 15. Zemrak F, Ambale‐Venkatesh B, Captur G, et al. Left atrial structure in relationship to age, sex, ethnicity, and cardiovascular risk factors: MESA (Multi‐Ethnic Study of Atherosclerosis). Circ Cardiovasc Imaging. 2017;10(2):e005379. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Tsang TS, Barnes ME, Bailey KR, et al. Left atrial volume: important risk marker of incident atrial fibrillation in 1655 older men and women. Mayo Clin Proc. 2001;76(5):467‐475. [DOI] [PubMed] [Google Scholar]
- 17. Chang SS, Dong JZ, Ma CS, et al. Current status and time trends of oral anticoagulation use among Chinese patients with nonvalvular atrial fibrillation: the Chinese atrial fibrillation registry study. Stroke. 2016;47(7):1803‐1810. [DOI] [PubMed] [Google Scholar]
- 18. Yao X, Abraham NS, Alexander GC, et al. Effect of adherence to oral anticoagulants on risk of stroke and major bleeding among patients with atrial fibrillation. J Am Heart Assoc. 2016;5(2):e003074. [DOI] [PMC free article] [PubMed] [Google Scholar]