Detection of Atrial Fibrillation After Stroke and the Risk of Recurrent Stroke

Detection of Atrial Fibrillation After Stroke and the Risk of Recurrent Stroke Hooman Kamel, MD,* Derek R. Johnson, MD,* Manu Hegde, MD, PhD,* Alan S. Go, MD,†‡x Stephen Sidney, MD, MPH,x Michael Sorel, MPH,x Nancy K. Hills, PhD,* and S. Claiborne Johnston, MD, PhD*

Failure to expeditiously diagnose atrial fibrillation (AF) as the cause of ischemic stroke has unclear consequences. We studied the association between detection of AF after discharge and the risk of recurrent stroke. We followed a prospectively assembled cohort of patients hospitalized for stroke for 1 year for new diagnoses of AF and recurrent stroke. We compared rates of recurrent stroke in patients with a new diagnosis of AF and those without a new diagnosis of AF after discharge using Kaplan‒Meier survival statistics. We conducted Cox proportional hazards analysis of the diagnosis and timing of AF and recurrent stroke after adjustment for age, sex, race, preexisting AF, hypertension, dyslipidemia, diabetes, previous stroke, and use of antithrombotic and statin medications. Among 5575 patients with stroke, 113 (2.0%) received a new diagnosis of AF after discharge, and 221 (4.0%) had recurrent stroke. At 1 year, the Kaplan‒ Meier rate of recurrent stroke was 18.9% in those with a new diagnosis of AF and 4.5% in others, including those with AF diagnosed before or during the index hospitalization (P 5 .001). The association between a new diagnosis of AF and stroke recurrence persisted after adjustment for potential confounders (hazard ratio, 5.6; 95% confidence interval, 3.4-9.1). A new diagnosis of AF after discharge for stroke is associated with an increased risk of recurrent stroke, even compared with patients with known AF. These findings identify a subset of patients at high risk for recurrent stroke and highlight the importance of timely detection of AF in patients with stroke. Key Words: Atrial fibrillation—diagnosis—electrocardiography—embolism—Stroke. Ó 2012 by National Stroke Association

Atrial fibrillation (AF) affects more than 2 million Americans and causes approximately 75,000 strokes in the United States each year.1-3 It is an important stroke risk factor to recognize, given that treatment of AF with From the *Departments of Neurology; †Epidemiology and Biostatistics; ‡Medicine, University of California San Francisco, San Francisco, California; and xDivision of Research, Kaiser Permanente Northern California, Oakland, California. Received March 2, 2011; accepted March 21, 2011. Supported by Centers for Disease Control and Prevention Grant MM-0620-04/04, administered through the Association of American Medical Colleges. Address correspondence to Hooman Kamel, MD, 505 Parnassus Avenue, Box 0114, San Francisco, CA 94143. E-mail: hooman.kamel@ ucsf.edu. 1052-3057/$ - see front matter Ó 2012 by National Stroke Association doi:10.1016/j.jstrokecerebrovasdis.2011.03.008

726

oral vitamin K antagonists (eg, warfarin) or direct thrombin inhibitors (eg, dabigatran) substantially reduces the risk of primary and recurrent stroke.4-8 Because AF is frequently paroxysmal and asymptomatic, patients with AF may remain undiagnosed even after presenting with ischemic stroke.9 To increase the chance of detecting underlying AF, current consensus-based guidelines recommend at least 24 hours of continuous cardiac monitoring in patients admitted with ischemic stroke.10 However, numerous studies have shown that this screening approach misses AF in at least 5% of patients with ischemic stroke.11-21 Prolonging the duration of monitoring detects more cases, but there is no consensus supporting a specific duration of extended monitoring.22 Several factors may explain why prolonged cardiac monitoring is not more widely used after stroke. First, most studies supporting its use are small. Furthermore,

Journal of Stroke and Cerebrovascular Diseases, Vol. 21, No. 8 (November), 2012: pp 726-731

ATRIAL FIBRILLATION AND RECURRENT STROKE

concerns have been raised about the cost of this monitoring,22 although a formal analysis indicated that outpatient cardiac monitoring after stroke is cost-effective.23 In addition, whether expeditiously identifying underlying AF in patients with stroke improves clinical outcomes is unclear. Patients with known cardioembolic stroke do not have a markedly higher risk of recurrent stroke; in fact, their risk may be lower compared with patients with stroke caused by atherothrombotic disease,24-26 although the temporal profile and severity of recurrence may differ between the 2 mechanisms. As a result, it has been suggested that prolonged cardiac monitoring in patients with stroke can be safely deferred, allowing patients with underlying AF to eventually declare themselves.27,28 On the other hand, patients with undiagnosed AF who present with stroke may be at a greater risk for recurrent stroke than those with known AF, supporting the timely diagnosis and treatment of AF. To study this question, we examined the association between detection of AF after ischemic stroke and the risk of recurrent stroke.

Subjects and Methods Design We identified a cohort of all patients admitted with ischemic stroke over a 3-year period within Kaiser Permanente Northern California (KPNC), a large integrated healthcare delivery system in northern California caring for more than 3 million members who are highly representative of the local and statewide population.3 This cohort of patients was assembled as part of the Quality Improvement in Stroke Prevention (QUISP) study,29 a randomized trial of standardized discharge orders after ischemic stroke. The QUISP trial used a matched-pair design to randomize hospitals within the KPNC system to continue usual care or adopt standardized discharge orders for the care of patients admitted with ischemic stroke. Patients hospitalized at participating hospitals were followed over time to assess the impact of the intervention on the rate of use of secondary stroke prevention measures. For our analysis, we studied this QUISP cohort to examine the association between receipt and timing of a diagnosis of AF after discharge from the index hospitalization and the risk of recurrent ischemic stroke. Our study was approved by the Institutional Review Board of the Kaiser Foundation Research Institute and the Committee on Human Research at the University of California San Francisco. The requirement for informed consent and HIPAA subject authorization were waived because of the minimal risk to patients.

Subjects Starting in July 2004, 12 of the 16 KPNC hospitals were randomized using a matched-pair design to continue usual care or to implement standardized discharge orders

727

for patients with ischemic stroke. The impact of this intervention was assessed in a cohort of all plan members aged .40 years admitted with ischemic stroke to these 12 hospitals and 2 other KPNC hospitals between January 1, 2004, and December 31, 2006. Eligible patients were identified through a search of an automated database for primary discharge diagnoses of ischemic stroke (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] codes 433.01, 433.11, 433.21, 433.31, 433.81, 433.91, 434.01, 434.11, and 434.91). Computed tomography (CT) or magnetic resonance imaging (MRI) of the brain was required to exclude patients with hemorrhagic stroke. In a validation study, a nurse medical record analyst reviewed the medical records of 761 patients (14%), and the diagnosis of ischemic stroke was confirmed in all cases.

Measurements We used electronic medical records and previously validated diagnostic codes30 to obtain patients’ demographic characteristics; relevant medical history, including a history of preexisting AF, previous ischemic stroke, hypertension, diabetes, chronic heart failure, coronary artery disease, and dyslipidemia; disposition at discharge from the initial hospitalization; the use of antithrombotic and statin medications on discharge; and the number of outpatient visits and hospitalizations during the year after discharge. Patients diagnosed with AF before or during the hospitalization for the index stroke were considered to have preexisting AF. Patients received routine clinical follow-up after discharge, and we reviewed electronic medical records for 1 year after discharge. Using previously validated methods,3 we identified new documented diagnoses of AF after the index hospitalization by searching databases containing diagnosis codes from clinic visits, hospitalizations, and reports of any electrocardiograms obtained during routine clinical care. Patients with stroke admitted to KPNC undergo 24 hours of inpatient cardiac monitoring, in accordance with clinical guidelines; more extended cardiac monitoring is not routinely performed, but may be done at the discretion of the treating physician. Our second outcome was a validated recurrent stroke. All potential recurrent strokes were identified by searching for ICD-9-CM codes 430.xx‒ 434.xx in the primary hospital discharge code position in an automated hospital discharge diagnosis database, as well as a billing claims database that includes a comprehensive listing of hospitalizations outside the KPNC system, because KPNC is financially responsible for such admissions. To ensure that these diagnoses of recurrent stroke were indeed new strokes and not coding errors or complications of the first stroke, 2 neurologists (M.H. and D.R.J.) independently reviewed medical charts from the hospitalization for recurrent stroke, with a third neurologist resolving any disagreements (H.K.).

H. KAMEL ET AL.

728

Table 1. Baseline characteristics of the study sample, overall and stratified by recurrent stroke status All Mean (SD) Age, years Number (%) Women Known AF Diagnosed hypertension Dyslipidemia Coronary artery disease Chronic heart failure Previous ischemic stroke Diabetes mellitus

Recurrent stroke

No recurrent stroke

P

73 (13)

72 (13)

73 (13)

.05

2956 (53) 1338 (24) 4397 (79) 845 (15) 1624 (29) 896 (16) 1021 (18) 1989 (36)

129 (58) 48 (22) 185 (84) 48 (22) 72 (33) 35 (16) 52 (24) 93 (42)

2827 (53) 1290 (24) 4212 (79) 797 (15) 1552 (29) 861 (16) 969 (18) 1896 (35)

.10 .42 .07 .01 .25 .92 .04 .04

Statistical Analysis Descriptive statistics using exact binomial confidence intervals were used to calculate the proportion of patients with new diagnoses of AF and recurrent stroke. The rates of new diagnoses of AF and recurrent stroke also were calculated using Kaplan‒Meier survival statistics. Follow-up was censored on the day on which patients died or on the last day of the last month of active KPNC membership, as applicable. The survival curve for recurrent stroke was stratified by AF status (ie, new diagnosis of AF after discharge, preexisting AF, or no diagnosis of AF). These curves were compared using the log-rank test. Cox proportional hazards analysis was used to identify independent predictors of recurrent stroke, including preexisting and new diagnoses of AF, demographic and medical history data listed above, and the discharge disposition. To control for possible detection bias resulting from patients with recurrent stroke undergoing additional electrocardiography, we also controlled for the number of outpatient visits and repeat hospitalizations before a new diagnosis of AF. Because we could not determine what proportion of AF diagnoses represented newly developed AF versus undiagnosed preexisting AF, our primary analysis included all new diagnoses of AF made in the year after discharge regardless of the temporal relationship to a diagnosis of recurrent stroke. This approach assumed that some proportion of new diagnoses reflected preexisting disease; to examine the effect of that assumption on our results, we performed the same analysis excluding new diagnoses of AF made after a recurrent ischemic stroke. We constructed our models by beginning with all covariates and eliminating in a stepwise fashion those that did not approach statistical significance (P , .20). We used log-log plots and Schoenfeld residuals to verify the proportionality assumption. We performed a similar multivariable analysis to identify independent predictors of a new diagnosis of AF.

Results During the study period, a total of 5575 patients were hospitalized with confirmed ischemic stroke at the 14 study sites. The demographic characteristics and medical comorbidities of these patients were consistent with previously published epidemiologic profiles of patients with stroke (Table 1).2 Within 1 year of discharge, 113 patients (2.0%; 95% confidence interval [CI], 1.7%-2.4%) received a new diagnosis of AF. In multivariable analysis, older age (hazard ratio [HR] per decade, 1.3; 95% CI, 1.1-1.6) was the only independent risk factor for receiving a new diagnosis of AF after stroke. Within 1 year of their index stroke, 221 patients (4.0%; 95% CI, 3.5%-4.5%) were hospitalized for recurrent stroke. Of these patients, 19 received a new diagnosis of AF after their initial hospitalization, constituting 8.6% of those with recurrent stroke (95% CI, 4.9%-12.3%). The Kaplan‒Meier rate of recurrent stroke was 18.9% (95% CI, 12.4%-28.2%) in those with a new diagnosis of AF, compared with 4.9% (95% CI, 3.7%-6.4%) in those with preexisting AF and 4.3% (95% CI, 3.7%-5.1%) in those without AF (P 5 .001, log-rank test) (Fig 1). A new diagnosis of AF remained an independent predictor of

Figure 1.

Survival free of recurrent stroke stratified by type of AF.

ATRIAL FIBRILLATION AND RECURRENT STROKE

recurrent stroke after adjustment for age, sex, race, medical comorbidities, antithrombotic and statin use, discharge disposition, and number of outpatient visits and hospitalizations (HR, 5.6; 95% CI, 3.4-9.1) (Table 2). Among patients with both a new diagnosis of AF and recurrent stroke, AF was documented before the recurrent stroke in 16%, at the time of recurrent stroke in 53%, and after the recurrent stroke in 32%. For the 16% of patients with AF detected before recurrent stroke, two-thirds were treated with warfarin before the recurrent stroke. After exclusion of AF cases diagnosed after recurrent ischemic stroke, newly diagnosed AF remained an independent predictor of stroke recurrence (HR, 3.9; 95% CI, 2.2-7.0).

Discussion In a large community-based cohort of patients hospitalized with ischemic stroke, a diagnosis of AF made after discharge was associated with a 5-fold higher risk of recurrent stroke. This result held after controlling for potential confounding factors. Among patients with AF, the risk of recurrent stroke was greater in those with AF diagnosed after hospital discharge compared with those with preexisting AF or AF discovered at the time of hospitalization for the index stroke, suggesting that not all cases of underlying AF are being recognized and optimally treated during the initial hospitalization. Two lines of evidence support this possibility. First, new-onset AF has been shown to be associated with an especially high rate of stroke, and thus some patients presenting with stroke may harbor newly developed AF that has not yet been detected clinically.31 Second, numerous studies have highlighted the limitations of the current standard of care for screening stroke patients for this type of paroxysmal AF, 24 hours of cardiac monitoring. Studies using between 48 hours and 1 week of continuous cardiac monitoring after stroke detected new cases of AF in at least 5% of patients.11-13,20,32 Studies using longer periods of continuous monitoring found that up to 28% of patients had previously undetected AF.15-17,19,21,33 In addition, new diagnoses of AF were made in approximately 7% of patients followed clinically for 3 months in randomized trials of therapies for acute ischemic stroke.18 Table 2. Independent risk factors for recurrent ischemic stroke Adjusted HR (95% CI) Age per decade Female sex History of previous ischemic stroke New diagnosis of AF

0.9 (0.8-1.0) 1.3 (1.0-1.8) 1.5 (1.1-2.1) 5.6 (3.4-9.1)

729

Our study reflects the real-world interaction between AF and recurrent stroke in a community setting, being performed in a health care delivery system using the current standard of care, which does not include routine extended outpatient cardiac monitoring after stroke. Thus, although we cannot be certain, many of the new diagnoses of AF in our study may reflect preexisting but undetected disease. If this is so, then this type of latent and difficult-to-detect AF appears to pose an especially high risk for recurrent stroke, given that patients with known AF who are not receiving anticoagulation therapy have a 6%-10% annual risk of recurrent stroke,7,34 compared with the 19% annual risk seen in the present study. The high risk of recurrent stroke seen in this population defines a vulnerable subset of patients who should be identified expeditiously and treated aggressively. Our findings should be viewed in light of several study limitations. First, the association between new diagnoses of AF and the rate of recurrent stroke might be explained in part by detection bias, because patients hospitalized with recurrent stroke would have undergone more screening for AF than those without recurrent stroke. This likely does not entirely account for the association, however, given that we controlled for the frequency of outpatient clinic visits and hospitalizations between the index stroke and the time of a new diagnosis of AF. Furthermore, patients admitted with recurrent stroke typically would have undergone 24 hours of inpatient cardiac monitoring, which has a yield of ,4%,35 whereas 9% of our patients with recurrent stroke had newly diagnosed AF. In addition, new diagnosis of AF remained an independent predictor of recurrent stroke even after cases of AF diagnosed after stroke recurrence were excluded from the analysis. Second, although we were able to study a large number of patients in a community setting by using the electronic databases of a large integrated healthcare delivery system, relying on such computerized data may have drawbacks compared with prospective, systematic screening approaches, including inadequate case finding and inaccurate classification of cases.36 But because KPNC pays for hospitalizations outside its system, the databases that we used reliably capture diagnoses made outside the KPNC system and thus are very likely to capture most clinically recognized diagnoses of stroke and AF. In addition, we used rigorously validated methods to ascertain new diagnoses of AF,3 along with independent adjudication by neurologists to confirm the diagnosis of recurrent stroke. Furthermore, failure to capture new diagnoses of AF and diagnoses of recurrent stroke would tend to bias the study toward the null hypothesis, likely providing a more conservative estimate in our study. Third, our study was conducted simultaneously with a trial in which some KPNC hospitals were randomly assigned to develop standardized stroke discharge orders or to continue usual care in hospitalized stroke patients. This is unlikely to have biased our study

H. KAMEL ET AL.

730

findings, however, because randomization should not have affected the overall association between new diagnoses of AF and recurrent stroke. Fourth, we lacked information regarding anticoagulant use that was sufficiently detailed to allow comparisons of the rates and effects of anticoagulation therapy over time in patients with preexisting AF versus newly diagnosed AF. Similarly, we lacked reliable data on the mechanisms of stroke in our cohort. This is significant, because patients with a known noncardioembolic mechanism, such as lacunar disease, may be less likely to be diagnosed with AF later.21 Such information may be especially helpful in identifying stroke patients at high risk of harboring paroxysmal AF, because we were unable to find robust enough clinical predictors to allow targeted screening of stroke patients with more prolonged cardiac monitoring. In summary, a diagnosis of AF made after ischemic stroke was independently associated with a 5-fold increased risk of recurrent stroke in a large communitybased cohort of patients, accounting for 9% of recurrent strokes within the first year. This finding underscores the importance of making a timely diagnosis of AF in patients with otherwise unexplained stroke if AF is the cause, and supports further study of improved systematic methods of detecting paroxysmal AF occurring after stroke.

9.

10.

11.

12.

13.

14.

15.

References 1. Hart RG, Halperin JL. Atrial fibrillation and thromboembolism: A decade of progress in stroke prevention. Ann Intern Med 1999;131:688-695. 2. Fang J, Alderman MH, Keenan NL, et al. Declining US stroke hospitalization since 1997: National Hospital Discharge Survey, 1988‒2004. Neuroepidemiology 2007; 29:243-249. 3. Go AS, Hylek EM, Phillips KA, et al. Prevalence of diagnosed atrial fibrillation in adults: National implications for rhythm management and stroke prevention. The AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study. JAMA 2001;285:2370-2375. 4. Connolly S, Pogue J, Hart R, et al. Clopidogrel plus aspirin versus oral anticoagulation for atrial fibrillation in the Atrial Fibrillation Clopidogrel Trial with Irbesartan for prevention of Vascular Events (ACTIVE W): A randomised controlled trial. Lancet 2006;367:1903-1912. 5. Connolly SJ, Ezekowitz MD, Yusuf S, et al. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med 2009;361:1139-1151. 6. Aguilar MI, Hart R, Pearce LA. Oral anticoagulants versus antiplatelet therapy for preventing stroke in patients with non-valvular atrial fibrillation and no history of stroke or transient ischemic attacks. Cochrane Database Syst Rev 2007;CD006186. 7. Saxena R, Koudstaal PJ. Anticoagulants for preventing stroke in patients with nonrheumatic atrial fibrillation and a history of stroke or transient ischaemic attack. Cochrane Database Syst Rev 2004;CD000185. 8. Diener HC, Connolly SJ, Ezekowitz MD, et al. Dabigatran compared with warfarin in patients with atrial fibrillation and previous transient ischaemic attack or

16.

17.

18.

19.

20.

21.

22.

23.

stroke: A subgroup analysis of the RE-LY Trial. Lancet Neurol 2010;9:1157-1163. Savelieva I, Camm AJ. Clinical relevance of silent atrial fibrillation: Prevalence, prognosis, quality of life, and management. J Interv Card Electrophysiol 2000; 4:369-382. Adams HP Jr, del Zoppo G, Alberts MJ, et al. Guidelines for the early management of adults with ischemic stroke: A guideline from the American Heart Association/ American Stroke Association Stroke Council, Clinical Cardiology Council, Cardiovascular Radiology and Intervention Council, and the Atherosclerotic Peripheral Vascular Disease and Quality of Care Outcomes in Research Interdisciplinary Working Groups. Stroke 2007;38: 1655-1711. Schuchert A, Behrens G, Meinertz T. Impact of long-term ECG recording on the detection of paroxysmal atrial fibrillation in patients after an acute ischemic stroke. Pacing Clin Electrophysiol 1999;22:1082-1084. Barthelemy JC, Feasson-Gerard S, Garnier P, et al. Automatic cardiac event recorders reveal paroxysmal atrial fibrillation after unexplained strokes or transient ischemic attacks. Ann Noninvasive Electrocardiol 2003; 8:194-199. Jabaudon D, Sztajzel J, Sievert K, et al. Usefulness of ambulatory 7-day ECG monitoring for the detection of atrial fibrillation and flutter after acute stroke and transient ischemic attack. Stroke 2004;35:1647-1651. Liao J, Khalid Z, Scallan C, et al. Noninvasive cardiac monitoring for detecting paroxysmal atrial fibrillation or flutter after acute ischemic stroke: a systematic review. Stroke 2007;38:2935-2940. Wallmann D, Tuller D, Wustmann K, et al. Frequent atrial premature beats predict paroxysmal atrial fibrillation in stroke patients: An opportunity for a new diagnostic strategy. Stroke 2007;38:2292-2294. Tayal AH, Tian M, Kelly KM, et al. Atrial fibrillation detected by mobile cardiac outpatient telemetry in cryptogenic TIA or stroke. Neurology 2008;71:1696-1701. Elijovich L, Josephson SA, Fung GL, et al. Intermittent atrial fibrillation may account for a large proportion of otherwise cryptogenic stroke: A study of 30-day cardiac event monitors. J Stroke Cerebrovasc Dis 2009; 18:185-189. Kamel H, Lees KR, Lyden PD, et al. Delayed detection of atrial fibrillation after ischemic stroke. J Stroke Cerebrovasc Dis 2009;18:453-457. Ziegler PD, Glotzer TV, Daoud EG, et al. Incidence of newly detected atrial arrhythmias via implantable devices in patients with a history of thromboembolic events. Stroke 2010;41:256-260. Stahrenberg R, Weber-Kruger M, Seegers J, et al. Enhanced detection of paroxysmal atrial fibrillation by early and prolonged continuous holter monitoring in patients with cerebral ischemia presenting in sinus rhythm. Stroke 2010;41:2884-2888. Gaillard N, Deltour S, Vilotijevic B, et al. Detection of paroxysmal atrial fibrillation with transtelephonic EKG in TIA or stroke patients. Neurology 2010; 74:1666-1670. Bell C, Kapral M. Use of ambulatory electrocardiography for the detection of paroxysmal atrial fibrillation in patients with stroke. Canadian Task Force on Preventive Health Care. Can J Neurol Sci 2000;27:25-31. Kamel H, Hegde M, Johnson DR, et al. Cost-effectiveness of outpatient cardiac monitoring to detect atrial fibrillation after ischemic stroke. Stroke 2010;41:1514-1520.

ATRIAL FIBRILLATION AND RECURRENT STROKE 24. Sacco RL, Foulkes MA, Mohr JP, et al. Determinants of early recurrence of cerebral infarction. The Stroke Data Bank. Stroke 1989;20:983-989. 25. Moroney JT, Bagiella E, Paik MC, et al. Risk factors for early recurrence after ischemic stroke: The role of stroke syndrome and subtype. Stroke 1998;29:2118-2124. 26. Petty GW, Brown RD Jr, Whisnant JP, et al. Ischemic stroke subtypes: A population-based study of functional outcome, survival, and recurrence. Stroke 2000;31:1062-1068. 27. Kessler DK, Kessler KM. Is ambulatory electrocardiography useful in the evaluation of patients with recent stroke? Chest 1995;107:916-918. 28. Hankey GJ, Warlow CP. Cost-effective investigation of patients with suspected transient ischaemic attacks. J Neurol Neurosurg Psychiatry 1992;55:171-176. 29. Johnston SC, Sidney S, Hills NK, et al. Standardized discharge orders after stroke: Results of the Quality Improvement in Stroke Prevention (QUISP) cluster randomized trial. Ann Neurol 2010;67:579-589. 30. Sidney S, Sorel M, Quesenberry CP Jr, et al. COPD and incident cardiovascular disease hospitalizations and mortality: Kaiser Permanente Medical Care Program. Chest 2005;128:2068-2075.

731 31. Wolf PA, Kannel WB, McGee DL, et al. Duration of atrial fibrillation and imminence of stroke: The Framingham Study. Stroke 1983;14:664-667. 32. Lazzaro MA, Krishnan K, Prabhakaran S. Detection of atrial fibrillation with concurrent Holter monitoring and continuous cardiac telemetry following ischemic stroke and transient ischemic attack. J Stroke Cerebrovasc Dis 2012;21:89-93. 33. Sposato LA, Klein FR, Jauregui A, et al. Newly diagnosed atrial fibrillation after acute ischemic stroke and transient ischemic attack: Importance of immediate and prolonged continuous cardiac monitoring. J Stroke Cerebrovasc Dis 2012;21:210-216. 34. Connolly SJ, Pogue J, Hart RG, et al. Effect of clopidogrel added to aspirin in patients with atrial fibrillation. N Engl J Med 2009;360:2066-2078. 35. Hornig CR, Haberbosch W, Lammers C, et al. Specific cardiological evaluation after focal cerebral ischemia. Acta Neurol Scand 1996;93:297-302. 36. Ladouceur M, Rahme E, Pineau CA, et al. Robustness of prevalence estimates derived from misclassified data from administrative databases. Biometrics 2007; 63:272-279.