Frequent Early Cardiac Complications Contribute to Worse Stroke Outcome in Atrial FibrillationTu H.T.H.a, b · Campbell B.C.V.a, b · Churilov L.d, e · Kalman J.M.a, c · Lees K.R.f · Lyden P.D.g · Shuaib A.h · Donnan G.A.e · Davis S.M.a, b · on behalf of the VISTA collaborators
aUniversity Department of Medicine, Departments of bNeurology and cCardiology, The Royal Melbourne Hospital, University of Melbourne, dDepartment of Mathematics and Statistics, and eFlorey Neuroscience Institutes, University of Melbourne, Melbourne, Vic., Australia; fInstitute of Cardiovascular and Medical Sciences, University of Glasgow, Glasgow, UK; gCedars-Sinai Medical Center, Los Angeles, Calif., USA; hDivision of Neurology, Department of Medicine, University of Alberta, Edmonton, Alta., Canada Corresponding Author
Background: Atrial fibrillation (AF) is associated with worse outcomes following ischemic stroke and more frequent cardiac complications in the general population. We aimed to establish whether early cardiac complications contribute to the poorer ischemic stroke outcomes in patients with AF, independent of baseline differences in age, stroke severity and cardiovascular risk factors. This might have important implications for acute stroke management in patients with AF. Methods: We searched VISTA-Acute, an academic database containing standardized data for 28,131 patients from 30 randomized-controlled acute stroke trials and 1 stroke registry, for imaging-confirmed placebo-treated patients with complete documentation of baseline demographics, cardiovascular risk factors, presence or absence of AF, neurologic impairment [National Institutes of Health Stroke Scale (NIHSS)], cardiac complications and 3-month outcome (modified Rankin Scale). A total of 2,865 patients from 6 randomized-controlled trials met the selection criteria, of whom 819 had AF. Binary logistic regression modeling was used to determine the independent effect of AF on stroke outcome and serious cardiac adverse events (SCAE), a composite end point including acute coronary syndrome, symptomatic heart failure, cardiopulmonary arrest, ventricular tachycardia, ventricular fibrillation and cardiac mortality. Results: All patients were enrolled into the source trials within 24 h of stroke onset. At baseline, patients with AF were older (mean 75 vs. 67 years, p < 0.001) and had greater neurologic impairment (median NIHSS 15 vs. 13, p < 0.001). The median time to first cardiac adverse event was 3 days [median difference 0, 95% confidence interval (CI) 0–1, p = 0.06] for both patients with and without AF. SCAE occurred more frequently [14.2 vs. 6%, odds ratio (OR) = 2.58, 95% CI 1.97–3.37] in patients with AF, particularly cardiac mortality (4.9 vs. 2.6%, OR = 1.89, 95% CI 1.25–2.88), symptomatic heart failure (6.5 vs. 2.2%, OR = 3.01, 95% CI 2.01–4.50), and ventricular tachycardia and/or fibrillation (2.4 vs. 0.8%, OR = 3.18, 95% CI 1.64–6.16). At 3 months, AF was independently associated with SCAE (OR = 2.14, 95% CI 1.61–2.86) and early mortality (OR = 1.44, 95% CI 1.14–1.81) after adjusting for all baseline imbalances. Conclusion: Early SCAE are common after stroke and are independently associated with the presence of AF. Given that many cardiac complications are potentially remediable, these results highlight the need for more rigorous surveillance for cardiac complications in acute ischemic stroke patients with AF.
© 2011 S. Karger AG, Basel
Atrial fibrillation (AF) is the most frequent persistent cardiac arrhythmia in the developed world and is associated with 10–20% of all ischemic strokes . AF is present in 1–2% of the general population and the prevalence increases with age, rising from ≤0.5% in individuals aged ≤55 years to >10% in those aged ≧85 years [2,3]. The incidence of AF is increasing, even after adjusting for the effects of an aging population .
Patients with AF have worse outcomes following ischemic stroke compared to those without AF [4,5,6,7,8]. The adverse effects of AF on stroke outcome have been chiefly linked to neurological factors including greater baseline stroke size and severity [4,5,6,7,8]. Although cardiac complications are a major source of morbidity and mortality following ischemic stroke  and AF is consistently associated with increased cardiovascular mortality in the general population [10,11], only one previous study has linked AF with increased cardiac mortality after stroke . Furthermore, there has been no previous detailed report of early cardiac morbidity in AF and stroke.
We therefore tested the hypothesis that increased early cardiac complications contribute to the adverse effect of AF on the outcome of acute ischemic stroke, independent of other baseline prognostic determinants, using serial clinical data from a select sample of acute ischemic stroke patients from the Virtual International Stroke Trials Archive (VISTA). This might have important implications for acute stroke management in those with AF, particularly regarding the need for more intensive clinical surveillance for cardiac complications and the requirement for routine cardiac monitoring.
The VISTA is an international collaborative repository for stroke clinical trial data from acute stroke, rehabilitation, secondary prevention and observational studies . These data are collated in a standardized format and accessible in an anonymized form for novel exploratory analyses .
To be eligible for inclusion in VISTA, acute stroke trials must have a minimum of 100 patients, documented entry criteria, documented consent or waiver of consent approved by a local institutional review board, baseline assessment using a validated neurologic impairment scale within 24 h of stroke onset, confirmation of stroke diagnosis by cerebral imaging within 7 days, outcome assessment using a validated neurologic or functional impairment scale between 1 and 6 months after stroke onset, and documented monitoring procedures to validate data .
After obtaining permission from the VISTA Steering Committee, we searched VISTA for patient data from the placebo arm of acute ischemic stroke trials containing the following prespecified variables: age, gender, cardiovascular risk factors, presence or absence of AF, baseline blood glucose level, baseline neurologic assessment using the National Institutes of Health Stroke Scale (NIHSS), cardiac adverse events (CAE) that occurred within 3 months of stroke onset and outcome assessment at 3 months after stroke onset using the modified Rankin Scale (mRS).
A history of hypertension, diabetes, hyperlipidemia, current or past cigarette use, a past history of myocardial infarction and a past history of transient ischemic attack were the prespecified baseline cardiovascular risk factors for the search. Non-fatal episodes of acute coronary syndromes (unstable angina and acute myocardial infarction), symptomatic heart failure, ventricular tachycardia, ventricular fibrillation, cardiopulmonary arrest, as well as cardiac deaths were grouped into the composite end point of serious cardiac adverse events (SCAE) . As the pooled databases used in VISTA were anonymized for trial source, the precise definition for AF, each baseline cardiovascular risk factor and each type of cardiac complication could not be established with certainty.
At the time of our search in August 2009, the acute stroke component of VISTA contained data for 28,131 patients from 30 randomized-controlled trials and one stroke registry. Our selection criteria were met by 2,865 patients from 6 randomized-controlled trials, of whom 819 (29%) were documented as having AF. A qualitative comparison of baseline characteristics, including age, gender, time from stroke onset to trial enrolment, baseline NIHSS score, AF, a history of hypertension, diabetes and myocardial infarction with data published by VISTA  showed that the select sample of patients was representative of all acute ischemic stroke patients in VISTA (online suppl. table 1; for all online supplementary material, see www.karger.com/doi/10.1159/000332028).
The following variables were compared between patients with and without AF that met our selection criteria: age, gender, cardiovascular risk factors, baseline glucose level, baseline NIHSS score, SCAE and mortality within 3 months of incident stroke, as well as NIHSS and mRS scores at 3 months after stroke onset. Furthermore, additional comparisons between patients with and without AF in those aged 75 years or older and those younger than 75 years were made, because age has consistently been shown to be independently associated with the incidence of AF and worse stroke outcomes [1,2,3,4,5,6,7,8], to determine whether the impact of AF on stroke outcome differed with age.
All statistical analyses were performed using SPSS (Version 19, Chicago, Ill., USA). The data were described as mean with standard deviation or median with interquartile range and analyzed using the unpaired t test with unequal variance or the Mann-Whitney U test depending on the underlying data distribution. Categorical and dichotomized variables were described as percentages and analyzed using the Fisher’s exact test. Unadjusted outcome effect sizes were estimated as differences in mean, median (Hodges-Lehmann nonparametric shift) or odds ratio (OR) with 95% confidence interval (CI) as appropriate.
Binary logistic regression modeling was applied to determine the effect of AF on the occurrence of SCAE, all cause mortality and poor functional outcome (defined as mRS of 3–6 at 3 months after stroke onset) adjusting for baseline imbalances in age (in yearly increments), gender, cardiovascular risk factors and stroke severity (for each point on the NIHSS). The application of the ordinal logistic regression modeling was also considered for the mRS end point but was not adopted because the mRS end point data did not adequately fulfil the proportional odds assumption.
All 2,865 patients that met the selection criteria in VISTA-Acute were enrolled into the six source trials within 24 h of stroke onset, with a median time of 4.5 h (interquartile range 3.7–5.8). Patients with AF were enrolled slightly earlier after stroke onset compared to patients without AF (4.3 vs. 4.5 h, p < 0.001).
At baseline (table 1), patients with AF were older, more likely to be female, more likely to be a non-smoker, had higher neurologic impairment and higher glucose level. A history of hypertension and a past history of myocardial infarction were also more common in patients with AF. At 3 months after stroke, patients with AF had greater neurologic impairment, worse functional outcome and higher all cause mortality compared to patients without AF (table 2).
|Table 1. Baseline characteristics|
|Table 2. Outcomes of patients with and without AF|
A total of 502 CAE occurred in 428 (15%) patients within 3 months of stroke onset. The median time to first CAE was 3 days in both patients with and without AF (table 2). SCAE in general occurred more frequently in patients with AF (14.2 vs. 6%, OR = 2.58, 95% CI 1.97–3.37). These especially applied to cardiac death, symptomatic heart failure, ventricular tachycardia and fibrillation. Patients with AF were also more likely to suffer from multiple CAE within 3 months of incident stroke.
Patients with AF had more frequent early cardiac complications and worse early stroke outcomes regardless of age. However, the absolute number of SCAE (11.3 vs. 6.5%, OR = 1.83, 95% CI 1.40–2.39) and the severity of neurologic impairment at 3 months following incident stroke (median NIHSS 9 vs. 4, p < 0.001) were higher in patients aged 75 years or older compared to patients younger than 75 years (online suppl. tables 2–4).
After adjusting for baseline differences in age, gender, cardiovascular risk factors, and stroke severity, AF remained independently associated with SCAE (OR = 2.14, 95% CI 1.61–2.86) and early mortality (OR = 1.44, 95% CI 1.14–1.81) but not poor functional outcome (OR = 0.85, 95% CI 0.68–1.07) within 3 months of stroke onset (tables 3, 4, 5).
|Table 3. Multivariate modeling assessing the effect of AF on SCAE|
|Table 4. Multivariate modeling assessing the effect of AF on 3-month stroke mortality|
|Table 5. Multivariate modeling assessing the effect of AF on poor functional outcome (mRS3–6) 3 months after incident stroke|
This is the first study to investigate the effect of AF on early mortality, disability and cardiac complications following ischemic stroke using serial clinical data from randomized-controlled trials that included both the NIHSS score and the mRS score. Our results confirm that early cardiac complications contribute to the worse ischemic stroke outcome in patients with AF, even after adjusting for baseline differences in age, cardiovascular risk factors and neurologic impairment. Furthermore, the occurrences of CAE were frequently early and potentially remediable with close clinical and electrocardiographic monitoring.
CAE have always been significant but frequently neglected contributors to early mortality following ischemic stroke . We found a cardiac mortality rate of 3.3% (19.4% of the overall mortality) in our cohort of 2,865 acute ischemic stroke patients within the first 3 months after stroke. Some recent acute ischemic stroke treatment trials reviewed by the American Heart Association reported a cardiac mortality rate of between 2–6% within 3 months after stroke .
The presence of AF has been shown to almost double the risk of cardiovascular mortality, with and without stroke, and triple the risk of developing heart failure in the general population [10,11]. Our results suggest that the presence of AF is associated with a similar increase in the odds of developing SCAE in the early clinical course following stroke. Furthermore, the finding that baseline stroke severity was an independent prognostic variable for SCAE is consistent with the previously reported association between insular cortical strokes and adverse cardiac complications that has been attributed to increased cerebrogenic autonomic and neurohumoral dysregulation .
At present, little evidence is available to guide the identification of CAE following ischemic stroke [17,18]. The American Stroke Association’s acute ischemic stroke guidelines recommend the use of continuous cardiac monitoring for the first 24 h following stroke . The Acute Ischemic Stroke guidelines of the European Stroke Organization also recommend the routine use of cardiac monitoring after an acute cerebrovascular ischemic event to screen for serious cardiac arrhythmias . Our finding that CAE occur in about one in four acute ischemic stroke patients with AF, with a median time to the occurrence of the first CAE of 3 days, provides strong support for the role of continuous cardiac monitoring. However, no study investigating the optimal duration and benefits of continuous cardiac monitoring following ischemic stroke has been published to date [17,18]. A planned trial of 5,000 patients has recently commenced its pilot phase in the UK [K.R. Lees, personal commun.]. Patients with previous undiagnosed AF following ischemic stroke should certainly receive more intensive cardiac evaluation, as three out of four patients with AF have underlying cardiac abnormalities .
The VISTA collaboration has previously investigated the identification of early serious cardiac complications following ischemic stroke through a detailed analysis of 800 placebo-treated patients from a randomized controlled acute ischemic stroke treatment trial . A past history of heart failure, concomitant diabetes mellitus, serum creatinine >115 µmol/l, increasing stroke severity and presence of prolonged corrected QT interval or ventricular extrasystoles on ECG were found to be the five most predictive baseline variables for the occurrence of SCAE. The authors postulated that the effect of AF on the occurrence of SCAE was explained by a history of heart failure and greater baseline stroke severity. Even when all five variables were absent, there was a 6% risk of developing a SCAE. Further research to guide the optimal method and duration of monitoring for cardiac complications following ischemic stroke is clearly needed.
Although both stroke and myocardial infarction are associated with substantial morbidity and mortality in previous studies [4,5,6,7,19,20], AF was independently associated with mortality but not disability after adjusting for baseline differences in age, cardiovascular risk factors and neurologic impairment in our study. This finding suggests that the effects on morbidity and mortality differ between neurologic and cardiac impairments in stroke patients. A similar observation was made in a previous randomized-controlled trial . The differential independent effect of AF on ischemic stroke mortality and morbidity also explains the conflicting reports from previous studies that were only able to quantitatively assess stroke mortality [4,5,6,7].
The main limitation of this study is the retrospective nature of the analysis. Although the VISTA collaboration has a standardized method for data collation, we have not been able to fully compensate for all heterogeneity. For instance, as mentioned in the methodology, it is likely that the precise definition of AF and different types of cardiac complications differed among the six source trials. This information cannot be accessed from the pooled databases used in VISTA, as data were anonymized for trial source [12,13]. In addition, the CHADS2 (congestive heart failure, hypertension, age >75 years, diabetes and ischemic stroke) score was recently shown to have a strong independent association with stroke outcome . However, we did not prespecify a past history of congestive heart failure in our data search, as it was not available in many published acute stroke trials [23,24]. Hence, we were unable to include the CHADS2 score in our analysis. Furthermore, information on the premorbid use of antiplatelet, anticoagulant, antihypertensive and lipid lowering agents was only available in a minority of patients. Therefore, we were unable to adjust for their use in our multivariate models.
In conclusion, AF is independently associated with frequent occurrence of early SCAE and increased all cause mortality following acute ischemic stroke. Physicians should consider more intensive cardiac monitoring and evaluation in acute ischemic stroke patients with AF, although further research is needed to determine their benefit.
The authors would like to thank the VISTA Steering Committee for providing access to the data. Dr Myzoon Ali from the VISTA Coordinating Centre in Glasgow provided assistance with raw data transfer and interpretation.
This study received funding from The Royal Melbourne Hospital Neuroscience Foundation.
H.T.H. Tu, B.C.V. Campbell, L. Churilov and J.M. Kalman report no relevant disclosures. K.R. Lees has received honoraria from Astellas, Asubio, Bayer, Boehringer Ingelheim, Johnson & Johnson, Lundbeck, Mitsubishi, Photothera, Sanofi-Aventis, Servier, Talecris, and Wyeth. P.D. Lyden receives royalties from sale of the book Thrombolytic Therapy for Acute Stroke. A. Shuaib has received honoraria and research grants from Astra Zeneca, Boehringer Ingelheim, Pfizer, Roche, Bayer, Coaxia, and Photothera. G.A. Donnan has served on advisory boards for Boehringer Ingelheim, Servier, Sanofi-Aventis and Bristol Myers Squibb. S.M. Davis has served on advisory Boards for Boehringer Ingelheim, Ever Pharma, Pfizer and Sanofi-Aventis.
Stephen M. Davis
Department of Neurology, The Royal Melbourne Hospital
Parkville, VIC 3050 (Australia)
Tel. +61 3 9342 8448, E-Mail Stephen.Davis@mh.org.au
The VISTA Steering Committee members are: A. Alexandrov, P.W. Bath, E. Bluhmki, L. Claesson, J. Curram, S.M. Davis, G.A. Donnan, H.C. Diener, M. Fisher, B. Gregson, J. Grotta, W. Hacke, M.G. Hennerici, M. Hommel, M. Kaste, K.R. Lees (Chair), P.D. Lyden, J. Marler, K. Muir, R. Sacco, A. Shuaib, P. Teal, N.G. Wahlgren, S. Warach, and C. Weimar.
Received: April 28, 2011
Accepted: August 2, 2011
Published online: October 14, 2011
Number of Print Pages : 7
Number of Figures : 0, Number of Tables : 5, Number of References : 24
Additional supplementary material is available online - Number of Parts : 1
Vol. 32, No. 5, Year 2011 (Cover Date: November 2011)
Journal Editor: Hennerici M.G. (Mannheim)
ISSN: 1015-9770 (Print), eISSN: 1421-9786 (Online)
For additional information: http://www.karger.com/CED