Journal Mobile Options
Table of Contents
Vol. 32, No. 4, 2011
Issue release date: October 2011
Cerebrovasc Dis 2011;32:370–382
(DOI:10.1159/000330637)

Outcomes for Patients with Ischaemic Stroke and Atrial Fibrillation: The PRISM Study (A Program of Research Informing Stroke Management)

Gattellari M.a, c · Goumas C.a · Aitken R.d · Worthington J.M.b, c, e
aSchool of Public Health and Community Medicine, and bSouth Western Sydney Clinical School, The University of New South Wales, Sydney, N.S.W., cIngham Institute, Liverpool, N.S.W., dAcademic and Global Relations Division, The University of Newcastle, Newcastle, N.S.W., and eDepartment of Neurophysiology, Liverpool Health Service, Liverpool, N.S.W., Australia
email Corresponding Author

Abstract

Background: In the past decade the prevalence of atrial fibrillation (AF) has been increasing in ageing populations while stroke prevention and management have advanced. To inform clinician practice, health service planning and further research, it is timely to reassess the burden of AF-related ischaemic stroke. Methods: We identified patients aged 18+ years with a primary or stay diagnosis of ischaemic stroke (ICD-10-AM I63.x), from July 1, 2000 to June 30, 2006, using an administrative health dataset of all hospitalisations in New South Wales (population ∼7 million). Fact of death was determined to December 2007. Results: Of the 26,960 index cases of ischaemic stroke, 25.4% had AF recorded during admission. Median age for AF and non-AF patients was 80.4 and 75.2 years, respectively (p < 0.001). Mortality was significantly higher in patients with AF at 30 days (19.4 vs. 11.5%), 90 days (27.7 vs. 15.8%) and 365 days (38.5 vs. 22.6%) (p values <0.0001). Adjusting for age and co-morbidities reduced these differences, with 90-day mortality of 20.9% in AF patients versus 14.7% in non-AF patients (p value <0.0001). The effect of AF on outcomes appears stronger in younger stroke patients relative to patients without AF (p valueinteraction <0.0001). At 30 days, the relative risk of mortality due to AF was 3.16 (95% CI 1.92–5.25) amongst those younger than 50, 1.71 (95% CI 1.32–2.22) in patients aged 50–64 years, 1.39 (95% CI 1.16–1.66) in patients aged 65–74 years, 1.29 (95% CI 1.17–1.43) in those aged 75–84 years, and 1.23 (95% CI 1.13–1.33) in those aged 85+ years. AF patients, surviving admission, spent a median of 19.2 days (95% CI 18.4–20.1) in hospital compared with 14.5 days (95% CI 13.9–15.1) for patients without AF (p < 0.001), with differences in length of stay greatest in younger patients (p valueinteraction <0.0001). 90-Day stroke survivors with AF spent an average of 21.5 days (95% CI 20.6–22.4) in hospital versus 16.6 days (95% CI 15.9–17.2) in those without AF. AF patients accessed more in-hospital rehabilitation (36.6%; 95% CI 35.0–38.2) than patients without AF (31.8%; 95% CI 31.0–32.7) (p value <0.0001), and differences in the proportion of AF versus non-AF patients accessing rehabilitation was greatest in younger patients (p valueinteraction <0.0006). Conclusions: Ischaemicstroke patients with AF have substantially worse outcomes than patients without AF, which can be partly explained by older age and greater co-morbidities. We have quantified the large effect of AF in younger patients and our results strongly argue for new antithrombotic research in young AF patients.


 Outline


 goto top of outline Key Words

  • Atrial fibrillation
  • Ischaemic stroke
  • Mortality
  • Population-based studies
  • Outcome
  • CHADS2 score
  • CHA2DS2-VASc score

 goto top of outline Abstract

Background: In the past decade the prevalence of atrial fibrillation (AF) has been increasing in ageing populations while stroke prevention and management have advanced. To inform clinician practice, health service planning and further research, it is timely to reassess the burden of AF-related ischaemic stroke. Methods: We identified patients aged 18+ years with a primary or stay diagnosis of ischaemic stroke (ICD-10-AM I63.x), from July 1, 2000 to June 30, 2006, using an administrative health dataset of all hospitalisations in New South Wales (population ∼7 million). Fact of death was determined to December 2007. Results: Of the 26,960 index cases of ischaemic stroke, 25.4% had AF recorded during admission. Median age for AF and non-AF patients was 80.4 and 75.2 years, respectively (p < 0.001). Mortality was significantly higher in patients with AF at 30 days (19.4 vs. 11.5%), 90 days (27.7 vs. 15.8%) and 365 days (38.5 vs. 22.6%) (p values <0.0001). Adjusting for age and co-morbidities reduced these differences, with 90-day mortality of 20.9% in AF patients versus 14.7% in non-AF patients (p value <0.0001). The effect of AF on outcomes appears stronger in younger stroke patients relative to patients without AF (p valueinteraction <0.0001). At 30 days, the relative risk of mortality due to AF was 3.16 (95% CI 1.92–5.25) amongst those younger than 50, 1.71 (95% CI 1.32–2.22) in patients aged 50–64 years, 1.39 (95% CI 1.16–1.66) in patients aged 65–74 years, 1.29 (95% CI 1.17–1.43) in those aged 75–84 years, and 1.23 (95% CI 1.13–1.33) in those aged 85+ years. AF patients, surviving admission, spent a median of 19.2 days (95% CI 18.4–20.1) in hospital compared with 14.5 days (95% CI 13.9–15.1) for patients without AF (p < 0.001), with differences in length of stay greatest in younger patients (p valueinteraction <0.0001). 90-Day stroke survivors with AF spent an average of 21.5 days (95% CI 20.6–22.4) in hospital versus 16.6 days (95% CI 15.9–17.2) in those without AF. AF patients accessed more in-hospital rehabilitation (36.6%; 95% CI 35.0–38.2) than patients without AF (31.8%; 95% CI 31.0–32.7) (p value <0.0001), and differences in the proportion of AF versus non-AF patients accessing rehabilitation was greatest in younger patients (p valueinteraction <0.0006). Conclusions: Ischaemicstroke patients with AF have substantially worse outcomes than patients without AF, which can be partly explained by older age and greater co-morbidities. We have quantified the large effect of AF in younger patients and our results strongly argue for new antithrombotic research in young AF patients.

Copyright © 2011 S. Karger AG, Basel


goto top of outline Introduction

Five percent of the population over the age of 65 and as many as 1/10 over the age of 75 are at an increased risk of stroke due to atrial fibrillation (AF) [1]. Population-based studies have reported higher mortality and greater disability when AF is associated with ischaemic stroke [2,3,4,5,6,7,8,9]. Around 20–35% of patients with ischaemic stroke have AF, and strokes caused by AF are more likely to be fatal and disabling than other strokes [2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22].

To our knowledge, only a few studies since 2000 have reported outcomes for patients with AF beyond their hospital stay [3,4,6,7]. In three, the number of AF stroke patients was relatively small, with cohorts of between 177 and 741 patients [3,4,6]. The one large study of over 24,000 strokes in patients with AF, from 1980 to 2002, included haemorrhagic strokes and did not compare outcomes with non-AF stroke patients [7]. The prevalence of AF is increasing in ageing communities [6,23] and, in the past decade, there have been important changes in stroke prevention and management, justifying the reassessment of AF-related stroke burden in modern cohorts. Previous studies have demonstrated poorer outcomes in AF patients when compared to non-AF patients [e.g.,3,4,5].

We analysed outcomes for patients with ischaemic stroke by AF status up to 1 year after hospitalisation and explored whether the effect of AF on outcomes according to patient age, sex and co-morbidities. Previous studies presenting data on AF mortality by age are limited in geographic scope [10,14,15,18,19] and either included too few patients with AF to reliably estimate the effect of AF according to age [14,15,18,19], did not extend the follow-up period beyond the early acute period [10], or did not formally test for effect modification [10,14,15,18,19]. Thus, it is difficult to speculate on any possible effect modification between age and AF using existing research. We hypothesised that AF patients would have worse outcomes than non-AF patients and that differences between AF and non-AF patients would be similar in patient subgroups characterised by age, sex and co-morbidities.

 

goto top of outline Methods

goto top of outline Databases

Hospital separations were obtained from the Admitted Patient Data Collection (APDC), a census of all hospitalisations in New South Wales (NSW), Australia’s most populous state of ∼7 million people. The APDC records diseases and procedures using ICD-10-AM coding [24]. A unique stroke hospitalisation can comprise several episodes of care, admission and separation dates, primary and stay diagnoses, co-morbidities, complications and transfers within and between hospitals. Fact of death was obtained from The NSW Registry of Birth, Deaths and Marriages recording all deaths of NSW residents occurring within NSW.

Ischaemic Stroke Hospitalisations
We selected patients aged 18 years or older separated from hospital with primary or stay diagnoses of acute ischaemic stroke (ICD-10 I63.x) from July 1, 2000 until June 30, 2006. We excluded non-residents of NSW and out-of-state transfers, selecting the first or ‘index’ ischaemic stroke admission within the study period. The admission date was defined as the date of the first acute episode of care, whether or not stroke was coded, ensuring capture of a patient’s first acute presentation to hospital before a stroke diagnosis was made. Episodes of care separated by up to 24 h were considered to comprise a continuous hospital admission [3].

Ascertainment of Atrial Fibrillation
We identified diagnoses of AF (ICD-10 code I48) recorded during the hospital admission.

Ascertainment of Co-Morbidities
We identified ICD-10 codes for relevant stroke risk factors (table 1) and categorised Charlson co-morbidities, an index previously used to predict inpatient mortality in an ICD-10-coded administrative health dataset [25]. We excluded cerebrovascular disease codes as redundant, in an approach reported to predict 1-year ischaemic stroke mortality in ICD-9-coded datasets [26].

TAB01
Table 1. Demographic characteristics of patients separated from hospital with a diagnosis of ischaemic stroke (I63.x), by presence of AF (n = 26,960)

goto top of outline Outcomes

Selected Stroke Signs and Complications
Four stroke signs and nine complications of stroke care were selected to reflect stroke severity [3].

All-Cause Mortality
TheNSW Registry of Birth, Deaths and Marriages provided fact of death up to December 10, 2007. Mortality was determined 30, 90 and 365 days after hospital admission with ischaemic stroke.

Proxy Indicators of Disability
Length of stay (LOS) was calculated from the date of the first acute episode of care until separation of the last acute or non-acute episode of care comprising the hospital admission. As elsewhere [3], episodes of care separated by up to 24 h were considered to comprise an admission. We calculated the days spent in hospital over a 90- and 365-day period, adapting a validated measure relating time spent at home to the modified Rankin score [27]. We restricted analysis to patients surviving to discharge, 90 or 365 days. From the subgroup surviving hospital admission, we calculated the proportion with a rehabilitation episode of care and number of in-hospital rehabilitation days.

Ascertainment of Pre-Admission Stroke and Bleeding Risk
We sought to classify AF patients according to their pre-admission stroke risk using recognised AF risk stratification schemes, namely the CHADS2 and the CHA2DS2-VASc scores [28,29]. In the presence of AF, the CHADS2 score ascribes 1 point for congestive heart failure (CHF), hypertension and diabetes and 2 points for a prior stroke or TIA to produce a score ranging from 0 to 6. The CHA2DS2-VASc has been shown to more accurately define stroke risk modifying the CHADS2 score by adding 1 point each for vascular disease and female gender and 2 points for an age of 75+ years [29]. Our aim was to determine the percentage of AF patients whose pre-admission risk scores would have strongly predicted a clinically important stroke risk before the ischaemic stroke event. A score of 2 or more derived from either scheme predicts a ‘high’ annual stroke risk indicating oral anticoagulation. Aspirin or no antithrombotic treatment is recommended in patients at low risk of stroke, specifically in those with a CHA2DS2-VASc score of 0 [28,29].

We included an assessment of bleeding risk, modifying the HAS-BLED score [30] for comparative purposes. The HAS-BLED score includes hypertension, abnormal renal or liver function, prior stroke, a history of major bleeding or conditions that predispose patients to bleeding (e.g. anaemia), age 65+ years and a history of alcohol abuse. We excluded co-morbidities that could not be ascertained using administrative health data, namely labile international normalized ratio or concomitant medications that increase bleeding risk with anticoagulation. Our modified score therefore describes a ‘baseline risk’ when patients are not receiving anticoagulation. Patients with CHADS2 and CHA2DS2-VASc scores of 2 or more have a stroke risk that would outweigh the risk of major bleeding irrespective of HAS-BLED scores [31]. A HAS-BLED score of 3 or more is said to indicate a heightened bleeding risk requiring ‘caution and regular review of the patient’ [31].

ICD-10-AM codes for co-morbidities were identified from all recorded hospitalisations preceding a patient’s ischaemic stroke admission. For these descriptive analyses, the cohort was restricted to AF patients presenting to hospital with an ischaemic stroke in the last year of our dataset, maximising co-morbidity ascertainment with a 5-year ‘look-back’ through hospitalisation data.

goto top of outline Statistical Analysis

Bivariate associations between AF status and patient characteristics were assessed using χ2 tests for categorical outcomes and non-parametric Mann-Whitney U tests for age. As conventional Cox regression survival analysis did not meet the proportional hazards assumption and log-binomial models for relative risk did not always converge, adjusted rates of mortality and relative risk of death at 30, 90 and 365 days, and adjusted rates of stroke symptoms and complications were estimated using Cox regression models, assigning constant time at risk to all patients. This method has been shown to produce correct relative risk estimates and confidence intervals and is preferable to estimating odds ratios using logistic regression, as odds ratios overestimate relative risks when outcomes are common [32]. Adjusted LOS and number of days spent in hospital during the first 90 and 365 days after ischaemic stroke were estimated using linear regression models. A log transformation was performed as these outcomes were positively skewed. Adjusted geometric means were obtained by exponentiating the mean estimates. A Charlson co-morbidity propensity score for each outcome of interest was constructed by fitting a multivariate regression model including an indicator variable for each Charlson co-morbidity. Predicted values for each patient were used as their propensity scores, which were modelled as continuous variables. For mortality outcomes, increasing propensity scores indicated lower risk of mortality. For days spent in hospital, increasing propensity scores indicated a greater number of days in hospital. We did not adjust for propensity scores in our analysis of stroke symptoms and complications as some of these codes were included in Charlson co-morbidity codes.

The potential modifying effect of age, sex and Charlson co-morbidity on the association between AF and each outcome was tested by adding, in turn, two-way interaction terms to the models. Significant (p < 0.10) interactions were further explored by stratifying analyses by broad categories of the effect modifier.

All multivariate analyses were adjusted for age as a continuous variable (both linear and quadratic terms), sex, year of separation and co-morbidity. We adjusted for marital status when considering LOS, discharge destination and exposure to in-hospital rehabilitation, as the presence of a partner may influence discharge destination [33,34]. Standard errors were adjusted for the clustering effect of the hospital recording the first acute episode of care. Descriptive and multivariate analyses were undertaken using SPSS version 18.0.2 [35] and SAS version 9.1.3 [36].

 

goto top of outline Results

goto top of outline Patient Sample

There were 29,464 acute ischaemic stroke hospitalisations from July 1, 2000 to June 30, 2006 in NSW and 719 (2.4%) hospitalisations were excluded due to ineligible age (n = 7), residency outside NSW (n = 540) or out-of-state transfer (n = 160). Twelve hospitalisations were removed for suspected data linkage errors. We identified 26,960 ‘index’ cases from the remaining 28,745 eligible hospitalisations.

Characteristics
AF was recorded in 6,842 (25.4%) ‘index’ cases. The median age for AF patients was 80.4 years, compared with 75.2 years for those without AF (p < 0.001) (table 1). Females were more likely to have AF than males (29.2 vs. 21.8%) (p < 0.001) and women with AF were significantly older than men with AF (median 82.2 vs. 78.0 years) (p < 0.001).

Only 4.0 and 7.3% of stroke patients aged younger than 50 or between 50 and 64 years of age, respectively, had AF. 1/5 patients (22.1%) aged 65–74 years had AF, almost 1/3 (29.9%) aged 75–84 years and more than 35% (37.3%) over 85 years.

Co-Morbidities
Myocardial infarction, CHF, hypertension, pulmonary disease, peptic ulcer, cognitive impairment and renal disease were significantly more likely to be recorded in AF patients (p < 0.001; table 1). In contrast, smoking and hypercholesterolaemia were significantly less prevalent in patients with AF. Diabetes was noted in 21.4 and 22.8% of AF and non-AF patients, respectively (p = 0.02).

1/5 (20.7%) AF patients younger than 50 years had hypertension compared with around 60% over 50 years. CHF was rare in young AF patients (3.4%, <50 years) while almost 10% of those aged 50–64 years had CHF, and prevalence increased further with age. Five percent of young AF patients (<50 years) had two or more other major stroke risk factors compared with around 25% aged 50–74 years and 75% of those over 75 (fig. 1).

FIG01
Fig. 1. AF patients and prevalence of stroke risk factors (n = 6,842). a Including being over 75 years of age.

goto top of outline Outcomes

Stroke Symptoms and Signs
After adjusting for age, sex and year, two-thirds (67.9%) of patients with AF had a recorded diagnosis of hemiplegia compared with 59.9% without AF (table 2). Speech, visual disturbances and dysphagia were significantly more prevalent in AF-related strokes (p < 0.001).

TAB02
Table 2. Stroke symptoms and indicators of a poor prognosis by AF status

Complications
Patients with AF were significantly more likely to be in a coma, receive intensive care or have mechanical ventilation (adjusted proportions: 11.6 vs. 5.9%) (p < 0.001) (table 2). AF strokes had a significantly higher prevalence of pneumonia (14.6 vs. 8.4%), sepsis (3.6 vs. 1.9%), deep vein thrombosis (1.9 vs. 1.4%), decubitus ulcer (4.0 vs. 2.3%), urinary incontinence (10.5 vs. 7.5%) and urinary tract infection (14.7 vs. 11.4%) (p < 0.001). Over 40% (43.1%) of patients with AF experienced at least one of these complications during hospitalisation, compared with 30.8% of those without AF (p < 0.001).

Mortality
A significantly greater proportion of patients with AF died, indicating a main effect of AF on mortality. For example, almost 40% of patients with AF (38.5%) died by 1 year compared with 22.6% of those without AF. Within 30 days, 19.4% of patients with AF died compared with 11.5% without AF (p < 0.001). As patients with AF were significantly older and had greater co-morbidities, adjusting mortality rates for age and co-morbidities reduced mortality differences between patients with and without AF (table 3). For example, at 90 days, the adjusted mortality rate of AF patients was 20.9% compared with the unadjusted mortality rate of 27.7%; in non-AF patients the adjusted and unadjusted mortality rates were similar (15.8 vs. 14.7%).

TAB03
Table 3. Mortality by AF status

In general, increasing age was associated with an increasing risk of mortality in both AF and non-AF stroke at 30, 90 and 365 days (table 3), indicating a main effect of age on mortality. For example, by 365 days, around 1/5 patients with AF younger than 50 and between 50 and 64 years of age had died, compared with 1/4 AF patients aged 65–74 years. In those aged 75–84 years, around 40% had died by 12 months, and over half (55.1%) aged 85 years or older had died. In patients without AF, around 1/10 aged younger than 50 years and between 50 and 64 years had died within 365 days, while 1/5 and 1/4 patients aged 65–74 and 75–84 years, respectively, had died. Around 40% of those aged 85 years or older without AF had died within 12 months.

There was a significant interaction between AF status and age (p valuesinteraction 0.002 to <0.0001). In younger patients, AF appeared to have a stronger relative influence on mortality when compared with older patients. For example, within 30 days, the adjusted mortality risk in patients younger than 50 years of age with AF was 11.8% compared with an adjusted mortality rate of 3.7% in patients without AF. This indicates that patients younger than 50 years of age with AF were around 3 times more likely to die than similarly aged patients without AF. In general, the relative effect of AF on mortality diminished with increasing age. In those patients aged 50–64, patients with AF had almost double the risk of dying within 30 days than those without AF (RR = 1.71), while patients aged 65–74 were around 1.4 times more likely to die if they had AF. In patients aged 75–84 years, those with AF were 1.3 times more likely to die than those without AF. In patients over 85 years of age, 26.7% of those with AF had died within 30 days compared with 21.8% of those without AF, yielding an adjusted relative risk of AF on mortality of around 1.2. This effect of decreasing relative risk with increasing age with AF was also observed at 90 and 365 days (table 3). Overall, increasing age was associated with increasing mortality; the interaction effect demonstrates that differences in mortality between AF and non-AF stroke decreased with increasing age.

The effect of AF on mortality at 30, 90 and 365 days was also modified by co-morbidity scores (table 4) (p valuesinteraction <0.0001). Although AF was associated with an elevated risk of mortality across all Charlson co-morbidity scores, the effect of AF appeared weaker in patients where the Charlson score predicted high mortality.

TAB04
Table 4. Adjusted mortality by weighted Charlson score

The effect of AF on mortality did not vary significantly between men and women at any stage of follow-up (p > 0.1).

LOS, Rehabilitation and Time Spent in Hospital
In patients discharged alive, those with AF had a longer median LOS of 19.2 days, compared with 14.5 days in patients without AF (p < 0.001). In the 3 months since their stroke, 90-day survivors with AF spent an average of 21.5 days in hospital, compared with 16.6 days (p < 0.001) in those without. One-year survivors, with or without AF, spent an average of 28.8 and 22.5 days in hospital, respectively (p < 0.0001). The effect of AF on days in hospital was significantly modified by age and co-morbidities (pinteraction 0.03 to <0.0001). In general, differences between AF and non-AF patients became smaller with increasing patient age and increasing co-morbidity scores (table 5).

TAB05
Table 5. Adjusted mean days spent in hospital, by ischaemic stroke admission, in the first 90 days and first 365 days from admission date (survivors)

In patients surviving admission, a higher proportion with AF had a rehabilitation episode of care following their acute ischaemic stroke episode of care (44.1 vs. 33.4%) (p < 0.001). The difference remained significant in the multivariate model (36.6 vs. 31.8%) (p < 0.0001) and was modified by age (p = 0.0006) and Charlson propensity scores (p < 0.0001) (table 5). In patients undertaking rehabilitation, the adjusted number of days spent in rehabilitation did not significantly differ between AF and non-AF patients (18.4 vs. 18.5 days, respectively) (p = 0.82) and was not modified by age or propensity scores (p > 0.1). The effect of AF was not significantly modified by gender for any of these outcomes (p > 0.1).

Ascertainment of Pre-Admission Stroke and Bleeding Risk
There were 1,293 AF patients in whom we ascertained pre-admission stroke and bleeding risk estimates using all available co-morbidity information in the 5 years preceding the ischaemic stroke admission. Overall, 1,217 or 94.1% had a CHA2DS2-VASc score of 2 or higher, indicating a high stroke risk, while significantly fewer (n = 957, 74.0%) (p < 0.0001) had CHADS2 scores of 2 or more (fig. 2).

FIG02
Fig. 2. Distribution of pre-admission stroke risk and bleeding risk scores in AF patients.

When comparing stroke risk scores by age, patients younger than 65 years were pooled owing to smaller numbers in the last year of our dataset. Across all age groups, CHA2DS2-VASc scores were significantly higher than CHADS2 scores (p < 0.0001) indicating that CHA2DS2-VASc scores ascribed patients a higher stroke risk, irrespective of age.

One half of AF-related stroke patients younger than 65 years had a CHA2DS2-VASc score of 2 or higher (51.3%) and 39.8% had a CHADS2 score of 2 or more. Applying the CHA2DS2-VASc scheme, almost all patients (91.3%) aged 65–74 years and all patients aged 75–84 and over 85 years had a score of 2 or more; using the CHADS2 scheme, only 42.1, 85.4 and 87.5%, respectively, had a score of 2 or more.

Around 1/5 patients with AF-related stroke younger than 65 years (18.6%) had a CHA2DS2-VASc score of 0, implying a ‘low risk’ of stroke. No patient over the age of 65 was considered ‘low-risk’ according to the CHA2DS2-VASc score. In contrast, using the CHADS2 score, a higher percentage of AF-related stroke patients under 65 (27.4%) and a sizeable proportion of patients aged 65–74 (20.0%) years were classified as ‘low risk’.

Overall, 25.7% of patients (n = 325) had a modified HAS-BLED score of 3 or higher (fig. 2). Only 3.5% of patients younger than 65 years had HAS-BLED scores of 3 or more, while 22.1, 29.7 and 28.6% of patients aged 65–74, 75–84 and 85+ years, respectively, had HAS-BLED scores equal to or exceeding 3. Overall, only 2.1% of patients had HAS-BLED scores that exceeded their CHA2DS2-VASc scores.

 

goto top of outline Discussion

This is the first Australian study of AF-related stroke outcomes and one of very few studies to assess AF-related stroke burden since major advances in stroke prevention and management. The strengths of the PRISM study are its comprehensive and representative capture of a large, consecutive population of ischaemic stroke patients over an entire health service with 1-year follow-up available for all patients. With over 6,000 AF patients, this is the first study to our knowledge to demonstrate that the effect of AF, on mortality and disability, varies by age and co-morbidities.

We found that over 1/4 hospitalized ischaemic strokes are AF-related and that the prevalence of AF increases with age. Mortality was significantly higher in AF-related than non-AF-related strokes, confirming the poor outcomes in this most common cause of cardioembolic stroke [3,4,6,8,9,10,11,12,13,14,18,19,20]. Increasing age was associated with increasing mortality in non-AF- and AF-related ischaemic stroke.

Almost 40% of patients with AF-related strokes died in the first year and 20% in the first 30 days, which is comparable to previous smaller studies [3,5]. Patients with AF were more likely than patients without AF to have documented hemiplegia, visual and speech disturbances and dysphagia as well as complications, such as pneumonia and venous thromboembolism [3,9,13]. AF-related stroke survivors spent substantially more time in hospital than non-AF patients.

Adjusting for age and co-morbidities reduced the effect of AF on outcomes. Patients with AF were significantly older than patients without AF and, in this and other studies, increasing age was associated with poorer outcomes in AF and non-AF ischaemic stroke. The poor prognosis in AF-related stroke patients can be partly attributed to increased age and to commonly associated co-morbidities [3].

Consistent with some [9,20], but not all studies [3,4,6,8,9,10,11,12,15], there were higher rates of recorded hypertension in our AF-related strokes. There were higher rates of cardiac disease [6,8,9,11,12,15,20], lower rates of smoking [3,4,8,9,10,11,20] and hypercholesterolaemia [3,4,8,10] and a similar rate of diabetes [3,6,8,12,15,20]. Recognising a distinct profile of co-morbidities which influence AF-related stroke risk and post-stroke prognosis highlights the importance of multifaceted stroke prevention.

goto top of outline AF-Related Stroke Outcomes by Age and Co-Morbidity

The PRISM study has a large cohort with extended follow-up, exploring stroke outcomes up to 1 year, according to AF status and age. More than 500 strokes were younger than 65 years of age and 1,960 were older than 85. Under the age of 65, fewer than 10% of ischaemic strokes were associated with AF, compared with almost 40% over 85 years. Importantly, AF was most strongly predictive of mortality and time spent in hospital in younger patients under 65 years.

The effect of AF and increasing age were independently associated with poorer outcomes with absolute mortality rates higher in AF than non-AF patients. However, we have shown that the relative effect of AF on mortality varies with patient age. Specifically, AF had a stronger effect on mortality in younger patients than in older patients. The adjusted relative risk of death in AF and non-AF at 30 days was 3.16 in those under 50, falling to 1.23 for those over 85 years. While absolute mortality rates were greater in older patients, the difference in mortality between AF and non-AF patients is greatest in younger stroke patients.

A previous study of 3,335 AF patients, including 600 patients younger than 65 and 500 patients older than 85 [10], was limited to 28-day outcomes and did not attempt to assess effect modification by age or extend follow-up beyond the acute stroke phase. Using their results, without adjustment for co-morbidity or age within broad age-bands, our analysis of their raw data also shows a higher relative risk estimate in younger patients.

We found four other studies presenting data on AF-related mortality according to age [14,15,18,19]; these studies did not test for an interaction effect between AF and age and all are now distant in time, limited in geographic scope and too small to allow meaningful conclusions. One Oxfordshire (UK) study, with follow-up to 1988, included 115 patients with AF. Only 21 were younger than 70 years and only 44 were older than 80 years [14]. In the Framingham Study there was an analysis of 103 AF patients with stroke, recruited from 1948 and followed to 1991 [18]. There were only 13 patients younger than 65 years of age and 59 older than 75 years. Two other studies in single centres between 1977–1981 and 1987–1989 included only 78 and 82 AF patients, respectively [15,19].

Several explanations may account for the effect modification between AF and age. Frailty in the elderly may greatly increase mortality and disability in non-AF-related strokes and may result in less provision of acute supportive care and rehabilitation. A previous Australian study of 395 stroke patients suggests that older patients are less likely to receive care in a stroke unit, where more intense investigations and treatments are used [37]. If older patients are less likely to reach a stroke unit, and receive less comprehensive diagnostic work-up, then AF may be underdiagnosed and the calculated effect of AF in older patients would be weakened through misclassification. This hypothetical limitation is based on a single small local study [37] and would not affect our conclusion that AF has a strong impact on mortality in younger strokes.

We cannot comment on anticoagulant rates with this data and anticoagulant use at the time of AF-related ischaemic stroke is known to lower risk of death and disability [3,5]. By guidelines, younger patients under 65 years are generally less likely to be anticoagulated and with lower rates of anticoagulation, younger stroke patients with AF may have relatively worse outcomes.

goto top of outline Pre-Stroke Risk in AF-Related Stroke

Treatment with antithrombotics is often determined according to stroke risk calculated using risk scores such as CHADS2 and CHA2DS2-VASc, which differ in detail. Oral anticoagulation is generally recommended in the presence of one ‘major’ risk factor, specifically prior stroke, TIA or systemic embolism or older age (75+ years) (CHA2DS2VASC) or in the presence of two ‘clinically relevant non-major risk factors’, such as heart failure and hypertension. Either aspirin or anticoagulation is recommended in the presence of one ‘non-major’ risk factor. Aspirin or no antithrombotic treatment is recommended in patients at low risk of stroke, specifically in those with a CHA2DS2-VASc score of 0 [38].

We carried out a descriptive analysis to demonstrate the pre-admission stroke risk of AF patients according to CHA2DS2-VASc and CHADS2 scores. Our results demonstrate that the majority of patients would have been identified as having a ‘high risk’ of stroke according to either scheme. However, the two scores did not perform equally. The CHA2DS2-VASc score would have categorised a greater number of our patients as ‘high-risk’ than CHADS2.

In AF-related strokes under age 65, 20 and 30% of patients destined to have an ischaemic stroke would have been classified as having a ‘low’ stroke risk, according to the CHA2DS2-VASc and CHADS2 scores, respectively. This result suggests that younger patients are more likely to be misclassified as having a low stroke risk, underscoring the limitations of existing risk stratification schemes in younger AF patients.

In the overwhelming majority of cases, bleeding risk scores were exceeded by stroke risk scores. While our modified HAS-BLED score would have underestimated bleeding risk, the majority of patients had high stroke risk scores that would likely outweigh bleeding risk. This suggests that anticoagulation is justified in the majority of AF patients who eventually have an ischaemic stroke.

goto top of outline ‘Young’ AF-Related Stroke

While the population benefit of preventing AF-related stroke in younger people is likely to be small, individual and community benefits may be considerable with large potential benefits in quality-of-life-years in these working-aged people. The prevalence of AF in patients younger than 65 years is low (<1%) [1], yet AF was associated with 9.8% of all strokes under 65, and accounted for 19% of all 30-day stroke deaths under 65, with a crude 30-day mortality of 12%. In those surviving 90 days, AF patients under 65 years had as many hospital days, or more, than older patients with AF. Young AF stroke survivors, compared with non-AF patients, spent relatively more days in hospital, our proxy indicator of disability, adapted from a validated measure [27].

For patients younger than 65, there is considerable evidence for the best management of cardiovascular risks such as diabetes and hypertension. There is little evidence to clearly guide best antithrombotic management in young patients with non-valvular AF. Our estimates of stroke and bleeding risk, identifiable before ischaemic stroke admission, suggest that a review of current risk-benefit assessment is needed. In the absence of one ‘major’ or two ‘non-major’ stroke risk factors, the stroke risk is deemed ‘low’ and aspirin may be preferred [38] as the bleeding risk is perceived to outweigh the benefits of anticoagulation. Using these recognised criteria, more than 1/5 younger AF stroke patients were estimated to have a ‘low’ stroke risk before their eventual ischaemic stroke. The estimated bleeding risk was low in the cohort of young AF-related stroke patients. Existing risk-benefit estimates may need review, to ensure we maximize stroke prevention in young patients with AF.

Warfarin has been shown to be very effective in older patients with AF, including ‘low risk’ patients where anticoagulation was not previously considered [39]. Newer anticoagulants may offer potential quality of life and risk benefits for younger patients [40] and aspirin and clopidogrel, in combination, may be useful in younger AF patients [41]. Considering recent evidence, emerging anticoagulants and our own findings, there is a strong case for reassessment of risk stratification and randomised control trials of new and existing antithrombotics in young AF patients.

goto top of outline Limitations

The ICD-10 code for AF (I48) includes atrial flutter and inclusion of this less common and possibly less serious diagnosis, may have underestimated the morbidity of AF [42]. Although, administrative datasets do not provide disability measures we adapted a validated measure of disability, using the days spent in hospital as a proxy indicator [27].

A failure to recognise a past history of AF or to detect elusive paroxysmal AF is a systematic problem across all studies and a problem in clinical practice. Underascertainment, particularly of elusive paroxysmal AF, on history or testing, would underestimate the effect of AF on outcomes. However, we note that the prevalence of AF in our study is comparable with that reported in other studies using different methodologies [e.g. [3,4,8,10,11,12,21] ].

Coding standards prescribe that co-morbidities affecting patient care are recorded on administrative datasets and the coders use clinician notes to identify relevant diagnoses such as AF. Strokes are assigned an emergency department triage category where an initial 12-lead electrocardiograph (ECG) and cardiac monitoring are expected care. Specific information on ECGs or cardiac event monitoring is not available on the dataset but AF recorded in the clinical notes, from history-taking, observations and clinical examinations as well as reports of ECGs or cardiac monitoring are the basis for coding. If AF has been detected on history or during the hospital admission, we would expect it to be recorded in the dataset.

Sources of error in administrative health data may include a lack of standardisation of coding over time and between hospitals and undernumeration of co-morbidities [43,44]. Australian datasets are a basis for funding and quality assurance is closely administered with audits and the coders undergo standardised training. Positive predictive values of principal diagnoses are high (>90%) in Australian studies [45,46]. Although ischaemic stroke coding has not been specifically validated in Australia, one recent study reported that hospital discharge primary diagnoses of stroke were correct in 38 out of 40 (95%) of records, validated against blinded medical record review by a stroke clinician [46]. Internationally, principal and secondary diagnoses of ischaemic stroke have acceptable levels of sensitivity, specificity and positive predictive values (>85%) [47].

The false-positive rate of co-morbidities is generally low (<1%) [44], and reliability of co-morbidity, procedure and cerebrovascular disease coding has been found to be high in an Australian hospitalisation dataset [48]. Our results confirm the utility of the Charlson co-morbidity index derived from administrative health data [25,26]. When we attempted to supplement co-morbidity codes with hypertension, smoking and hyperlipidaemia, they were associated with paradoxical improvement in survival and this has been noted in other analyses of administrative data [49]. The management and documentation of these conditions may be greater in those surviving the early days of stroke and be underrecorded in those with early death.

These study outcomes can only be generalised to hospitalised ischaemic stroke cases and our results cannot be used to infer incidence or stroke attack rates in our population. We aimed for a high level of specificity in patient selection to ensure valid and accurate determination of ischaemic stroke outcomes and therefore did not include unspecified stroke (I64). We note thereis an overuse of the unspecified stroke code in administrative datasets, and outcomes for unspecified stroke differ from coded subtypes [50]. The composition of unspecified stroke may alter with increasing use and quality of imaging [21,50]. While excluding unspecified strokes underestimates ischaemic stroke attack rates, their inclusion would capture stroke mimics and haemorrhagic strokes, reducing both specificity for ischaemic stroke and accuracy of outcome estimates.

 

goto top of outline Conclusion

Our results demonstrate that AF-related stroke accounts for at least 25% of ischaemic stroke and is still associated with high mortality and disability in modern-day care. Despite availability of thrombolysis, stroke units and other modern care, AF still accounts for a large number and a large proportion of fatal and disabling stroke and makes up a large part of the community burden of stroke. Our observations highlight the ongoing consequences of the wide evidence-practice gap seen in AF-related stroke prevention, which does not seem to go away [5,51].

Some of the poor outcomes in AF patients are explained by associated co-morbidities and greater age. Stroke patients with AF appear to have a distinct co-morbidity profile which predicts both stroke risk and poor outcomes and emphasises the importance of a multifaceted risk reduction strategy in AF patients.

Existing stroke risk stratification appears to work less well in younger AF patients, where many with stroke would have been classified as low-risk. Importantly, the effect of AF status was strongest in younger patients where effective prevention may save many quality-of-life years and where anticoagulation use may be relatively safe due to low bleeding risk. Our findings make a strong case for better stroke prevention in AF and urgent research on risk stratification and antithrombotic treatment in younger AF patients. The PRISM study provides information for clinical decision-making and health service planning and establishes a baseline to assess future stroke outcomes, using a method that is cost-effective for the study of large cohorts.

 

goto top of outline Acknowledgements

The authors express their gratitude to the CHeReL, Department of Health and NSW Registry of Births, Deaths and Marriages for data linkage services and data. We especially thank Ms. Katie Irvine, Ms. Kim Lim and Dr. Lee Taylor for their generous assistance. We extend our warmest thanks to Ms. Jasminka Muratidi, Manager, Clinical Coding Department, Liverpool Health Service, for her invaluable insights into stroke coding. Professor Chris Levi alerted us to the Quinn et al. [27] article validating ‘time at home’ as a measure of disability. This study received ethics approval from the New South Wales Population Health Services Research Ethics Committee (No. 2007/03/023). M.G. is funded by a Commonwealth Department of Health and Ageing PHCRED Senior Research Fellowship. The University of New South Wales provided funding for data linkage.

 

goto top of outline Disclosure Statement

None.


 goto top of outline References
  1. Wolf PA, Abbot RD, Kannel WB: Atrial fibrillation as an independent risk factor for stroke: the Framingham Study. Stroke 1991;22:983–988.
  2. Frost L, Andersen LV, Johnsen SP, Mortensen LS: Lost life-years attributable to stroke among patients with nonvalvular atrial fibrillation: a nationwide population-based follow-up study. Neuroepidemiology 2007;29:59–65.
  3. Thygesen SK, Frost L, Eagle KA, Johnsen SP: Atrial fibrillation in patients with ischemic stroke: a population-based study. Clin Epidemiol 2009;1:55–65.

    External Resources

  4. Hannon N, Sheehan O, Kelly L, Marnane M, Merwick A, Moore A, Kyne L, Duggan J, Moroney J, McCormack PME, Daly L, Fitz-Simon N, Harris D, Horgan G, Williams EB, Furie PL, Kelly PJ: Stroke associated with atrial fibrillation – incidence and early outcomes in the North Dublin Population Stroke Study. Cerebrovasc Dis 2010;29:43–49.
  5. Gladstone DJ, Bui E, Fang J, Laupacis A, Lindsay MP, Tu JV, Silver FL, Kapral MK: Potentially preventable strokes in high-risk patients with atrial fibrillation who are not adequately anticoagulated. Stroke 2009;40:235–240.
  6. Béjot Y, Ben Salem D, Osseby GV, Durier J, Marie C, Cottin Y, Moreau T, Giroud M: Epidemiology of ischaemic stroke from atrial fibrillation in Dijon, France, from 1985 to 2006. Neurology 2009;72:346–353.
  7. Frost L, Vukelic-Andersen L, Vestergaard P, Husted S. Mortensen LS: Trend in mortality after stroke with atrial fibrillation. Am J Med 2007;120:47–53.
  8. Marini C, De Santis F, Sacco S, Russo T, Olivieri L, Totaro R, Carolei A: Contribution of atrial fibrillation to incidence and outcome of ischaemic stroke: results from a population-based study. Stroke 2005;36:1115–1119.
  9. Lamassa M, Di Carlo A, Pracucci G, Basile AM, Trefoloni G, Vanni P, Spolveri S, Baruffi MC, Landini G, Ghetti A, Wolfe CDA, Inzitari D: Characteristics, outcome and care of stroke associated with atrial fibrillation in Europe: data from a multicentre multinational hospital-based stroke registry (The European Community Stroke Project). Stroke 2001;32:392–398.
  10. Kimura K, Minematsu K, Yamaguchi T: Atrial fibrillation as a predictive factor for severe stroke and early death in 15,831 patients with acute ischaemic stroke. J Neurol Neurosurg Psychiatry 2005:76:679–683.
  11. Steger C, Pratter A, Marinek-Bregel M, Avanzini M, Valentin A, Slany J, Stollberger C: Stroke patients with atrial fibrillation have a worse prognosis than patients without: data from the Austrian Stroke Registry. Eur Heart J 2004;25:1734–1740.
  12. Karataş M, Dilek A, Erkan H, Yavuz N, Sözay S, Akman N: Functional outcome in stroke patients with atrial fibrillation. Arch Phys Med Rehabil 2000;81:1025–1029.
  13. Miyasaka Y, Barnes ME, Gersh BJ, Cha SS, Seward JM, Bailey KR, Iwasaka Y, Tsang TSM: Time trends of ischemic stroke incidence and mortality in patients diagnosed with first atrial fibrillation in 1980 to 2000: report of a community-based study. Stroke 2005;36:2362–2366.
  14. Sandercock P, Banford J, Dennis M, Burn J, Slattery J, Jones L, Boonyakarnkul S, Warlow C: Atrial fibrillation and stroke: prevalence in different types of stroke and influence on early and long term prognosis (Oxfordshire Community Stroke Project). BMJ 1992;305:1460–1465.
  15. Friedman PJ: Atrial fibrillation after stroke in the elderly. Stroke 1991;22:209–214.
  16. Jørgensen HS, Nakayam H, Reith K, Raaschou HO, Olsen TS: Acute stroke with atrial fibrillation: The Copenhagen Stroke Study. Stroke 1996;27:1765–1769.
  17. Kaarisalo MM, Immonen-Raiha P, Marttila RJ, Salomaa V, Kaarsalo E, Salmi K, Sarti C, Sivenius J, Torppa J, Tuomilehto K: Atrial fibrillation and stroke: mortality and causes of death after the first acute ischaemic stroke. Stroke 1997;28:311–315.
  18. Lin HJ, Wolf PA, Kelly-Hayes M, Beiser AS, Kase CS, Benjamin EJ, D’Agostino RB: Stroke severity in atrial fibrillation: the Framingham Study. Stroke 1996;27:1760–1764.
  19. Lowe GD, Jaap AJ: Forbes CD. Relation of atrial fibrillation and high haematocrit to mortality in acute stroke. Lancet 1983;321:784–786.

    External Resources

  20. Dulli DA, Stanko H, Levine RL: Atrial fibrillation is associated with severe acute ischemic stroke. Neuroepidemiology 2003;22:118–123.
  21. Cadilhac DA, Pearce DC, Levi CR, Donnan GA: Improvements in the quality of care and health outcomes with new stroke care units following implementation of a clinician-led, health system redesign programme in New South Wales, Australia. Qual Saf Health Care 2008;17:329–333.
  22. Gattellari M, Worthington J, Jalaludin B, Mohsin M: Stroke unit care in a real-life setting: can results from randomized controlled trial be translated into every-day clinical practice? An observational study of hospital data in a large Australian population. Stroke 2009;40:10–17.
  23. Frost L, Vukelic-Andersen L, Vestergaard P, Husted S, Mortensen LS: Trends in risk of stroke in patients with a hospital diagnosis of nonvalvular atrial fibrillation: National Cohort Study in Denmark, 1980–2002. Neuroepidemiology 2006;26:212–219.
  24. ICD 10th Revision, Australian Modification: Commonwealth of Australia, 2000, ed II. Sydney, National Centre for Classification in Health, 2000.
  25. Sundararajan V, Henderson T, Perry C, Muggivan A, Quan H, Ghali WA: New ICD-10 version of the Charlson comorbidity index predicted in-hospital mortality. J Clin Epidemiol 2004;57:1288–1294.
  26. Goldstein LB, Samsa GP, Matchar DB, Horner RD: Charlson index comorbidity adjustment for ischemic stroke outcome studies. Stroke 2004;35:1941–1945.
  27. Quinn TJ, Dawson J, Lees JS, Chang TP, Walters MR, Lees KR: Time spent at home post-stroke: ‘home-time’ a meaningful and robust outcome measure for stroke trials. Stroke 2008;39:231–233.
  28. Gage BF, Waterman AD, Shannon W, Boechler M, Rich MW, Radford MJ: Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation. JAMA 2001;285:2864–2870.
  29. Lip GYH, Nieuwlaat R, Pisters R, Lane DA, Crijns HJGM: Refining clinical risk stratification for predicting stroke and thromboembolusm in atrial fibrillation using a novel risk factor-based approach: the Euro Heart survey on atrial fibrillation. Chest 2010;137:263–272.
  30. Lip GYH, Frison L, Halperin JL, Lane DA: Comparative validation of a novel risk score for predicting bleeding risk in anticoagulated patients with atrial fibrillation. The HAS-BLED score. J Am Coll Cardiol 2011;57:173–180.
  31. Lip GYH: Implications of the CHA2DS2-VASc and HAS-BLED scores for thromboprophylaxis in atrial fibrillation. Am J Med 2011;124:111–114.
  32. Barros AJD, Hirakata VN: Alternatives for logistic regression in cross-sectional studies: an empirical comparison of models that directly estimate the prevalence ratio. BMC Med Res Methodol 2004;3:21.

    External Resources

  33. Nguyen TA, Page A, Aggarwal A, Henke P: Social determinants of discharge destination for patients after stroke with low admission FIM instrument scores. Arch Phys Med Rehabil 2007;88:740–744.
  34. Candelise L, Pinardi G, Morabito A: Mortality in acute stroke with atrial fibrillation. Stroke 1991;22:169–174.
  35. SPSS for Windows, Release 18.0.2. Chicago, SPSS Inc, 2010.
  36. SAS software, version 9.2, of the SAS System for Windows 2002–2008 by SAS Institute Inc, Cary, USA.
  37. Cadilhac D, Pearce D, Mihalopoulos C, Carter R, Davis S, Donnan G: Economic evaluation of Australian stroke services: a prospective, multicenter study comparing dedicated stroke units with other care modalities. Stroke 2006;37:2790–2795.
  38. European Heart Rhythm Association/European Association for Cardio-Thoracic Surgery Camm AJ, Kirchhof P, Lip GY, Schotten U, Savelieva I, Ernst S, Van Gelder IC, Al-Attar N, Hindricks G, Prendergast B, Heidbuchel H, Alfieri O, Angelini A, Atar D, Colonna P, De Caterina R, De Sutter J, Goette A, Gorenek B, Heldal M, Hohloser SH, Kolh P, Le Heuzey JY, Ponikowski P, Rutten FH: Guidelines for the management of atrial fibrillation: the Task Force for the Management of Atrial Fibrillation of the European Society of Cardiology (ESC). Eur Heart J 2010;19:2369–2429.
  39. Healey JS, Hart RG, Pogue J, Pfeffer MA, Hohnsloer SH: Risks and benefits of oral anticoagulation compared with clopidogrel plus aspirin in patients with atrial fibrillation according to stroke risk: the atrial fibrillation clopidogrel trial with irbesartan for prevention of vascular events (ACTIVE-W). Stroke 2008;39:1482–1486.
  40. http://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/CardiovascularandRenalDrugsAdvisoryCommittee/UCM226011.pdf
  41. The ACTIVE Investigators: Effect of clopidogrel added to aspirin in patients with atrial fibrillation. N Engl J Med 2009;360:2066–2078.
  42. Biblo LA, Yuan Z, Quan KJ, Mackall JA, Rimm AA: Risk of stroke in patients with atrial flutter. Am J Cardiol 2001;87:346–349.
  43. Australian Institute of Health and Welfare (AIHW) 2006. Minimum dataset for Admitted Patient Care. Compliance evaluation 2001–2002 to 2003–2004. AIHW Cat. No. HSW 44. Health Services Series No. 28, 2006.
  44. Powell H, Lim LL, Heller RF: Accuracy of administrative data to assess comorbidity in patients with heart disease: an Australian perspective. J Clin Epidemiol 2001;54:687–693.
  45. Marsden DL, Spratt NJ, Walker R, Barker D, Attia J, Pollack M, Parsons MW, Levi CR: Trends in stroke attack rates and case fatality in the Hunter Region, Australia 1996–2008. Cerebrovasc Dis 2010;30:500–507.
  46. Tirshwell DL, Longstreth WT: Validating administrative data in stroke research. Stroke 2002;33:2465–2470.
  47. Preen DB, Holman CDJ, Lawrence DM, Baynham NJ, Semmens JB: Hospital chart review provided more accurate comorbidity information than data from a general practitioner survey or an administrative database. J Clin Epidemiol 2004;57:1295–1304.
  48. Henderson T, Shepheard J, Sundararajan V: Quality of diagnosis and procedure coding in ICD-10 administrative data. Med Care 2006;44:1011–1019.
  49. Smith RE, Shobha N, Dai D, Olson DM, Reeves MJ, Saver JL, Hernandez AF, Peterson ED, Fonarow GC, Schwamm LH: Risk score for in-hospital ischemic stroke mortality derived and validated within the Get With The Guidelines-Stroke Program. Circulation 2010;122:1496–1504.
  50. Béjot Y, Aouba A, de Peretti C, Grimaud O, Aboa-Eboulé, Woimant F, Jougla E, Giroud M: Time trends in hospital-referred stroke and transient ischaemic attack: results of a 7-year nationwide survey in France. Cerebrovasc Dis 2010;30:346–354.
  51. Gandolfo C, Balestrino M, Burrone A, Del Sette M, Finocchi C: Stroke due to atrial fibrillation and the attitude to prescribing anticoagulant prevention in Italy. A prospective study of a consecutive stroke population admitted to a comprehensive stroke unit. J Neurol 2008;255:796–802.

 goto top of outline Author Contacts

Assoc. Prof. John M. Worthington
South Western Sydney Clinical School and Department of Neurophysiology
Level 1 Clinical Services Building, Elizabeth Street
Liverpool, NSW 2170 (Australia)
Tel. +61 2 9828 3646, E-Mail John.Worthington@sswahs.nsw.gov.au


 goto top of outline Article Information

Received: January 31, 2011
Accepted: May 22, 2011
Published online: September 15, 2011
Number of Print Pages : 13
Number of Figures : 2, Number of Tables : 5, Number of References : 51


 goto top of outline Publication Details

Cerebrovascular Diseases

Vol. 32, No. 4, Year 2011 (Cover Date: October 2011)

Journal Editor: Hennerici M.G. (Mannheim)
ISSN: 1015-9770 (Print), eISSN: 1421-9786 (Online)

For additional information: http://www.karger.com/CED


Copyright / Drug Dosage / Disclaimer

Copyright: All rights reserved. No part of this publication may be translated into other languages, reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording, microcopying, or by any information storage and retrieval system, without permission in writing from the publisher or, in the case of photocopying, direct payment of a specified fee to the Copyright Clearance Center.
Drug Dosage: The authors and the publisher have exerted every effort to ensure that drug selection and dosage set forth in this text are in accord with current recommendations and practice at the time of publication. However, in view of ongoing research, changes in goverment regulations, and the constant flow of information relating to drug therapy and drug reactions, the reader is urged to check the package insert for each drug for any changes in indications and dosage and for added warnings and precautions. This is particularly important when the recommended agent is a new and/or infrequently employed drug.
Disclaimer: The statements, opinions and data contained in this publication are solely those of the individual authors and contributors and not of the publishers and the editor(s). The appearance of advertisements or/and product references in the publication is not a warranty, endorsement, or approval of the products or services advertised or of their effectiveness, quality or safety. The publisher and the editor(s) disclaim responsibility for any injury to persons or property resulting from any ideas, methods, instructions or products referred to in the content or advertisements.

Abstract

Background: In the past decade the prevalence of atrial fibrillation (AF) has been increasing in ageing populations while stroke prevention and management have advanced. To inform clinician practice, health service planning and further research, it is timely to reassess the burden of AF-related ischaemic stroke. Methods: We identified patients aged 18+ years with a primary or stay diagnosis of ischaemic stroke (ICD-10-AM I63.x), from July 1, 2000 to June 30, 2006, using an administrative health dataset of all hospitalisations in New South Wales (population ∼7 million). Fact of death was determined to December 2007. Results: Of the 26,960 index cases of ischaemic stroke, 25.4% had AF recorded during admission. Median age for AF and non-AF patients was 80.4 and 75.2 years, respectively (p < 0.001). Mortality was significantly higher in patients with AF at 30 days (19.4 vs. 11.5%), 90 days (27.7 vs. 15.8%) and 365 days (38.5 vs. 22.6%) (p values <0.0001). Adjusting for age and co-morbidities reduced these differences, with 90-day mortality of 20.9% in AF patients versus 14.7% in non-AF patients (p value <0.0001). The effect of AF on outcomes appears stronger in younger stroke patients relative to patients without AF (p valueinteraction <0.0001). At 30 days, the relative risk of mortality due to AF was 3.16 (95% CI 1.92–5.25) amongst those younger than 50, 1.71 (95% CI 1.32–2.22) in patients aged 50–64 years, 1.39 (95% CI 1.16–1.66) in patients aged 65–74 years, 1.29 (95% CI 1.17–1.43) in those aged 75–84 years, and 1.23 (95% CI 1.13–1.33) in those aged 85+ years. AF patients, surviving admission, spent a median of 19.2 days (95% CI 18.4–20.1) in hospital compared with 14.5 days (95% CI 13.9–15.1) for patients without AF (p < 0.001), with differences in length of stay greatest in younger patients (p valueinteraction <0.0001). 90-Day stroke survivors with AF spent an average of 21.5 days (95% CI 20.6–22.4) in hospital versus 16.6 days (95% CI 15.9–17.2) in those without AF. AF patients accessed more in-hospital rehabilitation (36.6%; 95% CI 35.0–38.2) than patients without AF (31.8%; 95% CI 31.0–32.7) (p value <0.0001), and differences in the proportion of AF versus non-AF patients accessing rehabilitation was greatest in younger patients (p valueinteraction <0.0006). Conclusions: Ischaemicstroke patients with AF have substantially worse outcomes than patients without AF, which can be partly explained by older age and greater co-morbidities. We have quantified the large effect of AF in younger patients and our results strongly argue for new antithrombotic research in young AF patients.



 goto top of outline Author Contacts

Assoc. Prof. John M. Worthington
South Western Sydney Clinical School and Department of Neurophysiology
Level 1 Clinical Services Building, Elizabeth Street
Liverpool, NSW 2170 (Australia)
Tel. +61 2 9828 3646, E-Mail John.Worthington@sswahs.nsw.gov.au


 goto top of outline Article Information

Received: January 31, 2011
Accepted: May 22, 2011
Published online: September 15, 2011
Number of Print Pages : 13
Number of Figures : 2, Number of Tables : 5, Number of References : 51


 goto top of outline Publication Details

Cerebrovascular Diseases

Vol. 32, No. 4, Year 2011 (Cover Date: October 2011)

Journal Editor: Hennerici M.G. (Mannheim)
ISSN: 1015-9770 (Print), eISSN: 1421-9786 (Online)

For additional information: http://www.karger.com/CED


Copyright / Drug Dosage

Copyright: All rights reserved. No part of this publication may be translated into other languages, reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording, microcopying, or by any information storage and retrieval system, without permission in writing from the publisher or, in the case of photocopying, direct payment of a specified fee to the Copyright Clearance Center.
Drug Dosage: The authors and the publisher have exerted every effort to ensure that drug selection and dosage set forth in this text are in accord with current recommendations and practice at the time of publication. However, in view of ongoing research, changes in goverment regulations, and the constant flow of information relating to drug therapy and drug reactions, the reader is urged to check the package insert for each drug for any changes in indications and dosage and for added warnings and precautions. This is particularly important when the recommended agent is a new and/or infrequently employed drug.
Disclaimer: The statements, opinions and data contained in this publication are solely those of the individual authors and contributors and not of the publishers and the editor(s). The appearance of advertisements or/and product references in the publication is not a warranty, endorsement, or approval of the products or services advertised or of their effectiveness, quality or safety. The publisher and the editor(s) disclaim responsibility for any injury to persons or property resulting from any ideas, methods, instructions or products referred to in the content or advertisements.

References

  1. Wolf PA, Abbot RD, Kannel WB: Atrial fibrillation as an independent risk factor for stroke: the Framingham Study. Stroke 1991;22:983–988.
  2. Frost L, Andersen LV, Johnsen SP, Mortensen LS: Lost life-years attributable to stroke among patients with nonvalvular atrial fibrillation: a nationwide population-based follow-up study. Neuroepidemiology 2007;29:59–65.
  3. Thygesen SK, Frost L, Eagle KA, Johnsen SP: Atrial fibrillation in patients with ischemic stroke: a population-based study. Clin Epidemiol 2009;1:55–65.

    External Resources

  4. Hannon N, Sheehan O, Kelly L, Marnane M, Merwick A, Moore A, Kyne L, Duggan J, Moroney J, McCormack PME, Daly L, Fitz-Simon N, Harris D, Horgan G, Williams EB, Furie PL, Kelly PJ: Stroke associated with atrial fibrillation – incidence and early outcomes in the North Dublin Population Stroke Study. Cerebrovasc Dis 2010;29:43–49.
  5. Gladstone DJ, Bui E, Fang J, Laupacis A, Lindsay MP, Tu JV, Silver FL, Kapral MK: Potentially preventable strokes in high-risk patients with atrial fibrillation who are not adequately anticoagulated. Stroke 2009;40:235–240.
  6. Béjot Y, Ben Salem D, Osseby GV, Durier J, Marie C, Cottin Y, Moreau T, Giroud M: Epidemiology of ischaemic stroke from atrial fibrillation in Dijon, France, from 1985 to 2006. Neurology 2009;72:346–353.
  7. Frost L, Vukelic-Andersen L, Vestergaard P, Husted S. Mortensen LS: Trend in mortality after stroke with atrial fibrillation. Am J Med 2007;120:47–53.
  8. Marini C, De Santis F, Sacco S, Russo T, Olivieri L, Totaro R, Carolei A: Contribution of atrial fibrillation to incidence and outcome of ischaemic stroke: results from a population-based study. Stroke 2005;36:1115–1119.
  9. Lamassa M, Di Carlo A, Pracucci G, Basile AM, Trefoloni G, Vanni P, Spolveri S, Baruffi MC, Landini G, Ghetti A, Wolfe CDA, Inzitari D: Characteristics, outcome and care of stroke associated with atrial fibrillation in Europe: data from a multicentre multinational hospital-based stroke registry (The European Community Stroke Project). Stroke 2001;32:392–398.
  10. Kimura K, Minematsu K, Yamaguchi T: Atrial fibrillation as a predictive factor for severe stroke and early death in 15,831 patients with acute ischaemic stroke. J Neurol Neurosurg Psychiatry 2005:76:679–683.
  11. Steger C, Pratter A, Marinek-Bregel M, Avanzini M, Valentin A, Slany J, Stollberger C: Stroke patients with atrial fibrillation have a worse prognosis than patients without: data from the Austrian Stroke Registry. Eur Heart J 2004;25:1734–1740.
  12. Karataş M, Dilek A, Erkan H, Yavuz N, Sözay S, Akman N: Functional outcome in stroke patients with atrial fibrillation. Arch Phys Med Rehabil 2000;81:1025–1029.
  13. Miyasaka Y, Barnes ME, Gersh BJ, Cha SS, Seward JM, Bailey KR, Iwasaka Y, Tsang TSM: Time trends of ischemic stroke incidence and mortality in patients diagnosed with first atrial fibrillation in 1980 to 2000: report of a community-based study. Stroke 2005;36:2362–2366.
  14. Sandercock P, Banford J, Dennis M, Burn J, Slattery J, Jones L, Boonyakarnkul S, Warlow C: Atrial fibrillation and stroke: prevalence in different types of stroke and influence on early and long term prognosis (Oxfordshire Community Stroke Project). BMJ 1992;305:1460–1465.
  15. Friedman PJ: Atrial fibrillation after stroke in the elderly. Stroke 1991;22:209–214.
  16. Jørgensen HS, Nakayam H, Reith K, Raaschou HO, Olsen TS: Acute stroke with atrial fibrillation: The Copenhagen Stroke Study. Stroke 1996;27:1765–1769.
  17. Kaarisalo MM, Immonen-Raiha P, Marttila RJ, Salomaa V, Kaarsalo E, Salmi K, Sarti C, Sivenius J, Torppa J, Tuomilehto K: Atrial fibrillation and stroke: mortality and causes of death after the first acute ischaemic stroke. Stroke 1997;28:311–315.
  18. Lin HJ, Wolf PA, Kelly-Hayes M, Beiser AS, Kase CS, Benjamin EJ, D’Agostino RB: Stroke severity in atrial fibrillation: the Framingham Study. Stroke 1996;27:1760–1764.
  19. Lowe GD, Jaap AJ: Forbes CD. Relation of atrial fibrillation and high haematocrit to mortality in acute stroke. Lancet 1983;321:784–786.

    External Resources

  20. Dulli DA, Stanko H, Levine RL: Atrial fibrillation is associated with severe acute ischemic stroke. Neuroepidemiology 2003;22:118–123.
  21. Cadilhac DA, Pearce DC, Levi CR, Donnan GA: Improvements in the quality of care and health outcomes with new stroke care units following implementation of a clinician-led, health system redesign programme in New South Wales, Australia. Qual Saf Health Care 2008;17:329–333.
  22. Gattellari M, Worthington J, Jalaludin B, Mohsin M: Stroke unit care in a real-life setting: can results from randomized controlled trial be translated into every-day clinical practice? An observational study of hospital data in a large Australian population. Stroke 2009;40:10–17.
  23. Frost L, Vukelic-Andersen L, Vestergaard P, Husted S, Mortensen LS: Trends in risk of stroke in patients with a hospital diagnosis of nonvalvular atrial fibrillation: National Cohort Study in Denmark, 1980–2002. Neuroepidemiology 2006;26:212–219.
  24. ICD 10th Revision, Australian Modification: Commonwealth of Australia, 2000, ed II. Sydney, National Centre for Classification in Health, 2000.
  25. Sundararajan V, Henderson T, Perry C, Muggivan A, Quan H, Ghali WA: New ICD-10 version of the Charlson comorbidity index predicted in-hospital mortality. J Clin Epidemiol 2004;57:1288–1294.
  26. Goldstein LB, Samsa GP, Matchar DB, Horner RD: Charlson index comorbidity adjustment for ischemic stroke outcome studies. Stroke 2004;35:1941–1945.
  27. Quinn TJ, Dawson J, Lees JS, Chang TP, Walters MR, Lees KR: Time spent at home post-stroke: ‘home-time’ a meaningful and robust outcome measure for stroke trials. Stroke 2008;39:231–233.
  28. Gage BF, Waterman AD, Shannon W, Boechler M, Rich MW, Radford MJ: Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation. JAMA 2001;285:2864–2870.
  29. Lip GYH, Nieuwlaat R, Pisters R, Lane DA, Crijns HJGM: Refining clinical risk stratification for predicting stroke and thromboembolusm in atrial fibrillation using a novel risk factor-based approach: the Euro Heart survey on atrial fibrillation. Chest 2010;137:263–272.
  30. Lip GYH, Frison L, Halperin JL, Lane DA: Comparative validation of a novel risk score for predicting bleeding risk in anticoagulated patients with atrial fibrillation. The HAS-BLED score. J Am Coll Cardiol 2011;57:173–180.
  31. Lip GYH: Implications of the CHA2DS2-VASc and HAS-BLED scores for thromboprophylaxis in atrial fibrillation. Am J Med 2011;124:111–114.
  32. Barros AJD, Hirakata VN: Alternatives for logistic regression in cross-sectional studies: an empirical comparison of models that directly estimate the prevalence ratio. BMC Med Res Methodol 2004;3:21.

    External Resources

  33. Nguyen TA, Page A, Aggarwal A, Henke P: Social determinants of discharge destination for patients after stroke with low admission FIM instrument scores. Arch Phys Med Rehabil 2007;88:740–744.
  34. Candelise L, Pinardi G, Morabito A: Mortality in acute stroke with atrial fibrillation. Stroke 1991;22:169–174.
  35. SPSS for Windows, Release 18.0.2. Chicago, SPSS Inc, 2010.
  36. SAS software, version 9.2, of the SAS System for Windows 2002–2008 by SAS Institute Inc, Cary, USA.
  37. Cadilhac D, Pearce D, Mihalopoulos C, Carter R, Davis S, Donnan G: Economic evaluation of Australian stroke services: a prospective, multicenter study comparing dedicated stroke units with other care modalities. Stroke 2006;37:2790–2795.
  38. European Heart Rhythm Association/European Association for Cardio-Thoracic Surgery Camm AJ, Kirchhof P, Lip GY, Schotten U, Savelieva I, Ernst S, Van Gelder IC, Al-Attar N, Hindricks G, Prendergast B, Heidbuchel H, Alfieri O, Angelini A, Atar D, Colonna P, De Caterina R, De Sutter J, Goette A, Gorenek B, Heldal M, Hohloser SH, Kolh P, Le Heuzey JY, Ponikowski P, Rutten FH: Guidelines for the management of atrial fibrillation: the Task Force for the Management of Atrial Fibrillation of the European Society of Cardiology (ESC). Eur Heart J 2010;19:2369–2429.
  39. Healey JS, Hart RG, Pogue J, Pfeffer MA, Hohnsloer SH: Risks and benefits of oral anticoagulation compared with clopidogrel plus aspirin in patients with atrial fibrillation according to stroke risk: the atrial fibrillation clopidogrel trial with irbesartan for prevention of vascular events (ACTIVE-W). Stroke 2008;39:1482–1486.
  40. http://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/CardiovascularandRenalDrugsAdvisoryCommittee/UCM226011.pdf
  41. The ACTIVE Investigators: Effect of clopidogrel added to aspirin in patients with atrial fibrillation. N Engl J Med 2009;360:2066–2078.
  42. Biblo LA, Yuan Z, Quan KJ, Mackall JA, Rimm AA: Risk of stroke in patients with atrial flutter. Am J Cardiol 2001;87:346–349.
  43. Australian Institute of Health and Welfare (AIHW) 2006. Minimum dataset for Admitted Patient Care. Compliance evaluation 2001–2002 to 2003–2004. AIHW Cat. No. HSW 44. Health Services Series No. 28, 2006.
  44. Powell H, Lim LL, Heller RF: Accuracy of administrative data to assess comorbidity in patients with heart disease: an Australian perspective. J Clin Epidemiol 2001;54:687–693.
  45. Marsden DL, Spratt NJ, Walker R, Barker D, Attia J, Pollack M, Parsons MW, Levi CR: Trends in stroke attack rates and case fatality in the Hunter Region, Australia 1996–2008. Cerebrovasc Dis 2010;30:500–507.
  46. Tirshwell DL, Longstreth WT: Validating administrative data in stroke research. Stroke 2002;33:2465–2470.
  47. Preen DB, Holman CDJ, Lawrence DM, Baynham NJ, Semmens JB: Hospital chart review provided more accurate comorbidity information than data from a general practitioner survey or an administrative database. J Clin Epidemiol 2004;57:1295–1304.
  48. Henderson T, Shepheard J, Sundararajan V: Quality of diagnosis and procedure coding in ICD-10 administrative data. Med Care 2006;44:1011–1019.
  49. Smith RE, Shobha N, Dai D, Olson DM, Reeves MJ, Saver JL, Hernandez AF, Peterson ED, Fonarow GC, Schwamm LH: Risk score for in-hospital ischemic stroke mortality derived and validated within the Get With The Guidelines-Stroke Program. Circulation 2010;122:1496–1504.
  50. Béjot Y, Aouba A, de Peretti C, Grimaud O, Aboa-Eboulé, Woimant F, Jougla E, Giroud M: Time trends in hospital-referred stroke and transient ischaemic attack: results of a 7-year nationwide survey in France. Cerebrovasc Dis 2010;30:346–354.
  51. Gandolfo C, Balestrino M, Burrone A, Del Sette M, Finocchi C: Stroke due to atrial fibrillation and the attitude to prescribing anticoagulant prevention in Italy. A prospective study of a consecutive stroke population admitted to a comprehensive stroke unit. J Neurol 2008;255:796–802.