Acute Stroke With Atrial Fibrillation

Atrial fibrillation is a common arrhythmia and a potent risk factor for stroke. The prevalence of AF is approximately 1% in the general population; in those older than 65 years, it is 5.9%.¹ The relative risk of stroke in patients with AF ranges from 5.6 to 17.6 of that observed in individuals without AF, depending on whether associated valvular disease is present.² AF is present in 15% to 21% of the patients with stroke.^3456789

Knowledge of AF as a prognostic factor in stroke is sparse. An increased mortality rate in stroke patients with AF has been reported in most³⁵⁶⁷⁸⁹ but not all¹⁰¹¹ previous studies. Whether the increased mortality rate is explained by a higher age in patients with AF, larger lesions, more-severe strokes, or by an effect of AF on the process of recovery is unsettled. The influence of AF on neurological and functional outcomes, length of hospital stay, and the need for institutional care after completed acute care and rehabilitation remains to be explored. A characterization of the stroke in patients with AF may also increase our understanding of the pathogenetic mechanisms of AF-induced stroke.

Recent randomized trials have demonstrated a pronounced effect of anticoagulant treatment on the stroke rate in patients with AF.¹² This treatment is now recommended for both primary and secondary stroke prevention. Despite the results of these randomized, controlled trials, many physicians remain reluctant to provide anticoagulant therapy especially for elderly patients.¹ Detailed knowledge of stroke in patients with AF sampled from a community-based stroke population is required to estimate the effect of stroke prevention in terms of reduction in stroke frequency, mortality, need of poststroke institutional care, and reduction in the number of hospital beds used by stroke patients.

Recently it has been questioned whether AF is a risk factor for stroke in both men and women, or if the risk is increased in women only.¹³¹⁴ Another controversy regards the relation between AF and silent cerebral infarcts. A high frequency of silent infarcts in patients with AF has been reported from retrospective studies.¹⁵¹⁶¹⁷

To address these issues, we characterized stroke in patients with AF and in patients with sinus rhythm in the Copenhagen Stroke Study. We compared associated risk factors, initial neurological stroke severity, CT characteristics, mortality, neurological and functional recovery, duration of acute care and rehabilitation in the hospital, neurological impairment and functional disability after completed treatment, and the need for poststroke institutional care. Multivariate analyses were used to test the specific influence of AF on these factors.

Patients and Methods

The Copenhagen Stroke Study is a prospective, community-based study of 1197 consecutive acute stroke patients treated on a single, 63-bed stroke unit from the time of acute admission to the end of rehabilitation. This stroke unit receives all stroke patients admitted from a well-defined catchment area (population, 238 886) of Copenhagen City, Denmark. The stroke unit handles all stages of acute care, workup, and all stages of rehabilitation in all patients, regardless of the age of the patient, the severity of the stroke, and the condition of the patient prior to the stroke. The stroke admission rate in the area is high, 88%.¹⁸ Inclusion started September 1, 1991, and ended September 30, 1993. Twelve patients were excluded because of insufficient ECG data on admission.

CT was performed with a Siemens Somatom DRG scanner. All scans were described by the same radiologist (H.O.R.) blinded to clinical data. Type of lesion (infarct/hemorrhage¹⁹ ), size (largest diameter in mm), cortical involvement, leukoaraiosis (diffuse or patchy lucencies of the white matter or centrum ovale²⁰ ), infarcts from previous strokes, silent infarcts,²¹ and whether an infarct was lacunar (noncortical, diameter <15 mm) were noted.

AF was diagnosed if the ECG obtained on admission revealed AF.

The initial neurological stroke severity was assessed with the SSS²²²³ at the time of the acute admission. The SSS evaluates neurological deficits (level of consciousness; paresis of arm, hand, and leg; aphasia, etc), and the total score ranges from 0 to 58 (maximum) points. The initial level of functional disability was assessed with the BI²⁴ during the first week after the acute admission. The BI scores the ADL. It evaluates 10 abilities (eating, grooming, walking, etc), and the total score ranges from 0 to 100 (maximum) points. The neurological deficit (neurological outcome) and the ADL level (functional outcome) after completed rehabilitation were assessed with the SSS and BI, respectively, at the time of discharge. Stroke in progression,²⁵ recurrent stroke during hospital stay, mortality rate, length of hospital stay, and discharge rates to nursing homes and to the patient's own home were recorded.

Statistical analysis was performed using the SPSS package.²⁶ In univariate analysis, the Student's t test was used for continuous data and the χ² test was used for noncontinuous data. Multiple linear and logistic regression models were used to test and quantify relations independent of possible confounding factors. Only factors with a potential relationship to the dependent variable were included in the model (age, sex, hypertension, diabetes,²⁷ intermittent claudication, ischemic heart disease in the analyses of CT findings, and, in addition, information of comorbidity,²¹ stroke in progression, and recurrent stroke in outcome analyses). All variables of interest were tested by the backward procedure. The level of significance was chosen to be P<.05. The study was approved by the Ethics Committee of Copenhagen, approval number V. 100.2263/91.

Results

Of the 1185 stroke patients, 217 (18.3%) had AF on admission. The frequency of AF increased steeply with age, from 2% in stroke patients <50 years old, to 28% in stroke patients in their 80s, and to 40% in stroke patients ≥90 years of age (Figure). Of the stroke patients with AF, 88% were ≥70 years of age (compared with 69% of the patients with sinus rhythm), 30% were between 70 and 80, and 47% were between 80 and 90.

Outcome

Stroke was more severe in patients with AF (Table 1). A multiple linear regression analysis showed that AF was associated with a decrease in initial SSS score by 5.7 points on average (B=−5.7; SE(B)=1.3, P<.0001). Initial disability was also more severe in patients with AF. A multiple linear regression analysis showed that AF was associated with a decrease in initial BI score by 9.5 points on average (B=−9.5; SE(B)=3.2, P=.003).

Patients with AF had higher mortality rates, longer hospital stays, and a lower discharge rate to their own homes (Table 1). Multiple regression analyses showed (1) that AF was associated with an increase in the relative risk of in-hospital death by a factor of 1.7 (OR, 1.7; 95% CI, 1.2 to 2.5, P=.005), (2) an increase in the length of hospital stay by 7.6 days on average (B=7.6; SE(B)=4.0, P=.04), and (3) a decrease in the chance of discharge back to the patient's own home by a factor of 0.71 (OR, 0.60; 95% CI, 0.44 to 0.85, P=.003). However, when the initial stroke severity (SSS on admission) was included in the analyses, the influence of AF became insignificant (in-hospital mortality: OR, 1.1; 95% CI, 0.7 to 1.8, P=.59; length of hospital stay: B=−0.5; SE(B)=3.7, P=.88; discharge to own home: OR, 0.85; 95% CI, 0.56 to 1.30, P=.47).

The neurological deficit was more pronounced and the functional level (ADL) lower after completed rehabilitation in patients with AF (Table 1). Multiple linear regression analyses showed that AF was associated with a decrease in neurological score after completed rehabilitation by 3.0 points (B=−3.0; SE(B)=1.1, P=.01), and a decrease in ADL score by 6.6 points (B=−6.6; SE(B)=3.2, P=.04). However, when the initial stroke severity was included in the analysis, the influence of AF on the neurological outcome became insignificant (B=0.4; SE(B)=0.9, P=.65), and when the initial functional ADL level (initial BI score) was included in the analysis, the influence of AF on functional outcome became insignificant (B=−1.3; SE(B)=2.1, P=.51). Stroke in progression and recurrent stroke (new stroke during the hospital stay) were equally frequent in patients with AF and in patients with sinus rhythm (Table 2).

Associated Risk Factors

Table 2 shows univariate comparisons of basic characteristics and risk factor distribution in patients with AF versus sinus rhythm. The multivariate logistic regression analysis revealed that the observed univariate relations found between AF and female sex and between AF and comorbidity disappeared after adjustment for difference in age between patients with and without AF; in the multivariate analysis, AF was associated only with increasing age (OR, 2.0 per 10-year increase; 95% CI, 1.6 to 2.6), ischemic heart disease (OR, 3.4; 95% CI, 2.4 to 4.8), previous stroke (OR, 1.8; 95% CI, 1.2 to 2.6), and inversely with systolic blood pressure on admission (OR, 0.93 per 10-mm Hg increase; 95% CI, 0.88 to 0.99).

CT Findings

Patients with AF had larger infarcts and had less-frequent cerebral hemorrhage (Table 3). A multiple linear regression analysis showed that AF increased infarct size by 9.3 mm on average (B=9.3; SE(B)=3.0, P<.001). Cortical involvement of the infarct was more frequent in patients with AF. A multiple logistic regression analysis showed that AF increased the relative risk of an infarct being cortical by a factor of 2.7 (OR, 2.7; 95% CI, 1.8 to 4.2, P<.0001). Lacunar infarcts were less frequent in patients with AF. A multiple logistic regression analysis showed that patients with AF were less likely to have lacunar infarcts (OR, 0.50; 95% CI, 0.26 to 0.98).

An infarct from a previous stroke was found more frequently in patients with AF. The multiple logistic regression analysis showed that AF increased the relative risk of CT revealing an infarct from a previous known stroke by a factor of 2.0 (OR, 2.0; 95% CI, 1.1 to 3.5, P<.01). Patients with AF had leukoaraiosis less frequently. A multiple logistic regression analysis showed that AF decreased the relative risk of leukoaraiosis by a factor of 0.33 (OR, 0.33; 95% CI, 0.18 to 0.60, P<.0001).

Silent infarcts were not more frequent in patients with AF. A multiple logistic regression analysis confirmed that AF per se did not increase the relative risk of silent infarcts in stroke patients (OR, 0.99; 95% CI, 0.65 to 1.5, P=.97).

Discussion

The Copenhagen Stroke Study is the first study to measure in detail the consequences of stroke in patients with AF. Using multiple regression analysis, we studied the specific influence of AF on mortality rates, on neurological and functional outcomes after completed rehabilitation, and on the length of hospital stay, all independent of other influencing factors. AF was associated with a 70% increase in mortality, a 40% decrease in the relative chance of discharge to a patient's own home, a 20% increase in the length of hospital stay, and a marked increase in impairment and disability in survivors. These detrimental effects were explained exclusively by one factor: Patients with AF had more-severe initial strokes. AF per se did not affect the process of recovery from stroke. Further, the rates of recurrent strokes were similar in patients with AF and in patients with sinus rhythm, as also reported in most previous studies,³⁷⁸²⁸ but not in all.¹⁰²⁹ Most previous studies have reported AF to be associated with an increased mortality rate in stroke patients.^35678911 Some studies have found the increased mortality rate to be due solely to increased age, larger lesion size, or more-severe initial neurological impairment in patients with AF⁶¹¹ ; others have found an additional detrimental effect of AF on the recovery from stroke.⁵⁸⁹ One study⁶ reported functional outcome to be similar in stroke patients with AF and sinus rhythm, whereas another study¹¹ found functional dependence, but not mortality rate, related to AF.

It is well established from randomized trials that proper anticoagulant treatment of patients with AF can prevent approximately 68% of all strokes with virtually no increase in major bleeding.¹² Despite the results of the trials, many physicians remain reluctant to provide anticoagulant therapy, especially for older patients with AF.¹ However, most individuals with AF are old—an estimated one third of the individuals with AF in the United States are older than 80 years.¹ In the unselected stroke population of the Copenhagen Stroke Study, patients with AF were considerably older than those with sinus rhythm: mean age was 80 versus 73 years. The markedly poorer outcome, increased length of hospital stay, and increased discharge rate to nursing homes found in this study for stroke patients with AF may suggest a substantial beneficial impact from the prevention of AF-induced stroke.

The incidence of AF found in this study is comparable with the incidence reported in other community-based studies.⁷⁸⁹ Because stroke in itself can sometimes be arrhythmogenic, some patients with AF may have had AF as a consequence of the stroke at the time of acute admission. Vingerhoets et al³⁰ found that 14% of stroke patients with AF diagnosed on admission did not have a history of AF. AF is, however, often asymptomatic and may merely have been undiagnosed before the stroke in these cases. The Framingham Study³¹ reported that AF persisted in 92% of the stroke patients with AF undiagnosed prior to the stroke. Thus, in the vast majority of cases it may be assumed that AF at the time of acute admission is not a short-lived, stroke-induced arrhythmia but a marker of AF at the onset of stroke.

The fact that patients with AF have more-severe strokes than patients with sinus rhythm supports the hypothesis that the pathogenesis of stroke may be different. Stroke in patients with AF may chiefly be of cardioembolic origin, causing occlusion of larger cerebral arteries and, subsequently, more-severe strokes. Most findings in this study are in line with this: patients with AF more often had large cortical infarcts on CT, consistent with other studies,³⁸³²³³ and less frequently they had lacunar infarcts, as previously reported.⁶ Stroke patients with AF had a higher frequency of previous symptomatic (large) infarcts but not a higher frequency of previous transient ischemic attack or of silent (small) infarcts. Furthermore, diffuse leukoaraiosis, which may be a marker of cerebral small vessel disease, was less frequent in patients with AF, as also previously reported from the Copenhagen Stroke Study.²⁰ AF in patients with small, lacunar strokes may merely be a coincidental finding; AF was rarely the only risk factor for stroke in our population. Also, intracerebral hemorrhage was a significantly less frequent cause of stroke in patients with AF, as reported previously.⁸ AF in these patients may also be merely coincidental.

This study confirms that age and ischemic heart disease are related to the presence of AF.³⁵⁶¹³ Retrospective reports of a higher frequency of silent infarcts in patients with AF¹⁵¹⁶¹⁷ could not be confirmed by this study or by other recent, prospective, unselected studies of stroke patients²¹³⁴³⁵ ; AF is not an independent risk factor for silent infarcts in the stroke population. It has been suggested that AF is a risk factor for stroke in women but not in men.¹³¹⁴ This could not be confirmed by this study. Although AF was more frequent in women, this univariate relation was due to the confounding effect of age: female patients were older. The multiple logistic regression analysis showed that AF was equally related to both sexes as a risk factor for stroke.

A chronically reduced cerebral blood flow in patients with AF compared with flow in patients with sinus rhythm has previously been reported,³⁶ and decreased cerebral blood flow during AF has been found to normalize after electrical cardioversion to sinus rhythm.³⁷ We found systolic blood pressure on admission to be significantly lower in patients with AF. The combined effect of chronically decreased cerebral blood flow and a relatively low systemic blood pressure in the acute phase of stroke may contribute to the severity of stroke in patients with AF by further reducing blood flow in the ischemic penumbra. Perfusion in the ischemic penumbra is known to be dependent on the systemic blood pressure in the acute stage of stroke.³⁸

In conclusion, the results from the Copenhagen Stroke Study support the hypothesis that stroke caused by AF is usually of embolic origin, leading to occlusion of larger cerebral arteries, and subsequently more-severe strokes. Stroke in patients with AF is devastating and much more severe than in patients with sinus rhythm. This emphasizes the importance of active tracing and subsequent anticoagulant treatment of individuals with AF in the general population. A lower blood pressure in patients with AF in the acute stage of stroke may also contribute to the increased stroke severity.

Selected Abbreviations and Acronyms

• Download figure
• Download PowerPoint

Footnotes