Authors: Jiongxing Wu, Xinmao Wu, Yanan Wang, Yuxi Jin, Wen Guo, Yajun Cheng, Deren Wang, Junfeng Liu
Categories: Research, Atrial fibrillation, Ischemic stroke, Echocardiographic
Source: BMC Cardiovascular Disorders
Authors: Jiongxing Wu, Xinmao Wu, Yanan Wang, Yuxi Jin, Wen Guo, Yajun Cheng, Deren Wang, Junfeng Liu
Newly detected atrial fibrillation after stroke (AFDAS) is a specific type of AF and have a different pathophysiology compared to patients with previously known AF before a stroke (KAF). However, the characteristics and outcomes between AFDAS and KAF have not been well studied. We aimed to further explore the clinical characteristics and long-term functional outcomes between AFDAS and KAF.
We retrospectively analyzed acute ischemic stroke patients admitted to our hospital between 2010 and 2017, who was also diagnosed with AF. The poor outcome was defined by the combination of death and any disability as scored by the modified Rankin Scale (mRS) score at 3 months and 12 months.
Among the final sample of 698 patients, 370 (53%) were classified into KAF and 328 (47%) were AFDAS. Compared to KAF, patients with AFDAS had a lower prevalence of hypertension (P = 0.01), previous stroke (P = 0.02), higher prevalence of valvular heart disease (P = 0.02), less prescribed with anticoagulants during the hospitalization (P < 0.01), and higher mortality at 3 months (OR 1.77, 95%CI 1.07–2.92, P = 0.03). Patients with AFDAS had an increased trend of mRS score compared to KAF (3 months, OR 1.37, 95% CI 1.04–1.81, P = 0.02; 12 months, OR 1.32, 95%CI 1.01–1.75, P = 0.04).
Patients with AFDAS had a higher risk of death and an increased trend of mRS score compared to KAF. Lower utilization of anticoagulation may be a major reason of worse outcome in AFDAS. Further investigation is required to identify whether the differences of the outcomes between AFDAS and KAF is confounded by pre-existing anticoagulation.
Atrial fibrillation is the most common cardiac arrhythmia and a major risk of acute ischemic stroke (AIS) [1, 2]. With the improvement of cardiac screening technologies, the detection of AF after stroke has improved substantially [3]. Among patients with AF-related ischemic stroke, around 30% of patients were diagnosed with AF before stroke (KAF) [4], and 24% were newly diagnosed AF after stroke during hospitalization (AFDAS) [5–7].
Compared to KAF, AFDAS may have a distinct pathological mechanism, including cardiogenic and neurogenic types, as well as a combination of both [8]. Cardiac structure and function changes such as atrial remodeling, are frequently observed in patients with previously known AF [9–11]. However, it is not fully clear whether these changes can also be found in patients with AFDAS. Several studies have explored the outcomes between KAF and AFDAS, though the results were not consistent. A recent study reported that the AFDAS patients had the higher risk of death, compared to KAF [12]. Conversely, another study indicated that AFDAS had similar risk of 1-year ischemic stroke recurrence and mortality when compared to KAF [13].
Considering the inconsistency of previous findings, it is necessary to further investigate the clinical characteristics, echocardiographic parameters and long-term functional outcomes between AFDAS and KAF.
We retrospectively analyzed acute ischemic stroke (AIS) patients admitted to the department of neurology, West China hospital, Sichuan University from 2010 to 2017. Ischemic stroke was diagnosed according to World Health Organization criteria [14], and confirmed by computed tomography (CT) or magnetic resonance imaging (MRI). The present study included patients with AIS who had AF defined according to the guidelines from the European Society of Cardiology [15]. A standard 12-lead ECG recording or a single-lead ECG tracing of > 30 s showing heart rhythm with no discernible repeating P waves and irregular RR intervals is diagnosed with AF. Based on their history of AF and the results of ECG monitoring in our hospital, we classified the patients into 2 (1) AFDAS (no history of AF but AF detected during hospitalization after ischemic stroke), (2) KAF (history of AF known before ischemic stroke). Patients were excluded if they (1) were younger than 18 years; (2) had malignant tumor, autoimmune disease or stroke due to brain trauma or neoplasm; (3) were lost to follow-up. The scientific use of the data collected in the study was approved by biomedical ethics committee of West China hospital (2019[362]).
Using a standardized form, we collected patients’ data as age, sex, stroke severity, vascular risk factors, and treatment in hospital [including antiplatelet, anticoagulation, statin, and reperfusion therapies (thrombolysis or endovascular therapy)]. Stroke severity was assessed based on the National Institutes of Health Stroke Scale (NIHSS) score on admission [16]. Treatments were recorded regardless of dose or duration. All patients completed a 12-lead ECG on admission, and some of the patients with high possibility of cardioembolic stroke received 24-hour Holter cardiac rhythm recording during hospitalization. All patients underwent standard echocardiography on a commercially available system (iE33, Philips Medical Systems, Bothell, WA, USA). We measured the left atrial and ventricular diameters, right atrial and ventricular diameters. Left ventricular ejection fraction (LVEF) was calculated using the biplane Simpson’s method after obtaining apical images from the 4-chamber and 2-chamber sections. Cardiac function was assessed by the attending physician according to New York Heart Association (NYHA) staging and patients’ clinical presentation during hospitalization.
The primary outcome was the death and poor outcome at 3 months and 12 months. Patients were followed up at 3 months and 12 months by telephone interview with the patients themselves or, if this was impossible, with a relative. Poor outcome was defined as a score of 3–6 according to the modified Rankin Scale (mRS) score [17] at 3 months and 12 months. Secondary outcome was the ordinal shift of mRS score. A patient was classified as lost to follow-up after at least three attempts to contact patients or their families by telephone.
Frequencies and descriptive statistics were used to describe the patients’ baseline characteristics. We compared the differences of baseline characteristics using Student’s t-test or the Mann–Whitney U-test for continuous variables and χ2 or Fisher’s exact test for categorical variables. To estimate the differences of the outcomes at 3 months and 12 months between KAF and AFDAS, both multivariate logistic regression models and ordinal shift analysis were performed. All statistical analyses were performed by using the R statistical programming environment (version 3.4.1. http://www.R-project.org). A P value < 0.05 was considered to indicate a significant difference. The data that support the findings of this study are available from the corresponding author on reasonable request.
During 2010 and 2017, a total of 734 AIS patients with AF met our inclusion criterion. We excluded 28 patients without follow-up, 8 patients with malignant tumor. In the end, 698 patients(mean age, 70 ± 11.48 years; male,294 patients) were included in our analysis. Among those patients, 370 (53%) patients were classified into KAF and 328 (47%) patients were AFDAS (Table 1).
Table 1Differences in baseline characteristics of stroke patients with KAF and AFDASOverallKAFAFDAS p n698370328Sex (male), n (%)294 (42.1)143 (38.6)151 (46.0)0.06Age (years), mean ± SD70.00 (11.48)70.66 (11.03)69.26 (11.95)0.11Age group0.21≤ 5692 (13.2)41 (11.1)51 (15.5)57–69201 (28.8)107 (28.9)94 (28.7)≥ 70405 (58.0)222 (60.0)183 (55.8)Hypertension, n (%)348 (49.9)202 (54.6)146 (44.5)0.01Diabetes, n (%)136 (19.5)77 (20.8)59 (18.0)0.4Hyperlipidemia, n (%)179 (25.6)86 (23.2)93 (28.4)0.15Acute myocardial infarction, n(%)14 (2.0)9 (2.4)5 (1.5)0.56Coronary heart disease, n (%)98 (14.0)60 (16.2)38 (11.6)0.1Rheumatic heart disease, n(%)180 (25.8)87 (23.5)93 (28.4)0.17Valvular heart disease, n(%)378 (54.2)185 (50.0)193 (58.8)0.02Previous stroke, n(%)132 (18.9)83 (22.4)49 (14.9)0.02Alcohol consumption, n(%)149 (21.3)76 (20.5)73 (22.3)0.65Smoking, n(%)103 (14.8)53 (14.3)50 (15.2)0.81NIHSS score on admission, median (IQR)9.00 [4.00, 15.00]9.00 [4.00, 15.00]9.00 [4.00, 15.00]0.91NIHSS groups on admission, n(%)0.9Mild (≤ 3)161 (23.1)84 (22.7)77 (23.5)Moderate (4–14)344 (49.3)181 (48.9)163 (49.7)Severe (≥ 15)193 (27.7)105 (28.4)88 (26.8)antiplatelet, n(%)606 (86.8)317 (85.7)289 (88.1)0.4Anticoagulant in hospital, n(%)177 (25.4)111 (30.0)66 (20.1)< 0.01Anticoagulant at 3 month, n(%)163 (23.4)100 (27.0)63 (19.2)0.019Stains, n(%)512 (73.4)279 (75.4)233 (71.0)0.22Reperfusion treatment, n(%)44 (6.3)28 (7.6)16 (4.9)0.19Death/disability at 3 month, n(%)362 (51.9)191 (51.6)171 (52.1)0.95Death/disability at 12 month, n(%)322 (46.1)167 (45.1)155 (47.3)0.63Death at 3 month, n(%)88 (12.6)38 (10.3)50 (15.2)0.06Death at 12 month, n(%)135 (19.3)66 (17.8)69 (21.0)0.33KAF Know after Atrial Fibrillation, AFDAS Atrial Fibrillation Detected After Stroke, NIHSS National Institutes of Health Stroke Scale
Compared to KAF, patients with AFDAS were more likely to be male (46.0% vs. 38.6%, P = 0.06), and had a higher prevalence of valvular heart disease (58.8% vs. 50.0%, P = 0.02), but a lower prevalence of hypertension (44.5% vs. 56.4%, P = 0.01) and previous stroke (14.9% vs. 22.4%, P = 0.015). In addition, the prescription rate of anticoagulants during hospitalization was lower in patients with AFDAS compared to those with KAF (20.1% vs. 30.0%, P < 0.01), as was at 3 month after discharge (19.2% vs. 27.0, P = 0.019). There were no differences between the 2 groups in terms of age, diabetes, hyperlipidemia, acute myocardial infarction, coronary heart disease, rheumatic heart disease, alcohol consumption, smoking, and NIHSS score at admission. There were also no significant differences between KAF and AFDAS in regard to echocardiographic parameters and cardiac function that including LVEF, left atrial (LA); left ventricular (LV); right atrial (RA); right ventricular (RV) and NYHA stage III or IV (All P > 0.05, Table 2).
Table 2Echocardiographic characteristics and cardiac function of stroke patients with KAF and AFDASOverallKAFAFDAS p n601331270Sex (male), n (%)250 (41.6)129 (39.0)121 (44.8)0.17Age (years), mean ± SD69.64 ± 11.3070.27 ± 10.9568.87 ± 11.690.13NYHA class III or IV, n(%)124 (20.6)69 (20.8)55 (20.4)0.97LVEF, mean ± SD61.72 ± 9.7461.98 ± 9.2061.40 ± 10.370.46LA, mean ± SD43.67 ± 8.9543.56 ± 8.9643.80 ± 8.960.74LV, mean ± SD46.44 ± 6.2246.53 ± 6.0846.33 ± 6.400.69RA, mean ± SD46.40 ± 10.7246.86 ± 11.1845.83 ± 10.120.24RV, mean ± SD21.15 ± 3.6521.29 ± 3.8920.99 ± 3.340.31NYHA New York Heart Association, LVEF Left ventricular ejection fraction, LA Left atrial, LV Left ventricular, RA Right atrial, RV Right ventricular
Overall, 88 (12.6%) patients and 135 (19.3%) patients were found to be dead at 3 and 12 months, respectively. The rate of poor outcome was 51.9% (362/698) at 3 months and 46.1% (322/698) at 12 months.
Patients with AFDAS exhibited a higher risk of mortality at 3 months and 12 months compared to KAF(OR 1.57, 95%CI 1–2.47.47, P = 0.049). After adjusting for age, sex, valvular heart disease, NIHSS score on admission, and anticoagulants during hospitalization, patients with AFDAS had a consistently higher risk of death at 3 months (adjusted OR 1.77, 95%CI 1.07–2.92, P = 0.03, Table 3), but a similar risk of death at 12 months compared to KAF (adjusted OR 1.39, 95% CI 0.91–2.12, P = 0.13, Table 3).
Table 3Multivariable logistic regression models testing associations between type of AF (AFDAS vs. KAF) and death at 3 month and 12 monthVariablesDeath at 3 monthDeath at 12 monthOR (95%CI)p ValueOR (95%CI)p ValueKAFreferencereferencereferencereferenceAFDAS1.77 (1.07–2.92)0.031.39 (0.91–2.12)0.13Model was adjusted for age, sex, VHD NIHSS score on admission, use anticoagulants and statins in hospitalCI Confidence interval, OR Odds ratio
We conducted further analysis to assess the risk of poor outcome between AFDAS and KAF. After adjusting for confounders, there was no significant difference in the risk of poor outcome between AFDAS and KAF at 3 months (adjusted OR 1.18, 95%CI 0.76–1.63, P = 0.57, Table 4) and 12 months (adjusted OR 1.28, 95CI% 0.88–1.96, P = 0.20, Table 4). However, in the ordinal analysis, patients with AFDAS had an increased trend of mRS score at 3 months (adjusted OR 1.37, 95% CI 1.04–1.81, P = 0.02, Table 5; Fig. 1) and 12 months compared to KAF (adjusted OR 1.32, 95%CI 1.01–1.75, P = 0.04; Table 5; Fig. 2).
Table 4Multivariable logistic regression models testing associations between type of AF (AFDAS vs. KAF) and poor outcome at 3 month and 12 monthVariablesPoor outcome at 3 monthPoor outcome at 12 monthOR (95%CI)p ValueOR (95%CI)p ValueKAFreferencereferencereferencereferenceAFDAS1.18 (0.76–1.63)0.571.28 (0.88–1.86)0.20Model was adjusted for age, sex, ACS, VHD, NIHSS score on admission, and use anticoagulants in hospitalCI Confidence interval, OR Odds ratio
Table 5Multivariable ordinal logistic regression models testing associations between type of AF (AFDAS vs. KAF) and mRS score shift at 3 month and 12 monthVariablesmRS score shift at 3 monthmRS score shift at 12 monthOR (95%CI)p ValueOR (95%CI)p ValueKAFreferencereferencereferencereferenceAFDAS1.37 (1.04–1.81)0.021.32 (1.01–1.75)0.04Model was adjusted for age, sex, ACS, VHD, NIHSS score on admission, and use anticoagulants in hospitalCI Confidence interval, OR Odds ratio
Fig. 1Distribution of mRS scores at 3 months in Patients With KAF and AFDAS
Fig. 2Distribution of mRS scores at 12 months in Patients With KAF and AFDAS
In this study, we found that compared to KAF patients, AFDAS patients had lower frequency of hypertension, previous stroke and were less prescribed with anticoagulants during hospitalization. However, there were no significant differences in echocardiographic parameters between KAF and AFDAS patients. Although both KAF and AFDAS had similar poor outcomes at 3 months and 12 months, patients with AFDAS had an increased trend of mRS score in the ordinal shift analysis. In addition, AFDAS patients had a higher risk of mortality at 3 months compared to KAF patients.
Atrial cardiomyopathy (ACM) is more commonly observed in patients with AF and left atrial enlargement, and is associated with a potentially increased risk of thromboembolic events, which may contribute to a worse clinical prognosis [18, 19]. Left atrial enlargement is often detected in patients with previously known AF [20, 21]. Previous studies suggest that AFDAS patients are comprised of both neurogenic and cardiogenic mechanisms [8, 22], while KAF are primarily cardiogenic with minimal or no participation of central neurogenic mechanisms. In the study, AFDAS patients showed a similar echocardiographic features including LVEF, LA, LV, RA and RV compared to KAF, which is accordance with previous studies [23]. There were no differences in cardiac structure and function between patients with AFDAS and KAF. This suggests that a significant proportion of patients in the AFDAS group may have had cardiogenic AF type, representing AF that existed before the stroke but remained undiagnosed until after the occurrence. This group of patients lacked relevant treatment before the detection of atrial fibrillation, so they may be more likely to be complicated with atrial cardiomyopathy. However, due to the current complexity of ACM diagnosis, it remains unclear whether AFDAS is associated with a higher prevalence of ACM. This potential association may represent an important factor warranting further investigation in future research.
Our study indicated that AFDAS patients had higher risk of mortality at 3months, as well as an increased trend in mRS score at both 3 and 12 months compared to KAF patients. These findings suggest the distinct mechanism between AFDAS and KAF. However, whether neurogenic and cardiogenic AFDAS have different prognosis is unknown. Unfortunately, adequate distinction between neurogenic and cardiogenic AFDAS are not available, which makes this comparison impossible. Therefore, identifying specific markers of neurogenic AFDAS may contribute to our understanding of the pathological mechanism involved and facilitate individualized implementation of secondary prevention strategies in the future.
Additionally, the lower utilization rate of anticoagulation therapy during hospitalization among AFDAS observed in the study, this may be attributed to considerations regarding the risk of bleeding and other comorbidities [24, 25]. It may partially explain the higher risk of 3-month mortality in AFDAS compared to KAF. Further investigation is needed to identify whether the differences of the outcomes between AFDAS and KAF is confounded by pre-existing anticoagulation.
In addition to these findings, our study revealed interesting sex-related patterns that KAF with more female (61.4% vs. 54%, P = 0.06). This finding showed that sex differences play an important role in stroke risk stratification and outcomes in atrial fibrillation patients. Recent studies have demonstrated that female sex represents a complex risk factor in AF-related stroke, with the stroke risk varying significantly by age among women compared to men [26]. In AF patients treated with oral anticoagulants, women showed higher rates of stroke/thromboembolism events (1.33 vs. 0.94 per 100 person-years), though this increased risk was not statistically significant after adjusting for confounders and competing risk of death. The sex differences observed in our cohort may have clinical implications for risk stratification and management decisions. While our study was not specifically powered to analyze sex-specific outcomes in AFDAS vs. KAF, the trend toward more males in the AFDAS group could suggest different underlying pathophysiological mechanisms. Future studies should specifically examine whether the distinct neurogenic and cardiogenic mechanisms proposed for AFDAS manifest differently between sexes, and whether sex-specific approaches to anticoagulation and monitoring are warranted in newly detected post-stroke AF.
There are several limitations in our study. First, the lack of standardized AF detection strategies and cardiac monitoring may have led to an underestimation of AFDAS in our cohort. Secondly, the diagnosis of KAF was based on medical history, which could lead to recall bias. However, it is important to note that ascertainment of KAF in actual clinical settings is usually according to patient recall and review of medical records. Additionally, our study was not specifically designed to identify AF after ischemic stroke, patients with AF detected by routine ECG or a confirmed history of AF did not undergo Holter monitoring, thus preventing us from analyzing the proportion of patients with paroxysmal versus sustained AF in each group. Finally, although anticoagulant adherence is relatively good in the short term, data on long-term anticoagulant adherence and delayed initiation of anticoagulation are lacking. This may prevent us from comprehensively evaluating the impact of anticoagulant therapy on outcome. due to a lack of data regarding adherence to anticoagulant therapy after discharge, we were unable to evaluate the effect of anticoagulation therapy on KAF and AFDAS after discharge. Future studies should address these limitations through several approaches. First, implementing standardized AF detection protocols with prolonged cardiac monitoring could improve AFDAS identification accuracy. Second, larger multi-center studies are needed to validate our findings and enhance generalizability across different healthcare settings and populations. Third, prospective studies with systematic long-term follow-up could better evaluate anticoagulant adherence patterns and their impact on clinical outcomes. Finally, studies designed specifically to distinguish paroxysmal from sustained AF patterns in both KAF and AFDAS groups would provide valuable insights for clinical management.
Despite these limitations, we included longitudinal data of patients with ischemic stroke admitted to West China hospital over an 8-year period. It reflect the real-world data of outcomes in patients with KAF and AFDAS admitted to our hospital.
In the study, patients with AFDAS showed a higher risk of 3-month mortality. Additionally, there was an observed upward trend in mRS scores at both 3 months and 12 months for AFDAS patients. Notably, lower utilization of anticoagulation during hospitalization was identified among AFDAS patients, which may significantly contribute to their worse outcomes. Further investigation is required to identify whether the differences of the outcomes between AFDAS and KAF is confounded by pre-existing anticoagulation.