CHADS2 and CHA2DS2-VASc Scores Can Guide the Evaluation of Cryptogenic Ischemic Stroke
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.333477
Source of Support: None, Conflict of Interest: None
Keywords: CHADS2, CHA2DS2-VASc, cryptogenic stroke
In a substantial number of ischemic strokes (IS), the underlying etiology remains undetermined despite adequate investigational work-up. Paroxysmal atrial fibrillation (PAF), arterial thromboembolism, patent foramen ovale (PFO) or other cardiac structural abnormalities are proposed to be among the underlying causes of cryptogenic strokes.,, Current guidelines recommend prolonged cardiac monitoring in cryptogenic ischemic strokes (CIS) to detect cardiac arrhythmias.,, However, etiology remains mostly unidentified and patients with CIS are treated with antiplatelets for secondary prevention in routine clinical practice.
Recently the term “Embolic stroke of undetermined source (ESUS)” was proposed to define CIS with presumed embolic etiology. But ESUS still comprises many different etiologies, such as non-stenotic large artery disease, and stroke prevention strategies are not well defined in this heterogeneous group. The introduction of this term mainly facilitated the conduction of randomized controlled trials in this population. Three recent clinical trials investigated the efficacy of novel oral anticoagulants (NOACs) for secondary prevention of recurrent ESUS,,, but anticoagulation was not found to be superior to antiplatelet therapy.
Although CHADS2 and CHA2DS2-VASc scores are initially developed to estimate the risk of stroke in patients with known AF,,, recent studies show that they have a strong predictive value for future diagnosis of AF and IS or transient ischemic attack (TIA) recurrence.,,
In this study, we aimed to investigate whether CHADS2 or CHA2DS2-VASc may be used to predict higher risk of ischemic stroke despite antiplatelet therapy in patients with CIS and to identify the patients who require more detailed investigational workup. Our second aim was to assess the risk factor profile of patients with first ever CIS who were already on antiplatelet(s).
Patients diagnosed with first ever CIS according to TOAST classification with complete data for CHADS2 and CHA2DS2-VASc scores calculation were included in this study. We excluded the patients with multiple stroke etiologies. Patient data were retrieved from the prospective hospital-based stroke registry; Istanbul Medical School Stroke Registry (IMSSR), which includes all stroke patients, hospitalized in the Neurology Department between 1996 and 2014. Diagnostic workup includes at least one brain imaging [magnetic resonance imaging (MRI) or computed tomography (CT)], cervical arterial imaging (MR angiography, CT angiography and/or Doppler ultrasonography), transthoracic echocardiography, 12 -lead ECG and standard blood work-up. A 24-hour Holter ECG monitoring was routinely done in patients admitted after 2008 and was performed in patients with suspected cardiac arrhythmia who were admitted before 2008. Screening for vasculitis, hereditary or acquired thrombophilia were performed only in young stroke patients (<50 years old) or in older patients when it was deemed necessary by the stroke neurologist in charge. A local ethics committee application was not made for retrospective collection of observational data (all of the patients were admitted before General Data Protection Regulation enforcement).
CHA2DS2-VASc score and the individual items of the score was calculated for all patients as previously described: 2 points each were assigned for age ≥75 years (A2) and for history of stroke, TIA, or thromboembolism (S2) and 1 point was assigned for each of the following items including congestive heart failure (C), hypertension (H), diabetes mellitus (D), age 65-75 years (A), vascular disease (VA) (defined as previous myocardial infarction, complex aortic plaque, carotid stenosis, and peripheral artery disease) and female sex category (Sc). Two points were given for history of stroke, TIA, and one point for each of the remaining items to calculate CHADS2 score. Clinical features including history of vascular risk factors (hypertension, diabetes, hyperlipidemia, tobacco or alcohol use), congestive heart failure, stroke onset during physical activity or sleep, the National Institutes of Health Stroke Scale (NIHSS) and modified Rankin Scale (mRS) scores on admission, the presence of acute multiple territory infarcts and abnormal sedimentation rate were recorded. CHA2DS2-VASc scores, demographic and clinical characteristics of patients diagnosed with first ever CIS on regular antiplatelet(s) were compared to those without prior antiplatelet therapy.
All patients recorded to the IMSSR are regularly followed-up by the stroke neurologists as outpatients in the IMSS unit. Long-term follow up data on current medication, recurrent stroke and mRS scores were retrieved from the existing outpatient files and accessible patients were interviewed by telephone to complete missing data.
Chi square test, Fisher's exact test and Mann-Whitney test were used for group comparisons where appropriate. Receiver operating characteristics (ROC) curve analysis was performed to analyze the predictive value of CHA2DS2-VASc for recurrent IS in the AP group. SPSS 15 software (SPSS Inc, Chicago, IL, U.S.A.) was used for statistical analysis and the significance level was set at P < 0.05.
Of the 3586 IS patients, 886 patients with first ever IS had complete data for CHADS2 and CHA2DS2-VASc scores calculation. Seventy-five (39 women) of them, with a mean age of 60.6 ± 16.4 years (median 61 years, range 18-98) on admission, were diagnosed with CIS as seen in[Figure 1] showing the flowchart of patient selection of the study. Prolonged cardiac rhythm monitoring could not be done in 17 patients and vascular work-up was done only with Doppler ultrasonography (instead of MRA or CTA) in 16 patients [Figure 1]. Screenings for vasculitis, hereditary or acquired thrombophilia done in 17 young stroke, and 2 older patients were negative.
Nineteen patients were using regular antiplatelet monotherapy prescribed by another outpatient clinic before their qualifying stroke (AP group). Demographic and clinical characteristics of CIS patients, on regular antiplatelets (19 patients) and those not taking any antiplatelet(s) (non-AP group) (56 patients) before the qualifying CIS are presented in [Table 1]. None of them had history of TIA or congestive heart failure. CHA2DS2-VASc scores were significantly higher in the AP group in comparison to the non-AP group. ROC curve analysis showed that a CHA2DS2-VASc score of ≥4 identified the occurrence of qualifying CIS under regular antiplatelet therapy with a sensitivity of 84% and specificity of 50%, P = 0.01. Area under the curve was 0.699 for the CHA2DS2-VASc score (CI 0.57-0.82). Similarly, CHADS2 scores were significantly higher in the AP group (ROC curve analysis showing an AUC of 0.705, with a sensitivity of 86% and a specificity of 37% (P = 0.008) for CHADS2 score ≥3 (CI 0.57-0.84)) [Figure 2].
In the presented study, 46 patients had all diagnostic tests required for ESUS. CHA2DS2-VASc scores were higher in 13 patients with ESUS in the AP group when compared to the remaining 33 patients in the non-AP group (4.5 ± 1.3 versus 3.9 ± 1.5, respectively), without reaching statistical significance. Patients with ESUS in the AP group had significantly higher CHADS2 scores in comparison to those in the non-AP group (3.38 ± 0.6 versus 2.88 ± 0.8, P = 0.047).
Follow-up information was available for 49 patients (15 patients in the AP group and 34 patients in the non-AP group), with a follow-up duration of 4.17 ± 2.1 years (mean ± SD). Antiplatelet monotherapy was given to the patients in the non-AP group and dual antiplatelet therapy was prescribed to the AP group for secondary prevention. Among the 15 CIS patients in the AP group, two had recurrent IS while receiving regular dual antiplatelet therapy. Persistent AF was detected during the evaluation of the recurrent stroke and anticoagulant therapy was started. These two patients' CHADS2 and CHA2DS2-VASc scores were 3 and 4; 4 and 6, respectively. In the non-AP CIS group, further rhythm Holter monitoring of 24 hours (apart from the first cardiac monitoring of 24 hours) could be performed in nine patients during routine follow-up visits. PAF was detected in four of these patients, by rhythm Holter recordings done in the first 2 years following the qualifying stroke. None of them reported any neurological complaints at that timeframe. CHADS2 and CHA2DS2-VASc scores were 4 and 6 respectively, in 3 of these 4 patients; and the fourth patient's CHADS2 and CHA2DS2-VASc scores were 3. Twenty-six patients did not come to routine outpatient visits and could not be reached by phone calls.
Among patients with CIS, occurrence of stroke in the AP group raises a possibility of embolic etiology, which may respond better to anticoagulation. In this study, we found that CHADS2 and CHA2DS2-VASc scores were significantly higher in the AP group in comparison to the non-AP group. ROC curves revealed that the patients with a CHADS2 score of ≥3 and CHA2DS2-VASc Score ≥4, may be at higher risk of stroke despite antiplatelet therapy, suggesting the need for further investigations in this subgroup and consideration of anticoagulant therapy.
Cryptogenic strokes still account for almost 25% of all ischemic stroke patients and recurrence rates are reported as approximately 5% per year.,, The term cryptogenic stroke also includes patients with multiple stroke etiologies or incomplete diagnostic work-up and therefore is not synonymous with ESUS. In this study, we included patients with CIS who had adequate investigational work-up at time of qualifying stroke, excluding those with multiple stroke etiologies. The most frequent secondary prevention used in stroke patients with undetermined cause is antiplatelet therapy. Some patients may have recurrent stroke despite regular use of antiplatelet(s), therefore determining the putative embolic source is important for selecting the optimal secondary prevention treatment. In a recent study, acetylsalicylic acid (ASA) use prior to the qualifying stroke was found to be independently associated with stroke recurrence in ASA-assigned patients with ESUS emphasizing that antiplatelet agents are not sufficient for stroke prevention in some patients. But the NAVIGATE-ESUS and RE-SPECT-ESUS trials exploring the efficacy of NOACs in the patients with ESUS, concluded that oral anticoagulation was not superior to ASA in preventing stroke recurrence in ESUS. The definition of ESUS includes heterogeneous causes which might influence the results of these recent clinical trials showing similar results both for anticoagulation and antiplatelet therapy., AF, left ventricular disease, PFO or cancer are some of the embolic sources, leading to formation of red thrombi which may respond better to anticoagulation. But, atherosclerotic disease also meets the current criteria for ESUS and benefits from ASA. A sub analysis of NAVIGATE-ESUS reported that ASA use prior to the qualifying stroke was more frequent in patients without branch atherosclerotic disease (BAD) in comparison to those with BAD. In addition, subgroup analyses of the same trial revealed a benefit with rivaroxaban in the participants with a left atrial diameter greater than 4.6 cm. Currently, early short term dual antiplatelet therapy is advised for patients with TIA or noncardioembolic stroke and antiplatelets are used for secondary prevention in the majority of patients with CIS, while some authors suggest anticoagulants. But, it is difficult for clinicians to decide the appropriate preventive therapy when the cause of the stroke remains undetermined.
The CHADS2 and CHA2DS2-VASc scores have correlation with various conditions such as coronary and peripheral arterial disease. They are also reported to be predictive of stroke and mortality in AF and non-AF cohorts. A recent study revealed an increased risk of embolic stroke with higher CHA2DS2-VASc score in a large cohort of middle-aged, community-dwelling individuals without AF. The 10-year incidence of embolic stroke increased from 0.15 per 1,000 person-years (for score 0 to 1) to 2.2 per 1,000 person-years (for score ≥4). The authors concluded that there was need for further research to determine whether anticoagulant use in certain individuals without AF but with high CHA2DS2-VASc risk scores would reduce the risk of ischemic stroke. In our study, patients with CIS using antiplatelet therapy prior to qualifying stroke (AP Group) had significantly higher CHADS2 and CHA2DS2-VASc scores. These findings suggest that the etiology was most probably embolic and the patients with higher scores may need anticoagulation for secondary prevention. We found similar results in our patients meeting ESUS criteria: CHADS2 scores were significantly higher and CHA2DS2-VASc scores were higher in the AP group in comparison to non-AP group, without reaching statistical significance which may be related to the sample size. According to these findings, we may suggest that CHADS2 and CHA2DS2-VASc scoring systems may be used to define study groups in the future studies investigating the beneficial effect of anticoagulation over antiplatelets in ESUS.
The patients having CIS are screened for AF with various methods including repeated Holter ECGs or prolonged heart rhythm monitoring, however long-term cardiac monitoring devices are expensive and not tolerated by all patients. Therefore, some authors recommended additional cardiac investigations only in patients with a high probability to develop AF including those with a high virtual CHA2DS2-VASc score. In this study, the repeated rhythm Holter monitoring detected PAF in four patients in the non-AP CIS group, but none of them had any neurological complaints following the qualifying stroke. The causal association between AF detected during long-term follow-up and the index stroke in stroke patients with undetermined cause remains still controversial. The rate of occurrence of subclinical and short-lasting AF was found to be similar among elderly patients with and without history of stroke. Furthermore, a recent study investigating the results of a transesophageal echocardiography registry reported that large aortic arch plaques should also be considered in CIS patients with CHADS2 and CHA2DS2-VASc scores ≥2, in addition to PAF. Our findings show that scoring systems (such as CHADS2 and CHA2DS2-VASc) could have implications for the investigational strategy in patients with CIS for defining further cardiac work-up candidates.
CHADS2 and CHA2DS2-VASc scores incorporate most of the risk factors for stroke occurrence. In the presented study, vascular disease, diabetes, hypertension and hyperlipidemia were more frequent in the AP group, who had significantly higher CHADS2 and CHA2DS2-VASc scores in comparison to the non-AP group. A recent study analyzing possible ESUS patient subgroups showed that younger patients with ESUS had a lower prevalence of dyslipidemia, lower median value of CHA2DS2-VASc and lower prevalence of left atrial enlargement, but a higher proportion of PFO. The same study also reported evidence of non-significant atherosclerosis, and/or cardioembolic source in older patients. In our study, the AP group with higher CHADS2 and CHA2DS2-VASc scores were significantly older in comparison to the non-AP group. This finding may suggest that cardioembolic source of embolism (CSE) was a more probable etiology in this subgroup. Multiple territorial infarcts are reported to be a radiological imaging-pattern associated with cardioembolism.,, In our study, acute infarcts in multiple arterial territories were more frequently observed in the AP group, further strengthening the above-mentioned point suggesting the presence of CSE.
Our study has limitations such as the retrospective design, small sample size and limited follow-up data. Twenty-six patients did not come to the clinic for routine visits after discharge, nor were they reachable by phone. Some of these patients may have been evaluated in another health center because of recurrent stroke. Only two of all the 49 patients maintained exclusively on antiplatelet therapy, developed stroke during the follow-up period showing a rather low incidence of recurrence. But, the lack of follow-up data on one third of our cohort precludes a definite comment on secondary prevention of CIS.
Secondly, patients registered before 2008 did not meet ESUS criteria due to incomplete diagnostic work-up. However, the study sample was a highly selected group of patients retrieved from a large prospective single-center registry of 4466 patients, with long-term follow-up in the same institution.
There is a need for additional tools to identify subgroups of patients with CIS for individualizing the diagnostic approach, planning workup and secondary prevention treatment. Our results support that CHADS2 and CHA2DS2-VASc scores may be used for this purpose. It seems reasonable to propose that future clinical trials investigating the efficacy of oral anticoagulation over antiplatelets in patients with ESUS, should preferentially include patients with a strong likelihood of embolic source, identified by predictive scoring tools.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2]