New Approach to Stroke Subtyping: The A-S-C-O (Phenotypic) Classification of StrokeAmarenco P.a · Bogousslavsky J.b · Caplan L.R.c · Donnan G.A.d · Hennerici M.G.e
aDepartment of Neurology and Stroke Center, INSERM U-698 and Paris-Diderot University, Bichat University Hospital, Paris, France; bDepartment of Neurology, Genolier Swiss Medical Network, Valmont-Genolier, Glion-sur-Montreux, Switzerland; cDivision of Cerebrovascular/Stroke, Beth Israel Deaconess Medical Center, Boston, Mass., USA; dNational Stroke Research Institute, Austin Health, University of Melbourne, Melbourne, Vic., Australia; eDepartment of Neurology, University of Heidelberg, Universitätsklinikum Mannheim, Mannheim, Germany Corresponding Author
We now propose a new approach to stroke subtyping. The concept is to introduce a complete ‘stroke phenotyping’ classification (i.e. stroke etiology and the presence of all underlying diseases, divided by grade of severity) as distinguished from past classifications that subtype strokes by characterizing only the most likely cause(s) of stroke. In this phenotype-based classification, every patient is characterized by A-S-C-O: A for atherosclerosis, S for small vessel disease, C for cardiac source, O for other cause. Each of the 4 phenotypes is graded 1, 2, or 3. One for ‘definitely a potential cause of the index stroke’, 2 for ‘causality uncertain’, 3 for ‘unlikely a direct cause of the index stroke (but disease is present)’. When the disease is completely absent, the grade is 0; when grading is not possible due to insufficient work-up, the grade is 9. For example, a patient with a 70% ipsilateral symptomatic stenosis, leukoaraiosis, atrial fibrillation, and platelet count of 700,000/mm3 would be classified as A1-S3-C1-O3. The same patient with a 70% ipsilateral stenosis, no brain imaging, normal ECG, and normal cardiac imaging would be identified as A1-S9-C0-O3. By introducing the ‘level of diagnostic evidence’, this classification recognizes the completeness, the quality, and the timing of the evaluation to grade the underlying diseases. Diagnostic evidence is graded in levels A, B, or C: A for direct demonstration by gold-standard diagnostic tests or criteria, B for indirect evidence or less sensitive or specific tests or criteria, and C for weak evidence in the absence of specific tests or criteria. With this new way of classifying patients, no information is neglected when the diagnosis is made, treatment can be adapted to the observed phenotypes and the most likely etiology (e.g. grade 1 in 1 of the 4 A-S-C-O phenotypes), and analyses in clinical research can be based on 1 of the 4 phenotypes (e.g. for genetic analysis purpose), while clinical trials can focus on 1 or several of these 4 phenotypes (e.g. focus on patients A1-A2-A3).
© 2009 S. Karger AG, Basel
This classification system defines clinically and pathogenically meaningful groups, while losing as little information as possible. It is factual and not interpretative, and based on the minimum number of necessary investigations and a timely work-up to make classification possible. To enable both researchers to use this classification and peer-reviewers of scientific journals to evaluate submitted manuscripts, the level of evidence required for a given patient to be classified into a group must be taken into account.
This new classification system recognizes that many patients belong to several categories; some categories may be causally related to the index stroke, whereas others are simply concurrent. By introducing the ‘level of diagnostic evidence’, this classification recognizes the completeness, the quality, and the timing of the evaluation to grade the underlying diseases.
The methods of this classification are clearly laid out in 1, 2, 3. Patients are evaluated for 4 predefined phenotypes, i.e. atherosclerosis (A), small vessel disease (S), cardiac disease (C), and other causes (O).
|Table 1. Methods of classification|
|Table 2. Grades of pathology (phenotyping)|
|Table 3. Levels of diagnostic evidence|
The first step is to ‘grade’ every patient in each of the 4 main ischemic groups (i.e. atherothrombotic, cardioembolic, small-vessel disease, and other causes) as we do in everyday clinical practice (table 2; fig. 1).
|Fig. 1. Example of the grading system (phenotyping) in 1 patient with an ipsilateral carotid stenosis >70% (atherothrombosis, grade 1) and atrial fibrillation (cardioembolism, grade 1), leukoaraiosis and microbleeds (small vessel disease, grade 3), and platelet count of 700,000/mm3 (other causes, grade 3). Stroke subtype is A1-S3-C1-03.|
Three grades of likelihood are considered (table 1):
• Grade 1: definitely a potential cause of the index stroke;
• Grade 2: causality uncertain;
• Grade 3: unlikely a direct cause of the index stroke (but disease is present).
In addition, when no disease is present patients are graded 0; patients who cannot be graded because no tests were performed are graded 9.
Three levels are added for the diagnostic instruments used:
• Level A: direct demonstration by gold-standard diagnostic tests or criteria;
• Level B: indirect evidence or less sensitive or specific tests or criteria;
• Level C: weak evidence in the absence of specific tests or criteria.
Every patient is identified by A-S-C-O: for example a patient with a 70% ipsilateral symptomatic stenosis, leukoaraiosis, atrial fibrillation and platelet count of 700,000/mm3 will be classified A1-S3-C1-O3 (fig. 1). Given the criteria and definitions in table 2, an electronic algorithm can be developed for bedside use (available on www.asco-classification.org).
The next step is optional. According to the specific A-S-C-O pattern, the patient may have only 1 potential cause (corresponding to grade 1 in 1 of the 4 A-S-C-O phenotypes) or may have several potential causes (coexisting cause group), provided this is based on A or B evidence diagnostic test.
When none of the A-S-C-O are 1 (A or B evidence), the cause is unknown, but underlying pathologies may exist (e.g. A2 or A3, or S2 or S3, or C2 or C3, or O2 or O3).
In cases where all A-S-C-O are 0, the cause is completely unknown.
When the patient cannot be classified into 1 of the 4 categories because of insufficient work-up, this stroke patient should not be subtyped (unclassifiable group, graded 9); as an example, a patient with carotid stenosis ≥70% on the symptomatic side who was not evaluated for cardiac source of embolism and had no brain imaging would be identified by A1-S9-C9-O0; if the same patient had an atrial fibrillation on ECG but no cardiac imaging performed, he would be identified by A1-S9-C1-O0; if the same patient had normal ECG but no cardiac imaging performed, he would be classified A1-S9-C9-O0. The higher the proportion of subjects with grade 9, the smaller the overall quality of the study.
Compared to the classifications reviewed in the companion paper , the main advantages of this new stroke subtyping system are that:
• It follows daily clinical practice: every day at the bedside, we ‘grade’ or estimate the presence of atherosclerotic disease, small vessel disease, cardiac disease, and other causes of stroke (table 2; fig. 1) for a given patient in order to finally decide which cause(s) is (are) the most likely;
• It retains the best available information, recognizing the overlap between the 4 main categories (fig. 1);
• It informs us about the level of diagnostic evidence (table 3);
• It is highly flexible for use in different purposes including meta-analysis. For example, in genetic studies, phenotyping must be very precise; by using this classification, the researcher can clearly state whether they are addressing small vessel disease grade 1–3 or only small vessel disease grade 1. In clinical trials, this classification better defines the selected population for recruitment; for example, like in most recent antiplatelet trials, the patients recruited would be patients with atherothrombotic stroke grades A1A2A3 in the present classification, small vessel disease grades S1S2S3, C2C3, O3, A0-S0-C0-O0 and with exclusion of patients with cardioembolic stroke grade C1 and other causes grade O1 and O2.
In a manner similar to the success of the classification system for headache put forward by the International Headache Society, the approach chosen in this new stroke classification provides a definition for each category and allows both clinicians and researchers to use a common terminology.
The authors thank the International Expert Opinion Committee who reviewed the early draft of the classification and gave valuable comments and suggestions. The views of the article are not necessarily those of the experts listed: H.P. Adams, G.W. Albers, O. Benavente, J. Broderick, H. Chabriat, A. Chamoro, H.C. Diener, J. Ferro, L.B. Goldstein, W. Hacke, G.J. Hankey, D.F. Hanley, A. Hillis, S.C. Johnston, M. Kaste, S.J. Kittner, G.L. Lenzi, S.R. Levine, D. Leys, J.P. Mohr, H. Markus, J.L. Mas, B. Norrving, J.M. Orgogozo, P.M. Rothwell, R.L. Sacco, J.L. Saver, S. Uchiyama, S. Warach, R.J. Wityk, L.K.S. Wong.
Prof. Pierre Amarenco, MD
Department of Neurology and Stroke Centre, Bichat University Hospital
46, rue Henri Huchard
FR–75018 Paris (France)
Tel. +33 1 4025 8726, Fax +33 1 4025 7198, E-Mail firstname.lastname@example.org
Received: February 2, 2009
Accepted: February 2, 2009
Published online: April 3, 2009
Number of Print Pages : 7
Number of Figures : 1, Number of Tables : 3, Number of References : 1
Vol. 27, No. 5, Year 2009 (Cover Date: April 2009)
Journal Editor: Hennerici M.G. (Mannheim)
ISSN: 1015-9770 (Print), eISSN: 1421-9786 (Online)
For additional information: http://www.karger.com/CED