Optimizing the Definitions of Stroke, Transient Ischemic Attack, and Infarction for Research and Application in Clinical Practice

Abstract

Background and purpose: Until now, stroke and transient ischemic attack (TIA) have been clinically based terms which describe the presence and duration of characteristic neurological deficits attributable to intrinsic disorders of particular arteries supplying the brain, retina, or (sometimes) the spinal cord. Further, infarction has been pathologically defined as death of neural tissue due to reduced blood supply. Recently, it has been proposed we shift to definitions of stroke and TIA determined by neuroimaging results alone and that neuroimaging findings be equated with infarction.

Methods: We examined the scientific validity and clinical implications of these proposals using the existing published literature and our own experience in research and clinical practice.

Results: We found that the proposals to change to imaging-dominant definitions, as published, are ambiguous and inconsistent. Therefore, they cannot provide the standardization required in research or its application in clinical practice. Further, we found that the proposals are scientifically incorrect because neuroimaging findings do not always correlate with the clinical status or the presence of infarction. In addition, we found that attempts to use the proposals are disrupting research, are otherwise clinically unhelpful and do not solve the problems they were proposed to solve.

Conclusion: We advise that the proposals must not be accepted. In particular, we explain why the clinical focus of the definitions of stroke and TIA should be retained with continued sub-classification of these syndromes depending neuroimaging results (with or without other information) and that infarction should remain a pathological term. We outline ways the established clinically based definitions of stroke and TIA, and use of them, may be improved to encourage better patient outcomes in the modern era.

Keywords: asymptomatic carotid stenosis; infarction; public health practice; stroke; transient ischaemic attack.

Figure 1

Continued fall in the rate of ipsilateral stroke in patients with >50% carotid stenosis given medical treatment alone since 2007. Sixty-seven percent relative (or 1.7% absolute) fall in the reported average annual ipsilateral stroke rate in patients with >50% ACS given medical intervention alone from 1985 to 2013. Black diamonds are study results with corresponding sample sizes. As in 2009, the Ryan-Holm stepdown Bonferroni correction was made for multiple comparisons (converting raw P values to P′ values, SYSTAT 13, SYSTAT Software Inc.) (3). UPL and LPL, respectively, upper and lower 95% prediction limits for new population rate estimates; UCL and LCL (dashed lines), respectively = upper and lower 95% confidence limits for the population regression line; WRL, weighted regression line; ACAS, Asymptomatic Carotid Atherosclerosis Study (112); CREST1, Carotid Revascularization Endarterectomy versus Stenting Trial (113). ** indicates two studies including a small minority with remote ipsilateral stroke/transient ischemic attack at baseline (17, 108). These were included in this analysis because the regression line was not very different with (y = 2.861 – 0.061x, P′ for slope and y intercept <0.00000 and r² = 0.379) or without them (y = 2.720 – 0.051 x, P′ for slope and y intercept <0.00000 and r² = 0.275). Both analyses showed a statistically significant fall in stroke risk from 1985 to 2013 although it was a little less in magnitude when the studies by Markus et al. (17) and Madani et al. (108) were excluded (57% relative and 1.4% absolute fall in rate). The outlier result (white diamond) (106) was not included in these analyses (see text). Abbreviations: CEA, carotid endarterectomy; CAS, carotid stenting.

Figure 2

Diagnostic and classification framework for syndromes due to focal, intrinsic, central-neurovascular compromise*^,**^,#. *These are the sequential layers of information that should be specified as far as possible when describing syndromes due to focal, intrinsic central-neurovascular compromise. These are syndromes referable to intrinsic disorders of particular blood vessels supporting the brain, retina, or spinal cord. This classification may be expanded when there is reason. As far as possible, qualifying terms such as “likely” or “confirmed” should be included to improve clarity. **This framework recognizes that the presence of a condition does not always mean causation (124). The mechanism of stroke or transient ischemic attack (TIA) in an individual is always about assigning probabilities (0–100%). Therefore, it is more accurate to speak of risk factors and their likeliness of contributing causation. ^#This is a classification of clinical syndromes (symptoms and/or signs). Asymptomatic neuroimaging findings should be classified separately and according to whether or not pathological characterization has occurred. Pathologically unconfirmed asymptomatic neuroimaging findings, such as suspected asymptomatic infarcts, should be classified as such with the reasons for suspecting a particular pathology. Asymptomatic neuroimaging findings can be sub-classified using the same principles as presented in Figure 2 where relevant, when known and when there is purpose. ⁺The most commonly recognized “characteristic” “focal” clinical deficits associated with stroke and TIA include sudden facial weakness, weakness or incoordination involving one or more limbs, dysphasia, dysarthria, visual impairment, vestibular or cranial nerve dysfunction which are manifested according to the particular vascular territory involved and neural tissue being secondarily compromised (7, 12, 13, 15). ⁺⁺For computed tomography, these include hypo-attenuation relative to “normal” areas, loss of white/gray matter differentiation, focal cortical swelling, hyper-dense arterial signs, and/or compression of adjacent structures (25, 26). For magnetic resonance imaging these include diffusion-weighted, T2, and fluid attenuation inversion recovery hyperintensity (26, 39, 46). ^{^}TIA is rapidly developed clinical symptoms and/or signs of cerebral, retinal, or spinal cord dysfunction lasting <24 h, with no apparent cause other than of focal neurovascular origin where resolution is swift and leaves no detectable permanent neurologic deficit. It is recognized that transient ischemic attacks commonly last 2–15 min. Adapted from the National Institutes of Health 1975 publication (12). ^{^^}Stroke is rapidly developed clinical symptoms and/or signs of cerebral, retinal, or spinal cord dysfunction lasting >24 h or leading to death, with no apparent cause other than of focal neurovascular origin. It is recognized that the deficits typically appear suddenly but may progress or fluctuate (not resolving), particularly over minutes to hours after onset. Adapted from Aho et al. (7). ^{^^^}Infarction is a pathological term used to describe tissue that has lost its blood supply for long enough to undergo ischemic necrosis (death) with characteristic macroscopic and microscopic findings (21). ^##Severity may be measured according to meaningful scales such as the modified Rankin (for activities of daily living) (125). “Fully resolved, mild, moderate, severe, or fatal” should be used in preference to “disabling” versus “non-disabling” to describe severity because even mild strokes are associated with disability unless all measurable deficits have resolved.