Downloaded from http://jnnp.bmj.com/ on February 6, 2016 - Published by group.bmj.com EDITORIAL 617 EDITORIAL than in western white populations, but ....................................................................................... this remains to be clarified. Only 6% of all particles (even the smallest) injected into the carotid arteries in an experi- What causes lacunar stroke? mental model ended up in the lenticu- lostriate arteries, the rest going to the J M Wardlaw cortical arteries or their cortical branches.9 Studies that suggested stron- ................................................................................... ger associations between lacunar stroke and emboli cited carotid stenoses as mild as 25%16 or cardiac abnormalities not clearly associated with embolism quarter of all ischaemic strokes type of ischaemic stroke. Finally, some (for example, left ventricular hypertro- (a fifth of all strokes) are lacunar classifications, such as the Trial of Org phy),13 or they had no, or an inappropri- type.1 Lacunar infarcts are small 10172 in Acute Stroke Treatment ate control group. It would certainly be A 8 infarcts (2–20 mm in diameter) in the (TOAST) method actually use risk useful to be able to infer the likely deep cerebral white matter, basal gang- factors (such as embolic sources or underlying mechanism of lacunar lia, or pons, presumed to result from the hypertension) to determine the stroke ischaemic stroke (that is, to identify occlusion of a single small perforating type, thus potentially biasing studies of the 10–15% of embolic/stenotic versus artery supplying the subcortical areas of differences in risk factors. Hence, inad- other intrinsic small vessel strokes) the brain.2 Although a recognised stroke vertent misdiagnosis of lacunar as cor- from the appearance of the brain lesion, subtype for over 50 years, the cause tical stroke, and vice versa, the paucity as that might help target effective of lacunar ischaemic stroke,2 and whe- of pathological material, and bias in secondary prevention regimens, but ther it is different to cortical ischae- some clinical classification systems may much more work is required to see mic stroke, remains under debate.34 have confounded previous pathology, whether different patterns actually Furthermore, lacunar stroke is not prognosis, and risk factor studies. exist, before determining how closely benign; 30% of patients are left depen- these relate to the underlying mechan- dent,5 and scant long term data suggest ism. IS THE CAUSE OF LACUNAR that up to 25% of patients have a second Hypertension and diabetes are said to DIFFERENT FROM CORTICAL stroke within 5 years.6 Therefore, the be strongly associated with lacunar ? prevention and treatment of this com- ISCHAEMIC STROKE ischaemic stroke. However, in studies The lacunar hypothesis supports the mon stroke subtype may be less than using risk factor free ischaemic stroke concept that lacunar ischaemic stroke ideal. subtype definitions, there was only a is due to an intrinsic cerebral small marginal excess of hypertension with arteriolar abnormality,2 in contrast to lacunar versus non-lacunar infarction PROBLEMS IN THE STUDY OF cortical ischaemic stroke, which is com- (RR 1.11; 95% CI 1.04 to 1.19) and no monly due to embolism from the heart LACUNAR STROKE difference for diabetes (RR 0.95; 95% CI Several factors have hampered the study or large arteries. Although some studies 0.83 to 1.09).10 Nor was there clear of lacunar stroke. Firstly, few patients suggest that many lacunar strokes are evidence of any association between die from lacunar stroke; if they do, caused by emboli, and while it is smoking, prior transient ischaemic death may occur long after the stroke, perfectly possible for a small embolus attack, excess alcohol consumption, or making autopsy material scant and to enter and occlude a lenticulostriate raised cholesterol in lacunar compared difficult to interpret, and the small artery,9 a systematic review of risk with non-lacunar infarction.10 cerebral vessels require meticulous dis- factors including only studies using section. Studies of risk factors and subtype definitions for ischaemic stroke causation have predominantly used a free of risk factors found that atrial IS THERE OTHER EVIDENCE OF A clinical diagnosis of stroke type, prob- fibrillation and carotid stenosis were DIFFERENT MECHANISM IN ably resulting in some misclassification. associated more with non-lacunar than LACUNAR ISCHAEMIC STROKE? Although lacunar infarcts are associated lacunar infarction (relative risk (RR) of After lacunar ischaemic stroke, a recur- with specific neurological syndromes, lacunar versus non-lacunar infarction: rence is more likely to be lacunar then and most patients with a clinical lacu- atrial fibrillation 0.51, 95% confidence cortical; 47% of recurrences after a nar syndrome have a small deep sub- interval (CI) 0.42 to 0.62; ipsilateral lacunar stroke were lacunar compared cortical infarct on brain imaging (if carotid stenosis 0.35, 95% CI 0.28 to with 15% of recurrences after a non- visible), 10–20% actually have a recent 0.44).10 Indeed, common causes of large lacunar stroke).19 If lacunar and cortical small cortical infarct in a location that artery (cortical) infarction, such as ischaemic strokes were due to the same explains their stroke presentation.7 emboli from the large arteries or mechanisms, then recurrent strokes Similarly, 10–20% of patients with a heart,11–13 or intracranial large artery would be equally likely to be cortical clinical mild cortical stroke actually atheromatous stenoses,12 appear un- after a lacunar stroke as lacunar, which have a recent relevant lacunar infarct likely to cause more than 10–15% of appears not to be the case. on imaging.7 Epidemiologically, these lacunar strokes.14–17 Perhaps lacunar Lacunar ischaemic stroke also appears patients behave more like the lesion infarcts due to emboli or middle cerebral to be more closely associated with white type on imaging than the clinical syn- artery stenoses are recognisable by being matter lesions (WMLs) than does cor- drome of the lesion they actually have.7 larger than non-embolic/stenotic tical ischaemic stroke. WMLs are abnor- Many studies have used an inappropri- lacunes,12 18 possibly because the embo- mal areas of hypodensity (on computed ate control group; age matched normal lus/stenosis occluded the origin of tomography scans) or hyperintensity controls can only indicate whether or several lenticulostriate arterioles simul- (on T2 weighted magnetic resonance not a particular risk factor is associated taneously. There is a suggestion that imaging (MRI)) in the deep hemi- with stroke, whereas identification of lacunar stroke secondary to middle spheric and periventricular white matter associations specific to lacunar stroke cerebral artery stenosis may be more and brain stem.20 They are in turn requires a control group with a different common in south Asian populations associated with cognitive decline,21 and www.jnnp.com Downloaded from http://jnnp.bmj.com/ on February 6, 2016 - Published by group.bmj.com 618 EDITORIAL increased risk of future stroke, particu- weak (RR 1.11; 95% CI 1.04 to 1.19).10 with microhaemorrhages (small blood larly lacunar type.22 WMLs also progress Thickened vessels may narrow or leaks).27 rapidly after lacunar stroke.23 After occlude the lumen, leading to ischae- Is there other evidence of endothelial lacunar stroke, 15–20% develop demen- mia, but relatively few truly occluded failure? Patients with isolated lacunar tia (more than might reasonably be vessels have been seen pathologi- infarction, or lacunar infarction plus attributed to the amount of brain cally.29 30 Thickened vessel walls may be WMLs, have elevated systemic plasma damage caused by a lacunar stiff and impair autoregulation, and markers of endothelial activation infarct).62124 After lacunar ischaemic indeed patients with WMLs may have (plasma intercellular adhesion molecule stroke, new ‘‘silent lacunar infarcts’’ impaired autoregulation.34 Some studies 1, thrombomodulin, and tissue factor occur on follow up imaging.25 found reduced cerebral blood flow pathway inhibitor) compared with age Asymptomatic small deep white matter (CBF) in patients with WMLs,35 but matched normal controls.43 However, in infarcts, in addition to the symptomatic others did not.36 CBF is difficult to the absence of non-lacunar controls, lesion, have been seen on MR diffusion quantify: ‘‘reduced’’ CBF may be arte- it is unclear if these patterns are specific imaging at presentation with lacunar factual,37 or simply the consequence of a to lacunar stroke. Several studies ischaemic stroke.26 The imaging appear- reduction in normal tissue to supply. have identified retinal microvascular ance of these asymptomatic lacunar None of this explains what starts the abnormalities that were associated with infarcts suggests that they are slightly vessel abnormality, only the features or increased risk of stroke, cognitive older than the presenting lesion. behaviour of the vessels and the possible impairment, and white matter lesions. Cerebral microhaemorrhages, which mechanisms for damaging the brain, Those most closely related to these are tiny, apparently asymptomatic once the abnormality is established. cerebral abnormalities (microaneur- bleeds (or ‘‘leaks’’) in the brain are also ysms, retinal haemorrhages, and soft associated with lacunar stroke and exudates) were most commonly seen WMLs.27 These observations suggest IF NOT OCCLUSION AND when there was