We have shown that among patients presenting with acute infarction on DWI, those with stroke symptoms that resolve within 24 hours have smaller leukoaraiosis volume compared to patients with persistent symptoms indicating an ischemic stroke. The median normalized leukoaraiosis volume was approximately 3 times higher in patients with ischemic stroke as compared to TSI. In addition to leukoaraiosis volume, infarct size and infarct location also differed between TSI and ischemic stroke; the probability of ischemic stroke increased with rising infarct volume. Infarcts that included critical structures with compact anatomic organization such as the brainstem or in locations where typical lacunar infarcts occurred were also more likely to be associated with ischemic stroke.
While infarct size and location are important, they together explain only half of the variability in functional outcome in patients with ischemic stroke.24
Our data suggest that the brain's intrinsic capacity to recover from ischemic brain injury may be in part determined by the extent of leukoaraiosis, which is also known to contribute to the variability in outcome.6–9
Pathology findings in leukoaraiosis include, among other things, axonal changes ranging from mild demyelination to severe axonal disruption.25
The severity of these pathologic changes correlates with the extent of leukoaraiosis on MRI.26
Published evidence from studies using transcranial magnetic stimulation, fMRI, and diffusion tensor imaging suggests that recruitment and reorganization of ipsilesional and contralesional brain regions during poststroke recovery requires the presence of intact connections between different parts of the brain.27–29
An in silico model of activity-dependent plasticity has supported this view, showing that axonal dysfunction caused by blocking of the propagation of action potentials between neurons results in deficits in pruning and recovery of functional synapses and this in turn produces a negative impact on adaptive plasticity that is much greater than the impact of gray matter injury.30
The view that leukoaraiosis impairs the brain's ability to compensate for the lost function is supported by multiple recent studies indicating that the extent of leukoaraiosis correlates with poor functional outcome and low quality of life after ischemic stroke.6–8
We have previously shown that leukoaraiosis volume is a predictor of modified Rankin Scale score at 6 months following ischemic stroke after controlling for age, infarct volume, etiologic stroke mechanism, initial stroke severity, and preventive stroke treatment; the highest quartile of leukoaraiosis volume is associated with a 1.5-fold higher modified Rankin Scale score as compared with the lowest quartile.9
The current data offer additional evidence that leukoaraiosis volume is a marker for the brain's capacity for rapid and complete recovery of the lost function.
It is notable that clinical recovery in TSI is a strikingly rapid process. In the present dataset, 46% of TSIs lasted less than 1 hour and another 26% lasted between 1 and 2 hours. The majority of TSI events lasting less than 1 hour lasted often only 1 to 10 minutes (66%). It is thus plausible to presume that most small infarcts (i.e., those within the region of overlap in ) either recover completely and very rapidly often within minutes or recover gradually and often only partially in longer than 24 hours. This tendency toward bimodal distribution of symptom duration in small infarcts suggests that there is a threshold in the brain's capacity to regain the lost function, which, when exceeded, symptoms that would otherwise last only for a few minutes could persist for more than 24 hours. The association between ischemic stroke and increasing leukoaraiosis volume may suggest that leukoaraiosis impairs the brain's reserve capacity so that small infarcts could easily overcome the threshold to result in lasting ischemic stroke symptomatology.
Leukoaraiosis is not only a predictor of functional outcome but also has been reported to predict future episodes of stroke in individuals with or without previous stroke.31–33
It has been suggested that leukoaraiosis reflects an increased burden of stroke risk factors and therefore identifies individuals at high risk of stroke.31
Data from the Atherosclerosis Risk in Communities study, on the other hand, challenges this view by showing that the association between leukoaraiosis and stroke risk is independent of conventional stroke risk factors such as hypertension, diabetes, and cigarette smoking.33
Our findings may shed light into the mechanism of increased stroke risk in individuals with leukoaraiosis by suggesting that leukoaraiosis-mediated impaired capacity to compensate for injury results in occurrence of symptomatic stroke in an event of acute infarction. This view is further supported by evidence indicating that leukoaraiosis confers an increased perioperative symptomatic stroke risk after procedures that are known to be associated with high rate of brain embolism such as carotid artery stenting,34
intraoperative shunt placement during carotid endarterectomy,35
and total aortic arch replacement.36
Future research examining leukoaraiosis load-by-treatment efficacy interaction in conditions associated with small brain infarcts such as those occurring following carotid endarterectomy and stenting might demonstrate practical utility of the current findings. Leukoaraiosis information could also be used to generate predictive algorithms (along with other independent predictors of clinical outcome) to optimize benefit and minimize risk from thrombolytic therapy. Such models might identify individuals with high likelihood of early spontaneous recovery in whom withholding thrombolytic therapy could be justified. In contrast, in patients with extensive leukoaraiosis, the high risk of intracranial hemorrhage might outweigh the benefit by thrombolysis that accrues from reduced risk of developing disabling permanent deficit.37
Our findings are subject to a number of limitations. First, time-based categorization of clinical status (TSI vs ischemic stroke) might have introduced ascertainment bias in subjects with small infarcts who had symptoms lasting for around 24 hours. Nevertheless, the observed association between leukoaraiosis and clinical phenotype was strong and persisted in sensitivity analyses in subsets with symptoms that lasted for <1 hour vs >24 hours. Second, although we stratified infarct location into anatomically distinct and clinically important categories, our approach did not take into account the regional differences within each category. A more detailed segmentation of the brain employing a voxel-by-voxel approach might provide more accurate estimation of the impact of location on clinical phenotype, and highlight specific regions of the brain where the impact of leukoaraiosis on persistence of symptoms might be highest. Such methods demand large datasets for adequate sampling of each brain compartment. Finally, patients who had MRI with motion artifacts were excluded from the study. This, however, did not appear to have caused a significant selection bias as none of the baseline clinical features listed in were significantly different between the study population and the excluded patients.
These data show that rapid functional recovery after small infarcts appears to be associated with integrity of white matter tracts as manifested by the amount of leukoaraiosis detected by MRI. The finding that individuals with increasing leukoaraiosis are at higher risk of developing ischemic stroke (rather than TSI) supports further studies to uncover the added value of baseline leukoaraiosis burden information in patient selection for clinical stroke trials.