Luciana Catanese, MD
Stroke continues to be a leading cause of disability worldwide. Indeed, a large number of patients who have suffered an ischemic stroke are unable to return to their homes, requiring disposition to acute rehabilitation facilities. The imaging predictors of good outcomes in acute ischemic stroke (AIS) patients undergoing convalescent rehabilitation are yet to be determined.
In this issue of Stroke, Senda et al. investigated the factors influencing convalescent rehabilitation outcomes in AIS patients focusing on imaging parameters such as degree and type of leukoaraiosis and stroke subtypes. Retrospective data from consecutive AIS subjects, who were hospitalized in a single rehabilitation center in Japan from January 2008 to December 2013, was included. The enrolled individuals were right-handed, completely independent (mRS 0 and Barthel Index of 100) at baseline, had an available brain MRI/MRA, were not discharged prematurely and did not carry a diagnosis of dementia. The Functional Independence Measure (FIM) score both on admission and at discharge was used as the measure of functional outcome. The authors classified stroke subtypes by etiology including lacunar infarcts (LI), atherothrombotic (AT), artery-artery embolism (A-to-A), cardioembolic (CE), undetermined embolism (UN) and “other”, if a specific mechanism such as vasculitis or postoperative stroke was suspected. Baseline demographic data was obtained from the initial hospitalization records. White matter lesions were graded on T2 or FLAIR sequences as periventricular (PV) or deep white matter hyperintensity (DWMH) using the Fazekas scale. Finally, the presence of ≥ 50% stenosis in the intracranial cerebral vasculature was considered “stenosis positive”. A stepwise multiple regression analysis was performed to characterize the relationships between the independent variables (total, cognitive and motor FIM scores at baseline, prior medications, age, sex, history of stroke, heart disease and tobacco, side of stroke, presence of unilateral or bilateral strokes, PVH and DWMH grade, atrial fibrillation and presence of significant stenosis on MRA) and the total, motor and cognitive FIM scores on both admission and at discharge. UN and other ischemic stroke groups were excluded from this analysis.
A total of 520 patients (317 men, 203 women; mean age 72.8 years) were included. The average number of days from symptom onset to transfer to rehabilitation was 32.8 +/- 10.3 days and the mean inpatient rehabilitation time was 112 +/- 20 days. Overall, hypertension was overrepresented in the LI group, diabetes in the A-to-A group and higher grades of PHV and DWMH in the LI and A-to-A group. As expected, positive arterial stenosis was more frequently seen in the A-to-A group. There was no significant difference in FIM scores among stroke subtypes. When looking at the total FIM scores at discharge, higher grades of PVH, age and the FIM scores on admission were strongly associated with worse outcomes. In the subgroup analysis of individual stroke subtypes, history of stroke was a significant predictor of worse rehabilitation outcomes in the A-to-A group. In regard to FIM motor scores, PVH grade, presence of bilateral ischemic strokes and history of stroke were factors significantly associated with worse outcomes. In the subgroup analysis, the presence of significant stenosis and PVH grade were predictors of poor outcomes in the AT and A-to-A group, respectively. Finally, age and hypertension medications correlated with worse outcomes as measured by the FIM cognitive scores. In the stroke subtypes analysis, history of tobacco and DWMH grade were significantly associated with worse outcomes in the CE and A-to-A groups.
Overall, the degree of leukoaraiosis on MRI as measured by the Fazekas scale appears to strongly correlate with worse functional outcomes in convalescent acute ischemic stroke patients. This data is not only in line with available literature such as that shown in the Leukoaraiosis and disability (LADIS) cohort study, but it also supports the usefulness of MRI when estimating functional outcomes. In addition to that, the authors found that PVH grade appeared to be associated with motor-function outcomes whereas DWMH correlated with cognitive function outcome. This study is not without limitations, the retrospective and single-center design, multiple subgroup analysis without correction for multiple testing and potential for residual confounding – to name a few – limit the interpretation of the findings and highlight the need for long-awaited prospective trials on the subject. Predicting functional outcomes in the convalescent rehabilitation period is of utmost importance for our stroke patients, and systems resource management.