| Background and PurposeStroke is one of the most common causes of disability and death in adults. In China, the prevalence of stroke increases year by year, which has a significant impact on Chinese economy. Majority of stroke patients may miss the opportunity of intravenous thrombolytic therapy due to the limited time window. Therefore, finding effective therapy and intervention beyond thrombolytic time window is necessary. Further understanding the relationship between pathophysiological change of stroke outside the time window and prognosis, can help clinician to take feasible treatment and effective rehabilitation.After stroke, the ischemic areas in some patients may have increased perfusion through some regulatory mechanisms, which may be beneficial for clinical outcome. However, it is still be controversial because it would increase the risk of hemorrhage. Stroke may also affect the blood brain barrier (BBB) beyond the time window, resulting in increased BBB permeability or BBB disruption. BBB disruption has been found to be a predictor of intracranial hemorrhage and edema, so it is harmful for stroke prognosis. With increasing time corse of stroke, abnormality of distal white matter (WM) connecting with stroke lesion could be caused by anterograde degeneration, which affects stroke recovery.Diabetes mellitus is an important risk factor for stroke prognosis, because it would stimulate the release of matrix metalloproteinases (MMP), promoting BBB breakdown. Therefore, BBB disruption should be commoner in diabetic stroke. On the other hand, it has been reported that diabetes can injury WM tracts in diabetic patients without stroke. The damaged WM tracts are related with clinical symptoms. If diabetes affects some WM tracts related with neurological deficits before stroke, the injuryed WM tracts may involve in the acute stage of stroke symptoms and influence stroke outcome.NIHSS is widely used in evaluating stroke severity in the acute stage in clinical work, but it cannot provide the pathophysiological information. MRI technique compensates for the limitation of NIHSS and provides information of changes in hemodynamics, BBB and WM tracts. SWI can indirectly reflect the hemodynamics after stroke according to the number of medullary veins. Delayed post-contrast T1WI can semiquantitatively assess BBB disruption according to contrast agent leakage into brain parenchyma. DTI can evaluate the microstructure of WM tracts in ischemic lesion and outside ischemic lesion. However, wthether the pathophysiological information reflected by MRI is related with prognosis in stroke and diabetic stroke has not been systematically investigated.Therefore, our study used medullary veins on SWI to reflect the hemodynamic information and analyzed its relationship with clinical outcome in stroke patients (Experiment 1); used PE on post-enhanced T1WI to detect BBB disruption and evaluated its relationship with prognosis in diabetic stroke patients (Experiment 2); and used FA map derived from DTI to indicate microstructural integrity of white matter and assessed its relationship with acute neurological deficits and outcome in diabetic stroke patients (Experiment 3).Material and MethodsExperiment 1:Forty-three stroke patients without thrombolytic or recanalization therapy in the middle cerebral artery territory were enrolled. Clinical outcome was classified as poor if the modified Rankin score (mRS) was> 2. The subtypes of AMV was classified into ipsilateral MV (MVi), contralateral MV (MVc) and normal appearances (Norm) according to the difference number of medullary veins between bilateral hemispheres on multi-echo SWI images. Perfusion weighted imaging (PWI) was used to calculate differences in hemispheric cerebral blood flow (ACBF) from corresponding areas. ΔCBF=(CBFaffected side-CBFnormal side)/CBFnormal side × 100%. Logistic regression model was used to analyze whether AMV was an independent factor associated stroke prognosis after including clinical and imaging variables. The relationships between perfusion change and different types of AMV were also analyzed using logistic regression model.Experiment 2:Sixty-two stroke patients without thrombolytic or recanalization therapy in the middle cerebral artery territory were enrolled. Clinical outcome was classified as unfavorable if the modified Rankin score (mRS) was 1. In ischemic lesion, BBB disruption was detected by parenchymal enhancement (PE) on 5 minutes delayed post-contrast T1WI. The frequency of PE between diabetic and non-diabetic stroke patients was compared. The size of PE was further segmented and then the correlation between glucose concentration and PE volume was analyzed in diabetic stroke. In diabetic stroke, logistic regression model was used to analyze whether BBB disruption was an independent factor associated outcome after including clinical and imaging variables. The same way was used to analyze the factors associated with clinical outcome in non-diabetic stroke,.Experiment 3:Thirty-six stroke patients without thrombolytic or recanalization therapy in the middle cerebral artery territory were enrolled. Clinical outcome was classified as unfavorable if the modified Rankin score (mRS) was 2. FA map derived from DTI was used to evaluate the integrity of WM. All the FA maps were normalized and excluded the ischemic regions. For patients with right lesions, these index maps were right-left flipped. The difference in FA map between diabetic and non-diabetic stroke was compared using two sample independent t-test in SPM. P value< 0.005 with cluster size> 100 was considered significant. In diabetic stroke, the relationships between FA of WM tracts with NIHSS total score and NIHSS motor score were analyzed using Pearson correlation. Then, logistic regression model was used to analyze the relationship between injuryed WM tracts and clincal outcome.ResultsExperiment 1:AMV was observed in 24 of 43 (55.8%) stroke patients with 14 (32.6%) MVi and 10 (23.2%) MVc. In addition to age and NIHSS score, MVi was independently associated with poor stroke outcome (OR,11.19; 95% CI,1.04-120.03; P =0.046). MVc was not independently predictive of outcome but was significantly more common in patients with good outcome (37.5% versus 5.3%, P= 0.026). A mean 64.5% decrease and 52.4% increase of ΔCBF was found in MVi and MVc, respectively and ΔCBF was independently related with MVi and MVc.Experiment 2:Diabetic patients had a significant higher frequency of PE than non-diabetic patients (58.6% versus 27.3%, p= 0.013). In diabetic stroke, there was a trend towards significant correlation between FPG and volume of PE (r= 0.336, p= 0.086). Logistic regression revealed that PE was independently associated with unfavorable outcome in diabetic stroke (OR,7.04; 95% CI,1.20-41.52; p= 0.031) and admission NIHSS score was independently associated with unfavorable outcome in non-diabetic stroke (OR,1.71; 95% CI,1.10-2.66; p= 0.017).Experiment 3:Compared to non-diabetic stroke, diabetic patients had a significant decreased FA in the right superior longitudinal fasciculus I (SLF-I), right posterior cingulum, posterior corpus callosum, right external capsule, right uncinate fasciculus, bilateral corticospinal tract (at the level of cerebral peduncle), right corticospinal tract (at the level of pons) and right middle cerebellar peduncle. In diabetic stroke, FA of posterior corpus callosum was negatively associated with both NIHSS total score and NIHSS motor (r=-0.518, P= 0.028; r=-0.444, P=0.065). Only FA of right SLF-I was found to be independently related with poor outcome in diabetic stroke (OR,0.705, 95% CI,0.505-0.984; P=0.040).ConclusionsIn stroke patients, AMV is related with perfusion change. Ipsilateral prominet MV reflects hypoperfusion and related with poor outcome, and contralateral prominet MV reflects hyperperfusion and is beneficial for good outcome. PE was found to be more common in diabetic stroke than in non-diabetic stroke, suggesting BBB was susceptible to be injuryed in diabetic patients, and BBB disruption was independently related with unfavorable outcome in diabetic stroke. In addition, different degree of WM damage beyond ischemic lesion was found in diabetic stroke, compared to non-diabetic stroke. The decreased integrity of posterior corpus callosum was associated with acute neurological deficits and decreased integrity of contralateral SLF-I was related with poor outcome after stroke. These findings provide objective evidence related with stroke prognosis, which may play an important role in guiding treatment for stroke patients, as well as those with diabetes. |
PDF Full Text Request