| Background and objectives:Large hemispheric infarction(LHI)is one of the most devastating condition with high mortality and disability rate among acute ischemic stroke(AIS)patients.Although decompressive hemicraniectomy(DHC)within 48 hours has been proven to benefit LHI patients with malignant brain edema(MBE)by randomized controlled trial(RCT),only highly selected cases would be eligible for DHC according to guidelines.Furthermore,because of its invasive nature and the need for multidisciplinary cooperation,DHC is also underused worldwide.Therefore,identifying feasible and practical strategies such as prevention and management of major complications,reperfusion therapy,and optimal use of statins is of great importance for improving clinical outcomes of LHI.Acute complications have a great influence on the death and unfavorable outcomes of LHI patients.It is common for LHI patients who usually suffered severe neurological deficit to experience multiple medical and neurological complications which may have potential interrelationship during hospitalization.However,which complication might play a significant role in the poor outcome of LHI patients remains unclear.Since most complications especially medical complications are potentially preventable or treatable,identifying the major complications and combination of them in LHI patients would help in the prevention,early detection and treatment of those complications independently associated with unfavorable outcome.Meanwhile,prediction of MBE using clinical and radiological factors has been systematically investigated previously,however,predictive model of MBE at early stage of stroke is not fully equivalent to the predictive model of primary endpoint of LHI patients.Most of the pre-existing prognostic prediction models are derived and validated in stroke cohort including minor stroke and have never been validated in LHI patients.Moreover,most of the commonly used models are composed of baseline characteristics and did not consider the impact of major complications and crucial interventions on the prognosis of LHI.Until recently,there is a lack of simple and practical prognostic models specified for LHI patients which can be widely applicated in medical institutions,and it is unclear whether the pre-existing prognostic models for stroke are applicable for predicting clinical outcome of LHI patients.Early reperfusion therapy,including intravenous thrombolysis(IVT)and endovascular interventions(EVT)is the most effective treatment in AIS and is widely used in nowadays clinical practice.A substantial proportion of LHI patients received IVT or EVT at hyperacute stage of stroke,however,the impact of early reperfusion therapy on the development of MBE and the clinical outcome of LHI patients has not been fully elucidated.Previous studies focus on the relationship between reperfusion and cerebral edema have not reached a consensus.Post-treatment recanalization might be a double-edged sword to LHI patients,arresting infarct growth and rescuing ischemic tissue at risk but leading to severe cerebral ischemia-reperfusion injury at the same time.To date,the effect of recanalization status on the development of MBE and the clinical outcome of LHI patients has not been clearly elucidated.On the other hand,although EVT is a more effective method in recanalization of large vessel occlusions,some families choose for IVT alone or even conservative treatments because of financial concerns,religion,and other values.Different reperfusion therapies might result in different incidence rate of MBE and different clinical outcomes.Reperfusion and neuroprotection are the two major strategies which have been developed to treat ischemic stroke.However,none of the neuroprotective agents has been proven effective in clinical trials and recommended by recent guidelines.Statins have been demonstrated to have pleiotropic effects on the neuroprotection of acute ischemic stroke in several studies,but clinical evidence of acute poststroke statin treatment as neuroprotective agents in AIS is limited.Based on available evidence,it is recommended to continue with prestroke statins use or newly initiate statins in patients with AIS presumed to be of atherosclerotic origin or in patients≤75 years of age who have clinically evident ASCVD in the acute phase of AIS.Since cardio-embolism caused by atrial fibrillation are most common stroke etiology in LHI patient,there is a lack of strong indication for acute poststroke statin treatment.Routine prescription of statins in the acute phase of AIS is still controversial.Although lack of sufficient evidence,statins are still commonly used off label in LHI patients in recent years.As a result,it is necessary to explore the safety and effectiveness of acute poststroke statin use in LHI patients according to real-world experience,and to clarify the feasibility and appropriate administration strategy of acute poststroke statin use as a neuroprotective agent for LHI patients.Therefore,the purpose of this study was to:(1)explore and identify the major complications independently associated with clinical outcome of LHI patients,establish and validate a simple and practical prognostic models specified for LHI patients based on major complications,and compare the predictive value of3-month unfavorable outcomes in LHI patients between the newly proposed predictive model and commonly used prognostic models of stroke patients,(2)explore the association between different reperfusion therapies and recanalization status with the development of MBE and the clinical outcome of LHI patients,and to identify the risk factors of MBE in LHI patients following early reperfusion therapy,(3)explore the safety and effectiveness of acute poststroke statin use in LHI patients according to real-world experience,and to clarify the feasibility and appropriate administration strategy of acute poststroke statin use as a neuroprotective agent for LHI patients,and to identify LHI patients who might benefit more from acute poststroke statin treatment.Materials and Methods:This study was performed using prospective data from the Chengdu Stroke Registry.AIS patients who were admitted to West China Hospital,Sichuan University,within 24 hours from symptoms onset were consecutively registered from Jan 1,2009 to Sep 31,2018.We enrolled LHI patients with computed tomography(CT)and/or magnetic resonance imaging(MRI)evidence of supratentorial cerebral infarction involving more than 50% of middle cerebral artery(MCA)region,with or without the involvement of the adjacent anterior cerebral artery(ACA)or posterior cerebral artery(PCA),and screened as a derivation cohort.LHI patients admitted to People’s Hospital of Deyang City within 24 hours from symptoms onset between Jan 1,2017 and Sep 31,2019 were screened as a validation cohort.Demographic data,clinical manifestations and admission delay,initial stroke severity was assessed by the National Institutes of Health Stroke Scale(NIHSS)score,baseline systolic and diastolic blood pressure,serum glucose on admission,vascular risk factors,stroke etiology,imaging findings,in-hospital treatments,and acute complications during hospitalization were collected,especially for detailed information of 15 common stroke-related complications,reperfusion therapy,and prestroke and poststroke statins treatment.And then a database of LHI patients was established.Patients were followed up at 1 week,1-month and 3-month after stroke onset for any cause of death or unfavorable outcome [defined as a modified Rankin Scale(m RS)score of 4 to 6,and 6 indicates death].(1)Part Ⅰ: The univariate analysis was used to compare the difference in the incidence of the primary outcome(3-month unfavorable outcome)and secondary outcomes(1-month and 3-month mortality)between each complication exposed and unexposed group to screen the potential complications associated with death and unfavorable outcome of LHI patients.The selected potential complications and confounders were added into the multivariate logistic regress model to confirm the major complications independently associated with 3-month unfavorable outcome as well as 1-month and 3-month mortality of LHI patients.And then we separately calculated the area under the curves(AUCs)of individual major complication and combination of major complications for predicting 3-month unfavorable outcome using the receiver operating characteristic(ROC)curves.In addition,the selected combination of major complications and other factors independently associated with3-month unfavorable outcome were added into the multivariate logistic regress model to determine the components of the new prognostic prediction model.We assigned points for each component and then derived the new predictive model for LHI patients based on the weight of its predictive value.In the derivation and the validation cohorts,the AUC was used to measure the discrimination of the new unfavorable outcome prediction models,Hosmer-Lemeshow goodness of fit and calibration plot was used to test the calibration of the new predictive model.Finally,we validated the predictive value of three commonly used prognostic models of stroke patients(including ASTRAL,THRIVE,and SPAN-100)for 3-month unfavorable outcome in our derivation cohort,and compared them with our newly proposed predictive model using De Long’s test.(2)Part Ⅱ:We conduct the Part II of this thesis in a cohort of LHI patients who had received reperfusion therapy at hyperacute stage.The cohort was divided into IVT group(receiving IVT alone)and EVT group(direct mechanical thrombectomy or bridging therapy)according to different reperfusion therapies at hyperacute stage.The included patients were also divided into successful recanalization group and unsuccessful recanalization group according to the post-treatment recanalization status.The primary outcome measures were 3-month unfavorable outcome,and the secondary outcome measure was MBE during hospitalization and 3-month mortality.The univariate analysis was used to compare the difference in baseline characteristics,in-hospital treatments,MBE and other major complications,and the incidence of 3-month unfavorable outcome and mortality between different reperfusion therapies and recanalization status groups.We separately calculated3-month survival curves estimated by the Kaplan-Meier method and performed a log-rank test for survival comparisons between patient groups.Multivariate logistic regress was performed to explore the association between different reperfusion therapies and recanalization status with the development of MBE and the clinical outcome of LHI patients,and to identify the risk factors of MBE in LHI patients following early reperfusion therapy.(3)Part Ⅲ:The patients were divided into two groups(statin user and statin nonuser)based on the statin treatment regimen,whether they initiated poststroke statin therapy within 72 hours from stroke onset or not.The primary outcome measure was 3-month unfavorable outcome.The secondary outcome measures were1-week,1-month and 3-month mortality,a composite of cardiac events during hospitalization,early stroke recurrence,and the overall hemorrhagic transformation and symptomatic hemorrhagic transformation.We performed a propensity score matching(PSM)algorithm(logistic regression)including all baseline characteristics to calculate the propensity score for each patient.Then the LHI patients with early statin administration(the statin user group)were matched with statin nonuser group via using the nearest neighbor approach to minimize potential imbalances in the distribution of potential confounders between statin users and nonusers.All outcomes were compared between the two groups before and after PSM via univariate analysis.Multivariate logistic regression analyses adjusting for confounders and the propensity score were further performed before and after PSM,to explore whether early poststroke statin administration was association with better clinical outcomes of LHI patients.In addition,subgroup analyses were conducted according to different statin administration strategies to determine the influence of different statin administration strategies on the clinical outcomes of LHI patients.Finally,Stratified analyses and interaction test were conducted according to different age group,initial stroke severity,reperfusion therapy/DHC/mechanical ventilation administration or not,cardio-embolism or not,with clinical ASCVD or not,to identify variables that might modify the association between early poststroke statin administration and outcomes,and identify LHI patients who might benefit more from early poststroke statin administration.Results:Part Ⅰ:(1)During the study period,a total of 643 LHI patients admitted to hospital within 24 hours from symptoms onset were enrolled.Among the entire cohort,82(12.8%)died during hospitalization,96(14.9%)died at 1 week,185(28.8%)died at 1 month,219(34.1%)died at 3 and 437(68.0%)patients had unfavorable outcome at 3 months.618(96.1%)had at least one of the common stroke-related complications during hospitalization.The incidence of common acute complications were ranked from high to low as follows: pulmonary infection(73.4%),electrolyte disorder(57.9%),overall hemorrhagic transformation(46.2%,symptomatic hemorrhagic transformation 6.7%),urinary incontinence(44.2%),hypoalbuminemia(41.8%),MBE(40.3%),gastrointestinal bleeding(31.9%),acute heart failure(21.6%),acute renal failure(19.0%),urinary tract infection(17.4%),central hyperthermia(11.5%),post-stroke seizures/epilepsy(10.0%),deep vein thrombosis/pulmonary embolism(8.7%),bedsore(2.6%),and early recurrent stroke(1.6%).(2)We identified 5 major complications independently associated with3-month unfavorable outcomes of LHI patients: MBE(OR 4.28,95%CI 1.97-9.28,P<0.001),central hyperthermia(OR 10.61,95%CI 1.23-91.49,P=0.032),pulmonary infection(OR 1.76,95%CI 1.05-2.95,P=0.031),acute renal failure(OR3.00,95%CI 1.33-6.81,P=0.008),and hypoalbuminemia(OR 2.01,95%CI1.16-3.48,P=0.013).Although central hyperthermia had a low AUC(0.57)and sensitivity(15.8%),it had a high specificity(97.9%)to predict the 3-month unfavorable outcomes of LHI patients with the positive predictive value of 94.5%.Overall,the predictive value of individual major complication for 3-month unfavorable outcomes of LHI patients was limited(AUC 0.57-0.69),among these 5major complications,MBE showed the highest predictive value(AUC 0.69,95% CI0.65-0.73).The combination of major complications showed better predictive value for 3-month unfavorable outcomes of LHI patients compared with the individual major complication,among which a combination of MBE,pulmonary infection,acute renal failure,and hypoalbuminemia had the highest predictive ability(AUC0.81,95%CI 0.77-0.84)and constituted important components of a comprehensive predictive model for unfavorable outcomes of LHI patients.(3)We identified 3 major complications independently associated with1-month mortality of LHI patients: MBE(OR 9.83,95%CI 4.86-19.88,P<0.001),central hyperthermia(OR 7.98,95%CI 2.81-22.70,P<0.001),and acute renal failure(OR 3.45,95%CI 1.71-6.93,P=0.001).We also identified 4 major complications independently associated with 3-month mortality of LHI patients: MBE(OR 7.23,95%CI 3.68-14.21,P<0.001),symptomatic hemorrhagic transformation(OR 3.93,95%CI 1.43-10.81,P=0.008),central hyperthermia(OR 5.34,95%CI 1.92-14.82,P=0.001),and acute renal failure(OR 3.25,95%CI 1.60-6.62,P=0.001).(4)We constructed a simple and practical predictive model for unfavorable outcomes of LHI patients with a total score of 9 based on major complications,which consisted of age≥70 years(yes=2 point;no=0 point),NIHSS score≥20(yes=1point;no=0 point),MBE(yes=3 points;no=0 point),pulmonary infection(yes=1point;no=0 point),acute renal failure(yes=1 point;no=0 point),and hypoalbuminemia(yes=1 point;no=0 point).In the derivation cohort,the incidence of 3-month unfavorable outcome increased gradually with the increased score of our newly proposed predictive model.The risk of 3-month unfavorable outcome in patients with LHI was 2.00 times higher for each 1-point increase in the total score of our new model.The AUC was 0.84(95% CI 0.81-0.87)for 3-month unfavorable outcome.In LHI patients who scored >3,the positive predictive value for 3-month unfavorable outcome was 87.4%,with a sensitivity of 79.2% and a specificity of74.4%.Model calibration assessed by the Hosmer-Lemeshow goodness of fit test(P=0.877)and calibration plot was satisfactory in the derivation cohort.In the validation cohort,the AUC was 0.87(95% CI 0.81-0.94)for 3-month unfavorable outcome,and the new predictive model also demonstrated good calibration assessed by the Hosmer-Lemeshow goodness of fit test(P=0.887)and calibration plot.The three commonly used prognostic models for stroke patients including ASTRAL,THRIVE and SPAN-100 yielded AUCs of 0.73,0.69,and 0.64,respectively,for3-month unfavorable outcome of LHI patients in the derivation cohort.The De Long’s test showed that our newly proposed predictive model had a higher AUC to predict 3-month unfavorable outcome of LHI patients than any of the three commonly used prognostic models for stroke(all P<0.001).Part Ⅱ:In a cohort of 114 LHI patients who had received reperfusion therapy at hyperacute stage of stroke,77 were treated with IVT alone,47 were treated with EVT.Successful recanalization was achieved in 55(44.4%)patients,while complete recanalization was achieved in 39(31.5%)patients.During hospitalization,58(46.8%)cases developed MBE following early reperfusion therapy.The median time from stroke onset to the development of MBE is 2 days((interquartile range[IQR] 1-2 days),89.7%(52/58)patients developed MBE within 3 days from stroke onset.The incidence of MBE was 37.7% in the IVT group and 67.7% in the EVT group.While the incidence of MBE was 43.6% in LHI patient with successful recanalization and 49.3% in patients with unsuccessful recanalization.Multivariate analysis found that neither different reperfusion therapies(EVT versus IVT alone)nor different recanalization status(successful recanalization versus.unsuccessful recanalization)was independently associated with the development of MBE in LHI patients who had received reperfusion therapy(all P>0.05).The multivariate analysis identified the following independent factors associated with MBE in LHI patients following reperfusion therapy: age(OR 0.95,95%CI 0.91-0.98),diastolic blood pressure on admission(OR 1.04,95%CI 1.01-1.07),baseline Alberta Stroke Program Early CT Score(ASPECTS)(OR 0.69,95%CI 0.50-0.96),right hemisphere infarction(OR 4.29,95%CI 1.38-13.40),hypodensity >1/3 of the MCA territory on baseline CT scan(OR 3.56,95%CI 1.11-11.49),infarction involving the ipsilateral ACA territory(OR 7.33,95%CI 1.37-39.32)(all P<0.05).Among the cohort of LHI patients following early reperfusion therapy,42(33.9%)patients died and 84(67.7%)patients had unfavorable outcome at 3 months.The multivariate analysis identified that EVT was associated with a lower risk of 3-month mortality in LHI patients following reperfusion therapy,compared with IVT alone(EVT versus IVT: OR 0.10,95%CI 0.01-0.84,P=0.034).Meanwhile,both successful recanalization(OR 0.21,95%CI 0.07-0.70,P=0.010)and complete recanalization(OR 0.18,95%CI 0.05-0.64,P=0.008)were associated with a lower risk of 3-month unfavorable outcome in LHI patients following reperfusion therapy.Part Ⅲ: In an entire cohort of 643 patients admitted within 24 hours from onset,327(50.9%)patients initiated poststroke statin therapy within 72 hours from stroke onset(acute statin user group)and 316(49.1%)patients did not treated with statin within 72 hours(acute statin nonuser group).In univariate analyses before and after PSM,LHI patients with early poststroke statin administration showed lower rates of1-week mortality(4.2% vs.19.2% after PSM,P<0.001),1-month mortality(18.6%vs.35.3% after PSM,P=0.001)and 3-month mortality(22.2% vs.40.1% after PSM,P<0.001).Meanwhile,acute statin users showed a trend of lower incidence of3-month unfavorable outcome(60.5% vs.70.1% after PSM,P=0.086).It is worth noting that the acute statin user did not have a significant higher rate of the overall HT and s HT(all P>0.05)before and after PSM,however,the acute statin user had a significant higher rate of acute renal failure(22.3% vs.15.5% before PSM,P=0.028;24.6% vs.15.0% after PSM,P=0.028)which was more common in rosuvastatin users(before PSM: OR 2.32,95%CI 1.28-4.21,P=0.008;after PSM: OR 2.95,95%CI 1.28-6.80,P=0.006).In multivariate analyses adjusting for confounders associated with outcome measures,prestroke statin use and propensity score before and after PSM,early poststroke statin administration was associated with significant lower risk of 1-week mortality(after PSM: HR 0.23,95%CI 0.10-0.54,P=0.001),1-month mortality(after PSM: HR 0.48,95%CI 0.31-0.76,P=0.002)and 3-month mortality(after PSM: HR 0.49,95%CI 0.32-0.75,P=0.001).Meanwhile,acute statin users showed a trend of lower risk of 3-month unfavorable outcome in the propensity score-matched cohort(OR 0.55,95%CI 0.29-1.03,P=0.062).Subgroup analyses found that different statin administration strategies including statin initiation strategy(prestroke statin continued or new initiation of statin treatment within 3 days),intensity of statin use(low-to-moderate intensity or high-intensity),and time of poststroke statin initiation(initiated within 24 hours from symptom onset or not),did not change the effect of early poststroke satin treatment on lowering the 3-month mortality in LHI patients,and moreover,LHI patients with newly initiated statin,statin initiated within 24 hours from symptom onset got more benefit.Stratified analyses and interaction test suggested that the association between early poststroke statin treatment and lower mortality of LHI patients had not been change by different age group(≤75 vs.>75 years),initial stroke severity(baseline NIHSS score <20 vs ≥20),receiving reperfusion therapy or not,DHC administration or not,mechanical ventilation or not,cardio-embolism or not,with clinical ASCVD or not(all P for interaction > 0.05).Meanwhile,In LHI patients who were aged ≤ 75 years(OR 0.50,95%CI 0.26-1.00,P=0.050),reperfusion therapy nonusers(OR 0.55,95%CI 0.27-1.10,P=0.091),DHC nonuser(OR 0.55,95%CI 0.29-1.05,P=0.072),non-cardioembolic infarction(OR 0.32,95%CI0.11-0.90,P=0.030),clinical ASCVD(OR 0.40,95%CI 0.17-0.92,P=0.031),early poststroke statin treatment showed a lower risk of 3-month unfavorable outcome in the propensity score-matched cohort.Conclusions:(1)In the present study,we identified 5 major complications independently associated with 3-month unfavorable outcomes of LHI patients: MBE,central hyperthermia,pulmonary infection,acute renal failure,and hypoalbuminemia.We also identified 3 major complications independently associated with 1-month mortality and 4 major complications independently associated with 3-month mortality.Meanwhile,we identified the combination of major complications showed highest predictive ability for 3-month unfavorable outcomes of LHI patients.We innovatively derived a simple and practical predictive model for 3-month unfavorable outcomes of LHI patients based on major complications.The discrimination and calibration of our newly proposed predictive model was satisfactory in the derivation cohort,and the predictive ability of new model is higher than any of the three commonly used prognostic models for stroke.The present study provides clear evidence for prevention and management of major complications to improve the clinical outcomes of LHI patients,and also helps clinicians,patients and their families estimate long-term clinical outcomes of LHI patients when they attempt to make an appropriate decision.(2)Nearly one half of LHI patients following early reperfusion therapy developed MBE and the median time from stroke onset to the development of MBE is 2 days.Neither different reperfusion therapies nor different recanalization status was independently associated with the development of MBE in LHI patients who had received reperfusion therapy,while younger age,higher baseline diastolic blood pressure,lower baseline ASPECTS,right hemisphere infarction,hypodensity >1/3of the MCA territory on baseline CT scan,infarction involving the ipsilateral ACA territory were independent risk factors associated with MBE in LHI patients following reperfusion therapy.In our study,both successful recanalization and complete recanalization were associated with a lower risk of 3-month unfavorable outcome in LHI patients following reperfusion therapy.Meanwhile,EVT was associated with a lower risk of 3-month mortality,compared with IVT alone.The results of the present study are helpful for early identification of LHI patients who were at high risk of MBE following reperfusion therapy,help individualized monitoring and giving effective interventions such as DHC.Our study also provides preliminary evidence in real world for the feasibility of reperfusion therapy in improving clinical outcomes of LHI patients.(3)Our study contributes preliminary evidence in real world for the safety and effectiveness of early poststroke statin treatment in improving clinical outcomes of LHI patients.The current study found that early poststroke statin treatment within 72 hours from stroke onset was associated with a significant lower risk of 1-week,1-month and 3-month mortality.Meanwhile,acute statin users showed a trend of lower incidence of 3-month unfavorable outcome.It is worth noting that acute statin user did not have a significant higher rate of the overall HT and s HT,however,the acute statin user had a significant higher rate of acute renal failure which was more common in rosuvastatin users.Subgroup analyses suggested that LHI patients with newly initiated statin,statin initiated within 24 hours from symptom onset got more benefit.Stratified analyses and interaction test suggested that the association between early poststroke statin treatment and lower death rate of LHI patients did not change by different age group,initial stroke severity,administration of reperfusion therapy/DHC/mechanical ventilation or not,cardio-embolism or not,with clinical ASCVD or not.Future clinical trials with large sample size are needed to verify the findings of our study. |
PDF Full Text Request