Predicting ischemic stroke patients' prognosis changes using machine learning in a nationwide stroke registry.

Accurately predicting the prognosis of ischemic stroke patients after discharge is crucial for physicians to plan for long-term health care. Although previous studies have demonstrated that machine learning (ML) shows reasonably accurate stroke outcome predictions with limited datasets, to identify specific clinical features associated with prognosis changes after stroke that could aid physicians and patients in devising improved recovery care plans have been challenging. This study aimed to overcome these gaps by utilizing a large national stroke registry database to assess various prediction models that estimate how patients' prognosis changes over time with associated clinical factors. To properly evaluate the best predictive approaches currently available and avoid prejudice, this study employed three different prognosis prediction models including a statistical logistic regression model, commonly used clinical-based scores, and a latest high-performance ML-based XGBoost model. The study revealed that the XGBoost model outperformed other two traditional models, achieving an AUROC of 0.929 in predicting the prognosis changes of stroke patients followed for 3 months. In addition, the XGBoost model maintained remarkably high precision even when using only selected 20 most relevant clinical features compared to full clinical datasets used in the study. These selected features closely correlated with significant changes in clinical outcomes for stroke patients and showed to be effective for predicting prognosis changes after discharge, allowing physicians to make optimal decisions regarding their patients' recovery.

Patient and hospital characteristics associated with do-not-resuscitate/do-not-intubate orders: a cross-sectional study based on the Taiwan stroke registry.

BACKGROUND: Previous studies of do-not-resuscitate (DNR) or do-not-intubate (DNI) orders in stroke patients have primarily been conducted in North America or Europe. However, characteristics associated with DNR/DNI orders in stroke patients in Asia have not been reported. METHODS: Based on the Taiwan Stroke Registry, this nationwide cross-sectional study enrolled hospitalized stroke patients from 64 hospitals between 2006 and 2020. We identified characteristics associated with DNR/DNI orders using a two-level random effects model. RESULTS: Among the 114,825 patients, 5531 (4.82%) had DNR/DNI orders. Patients with acute ischemic stroke (AIS) had the highest likelihood of having DNR/DNI orders (adjusted odds ratio [aOR] 1.76, 95% confidence interval [CI] 1.61-1.93), followed by patients with intracerebral hemorrhage (ICH), and patients with subarachnoid hemorrhage (SAH) had the lowest likelihood (aOR 0.53, 95% CI 0.43-0.66). From 2006 to 2020, DNR/DNI orders increased in all three types of stroke. In patients with AIS, women were significantly more likely to have DNR/DNI orders (aOR 1.23, 95% CI 1.15-1.32), while patients who received intravenous alteplase had a lower likelihood (aOR 0.74, 95% CI 0.65-0.84). Patients with AIS who were cared for by religious hospitals (aOR 0.55, 95% CI 0.35-0.87) and patients with SAH who were cared for by medical centers (aOR 0.40, 95% CI 0.17-0.96) were significantly less likely to have DNR/DNI orders. CONCLUSIONS: In Taiwan, DNR/DNI orders increased in stroke patients between 2006 and 2020. Hospital characteristics were found to play a significant role in the use of DNR/DNI orders.

Renal function is associated with one-month and one-year mortality in patients with intracerebral hemorrhage.

OBJECTIVE: This study evaluated short-term (1-month) and long-term (1-year) mortality risks associated with the glomerular filtration rate (eGFR) on admission for patients with intracerebral hemorrhage. METHODS: From the Taiwan Stroke Registry data from April 2006 to December 2016, we identified and stratified patients with intracerebral hemorrhage into five subgroups by the eGFR level on admission: ≥90, 60-89, 30-59, 15-29, and <15 mL/min/1.73 m2 or on dialysis. Risks for 1-month and 1-year mortality after intracerebral hemorrhage were compared by the eGFR levels. RESULTS: Both the 1-month and 1-year mortality rates progressively increased with the decrease in eGFR levels. The 1-month mortality rate in patients with eGFR < 15 mL/min/1.73 m2 or on dialysis was approximately 5.5-fold greater than that in patients with eGFR ≥ 90 mL/min/1.73 m2 (8.31 versus 1.50 per 1000 person-days), with an adjusted hazard ratio (HR) of 4.59 [95% confidence interval (CI) = 2.71-7.78]. Similarly, the 1-year mortality in patients with eGFR < 15 mL/min/1.73 m2 or on dialysis was 7.5 times that in patients with eGFR ≥ 90 mL/min/1.73 m2 (2.34 versus 0.31 per 1000 person-days), with an adjusted HR of 4.54 (95% CI 2.95-6.98). CONCLUSION: Impairment of renal function is an independent risk factor for mortality in patients with intracerebral hemorrhage in a gradual way. The eGFR level is a prognostic indicator for patients with intracerebral hemorrhage.

Smoking Status and Functional Outcomes in Young Stroke.

Objective: Stroke in young adults is uncommon, and the etiologies and risk factors of stroke in young adults differ from those in older populations. Smoker's paradox is an unexpected favorable outcome, and age difference is used to explain the association between smoking and the favorable functional outcome. This study aimed to investigate the existence of this phenomenon in young stroke patients. Methods: We analyzed a total of 9,087 young stroke cases registered in the nationwide stroke registry system of Taiwan between 2006 and 2016. Smoking criteria included having a current history of smoking more than one cigarette per day for more than 6 months. After matching for sex and age, a Cox model was used to compare mortality and function outcomes between smokers and non-smokers. Results: Compared with the non-smoker group, smoking was associated with older age, higher comorbidities, and higher alcohol consumption. Patients who report smoking with National Institutes of Health Stroke Scale scores of 11-15 had a worse functional outcome (adjusted odds ratio, 0.81; 95% confidence interval, 0.76 - 0.87). Conclusion: Smokers had a higher risk of unfavorable functional outcomes at 3 months after stroke, and therefore, we continue to strongly advocate the importance of smoking cessation.

Low-Density Lipoprotein Cholesterol and Mortality in Patients With Intracerebral Hemorrhage in Taiwan.

Objective: Lower serum low-density lipoprotein cholesterol (LDL-C) levels are associated with increased intracerebral hemorrhage (ICH) risk. However, reverse causality and residual confounding has not attracted public attention. Therefore, we assessed whether people with LDL-C have increased risk of mortality adjusting for potential confounders using two large Taiwan cohorts. Methods: The Mei-Jhao (MJ) cohort has 414,372 adults participating in a medical screening program with 378 ICH deaths within 15 years of follow-up (1994-2008). Cox proportional hazards regressions estimated hazard death ratios according to LDL-C levels. We identified 4,606 ICH patients from the Taiwan Stroke Registry (TSR) and analyzed the impact of LDL-C on 3-month mortality. Results: Low cholesterol (LDL-C <100 mg/dL), found in 1/4 of the MJ cohort, was highly prevalent (36%) among young adults (age 20-39). There was a graded relationship between cholesterol and mortality for ICH [Hazard ratio, 1.56; 95% confidence interval (CI), 1.13-2.16]. Compared with patients with an LDL-C of 110-129 mg/dL in TSR, the risk for mortality was 1.84 (95% CI, 1.28-2.63) with an LDL-C of <100 mg/dL. Conclusion: Lower serum LDL-C level independently predicts higher mortality after acute ICH. While its causative role may vary, low cholesterol may pose potential harms in Taiwan.

Comparison of outcome prediction models post-stroke for a population-based registry with clinical variables collected at admission vs. discharge.

Aim: The ability to predict outcomes can help clinicians to better triage and treat stroke patients. We aimed to build prediction models using clinical data at admission and discharge to assess predictors highly relevant to stroke outcomes. Methods: A total of 37,094 patients from the Taiwan Stroke Registry (TSR) were enrolled to ascertain clinical variables and predict their mRS outcomes at 90 days. The performances (i.e., the area under the curves (AUCs)) of these independent predictors identified by logistic regression (LR) based on clinical variables were compared. Results: Several outcome prediction models based on different patient subgroups were evaluated, and their AUCs based on all clinical variables at admission and discharge were 0.85-0.88 and 0.92-0.96, respectively. After feature selections, the input features decreased from 140 to 2-18 (including age of onset and NIHSS at admission) and from 262 to 2-8 (including NIHSS at discharge and mRS at discharge) at admission and discharge, respectively. With only a few selected key clinical features, our models can provide better performance than those previously reported in the literature. Conclusion: This study proposed high performance prognostics outcome prediction models derived from a population-based nationwide stroke registry even with reduced LR-selected clinical features. These key clinical features can help physicians to better focus on stroke patients to triage for best outcome in acute settings.

Smoking Paradox in Stroke Survivors?: Uncovering the Truth by Interpreting 2 Sets of Data.

Background and Purpose- The observation that smokers with stroke could have better outcome than nonsmokers led to the term "smoking paradox." The controversy of such a complex claim has not been fully settled, even though different case mix was noted. Analyses were conducted on 2 independent data sets to evaluate and determine whether such a paradox truly exists. Methods- Taiwan Stroke Registry with 88 925 stroke cases, and MJ cohort with 541 047 adults participating in a medical screening program with 1630 stroke deaths developed during 15 years of follow-up (1994-2008). Primary outcome for stroke registry was functional independence at 3 months by modified Rankin Scale score ≤2, for individuals classified by National Institutes of Health Stroke Scale score at admission. For MJ cohort, mortality risk by smoking status or by stroke history was assessed by hazard ratio. Results- A >11-year age difference in stroke incidence was found between smokers and nonsmokers, with a median age of 60.2 years for current smokers and 71.6 years for nonsmokers. For smokers, favorable outcome in mortality and in functional assessment in 3 months with modified Rankin Scale score ≤2 stratified by the National Institutes of Health Stroke Scale score was present but disappeared when age and sex were matched. Smokers without stroke history had a ≈2-fold increase in stroke deaths (2.05 for ischemic stroke and 1.53 for hemorrhagic stroke) but smokers with stroke history, 7.83-fold increase, overshadowing smoking risk. Quitting smoking at earlier age reversed or improved outcome. Conclusions- "The more you smoke, the earlier you stroke, and the longer sufferings you have to cope." Smokers had 2-fold mortality from stroke but endured stroke disability 11 years longer. Quitting early reduced or reversed the harms.

CO-releasing molecules and increased heme oxygenase-1 induce protein S-glutathionylation to modulate NF-κB activity in endothelial cells.

Protein glutathionylation is a protective mechanism that functions in response to mild oxidative stress. Carbon monoxide (CO) can increase the reactive oxygen species concentration from a low level via the inhibition of cytochrome c oxidase. We therefore hypothesized that CO would induce NF-κB-p65 glutathionylation and then show anti-inflammatory effects. In this study, we found that CO-releasing molecules suppress TNFα-induced monocyte adhesion to endothelial cells (ECs) and reduce ICAM-1 expression. Moreover, CO donors were further found to exert their inhibitory effects by blocking NF-κB-p65 nuclear translocation, but do so independent of IκBα degradation, in TNFα-treated ECs. In addition, p65 protein glutathionylation represents the response signal to CO donors and is reversed by the reducing agent dithiothreitol. Thiol modification of the cysteine residue in the p65 RHD region was required for the CO-modulated NF-κB activation. The suppression of p65 glutathionylation by a GSH synthesis inhibitor, BSO, and by catalase could also attenuate TNFα-induced p65 nuclear translocation and ICAM-1 expression. CO donors induce Nrf2 activation and Nrf2 siRNA suppresses CO-induced p65 glutathionylation and inhibition. Furthermore, we found that the CO donors induce heme oxygenase-1 (HO-1) expression, which increases p65 glutathionylation. In contrast, HO-1 siRNA attenuates CO donor- and hemin-induced p65 glutathionylation. Our results thus indicate that the glutathionylation of p65 is likely to be responsible for CO-mediated NF-κB inactivation and that the HO-1-dependent pathway may prolong the inhibitory effects of CO donors upon TNFα treatment of ECs.

Rivastigmine-induced REM sleep behavior disorder (RBD) in a 88-year-old man with Alzheimer's disease.

We report the case of an 88-year-old man with Alzheimer's disease (AD) of 8 years duration (emerging shortly after the de novo onset of sleeptalking) who developed REM sleep behavior disorder (RBD) after increasing the nightly dose of rivastigmine, an acetylcholinesterase inhibitor, from 1.5 mg to 3 mg (total daily dose, 4.5 mg), as therapy for his dementia. His family then became aware of recurrent nocturnal episodes arising from sleep of his leaving bed, and he sustained multiple abrasion injuries from falling down. Polysomnography (PSG), utilizing a seizure montage with fast paper speed, conducted with the patient taking rivastigmine 3 mg at bedtime, documented 3 abrupt episodes of bilateral arm-waving with moaning and shouting that emerged exclusively during each of the 3 REM sleep periods, with the duration of the episodes lasting 8 to 25 seconds. No epileptiform discharge appeared with the onset of these REM sleep behaviors. Therapy with clonazepam, 0.5 mg at bedtime (with ongoing 3 mg bedtime and 4.5 mg total daily rivastigmine therapy), fully suppressed the sleep-related events, with prompt relapse whenever clonazepam was not taken. This is the second reported case (both males with AD) of rivastigmine-induced RBD, and the oldest reported case of RBD; and it represents reversible, medication-induced, acute RBD.