A Review of Risk Factors and Predictors for Hemorrhagic Transformation in Patients with Acute Ischemic Stroke

Development and validation of an explainable machine learning prediction model of hemorrhagic transformation after intravenous thrombolysis in stroke

Objective

To develop and validate an explainable machine learning (ML) model predicting the risk of hemorrhagic transformation (HT) after intravenous thrombolysis.

Methods

We retrospectively enrolled patients who received intravenous tissue plasminogen activator (IV-tPA) thrombolysis within 4.5 h after symptom onset to form the original modeling cohort. HT was defined as any hemorrhage on head CT scan completed within 48 h after IV-tPA administration. We utilized the Random Forest (RF), Multilayer Perceptron (MLP), Adaptive Boosting (AdaBoost), and Gaussian Naive Bayes (GauNB) algorithms to develop ML-HT models. The models' predictive performance was evaluated using confusion matrix (including accuracy, precision, recall, and F1 score), and discriminative analysis (area under the receiver-operating-characteristic curve, ROC-AUC) in the original cohort, followed by validation in an independent external cohort. The models' explainability was assessed using SHapley Additive exPlanations (SHAP) global feature plot, SHAP Summary Plot, and Partial Dependence Plot.

Results

A total of 1,007 patients were included in the original modeling cohort, with an HT incidence of 8.94%. The RF-based ML-HT model showed metrics of 0.874 (accuracy), 0.972 (precision), 0.890 (recall), 0.929 (F1 score); with ROC-AUC of 0.7847 in the original cohort and 0.7119 in the external validation cohort. The MLP model showed 0.878, 0.967, 0.989, 0.978, 0.7710, and 0.6768, respectively. The AdaBoost model showed 0.907, 0.967, 0.989, 0.978, 0.7798, and 0.6606, respectively. The GauNB model showed 0.848, 0.983, 0.598, 0.716, 0.6953, and 0.6289, respectively. The explainable analysis of the RF-based ML model indicated that the National Institute of Health Stroke Scale (NIHSS) score, age, platelet count, and atrial fibrillation were the primary determinants for HT following IV-tPA thrombolysis.

Conclusion

The RF-based explainable ML model demonstrated promising predictive ability for estimating the risk of HT after IV-tPA thrombolysis and may have the potential to assist the clinical decision-making in emergency settings.

Integrating Neutrophil-To-Albumin Ratio and Triglycerides: A Novel Indicator for Predicting Spontaneous Hemorrhagic Transformation in Acute Ischemic Stroke Patients

BACKGROUND

Hemorrhagic transformation (HT) is a tragic complication of acute ischemic stroke (AIS), with spontaneous HT (sHT) occurring even without reperfusion therapies. Despite evidence suggesting that several inflammation biomarkers are closely related to HT, its utility in sHT risk stratification remains unclear. This study aimed to identify and integrate effective inflammatory biomarkers associated with sHT and to develop a novel nomogram model for the early detection of sHT.

METHODS

We conducted a retrospective observational cohort study of AIS patients receiving conventional medical treatment solely from March 2022 to March 2023, using a prospectively maintained database. All patients underwent CT follow-up within 7 days after admission, with sHT occurrence within this period as the outcome. Data on demographics, clinical information, laboratory results, and imaging were collected. The cohort was divided into training and validation sets (7:3). Least absolute shrinkage and selection operator (LASSO) regression selected inflammatory biomarkers for a novel index. Univariable and multivariable logistic regressions were conducted to identify independent sHT risk factors. Receiver operating characteristic (ROC) analysis determined optimal cut-off values for continuous factors. A nomogram was developed and validated internally and externally. Predictive accuracy was assessed using the area under the ROC curve (AUC) and calibration plots. Decision curve analysis (DCA) evaluated clinical usefulness.

RESULTS

Of 803 AIS patients, 325 were included in the final analysis. sHT was found in 9.5% (31 patients). Training (n = 228) and validation (n = 97) cohorts showed no significant demographic or clinical differences. LASSO regression integrated neutrophil-to-albumin ratio (NAR) and triglycerides (TGs) into a novel index-NATG. Independent sHT risk factors included baseline National Institute of Health Stroke Scale (NIHSS) (OR = 1.09, 95% CI (1.02, 1.16), p = 0.0095), NATG (OR = 1534.87, 95% CI (5.02, 469638.44), p = 0.0120), D-dimer (DD) (OR = 1.12, 95% CI (1.01, 1.25), p = 0.0249), and total cholesterol (TC) (OR = 1.01, 95% CI (1.00, 1.01), p = 0.0280), with their respective optimal cut-off values being 13, 0.059, 0.86, and 3.6. These factors were used to develop the nomogram in the training cohort, which achieved an AUC of 0.804 (95% CI, 0.643-0.918) in the training cohort and 0.713 (95% CI, 0.499-0.868) in the validation cohort, demonstrating consistent calibration. DCA confirmed the nomogram's clinical applicability in both cohorts.

CONCLUSIONS

A novel indicator combining NAR and TG is positively associated with sHT in AIS patients. The constructed nomogram, integrating this novel indicator with other risk factors, provides a valuable tool for identifying sHT risk, aiding in clinical decision-making.

Characterization and individual-level prediction of cognitive state in the first year after 'mild' stroke

BACKGROUND

Mild stroke affects more than half the stroke population, yet there is limited evidence characterizing cognition over time in this population, especially with predictive approaches applicable at the individual-level. We aimed to identify patterns of recovery and the best combination of demographic, clinical, and lifestyle factors predicting individual-level cognitive state at 3- and 12-months after mild stroke.

METHODS

In this prospective cohort study, the Montreal Cognitive Assessment (MoCA) was administered at 3-7 days, 3- and 12-months post-stroke. Raw changes in MoCA and impairment rates (defined as MoCA<24 points) were compared between assessment time-points. Trajectory clusters were identified using variations of ≥1 point in MoCA scores. To further compare clusters, additional assessments administered at 3- and 12-months were included. Gamma and Quantile mixed-effects regression were used to predict individual MoCA scores over time, using baseline clinical and demographic variables. Model predictions were fitted for each stroke survivor and evaluated using model cross-validation to identify the overall best predictors of cognitive recovery.

RESULTS

Participants' (n = 119) MoCA scores improved from baseline to 3-months (p<0.001); and decreased from 3- to 12-months post-stroke (p = 0.010). Cognitive impairment rates decreased significantly from baseline to 3-months (p<0.001), but not between 3- and 12-months (p = 0.168). Nine distinct trajectory clusters were identified. Clinical characteristics between clusters at each time-point varied in cognitive outcomes but not in clinical and/or activity participation outcomes. Cognitive performance at 3- and 12-months was best predicted by younger age, higher physical activity levels, and left-hemisphere lesion side.

CONCLUSION

More than half of mild-stroke survivors are at risk of cognitive decline one year after stroke, even when preceded by a significantly improving pattern in the first 3-months of recovery. Physical activity was the only modifiable factor independently associated with cognitive recovery. Individual-level prediction methods may inform the timing and personalized application of future interventions to maximize cognitive recovery post-stroke.

The differences of fibrinogen levels in various types of hemorrhagic transformations

Introduction

Hemorrhagic transformation (HT) is a serious complication that can occur spontaneously after an acute ischemic stroke (AIS) or after a thrombolytic/mechanical thrombectomy. Our study aims to explore the potential correlations between fibrinogen levels and the occurrence of spontaneous HT (sHT) and HT after mechanical thrombectomy (tHT).

Methods

A total of 423 consecutive AIS patients diagnosed HT who did not undergone thrombolysis and 423 age- and sex-matched patients without HT (non-HT) were enrolled. Fibrinogen levels were measured within 24 h of admission after stroke. The cohorts were trisected according to fibrinogen levels. The HT were further categorized into hemorrhagic infarction (HI) or parenchymal hematoma (PH) based on their imaging characteristics.

Results

In sHT cohort, fibrinogen levels were higher in HT patients than non-HT patients (p < 0.001 versus p = 0.002). High fibrinogen levels were associated with the severity of HT. HT patients without atrial fibrillation (AF) had higher levels of fibrinogen compared to non-HT (median 3.805 vs. 3.160, p < 0.001). This relationship did not differ among AF patients. In tHT cohort, fibrinogen levels were lower in HT patients than non-HT patients (p = 0.002). Lower fibrinogen levels were associated with the severity of HT (p = 0.004). The highest trisection of fibrinogen both in two cohorts were associated with HT [sHT cohort: OR = 2.515 (1.339-4.725), p = 0.016; that cohort: OR = 0.238 (0.108-0.523), p = 0.003].

Conclusion

Our study suggests that lower fibrinogen level in sHT without AF and higher fibrinogen level in tHT are associated with more severe HT.

Association between daily screen time and risk of stroke among middle-aged and elderly people: research based on China health and nutrition survey

Background

We aimed to explore the independent associations between screen time and the risk of stroke among Chinese adults based on the China Health and Nutrition Survey (CHNS).

Methods

Data on Chinese adults aged older than 40 years from the CHNS in during 2004-2009 were selected. A total of 4,587 individuals were included in 2009, including screen time and the risk of stroke. Simultaneously, we traced the previous screen time to 2004 for those with outcome measures in 2009 (n = 2,100). Basic information, lifestyle, and screen behavior were obtained through face-to-face interviews and self-completed questionnaires. Anthropometric data collected included blood pressure, body weight, height, hip circumference, and waist circumference. Fasting blood was obtained for measurements of lipid and glucose levels. Cross-sectional analysis and cohort analysis were both performed using multivariate logistic regression.

Results

Of all participants, 3,004 (65.49%) participants spent more than 2 h per day on screen time. Taking the men who spent less than 2 h on screen per day as reference, the crude odds ratio (OR) of the high risk of stroke was 1.53 [95% confidence interval (CI), 1.20-1.95] for the men who spent 2-3 h per day on screen and 2.37 (95% CI, 1.78-3.16) for the men who spent more than 3 h per day on screen. This difference remained significant after adjusting for confounding factors. No association was observed among women. However, in the cohort analysis with screen time in 2006 as the independent variable, the association between screen time and stroke risk was found both in men [OR, 1.83 (95% CI, 1.19-2.82)] and women [OR, 1.48 (95% CI, 1.10-1.99)]).

Conclusion

We found that the high screen time was associated with an increased stroke risk, which was pronounced in men, warranting a universal need to limit screen time in order to improve health.

Key CT markers for predicting haemorrhagic transformation after ischaemic stroke: a prospective cohort study in China

OBJECTIVES

Limited studies have systematically addressed the CT markers of predicting haemorrhagic transformation (HT). We aimed to (1) investigate the predictive ability of the imaging factors on multimodal CT for HT and (2) identify the key CT markers that can accurately predict HT while maintaining easy and rapid assessment in the early stage of stroke.

DESIGN AND SETTING

This was a prospective cohort study conducted in a tertiary hospital in Southwest China.

PARTICIPANTS

Patients with ischaemic stroke admitted within 24 hours after onset were included.

OUTCOME MEASURES

The primary outcome was measured as the overall HT. The secondary outcomes were the presence of parenchymal haematoma, symptomatic HT and spontaneous HT.

RESULTS

A total of 763 patients were included. The early hypodensity >1/3 of the middle cerebral artery (MCA) territory, Alberta Stroke Programme Early CT Score≤7, midline shift, hyperdense middle cerebral artery sign (HMCAS), poor collateral circulation, infarct core and penumbra was independently associated with the increased risk of HT (all p < 0.05). The sensitivity of midline shift for predicting HT was only 3.5%, whereas its specificity was 99.8%. The combination of the early hypodensity >1/3 of the MCA territory, midline shift and HMCAS showed a good predictive performance for HT (area under the curve 0.80, 95% CI 0.75 to 0.84).

CONCLUSIONS

Seven imaging factors on multimodal CT were independently associated with HT. The high specificity of midline shift suggests the need to consider it as an imaging indicator when assessing the risk of HT. The early hypodensity >1/3 of the MCA territory, midline shift and HMCAS was identified as the key CT markers for the early prediction of HT. The coexistence of the three key factors might be a valuable index for identifying individuals at high bleeding risk and guiding further treatments.

Association between serum globulin and cognitive impairment in older American adults

Background and aims

Cognitive impairment is on the rise around the world, with profound economic and social consequences. Serum globulin, a marker of liver function, may also play a role in cognitive function. Unfortunately, no consistent conclusion exists regarding the association between serum globulin and cognitive function.

Methods

Data from the 2011 to 2014 National Health and Nutrition Examination Survey were used to assess the association between serum globulin and cognitive impairment. Cognitive function was assessed by three tests: Consortium to Establish a Registry for Alzheimer's Disease (CERAD), Animal Fluency (AF), and Digit Symbol Substitution Test (DSST). Furthermore, the breakthrough point of cognitive impairment correlated with CERAD < 5, AF < 14, and DSST < 34. A weighted multiple logistics regression model was used to verify the association between serum globulin and cognitive impairment. Generalized additive models (GAMs) and a smooth curve fit (penalty spline method) were used to determine a non-linear relationship between serum globulin and cognitive impairment. Finally, subgroup analysis and interaction tests were conducted to further verify the association between serum globulin and cognitive impairment.

Results

Data from 2,768 participants aged ≥60 (in accordance with the study design) were collected for the final analysis. Data suggested that serum globulin levels were associated with an elevated cognitive impairment based on the AF [full adjustment, OR = 1.05, 95% CI: 1.01-1.08] and DSST [full adjustment, OR = 1.06, 95% CI: 1.02-1.10] tests. Eventually, the GAM and smooth curve fit model was conducted to confirm that the association between serum globulin and cognitive impairment was non-linear. Moreover, the inflection point was 27 g/L serum globulin based on the CERAD test and 35 g/L serum globulin based on the AF test. Finally, the interaction term between serum globulin and cognitive impairment based on the AF test indicated no significant interactions among all variables (all p for interaction >0.05).

Conclusion

The association between serum globulin levels and cognitive impairment is non-linear. A threshold effect exists between serum globulin and cognitive impairment. Large-scale prospective clinical trials are needed to validate our findings.

Risk Factors Associated With Neurological and Extra-Neurological Complications and Mortality in Patients With Stroke

Introduction Understanding when and how systemic complications can occur after an acute stroke is crucial. It is important to identify patients who are at higher risk for these complications. Early and effective treatment based on this knowledge can significantly improve patient outcomes. The objective of this study was to identify the risk factors associated with neurological and extra-neurological complications and mortality in stroke patients treated at a secondary care hospital. Methods Of a total of 170 patients diagnosed with hemorrhagic/ischemic stroke and transient cerebral ischemia at a secondary care hospital in Mexico, the records of 125 were reviewed and of these, 86 were included in the study. The study group comprised 86 adult patients (> 18 years of age) diagnosed with ischemic or hemorrhagic stroke or transient cerebral ischemia. Their demographics, clinical characteristics, in-hospital complications, and mortality were retrospectively analyzed.  Results Of the 86 patients examined, 34.9% experienced complications, regardless of the type of stroke. The most significant factor associated with mortality and complications during hospitalization in patients with stroke was previous diseases. Other factors that were linked to higher mortality were pre-existing medical conditions. The most common neurological complication among patients with stroke during hospitalization was intracranial hypertension (3.5%). As for extra-neurological complications, pressure ulcers and nosocomial pneumonia had an occurrence rate of 4.7%. Conclusions The main neurological complication during hospitalization of patients with stroke was intracranial hypertension, while the extra neurological complications were pressure ulcers and nosocomial pneumonia.

Serum biomarkers to predict hemorrhagic transformation and ischemic stroke outcomes in a prospective cohort study

Ischemic stroke (IS) remains one of the most frequent causes of death and disability worldwide. Identifying possible prognosis factors for IS outcomes, including hemorrhagic transformation (HT), could improve patients' recovery. This study aimed to investigate the potential prognosis role of non-specific laboratory data at admission and baseline MMP-2 and MMP-9 serum levels in predicting HT risk, discharge, and 3-month follow-up status of IS patients. Data from 150 successive acute cerebral infarction patients were analyzed in a prospective cohort study. The active group included patients who developed HT during hospitalization (55 persons). There were no significant differences in age, gender distribution, time to admission, or time to blood sample collection for MMPs measurement between patients in the active and control groups. IS patients from the active group had a significantly higher rate of AF (atrial fibrillation) in the past (p=0.003), while differences in other factors such as diabetes, hypertension, myocardial infarction, previous stroke, obesity, smoking, and alcohol were not significant. Admission NIHSS score and mRS (modified Rankin Scale) values (at discharge and 90 days) were significantly worse in the active group (p<0.001). Among the analyzed admission laboratory factors (glycemia, lipid profile, coagulation panel, inflammatory reaction parameters, MMP-2, MMP-9), INR presented an inverse correlation, with lower values in the HT cohort (univariate analysis - p=0.01, OR=0.11; multivariate analysis - p=0.03, OR=0.09). Further research on larger cohorts is warranted to determine the specific laboratory biomarkers for predicting hemorrhagic transformation and ischemic stroke outcomes.

Vascular contributions to cognitive impairment/dementia in diabetes: role of endothelial cells and pericytes

Vascular contributions to cognitive impairment/dementia (VCID) are a leading cause of dementia, a known neurodegenerative disorder characterized by progressive cognitive decline. Although diabetes increases the risks of stroke and the development of cerebrovascular disease, the cellular and vascular mechanisms that lead to VCID in diabetes are yet to be determined. A growing body of research has identified that cerebrovascular cells within the neurovascular complex display an array of cellular responses that impact their survival and reparative properties, which plays a significant role in VCID development. Specifically, endothelial cells and pericytes are the primary cell types that have gained much attention in dementia-related studies due to their molecular and phenotypic heterogeneity. In this review, we will discuss the various morphological subclasses of endothelial cells and pericytes as well as their relative distribution throughout the cerebrovasculature. Furthermore, the use of diabetic and stroke animal models in preclinical studies has provided more insight into the impact of sex differences on cerebral vascularization in progressive VCID. Understanding how cellular responses and sex differences contribute to endothelial cell and pericyte survival and function will set the stage for the development of potential preventive therapies for dementia-related disorders in diabetes.

Net clinical benefit of anticoagulant therapy from a neurologist's perspective: A review

In this article the concept of "net clinical benefit" (NCB) is considered in the context of stroke, the role of direct oral anticoagulants in secondary prevention of cardioembolic stroke and NCB from their prescription. Practical aspects of NCB evaluation from the neurologist's point of view are presented, taking into account such factors as stroke duration and severity, features of neurological deficit (severity of residual limitations, dysphagia, motor disorders and risk of falls), neuroimaging characteristics of stroke (focal size and hemorrhagic transformation) and concomitant neurological diseases (epilepsy and dementia). In all these situations, the use of oral anticoagulants has a significant clinical benefit, which justifies the inadmissibility of refusal to prescribe them.