The Role of VASOGRADE as a Simple Grading Scale to Predict Delayed Cerebral Ischemia and Functional Outcome After Aneurysmal Subarachnoid Hemorrhage

Grading Scores for Identifying Patients at Risk of Delayed Cerebral Ischemia and Neurological Outcome in Spontaneous Subarachnoid Hemorrhage: A Comparison of Receiver Operator Curve Analysis

BACKGROUND

Numerous grading scales were proposed for subarachnoid hemorrhage (SAH) to assess the likelihood of unfavorable neurological outcomes (UO) and the risk of delayed cerebral ischemia (DCI). We aimed to validate the Hemorrhage, Age, Treatment, Clinical Status, and Hydrocephalus (HATCH) score and the VASOGRADE, a simple grading scale for prediction of DCI after aneurysmal SAH.

METHODS

This was a retrospective single-center study of patients with nontraumatic SAH (January 2016 to December 2021) admitted to the intensive care unit. We performed a receiver operating characteristic (ROC) curve analysis to assess the discriminative ability of the HATCH and the VASOGRADE to identify patients who had UO at 3 months (defined as Glasgow Outcome Scale score of 1-3), hospital mortality, and DCI and compared their performance with the World Federation of Neurosurgical Surgeons, the modified Fisher, the Sequential Organ Failure Assessment, and the Acute Physiology and Chronic Health Evaluation II scales. We performed a multivariate logistic regression analysis to assess the association between HATCH and UO at 3 months and between VASOGRADE and DCI.

RESULTS

We included 262 consecutive patients with nontraumatic SAH. DCI was observed in 82 patients (31.3%), whereas 78 patients (29.8%) died during hospital stay and 133 patients (51%) had UO at 3 months. HATCH was independently associated with UO (odds ratio 1.61, 95% confidence interval [CI] 1.36-1.90) and had an area under the ROC curve (AUROC) of 0.83 (95% CI 0.77-0.88), comparable to the Acute Physiology and Chronic Health Evaluation II (AUROC 0.84, 95% CI 0.79-0.89) and Sequential Organ Failure Assessment (AUROC 0.83, 95% CI 0.77-0.88).

CONCLUSIONS

Hemorrhage, Age, Treatment, Clinical Status, and Hydrocephalus and VASOGARDE scores had a good performance to predict UO or in-hospital mortality and DCI, respectively; however, their performance did not outperform nonspecific routinely used scores.

Aneurysmal Subarachnoid Hemorrhage Risk Assessment Model Identifies Patients for Safe Early Discharge at Day 15-The SAFE-SaH Score

BACKGROUND

Patients with aneurysmal subarachnoid hemorrhage (aSAH) are often hospitalized for 21 days after aneurysm rupture due to the risk of complications. However, some never experience complications and are unlikely to benefit from a prolonged hospitalization. The aim of this study is to derive a risk assessment model (RAM) using data from the first 14 days of hospitalization to identify low-risk patients for early discharge, at day 15 or after.

METHODS

Patients ≥ 18 years old with an acute aSAH at a Comprehensive Stroke Center from 2017 to 2024 were included. Baseline demographics, aSAH grading scales, and in-hospital complications requiring intervention were characterized. Complications included: vasospasm, delayed cerebral ischemia (DCI), cerebral salt wasting (CSW), cerebral edema, seizures, arrhythmias, respiratory failure, and hydrocephalus. Binary logistic regression with leave-one-out cross validation (LOOCV) was used to identify an optimal RAM.

RESULTS

Of 165 patients, the mean Hunt Hess Score (HHS) was 2.5 (standard deviation, SD 1.2), modified Fisher Score (mFS) was 3.1 (SD 1), endovascular therapy was used for aneurysm securement in 73% of patients, and 54.5% of patients experienced complications during days 15-21. In bivariate analyses, days 0-14 variables associated with days 15 + complications were the following: HHS, mFS, middle cerebral artery (MCA) aneurysm, clinical or radiologic vasospasm, endovascular therapies, intraventricular hemorrhage, hydrocephalus, presence of external ventricular drain (EVD), mechanical ventilation, vasopressors, hypertonic solutions, antiseizure medications, milrinone, and fludrocortisone (all p < 0.05). LOOCV regression for a best fit RAM included six variables: Sum of Vasopressors, Artery (MCA aneurysm), Fludrocortisone, EVD, Scale (mFS and HHS), "SAFE-SaH" and had an area under the receiver operator characteristics curve of 0.90 (95% confidence interval 0.85-0.95), sensitivity of 0.94, specificity of 0.69, positive predictive value of 79%, and negative predictive value of 91% for predicting complications on day 15 + .

CONCLUSIONS

This is the first ever RAM to incorporate clinical data from the first 14 days of hospitalization to identify patients with an aSAH at low risk for complications after day 14. With 94% sensitivity, the RAM classifies patients who will not have complications and may assist in earlier disposition on day 15 or after.

Predicting Mortality in Subarachnoid Hemorrhage Patients Using Big Data and Machine Learning: A Nationwide Study in Türkiye

Background/Objective: Subarachnoid hemorrhage (SAH) is associated with high morbidity and mortality rates, necessitating prognostic algorithms to guide decisions. Our study evaluates the use of machine learning (ML) models for predicting 1-month and 1-year mortality among SAH patients using national electronic health records (EHR) system. Methods: Retrospective cohort of 29,274 SAH patients, identified through national EHR system from January 2017 to December 2022, was analyzed, with mortality data obtained from central civil registration system in Türkiye. Variables included (n = 102) pre- (n = 65) and post-admission (n = 37) data, such as patient demographics, clinical presentation, comorbidities, laboratory results, and complications. We employed logistic regression (LR), decision trees (DTs), random forests (RFs), and artificial neural networks (ANN). Model performance was evaluated using area under the curve (AUC), average precision, and accuracy. Feature significance analysis was conducted using LR. Results: The average age was 56.23 ± 16.45 years (47.8% female). The overall mortality rate was 22.8% at 1 month and 33.3% at 1 year. One-month mortality increased from 20.9% to 24.57% (p < 0.001), and 1-year mortality rose from 30.85% to 35.55% (p < 0.001) in the post-COVID period compared to the pre-COVID period. For 1-month mortality prediction, the ANN, LR, RF, and DT models achieved AUCs of 0.946, 0.942, 0.931, and 0.916, with accuracies of 0.905, 0.901, 0.893, and 0.885, respectively. For 1-year mortality, the AUCs were 0.941, 0.927, 0.926, and 0.907, with accuracies of 0.884, 0.875, 0.861, and 0.851, respectively. Key predictors of mortality included age, cardiopulmonary arrest, abnormal laboratory results (such as abnormal glucose and lactate levels) at presentation, and pre-existing comorbidities. Incorporating post-admission features (n = 37) alongside pre-admission features (n = 65) improved model performance for both 1-month and 1-year mortality predictions, with average AUC improvements of 0.093 ± 0.011 and 0.089 ± 0.012, respectively. Conclusions: Our study demonstrates the effectiveness of ML models in predicting mortality in SAH patients using big data. LR models' robustness, interpretability, and feature significance analysis validate its importance. Including post-admission data significantly improved all models' performances. Our results demonstrate the utility of big data analytics in population-level health outcomes studies.

Machine Learning for the Early Prediction of Delayed Cerebral Ischemia in Patients With Subarachnoid Hemorrhage: Systematic Review and Meta-Analysis

BACKGROUND

Delayed cerebral ischemia (DCI) is a primary contributor to death after subarachnoid hemorrhage (SAH), with significant incidence. Therefore, early determination of the risk of DCI is an urgent need. Machine learning (ML) has received much attention in clinical practice. Recently, some studies have attempted to apply ML models for early noninvasive prediction of DCI. However, systematic evidence for its predictive accuracy is still lacking.

OBJECTIVE

The aim of this study was to synthesize the prediction accuracy of ML models for DCI to provide evidence for the development or updating of intelligent detection tools.

METHODS

PubMed, Cochrane, Embase, and Web of Science databases were systematically searched up to May 18, 2023. The risk of bias in the included studies was assessed using PROBAST (Prediction Model Risk of Bias Assessment Tool). During the analysis, we discussed the performance of different models in the training and validation sets.

RESULTS

We finally included 48 studies containing 16,294 patients with SAH and 71 ML models with logistic regression as the main model type. In the training set, the pooled concordance index (C index), sensitivity, and specificity of all the models were 0.786 (95% CI 0.737-0.835), 0.77 (95% CI 0.69-0.84), and 0.83 (95% CI 0.75-0.89), respectively, while those of the logistic regression models were 0.770 (95% CI 0.724-0.817), 0.75 (95% CI 0.67-0.82), and 0.71 (95% CI 0.63-0.78), respectively. In the validation set, the pooled C index, sensitivity, and specificity of all the models were 0.767 (95% CI 0.741-0.793), 0.66 (95% CI 0.53-0.77), and 0.78 (95% CI 0.71-0.84), respectively, while those of the logistic regression models were 0.757 (95% CI 0.715-0.800), 0.59 (95% CI 0.57-0.80), and 0.80 (95% CI 0.71-0.87), respectively.

CONCLUSIONS

ML models appear to have relatively desirable power for early noninvasive prediction of DCI after SAH. However, enhancing the prediction sensitivity of these models is challenging. Therefore, efficient, noninvasive, or minimally invasive low-cost predictors should be further explored in future studies to improve the prediction accuracy of ML models.

TRIAL REGISTRATION

PROSPERO (CRD42023438399); https://tinyurl.com/yfuuudde.

Proposal of a Machine Learning Based Prognostic Score for Ruptured Microsurgically Treated Anterior Communicating Artery Aneurysms

Background: Aneurysmal subarachnoid hemorrhage (aSAH) carries significant mortality and disability rates, with rebleeding posing a grave risk, particularly in anterior communicating artery (AcoA) aneurysms. This retrospective study aims to analyze preoperative and intraoperative variables of patients with ruptured AcoA aneurysms, evaluating the association of these variables with patient outcomes using machine learning techniques, proposing a prognostic score. Materials and Methods: A retrospective study was conducted on 50 patients who underwent microsurgical clipping for a ruptured AcoA aneurysm at San Giovanni Bosco Hospital, Turin, Italy. The clinical and aneurysmal data-including clinical evaluations, risk factors, aneurysmal characteristics, and intra- and postoperative details-were examined. The study population was analyzed using machine learning techniques such as the MRMR algorithm for feature selection, and the LASSO method was employed to construct linear predictive models based on these features. Results: The study cohort had a mean age of 54 years, with 26 female and 24 male patients. Temporary clipping of main vessels was performed in 96% of procedures, with a mean duration of 3.74 min. Postoperatively, the mean Intensive Care Unit (ICU) stay was 7.28 days, with 14% mortality at 30 days and 4% within the first week. At the six-month follow-up, 63% of discharged patients had a Glasgow outcome scale (GOS) of 5, with radiological confirmation of complete aneurysm exclusion in 98% of cases. Machine learning techniques identified the significant predictors of patient outcomes, with LASSO algorithms generating linear models to predict the GOS at discharge and at 6 months follow-up. Conclusions: Preoperative factors like the BNI score, Vasograde, and preoperative cerebral edema demonstrate significant correlations with patient outcomes post-clipping. Notably, intraoperative bleeding and extended temporary clipping durations (over 3 min) emerge as pivotal intraoperative considerations. Moreover, the AcoA prognostic score shows promise in predicting patient outcomes, discharge plans, and ICU duration.

Leveraging machine learning algorithms to forecast delayed cerebral ischemia following subarachnoid hemorrhage: a systematic review and meta-analysis of 5,115 participants

It is feasible to predict delayed cerebral ischemia (DCI) after aneurysmal subarachnoid hemorrhage (aSAH) using Artificial intelligence (AI) algorithms, which may offer significant improvements in early diagnosis and patient management. This systematic review and meta-analysis evaluate the efficacy of machine learning (ML) in predicting DCI, aiming to integrate complex clinical data to enhance diagnostic accuracy. We searched PubMed, Scopus, Web of science, and Embase databases without restrictions until June 2024, applying PRISMA guidelines. Out of 1498 studies screened, 10 met our eligibility criteria involving ML approaches in patients with confirmed aSAH. The studies employed various ML algorithms and reported differential ML metrics outcomes. Meta-analysis was performed on eight studies, which resulted in a pooled sensitivity of 0.79 [95% CI: 0.63-0.89], specificity of 0.78[95% CI: 0.68-0.85], positive DLR of 3.54 [95% CI: 2.22-5.64] and the negative DLR of 0.28 [95% CI: 0.15-0.52], diagnostic odds ratio of 12.82 [95% CI: 4.66-35.28], the diagnostic score of 2.55 [95% CI: 1.54-3.56], and the area under the curve (AUC) of 0.85. These findings show significant diagnostic accuracy and demonstrate the potential of ML algorithms to significantly improve the predictability of DCI, implying that ML could impart a significant role on improving clinical decision making. However, variability in methodological approaches across studies shows a need for standardization to realize the full benefits of ML in clinical settings.

Treatment factors to suppress delayed cerebral ischemia after aneurysmal subarachnoid hemorrhage based on VASOGRADE: multicenter cohort study

Delayed cerebral ischemia (DCI) is one of the most important outcome determinants for aneurysmal subarachnoid hemorrhage (aSAH). VASOGRADE, which combines World Federation of Neurological Surgeons grade and modified Fisher grade, is a useful scale for predicting DCI after aSAH. However, no studies have investigated whether VASOGRADE influences the treatment options. We retrospectively analyzed 781 aSAH patients who were prospectively enrolled in 9 primary stroke centers from 2013 to 2021. The total cohort consisted of 76 patients (9.7%) with VASOGRADE-Green, 390 patients (49.9%) with VASOGRADE-Yellow, and 315 patients (40.3%) with VASOGRADE-Red. Worse VASOGRADE had higher incidences of DCI, which occurred in 190 patients (24.3%). As only 5 patients (6.6%) with VASOGRADE-Green developed DCI, we searched for DCI-associated factors in patients with VASOGRADEs-Yellow and -Red. Multivariate analyses revealed independent treatment factors suppressing DCI as follows: no postoperative hemorrhagic complication, combined administration of fasudil hydrochloride and cilostazol, combination of clipping and cisternal drainage, and coiling for VASOGRADE-Yellow; and clipping, and administration of fasudil hydrochloride with or without cilostazol for VASOGRADE-Red. The findings suggest that treatment strategies should be determined based on VASOGRADE to prevent DCI after aSAH.

Sex-Related Differences in Mortality, Delayed Cerebral Ischemia, and Functional Outcomes in Patients with Aneurysmal Subarachnoid Hemorrhage: A Systematic Review and Meta-Analysis

Background/Objective: Sex-related differences among patients with aneurysmal subarachnoid hemorrhage (aSAH) and their potential clinical implications have been insufficiently investigated. To address this knowledge gap, we conduct a comprehensive systematic review and meta-analysis. Methods: Sex-specific differences in patients with aSAH, including mortality, delayed cerebral ischemia (DCI), and functional outcomes were assessed. The functional outcome was dichotomized into favorable or unfavorable based on the modified Rankin Scale (mRS), Glasgow Outcome Scale (GOS), and Glasgow Outcome Scale Extended (GOSE). Results: Overall, 2823 studies were identified in EMBASE, MEDLINE, PubMed, and by manual search on 14 February 2024. After an initial assessment, 74 studies were included in the meta-analysis. In the analysis of mortality, including 18,534 aSAH patients, no statistically significant differences could be detected (risk ratio (RR) 0.99; 95% CI, 0.90-1.09; p = 0.91). In contrast, the risk analysis for DCI, including 23,864 aSAH patients, showed an 11% relative risk reduction in DCI in males versus females (RR, 0.89; 95% CI, 0.81-0.97; p = 0.01). The functional outcome analysis (favorable vs. unfavorable), including 7739 aSAH patients, showed a tendency towards better functional outcomes in men than women; however, this did not reach statistical significance (RR, 1.02; 95% CI, 0.98-1.07; p = 0.34). Conclusions: In conclusion, the available data suggest that sex/gender may play a significant role in the risk of DCI in patients with aSAH, emphasizing the need for sex-specific management strategies.

Impact of Medical and Neurologic Complications on the Outcome of Patients with Aneurysmal Subarachnoid Hemorrhage in a Middle-Income Country

OBJECTIVE

Almost two thirds of the world's aneurysmal subarachnoid hemorrhage (aSAH) are in low- and middle-income countries. Herein, we aimed to evaluate the impact of complications on the outcome of aSAH in a middle-income country.

METHODS

Baseline data (age, sex, World Federation of Neurosurgical Society, time ictus-treatment, treatment modality) and medical and neurologic complications from a cohort in Brazil (2016-2019) were evaluated: delayed cerebral ischemia; hydrocephalus; meningitis; seizures; intracranial hypertension; infections (pneumonia, bloodstream, urinary tract infection infection of undetermined source); sodium disturbances; acute kidney injury; and cardiac and pulmonary complications. The primary outcome was the modified Rankin scale (mRS) at hospital discharge. Univariate and multivariate models were employed.

RESULTS

From 212 patients (71.7% female, age 52.7 ± 12.8), 92% developed at least 1 complication (any infection-43.9%, hydrocephalus-34.4%, intracranial hypertension-33%, infection of undetermined source-20.8%, hypernatremia-20.8%, hyponatremia-19.8%, delayed cerebral ischemia-related infarction-18.7%, pneumonia-18.4%, acute kidney injury-16.5%, and seizures-11.8%). In unadjusted analysis, all but hyponatremia and urinary tract infection were associated with mRS 3-6 at discharge; however, complications explained only 12% of the variation in functional outcome (mRS). Most patients were treated by clipping (66.5%), and 15.6% (33 patients) did not receive a definitive treatment. The median time ictus-admission and ictus-treatment were 5 and 9 days, respectively.

CONCLUSIONS

While medical and neurologic complications are a recognized opportunity to improve aSAH care, low- and middle-income countries comprise 70% of the world population and still encounter difficulties concerning early definitive aneurysm treatment, rebleeding, and human and material resources.

The post-resuscitation VASOGRADE: a more accurate scale to predict delayed cerebral ischemia after aneurysmal subarachnoid hemorrhage

BACKGROUND

Predicting the occurrence of delayed cerebral ischemia (DCI) after aneurysmal subarachnoid hemorrhage is of interest to adjust the level of care. The VASOGRADE, a simple grading scale using admission World Federation of Neurosurgical Societies (WFNS) grading score and modified Fisher scale (mFS) on first CT scan, could help to select patients at risk of DCI. However, using data after initial resuscitation (initial complication treatment, aneurysm exclusion) may be more relevant.

METHODS

We calculated a post-resuscitation VASOGRADE (prVG) using WFNS grade and mFS after early brain injury treatment and aneurysm exclusion (or at day 3). Patients were categorized as green, yellow, or red.

RESULTS

Using our prospective observational registry, 566 patients were included in the study. Two hundred six (36.4%) were classified as green, 208 (36.7%) as yellow, and 152 (26.9%) as red, and DCI was experienced in 22 (10.7%), 67 (32.2%), and 45 (29.6%) cases respectively. Patients classified as yellow had higher risk of developing DCI (OR 3.94, 95% CI 2.35-6.83). Risk was slightly lower in red patients (OR 3.49, 95% CI 2.00-6.24). The AUC for prediction was higher with prVG (0.62, 95% CI 0.58-0.67) than with VASOGRADE (0.56, 95% CI 0.51-0.60) (p < 0.01).

CONCLUSION

By using simple clinical and radiological scale evaluated at subacute stage, prVG is more accurate to predict the occurrence of DCI.

Diagnosis of Delayed Cerebral Ischemia in Patients with Aneurysmal Subarachnoid Hemorrhage and Triggers for Intervention

INTRODUCTION

Delayed cerebral ischemia (DCI) is a major determinant for poor neurological outcome after aneurysmal subarachnoid hemorrhage (aSAH). Detection and treatment of DCI is a key component in the neurocritical care of patients with aSAH after initial aneurysm repair.

METHODS

Narrative review of the literature.

RESULTS

Over the past 2 decades, there has been a paradigm shift away from macrovascular (angiographic) vasospasm as a main diagnostic and therapeutic target. Instead, the pathophysiology of DCI is hypothesized to derive from several proischemic pathomechanisms. Clinical examination remains the most reliable means for monitoring and treatment of DCI, but its value is limited in comatose patients. In such patients, monitoring of DCI is usually based on numerous neurophysiological and/or radiological diagnostic modalities. Catheter angiography remains the gold standard for the detection of macrovascular spasm. Computed tomography (CT) angiography is increasingly used instead of catheter angiography because it is less invasive and may be combined with CT perfusion imaging. CT perfusion permits semiquantitative cerebral blood flow measurements, including the evaluation of the microcirculation. It may be used for prediction, early detection, and diagnosis of DCI, with yet-to-prove benefit on clinical outcome when used as a screening modality. Transcranial Doppler may be considered as an additional noninvasive screening tool for flow velocities in the middle cerebral artery, with limited accuracy in other cerebral arteries. Continuous electroencephalography enables detection of early signs of ischemia at a reversible stage prior to clinical manifestation. However, its widespread use is still limited because of the required infrastructure and expertise in data interpretation. Near-infrared spectroscopy, a noninvasive and continuous modality for evaluation of cerebral blood flow dynamics, has shown conflicting results and needs further validation. Monitoring techniques beyond neurological examinations may help in the detection of DCI, especially in comatose patients. However, these techniques are limited because of their invasive nature and/or restriction of measurements to focal brain areas.

CONCLUSION

The current literature review underscores the need for incorporating existing modalities and developing new methods to evaluate brain perfusion, brain metabolism, and overall brain function more accurately and more globally.

The Role of Early Serum Biomarkers and Clinical Rating Scales in the Prediction of Delayed Cerebral Ischaemia and Short-Term Outcome after Aneurysmal Subarachnoid Haemorrhage: Single Centre Experience

Considering the variety of complications that arise after aneurysmal subarachnoid haemorrhage (aSAH) and the complex pathomechanism of delayed cerebral ischaemia (DCI), the task of predicting the outcome assumes a profound complexity. Therefore, there is a need to develop early predictive and decision-making models. This study explores the effect of serum biomarkers and clinical scales on patients' outcomes and their interrelationship with DCI and systemic complications in aSAH. This was a retrospective analysis including aSAH patients admitted to the Wroclaw University Hospital (Wrocław, Poland) from 2011 to 2020. A good outcome was defined as a modified Rankin Scale (mRS) score of 0-2. The prediction of the development of DCI and poor outcome was conducted using logistic regression as a standard model (SM) and random forest as a machine learning method (ML). A cohort of 174 aSAH patients were included in the analysis. DCI was diagnosed in 79 (45%) patients. Significant differences between patients with poor vs. good outcome were determined from their levels of albumin (31 ± 7 vs. 35 ± 5 (g/L); p < 0.001), D-dimer (3.0 ± 4.5 vs. 1.5 ± 2.8 (ng/mL); p < 0.001), procalcitonin (0.2 ± 0.4 vs. 0.1 ± 0.1 (ng/mL); p < 0.001), and glucose (169 ± 69 vs. 137 ± 48 (nmol/L); p < 0.001). SM for DCI prediction included the Apache II scale (odds ratio [OD] 1.05; 95% confidence interval [CI] 1.00-1.09) and albumin level (OD 0.88; CI 0.82-0.95). ML demonstrated that low albumin level, high Apache II scale, increased D-dimer and procalcitonin levels had the highest predictive values for DCI. The integration of clinical parameters and scales with a panel of biomarkers may effectively facilitate the stratification of aSAH patients, identifying those at high risk of secondary complications and poor outcome.