Vasospasm following aneurysmal subarachnoid hemorrhage: Thrombocytopenia a marker

Machine learning predicts cerebral vasospasm in patients with subarachnoid haemorrhage

BACKGROUND

Cerebral vasospasm (CV) is a feared complication which occurs after 20-40% of subarachnoid haemorrhage (SAH). It is standard practice to admit patients with SAH to intensive care for an extended period of resource-intensive monitoring. We used machine learning to predict CV requiring verapamil (CVRV) in the largest and only multi-center study to date.

METHODS

Patients with SAH admitted to UCLA from 2013 to 2022 and a validation cohort from VUMC from 2018 to 2023 were included. For each patient, 172 unique intensive care unit (ICU) variables were extracted through the primary endpoint, namely first verapamil administration or no verapamil. At each institution, a light gradient boosting machine (LightGBM) was trained using five-fold cross validation to predict the primary endpoint at various hospitalization timepoints.

FINDINGS

A total of 1750 patients were included from UCLA, 125 receiving verapamil. LightGBM achieved an area under the ROC (AUC) of 0.88 > 1 week in advance and ruled out 8% of non-verapamil patients with zero false negatives. Our models predicted "no CVRV" vs "CVRV within three days" vs "CVRV after three days" with AUCs = 0.88, 0.83, and 0.88, respectively. From VUMC, 1654 patients were included, 75 receiving verapamil. VUMC predictions averaged within 0.01 AUC points of UCLA predictions.

INTERPRETATION

We present an accurate and early predictor of CVRV using machine learning with multi-center validation. This represents a significant step towards optimized clinical management and resource allocation in patients with SAH.

FUNDING

Robert E. Freundlich is supported by National Center for Advancing Translational Sciences federal grant UL1TR002243 and National Heart, Lung, and Blood Institute federal grant K23HL148640; these funders did not play any role in this study. The National Institutes of Health supports Vanderbilt University Medical Center which indirectly supported these research efforts. Neither this study nor any other authors personally received financial support for the research presented in this manuscript. No support from pharmaceutical companies was received.

Machine Learning Predicts Cerebral Vasospasm in Subarachnoid Hemorrhage Patients

Background

Cerebral vasospasm (CV) is a feared complication occurring in 20-40% of patients following subarachnoid hemorrhage (SAH) and is known to contribute to delayed cerebral ischemia. It is standard practice to admit SAH patients to intensive care for an extended period of vigilant, resource-intensive, clinical monitoring. We used machine learning to predict CV requiring verapamil (CVRV) in the largest and only multi-center study to date.

Methods

SAH patients admitted to UCLA from 2013-2022 and a validation cohort from VUMC from 2018-2023 were included. For each patient, 172 unique intensive care unit (ICU) variables were extracted through the primary endpoint, namely first verapamil administration or ICU downgrade. At each institution, a light gradient boosting machine (LightGBM) was trained using five- fold cross validation to predict the primary endpoint at various timepoints during hospital admission. Receiver-operator curves (ROC) and precision-recall (PR) curves were generated.

Results

A total of 1,750 patients were included from UCLA, 125 receiving verapamil. LightGBM achieved an area under the ROC (AUC) of 0.88 an average of over one week in advance, and successfully ruled out 8% of non-verapamil patients with zero false negatives. Minimum leukocyte count, maximum platelet count, and maximum intracranial pressure were the variables with highest predictive accuracy. Our models predicted "no CVRV" vs "CVRV within three days" vs "CVRV after three days" with AUCs=0.88, 0.83, and 0.88, respectively. For external validation at VUMC, 1,654 patients were included, 75 receiving verapamil. Predictive models at VUMC performed very similarly to those at UCLA, averaging 0.01 AUC points lower.

Conclusions

We present an accurate (AUC=0.88) and early (>1 week prior) predictor of CVRV using machine learning over two large cohorts of subarachnoid hemorrhage patients at separate institutions. This represents a significant step towards optimized clinical management and improved resource allocation in the intensive care setting of subarachnoid hemorrhage patients.

Biomarkers in aneurysmal subarachnoid hemorrhage: A short review

Poor outcomes of aneurysmal subarachnoid hemorrhage (aSAH) can be the result of the initial catastrophic event or the many acute or delayed neurological complications. Recent evidence suggests that some molecules play a critical role in both events, through some unknown pathways involved. Understanding the role of these molecules in these events could allow to improve diagnostic accuracy, guide management, and prevent long-term disability in aSAH. Here we present the studies on aSAH biomarkers present in current medical literature, highlighting their roles and main results.

Role of platelets in the pathogenesis of delayed injury after subarachnoid hemorrhage

Aneurysmal subarachnoid hemorrhage (aSAH) patients develop delayed cerebral ischemia and delayed deficits (DCI) within 2 weeks of aneurysm rupture at a rate of approximately 30%. DCI is a major contributor to morbidity and mortality after SAH. The cause of DCI is multi-factorial with contributions from microthrombi, blood vessel constriction, inflammation, and cortical spreading depolarizations. Platelets play central roles in hemostasis, inflammation, and vascular function. Within this review, we examine the potential roles of platelets in microthrombi formation, large artery vasospasm, microvessel constriction, inflammation, and cortical spreading depolarization. Evidence from experimental and clinical studies is provided to support the role(s) of platelets in each pathophysiology which contributes to DCI. The review concludes with a suggestion for future therapeutic targets to prevent DCI after aSAH.

miRNA expression profiling of cerebrospinal fluid in patients with aneurysmal subarachnoid hemorrhage

OBJECTIVE MicroRNAs (miRNAs) regulate gene expression and therefore play important roles in many physiological and pathological processes. The aim of this pilot study was to determine the feasibility of extraction and subsequent profiling of miRNA from CSF samples in a pilot population of aneurysmal subarachnoid hemorrhage patients and establish if there is a distinct CSF miRNA signature between patients who develop cerebral vasospasm and those who do not. METHODS CSF samples were taken at various time points during the clinical management of a subset of SAH patients (SAH patient samples without vasospasm, n = 10; SAH patient samples with vasospasm, n = 10). CSF obtained from 4 patients without SAH was also included in the analysis. The miRNA was subsequently isolated and purified and then analyzed on an nCounter instrument using the Human V2 and V3 miRNA assay kits. The data were imported into the nSolver software package for differential miRNA expression analysis. RESULTS From a total of 800 miRNAs that could be detected with each version of the miRNA assay kit, a total of 691 miRNAs were communal to both kits. There were 36 individual miRNAs that were differentially expressed (p < 0.01) based on group analyses, with a number of miRNAs showing significant changes in more than one group analysis. The changes largely reflected differences between non-SAH and SAH groups. These included miR-204-5p, miR-223-3p, miR-337-5p, miR-451a, miR-489, miR-508-3p, miR-514-3p, miR-516-5p, miR-548 m, miR-599, miR-937, miR-1224-3p, and miR-1301. However, a number of miRNAs did exclusively differ between the vasospasm and nonvasospasm SAH groups including miR-27a-3p, miR-516a-5p, miR-566, and miR-1197. CONCLUSIONS The findings indicate that temporal miRNA profiling can detect differences between CSF from aneurysmal SAH and non-SAH patients. Moreover, the miRNA profile of CSF samples from patients who develop cerebral vasopasm may be distinguishable from those who do not. These results provide a foundation for future research at identifying novel CSF biomarkers that might predispose to the development of cerebral vasospasm after SAH and therefore influence subsequent clinical management.

Impact of Early Leukocytosis and Elevated High-Sensitivity C-Reactive Protein on Delayed Cerebral Ischemia and Neurologic Outcome After Subarachnoid Hemorrhage

BACKGROUND

The role of inflammatory response in the pathophysiology of subarachnoid hemorrhage (SAH) is being increasingly recognized. This study analyzed the impact of cellular and biochemical markers of early inflammatory response to ictus on outcome after SAH.

METHODS

Patients with SAH were prospectively studied for markers of early cellular, biochemical, and cytotoxic inflammatory response, including total leukocyte count (TLC), high-sensitivity C-reactive protein (hs-CRP), and lactate dehydrogenase. The relationship of these markers to delayed cerebral ischemia (DCI), new infarct, and Glasgow Outcome Scale (GOS) score at 3 months was studied.

RESULTS

The study comprised 246 patients. Of patients, 94 who developed DCI had a significantly higher TLC [± SD] (11.2 × 10(3)/mm(3) [± 4.0] vs. 9.4 × 10(3)/mm(3) [± 2.9], P = 0.001) and 62 with new infarct had significantly higher TLC (11.0 × 10(3)/mm(3) [± 3.6] vs. 9.8 × 10(3)/mm(3) [± 3.4], P = 0.05). GOS score had a significant inverse relationship to TLC at admission. The mean TLC [± SD] was 12.7 × 10(3)/mm(3) [± 4.2], 11.7 × 10(3)/mm(3) [± 3.1], 10.2 × 10(3)/mm(3) [± 3.4], and 9.3 × 10(3)/mm(3) [± 2.8] among patients with GOS scores of 1, 3, 4, and 5 (P < 0.001). hs-CRP showed a trend of an inverse relationship to GOS score in univariate analysis. Lactate dehydrogenase had no relationship with any outcome parameter. In multivariate analysis, higher admission TLC had a significant association with DCI (P = 0.01) and poorer GOS score (P < 0.001), and higher hs-CRP had a significant association with poorer GOS score (P = 0.05).

CONCLUSIONS

A leukocytosis response to ictus seems to have a significant independent association with both DCI and poor GOS score, and hs-CRP level had a significant independent association with poor GOS score, indicating preeminence of early cellular response in SAH pathophysiology.