BACKGROUND:Intracerebral haemorrhage with ventricular extension is associated with poor neurological outcome and high mortality. External ventricular drainage (EVD) is a standard intervention for acute hydrocephalus in such patients, yet the role of an additional early lumbar drainage (LD) remains uncertain. AIM:To evaluate whether early LD leads to improved functional outcomes in patients suffering from ICH with intraventricular haemorrhage (IVH) who require an EVD, compared to patients not receiving early LD. STUDY DESIGN:DRAIN IVH is a multi-centre, parallel-group, open-label randomised controlled trial with blinded endpoint evaluation. Key inclusion criteria are ICH with IVH in the third and/or fourth ventricle with the need for EVD placement due to acute hydrocephalus, age ≥ 18 years and LD insertion within 72 h after symptom onset or last-seen-well. SAMPLE SIZE:A total of 354 patients will be randomised for either early LD plus standard of care (comprising EVD) or standard of care alone. STUDY ENDPOINTS:The primary outcome is the rate of modified Rankin Scale score 0-3 at 180 days. Secondary outcomes include need for VP shunt, need for tracheostomy and EQ-5D. Safety outcomes include mortality, bacterial ventriculitis/meningitis, EVD complications and LD complications. SUMMARY:DRAIN IVH will provide evidence for physicians whether early LD in addition to standard of care treatment leads to better outcomes in ICH with IVH compared to standard of care alone. TRIAL REGISTRATION:Clinicaltrials.gov; NCT06510842.

The multicenter trial SETPOINT2 comparing early versus delayed tracheostomy in severe stroke was conducted in Germany and the USA. A pre-specified aim was to assess 6-month patient- and caregiver-reported outcomes to better inform clinical decision-making in the acute phase of care, irrespective of tracheostomy timing. Patients had acute ischemic, hemorrhagic, or subarachnoid hemorrhage (SAH) strokes, respiratory failure, and a predicted necessity for tracheostomy (SETscore > 10). Outcome measures assessed by telephone interview 6 months after hospital admission included the modified Rankin scale (mRS), EuroQol 5D-5L (EQ-5D-5L), EuroQol Visual Analog Scale, Burden Scale for Family Caregivers short form (BSFC-s), and questions related to satisfaction with the treatment results. Of 366 patients with available data at 6 months, 118 (32

Delayed cerebral ischemia (DCI) occurs several days after subarachnoid hemorrhage (SAH) and could therefore be mitigated by treatment intervention. As testament to the great challenge of designing effective trials, however, the prevalence and severity of DCI have remained unchanged for decades. A critical need in the development of novel prophylactic treatments is optimizing initial prediction for DCI, since improved risk stratification based on prognosis could increase the statistical power of randomized trials, thereby increasing the probability of success and reducing costs. The primary objective of this study, using secondary analysis of data from the DISCHARGE-1 trial, was to evaluate the power of initial variables (medical history, clinical examination, and pre-interventional CT) for predicting the volume of delayed infarcts in SAH survivors. In multivariate analysis of 164 early survivors with pre-interventional CT within 72 h of the initial hemorrhage, the modified Hijdra scale (mHS; β = -0.016, p < 0.001) was the only independent predictor of infarct volume due to DCI, outperforming common clinical scores und manually segmented hemorrhage volumes. Furthermore, early brain injury (EBI) volume identified high-risk patients for early death early on. Thus, we propose a combined approach using mHS and EBI volume for early risk stratification in randomized trials targeting DCI.

Intracranial hypertension (IH) is a major cause of secondary brain injury after traumatic brain injury (TBI) and is typically treated after onset. Machine learning (ML) approaches show promise for IH prediction, but methodological heterogeneity across studies limits comparability and clinical translation. We evaluated how key methodological choices influence ML performance for IH prediction. In this retrospective observational study, intracranial pressure (ICP), arterial blood pressure (ABP), and additional variables were collected from TBI patients in two Dutch hospitals and the CENTER-TBI high-resolution cohort. Six methodological categories were varied: IH definition (ICP ≥ 22 or ≥ 30 mmHg), input data configuration, labelling strategy (IH onset versus continuous labelling), temporal separation of control segments, observation window length (10–600 min), and model class (eXtreme Gradient Boosting, random forest, logistic regression, Gaussian naïve Bayes, and long short-term memory). Models were evaluated using patient-level tenfold cross-validation with area under the receiver operating characteristic curve (AUC) as primary outcome. We included 316 patients. Prediction performance varied substantially across methodological choices. Continuous labelling inflated performance, and optimal observation windows were 240–360 min. Across conditions, eXtreme Gradient Boosting achieved the highest performance (AUC 0.92). These findings highlight the need for standardized methodology and transparent reporting in IH prediction studies.

Aneurysmal subarachnoid hemorrhage is a critical condition with high case-fatality and lasting impacts on survivors. Acute events that are the direct result of aneurysm rupture, such as acute ischemia, elevated intracranial pressure, cerebral edema, seizures, and hydrocephalus, lead to early brain injury. A delayed cascade of processes, including a prominent systemic inflammatory response, may lead to secondary brain injury and delayed cerebral ischemia, which often further impairs recovery. Systemic complications, including cardiac and pulmonary dysfunction, fever, and electrolyte imbalances, arise in the interplay between early and secondary brain injury and challenge the clinical course. Early management focuses on the prevention of rebleeding mainly through aneurysm securement, amelioration of early brain injury through cerebrospinal fluid drainage, control of intracranial pressure, and organ support to avoid or attenuate secondary brain injury. Nimodipine remains the only pharmacological agent shown to reduce delayed cerebral ischemia, and lumbar drainage of cerebrospinal fluid to reduce subarachnoid blood may improve outcome. Management strategies for hemodynamic interventions, seizures, intracranial pressure control, large artery vasospasm, and electrolytes remain consensus-based and with large variation in practice. Several advances in understanding inflammation and delayed cerebral ischemia, as well as in monitoring and interventions hold promise, but robust trials are needed to refine protocols and improve patient recovery. Understanding and mitigating the cascade of damage from rupture to recovery is essential to reduce the burden of this devastating condition. In this review, we appraise the current understanding of the pathophysiology of post-rupture complications as well as scientific and management data, with a focus on recent advances.

INTRODUCTION:The management of severe traumatic brain injury (sTBI) in the intensive care unit (ICU) is focused on preventing secondary brain insults, by ensuring adequate cerebral perfusion, oxygenation and substrate delivery. Despite optimisation of intracranial pressure (ICP) and cerebral perfusion pressure (CPP) using evidence-based guidelines, brain tissue hypoxia can still occur and is strongly associated with adverse functional outcomes post sTBI. METHODS AND ANALYSIS:The Brain Oxygen Neuromonitoring in Australia and New Zealand Assessment - Global Trial (BONANZA-GT) is an international, two-arm, open-label, parallel group, randomised controlled trial comparing sTBI management incorporating early brain tissue oxygen (PbtO2) monitoring and optimisation, with ICP/CPP-based management alone. A total of 860 adults admitted to participating institutions with non-penetrating sTBI and requiring insertion of an ICP monitor (as determined by the treating neurosurgeon) will be enrolled. The primary outcome is the proportion of patients with favourable neurological outcomes, as defined by a Glasgow Outcome Score-Extended (GOS-E) >4, at 6 months following injury. Key secondary outcomes include all-cause mortality at ICU discharge, hospital discharge, adverse events, as well as hospital and ICU length of stay and GOS-E at 12 months. The BONANZA-GT will determine whether a protocolised therapeutic strategy guided by continuous PbtO2 monitoring in addition to ICP/CPP targets results in improved neurological outcomes when compared with standard care using ICP/CPP-guided management alone. ETHICS AND DISSEMINATION:Approval has been obtained from relevant ethics boards in every jurisdiction that is participating in the trial. Inclusion of adults who lack capacity for informed consent will be governed in accordance with the legal requirements of each participating site. Study findings will be presented at scientific meetings and disseminated via peer-review publications. TRIAL REGISTRATION NUMBER:Australian and New Zealand Clinical Trials Registry (ACTRN 12619001328167).

Sedatives play an important role in the management of patients with severe traumatic brain injury (sTBI) in the intensive care unit (ICU). Benzodiazepines are common for sedation (midazolam-based) but have been discouraged for non-brain-injured patients in the ICU. This study aimed to investigate the effect of midazolam-based sedation versus non-midazolam-based sedation on the need for intracranial pressure (ICP) lowering therapies in patients with sTBI in the ICU. We studied patients with sTBI (Glasgow Coma Sale ≤8) from 14 ICUs in Europe and Australia, who received ICP monitoring and continuous instrumental variable (IV) sedation for at least 24 h. We analyzed the association between sedation strategy and the need for ICP lowering therapies during the first 7 ICU days using a multivariable logistic regression model, adjusted for clinical markers of injury severity. We also analyzed the center as an IV in a random effects model to address potentially unmeasured confounding. Among 227 patients with sTBI, 152 (67%) received midazolam-based sedation. These patients had a lower age and higher median Glasgow Coma Scale on admission compared with 75 patients in the non-midazolam-sedated group. In logistic regression analyses, patients with midazolam-based sedation had higher odds of receiving hyperosmolar therapy (odds ratio [OR]: 3.4, 95% confidence intervals [CI]: 1.6-7.7). This effect could not be confirmed in the instrumental variable analysis (hyperosmolar therapy: OR: 1.3, 95% CI: 0.1-13.1). The mean ICU length of stay was significantly longer in the midazolam-based sedation group compared with the non-midazolam-based sedation group (19 vs. 13 days, hazards ratio 0.6, 95% CI: 0.4-0.8). Midazolam-based sedation was common for patients with sTBI without a significantly increased need for ICP therapies but an association with longer ICU stay. Larger prospective comparative effectiveness studies are needed regarding sedation strategies in critically ill patients with TBI.

Fever is associated with worse outcomes in patients with acute brain injury. Diclofenac, a non-steroidal anti-inflammatory drug, is commonly used as antipyretic therapy. As evidence emerged that short diclofenac infusions (< 1 h) decrease brain tissue oxygen (PtO2) and cerebral perfusion pressure (CPP), clinical practice has shifted to extended infusions (12 h). The purpose of this study was to investigate the effects of extended diclofenac infusion for the treatment of fever on cerebral perfusion and tissue oxygenation after acute brain injury. We conducted a retrospective study of prospectively collected data from a cohort of 18 patients with acute brain injury and PtO2 monitoring admitted between November 2018 and April 2024. The hour before and the 12 h during an extended diclofenac infusion were compared. Additionally, we compared the 12 h prior and 12 h during the diclofenac infusion. Cerebral autoregulation and metabolites obtained by microdialysis were assessed in a subgroup of patients. Thirty-nine interventions were analyzed. Core temperature decreased from 38.1°C in the hour before to 37.4 °C during an extended diclofenac infusion (p < 0.0001). ICP (11.0 vs 10.0 mmHg, p < 0.0001) and heart rate (84 vs. 77 bpm, p < 0.0001) decreased. CPP and PaCO2 did not vary significantly. PtO2 decreased from 23.1 mmHg (IQR 19.0–31.4) during fever peak to 21.7 mmHg (IQR 17.8–27.2) (p < 0.0001). Median PtO2 during the 12 h before diclofenac was 23.3 mmHg (IQR 18.9–30.5). In a multivariable analysis the effect of treatment was significantly influenced by heart rate and temperature (p < 0.0001). Extended diclofenac infusions for the treatment of fever in patients with acute brain injury achieve a clinically significant reduction in temperature but are associated with a small decrease in PtO2, even in the setting of maintained CPP.

BACKGROUND:We determined the predictive power of semi-automated blood-brain barrier assessment and other variables collected during neurocritical care for the outcome of 'epilepsy or late death' following aneurysmal subarachnoid haemorrhage. METHODS:This is a secondary analysis of the prospective, non-interventional, prognostic DISCHARGE-1-cohort from six university hospitals in Germany. All patients who underwent at least one contrast-enhanced MRI during neurocritical care were included. Initial clinical scores and Modified Rankin Scale at day 14 were available. Subdural electrocorticography was scored for seizures and spreading depolarisations. Two MRIs, one post-aneurysm occlusion and another post-neuromonitoring, were semi-automatically segmented into cerebrospinal fluid spaces, normal brain tissue, and abnormal brain tissue. Normal and abnormal tissue were further classified into tissue with "intact" or "dysfunctional" blood-brain barrier. Epilepsy and late death were determined at a median of 3.7 years. FINDINGS:Abnormal, barrier-dysfunctional tissue as a percentage of intracranial volume on post-monitoring MRI was the only independent predictor of early death within three weeks among 130 patients. In the 121 early survivors, this variable was also the only independent predictor of 'epilepsy or late death'. This result, obtained by a combination of imputation and the leaving-one-out method, was confirmed in two sensitivity analyses within smaller populations and with fewer missing values. INTERPRETATION:The study substantiates previous experimental evidence that blood-brain barrier dysfunction plays a key role in epileptogenesis after brain injuries. Contrast-enhanced MRI, a minimally invasive technique, highlighted abnormal, barrier-dysfunctional tissue as a stand-alone independent predictor, underscoring its potential as a 'precision medicine' tool in early diagnosis and intervention. FUNDING:JPD and AF report a grant from the Era-Net Neuron EBio2 with funds from BMBF 01EW2004 and CIHR Award No. NDD 168164. JPD reports a grant from DFG DR 323/10-2 (project number: 413848220) and EU Horizon MSCA-DN 101119916-SOPRANI. AF reports grants from the Canadian Institutes of Health Research (CIHR) PJT 148896 and Israel Science Foundation (ISF) 2254/20. NH is Berlin Institute of Health Clinical Fellow, funded by Stiftung Charité.

ImportanceCerebral vasospasm largely contributes to a devastating outcome after aneurysmal subarachnoid hemorrhage (aSAH), with limited therapeutic options.ObjectiveTo investigate the safety and efficacy of localized nicardipine release implants positioned around the basal cerebral vasculature at risk for developing proximal vasospasm after aSAH.Design, Setting, and ParticipantsThis single-masked randomized clinical trial with a 52-week follow-up was performed between April 5, 2020, and January 23, 2023, at 6 academic neurovascular centers in Germany and Austria. Consecutive patients with World Federation of Neurological Surgeons grade 3 or 4 aSAH due to a ruptured anterior circulation aneurysm requiring microsurgical aneurysm repair participated.InterventionDuring aneurysm repair, patients were randomized 1:1 to intraoperatively receive 10 implants at 4 mg of nicardipine each plus standard of care (implant group) or aneurysm repair alone plus standard of care (control group).Main Outcome and MeasuresThe primary end point was the incidence of moderate to severe cerebral angiographic vasospasm (aVS) between days 7 and 9 after aneurysm rupture as determined by digital subtraction angiography.ResultsOf 41 patients, 20 were randomized to the control group (mean [SD] age, 54.9 [9.1] years; 17 female [85%]) and 21 to the implant group (mean [SD] age, 53.6 [11.9] years; 14 female [67%]). A total of 39 patients were included in the primary efficacy analysis. In the control group, 11 of 19 patients (58%) developed moderate or severe aVS compared with 4 of 20 patients (20%) in the implant group (P = .02). This outcome was paralleled by a lower clinical need for vasospasm rescue therapy in the implant group (2 of 20 patients [10%]) compared with the control group (11 of 19 patients [58%]; P = .002). Between days 13 and 15 after aneurysm rupture, new cerebral infarcts were noted in 6 of 19 patients (32%) in the control group and in 2 of 20 patients (10%) in the implant group (P = .13). At 52 weeks, favorable outcomes were noted in 12 of 18 patients (67%) in the control group and 16 of 19 patients (84%) in the implant group (P = .27). The adverse event rate did not differ between groups.Conclusions and RelevanceThese findings show that placing nicardipine release implants during microsurgical aneurysm repair can provide safe and effective prevention of moderate to severe aVS after aSAH. A phase 3 clinical trial to investigate the effect of nicardipine implants on clinical outcome may be warranted.Trial RegistrationClinicalTrials.gov Identifier: NCT04269408

INTRODUCTION:Cerebral infarction from delayed cerebral ischemia (DCI) is a leading cause of poor neurological outcome after aneurysmal subarachnoid hemorrhage (aSAH). We performed an international clinical practice survey to identify monitoring and management strategies for cerebral vasospasm associated with DCI in aSAH patients requiring intensive care unit admission.METHODS:The survey questionnaire was available on the European Society of Intensive Care Medicine (May 2021-June 2022) and Neurocritical Care Society (April - June 2022) websites following endorsement by these societies.RESULTS:There were 292 respondents from 240 centers in 38 countries. In conscious aSAH patients or those able to tolerate an interruption of sedation, neurological examination was the most frequently used diagnostic modality to detect delayed neurological deficits related to DCI caused by cerebral vasospasm (278 respondents, 95.2%), while in unconscious patients transcranial Doppler/cerebral ultrasound was most frequently used modality (200, 68.5%). Computed tomography angiography was mostly used to confirm the presence of vasospasm as a cause of DCI. Nimodipine was administered for DCI prophylaxis by the majority of the respondents (257, 88%), mostly by an enteral route (206, 71.3%). If there was a significant reduction in arterial blood pressure after nimodipine administration, a vasopressor was added and nimodipine dosage unchanged (131, 45.6%) or reduced (122, 42.5%). Induced hypertension was used by 244 (85%) respondents as first-line management of DCI related to vasospasm; 168 (59.6%) respondents used an intra-arterial procedure as second-line therapy.CONCLUSIONS:This survey demonstrated variability in monitoring and management strategies for DCI related to vasospasm after aSAH. These findings may be helpful in promoting educational programs and future research.

The choice between clipping and coiling of ruptured cerebral aneurysms in subarachnoid hemorrhage (SAH) remains controversial. The recently published Earlydrain trial provides the opportunity to analyze the latest clip-to-coil ratio in German-speaking countries and to evaluate vasospasm incidence and explorative outcome measures in both treatment modalities. We performed a post hoc analysis of the Earlydrain trial, a multicenter randomized controlled trial investigating the use of an additional lumbar drain in aneurysmal SAH. The decision whether to clip or to coil the ruptured aneurysm was left to the discretion of the participating centers, providing a real-world insight into current aneurysm treatment strategies. Earlydrain was performed in 19 centers in Germany, Switzerland, and Canada, recruiting 287 patients with aneurysmal SAH of all severity grades. Of these, 140 patients (49

- BACKGROUND: Continuous bedside monitoring of brain tissue oxygen levels is a crucial component in the management of comatose patients suffering from acute brain injury on neurointensive care units. Ensuring sufficient brain oxygenation is recognized as an essential objective within neurocritical care, aimed at safeguarding patients from secondary ischemia. Hypoperfusion in occipital and the posterior watershed regions often remains undetected, as the placement of probes in these areas is challenging. A major concern is that patients would have to lie on the traditionally used implanted bolts due to the occipital entry point of the probes. Therefore, we present a novel techique compatible with magnetic resonance imaging that enables bedside placement of brain tissue oxygen probes without the use of a bolt in these areas. - METHODS: We conducted bedside implantations of Licox brain tissue oxygenation probes through Frazier 's point utilizing peripheral venous cannulas on burr holes eliminating the need for bolts. - RESULTS: A novel approach was successfully established for the bedside implantation of a Licox brain tissue oxygenation probe for occipital regions. - CONCLUSIONS: This technical note describes the feasibility of a novel, simple, and straightforward bedside technique for boltless implantation of Licox brain tissue oxygen probes leading to rigid fixation and compatibility with magnetic resonance imaging.

Objective: Detection of delayed cerebral ischemia (DCI) is challenging in comatose patients with poor-grade aneurysmal subarachnoid hemorrhage (aSAH). Brain tissue oxygen pressure (PbtO2) monitoring may allow early detection of its occurrence. Recently, a probe for combined measurement of intracranial pressure (ICP) and intraparenchymal near-infrared spectroscopy (NIRS) has become available. In this pilot study, the parameters PbtO2, Hboxy, Hbdeoxy, Hbtotal and rSO2 were measured in parallel and evaluated for their potential to detect perfusion deficits or cerebral infarction. Methods: In patients undergoing multimodal neuromonitoring due to poor neurological condition after aSAH, Clark oxygen probes, microdialysis and NIRS-ICP probes were applied. DCI was suspected when the measured parameters in neuromonitoring deteriorated. Thus, perfusion CT scan was performed as follow up, and DCI was confirmed as perfusion deficit. Median values for PbtO2, Hboxy, Hbdeoxy, Hbtotal and rSO2 in patients with perfusion deficit (Tmax > 6 s in at least 1 vascular territory) and/or already demarked infarcts were compared in 24- and 48-hour time frames before imaging. Results: Data from 19 patients (14 University Hospital Zurich, 5 Charite Universitatsmedizin Berlin) were prospectively collected and analyzed. In patients with perfusion deficits, the median values for Hbtotal and Hboxy in both time frames were significantly lower. With perfusion deficits, the median values for Hboxy and Hbtotal in the 24 h time frame were 46,3 [39.6, 51.8] mu mol/l (no perfusion deficits 53 [45.9, 55.4] mu mol/l, p = 0.019) and 69,3 [61.9, 73.6] mu mol/l (no perfusion deficits 74,6 [70.1, 79.6] mu mol/l, p = 0.010), in the 48 h time frame 45,9 [39.4, 51.5] mu mol/l (no perfusion deficits 52,9 [48.1, 55.1] mu mol/l, p = 0.011) and 69,5 [62.4, 74.3] mu mol/l (no perfusion deficits 75 [70,80] mu mol/l, p = 0.008), respectively. In patients with perfusion deficits, PbtO2 showed no differences in both time frames. PbtO2 was significantly lower in patients with infarctions in both time frames. The median PbtO2 was 17,3 [8,25] mmHg (with no infarctions 29 [22.5, 36] mmHg, p = 0.006) in the 24 h time frame and 21,6 [11.1, 26.4] mmHg (with no infarctions 31 [22,35] mmHg, p = 0.042) in the 48 h time frame. In patients with infarctions, the median values of parameters measured by NIRS showed no significant differences. Conclusions: The combined NIRS-ICP probe may be useful for early detection of cerebral perfusion deficits and impending DCI. Validation in larger patient collectives is needed.

OBJECTIVE: Aneurysmal subarachnoid hemorrhage (aSAH) from a ruptured intracranial aneurysm is a severe, life-threatening condition, with high morbidity and mortality. The current treatment often involves surgical clipping or endovascular treatment within the first 24 - 48 hours. Although there is ample evidence of complications in treating unruptured aneurysms, similar data in patients with acutely ruptured aneurysms are limited. The recently completed EARLYDRAIN trial showed improved neurologic results from lumbar drainage after aneurysm treatment in patients with aSAH. Using this data set, we aim to study the frequency and effects of complications and identify associated risk factors. - METHODS: A substudy was carried out of the prospective multicenter randomized controlled EARLYDRAIN trial. We analyzed treatment-associated complications (bleeding and/or infarctions) detected on computed tomography on day 1 after aneurysm occlusion. Outcomes were the occurrence of postprocedural complications, secondary infarctions in the acute phase, and the modified Rankin Scale score after 6 months. - RESULTS: The EARLYDRAIN trial recruited 287 patients in 19 centers. Of these patients, 56 (19.5%) experienced a treatment complication. Twenty-five patients (8.7%) experienced postprocedural intracranial hemorrhage and 34 patients (11.8%) experienced a treatment-associated infarction. Patients with a complication showed more secondary infarctions ( P = 0.049) and worse neurologic outcomes after 180 days ( P = 0.025) compared with patients with no complication. Aneurysm location, rebleeding before the treatment, number of patients recruited per center, and the day of the treatment were independent risk factors for the occurrence of complications. - CONCLUSIONS: The present study shows that patients with aSAH frequently experience intervention-associated complications associated with aneurysm occlusion required to prevent recurrent hemorrhage. Consequently, patients with aSAH with treatment-related complications more often experience a worse clinical course and poor outcome.

After aneurysmal subarachnoid hemorrhage (aSAH), elevated intracranial pressure (ICP) due to disrupted cerebrospinal fluid (CSF) dynamics is a critical concern. An external ventricular drainage (EVD) is commonly employed for management; however, optimal strategies remain debated. The randomized controlled Earlydrain trial showed that an additional prophylactic lumbar drainage (LD) after aneurysm treatment improves neurological outcome. We performed a post hoc investigation on the impact of drainage volumes and critical ICP values on patient outcomes after aSAH. Using raw patient data from Earlydrain, we analyzed CSF drainage amounts and ICP measurements in the first 8 days after aSAH. Outcomes were the occurrence of secondary infarctions and the score on the modified Rankin scale after 6 months, dichotomized in values of 0–2 as favorable and 3–6 as unfavorable. Repeated measurements were considered with generalized estimation equations. Earlydrain recruited 287 patients, of whom 221 received an EVD and 140 received an LD. Higher EVD volumes showed a trend to more secondary infarctions (p = 0.09), whereas higher LD volumes were associated with less secondary infarctions (p = 0.009). The mean total CSF drainage was 1052 ± 659 mL and did not differ concerning infarction and neurological outcome. Maximum ICP values were higher in patients with poor outcomes but not related to drainage volumes via EVD. After adjustment for aSAH severity and total CSF drainage, higher LD volume was linked to favorable outcome (per 100 mL: odds ratio 0.61 (95

OBJECTIVE In neurocritical care, data from multiple biosensors are continuously measured, but only sporadically acknowledged by the attending physicians. In contrast, machine learning (ML) tools can analyze large amounts of data continuously, taking advantage of underlying information. However, the performance of such ML-based solutions is limited by different factors, for example, by patient motion, manipulation, or, as in the case of external ventricular drains (EVDs), the drainage of CSF to control intracranial pressure (ICP). The authors aimed to develop an ML-based algorithm that automatically classifies normal signals, artifacts, and drainages in high-resolution ICP monitoring data from EVDs, making the data suitable for real-time artifact removal and for future ML applications. METHODS In their 2-center retrospective cohort study, the authors used labeled ICP data from 40 patients in the first neurocritical care unit (University Hospital Zurich) for model development. The authors created 94 descriptive features that were used to train the model. They compared histogram-based gradient boosting with extremely randomized trees after building pipelines with principal component analysis, hyperparameter optimization via grid search, and sequential feature selection. Performance was measured with nested 5-fold cross-validation and multiclass area under the receiver operating characteristic curve (AUROC). Data from 20 patients in a second, independent neurocritical care unit (Charité - Universitätsmedizin Berlin) were used for external validation with bootstrapping technique and AUROC. RESULTS In cross-validation, the best-performing model achieved a mean AUROC of 0.945 (95% CI 0.92–0.969) on the development dataset. On the external validation dataset, the model performed with a mean AUROC of 0.928 (95% CI 0.908–0.946) in 100 bootstrapping validation cycles to classify normal signals, artifacts, and drainages. CONCLUSIONS Here, the authors developed a well-performing supervised model with external validation that can detect normal signals, artifacts, and drainages in ICP signals from patients in neurocritical care units. For future analyses, this is a powerful tool to discard artifacts or to detect drainage events in ICP monitoring signals.