Effects of Cerebrospinal Fluid Drainage on Cerebral Hemodynamics

Evaluating the Neuroprotective Effects of Continuous Lumbar Drainage on the Cerebral Perfusion in Patients With an Acute Ischemic Stroke and Hemodynamic Failure Type 2: a Pilot Study

This pilot interventional study investigates whether continuous cerebrospinal fluid (CSF) drainage via lumbar drain can improve cerebrovascular reactivity (CVR) in patients with acute ischemic stroke and hemodynamic failure, as assessed by blood oxygenation level dependent (BOLD) MRI.

Patients with acute ischemic stroke frequently exhibit impaired cerebral autoregulation and reduced vascular reserve. In such cases, a low BOLD-CVR response is associated with a high (~40%) risk of recurrent ischemic stroke despite optimal medical therapy. Hemodynamic failure stage 2 (HF2), defined by severely reduced CVR (hemispheric BOLD-CVR ≤ 0.1041), indicates exhausted compensatory capacity and poor perfusion.

Evidence from other neurological conditions suggests that CSF drainage via lumbar drain may transiently lower intracranial pressure, improve perfusion, and enhance autoregulatory function. This study aims to apply these findings to acute stroke patients with hemodynamic compromise.

Eligible patients will undergo BOLD-CVR MRI to assess cerebrovascular reserve. Those meeting HF2 criteria and without contraindications will be offered inclusion. After informed consent (or proxy consent), a lumbar drain will be placed, and continuous CSF drainage performed over ~48 hours. A follow-up BOLD-CVR MRI will assess changes in CVR. In addition, 20 ml of CSF will be extracted to evaluate acute effects on CVR. A historical cohort will be used to control for spontaneous CVR variation over 48 hours.

Primary Objective:

To determine whether continuous CSF drainage improves BOLD-CVR in patients with HF2 following acute ischemic stroke.

Secondary Objectives:

To compare longitudinal BOLD-CVR changes to a historical cohort. To evaluate safety and feasibility of lumbar drainage in this setting. To assess short-term clinical outcomes (NIHSS, mRS) and recurrent events. To assess BOLD-CVR change after 20 ml CSF extraction.

Study Design:

Single-center, open-label, prospective pilot study. Participants serve as their own control, with pre- and post-intervention imaging to evaluate physiological effects. Safety and exploratory clinical data will also be collected.

This proof-of-concept trial aims to generate preliminary evidence for a potential therapeutic strategy in high-risk stroke patients with poor autoregulation. Positive findings could inform future randomized controlled trials.

This pilot interventional study explores the physiological and clinical impact of continuous cerebrospinal fluid (CSF) drainage via lumbar drain on cerebrovascular reactivity (CVR) in patients with acute ischemic stroke who exhibit hemodynamic failure, as measured by blood oxygenation level-dependent (BOLD) functional MRI.

Background and Rationale In patients with acute ischemic stroke, failure of autoregulatory capacity and exhaustion of cerebrovascular reserve represent major contributors to secondary ischemic injury and poor outcomes. Despite optimal medical management, patients with documented hemodynamic failure remain at high risk for recurrent ischemic stroke, often exceeding 30-40% within the first months post-event.

BOLD-CVR is a non-invasive MRI-based technique that maps the cerebrovascular response to a vasodilatory stimulus (commonly CO₂ or breath-hold induced hypercapnia). When interpreted using standardized processing pipelines, it offers a quantitative assessment of cerebral hemodynamics and can differentiate between regions with preserved, reduced, or exhausted vascular reserve. In previous research, a hemispheric mean BOLD-CVR ≤ 0.1041 or a middle cerebral artery (MCA) territory mean BOLD-CVR ≤ 0.0778 has been associated with hemodynamic failure stage 2 (HF2).

There is preliminary evidence from other neurological conditions (e.g., idiopathic intracranial hypertension, traumatic brain injury, and subarachnoid hemorrhage) that CSF diversion through lumbar drainage can lead to improved cerebral perfusion and functional outcomes. The proposed mechanism involves a transient reduction in intracranial pressure (ICP), improved compliance, and favorable shifts in cerebral perfusion pressure gradients. However, this strategy has not been evaluated in the context of acute ischemic stroke in patients without hydrocephalus but with severe autoregulatory impairment.

This study hypothesizes that temporary CSF drainage in such patients will improve BOLD-CVR as an imaging marker of vascular reserve and possibly influence clinical outcomes.

Study Objectives

Primary Objective:

To assess whether continuous CSF drainage via lumbar drain results in measurable improvement of cerebrovascular reactivity (BOLD-CVR) in acute ischemic stroke patients with HF2.

Secondary Objectives:

To evaluate the safety and technical feasibility of lumbar drain placement in acute stroke patients without hydrocephalus.

To explore changes in neurological function (NIH Stroke Scale and modified Rankin Scale) at discharge and during follow-up.

To monitor for recurrent ischemic stroke events during the subacute period post-intervention.

To compare observed CVR changes with those occurring naturally over time in a historical cohort with matched inclusion criteria but no CSF drainage.

Study Design and Methodology This is a prospective, single-center, open-label, pilot interventional trial. Patients admitted with acute ischemic stroke will undergo BOLD-CVR MRI within a defined time window (typically within 3-7 days post-event) to assess cerebrovascular reserve.

Inclusion criteria require radiologically confirmed acute ischemic stroke and a BOLD-CVR profile consistent with HF2. Patients with hydrocephalus, mass effect, active CNS infection, or contraindications to lumbar puncture are excluded.

Intervention: Eligible and consenting patients will undergo placement of a lumbar drain by experienced neurosurgical staff. Continuous drainage will be performed over a period of 48-72 hours using a pressure-controlled closed CSF system. Drainage volumes and pressures will be recorded in standardized templates.

Imaging Protocol: BOLD-CVR imaging will be performed at two time points:

Baseline scan: prior to lumbar drainage.

Follow-up scan: 12-24 hours after completion of drainage.

All imaging will be performed on the same scanner with identical acquisition parameters and stimulus paradigm (e.g., breath-hold or fixed hypercapnic challenge with end-tidal CO₂ monitoring). Data will be processed using a standardized, validated CVR quantification pipeline incorporating motion correction, physiological noise regression, and voxelwise modeling of the BOLD signal in response to the vasoactive stimulus.

Control Strategy: Historical Comparison Given the exploratory and interventional nature of this study, a randomized control group is not feasible within this pilot phase. Instead, a historical control cohort will be utilized to contextualize observed changes in BOLD-CVR following lumbar drainage.

The historical cohort includes previously scanned acute ischemic stroke patients with HF2 profiles who underwent BOLD-CVR MRI at two time points (typically 3-7 days apart) but without any interventional procedure between scans. These patients were matched for age, vascular risk factors, infarct location, and NIHSS at baseline.

This allows for a comparison of temporal BOLD-CVR changes occurring naturally in similar patients, thereby enabling preliminary inference about the causal effect of CSF drainage on CVR dynamics.

Outcome Measures

Primary Endpoint:

Change in mean hemispheric and/or MCA-territory BOLD-CVR between pre- and post-drainage imaging.

Secondary Endpoints:

Incidence of adverse events related to lumbar drain (e.g., headache, infection, CSF leak).

Change in NIHSS from baseline to discharge.

Change in mRS at 3-month follow-up.

Incidence of recurrent ischemic stroke within 90 days.

Statistical and Analytical Considerations Given the exploratory nature of this pilot study, no formal power calculation has been performed. However, a target sample size of 10-15 interventional cases has been selected based on feasibility and expected enrollment rates.

Statistical approach:

Pre/post comparisons will be performed using paired statistical tests (e.g., Wilcoxon signed-rank or paired t-test depending on normality) to evaluate changes in BOLD-CVR.

The historical cohort (n ≈ 20-30) will serve as a reference distribution. Between-group comparisons will use mixed-effects models or ANCOVA adjusting for baseline CVR and time between scans.

Exploratory ROC and AUC analyses will be conducted to evaluate whether baseline CVR thresholds predict benefit from CSF drainage.

Adverse event rates and tolerability measures will be reported descriptively.

Missing data will be handled using appropriate methods (e.g., multiple imputation or last observation carried forward), depending on extent and distribution.

Technical Considerations BOLD-CVR acquisition will use echo-planar imaging sequences optimized for T2*-sensitivity, with whole-brain coverage and temporal resolution ≤2s.

Stimulus control will be ensured via synchronized end-tidal gas monitoring or standardized breath-hold timing via visual cueing.

Data analysis will include use of validated toolboxes (e.g., FSL, AFNI, or custom MATLAB pipelines), including individual voxel z-scoring and spatial normalization to MNI space.

Quality control metrics (e.g., framewise displacement, temporal signal-to-noise ratio) will be logged to assess scan integrity.

All scans will be independently reviewed by a neuroradiologist and a neuroimaging researcher.

Data Handling and Ethics All study data will be pseudonymized and stored on secure hospital servers with limited access. BOLD-CVR imaging data, drain parameters, clinical metrics, and safety events will be recorded in a pre-defined electronic case report form (eCRF). Monitoring will be performed according to ICH-GCP principles. This study has been submitted for ethical approval to the responsible Cantonal Ethics Committee (KEK) and notified to Swissmedic as a Category B interventional clinical trial.

Informed consent will be obtained prior to inclusion. For patients lacking decision-making capacity, legal proxy consent will be used, followed by re-consent as soon as feasible. Participation can be withdrawn at any time, and imaging or data will be excluded upon request.

Anticipated Impact This study provides a first step toward evaluating an innovative, low-risk intervention in a high-risk stroke population for whom current treatments are insufficient. If successful, lumbar drainage could become a temporary, targeted approach to improve cerebral perfusion and reduce recurrence risk in patients with exhausted autoregulatory capacity.

The study is expected to yield mechanistic insights and generate effect size estimates to inform a future randomized controlled trial.

Inclusion Criteria:

• Male and female patients > 18 years old
• Acute ischemic stroke related to a new uni- or bilateral occlusion of the ICA, MCA (M1- and or M2 segment) or both
• Persisting vascular pathology after the initial treatment.
• Hemodynamic impairment on the BOLD-CVR done within 48 hours after stroke onset.
• Capable of providing informed consent as documented by signature, or if unable, the next of kin (legally authorized representative) must be available for medical decision-making.

Exclusion Criteria:

• Inclusion in any other running scientific study

• Contraindications for MRI, including but not limited to:

  o Pacemakers, metallic implants/prostheses, metallic tattoo dyes.

• Severe glaucoma.

• Unwillingness or inability to perform breathing maneuvers required for BOLD-CVR assessment.

• Major cardiopulmonary diseases, including:

  • Severe uncontrolled asthma bronchiale.
  • Severe chronic obstructive pulmonary disease (COPD, GOLD Stage >2).
  • Diffuse interstitial lung disease.
  • Recent myocardial infarction (acute/subacute).
  • Severe heart failure (NYHA class >2).

• Symptomatic increased intracranial pressure (ICP) or related conditions:

  • Absent/compressed basal cisterns.
  • Compression or displacement of the fourth ventricle.
  • Cerebellar tonsillar herniation (>5mm below the foramen magnum) on initial CT/MRI or on MRI at Day 1.

• Obstructive hydrocephalus, including:

  • Aqueductal stenosis.
  • Enlarged lateral and third ventricles with a normal fourth ventricle.

• Presence of intracranial hemorrhage class >1 (per Heidelberg classification).

• Unilateral or bilateral subdural hematomas.

• Malignant infarction syndrome (supratentorial or infratentorial).

• Decreased level of consciousness at enrollment, not due to aphasia

• New-onset seizures after stroke onset.

• Extensive lumbar surgery history.

• Local or systemic infection:

• Infection at the lumbar drain insertion site.

• Systemic infection at the time of enrollment.

• Severe coagulopathy, including:

  • Moderate/severe thrombocytopenia (<100,000 platelets/µL).
  • Hemophilia or other coagulation disorders.
  • Use of anticoagulants (except for platelet aggregation inhibitors) that cannot be stopped.
  • Within 24 hours after intravenous thrombolysis

• Structural spinal abnormalities, including:

  • Spina bifida.
  • Scoliosis affecting lumbar puncture feasibility.

• Pregnant (confirmed or verbally suspected) or lactating women.

• Inability to follow the procedures of the study, including but not limited to:

  • Severe cognitive impairment (e.g., dementia).
  • Psychiatric disorders affecting cooperation.
  • Severe language barriers preventing study compliance.

• Failure to insert the lumbar drain.