Targeting Hemoglobin to Reduce Delayed Cerebral Ischemia After Subarachnoid Hemorrhage

Abstract

Delayed cerebral ischemia (DCI) continues to be a sequela of aneurysmal subarachnoid hemorrhage (aSAH) that carries significant morbidity and mortality. Aside from nimodipine, no therapeutic agents are available to reduce the incidence of DCI. Pathophysiologic mechanisms contributing to DCI are poorly understood, but accumulating evidence over the years implicates several factors. Those have included microvessel vasoconstriction, microthrombosis, oxidative tissue damage, cortical spreading depolarization as well as large vessel vasospasm. Common to these processes is red blood cell leakage into the cerebrospinal fluids (CSF) and subsequent lysis which releases hemoglobin, a central instigator in these events. This has led to the hypothesis that early blood removal may improve clinical outcome and reduce DCI. This paper will provide a narrative review of the evidence of hemoglobin as an instigator of DCI. It will also elaborate on available human data that discuss blood clearance and CSF drainage as a treatment of DCI. Finally, we will address a recent novel device that is currently being tested, the Neurapheresis CSF Management System™. This is an automated dual-lumen lumbar drainage system that has an option to filter CSF and return it to the patient.

Introduction

Delayed cerebral ischemia (DCI) after aneurysmal subarachnoid hemorrhage (aSAH) affects around one third of patients and continues to be a poorly understood phenomenon.[1] DCI presents as a clinical syndrome including focal neurological or cognitive deficits, with an onset between 3-14 days after hemorrhage.[2] DCI is associated with severe morbidity and up to three times increased mortality rate during the first two weeks of aSAH.[3] Nevertheless, it continues to have limited treatment options, demanding an urgent need for developing novel and effective therapeutic interventions.[4]

Historically, the underlying etiology of DCI was thought to be large vessel vasospasm leading to ischemia. Starting in the early 1940s, angiographic evidence of cerebral vasospasm started to emerge as a cause of cerebrovascular insufficiency.[5, 6] By late 1940s-mid 1950s, there were numerous reports of cerebral vasospasm contributing to delayed deficits after aSAH.[7, 8] Observations at that time suspected the irritative effect of subarachnoid blood as the etiology of cerebral vasospasm. In one description of remote vasospasm, Pool reports, “vasospasm that so often spreads distally and sometimes proximally from the aneurysmal site could also be caused by irritative, chemical or inflammatory effects of escaped perivascular blood.”[7]

Over time, further research has demonstrated that cerebral vasospasm is certainly one contributor to DCI. However, treating angiographic vasospasm does not appear to reduce DCI consistently, and regions of cerebral ischemia do not always correlate with angiographic spasm. [9, 10]

Current data points toward a multifactorial etiology for DCI including microvessel vasoconstriction, microthrombosis, oxidative tissue damage, cortical spreading depolarization as well as large vessel vasospasm.[1] This cascade of events is triggered by the initial brain injury following aneurysmal rupture, however it is maintained and exacerbated by the blood that is released into the subarachnoid space. After the bleeding stops, RBC hemolysis and release of hemoglobin (Hgb) drive the above phenomena. Free Hgb, unless bound and scavenged by the normal mechanisms of the body, will oxidize into various forms, and become a potent pro-inflammatory and cytotoxic molecule.[11]

Animal studies have shown that Hgb (and its byproducts) can lead to endothelial damage and apoptosis, neuroinflammation, blood-brain barrier disruption and arterial thrombosis.[12] Human studies have also corroborated such roles. Furthermore, human studies have shown a strong relationship between the volume and thickness of subarachnoid blood and the development of vasospasm.[13, 14] Taken together, the current understanding is that the pathology of cerebral vasospasm and DCI stems from the presence of subarachnoid blood and resultant injury to nearby cells (neuronal, glial, and endothelial).

Numerous papers have addressed the pathophysiology of DCI with a focus on blood and blood degradation products.[3, 11, 15–25] Furthermore, evidence, albeit still weak, now exists for removing cerebrospinal fluid (CSF) blood as a preventative treatment against DCI in humans. This paper will review the evidence and will elaborate on the current clinical trials for removal of subarachnoid blood as a treatment for DCI.

Methods

This paper was synthesized as a narrative review of the topic. The search engines of MEDLINE/PubMed. Keywords searched included cerebrospinal fluid filtration, subarachnoid hemorrhage, vasospasm, delayed cerebral ischemia, hemoglobin, and ruptured aneurysm. Identified abstracts were screened by the authors. Further relevant references were identified from the bibliography of extracted articles as needed.

Animal studies and experimental evidence for Blood/Hgb as the primary mediator for DCI

Animal studies have established a prominent role for lysed erythrocytes’ products within the subarachnoid space in the pathogenesis of DCI/vasospasm. Blood is a major instigator of DCI through various mechanisms of actions. Blood and blood-degradation products provoke cerebral vasospasm, inflammation, microthrombosis and blood-brain barrier dysregulation.[26] A comprehensive review of each entity is beyond the scope of this paper, however we will briefly summarize the role of Hgb in each pathophysiological event.

Hgb and Vasospasm

Hgb can contribute to DCI by directly inducing vasospasm. Vasospasm is an outcome of endothelial cell (EC) dysfunction, smooth muscle contraction and changes in vascular autoregulation. Hgb, and particularly oxy-Hgb, has been identified as a key instigator of cerebral vasospasm based on both in-vitro and in-vivo studies.[16, 27] Experimental papers have shown that endothelial nitric oxide (NO) signaling is disrupted after aSAH by unbound Hgb with resulting spasm in arteries and smaller arterioles. NO is a potent vasodilator and following aSAH, NO and its signaling can be disrupted by various mechanisms. OxyHgb reacts with NO thereby depleting the vasodilator within the vascular wall preventing blood vessel relaxation.[28] Other observations report that neurons expressing NO synthase (NOS) are not present after aSAH because of heme and Hgb toxicity.[29] There are also NO-independent mechanisms which can lead to vasospasm.[26, 30, 31] Bilirubin oxidation accumulates toxic molecules that promote vasoconstriction by acting on large potassium channels. Furthermore, bilirubin products have also been shown to inhibit NOS.[19, 32] In support of these findings, NO-based therapies reversed aSAH-associated vasospasm in multiple animal models, including monkeys, rabbits, and rodents.[33] Other reported mechanisms of Hgb-induced vasospasm involve eicosanoids. Eicosanoids actions have been shown to potentiate vasoconstricting nerves and inhibit vasodilator nerves.[16] Finally, free heme has direct effects including increasing endothelin-1 (ET-1) levels,[34] lipid peroxidation of cell membranes, and direct oxidative stress on smooth muscle cells,[35] all of which contribute to vasoconstriction.[36] Experimental models have confirmed the role of ET-1 in vasospasm and ET antagonism reduces vasospasm severity in experimental SAH models.[3, 33] As such, the role of ET-1 has been the target of treatment in several clinical trials. [36–41]

Hgb and Inflammation

A pro-inflammatory state ensues following aSAH with increasing cytokines in both CSF and serum. This state has been linked to worse neurological outcome, neurological decline, and increased vasospasm.[11] Several cytokines have been repeatedly identified in independent studies including IL-1ß, TNF⍺, and IL-6.[42, 43] The Toll-like receptors (TLRs) are fundamental players in innate immunity and inflammatory cascades and TLR-4 has been specifically studied in SAH. Activation of TLR-4 results in activation of a pro-inflammatory cascade and recruitment of immune cells. Recent studies have confirmed that TLR-4 plays a major role in inflammatory reactions, blood-brain barrier disruption, and vasospasm following experimental SAH and identified TLR-4 as a possible therapeutic target.[44, 45] Products of erythrocyte lysis are well known to be strong activators of TLR-4 and a main instigator of TLR-4 signaling.[46] Following TLR-4 binding, cell adhesion molecules are upregulated on ECs which in turn recruits macrophages and neutrophils into the subarachnoid space. While these cells help in scavenging degraded blood, they can remain trapped, degranulate, and inadvertently propagate a cycle of inflammation and secondary damage.[11]

Free heme also mediates the activation of neutrophils causing their oxidative burst and release of superoxide contributing to oxidative damage.[47] Neutrophils have been suggested to be involved in the outcomes following aSAH in humans.[48] In a mouse model of SAH, myeloid cell depletion ameliorated vasospasm after SAH.[49] In a rabbit model of SAH, antibodies against neutrophil adhesion molecules also resulted in decreased vasospasm.[50] More recently, neutrophil depletion after SAH in mice was shown to improve memory through reducing tissue inflammation and cerebral vasoconstriction.[51]

Hgb and Microthrombosis

Erythrocyte lysis and the subsequent free Hgb result in NO scavenging and EC injury (further reducing EC NO).[11] Since platelet activation is strongly inhibited by NO, the strong reduction in NO likely contributes to the pro-coagulable state that follows SAH. The inflammatory cascade is also a potent platelet activator.[52] The presence of cerebrovascular microthrombi after SAH has been documented as early as 1983,[53] and has been confirmed in both human and animal models of SAH.[30, 54] While the exact mechanism of formation of microthrombosis is unknown, it is thought to be an interplay between damaged ECs, activation of platelets or the coagulation cascade, and impairment of the fibrinolytic system. Studies have shown that microthrombi count positively correlates with apoptotic neurons, infarction volume, and delayed deficits.[55] Animal subjects have shown that microthrombi and occluded microvasculature occur throughout the brain between 2 to 7 days after SAH.[56] Recent evidence has emphasized the role of platelets in microthrombosis with some studies suggesting that platelet activation occurs prior to vasospasm after aSAH.[55] Clinical trials targeting hypercoagulability have shown non-significant trends toward better outcomes, and trials utilizing heparin and other agents are ongoing.[57–60]

Removal of hemoglobin or its by-products as a treatment for DCI

Because erythrolysis and Hgb release instigate brain injury after SAH, it follows that mitigating Hgb toxicity will serve as a treatment. Haptoglobin (Hp) is the typical Hgb scavenger in the intravascular space where it irreversibly binds cell-free Hgb, blocking its toxic effects and aiding in its clearance.[61] However, after SAH, the physiologic quantities of Hp in the CSF are not sufficient to bind free Hgb. This has prompted studies to evaluate the use of Hp as a therapy for DCI.[62, 63] While studies using Hp to treat SAH are limited, Hp has been well-studied for treatment in models of intracerebral hemorrhage (ICH). In one study, Hp overexpression was able to reduce the neurological deficits seen after an experimental model of ICH.[64] Other animal studies have supported similar findings.[65] Specifically for SAH, prolonged intrathecal exposure to uncomplexed Hgb was sufficient to model outcomes seen after SAH including behavioral, vascular, cellular, and molecular changes, and Hp administration was safe and reversed the neurological deficits after Hgb injection.[63] It was also shown that Hp administration attenuates iron staining in the parenchyma.[66] In another paper, CSF derived from patients with DCI after SAH was shown to induce arterial vasoconstriction when added to porcine basilar arteries. Addition of Hp to patient-derived CSF restored the vasodilatory response to intrinsic and extrinsic NO.[63]

Additionally, iron chelators, particularly 2,2′-dipyridyl,[16, 17, 67] as well as deferoxamine and lactoferrin have been tested as possible treatments of early brain injury after SAH and ICH.[22, 68–70] Administration of 2,2′-dipyridyl or deferoxamine improved neurological outcomes and reduced large vessel (middle cerebral artery and basilar artery) vasospasm.[17, 67] Other murine studies have also shown neuroprotection after SAH with intraventricular deferoxamine, however the mechanism was independent of attenuating vasospasm.[22, 68, 69]

Historical clinical studies for CSF drainage and diversion

Following SAH, RBCs are released into the subarachnoid space. Over 2-3 days, RBCs break down releasing Hgb most-likely with the downstream effects detailed above. While mechanisms are in place to attempt clearance of such products, the number of RBCs in the subarachnoid space overwhelms this system.[11] The fact that RBCs break down over days provides a therapeutic window during which blood removal has the potential to reduce overall Hgb release. This has led to the hypothesis that early blood removal may improve clinical outcome and reduce DCI. This clinical goal has been attempted by employing one of several modalities, either alone or in combination, including EVD, lumbar drain (LD), and cisternal flushing with or without the use of anticlotting agents.

Direct Surgical Removal:

If clipping of the aneurysm or other open surgical intervention is performed, blood removal and improved CSF flow can be achieved intraoperatively by irrigating the blood from the basal cisterns. However, studies addressing the efficacy of this measure for decreasing shunt dependency have produced inconclusive results.[71] Clinical reports of clot removal during open surgery have been reported since early 1980s. In one of the earlier studies including 64 patients, the authors concluded that it “is possible to prevent intracranial arterial spasm and associated neurological deterioration by early operation and removal of clotted blood from the subarachnoid space”.[72] A similar conclusion was found in another study involving 45 cases.[73] However, following the advent of endovascular techniques, studies emerged comparing various outcomes including vasospasm. In one large study involving 415 treated patients (339 by clipping and 76 by coiling), the incidence of vasospasm was reduced in low grade patients treated with coiling and not different in higher grades.[74] Furthermore, other studies have failed to provide evidence of decreased vasospasm at the site of clot removal compared to the contralateral side.[75, 76] In a more recent study which quantified blood volume in the subarachnoid space, the authors concluded that while clot clearance during the operation is more rapid, the percent reduction of the clot after days 3 through 5 is similar, primarily due to the presence of external CSF drainage. Furthermore, the incidence of symptomatic vasospasm did not differ significantly between these two groups.[77]

From a practical standpoint, surgically exploring the sylvian fissure and basal cisterns may carry its own risks including additional intraoperative bleeding, venous injury, and iatrogenic spasm by manipulation of cerebral vessels. Moreover, many areas of the brain will not be accessible for irrigation (i.e contralateral sylvian fissure, posterior fossa).

External Ventricular Drain:

CSF drainage via an EVD has a long-standing history, but the literature suggests that it can lead to a higher incidence of post hemorrhagic shunt dependency.[78–80] This is presumed to be because blood settles and clots in the basal cisterns as the EVD drains the less dense CSF present in the lateral ventricles.[78–80] Because cisternal CSF circulation is reduced due to partial reversal of CSF flow, the ability of EVDs to effectively decrease SAH blood remains unclear. A large study involving CSF draining via both LD and EVD reported that drainage via LD was more successful in removing blood.[79] Another study that compared EVD to LD found that clearance of subarachnoid clots was more rapid in patients treated with LD (n = 34) compared to those treated with EVD (n = 17).[81] Unfortunately, there are no prospective studies to date that evaluate whether EVD placement reduces DCI.[82]

Lumbar Drain:

Application of a LD has been proposed as an alternative approach to address clotting of blood in the basal cisterns and to mobilize and reduce total blood burden in the CSF space (Table 1). Draining the lumbar cistern maintains a more normal CSF circulation and can immediately clear the blood that has settled.[77, 79] The safety of this approach was shown in two retrospective studies of patients with aSAH.[79, 83] Adoption of the stringent exclusion criteria for the use of lumbar drains in aSAH developed by Klimo and Schmidt[79] allows the avoidance of the well-recognized but generally uncommon complications of LDs.[84] Such complications included risk of over-drainage with associated formation of subdural hemorrhages and downward herniation. As an additional safety measure, the authors suggest the conversion of patients with an EVD to a LD after securement of the aneurysm, after the patients’ neurological exam has had time to improve, and generally before the period of greatest risk of vasospasm. It has subsequently been suggested that poor grade patients may also benefit from the use of LD in combination with EVD.[85]

Table 1.

Studies of lumbar drainage and outcomes in aneurysmal subarachnoid hemorrhage.

Study Name or Author/year	Number of Subjects	Study Type/groups	Key Finding/Interpretation
Klimo et al, 2004[79]	167	Retrospective	Lumbar drainage after SAH markedly reduced the risk of clinically evident vasospasm and its sequelae, shortened hospital stay, and improved outcome.
Kwon et al, 2008[83]	107	RCT	Lumbar CSF drainage showed reduction of clinical vasospasm following endovascular coiling on aneurysmal subarachnoid hemorrhage.
Hanggi et al, 2008[89]	40	RCT	Combination of lumboventricular lavage and mechanical head motion reduced vasospasm on TCD ultrasonography, the incidence of delayed ischemic neurological deficits, and secondary infarctions on computed tomography and improved clinical outcome. No obvious effect could be found on the rate of angiographic vasospasm.
LUMAS, 2012[86]	210	RCT	Reduced rates of delayed ischemic neurological deficit and improved early outcomes with lumbar drainage compared to standard care alone
Park et al, 2015[90]	234	RCT	LD after aneurysmal SAH shows marked reduction of clinical vasospasm and need for angioplasty, with this favorable GOS score at 6 month follow-up.
Borkar et al, 2018[87]	60	RCT	Lumbar CSF drainage reduced clinical and radiographic vasospasm and showed improved outcomes and a trend toward a shorter hospital stay.
Fang et al, 2020[88]	193	Retrospective	In the higher modified Fisher group, patients who received LD had significantly lower risk of cerebral vasospasm, delayed cerebral infarction, and hydrocephalus; the GOS score was significantly higher in the LD group at discharge and at 3 months of follow-up.

Abbreviations: RCT: randomized controlled trial; SAH: subarachnoid hemorrhage; TCD: transcranial doppler; LD: Lumbar drainage; GOS: Glasgow outcome scale

Lumbar drainage studies have shown a reduced incidence of angiographic vasospasm and/or improvement in clinical outcomes.[79, 83, 86–90] One of the larger studies included 81 patients who had LDs placed for CSF diversion versus a group of 86 patients who either received no form of CSF drainage or were treated solely with an EVD.[79] Lumbar CSF drainage reduced the incidence of clinical vasospasm from 51% to 17%, the need for endovascular vasospasm treatment from 45% to 17%, and the risk of cerebral infarction from 27% to 7%.[79] A systematic review and meta-analysis showed that LD is a promising measure for the prevention of delayed aSAH-related ischemic complications; in this analysis, LD was associated with a statistically significant decrease in the risk of DCI-related complications (cerebral infarctions and clinical deterioration), as well as the risk of severe disability.[91]

In the largest and only randomized trial (LUMAS, Lumbar Drain vs. No CSF diversion, N=210), al-Tamimi et al. demonstrated that the use of LD significantly reduced the prevalence of delayed ischemic neurological deficit (35.2% vs. 21.0%) and improved early clinical outcome.[86] There was no significant difference in clinical outcome at 6 months or in the need for permanent CSF shunting, although there was a nonsignificant decrease in radiologically confirmed infarct. However, it is well recognized that clinical trials powered to detect changes in DCI occurrence often report difficulties in detecting patient-centered “downstream” clinical outcome.[92] This study supports the use of LD in patients with Fisher Grade 3 aSAH in which it may confer benefit by reducing CSF blood volume. Table 1 summarizes some of the main studies reporting the use of LD after aSAH and outcomes.

Intrathecal thrombolytics

Thrombolytics as an adjuvant in the removal of blood products from aSAH patients has also been an active area of study. Intraventricular fibrinolysis (IVF) with low-dose recombinant tissue plasminogen activator (tPA), and less commonly urokinase, has been employed using both EVD or intra-cisternal access with positive results on DCI and mortality.[93]

One randomized study giving intrathecaltPA bolus 5 mg every 12 h combined with a lateral rotational therapy (RotoRest(®)) showed that it is safe, improves clot clearance, and decreases incidence of DCI.[94] Another study employing intra-ventricular tPA noted no significant improvement in the long-term functional recovery following aSAH with tPA, but found that a high dosage of tPA might reduce the incidence of angiographic vasospasm[95]. Delivery techniques vary and include EVDs, basal cistern tubes, lumbar cisterns, cisterna magna tubes, and even chiasmatic, prepontine, and sylvian cisterns among others. Reported dosing of tPA given through an EVD ranges from 1 to 5mg per dose for a total of 5 to 12 doses.[93]

Meta-analyses have been supportive of intrathecal fibrinolytics. In one meta-analysis of eight RCTs, the use of fibrinolytic agents was significantly associated with lower rates of vasospasm and hydrocephalus after SAH.[96] In another large meta-analysis, intrathecal administration of thrombolytic agents to break the subarachnoid clot revealed a clinically relevant and statistically significant beneficial effect on DCI and mortality.[97] Hemorrhagic complications compared to controls were not found to be increased in a pooled analysis of randomized controlled studies.[93] While recent data strongly support intrathecal fibrinolytics, current guidelines have not clearly advocated their use.[4] This could be as the guidelines were written before much of the data cited above were published. The lack of endorsement of such intervention by governing committees, coupled with the risks of cisternal access and injections along with the perceived risk of fibrinolytics, limit the popularity of this intervention.

Ongoing Clinical trials for Blood Clearance

Several ongoing clinical trials for blood clearance are underway. Evaluating fibrinolysis, FIVHeMA (Intraventricular Fibrinolysis for Aneurysmal Subarachnoid Hemorrhage, NCT03187405) is a phase III trial in France enrolling 440 patients for randomization and administration of alteplase through an indwelling EVD .[98] The primary endpoint is mRS at 6 months. The SPLASH trial is an ongoing phase II trial in Germany that will randomize patients to either standard of care only (EVD) or standard of care plus cisternal lavage. Cisternal lavage with urokinase, nimodipine, and saline will be performed through a catheter placed stereotactically into the basal cisterns. The primary endpoint is mRS at 6 months.[99]

Two studies are assessing lumbar drainage as a mechanism to improve outcome. Lumbar Drain vs Extraventricular Drain to Prevent Vasospasm in Subarachnoid Hemorrhage (NCT03065231), is a prospective randomized study at UC San Diego that will compare LD vs EVD. Enrollment will be 100 patients per group and the measured outcomes will be clinical vasospasm, decreasing subarachnoid blood on CT, overall ICU stay, and reducing the need for a permanent ventriculoperitoneal shunt. EARLYDRAIN is a German randomized trial where aSAH patients are assigned to early continuous lumbar drainage (+/− EVD) vs. standard of care. The primary endpoint is disability at 6 months, and secondary endpoints include vasospasm.[100] However, while the study design was published in 2011, there has been no further publications, with the last update being in 2017 on clinicaltrials.gov.

Neurapheresis CSF Management System Setup and PILLAR Trial

While there are promising results of the use of lumbar drainage to improve outcomes in aSAH, bloody CSF removal from the lumbar spine is limited by the volume of production of CSF, i.e the rate of which one can draw out CSF should be well below the rate of CSF production. One novel device currently being tested is an automated dual-lumen lumbar drainage system that has an option to filter RBCs from the CSF and return “clean” CSF to the patient (Neurapheresis CSF Management System™, Minnetronix, St. Paul MN).[101, 102] Due to the fact that clean CSF is returned, the net negative volume loss is reduced and filtration rates can be increased above physiologic CSF production rates. As a result, this system has the potential to more promptly reduce blood burden following aSAH prior to hemolysis compared to EVD or LD.[103]

The Neurapheresis CSF Management System is composed of a dual lumen catheter that is of similar size as a typical lumbar drain. It is placed under fluoroscopic guidance into the intrathecal lumbar space, and advanced into the upper thoracic spine (Figure 1). Using a pressure-controlled pump system, bloody CSF is removed via catheter fenestrations in the lumbar spine and filtered. The filtered CSF is then returned to the patient via fenestrations located at the distal catheter tip and the retentate is routed to a waste bag.

Figure 1.

Schematic representation of the Neurapheresis CSF Management System showing dual lumen catheter with bloody CSF removed via catheter fenestrations in the lumbar spine and “clean” CSF returned to the patient via fenestrations located at the distal catheter tip.

In a recent prospective clinical trial - Prospective Trial of Cerebrospinal Fluid Filtration After Aneurysmal Subarachnoid Hemorrhage via Lumbar Catheter (PILLAR) - this device was placed and tested in 13 patients. CSF filtration was performed for up to 24 hours (mean 15:07 hours, range 5:32 – 24:00 hours). Although the device returns the majority of CSF to the patient, there is a drainage effect due to removal of RBCs as well as accompanying CSF. A mean of 632.0 mL of CSF was filtered, with 439.0 mL of CSF returned to the patient and mean 193.0 mL of concentrated waste removed. From start to finish of the treatment with the Neurapheresis System this equated to an average effective waste removal rate of 7.24 mL/hour (range 3.88 to 11.34 mL/hour). In 92.3% of patients (12/13), the neurological (GCS, WFNS, Hunt & Hess) and peripheral checks remained stable or improved during treatment with the Neurapheresis System. In one patient, there was a decline in GCS score secondary to the development of focal cerebral edema determined to be related to the clipping of an MCA aneurysm (not an adverse event caused by the device per the Data and Safety Monitoring Board). CSF samples before, during, and after filtration confirmed significant reduction in RBCs and protein.[101] (Figure 2)

Figure 2.

CSF samples before and after filtration confirming significant reduction in RBCs and protein.

The results of this study indicate significant blood removal. However, several practical limitations remain unknown including whether filtration of lumbar/thoracic CSF reduces blood burden in the supratentorial space and would ultimately confer a clinical benefit. Early data suggests that there is reduction in cisternal blood using the Neurepharesis System, but how this compares to natural history of absorption is still undergoing analysis. (Figure 3)

Figure 3.

Significant Reduction in Hijdra cisternal blood score before and after filtration using the Neurepharesis CSF Management

Following the PILLAR trial completion in 2018, the device was further evaluated under a second clinical trial, the PILLAR-XT study (NCT03607825), with increased filtration times (up to 72 hours). In PILLAR-XT, 33 patients were enrolled, with 27 being treated. The safety data and filtration results are currently being evaluated.

Complexities for Clinical Trials Evaluating Early Blood Removal

Despite at least three decades of research investigating CSF diversion and blood removal, there is no definitive recommendation on the preferred method or whether fibrinolytics are clearly beneficial.[4] Scientific evidence from both animal models as well as human studies remains supportive that early clearance of blood will reduce DCI and improve overall outcomes if it can be carried out in a safe and timely fashion. Nevertheless, there is a relatively low population incidence of aneurysmal SAH, and variability of comorbidities and presentation severity complicate trial enrollment and analysis.[104] Perhaps the largest barrier is that only 20-30% of SAH patients experience DCI[105] and therefore clinical trials assessing therapeutic efficacy are diluted by patients who never have DCI necessitating large patient samples to establish adequate statistical power.[104] A biomarker predicting patients at risk for developing DCI would facilitate a focused cohort of SAH patients (by improving treatment effect), however such a biomarker remains elusive and has hindered SAH research.[104] Nevertheless, trials on blood removal remain exciting as technology for CSF filtration advances and evidence for the safety of fibrinolytics accrues.

Conclusion

Although the factors contributing to poor clinical outcome after aSAH are multifactorial, RBC hemolysis in the CSF is the primary instigator of the pathophysiologic processes contributing to delayed complications. Since Hgb toxicity triggers numerous injury pathways, therapeutic treatment targeting one (e.g. inflammation, hypercoagulability, vasospasm) may have limited clinical impact. In contrast, therapeutic treatment to reduce RBC volume prior to hemolysis has the potential to prevent post-aSAH pathology across multiple mechanistic pathways. As a result, rapid and early removal of RBCs is an attractive concept to prevent the downstream effects of free Hgb and warrants continued clinical investigation.

Data availability statement:

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.