Continuous irrigation with thrombolytics for intraventricular hemorrhage: Case-Control Study

Abstract

Background:

Intraventricular hemorrhage (IVH) is a complication of a spontaneous intracerebral hemorrhage. Standard treatment is with external ventricular drain (EVD). Intraventricular thrombolysis may improve mortality but does not improve functional outcomes. We present our initial experience with a novel irrigating EVD (IRRAflow) that automates continuous irrigation with thrombolysis.

Methods:

Single-center case-control study including patients with IVH treated with EVD compared to IRRAflow. We compared standard demographics, treatment, and outcome parameters between groups. We developed a brain phantom injected with a human clot and assessed clot clearance using EVD/IRRAflow approaches with CT imaging.

Results:

21 patients were treated with standard EVD and 9 patients with IRRAflow. Demographics were similar between groups. 33% of patients with EVD also had at least one dose of t-PA and 89% of patients with IRRAflow received irrigation with t-PA (p=0.01). Mean drain days were 8.8 for EVD versus 4.1 for IRRAflow (p=0.02). Days-to-clearance of ventricular outflow was 5.8 for EVD versus 2.5 for IRRAflow (p=0.02). Overall clearance was not different. 37% of EVD patients achieved good outcome (mRS≥3) at 90 days versus 86% of IRRAflow patients (p=0.03). Assessing only t-PA, reduction in mean days to clearance (p=0.0004) and ICU days (p=0.04) was observed. In the benchtop model, the clot treated with IRRAflow and t-PA showed a significant reduction of volume compared to control.

Conclusion:

Irrigation with IRRAflow and t-PA is feasible and safe for patients with IVH. Improving clot clearance with IRRAflow may result in improved clinical outcomes and should be incorporated into randomized trials.

INTRODUCTION

Intraventricular hemorrhage (IVH) as an extension of a spontaneous intracerebral hemorrhage (sICH) is a frequent complication[1] that aggravates the outcomes of this type of stroke.[1–7] One of the major complications of IVH is the mechanical obstruction of the 3^rd and 4^th ventricles leading to obstructive hydrocephalus as well as secondary brain injury from blood breakdown products.[8, 9] The mainstay of treatment for obstructive hydrocephalus involves the placement of an external ventricular drain (EVD) for intracranial pressure (ICP) monitoring and cerebrospinal fluid (CSF) diversion as necessary, in preference to medical management alone.[10] Such an approach is frequently complicated by drain clotting with some series reporting as high as 60% obstruction rates[11], need for drain replacement[12] or bilateral EVD placement for casted ventricles.[13] Mortality is greater than 50% with only 20–40% of survivors having good functional outcomes.[13–15] The need for long-term interventions such as ventriculoperitoneal shunting (VPS) is around 20%.[13]

Intraventricular thrombolysis through the EVD using intermittent boluses of recombinant tissue plasminogen activator (t-PA) has been tested compared to saline control injections and has been shown to be safe with a mortality benefit.[10] Despite this benefit, a definite impact on functional outcome has not been established.[13] Nonetheless, these data suggest that there may be an association between enhanced IVH clearance and better functional outcomes.[13, 16, 17] From a practical perspective, intraventricular thrombolysis is widely used to maintain catheter patency.[18]

Recent approaches to improve care for patients with IVH have focused on the potential benefit of enhancing clot volume reduction, which has been associated with better outcomes.[13, 16, 17] In addition, better delivery methods that limit the need for continuous breaking of the sterile circuit may improve the safety and efficacy of intraventricular thrombolysis.[19, 20] The recent development of an automatic irrigating catheter (IRRAflow) that uses continuous ICP measurements as a feedback mechanism offers the potential to deliver continuous boluses of irrigation alone or irrigation with t-PA.[19] This cycling feedback process of ICP measurement, irrigation and drainage is intended to prevent overirrigation using a bedside control unit and has been shown to be safe in IVH associated to sICH[19] and subarachnoid hemorrhage.[20]

We present a case-control study comparing patients who underwent placement of a standard EVD with and without intraventricular t-PA to those who underwent placement of the IRRAflow catheter with continuous irrigation with and without t-PA, assessing relevant safety and efficacy outcomes. In addition, we show a benchtop human clot model embedded in a brain phantom made of Agar comparing clot clearance rates using EVD with intermittent boluses of t-PA versus constant irrigation with t-PA using the IRRAflow device. We hypothesize that the use of IRRAflow with continuous irrigation with t-PA is feasible and safe providing a predictable method of severe IVH removal with a high burden of clot reduction when compared to the standard of care, which can be used in larger studies to assess functional outcomes.

METHODS

Study design, procedure, and participants

We conducted a single-center case-control study with patients who were diagnosed with spontaneous intracerebral or cerebellar hemorrhage and intraventricular extension causing hydrocephalus requiring CSF diversion who were treated with the IRRAflow device or standard EVD from January 2019 through September 2022. Patients who had a diagnosis of cerebral amyloid angiopathy, intracerebral tumor or trauma were excluded, as well as those who required surgical decompression. Lobar hemorrhages were excluded to focus only on deep sICH with IVH.

EVD or IRRAflow was placed per clinical standard of care based on the preference of the neurosurgery attending. Briefly, after a standard twist drill hole was created, the IRRAflow catheter was reverse-tunneled using a plastic sheath. In cases where stereotactic navigation was used, a Peel-Away Sheath Introducer was placed using navigation and the IRRAflow catheter was placed through the peel-away sheath, identical to the standard MISTIE approach for ICH.[21]

Patients who were treated with the IRRAflow device received intraventricular irrigation with alteplase (Activase, Genentech Inc – USA, NDC 50242–085-25) or saline alone. Briefly, the goal was to set the infusion rate on the IRRAflow device to deliver a dose of alteplase equivalent to what has been described in earlier CLEAR trials.[15] Patients treated with EVD received intermittent boluses of alteplase based on the clinical judgement of the treating physician, in doses of 3 mg every 12 hours or 2 mg every 8 hours, delivering a total of 6 mg per day. For the IRRAflow group, using a pharmacy approved protocol, we admixed alteplase 4mg in 500 mL of preservative-free normal saline and irrigated at a rate of 30 mL/h as we have previously described[19], requiring a total of 720 mL per day, equivalent to 5.76 mg of alteplase per day. In practice, given the cycling nature of the active fluid exchange mechanism of the IRRAflow device (measure ICPs, drain, irrigate), the volume delivered was less, affecting the dose of alteplase that was effectively administered. Informed consent was obtained prior to either procedure. IRRAflow is an FDA-approved device, therefore not considered an investigational device.

We retrospectively collected IRRAflow cases and patients who required CSF diversion with standard-of-care EVD during the same period under an IRB approved protocol (UNM HRPO# 21–051). We recorded standard demographics and characteristics of clinical presentation. We documented outcome parameters of the interventions, including modified Graeb score (mGS) before and after removal of drain (calculated as described previously[22]), number of days with drain, successful clearance of 3rd and 4th ventricles (defined as computed tomography (CT) evidence of interval absence of hyperdensity in the 3rd and 4th ventricles), number of days from the placement of the drain to successful clearance of the 3rd and 4th ventricles, need for ventriculo-peritoneal shunt (VPS) placement as a long-term treatment for CSF diversion, need for EVD replacement, diagnosis of ventriculitis requiring antibiotic treatment, presence of Grade 3 catheter tract hemorrhage as described previously,[23] days of stay in the intensive care unit, modified Rankin Score (mRS) upon discharge and at 90 days after discharge. Continuous data were compared with unpaired t-tests and categorical data were compared using Fisher’s exact test or chi-square test. Significance was set at p=0.05. Logistic regression was performed for the primary outcomes of good outcome (mRS 0–3) and mortality at discharge and around 90 days. All calculations were done using PRISM 9.5.1 (GraphPad Software Inc - MA, USA). No correction for multiple comparisons was used given the exploratory nature of the analysis and <20 variables assessed. All data are presented as mean or percentage, unless otherwise specified. This study was made following the recommendations of the STROBE statement.[24]

Bench study

We developed a bench model to test a human clot under different thrombolytic conditions. First, we made a brain phantom using agar gel as previously described.[25, 26] Briefly, we mixed agar (Sigma – USA, Cat No. A7921) with 1x TBE buffer (Promega - USA. Cat No. V4251) at a concentration of 0.6% w/v and allowed it to boil for 5 minutes while mixing. When the mixture reached 50°C, we poured it in a 1 L polymethyl methacrylate container and let it cool on an ice block. Once the center of the phantom reached 25°C, we infused 50–60 mL of the previously made clot mixture, using a 50 mL syringe attached to a 14G IV catheter. To generate the clot, we salvaged an expired unit of whole-blood and one of plasma from the blood bank and centrifuged the whole-blood at 1200 rpm at 20°C for 20 minutes and then pipetted out the plasma to combine it with the platelets. We added thrombin (Pfizer – USA, NDC 60793–315-01) and mixed for 2 minutes and subsequently added CaCl 5% (Sigma – USA, Cat No. C1016) to the mixture of red blood cells (RBCs) and buffy coat, using a ratio of RBCs + buffy coat to plasma + platelets of 1:10, as described previously.[27] The embedded clots were allowed to rest for 6 hours and were imaged using CT every 24 hours for 3 days. CTs were performed with a standard protocol on the same Siemens Biograph 64 PET/CT (CT only used) using energy of 100 peak kilovoltage (kVp) and tube current of 275 milliampere-seconds (mAs). Standard head CT reformats were made including 1 mm thick reformats used for measurements. Each clot was assigned to an intervention group: control (no intervention), EVD + t-PA, IRRAflow + saline and IRRAflow + t-PA. The EVD + t-PA group received intermittent boluses of t-PA every 8 hours, per CLEAR protocol.^[13] The IRRAflow + saline and IRRAflow + t-PA groups received constant irrigation at a rate of 30 mL/h with either saline alone or alteplase at the concentration described previously.[20] We calculated clot volumes using the ABC/2 formula through the local PACS platform (IntelliSpace Radiology Enterprise 4.5, Philips – Netherlands). Data were analyzed with Dunnett’s multiple comparisons test with a type I error rate of 0.05, using PRISM 6.01.

RESULTS

Thirty subjects were identified who met inclusion criteria: 21 were treated with standard EVD and 9 were treated with the IRRAflow device. Demographic information and baseline severity factors are summarized in Table 1 and were similar between groups (p>0.05). The mean Glasgow Coma Scale (GCS) was 10 in the EVD group and 12 in the IRRAflow group (p=0.21). Rates of ICH score of 3 or 4 were similar (38% in the EVD group and 33% in the IRRAflow group [p>0.99]). Graeb scores were also similar (p=0.21).

Table 1.

Demographics and outcomes

	Standard EVD (n=21)	IRRAflow (n=9)	p value
Demographics:
Age (mean)	59	55	0.44
Sex (%Female)	33%	44%	0.43
Baseline Graeb score (mean)	13	16	0.24
Baseline ICH score (% 3–4)	38%	33%	>0.99
Admission GCS score (mean)	10	12	0.21
Treatment Variables:
tPA used (%)	33%	89%	0.01
Graeb score after removal (mean)	2	3	0.61
Drain days (mean)	8.8	4.1	0.02
Successful clearance of 3rd and 4th ventricle (%)	52%	89%	0.10
Days to clearance (mean)	5.8	2.5	0.02
VPS (%)	15%	0%	0.5
Need for EVD replacement (%)	39%	0%	0.04
Antibiotic treated ventriculitis (%)	15%	0%	0.5
Track hemorrhage >5cc (%)	24%	22%	>0.99
ICU days (mean)	13.4	10.1	0.23
Good outcome (mRS ≥3) at DC (%)	19%	33%	0.40
Good outcome (mRS ≥3) at 90 days (%)	37%	86%	0.03

Variation in treatment approaches between groups was significant, with a lower threshold to use t-PA in the IRRAflow group (89% in the IRRAflow group versus 33% in the EVD group [p=0.01]). In addition, nearly all patients with IRRAflow who had a supratentorial hemorrhage had the drain placed ipsilateral to the clot (83%) while nearly all the EVD patients with supratentorial hemorrhage had the EVD placed contralateral to the clot (84%).

Outcomes

The use of IRRAflow was associated with improvements in multiple outcome variables (Table 1). The number of drain days was significantly less (4.1 versus 8.8 [p=0.02]). Days to clearance of the ventricular outflow was also less (2.5 versus 5.8 [p=0.02]). However, the rate of overall clearance of the 3rd and 4th ventricles was not statistically different. No EVD replacement or VPS were required in the IRRAflow group, however two patients with EVD required replacement and two required VPS. Functional outcomes were non-significantly better in the IRRAflow group at discharge (OR = 2.13, 95% CI [0.34–12.71]), however improved and became significantly better at 90 days, likely due to ongoing clinical improvement in survivors (OR = 10.29, 95% CI [1.372–216.4]) (Figure 1.)

Figure 1:

Functional outcomes at discharge and 90-days after discharge by treatment group. Stacked bar plots illustrate percentages based on modified Rankin Score (mRS), ranging from mRS 0 (no disability, not shown) to mRS 6 (deceased); Lost in follow up (LFU) patients who did not return for follow up were not included in the calculations for comparison, as shown by the total number of patients at the bottom of each bar (n). Good outcome was defined as mRS ≤ 3 and is highlighted with dotted lines to depict the difference between groups. For all the study population (left panel), the percentages included all patients who received intraventricular t-PA or drainage alone. The functional outcome was better in the IRRAflow group at 90-days after discharge (p=0.02). For the t-PA only subgroup (right panel), the analysis only included the patients who were treated with t-PA. The functional outcome was better in the IRRAflow group at 90-days after discharge (p=0.01)

In a subgroup analysis, selecting only patients who received treatment with t-PA, the trends were similar (Figure 2). The IRRAflow group showed fewer days to clearance of ventricular outflow (1.9 versus 7.4 [p=0.0004]), less drain days (3.7 versus 10.6 [p=0.0098] and less ICU days (9.9 versus 23.2 [p=0.04]).

Figure 2:

Subgroup analysis comparing outcomes in patients who received intraventricular thrombolysis with t-PA. Barplots showing variables measured in patients who were treated with EVD (blue) or IRRAflow (orange). IRRAflow group showed better outcomes for the following variables: a) Days-to-clearance (p=0.0004), b) Drain days (p=0.0098), c) ICU days (p=0.04)

Figure 3 demonstrates an illustrative case of this approach. In this subject, there was a small right caudate hemorrhage with casted right lateral ventricle. The IRRAflow catheter was placed using the procedure described in the Methods section. A peel away sheath was placed directly into the casted ventricle using stereotactic guidance. The IRRAflow catheter was placed down the peel away sheath. As expected, there was no initial CSF and minimal bloody drainage. During the first 24 hours after the placement of the IRRAflow catheter, irrigation with saline alone did not achieve any observable clearance of the IVH. After starting irrigation with t-PA, the ventricular outflow was restored within 24 hours. These results are consistent with the observations from our benchtop model.

Figure 3:

CT images obtained at different timepoints before and after starting irrigation with t-PA in a patient treated with IRRAflow. Coronal (up) and Axial (down) plane images are shown for each timepoint. Baseline images were obtained upon admission. 24h irrigation only was obtained after 24 hours with irrigation with saline. 24h irrigation + t-PA was obtained after 24 hours of irrigation with t-PA, which continued until the 3^rd ventricle was cleared, per our protocol. Drain was removed on hospital day (HD) #5 without any significant complications

In vitro clot reduction

We performed clot simulation in 4 simultaneous conditions as described above. IRRAflow with t-PA was the most effective strategy for clot reduction, reducing to a final volume of 10.6 mL (12% of initial clot volume). EVD with intermittent t-PA bolusing performed similarly to IRRAflow with saline irrigation. The control sample without drainage remained relatively stable over the 72 hours of treatment (Figure 4). Qualitatively, there was an area of hypodense “halo” observed within the clot surrounding the tip of the catheter in the IRRAflow with t-PA group that was not present in the other comparators (Figure 5), accounting for the enhanced clot clearance in this group.

Figure 4:

Line plot depicting clot volume reduction over time in the pre-clinical clot model. Four groups are shown: Control (grey), EVD + tPA (blue), IRRAflow + saline (orange) and IRRAflow + tPA (red). Volume (Y axis) is expressed as percentage (%) vs Time (X axis) expressed in hours (hr). Significance (p < 0.05) was calculated comparing the intervention groups (EVD + tPA, IRRAflow + saline, IRRAflow + tPA) with the control group using Dunnett’s multiple comparison test. A significant reduction in clot volume was achieved in the IRRAflow + tPA group, resulting in a greater than 50% clot reduction within the first 24 hours of intervention

Figure 5:

CT images of brain phantoms with embedded human clots that underwent different thrombolytic interventions after 24 hours of treatment. The picture on the left shows EVD + tPA, middle picture shows IRRAflow + saline and left picture shows IRRAflow + tPA. In dotted red line, highlighting the “fibrinolytic halo” only present in the IRRAflow + tPA group, showing enhanced clot clearance

DISCUSSION

IVH is a severe complication of ICH that is associated with worse outcomes.[28] This may present some of the strongest opportunities for neurosurgical therapies to improve outcomes. In most cases with hydrocephalus in patients who are deemed salvageable, an EVD is necessary.[10] CSF diversion is often limited due to immediate drain obstruction due to drain placement in a casted ventricle. Therefore, many advocate for placing the EVD in the contralateral ventricle, which was the standard practice in most of our EVD cases. The limitation of a contralateral drain placement is the possibility of worsening the midline shift and subfalcine herniation, and it may not effectively clear blood from the casted ventricle.[29] For that reason, some authors advocate for bilateral EVD placement in such cases.[30]

We developed a different solution to this problem like the clot-targeted approach used in MISTIE.[21] Using stereotactic guidance, the catheter is placed directly into the thickest part of the IVH. By using continuous irrigation, we did not experience any cases of catheter obstruction in this group. t-PA was typically started only after stability CT showing adequate position of the catheter in the clot. Using this approach, we achieved clearance of the ventricular outflow in less time than the standard EVD group, suggesting that this technique may allow for faster treatment of hydrocephalus and significantly enhanced clot clearance as shown in our bench study. The subgroup analysis of outcomes shows improvement in functional outcomes at 90-days with IRRAflow (Figure 1) suggesting a possible association between significantly shorter intervention times (Figure 2) and better functional outcomes. The rationale for this association may be related to the known mechanisms by which prolonged exposure to breakdown blood products in an IVH lead to secondary brain injury, impairing CSF dynamics and even causing hydrocephalus.[9] Although the difference in overall clearance of ventricles was not significant between groups at the end of the treatment, reducing the time-to-clearance may be beneficial and relevant to reduce complications related to prolonged drain times.[31, 32] In CLEAR III, nearly one-third of patients required ICP lowering therapies and similar numbers required bilateral EVD, suggesting the need for earlier ventricular decompression. Our approach efficiently overcomes most of the reported limitations of the contralateral EVD placement in IVH.

Despite only including the drain-associated complications in Table 1, we did not include other medical complications that arise in the intensive care unit (ICU) such as nosocomial infections and other complications inherent to our study population[33], that could account for a not statistically significant difference in overall ICU stay.

Furthermore, the use of EVD is associated with multiple complications that can also contribute to worse outcomes. In our small series, none of the patients treated with the IRRAflow required EVD replacement or permanent VPS, consistent with the published literature[34]; whereas the complication rate in the control arm was closer to 18%, as reported in CLEAR III. Interestingly, the track hemorrhage rate was similar. However, in the IRRAflow group, these were not clinically significant and resolved with irrigation, while many of these led to additional drain replacements in the EVD group. Multiple studies have demonstrated the relatively high incidence of complications with EVD placement.[23, 35–37] Therefore, improving the safety profile of the treatment itself by reducing complications that may require drain replacement is of preeminent need. Figure 4 summarizes the enhanced intraventricular clot clearance after irrigation with t-PA as early as 24 hours after starting the intervention, which clinically translates to improvement of hydrocephalus.

Our clot model data is consistent with our experience. The use of t-PA appears to be the fastest for ventricular clearance, however irrigation alone may have more efficacy than drainage alone. The rationale for this finding lies in enhancing clot clearance by combining chemical and mechanical clot erosion (irrigation with tPA) versus only mechanical erosion (irrigation with saline), in comparison to drainage alone.

Currently, other treatment alternatives include neuroendoscopic approaches with varied outcomes that mainly rely on surgical expertise.[38–44] Endoscopic removal of IVH may result in a more invasive procedure compared to constant irrigation with t-PA using an IRRAflow device. Even accounting for the learning curve involved in the reverse-tunnelling technique required for the IRRAflow catheter placement, it may be more accessible to the average surgeon when compared to other approaches.

LIMITATIONS

There are notable limitations to this preliminary study. Our sample size is underpowered, and the treatment approaches differed between groups. Though patients were generally treated based on physician preference, there was crossover between techniques or other unmeasured factors that could introduce bias. The use of t-PA was more frequent in the IRRAflow group, possibly due to an overall more aggressive approach. Prospective, randomized data are needed to better define the benefit of continuous irrigation with t-PA compared to standard EVD drainage for IVH.

CONCLUSION

The use of IRRAflow is feasible and safe in the setting of IVH. Moreover, it appears to be associated with fewer complications than standard EVD. IVH clearance overall is not significantly different, but the time to clearance is, and there may be efficacy in terms of improved clinical outcomes, favoring continuous irrigation with t-PA over intermittent boluses. A bench model confirms superior clot clearance with t-PA continuous irrigation when compared to controls of irrigation with saline alone, intermittent t-PA bolus and drainage only. These findings support future use in clinical trials to assess whether this treatment modality can be associated with better functional outcomes.

Non-standard Abbreviations and Acronyms:

IVH

intraventricular hemorrhage

sICH

spontaneous intracerebral hemorrhage

EVD

external ventricular drain

ICP

intracranial pressure

CSF

cerebrospinal fluid

VPS

ventriculoperitoneal shunt

t-PA

tissue plasminogen activator

mGS

modified Graeb score

CT

computed tomography

mRS

modified Rankin Score

GCS

Glasgow Coma Scale

MISTIE

Minimally Invasive Surgery with Thrombolysis in Intracerebral haemorrhage Evacuation

CLEAR III

phase 3 Clot Lysis: Evaluating Accelerated Resolution of Intraventricular Haemorrhage

STROBE

STrengthening the Reporting of OBservational studies in Epidemiology

PACS

Picture Archiving and Communication System

Availability of data and materials:

not applicable