Article Text

Download PDFPDF

Original research
Minimally invasive endoscopic hematoma evacuation vs best medical management for spontaneous basal-ganglia intracerebral hemorrhage
  1. Nitin Goyal1,
  2. Georgios Tsivgoulis1,2,
  3. Konark Malhotra3,
  4. Aristeidis H Katsanos2,
  5. Abhi Pandhi1,
  6. Khalid A Alsherbini1,
  7. Jason J Chang1,4,
  8. Daniel Hoit5,
  9. Andrei V Alexandrov1,
  10. Lucas Elijovich1,5,
  11. David Fiorella6,
  12. Christopher Nickele5,
  13. Adam S Arthur5
  1. 1 Department of Neurology, University of Tennessee Health Science Center, Memphis, Tennessee, USA
  2. 2 Department of Neurology, Attikon University General Hospital, School of Medicine, National & Kapodistrian University of Athens, Athens, Greece
  3. 3 Department of Neurology, West Virginia University Charleston Division, Charleston, West Virginia, USA
  4. 4 Medstar Washington Hospital Center, Washington, District of Columbia, USA
  5. 5 Department of Neurosurgery, University of Tennessee Health Science Center, Memphis, Tennessee, USA
  6. 6 Department of Neurosurgery, Stony Brook Medical Center, New York, USA
  1. Correspondence to Dr Adam S Arthur, Semmes-Murphey Clinic, Memphis, TN 38120, USA; aarthur{at}


Background We conducted a case-control study to assess the relative safety and efficacy of minimally invasive endoscopic surgery (MIS) for clot evacuation in patients with basal-ganglia intracerebral hemorrhage (ICH).

Methods We evaluated consecutive patients with acute basal-ganglia ICH at a single center over a 42-month period. Patients received either best medical management according to established guidelines (controls) or MIS (cases). The following outcomes were compared before and after propensity-score matching (PSM): in-hospital mortality; discharge National Institutes of Health Stroke Scale score; discharge disposition; and modified Rankin Scale scores at discharge and at 3 months.

Results Among 224 ICH patients, 19 (8.5%) underwent MIS (mean age, 50.9±10.9; 26.3% female, median ICH volume, 40 (IQR, 25–51)). The interventional cohort was younger with higher ICH volume and stroke severity compared with the medically managed cohort. After PSM, 18 MIS patients were matched to 54 medically managed individuals. The two cohorts did not differ in any of the baseline characteristics. The median ICH volume at 24 hours was lower in the intervention group (40 cm3 (IQR, 25–50) vs 15 cm3 (IQR, 5–20); P<0.001). The two cohorts did not differ in any of the pre-specified outcomes measures except for in-hospital mortality, which was lower in the interventional cohort (28% vs 56%; P=0.041).

Conclusions Minimally invasive endoscopic hematoma evacuation was associated with lower rates of in-hospital mortality in patients with spontaneous basal-ganglia ICH. These findings support a randomized controlled trial of MIS versus medical management for ICH.

  • endoscopy
  • hemorrhage
  • standards
View Full Text

Statistics from


Although spontaneous intracerebral hemorrhage (ICH) accounts for only 10%–15% of all strokes, it contributes significantly to worldwide morbidity and mortality, with 3- month mortality rates for these patients reaching 40%–50%.1 2 It is thought that early deterioration during the initial hours after ICH is due to hematoma expansion, whereas secondary injury results from perihematomal edema and hemotoxicity from blood degradation products.3 Trials of standard craniotomy and hematoma evacuation for spontaneous supratentorial ICH have failed to show any impact on functional outcomes or mortality.4 5 Various endoscopic guided minimally invasive surgery (MIS) procedures have evolved over the past decade based on encouraging results from pilot randomized controlled clinical trials6–8 and meta-analyses.9

One of these innovative technologies is the Apollo System (Penumbra Inc, Alameda, CA), which used a non-clogging aspiration and irrigation system. This device was initially approved by the US Food and Drug Administration for the evacuation of ventricular hemorrhage. This indication was later expanded to include the removal of hemorrhagic products during surgery of the cerebrum.3 In the present case-control study, we analyzed our outcomes using the Penumbra Apollo system compared with best medical management alone in patients with basal-ganglia hemorrhages.


Study population

We conducted a retrospective analysis of a prospectively collected database of consecutive patients with acute (<24 hours) spontaneous ICH in a tertiary-care stroke center from July 2014 to December 2017 as previously described.10–12 Inclusion criteria were as follows: spontaneous cause for ICH; medically managed or underwent Apollo-assisted endoscopic clot evacuation; adult age (>18 years old); and basal-ganglia location. Exclusion criteria included the following: nonspontaneous etiologies of ICH (including traumatic ICH, metastatic hemorrhagic cerebral lesions, ICH resulting from venous sinus thrombosis, ICH resulting from hemorrhagic transformation of ischemic strokes, and ICH resulting from underlying vascular lesions); ICH attributable to supratherapeutic international normalized ratio (INR) in the setting of pre-hospital anticoagulation or coagulopathy (threshold INR of 1.5 or greater) and thrombocytopenia (Platelets<50 000/mm3); age <18 years of age; location other than basal ganglia; and herniating ICH patients who required urgent decompressive hemicraniectomy. Institutional investigation review board (IRB) approval for this study was granted based on the acute ICH database (18–05826-XP). The need for patient consent was waived according to IRB recommendations.

Baseline characteristics and outcome measures

We obtained demographics, medical history, premorbid modified Rankin Scale (mRS) scores, and baseline radiological and clinical parameters as previously reported by our group.10–12 Baseline severity of neurological dysfunction was documented using National Institutes of Health Stroke Scale (NIHSS) score: baseline severity of ICH was quantified by ICH score.11 Admission hematoma volume was measured using the ABC/2 method.10–12 All patients underwent CT of the head at hospital admission. Follow-up CT was acquired within 6 to 24 hours of initial head CT.12

Outcome measures included in-hospital mortality, intensive care unit length of stay, discharge NIHSS score, mRS at discharge, discharge disposition (good (home or inpatient rehabilitation) vs poor (nursing facility or morgue)), mRS at 3 months, and good functional outcome (defined as mRS score of 0–3 at 3 months). Baseline characteristics and outcomes were compared between ICH patients who underwent MIS with Apollo endoscopic evacuation (intervention group), and medically managed individuals (control group) before and after propensity score matching (PSM).

Apollo endoscopic intracerebral hemorrhage evacuation

The use of Apollo for endoscopic-assisted evacuation of ICH was conducted in a hybrid operating room with intraprocedural imaging for the assessment of the hematoma evacuation. The imaging was first reviewed so that a general plan could be made for the proper placement of the burr hole. Stereotactic navigation was used intraprocedurally to ensure that the trajectory to the hematoma passes through the long axis of the lesion, which ensures that the endoscope sheath can be manipulated as little as possible. A cone beam CT was obtained and registered with the stealth system, which allows the entry point and incision to be marked, followed by prepping and draping the field in the usual sterile fashion. After incision, a burr hole was drilled and the dura opened in a cruciate fashion. The underlying pia was then coagulated and incised in preparation for placement of a 19 F endoscope sheath. The sheath was introduced with neuronavigation in a trajectory through the long axis of the clot to a depth approximately 1 cm short of the deep clot margin. Next, the Apollo device was placed through the working channel of the endoscope and actuated for controlled hematoma evacuation under direct endoscopic visualization. The sheath can be gently manipulated to reach the edges of the hematoma. As the evacuation proceeded (distal to proximal), the sheath was gradually withdrawn during the procedure. Typically, the intracerebral pressure functions to extrude the clot toward — and sometimes into — the sheath, and much of the Apollo usage is just at the end of the sheath rather than out in the brain cavity. After the sheath has been withdrawn to the proximal margin of the cavity, it can be removed, and a cone beam CT of the head can be performed to check progress. If a second pass is required, the new scan can be used to navigate the sheath into the residual hematoma. If hematoma removal is judged adequate and hemostasis is evident, the burr hole can be covered with a plate, and the wound closed.

Medical management

As per the protocol at our center, ICH patients are admitted to the neuro-critical care unit and managed by the neuro-intensivist service. Medical management is guided by international recommendations5 and individualized in specific patients as needed. These recommendations include blood pressure control, intracranial pressure control strategies in various combinations (elevation of head of bed, sedation/analgesia, normocarbia/hypocarbia, hyperosmolar treatment, cerebrospinal fluid (CSF) drainage via external ventricular drain, hypothermia or barbiturate coma), deep venous thrombosis prophylaxis, and management of secondary complications, such as seizures, hydrocephalus, or salt wasting.

Statistical analyses

Patients in the treatment group receiving endoscopic hematoma evacuation were matched to control group patients receiving medical management. For matching, we used a structured, iterative propensity score model with inclusion of all baseline characteristics, except for the treatment variable (endoscopic hematoma evacuation or medical management), with the primary objective of maximizing the balance in the distribution of possible confounders between the two aforementioned groups. The corresponding propensity score of the treatment variable was calculated for each subject and a nearest neighbor matching algorithm with a 1:3 allocation was subsequently implemented to match eligible patients in the treatment group (patients receiving endoscopic hematoma evacuation) to patients in the control group (patients receiving medical management) within 0.2 x SD of the logit of the propensity score, according to their treatment status. To determine whether the propensity score matching approach achieved balance in all potential confounders, we compared all baseline characteristics of patients in the treatment group to their control patients before and after propensity score matching.

Continuous variables with normal distributions are presented as means with SD, whereas those with skewed distributions are presented as medians with IQR. Categorical variables are presented as percentages. Statistical comparisons between the two groups (Apollo endoscopic hematoma evacuation or medically managed) were performed using the Pearson’ s χ2 test, unpaired t-test or Mann–Whitney U test where appropriate. The distribution of the mRS-scores between the two groups at both discharge and 3 months was assessed with the Cochran–Mantel–Haenszel test. The differences in all baseline characteristics and outcomes of interest between the two groups were tested under statistical significance hypotheses using an alpha value of 0.05. All statistical analyses were performed with the Stata Statistical Software Release 13 (College Station, TX, StataCorp LP).


A total of 572 consecutive ICH patients were evaluated. Of those, 224 ICH patients fulfilled study inclusion criteria — endoscopic clot evacuation (n=19), medical management (n=205). Baseline characteristics between the two groups are presented in online supplemental table 1. The intervention cohort had higher median admission ICH volume (40 cm3 (IQR, 25–51) vs. 9 cm3 (IQR, 3–18); P<0.001) and median admission NIHSS score (21 points (IQR, 18–24) vs 11 points (IQR, 4–17); P=0.017) compared with controls.

After PSM, 18 ICH patients in the intervention cohort were matched to 54 patients in the control cohort (1:3 matching). One patient in the treatment group was discarded after propensity score matching due to no available matches in the control group within the predefined caliper. Table 1 compares the baseline characteristics between the two propensity-matched cohorts. The two cohorts did not differ in any of the baseline characteristics, including median admission ICH volume (40 cm3 (IQR, 25–50) vs 33 cm3 (IQR, 17–47); P=0.424), median admission NIHSS score (22 points (IQR, 20–24) vs 21 points (IQR, 16–34); P=0.794), and mean admission ICH score (2.2±0.9 points vs 2.5±1.3 points; P=0.341). The median ICH volume at 24 hours was lower in the intervention group (40 cm3 (IQR, 25–50) vs 15 cm3 (IQR, 5–20); P<0.001). The median decrease in ICH volume in the intervention group was 24 cm3 (IQR 11–39).

Table 1

Baseline characteristics of propensity-matched (1:3) cohorts (Apollo-assisted minimally invasive endoscopic hematoma evacuation vs medical management)

Outcomes in the two groups after PSM are presented in table 2. The two cohorts did not differ in any of the pre-specified outcomes measures except for in-hospital mortality, which was lower in the intervention cohort (28% vs 56%; P=0.041). The 3-month mortality rates tended to be lower in the intervention cohort (38% vs 60%; P=0.107). The intensive care unit length-of-stay(LoS) was similar in the two groups (median LOS in days (IQR), intervention, 97–12 vs control, 96–12; P=0.497). The distribution of mRS scores at discharge and at 3 months did not differ between the intervention and the control cohort (P by Cochran–Mantel–Haenszel test >0.4). The following peri-procedural complications were documented in the intervention cohort: CSF leak (n=2); and recurrent ipsilateral or contralateral stroke (n=4).

Table 2

Univariable associations of various outcome measures in two propensity-matched groups (Apollo-assisted hematoma evacuation or medical management)


This case-control study indicates an absolute risk reduction of 28% and 22% in in-hospital and 3-month mortality in patients with spontaneous basal-ganglia ICH who underwent Apollo-assisted endoscopic clot evacuation compared with PSM patients treated with medical management. Additionally, the reduction in ICH volume at 24 hours was greater in the intervention group. Nevertheless, the two groups did not differ in the distribution of discharge and 3-month mRS scores, as well as the 3-month good clinical outcome (mRS scores of 0–3) rates.

To the best of our knowledge, this is the first comparative study to evaluate the efficacy of the Apollo-assisted endoscopic clot evacuation system compared with medical management. Prior case series in both cadaver models13 and humans14 15 have demonstrated the feasibility and safety of MIS using the Apollo system. Individual case examples have demonstrated favorable clinical outcomes in some patients. Spiotta et al reported results for 29 ICH patients (23 parenchymal and six with additional intraventricular hemorrhage) undergoing MIS with Apollo.14 These investigators reported a technical success rate of 91.7% with a mean reduction in ICH volume of 54%. They observed a low in-hospital mortality rate of 13.8% despite the inclusion of patients with large hemispheric bleeds (mean ICH volume, 45.4±30.8 cm3). These findings are in line with our observations that MIS may reduce discharge and 3-month mortality rates.

The Apollo system is an innovative technology that involves a stereotactic MIS approach for hematoma evacuation through a burr hole. It may confer major advantages in comparison to techniques used in the Minimally Invasive Surgery Plus Tissue-Type Plasminogen Activator for ICH Evacuation (MISTIE) trial, which involves the administration of a thrombolytic agent into the hematoma to effect passive drainage.16 The MISTIE technique requires several days to effect clearance of the hemorrhage and entails the additional risk of an indwelling parenchymal catheter that must be manipulated multiple times over several days for tissue plasminogen activator (tPA) injections. It also requires many follow-up CT scans to monitor catheter position and the hemorrhage evacuation, and is thus relatively time and resource-intensive.17 In contrast to the prolonged thrombolytic drainage required in MISTIE, the Apollo system achieves immediate hematoma evacuation during the procedure and does not require tPA administration. These advantages may result in shorter intensive care unit stays, and a reduction in iatrogenic re-hemorrhage rates. Another method of endoscopic hematoma evacuation that involves diffusion tensor imaging technology, the Brain Patch tubular retraction device (a high-definition visualization platform and a specially designed resection device), is being studied and has shown promising results.18 It is designed to avoid excessive damage to white matter tracts while accessing deep lesions.18 However, the experience with this new device is limited to smaller multicenter series, and large series and randomized trials are needed.

In the present series, we did not detect any improvement in functional outcomes in the MIS cohort. This was not unexpected given the small size of the intervention cohort (n=18). Moreover, the short 3-month follow-up duration prevented us from assessing functional outcomes (ie, mRS) at 6 or 12 months when clinical recovery following ICH is maximized.19 20

Future randomized controlled clinical trials are warranted to document whether MIS using mechanical aspiration under endoscopic guidance (Apollo (first generation) and now Artemis (second generation), Penumbra Inc, Alameda, CA) is associated with improved functional outcomes in patients with supratentorial ICH. Two studies — the prospective multicenter INVEST-FEASIBILITY study ( Identifier: NCT02661672) and the randomized controlled MIND trial (a prospective, multicenter study of Artemis, a Minimally Invasive Neuroevacuation Device, in the removal of intracerebral hemorrhage) ( Identifier: NCT03342664) — are currently recruiting patients with supratentorial ICH.

Our study has important limitations that need to be acknowledged. First, the retrospective analysis of our prospectively collected database confers important methodological shortcomings. Second, assessment of imaging for ICH volumes was not adjudicated. Third, the disparities in withdrawal of care between the two cohorts (medical arm (46%) vs MIS (28%); P=0.269) — although non-significant — may have influenced the reductions in mortality rates that were observed in the intervention cohort. In the medical arm, 40% of all deaths were secondary to withdrawal of care, compared with 10% of all deaths in the treatment arm. However, it is important to note that the early withdrawal of care is discouraged at our center, and it is usually proposed to families only after maximal medical and/or surgical options have been exhausted. Fourth, the relatively short clinical follow-up time may have reduced the ability to identify a beneficial effect on functional outcome after MIS. Fifth, some of our neurosurgeons are trained in minimally invasive surgery for clot evacuation while others are not. Consideration of this therapy was therefore limited based on which surgeons were on call. We cannot rule out selection bias in this cohort, but descriptive statistics seem to provide some reassurance that the patient cohorts are similar in many ways. Sixth, we only studied patients with basal-ganglia ICH location in our study because treating physicians at our center felt that the ICHs in deep basal-ganglia location is one subgroup among supratentorial locations that would potentially benefit most from endoscopic hematoma evacuation, especially in light of previously failed trials of open surgical hematoma evacuation. Having said that, we agree that the MIS technique should be studied for ICHs in all supra-tentorial locations. The results of ongoing randomized controlled trials will provide more information on this issue. Seventh, the mortality rates in our cohorts are high, explained by the inclusion of patients with high ICH scores and large admission hematoma volumes. The mortality rate in our study population is comparable to prior published literature on outcomes of patients with severe ICH.21 22 Finally, the small sample size of this pilot study makes it vulnerable to residual confounding.

In conclusion, our pilot case-control study provides preliminary evidence indicating that MIS using a mechanical aspiration system under endoscopic guidance is associated with greater reduction in hematoma volume at 24 hours and may reduce in-hospital mortality rates in patients with supratentorial ICH. A randomized controlled trial is justified, and underway, to definitively assess the potential clinical benefits of this promising technology.


The authors wish to thank Andrew J. Gienapp, BA (Neuroscience Institute, Le Bonheur Children’s Hospital, Memphis, TN and Department of Neurosurgery, University of Tennessee Health Science Center, Memphis, TN) for technical and copy editing, preparation of the manuscript, figures, and table for publishing, and publication assistance with this manuscript.


View Abstract


  • Contributors All authors of this work met ICMJE criteria for authorship and made substantial contributions to the conception and design, acquisition of data, analysis and interpretation of data, drafting, critical revising, and final approval of this manuscript.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests NG, GT, KM, AHK, AP, KA, JJC, AVA, and CN have no financial relationships to disclose. DH is a consultant for Codman, Neurovascular, Medtronic, MicroVention, Penumbra, Sequent, and Stryker. LE is a consultant for Codman Neurovascular, Medtronic, MicroVention, Penumbra, Sequent, and Stryker. DF, is an unpaid consultant and proctor for ev3/Chestnut Medical. ASA is a consultant for Leica, Medtronic, Microvention, Penumbra, Siemens, and Stryker, receives research support from Microvention, Penumbra, and Siemens, and is a shareholder in Bendit, Cerebrotech, Serenity, and Synchron.

  • Patient consent Not required.

  • Provenance and peer review Not commissioned; internally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.