Bringing doctors together

or

A free, private network for verified physicians

Leading voices

Dr. Jonathan Rosand

Neurocritical Care · Broad Institute

View

Dr. Jinsy Andrews

ALS · NYU Langone

View

Dr. Maisha Robinson

Hospice & Palliative Care · Mayo Clinic Florida

View

Dr. Michael Kentris

Epilepsy · St. Elizabeth Medical Center

View

Dr. Daniel Correa

Epilepsy · AECOM

View

Dr. Aaron Berkowitz

General Neurology · UCSF

View

Dr. Jose Hernan Posas

Sports Neurology · UVA

View

Dr. Michael Okun

Movement Disorders · UF

View

Dr. Joseph Sirven

Epilepsy · Mayo Clinic College of Medicine

View

Dr. Rachel Marie Salas

Sleep Medicine · Johns Hopkins

View
  • Discover your community

  • Stay current

  • A profile that represents you

  • Find your people

About RoonPrivacyTerms

    Overview

    Outcomes Following Minimally Invasive Surgery for Intracerebral Hemorrhage in the AHA Get With The Guidelines-Stroke Registry.

    Apr 3, 2025DOI: 10.1161/strokeaha.124.048650Journal ArticleAHA

    Written by: Santosh B MurthyCenai ZhangAndrew L A GartonBrian Mac GroryShreyansh ShahGregg C FonarowLee H SchwammDeepak L BhattEric E SmithGuido J FalconeSeyedmehdi PayabvashWendy C ZiaiJared KnopmanCharles C MatoukJ MoccoHooman KamelKevin N Sheth

    Find these authors on Roon

    SS
    Dr. Shreyansh Shah
    Neurocritical Care · Duke

    Introduction

    Acute non-traumatic supratentorial intracerebral hemorrhage (ICH) is associated with high morbidity and mortality.1 Recent, pragmatic randomized trials investigating interventions such as early intensive blood pressure reduction and administration of hemostatic therapies suggest improved outcomes after ICH.2–4 In addition, surgical interventions such as hematoma evacuation have emerged as a potential avenue for mitigating the downstream effects of hematoma expansion and secondary injury.5–7 Earlier surgical trials using open craniotomy with clot evacuation was studied in the STICH −1 and STICH-2 (Surgical Trial in Lobar Intracerebral Haemorrhage) trials and found no improvement in functional outcomes in patients randomized to surgery.8,9 In contrast, the MISTIE III (Minimally Invasive Surgery Plus Alteplase for Intracerebral Hemorrhage Evacuation) trial, revealed that stereotactic MIS with concurrent fibrinolysis reduced mortality, albeit without impacting disability.10There has since been a renewed interest in the exploration of MIS techniques for ICH. In fact, the two newer modifications of MIS currently performed include using an endoport device and the other, using an endoscopic evacuation system.11 Recently, the ENRICH trial (Early Minimally Invasive Removal of Intracerebral Hemorrhage), demonstrated the superiority of early endoport-assisted MIS in treating moderate to large, lobar hemorrhages, with noted reductions in mortality and functional recovery.12 In light of these results, the MIND trial (Artemis in the Removal of Intracerebral Hemorrhage), which assessed endoscopic MIS using the Artemis device, was recently stopped.13 Given this backdrop of limited clinical trial data, a comprehensive examination of national, real-world MIS effectiveness is needed. To fill this need, we leveraged a large, clinical practice-based, stroke quality improvement registry in the US in order to analyze the association of MIS compared to standard medical treatment in patients with acute ICH.

    Methods

    We performed a retrospective cohort study using the American Heart Association Get With The Guidelines (GWTG)-Stroke registry in the US between January 2011 and December 2021. GWTG is the largest quality improvement registry for cardio and neurovascular disease in the United States, with half of all stroke admissions in the US captured in the registry. Since that data were collected for quality improvement purposes, and not primarily for research, the dataset used in this analysis are restricted and cannot be shared directly with other investigators. However, interested researchers may formally request access to the dataset by submitting a proposal to the American Heart Association-GWTG committee. Each participating hospital in the GWTG-Stroke registry received either human research approval to enroll cases without individual patient consent under the common rule, or a waiver of authorization and exemption from subsequent review by their institutional review board. Advarra, the IRB for the American Heart Association, exempted this study from further review. Additionally, Weill Cornell Medicine’s institutional review board also approved this analysis. This study was performed in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for reporting observational studies.14
    GWTG-Stroke is an ongoing, national, hospital-level, stroke registry sponsored by the American Heart Association as part of a quality improvement initiative. Details of the GWTG-Stroke registry, including case ascertainment and data collection methodologies, have been described elsewhere.15,16 Briefly, using a patient management tool, trained hospital personnel collect data on consecutive patients admitted with a stroke at a participating hospital, which include demographics, comorbidities, medications, diagnostic testing, in-hospital treatment, and discharge outcomes. Prior studies have reported >90% accuracy and concordance for the GWTG data, in comparison with medical record review.17 IQVIA (Parsippany, New Jersey) serves as the data collection and coordination center.
    We included patients diagnosed with a non-traumatic ICH in the GTWG-Stroke registry between January 2011 and December 2021. We excluded patients with a concurrent diagnosis of subarachnoid hemorrhage, subdural hemorrhage, or ischemic stroke to limit the erroneous inclusion of patients who had an ICH as a complication of therapeutic interventions. Patients who were transferred out to another hospital were excluded to prevent double counting of participants, similar to prior analyses.18,19 We also excluded patients who underwent open surgical approaches such as hemicraniectomy and craniotomy with hematoma evacuation. The study exposure was MIS, defined as a composite of stereotactic surgical hematoma evacuation with instillation of fibrinolytic agent, and endoscopic surgical hematoma evacuation, as ascertained in the GWTG registry.
    The primary outcome was in-hospital mortality. Secondary outcomes included poor discharge outcome which was defined as a composite of in-patient mortality or discharge to hospice, discharge to inpatient rehabilitation, ambulatory status at discharge (ambulatory with or without assistance versus non-ambulatory), and modified Rankin Scale (mRS) score at discharge. 20
    Using standard descriptive statistics, we reported data as counts and percentage, or median and interquartile range, unless specified otherwise. To minimize confounding by indication, we performed overlap weighting propensity score matching to ensure balance of covariates between ICH patients treated with and without MIS.21 This method assigns to each patient a weight that is proportional to their probability of receiving the alternative treatment option given their baseline characteristics. The overlap weighting creates exact balance on the mean of every measured covariate when the propensity score is estimated by logistic regression, and therefore, mimics the attributes of a clinical trial.21 The main assumptions of this method are the absence of unmeasured confounders and the existence of comparable individuals between the treatment and control groups for matching. While the first assumption is not fully testable, given the possibility of variables that are inherently unmeasured, we nonetheless, tried to minimize unmeasured confounding by matched patients with and without surgery on demographic, clinical, and hospital variables that have been comprehensibly shown to influence outcomes after ICH. To test the second assumption, we used Love plots to inspect the distribution of the propensity scores after matching and to confirm sufficient overlap between the treatment and control groups. To create propensity scores, a multivariable logistic regression model was created with MIS vs. no MIS as the dependent variable. Model covariates were chosen a priori based on prior literature and included demographics (age, sex, race), vascular comorbidities (hypertension, diabetes mellitus, prior stroke/transient ischemic attack), ICH severity (National Institutes of Health (NIH) Stroke Scale and external ventricular drain use), prior antithrombotic therapy, and hospital characteristics (annual hospital ICH case volume, academic hospital status, and hospital teaching status).22–25 The NIH stroke scale was used as a marker for ICH severity, while external ventricular drain was considered a surrogate for intraventricular hemorrhage or hydrocephalus, since these measures were not available in the GWTG dataset. The rationale for matching patients on prior antithrombotic therapy was to account for the independent relationship between the use of these medications and mortality.26,27
    After matching, the differences in the weighted means of the matched groups were compared using the standardized mean difference, for demographics, comorbidities, and hospital characteristics.
    In the primary analyses of the propensity-weighted population, we assessed the relationship between MIS and in-hospital mortality using multivariable logistic regression and calculated the odds ratios (OR) and 95% confidence intervals (CI). Additionally, adjusted risk differences were calculated from the model. In secondary analyses of the matched cohort, we used logistic regression to evaluate the association between the type of MIS (stereotactic or endoscopic) and mortality. To address the high early mortality in ICH patients, we performed a sensitivity analysis, where we excluded patients who had initiation of palliative care measures, withdrawal of care, or death in the first 48 hours of admission, and then used multivariable logistic regression to examine the association of MIS with in-hospital mortality. Lastly, in pre-specified subgroup analyses, we used similar methods to assess the effect of MIS on in-hospital mortality across the following prespecified subgroups: age (<65 vs. ≥65 years), sex (male vs. female), baseline NIH stroke scale (stratified by the median), external ventricular drain use (present vs. absent), and prior antithrombotic therapy (present vs. absent). The subgroup analyses entailed evaluation of the same outcome (mortality) in different subpopulations. To minimize the possibility of inflated Type I error in the multiple comparisons, we used the False Discovery Rate approach to control for the rate of false positives. All analyses were two-tailed and performed on the American Heart Association Precision Medicine Platform using SAS (version 9.4) and R Foundation (version 4.0.2).
    In the GWTG dataset, hospital-level data were complete for all hospital-level characteristics, but patient-level variables had missing values. While we did not impute missing data in the main analyses, we did perform additional analyses where we used multiple imputation by chained equation to impute the missing baseline values, and then matched patients with and without MIS. Patients with missing data on outcomes were not imputed and were excluded from the models.

    Results

    The GWTG cohort included 5,976,408 stroke patients, of whom 684,467 (11.5%) had a non-traumatic ICH. After relevant exclusions, the final analytical cohort included 555,964 ICH patients (Figure 1). The mean age was 68 (SD, 15.3) years, and 262,999 (47.3%) were female. A total of 703 patients underwent MIS, of whom 391 (53.1%) had endoscopic hematoma evacuation and 312 (46.9%) had stereotactic MIS. The median time to surgery was 2 days (IQR, 1–3). In the unmatched cohort, patients who underwent MIS were younger and had lower ICH severity on admission, while medically managed ICH patients had more vascular risk factors including hypertension, diabetes mellitus, atrial fibrillation, carotid stenosis, coronary artery disease and prior stroke (Table 1). MIS occurred more commonly at academic hospitals, teaching hospitals, and those with a higher annual ICH case volume. Baseline characteristics of the unmatched cohort in the endoscopic and stereotactic MIS groups are shown in Table 2.
    Participants with and without MIS were propensity matched on age, sex, race, National Institutes of Health (NIH) Stroke Scale, prior antithrombotic therapy, and external ventricular drain use resulted in 446 patients in MIS group and 35,361 in the control group. After matching, we examined the distribution of propensity scores to confirm the adequate overlap between groups. There were no significant differences between the groups that received and did not receive MIS (Supplemental Table I). Similarly, the matched cohort comprised 228 patients and 26,882 patients in the endoscopic MI and conservatively managed groups respectively, and were similar in demographics and comorbidities, and hospital characteristics. On the other hand, a total of 218 patients who underwent stereotactic MIS were matched to 43,435 patients who did not have surgery. The groups were well balanced after matching.
    The matched cohort had 8321 (23.5%) in-hospital deaths in the non-surgical group, and 60 (13.5%) deaths in the MIS group (Table 3). In the logistic regression analyses of the matched cohort, MIS was associated with lower odds of in-hospital mortality (adjusted odds ratio [OR], 0.50; 95% confidence interval [CI], 0.39–0.65; adjusted risk difference, −0.10; 95% CI, −0.13 to −0.07) as observed in Table 3 with the corresponding Love plot shown in Supplemental Figure 1.
    MIS was associated with lower odds of unfavorable discharge disposition as defined by a combination of in-hospital mortality and discharge to a hospice facility (aOR, 1.93; 95% CI, 1.61–2.32; adjusted risk difference, −0.11; 95% CI, −0.15 to −0.08). Conversely, MIS was associated with higher odds of favorable discharge to inpatient rehabilitation (aOR, 1.93; 95% CI, 1.61–2.32; adjusted risk difference, 0.16; 95% CI, 0.12–0.21). There was no relationship between MIS and functional outcomes (ambulatory status or dichotomous mRS scores) (Table 3), and the corresponding Love plots are shown in Supplemental Figure 2.
    In additional analyses, lower rates of mortality were observed in endoscopic MIS (40, 17.5% vs. 6,675, 24.8%) and stereotactic MIS groups (20, 9.1% vs. 9,582, 22.1%), compared to their matched non-surgical counterparts (Table 4, Supplemental Figure 3). In the adjusted multiple logistic regression models, decreased odds of in-hospital mortality were observed with endoscopic MIS (aOR, 0.64, 95% CI, 0.47–0.89) and stereotactic MIS (aOR, 0.36, 95% CI, 0.23–0.56). There was no relationship between the type of MIS and ambulatory status at discharge. Lower mortality was consistently noted with MIS after the exclusion of patients who had death or withdrawal of care in the first 48 hours (OR, 0.47; 95% CI, 0.35–0.63; adjusted risk difference, −0.10; 95% CI, −0.13 to −0.07), and when missing patient-level data were imputed (OR, 0.51; 95% CI, 0.42–0.62; adjusted risk difference, −0.11; 95% CI, −0.14 to −0.08) (Table 4). Lastly, similar reductions in mortality were observed with MIS across pre-specified subgroups, with the exception of the NIH stroke scale < 8 subgroup, where there was no association with mortality (Figure 2).
    In post hoc analyses, we performed propensity matching using the inverse probability weighted method instead of the overlap method, for the primary and secondary outcomes. The results were similar to those observed with overlap propensity matching and are shown in the Supplemental Table II.

    Discussion

    In this large, diverse, representative, nationwide U.S. registry of patients with ICH, MIS was associated with reduced in-hospital mortality and favorable discharge disposition. Both, endoscopic and stereotactic MIS approaches, demonstrated similar associations with mortality benefits.
    Our real-world findings add important new evidence to the growing body of evidence indicating that MIS is a viable treatment alternative to traditional surgery or medical management for ICH. The emergence of MIS techniques has resulted in the re-evaluation of surgery as a definitive treatment option for ICH.28 Several prior studies suggested a benefit of MIS, compared to open craniotomy and conservative medical management.29,30 In the MISTIE III trial, a significant reduction in mortality was observed with stereotactic surgery and concurrent thrombolysis.10 The ENRICH trial demonstrated benefit of the endoport-assisted parafascicular technique in improving mortality and disability.12 While regional data suggest gradually increasing trends in the implementation of MIS31, data are lacking on the large-scale, real-world relationship of MIS with outcomes, particularly outside of clinical trials and single center patient populations. In this context, our study fills this gap by supporting MIS efficacy with evidence suggesting clinical effectiveness towards reduced mortality in a nationally representative sample. In conjunction with the results of the ENRICH trial, our findings add to a growing body of literature on MIS and may have important implications for systems of care and emergent transfer of ICH patients.
    The results of this analysis demonstrate the benefits of MIS on in-hospital mortality and provide a short window for evaluating functional outcome recovery. Consistent with recent data suggesting that functional outcome improvement requires longer term assessment and that ICH patients have a protracted trajectory of functional recovery, directionally favorable functional outcomes were not significant in this analysis since our assessment was at discharge.32 Prior literature has also shown an inverse relationship between time to MIS and good outcomes, attesting to the time is brain concept.33 Although the median time to surgery was 2 days (IQR, 1–3) in our study, this information was missing in over 50% of the patients, which precluded further analysis of the influence of surgical timing on outcomes. This study provides unique, early insights from a national registry about ICH patients likely to receive MIS. Specifically, patients with ICH who received MIS were younger, had lower prevalence of cardiovascular comorbidities such as hypertension, diabetes, coronary artery disease, and atrial fibrillation, had higher ICH severity, and were more likely to be treated at tertiary care hospitals with a high annual volume of ICH admissions. Notably, MIS did not offer any mortality benefit in low ICH severity (NIHSS < 8), since these patients were likely destined to have a good recovery regardless of surgery and these results mirror those of prior pooled analyses comparing open craniotomy to medical management.34 It is also noteworthy that previous AHA guideline recommendations (2010 and 2015) for the management of ICH did not recommend routine surgical clot evacuation, via either open or minimally invasive approaches, for supratentorial ICH.35,36 This likely explains the overall low rates of surgeries performed for supratentorial ICH in the GWTG registry. These procedures were mostly performed in tertiary care hospitals and teaching facilities. As MIS becomes more widespread in the community and is adopted as the standard of care, the true impact of this procedure on ICH outcomes can be ascertained.
    Our study has some limitations. First, the present study was potentially subject to confounding by indication in that clinicians offered MIS to ICH patients deemed likely to survive or have a meaningful recovery. To minimize confounding by indication, we matched the two groups on demographics and ICH severity, factors which have been consistently shown to influence outcomes22,23, and considered by clinicians for prognostication. While the inherent confounding by indication in our study can only be addressed by a randomized clinical trial to inform clinical decision making, results from the ENRICH trial demonstrating the benefit of MIS after ICH provide reassurance about the validity of our results.12 Second, our results may have limited generalizability to ICH patients treated at hospitals outside the GWTG-Stroke registry or in other countries. That said, GTWG-Stroke is the largest stroke registry that encompasses the majority of the tertiary hospitals, where ICH patients are likely to be treated, and is hence representative of the ICH treatment practices in the US. Third, we combined endoscopic and stereotactic MIS approaches in the primary analysis, even though they entail completely different operative techniques. Moreover, the stereotactic and endoscopic surgical approaches in the GWTG registry likely did not capture the endoport-assisted evacuation since the procedure involves open craniotomy with exoscopic visualization and conventional microsurgical techniques.37 Regardless, the results remained unchanged when the endoscopic and stereotactic MIS approaches were analyzed separately in secondary analyses. Future real world analyses should capture, compare, and contrast distinct operative techniques Fourth, we did not have granular data on ICH imaging characteristics including ICH location, hematoma volume, presence of intraventricular hemorrhage, or hematoma expansion, which are independently associated with poor outcomes.38
    In recent decades, comprehensive stroke centers have urged hospitals to create interdisciplinary stroke and critical care neurology units in order to provide a range of treatment options to ICH patients. These results, in combination with recent clinical trial data, suggest an important role for neurosurgical, critical care and vascular neurology collaboration in order to optimally identify and manage acute ICH care.39 In decades past, acute ICH was met with considerable nihilism40, but recent advances suggest that acute ICH is a ready population for quality improvement initiatives directed at implementation of evolving care pathways for ICH.41

    Conclusions

    In the largest quality improvement stroke registry in the US, MIS was associated with reduced in-hospital mortality after ICH. Studies with long-term follow functional outcome measures are needed to fully understand the impact of MIS on ICH treatment.
    SSDr. Shreyansh Shah

    Dr. Shreyansh Shah

    Neurology · Neurocritical Care
    Duke University Health System
    shreyansh