Recommendations for Clinical Trials in ICH

The first Hemorrhagic Stroke Academia Industry (HEADS) roundtable in 2017 established priorities for basic, translational, and clinical research in intracerebral hemorrhage (ICH) and benchmark standards for improving research in this globally serious and challenging condition.^1,2 Herein, we present updated recommendations from the second HEADS roundtable meeting in 2019, which focused on the methods of clinical trials in acute ICH.

The second HEADS roundtable (HEADS-2) was convened alongside the 7th World Intracranial Hemorrhage Conference in Granada, Spain in May 2019. The neutral results of several clinical trials in acute ICH published since the first HEADS meeting in 2017 have reinforced the challenges of advancing strategies for the management of this serious condition.^3–5 The focus of HEADS-2 was on the execution and conduct of clinical trials in ICH, with the aim of developing recommendations to improve their design and conduct. The format was similar to the first HEADS roundtable,^1,2 with Day 1 consisting of three sessions covering the following topics: (1) target selection and approaches to variables influencing outcomes including hematoma expansion, perihematomal edema (PHE), and secondary injury; (2) selecting the relevant eligible patient population for ICH trials, with an emphasis on timing, age and sex, hematoma volume and location, presence of intraventricular hemorrhage (IVH), and comorbidities; and (3) challenges in performing ICH trials from industry and academic perspectives. On Day 2, participants broke out into 3 working subgroups to address these issues. This paper consolidates the HEADS-2 recommendations with the aim of improving the rigor, efficiency, and success of future clinical trials to find effective therapies for ICH.

Intracerebral Hemorrhage Is Distinct From Ischemic Stroke

For many years, ICH has been the bridesmaid, trying to follow the footsteps of its ischemic counterpart.⁶ However, it has become increasingly evident that clinical research in ICH has clear distinctions from acute ischemic stroke (AIS). First, the global burden of ICH is higher, particularly in low- and middle-income countries, and Asia; and the loss of productive life years from ICH is far greater than from AIS.⁷ The high case fatality and morbidity after ICH, complexities and variability of care, and early withdrawal-of-treatment; all significantly limit the number of potential participants available for clinical trials. Improved understanding of the patterns and process of hematoma expansion and secondary injury in ICH and differences in immediate therapeutic goals after ICH compared with AIS also impact on trial design. While the primary objective in AIS is to resuscitate the penumbral tissue to minimize disability, the primary objective in ICH is to prevent rapid deterioration and early death. Although initial interventions that target hematoma growth in ICH might be more steeply time-dependent than reperfusion therapy in AIS, the pathophysiology of secondary injury in ICH also provides opportunities for a wider therapeutic window. These distinctive differences elucidate the need for greater global collaboration to increase recruitment and improve the efficiency of ICH trials.

Target Selection and Timing of Interventions in ICH Trials

Broadly, the potential therapeutic targets for ICH are prevention, limiting hematoma expansion, hematoma evacuation, or reduction, reducing ICH-induced secondary brain injury and improving repair. ICH prevention was not a focus of HEADS-2. ICH-induced brain damage involves both primary injury, including the physical disruption caused by the hemorrhage and its expansion, and secondary injury, including downstream injury cascades initiated by the initial physical injury, clot-derived neurotoxic factors such as hemoglobin and iron, and neuroinflammation^8,9; all of which are intimately related.

Hematoma Expansion

Hematoma volume and expansion are powerful predictors of death and disability after ICH. Approximately, one-third of patients demonstrate substantial hematoma expansion (HE) in the first few hours after ICH, which worsens outcomes.¹⁰ Timing of observation influences the frequency of HE; presentation within the first few hours is associated with a higher likelihood of HE.¹¹ Hematoma expansion is potentially modifiable, and limiting early HE has been an appealing therapeutic target for ICH. The therapeutic options to restrict HE can be medical (nonsurgical) or surgical, and the 2 approaches might be complementary. Previous trials of activated recombinant Factor VIIa¹² and tranexamic acid⁵ have shown that these interventions can modestly arrest hematoma growth. Minimally invasive clot evacuation techniques provide an opportunity to modify and reduce ICH volume.³ However, to-date, these interventions did not result in improved functional outcomes.^3,5,12 The neutral results of these ICH trials may have been influenced by time delays over patient selection and initiation of interventions, which may be most effective very soon after ICH onset.

Secondary Injury After ICH

Unlike HE, the secondary injury after ICH occurs over hours to weeks and is attributed to the toxic effects of blood degradation products emanating from hemolysis of red blood cells, thrombin production, and neuroinflammation.⁸ These processes lead to disruption of blood-brain barrier and cerebral edema. PHE, which can readily be detected using different imaging modalities, is thought to represent a radiological marker of secondary injury and is increasingly recognized as a therapeutic target or proof-of-concept surrogate marker in assessing the efficacy of interventions targeting the secondary injury in phase II ICH trials. However, whether reductions in PHE can be used as a marker of therapeutic efficacy is debated, as its assessment is complex, and formation and evolution cover several processes.¹³ Early PHE involves clot retraction and hydrostatic pressure change forcing serum into the perihematomal space to form vasogenic edema. Thrombin formation and activation of the coagulation cascade, erythrolysis and hemoglobin-mediated toxicity, and inflammation also contribute to the formation of vasogenic edema in the following days, while neuronal death and energy failure contribute to the formation of cytotoxic edema. These chronological pathophysiological changes must be taken into consideration, and correction algorithms should be used in the assessment of PHE. Furthermore, the relationship between PHE and long-term outcomes is uncertain. While malignant PHE may contribute to increased intracranial pressure, midline shift, and potentially herniation and death, the effects associated with small hematoma are less clear. Previous studies have been largely observational and retrospective and used different parameters, assessors, onset-to-imaging and onset-to-outcome assessment times, outcome definitions and end points, and imaging modalities.¹⁴

There are pros and cons to potential therapeutic interventions that may target both primary and secondary ICH-induced brain injury. Reducing hematoma volume and expansion may limit both primary and secondary injury, but such expansion is time-dependent and only occurs in a subset of patients with ICH.¹⁰ Therefore, the effects of hemostatic agents may be maximized by very early delivery after ictus (eg, within the first 2 hours). Early arrest of hematoma growth and its stabilization may also facilitate earlier clot evacuation. There are data to suggest that minimally invasive surgical techniques, which do not use thrombolytic agents, can be safely done within few hours of ICH onset.^15,16 Therefore, the timing of surgery and its relationship to ongoing hematoma expansion (benefit versus harm) are important targets for future research.

Drugs targeting secondary brain injury avoid the use of surgery and can have an approximate therapeutic time window of 24 to 72 hours, dependent on agent, in preclinical studies. However, they do not affect ICH-induced primary injury. In addition, secondary brain injury after ICH involves multiple pathways which evolve spatially and temporally.⁹ Indeed, the same pathway (eg, neuroinflammation) may have detrimental or beneficial effects depending on time.¹⁷ The sequence of events and the human time window requires further evaluation.

Imaging can provide insights into acute brain injury after ICH and the potential benefit of therapeutic strategies. One commonly assessed injury biomarker is PHE. However, there are several unanswered questions and opportunities for standardizing assessment of PHE to allow better evaluation of its role as a therapeutic target or surrogate marker. For example, the relationship between PHE and mass effect (assessed by midline shift, ventricular ablation, or intracranial pressure measurements) and outcome requires further examination. Also, the impact of age and age-related brain atrophy on these intermixed variables should be given consideration. There is a need to utilize new imaging technologies, such as magnetic resonance diffusion tensor imaging, to examine local versus global mass effect and the impact of PHE location on mass effect and outcome. Understanding the importance of PHE as a brain injury mechanism requires trials of specific anti-edema therapies. Examination of how PHE correlates with neuroinflammation, hematoma changes, and white matter injury may give further insight.

Recommendations

• The targets described above are not mutually exclusive; a multimodal approach is likely required for ICH (eg, hemostatic therapy to arrest hematoma expansion followed by hematoma evacuation, and combined with an agent to limit secondary brain injury).

• There are limited resources and a preclinical network directly comparing different potential therapies may inform decision-making over which therapies should go forward to clinical trial.

• Access to data from human hematoma samples should be utilized as one of multiple bidirectional data flow to better inform preclinical studies and future trial designs.

• Future studies of PHE should use rigorous methodologies and systematic and standardized approaches,¹⁴ with prespecified analyses, end points, and adequate statistical power.

Recovery and Timing of Outcome Assessments

Most ICH survivors are left with some degree of disability. Therefore, rehabilitation has an important role in promoting recovery and readjustment to the illness in ICH survivors. However, there are few data on the natural long-term course and degree of recovery after ICH, frequency of the various ICH-related impairments, and their evolution over time.¹⁸ Most ICH randomized trials have conventionally assessed functional outcomes at 90 days (as in AIS), but some have pushed this out to 365 days.^3–5 Emerging data from recent trials^3,4 confirm that recovery takes longer after ICH compared with AIS and that a 90-day outcome assessment is too early to capture the full extent of recovery.

In most large randomized controlled rehabilitation trials, patients with ICH have been included alongside patients with AIS. Injury mechanisms differ between hemorrhage and ischemia and recovery also seems to differ.¹⁹ This is a ripe area for research to help determine the duration and intensity of rehabilitation in the acute and chronic phases of ICH. In addition, such trials may be used to assess physiotherapeutic methods, robot-assisted technologies, and wearable sensors and devices to track motor recovery after ICH. There is also a great need to understand why patients vary greatly in their degree of recovery following ICH. To what extent is this dependent on the degree of initial injury, the site of injury, individual genetics, or presence of comorbidities? Advances in imaging modalities, such as diffusion tensor imaging or lesion mapping, and use of long-term and task-dependent imaging and resting-state functional MRI, could help address patient variability to predict recovery, and tailor patient-specific rehabilitation programs to maximize recovery. Our understanding of motor function deficits after ICH and ability to address them with rehabilitation are generally more advanced than with psychosocial dysfunction that also attends ICH. This is a very important area of study that merits resources. Cognitive decline occurs in many individuals after ICH with an enormous impact on the patient, family, and society; there is a need to identify rehabilitation interventions targeting cognition.

Recommendations

• Key outcomes for ICH—death and disability/dependency—might be best assessed at separate time points; death at 7 or 30 days, and disability at 6 to 12 months. This recommendation takes into account the pros and cons associated with increased study duration.

• There is a need for rehabilitation trials which specifically focus on patients with ICH.

Patient Selection and Adjustment for Prognostic Confounders

Patient selection in ICH trials influences the detection of a possible benefit and adequate evaluation of safety and has implications for the overall wider applicability of the trial results. As the characteristics of selected patients have prognostic impact on outcomes, it is important that randomization is representative of and balanced according to key baseline variables, such as age, IVH, hematoma location and volume, and important comorbidities. Adjustment for imbalances of biologically established covariates should be undertaken in final analyses.

Age

Many ICH randomized trials have limited enrollment to patients aged ≤80 years. As old age alone may have little impact on the effects of an intervention, excluding patients older than 80 years from trials limits the generalizability of the results to a large proportion of patients with ICH who have the potential to benefit. However, older age is a marker of accumulating comorbidities and may be associated with increased likelihood of advanced directives and early withdrawal-of-care after ICH.

Recommendations

• Substituting age with a functional threshold (eg, modified Rankin Scale score >1) or requiring an intent to provide aggressive care and postpone withdrawal-of-care orders (eg, 2–4 weeks) as criteria for enrollment in ICH trials could alleviate the preceding concerns.

• Epidemiological and trial data should be used to evaluate the relationship between age, hematoma volume, location, and outcome to inform trial designs, and to assess whether a specific upper age limit is justified as an eligibility criterion for ICH trials.

IVH

Although the presence and volume of IVH has important implications on prognosis in ICH,^19,20 these variables are rarely included in the randomization process of ICH trials. At a minimum, adjustment for these variables should be undertaken in the analyses of final results. The presence of IVH alone should not be an exclusion criterion. Validated scales or full quantification of IVH volume should be employed.^21,22

Recommendations

• Trial and epidemiological data should be developed to define if an upper limit for IVH volume exists to better inform exclusion and randomization processes.

Hematoma Location

The site of ICH has a clear impact on survival and functional outcome.^23,24 Infratentorial location, in particular, predicts higher likelihood of death, and is a current clinical indication for surgery in case of cerebellar ICH.²⁵ Consequently, most ICH trials exclude patients with infratentorial hemorrhage. Prognosis also varies according to deep and lobar location of supratentorial ICH²³; ICH involving the posterior limb of internal capsule or thalamus tends to have poorer outcomes compared with other locations.^24–26

Recommendations

• Future ICH trials should consider stratifying randomization based on ICH location, for example deep thalamo-capsular versus deep non thalamo-capsular versus lobar locations. This requires a standardized process and training to minimize variability and maximize reliability of determining hemorrhage location.

Hematoma Volume

Large ICH volume is a consistent predictor of poor outcome. As a result, most clinical trials use a cutoff threshold of ICH volume or low score on the Glasgow Coma Scale as a surrogate measure of ICH severity/volume to exclude poor prognosis patients. Epidemiological and trial data suggest that patients with ICH volume <30 mL have good potential for survival and recovery.²⁷ An argument can be made to exclude patients with ICH >30 mL from trials of medical therapies. Trials targeting hematoma growth may require even lower initial ICH volume because expansion may cause further damage. Situations unlikely to result in any clinically meaningful difference, such as a very large ICH, should be excluded from medical trials. Conversely, these patients are good candidates for inclusion in surgical trials because of the greater potential impact on outcome from hematoma evacuation in large ICH. Assessment of ICH volume should be incorporated into routine initial assessments of patients presenting with ICH.

Baseline Comorbidities

Various clinical conditions can influence outcome after ICH.²⁸ However, these variables are not systematically stratified at the time of randomization or later adjusted for when assessing outcomes in most ICH trials. Recent trials have employed adaptive randomization/minimization to achieve balance of 3 to 5 baseline factors.^3,4 The impact of key selected or range of comorbidities, such as diabetes mellitus, renal failure, chronic small vessel disease with attendant white matter abnormality, is under-evaluated and deserves further study.

The use of validated tools which combine comorbidities, such as the Charlson comorbidity index,^28,29 offers another approach to assess the influence of comorbidities on the treatment effect and outcomes in ICH trials.

Trial Design, Logistics, and Challenges

Clinical trials in ICH pose significant challenges related to difficulty with recruitment and ability to obtain timely informed consent. While the use of selective and stringent eligibility criteria to target patients who are most likely to benefit from the study intervention could potentially increase the chances of identifying a beneficial signal, it imposes significant screening burden to identify (and not miss) the infrequent eligible patient. Despite the soundness of this strategy, a decreased pool of eligible patients with ICH places increased demands on recruitment and requires a wide recruitment network. It has been estimated that a large proportion of patients with ICH are not eligible for any trial.³⁰ Potential solutions for improving recruitment into ICH trials are outlined below.

Co-Enrollment

Allowing co-enrollment in some ICH trials could increase recruitment and allow quicker completion of the trials. However, there are potential disadvantages. There is a potential for increased burden on participants due to additional consent and multiple follow-up assessments. There are also logistical and operational challenges related to complexities of statistical analyses, harmonization of administrative tasks, assessment of interactions between the interventions, and adjudicating relatedness of adverse events to a particular intervention. These drawbacks can be remedied by preplanned discussion among the leading investigators to synchronize trials activities and design and considering factorial designs.

Investigators should consider co-enrollment particularly to academic studies using a single or joint Data Safety and Monitoring Board across studies, and closely monitoring co-enrollment and its effect on recruitment, adherence to study protocols, and drop-out.

Formation of an ICH International Trials Consortium

Large trials mitigate many of the concerns regarding differences in baseline severity and comorbidities; they also increase precision of estimation. This approach has been successful in producing successful treatments in heart disease and cancer. Collaboration between multiple countries is required to complete any such medium-to-large size phase II/III randomized, controlled, trials in ICH. Establishment of an ICH International Trials Consortium could facilitate planning, coordination, and cooperation of such trials; increase generalizability of the results; allow sharing infrastructure for co-enrollment; and would generate enthusiasm to share and optimize research ideas. However, this will require increased awareness among funders of the need for international cooperation; skillful and difficult coordination between various funding bodies from different countries; goodwill, academic credits, and trust between the oversight committee and participating investigators; and streamlining standards-of-care across participating institutions to minimize the impact of variability of care on data quality. Advanced planning, cooperation with various networks such as the Global Alliance of Independent Networks in Stroke led by the National Institute of Neurological Disorders and Stroke StrokeNet, the European Stroke Organization Trials Alliance and the Canadian Stroke Trials Consortium, as well as industrial cooperation and co-funding, would all assist in success of an ICH International Trials Consortium.

Establishment of an ICH international registry or platform could also allow for assessing nonrandomized patients and data linkage for comprehensive cohorts data. Hardships related to funding and administrative oversight may be partially solved by building registries into trial designs and industrial support.

Simplify Trials to Align With Existing Practice, Consent, and Procedures

The variability in clinical practice and beliefs, complicated nature of the informed consent process and various regulatory requirements, and disabling neurological deficits in patients with ICH have long been recognized by investigators as hindrances to successful and timely recruitment. Communication, cognitive, and physical impairments among patients with ICH clearly lend to the requirement for simplification of the informed consent form and process in ICH trials to allow rapid recruitment for the testing of time-dependent treatments. Surrogate consent is not always possible in a timely manner. Community consent has been proposed for ICH trials but is rarely implemented. There is a need for a concise, easy-to-understand, stroke-specific, consent form to facilitate recruitment of patients with ICH into clinical trials while protecting their welfare and rights. Policies are needed to facilitate the use of digital technology, such as videotelephony and telemedicine to obtain consent.

Simplifying trial procedures, in particular, follow-up assessments, can significantly improve recruitment and retention. Trial planning must take into consideration local variations in practice, obtain consensus from investigators on common definitions and essential assessments, and align the study protocol with usual processes and standards-of-care at various participating sites. Follow-up procedures should take into consideration the unique nature of patients with ICH, their disabling deficits, and long rehabilitation. The use of voice-over-internet technology and videotelephony is encouraged as an alternative to in-person follow-up assessments after discharge from the hospital.

Selecting a translational paradigm for ICH with high potential for success is also important. One paradigm is to perform a mega-trial encompassing a wide range of patients, then identify which subgroups benefit in adequately powered statistical tests on interaction. Experience from AIS thrombolysis and endovascular trials offers an alternative paradigm; to target highly selected patient population to increase the chances of a clear signal of benefit and proof-of-concept then perform additional trials in expanded populations once an intervention is proved effective in carefully targeted population to potentially broaden its indication.

Mortality as a Primary Efficacy End Point in ICH Trials

The HEADS-2 participants considered whether early mortality should be used as the primary efficacy end point as opposed to a safety outcome for clinical trials in ICH. The question of death versus survival is the first question that most families ask caregivers. Arguments for using mortality as a primary efficacy end point are: survival is clinically meaningful and can be the primary early goal of medical intervention; it is an objective measure as opposed to the subjective nature of other measures such as health-related quality of life or functional outcome assessments; and it could increase statistical power of the study. Mortality shift if achieved in one or more trials has the potential to standardize approaches to rendering care for patients with ICH. Mortality reduction is more readily and more frequently achieved than improved functional outcomes. Although there were no significant differences in functional outcomes between treated patients and control groups in the Tranexamic acid for hyperacute primary ICH (TICH-2),⁵ the thrombolytic removal of intraventricular hemorrhage in treatment of severe stroke (CLEAR-III),²⁵ and the efficacy and safety of MISTIE-III trial (minimally invasive surgery with thrombolysis in ICH evacuation,³ treated patients had lower rates of early deaths. Early mortality is readily accepted as an end point in traumatic brain injury and hemorrhagic shock trials, and an earlier primary mortality end point may be appropriate in severely affected patients with ICH where the peak functional recovery may take several months; survival being essential for recovery. Conversely, improved survival should not be at the expense of persistent and severe disability, which would increase burden on the patients, families, society, and the overall health system. Importantly, survival may not be an acceptable outcome to some patients, families, and physicians.

Engagement with patient and consumer groups representative of the general population and inclusive of different cultural backgrounds is required to gain better information over the appropriateness of using reduction in early mortality as the primary measure of efficacy in future trials.

Conclusions

There are many challenges and opportunities to optimize clinical research and randomized trials in ICH. We hope these recommendations from HEADS-2 will facilitate improved rigor, efficiency, and success of future ICH trials to find effective therapies for this devastating condition (Table).

Acknowledgments

We thank Kenes Group for their help in organizing the Hemorrhagic Stroke Academia Industry (HEADS)-2 roundtable meeting.

Footnotes