Eligibility criteria

1. Types of studies

   RCTs with or without blinding will be included in this study. Non-RCTs, studies with the cross-over design, and uncontrolled clinical trials will be excluded.

2. Types of participants

   Adult patients with acute ICH (within seven days) confirmed by CT or MRI according to a guideline for healthcare professionals from the American Heart Association/American Stroke Association9 will be included. There will be no restrictions in terms of the patients’ age, gender, race, education or economic status. Patients with traumatic haemorrhagic stroke, primary intraventricular haemorrhage and subarachnoid haemorrhage will be excluded.

3. Types of intervention

   We will mainly focus on the intervention that edaravone was compared with the placebo or no treatment. Additionally, trials wherein routine treatments or cointerventions with edaravone were administered equally to all groups will also be included. However, there will be no restriction on the course of treatment.

4. Types of outcome measures

   As ICH is a life-threatening condition with a high rate of disability, we will pay more attention to mortality and the long-term functional status in this systematic review. Clinical studies that reported numerical data on one or more of the following outcomes will be considered.

1. Primary outcomes

   All-cause mortality and dependency at the end of the follow-up will be set as the primary outcomes. The functional status was assessed using clinical scales including the modified Rankin Scale (mRS), Glasgow Outcome Scale (GOS) and Barthel Index (BI). The dependency will be defined as mRS grades 3–6, GOS grades 1–3, or BI less than or equal to 60.

2. Secondary outcomes

   The secondary outcomes will include: (1) improvement of neurological impairment assessed using clinical scales including the National Institute of Health Stroke Scale, Canadian Neurological Scale, European Stroke Scale, Scandinavian Stroke Scale, Modified Edinburgh-Scandinavian Stroke Scale and other related scales, (2) the total efficiency rate including cure rate, obvious effective rate and effective rate, and (3) reduction in the haematoma volume.

3. Safety outcome

   Adverse effects of edaravone including impairment of kidney and liver function, skin irritation, nausea, will be evaluated.

Search strategy

We will conduct the comprehensive electronic searches of Medline, Embase and Cochrane Central Register of Controlled Trials. Given that edaravone is widely used in China, we will search the following Chinese databases: China National Knowledge Infrastructure, Chinese scientific periodical database of VIP INFORMATION, Wanfang Data and SinoMed from their respective inception dates to 23 April 2020. In addition, we will also search for clinical trial registers, dissertations and grey literature.35 We will develop the search strategy for Medline (see online supplementary appendix 1. Search Strategy Example) and the equivalent search words will be used in other databases as well. The registers which mainly include ongoing or unpublished trials are the following:

• WHO International Clinical Trials Registry Platform.

• ClinicalTrials.gov.

• The United Kingdoms’ ISRCTN registry (ISRCTN).

• Chinese Clinical Trial Registry.

• Australia and New Zealand Clinical Trials Registry.

• The Netherlands Trial Register.

• German Clinical Trials Register.

• Japan Primary Registries Network.

• Clinical Trials Registry—India.

• Iranian Registry of Clinical Trials.

• Sri Lanka Clinical Trials Registry.

Our research will be restricted to humans and clinical trials, with no language restrictions.

Screening and selection

Duplicate articles will be removed after identifying them by database searching. Two review authors (LF, TL) will independently screen the articles for titles and abstracts according to the inclusion criteria. The full text will be reviewed if necessary. In addition to this, the list of related studies from the references will be examined further to identify other potential studies to be included. The reviewers will exclude reports that are irrelevant to our research and retrieve full-text articles for the remaining references. The same two reviewers will independently screen these full-text articles to identify studies for inclusion, as well as determine and record reasons for the exclusion of ineligible studies. They will resolve any disagreements through discussion or, if required, they will consult a third review author (YG) to arbitrate when disagreements are not resolved. The excluded studies will be listed in a table with the proper reasons. The whole process of study screening and selection is shown in figure 1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Flow diagram of the study selection process. ICH, intracerebral haemorrhage.

Data extraction

Two review authors (QY, PJ) will independently extract data on methods, patients, interventions, outcomes and results from the included studies, using a preformulated data collection form. We will try to contact the corresponding authors for any missing data or clarification on unclear information.35

Quality assessment

The methodological quality assessment of the eligible studies will be independently conducted by two reviewers (LF, NL) for each study using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions.35 Disagreements will be resolved by discussion or by involving another review author (YG). The risk of bias will be assessed according to the following domains: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting and other bias. The risk of bias for each domain will be classified into three levels: low risk, high risk and unclear. Unclear items in studies will be inquired by contacting the corresponding authors for details. We will provide information from the study report together with a justification for our judgement, in the ‘risk of bias’ tables.

Data synthesis and management

1. Measures of treatment effect

   We plan to summarise the data using risk ratio calculations and 95% CIs for dichotomous outcomes and mean differences with 95% CIs for continuous outcomes in the final analysis. We will calculate standardised mean differences with 95% CIs when different scales were used to measure the same continuous outcome variable.

2. Dealing with missing data

   The corresponding author will be contacted by reviewers via email or telephone to obtain the missing data or information which was not clearly described. In case the missing data are unavailable, intention-to-treat and sensitivity analyses will be performed to address the potential impact of the missing data,36 37 which will then be discussed if necessary.

3. Assessment of heterogeneity and data synthesis

   Statistical heterogeneity among the included trials will be evaluated using the I² test. A meta-analysis will be conducted if there is no significant clinical (relating to the participants, interventions, controls and outcomes) and statistical heterogeneity (I² values are less than 75%) between the included trials. However, if the I² value is less than 25%, we will use a fixed-effect model to synthesise the data, and if it is between 25% and 75%, we will estimate the sources of the heterogeneity. If the statistical heterogeneity is explained successfully by sensitive analysis or subgroup analysis (ie, I² is less than 25%), we will also use the fixed-effect model to synthesise the data. Otherwise, a random-effect model will be applied. Data will not be synthesised if there is a significant level of statistical heterogeneity among the trials (ie, I² is greater than 75%) that cannot be explained or handled by subgroup analysis. All statistical analyses will be performed using Review Manager V.5.3 (The Cochrane Collaboration) software.

4. Analysis of subgroups or subsets

   If the data are available for the subgroup analyses, we will plan to compare:

   • Effects in patients with various dose of edaravone (less than 60 mg/day, 60 mg/day and more than 60 mg/day);

   • Effects in patients with various course of treatment (less than 14 days, 14 days and more than 14 days);

   • Effects in patients with various drug combinations (edaravone plus nimodipine, and edaravone plus other neuroprotective agents);

   • Effects in patients with various types of ICH based on Structural lesion, Medication, Amyloid angiopathy, Systemic/other disease, Hypertension, Undetermined aetiological classification;

   • Effects in patients with various course of disease (within 24 hours and after 24 hours from stroke onset);

   • Effects in patients with various haemorrhage sites (brain stem, cerebellum, basal ganglia region and other sites).

5. Assessment of reporting biases

   A funnel plot will be generated to explore the possibility of publication bias if 10 or more trials are included per comparison.

Confidence in cumulative evidence

The strength of the body of evidence in this review will be categorised as high, moderate, low or very low according to the Grading of Recommendations Assessment, Development and Evaluation system38 using the GRADEpro software.

Patient and public involvement

Not applicable. This protocol of systematic review and meta-analysis does not directly involve patients and the general public. Data will be collected from published articles retrieved from main databases and manual search.