Scolaris Content Display Scolaris Content Display

Cochrane Database of Systematic Reviews Protocol - Intervention

Intravenous immunoglobulin for the treatment of childhood encephalitis

This is not the most recent version

Collapse all Expand all

Abstract

This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:

To assess the safety and efficacy of intravenous immunoglobulin as add‐on treatment for children with encephalitis.

Background

Description of the condition

Encephalitis is a syndrome of neurological dysfunction that results from inflammation of the brain parenchyma. The worldwide annual incidence ranges from 3.5 to 7.4 per 100 000, rising to 16 per 100,000 in children, with the highest incidence in infants under one year of age (Thompson 2012). Encephalitis could result either from an infection of the brain (infectious encephalitis) and/or from autoantibodies that affect the brain (immune‐mediated encephalitis) (Zuliani 2012).

Infections have been considered to be the major cause of encephalitis and more than 100 different causative pathogens have been recognised. Viruses are the most common pathogens known to cause encephalitis. However, a host of other pathogens including bacteria and protozoa have also been implicated.

Immune mediated disorders such as acute disseminated encephalomyelitis are now recognised to contribute to a significant proportion of cases where no infective cause is identified (Granerod 2010). More recently, several well characterised immunological syndromes that are mediated by antibodies against central nervous system surface proteins such as N‐methyl‐D‐aspartate receptor and the voltage‐gated potassium channel‐complex and its associated proteins, have been identified in patients with encephalitis (Hacohen 2013; Hacohen 2014), and these account for 4 % and 7% of overall cases (Granerod 2010). However, despite routine investigations, no aetiology is found in up to 60% of cases of encephalitis (Davison 2003; Granerod 2010).

There is a high rate of mortality and morbidity from encephalitis, despite current standard of care. An exaggerated host immune response has been implicated in the pathogenesis of encephalitis and this has been shown to play a part in the disease pathogenesis (Cervantes‐Barragán 2012; de Aquino 2013; Lundberg 2008; Ramakrishna 2013).

While a mortality rate of up to 20% has been reported (Granerod 2010; Ramakrishna 2013), incomplete recovery after childhood encephalitis is common with persisting symptoms in up to 60% of affected children (Aygun 2001; Fowler 2010). A 12‐year prospective study of children with herpes simplex virus encephalitis demonstrated that neurological sequelae occurred in 63% of cases, including seizures in 44% and developmental delay in 25% (Fowler 2008). Long term complications such as severe physical impairment, behavioural, psychosocial and educational difficulties have also been reported (Dowell 2000).

Encephalitis also imposes a substantial economic and health care resource burden. Health, social and economic costs are also extended where families are left bereaved or with a child who has sustained disability. A 12 year American review reports an encephalitis‐associated hospitalisation rate of 6.9 per 100,000 persons. The death rate in the same study was 5.8% of all hospitalisations (Vora 2014). A four‐year review of paediatric intensive care unit United Kingdom national data showed a total of 353 admissions due to encephalitis, with an average length of stay of 4.3 days. A total of 85% of admitted children required ventilation, and some additionally required cardiovascular support (19%) and renal dialysis (6%) (unpublished observations). An American study reports approximately 19,000 hospitalisations (7.3 hospitalisations per 100,000 population) and 230,000 hospital days from encephalitis over a 10‐year period, with an estimated cost from encephalitis associated hospitalisations of USD$28,000 leading to an annual national cost of USD$650 million (Khetsuriani 2002).

Given this huge burden from encephalitis despite current standard of care, there is the need to identify other adjunctive treatment option that could be used in the management of children with encephalitis.

Irrespective of aetiology, the underlying pathogenesis in encephalitis is brain inflammation. The degree of brain inflammation seen in some forms of encephalitis has been shown to correlate with clinical outcomes (Ramakrishna 2013). Attenuation of such inflammation either as a direct effect or through modulation of the immune response may therefore be key to improving clinical outcomes. Biologic agents that possess both anti‐inflammatory and /or immunomodulatory properties may therefore be useful as adjunctive therapies and could improve outcomes (Ramakrishna 2013; Rozenberg 2013). Intravenous immunoglobulin is one such biological agents.

Description of the intervention

This review aims to assess the role of add‐on intravenous immunoglobulin treatment to routinely used treatment (aciclovir) for children with encephalitis. Although not routine treatment, corticosteroid therapy will also be considered in this review because it is often used in the treatment of encephalitis. The use of experimental therapies will not be considered.

Intravenous immunoglobulin

Intravenous immunoglobulin is a blood product, made from pooled collections of human plasma collected from thousands of blood donors. It comes as a ready‐to‐use liquid formulation for intravenous administration. It has a half life of three to four weeks. Intravenous immunoglobulin is being used increasingly in the management of a wide range of neurological conditions and its efficacy has been established in a few of these (Hughes 2009). Licensed indications include as replacement therapy for patients with primary and secondary antibody deficiency states, Kawasaki disease, haematological conditions (idiopathic immune thrombocytopaenic purpura, B‐cell chronic lymphocytic leukaemia) and neurological conditions (myasthenia gravis, multifocal motor neuropathy, Guillain‐Barré syndrome and chronic demyelinating polyneuropathy) (FDA 2013). Intravenous immunoglobulin is sometimes also used off‐label for the treatment of children with encephalitis.

The dose of Intravenous immunoglobulin for each indication varies. For primary and secondary antibody deficiency states the starting dose is between 0.4 ‐ 0.6 g/kg of body weight and needs to be adjusted based on clinical outcome. For neurological diseases two doses of 2 g/kg of body weight over a five days and given six weeks apart. For haematological conditions, a dose of 0.8 ‐ 1 g/kg is used.

Intravenous immunoglobulin treatment is safe, however adverse events such as chills, headache, fever, vomiting, allergic reaction, nausea, arthralgia, low blood pressure and low back pain may occur. Rarely, sudden fall in blood pressure, anaphylactic shock and thromboembolic reactions could occur. Cases of reversible aseptic meningitis and isolated cases of haemolytic anaemia have been observed as well as increase in serum creatinine level and /or acute renal failure. Since IVIG is a blood product, there is the risk of transmission of infectious agents such as the human immunodeficiency virus (HIV) and viral hepatitis by contaminated products (Looney 2006)

Aciclovir

Aciclovir is widely used in the treatment of herpesvirus infections, particularly herpes simplex and varicella‐zoster virus. It is a synthetic nucleoside analogue with in vitro and in vivo inhibitory activity against herpes simplex virus types 1 and 2 as well as varicella‐zoster virus (Elion 1983). Aciclovir is available in either oral, topical and intravenous forms. It is poorly water soluble and has poor oral bioavailability hence intravenous administration is necessary if high concentrations are required, such as in serious infections.The intravenous form is used in the treatment of herpes simplex and varicella‐zoster virus encephalitis and the dose is based on weight (kilograms) and varies with age:

  • neonate (20 mg/kg);

  • child 1‐3 months (10‐20 mg/kg);

  • child 3 months‐12 years (500 mg/m2 8 hourly);

  • child 12‐18 years (10 mg/kg 8 hourly).

Duration of treatment for encephalitis is 14‐21 days (British National Formulary Publications).

Side effects such as nausea, vomiting, diarrhoea, abdominal pain, hepatitis, and renal dysfunction have been reported with aciclovir use. Interractions have also been reported with the ciclosporin (increased risk of toxicity), mycophenolate (increased plasma concentration of aciclovir and inactive metabolite of mycophenolate), probenecid (reduced excretion of aciclovir and increased plasma concentration), tacrolimus (increased risk of nephrotoxicity), theophylline (increase plasma concentration of theophylline) (British National Formulary Publications).

Corticosteroids

The role of corticosteroid in the treatment of encephalitis is not yet established however they have been often used, especially for patients with herpes simplex virus encephalitis with marked cerebral oedema, brain shift or raised intracranial pressure. Information from experimental animal research (Sellner 2005; Thompson 2000;) and from clinical observations (Kamei 2005; Lizarraga 2013; Ramos‐Estebanez 2014) indicate a substantial benefit in outcomes for patients with HSVE treated with adjuvant dexamethasone. Corticosteroids are potent anti‐inflammatory agents. Their clinical use along with antiviral therapy has been advocated in patients with herpes simplex encephalitis and cerebral edema where they reduce brain swelling. Results of a prospective randomised controlled trial on the use of adjunctive corticosteroid therapy in herpes simplex virus encephalitis are awaited (Martinez‐Torres 2008).

How the intervention might work

Intravenous immunoglobulin

Intravenous immunoglobulin has multiple actions which may operate in concert with each other. For a particular disease, there may be one predominant mechanism of action depending on the underlying disease pathogenesis.The most relevant actions of intravenous immunoglobulin include the following:

Additional actions include the effect of intravenous immunoglobulin on superantigens and enhancement of remyelination (Dalakas 1998). Antiviral functions of intravenous immunoglobulin and its potential to inhibit viral infection has been demonstrated in vitro (Frenzel 2012; Kishimoto 2004; Krause 2002).

Aciclovir

The inhibitory activity of acyclovir is highly selective due to its affinity for the enzyme thymidine kinase encoded by herpes simplex and varicella‐zoster virus. This viral enzyme converts aciclovir into aciclovir monophosphate, a nucleotide analogue. The monophosphate is further converted into diphosphate by cellular guanylate kinase and into triphosphate by a number of cellular enzymes. In vitro, aciclovir triphosphate stops replication of herpes viral deoxyribonucleic acid (DNA). This is accomplished in three ways:

  1. competitive inhibition of viral DNA polymerase;

  2. incorporation into and termination of the growing viral DNA chain; and

  3. inactivation of the viral DNA polymerase (Elion 1983; Elion 1993; Kerpel‐Fronius 1983).

Corticosteroids

Corticosteroids are powerful endogenous immunosuppressors, especially for the innate immune response and the subsequent inflammatory reaction (Esposito 2012; McKay 1999). Experimental data indicate that they may attenuate central nervous system damage by reducing cytokine and prostaglandin production, and limiting the nitric oxide concentration induced by the increased expression of immunological nitric oxide synthase (Meyding‐Lamadé 2002; McKay 1999). Dexamethasone is able to repress lipopolysaccharide‐induced nuclear factor kappa‐light‐chain‐enhancer of activated B cells (NF‐κB) activation in the brain (Glezer 2003); cortisol can abolish stimulated interleukin (IL)‐1β and tumour necrosis factor(TNF)‐α gene expression in microglial cells. Ultimately, the effects of corticosteroids lead to a decrease in pro‐inflammatory signal transduction pathways and gene expression, which is an essential endogenic mechanism to avoid exaggerated responses during immunogenic challenges (Sergerie 2007).

Why it is important to do this review

There remains significant mortality and morbidity from childhood encephalitis despite standard treatment. Effective immunomodulatory strategies are required to reduce mortality and morbidity in children with encephalitis. Several previous reports point to a possible beneficial effect of intravenous immunoglobulin in different forms of encephalitis (Caramello 2006; Hacohen 2013; Titulaer 2013; Wang 2006). In a recent paediatric cohort study, intravenous immunoglobulin was used in only 35% (17/48), presenting with probable autoimmune encephalopathy (Dalakas 1998; Hacohen 2013); whilst in a recent prospective surveillance study only 15% (6/40) of children with acute disseminated encephalomyelitis received intravenous immunoglobulin (Absoud 2013; Marchioni 2013).

There however remains a lack of consensus amongst clinicians on the use of intravenous immunoglobulin in childhood encephalitis and practice varies widely. This review therefore aims to provide a detailed analysis of existing data on the use of intravenous immunoglobulin in the treatment of children with encephalitis and may provide additional supportive evidence to inform on its routine use in clinical practice.

Objectives

To assess the safety and efficacy of intravenous immunoglobulin as add‐on treatment for children with encephalitis.

Methods

Criteria for considering studies for this review

Types of studies

We will include only randomised controlled clinical trials (RCTs). In the absence of RCTs, we will provide a narrative summary of large (n ≥ 10 patients included) non‐randomised studies including controlled clinical trials, case‐control and cohort studies.

Types of participants

Inclusion criteria

Children with a clinical diagnosis of acute (symptoms present within 24 hours prior to hospitalisation) or sub‐acute (symptoms present between 24 hours and 4 weeks prior to hospitalisation) encephalitis who are aged 6 weeks to 17 years (before 18th birthday) at the time of diagnosis.

Since there is the likelihood of a variation in the diagnosis of encephalitis, where the eligibility of participants in an identified study (in terms of the diagnosis of encephalitis) is in doubt, we will apply a pre‐defined diagnostic criteria adapted from the Consensus Statement of the International Encephalitis Consortium in order to ascertain eligibility (Venkatesan 2013). A diagnosis of encephalitis will be agreed if the following features are present:

  • Altered mental state (reduced or altered conscious level, irritability, altered personality or behaviour, lethargy), and any two of the following:

    • brain imaging evidence consistent with encephalitis or immune‐mediated encephalopathy that appears acute in onset;

    • cerebrospinal fluid pleocytosis: cerebrospinal fluid white blood count of ≥ 5 cells/mm3;

    • presence of autoantibodies such as N‐methyl D‐aspartate receptor antibodies, and voltage‐gated potassium channel antibodies, in cerebrospinal fluid and/or blood;

    • generalised or partial seizures not fully attributable to a pre‐existing seizure disorder;

    • new onset focal neurological signs (including movement disorders);

    • abnormality on electroencephalogram (EEG) or cerebral function analysis monitor that is consistent with encephalitis and not attributable to another cause;

    • fever > 38oC within 72 hours before or after presentation to hospital.

Exclusion criteria

  • Children with chronic encephalitis i.e. where presenting symptoms have lasted longer than four weeks

  • Children with known hypersensitivity to intravenous immunoglobulin

Types of interventions

Experimental studies

  • Intravenous immunoglobulin plus standard care (experimental group) versus standard care alone (control group)

  • Intravenous immunoglobulin plus standard care (experimental group) versus standard care plus placebo (control group)

    • We define standard care as any of the following used either alone or in combination: intravenous aciclovir, intravenous corticosteroid therapy

Cohort and case control studies

  • Any treatment regimen that includes intravenous immunoglobulin in addition to standard treatment versus any regimen without intravenous immunoglobulin

For both RCTs and non‐RCTs, corticosteroid therapy will comprise any type of corticosteroid such as methylprednisolone, prednisone, prednisolone and dexamethasone.

We will assess all interventions independent of dosage, time interval between onset of symptoms and randomisation and duration of treatment.

Types of outcome measures

Primary outcomes

  • Proportion of participants with significant disability (='poor outcome') at six months after treatment

    • Significant disability using any validated disability assessment scale such as (but not limited to) the Glasgow Outcome Scale Extended (score of 2 or less) (Beers 2012), Liverpool Outcome Score (score of 2 or less) (Lewthwaite 2010), the modified Rankin Scale (score of 4 or more) (van Swieten 1988)

  • Proportion participants with at least one serious adverse event as defined in the trial

    • Where the definition of a serious adverse event is not clearly defined, we will use the definition from the International Conference on Harmonisation (ICH) Harmonised Tripartite Guideline (ICH 1994), which defines a serious adverse event as any adverse event that results in any of the following outcomes: life‐threatening, death, requires prolongation of hospital stay, persistent or significant disability/incapacity, congenital abnormality

Secondary outcomes
Short term (during hospital admission)

  • Length of hospital stay (either as continuous or categorised data, as reported in the trial)

  • Proportion of participants requiring invasive ventilation and duration (either as continuous or categorised data, as reported in the trial) of ventilation for ventilated children.

Long term (at six months post discharge from hospital)

  • Any cognitive impairment as defined in the trial (binary: Yes/No), major cognitive disability as defined by in the trial (binary: Yes/No) and mean cognitive scores using any validated age‐appropriate psychometric instrument

  • Poor adaptive functioning as defined in the trial (binary: Yes/No) or adaptive behaviour scores less than two standard deviations from the mean, using any validated age‐appropriate scale

  • Quality of life assessment scores obtained using any validated age‐appropriate tool

    • Pediatric Quality of Life Inventory (PedsQL) (Varni 2001)

    • Health‐related quality of life (Matza 2004)

  • Number of seizures per participant and the proportion of participants with new diagnosis of epilepsy

Search methods for identification of studies

Electronic searches

The Trials Search Co‐ordinator of the Cochrane Multiple Sclerosis and Rare Diseases of the Central Nervous System Group will run the initial searches for all prospectively registered and ongoing trial sin the following databases:

  • Cochrane Central Register of Controlled Trials (CENTRAL);

  • MEDLINE;

  • EMBASE;

  • Cumulative Index to Nursing and Allied Health Literature (CINAHL);

  • Latin American and Caribbean Health Science Information Database (LILACS);

  • ClinicalTrials.gov (http://clinicaltrials.gov);

  • the World Health Organization (WHO) International Clinical Trials Registry Portal (ICTRP) search portal (http://apps.who.int/trialsearch/).

In addition, two authors (MI, NGM) will also search the Global Health Library (Virtual Health Library) (http://www.globalhealthlibrary.net/php/index.php) and Science Citation Index Expanded (SCI‐EXPANDED) & Conference Proceedings Citation Index‐Science (CPCI‐S) (Web of Science) (1945 to date) (Appendix 1).

We will use a combination of free‐text and Medical Subject Headings (MeSH) terms to describe the population and intervention as illustrated in Appendix 1. We will apply the search filter 'Cochrane Highly Sensitive Search Strategy for identifying randomised trials in MEDLINE: sensitivity‐ and precision‐maximizing version (2008 revision); PubMed format' to MEDLINE and adaptations of it to the other databases except CENTRAL (Higgins 2011). We will apply no language or date limits to our search. We will identify non‐RCTs using the same search strategy, without the RCT/controlled clinical trial filters.

Searching other resources

We will screen reference lists of included studies and related reviews for additional trials. In addition, we will contact experts and individual researchers working in this field to enquire about any unpublished or missing data .

Data collection and analysis

Selection of studies

Two authors (MI, NGM) will independently assess the eligibility of studies to be included in this review using an eligibility form. We will perform screening at two levels. Level 1 screening will entail a broad screen of title and/or abstracts, as available. In all instances where the title and/or the abstract suggest that the report relates to a trial reporting on the use of intravenous immunoglobulin in the treatment of encephalitis in a child or group of children, we will retrieve the full‐text article for further assessment. Level 2 screening will involve a comprehensive assessment of the full text of retrieved articles to ascertain eligibility. Where eligibility is unclear, we will contact the trial authors to clarify this. We will compare multiple reports of the same study, and we will use the most comprehensive report. We will then link multiple publications together as companion reports, but true duplicates will be excluded. We will document the exclusion of any study from the review, and will describe the reasons for exclusion in the characteristics of excluded studies table. We will resolve any disagreements by discussion between the two review authors in the first instance, and if unresolved by consulting the other co‐authors (MA, AJP).

Data extraction and management

Two review authors (MI and NGM) will extract the data from studies independently using a standardised data extraction sheet, accompanied by data extraction guidelines that will be developed after piloting the data extraction form. The data extraction form will include information on outcomes as listed in the primary and secondary objectives of this review as well as data related to the study characteristics, participant characteristics, interventions and comparators, follow up details, declaration of interest for the primary investigators and also source of funding for each study. Where the relevant data is unclear or missing, we will contact the publication author to clarify this or provide the missing data. The two authors will correct any errors in data extraction after discussion. We will then link multiple publications together as companion reports, but true duplicates will be excluded. We will resolve any discrepancies in data extraction between the two authors in the first instance and if unresolved, by the involvement of the other co‐authors (MA, AJP). We will manage and analyse all data using the Cochrane Collaboration statistical software, Review Manager 2014.

Assessment of risk of bias in included studies

Experimental studies (RCTs)

Two independent authors (MI and NGM) will assess the risk of bias in the included trails. We will use the Cochrane Collaboration's tool for assessing risk of bias to evaluate the internal validity of the design and conduct of included studies (Higgins 2011). This tool describe a number of domains to be judged on the adequacy of each study.

  • Selection bias:

    • sequence generation;

    • allocation concealment.

  • Performance bias:

    • blinding of participants;

    • blinding of personnel.

  • Detection bias:

    • blinding of outcome assessors (for each outcome separately).

  • Attrition bias:

    • incomplete outcome data (for each outcome separately).

  • Reporting bias:

    • selective outcome reporting.

To classify the methodological quality of studies, we will consider bias for selection, detection and attrition. We will judge a high risk of attrition bias based on either the reasons for missing data or the difference between the percentage of participants missing in the control and treatment groups.

We will use the summary quality assessment at the analysis stage as a means of interpretation of the results. For each domain and for the summary assessment, we will assign the risk of bias categories as (Higgins 2011):

  • low risk of bias, plausible bias unlikely to seriously alter the results;

  • unclear risk of bias, plausible bias that raises some doubt about the results;

  • high risk of bias, plausible bias that seriously weakens confidence in the results.

We will rate a study as 'low risk of bias' when none of the three domains is affected, 'high risk of bias' when at least one of the three domains is affected and 'unclear risk bias in all other cases. We will consider the risk of bias in included studies when we interpret the review's results. If the study protocol is in the public domain, we will compare the outcomes listed in the protocol with the outcomes reported in the published manuscript. If the protocol is not publicly accessible, we will contact the authors and ask for a copy of the protocol. When there is suspicion of or direct evidence for selective outcome reporting, we will contact the study authors for additional information. We will resolve any disagreements by discussion between the two authors (MI and NGM) the first instance, and if unresolved by involvement of the other co‐authors (MA and AJP).

Controlled clinical trials, case‐control and cohort studies

We will evaluate the quality of case‐control and cohort studies (prospective and retrospective) using the appropriate Newcastle‐Ottawa Scales as follows (Appendix 2; Appendix 3).

  • Case control studies:

    • selection (case and control);

    • comparability;

    • assessment of exposure.

  • Cohort studies:

    • selection (exposed and non‐exposed) cohort;

    • comparability;

    • assessment of outcome.

To assess the methodological quality of non‐RCTs, we will consider the risk of bias in all three domains listed above. We will rate a study as 'high risk of bias' if a least one of the three domains are affected and low risk of bias where none of the domains is affected and unclear bias in all other cases. We will assess the quality of non‐randomised studies in relation to the presence of potential confounders that could make interpretation of the results difficult. We will consider a number of potential confounding factors: age, sex, type and severity of encephalitis (as classified by the trialists).

Criteria for assessing quality of safety data for all studies

We will assess risk of bias for serious adverse events by considering specific factors that may have had a large influence on the adverse events data. We will evaluate methods for monitoring and detecting a serious adverse event for each study as below:

  1. did the researchers actively monitor for serious adverse events (low risk of bias), or did they simply provide spontaneous reporting of serious adverse events that arose (high risk of bias)?

  2. did the authors define serious adverse events according to an accepted international classification and report the number of serious adverse events?

Measures of treatment effect

We will carry out statistical analyses using Review Manager 2014. For dichotomous data, we will report on relative risk (RR) with 95% confidence intervals (CIs).For continuous data, we will use weighted mean difference (WMD). If statistically significant reduction in risk difference (RD) is found, we will calculate the number needed to treat to benefit (NNTB). If a statistically significant increase in the RD is found, we will calculate the number needed to treat to harm (NNTH). We will analyse mortality as a hazard ratio (HR). We will use Parmar's method (Parmar 1998) if HRs have not been explicitly presented in the study.

Unit of analysis issues

In all studies, the individual child will be used as the unit of analysis. We do not anticipate that trials of intravenous immunoglobulin treatment for children with encephalitis using non‐standard designs such as cross‐over trials and cluster‐randomised trials have been conducted.

Dealing with missing data

We will contact authors where possible and record responses in the event that we identify any missing data. If possible, we will perform an intention‐to‐treat analysis.

Assessment of heterogeneity

We will test the results from the studies for heterogeneity using the chi‐square test when possible (Higgins 2003). We will assess statistical heterogeneity by visually inspecting forest plots and using the chi‐squared test for heterogeneity (with a P value < 0.10 for significance) and the I2 statistic as a measure of inconsistency across studies. Observed value of I2 and its interpretation will be as follows:

  • 0% to 40%: might not be important;

  • 30% to 60%: may represent moderate heterogeneity;

  • 50% to 90%: may represent substantial heterogeneity;

  • 75% to 100%: considerable heterogeneity.

If substantial statistical heterogeneity is found to be present according to the chi‐squared test for heterogeneity or I2 statistic, we will perform meta‐analysis with a random‐effects model rather than a fixed‐effect model and we will perform sensitivity analyses investigating differences in patient differences or study characteristics as sources of heterogeneity.

Assessment of reporting biases

The process of reporting bias will be as described in the Assessment of risk of bias in included studies section'. We will assess studies for reporting bias by analysing inclusion and exclusion criteria and blinding of participants and observers. We will also analyse primary and secondary outcomes for any reporting biases. We will assess funnel plot asymmetry if more than 10 studies are found. Where there are fewer studies, the power of the test is too low. Reasons for asymmetry include publication bias, outcome reporting bias, language bias, citation bias, poor methodological design and heterogeneity. We will assess and summarise these for each trial.

Data synthesis

We will compute pooled estimates using a fixed‐effect model meta‐analysis where trials are judged to be sufficiently homogenous (I2 < 50%). In addition, we will use the random‐effects model where there is clinical or methodological heterogeneity sufficient to suggest that treatment effects might differ between trials (I2 > 50%). For the primary and secondary outcomes of interest, where there is significant heterogeneity in the measurement scales and/or defined threshold values in th included studies, we will perform a qualitative description of results.

Where trials report binary outcomes such as disability (i.e. 'poor outcome'), cognitive impairment or major cognitive disability using different measurement scales and /or threshold value definitions, we will perform only a qualitative description of results.

For studies with multiple intervention groups, we will perform separate meta‐analyses synthesising each of the different comparisons of interventions for each of the different outcomes. Where it is impossible to perform a meta‐analyses such as when there is significant clinical heterogeneity across the included studies, we will provide the reason for this, and provide a descriptive analysis of these studies in text and tables.

Subgroup analysis and investigation of heterogeneity

If sufficient studies are available, we will conduct subgroup analyses according to a number of the pre‐specified factors as it is probable that given the heterogeneity of the causes of encephalitis, clinical outcomes may depend on the type of encephalitis; in addition, there will be immunological differences between immunocompetent and immunocompromised patients which may impact on the magnitude of response to intravenous immunoglobulin. Furthermore, the degree of an observed effect may depend on the dose of intravenous immunoglobulin used.

The pre‐specified factors are:

  • encephalitis types (Infective/immune mediated/unknown aetiology);

  • co‐morbidities (patients with primary or secondary immunodeficiency versus immunocompetent patients).

  • dose of intravenous immunoglobulin (low‐dose intravenous immunoglobulin (less than 1g per Kg in weight) versus high dose (≥1g/Kg);

  • time of initiation of intravenous immunoglobulin treatment (early (within five days of hospital admission) versus late (beyond five days of hospital admission);

  • time of initiation of standard therapy (early (within 48 hours of hospital admission) versus late (beyond 48 hours days of hospital admission).

Sensitivity analysis

We will perform sensitivity analysis using a summary table and we will assess whether the effect of intravenous immunoglobulin varied with the exclusion of studies rated as high or unclear risk of bias. If the results observed remain unchanged, then we will consider the evidence to be robust.

Summary of findings table

We will create a 'Summary of findings' table using the following outcomes:

  • significant disability;

  • length of hospital stay;

  • need for invasive ventilation;

  • serious adverse events;

  • cognition;

  • adaptive skills;

  • quality of life; and

  • new onset epilepsy.

We will use the five (GRADE) considerations (study limitations, consistency of effect, imprecision, indirectness and publication bias) to assess the quality of a body of evidence as it relates to the studies which contribute data to the meta‐analyses for the prespecified outcomes. We will use methods and recommendations described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011), using the GRADEpro software (GRADEpro 2014). We will justify all decisions to down‐ or up‐grade the quality of studies using footnotes and we will make comments to aid reader's understanding of the review where necessary.