Background There are data suggesting that intravenous immunoglobulin treatment has some benefit for certain forms of encephalitis but robust evidence from large randomised controlled trials in children with all-cause encephalitis is lacking. Objective To evaluate whether intravenous immunoglobulin treatment improves neurological outcomes in childhood encephalitis when given early in the illness. Design Phase 3b, investigator-initiated, randomised, double-blind, placebo-controlled trial of intravenous immunoglobulin for the treatment of encephalitis in children. Setting Twenty-one NHS Hospitals in the UK. Participants Children aged 6 months to 16 years with a diagnosis of acute or sub-acute encephalitis. Intervention Two doses (1 g/kg/dose) of either intravenous immunoglobulin or matching placebo, given 24–36 hours apart, in addition to standard treatment. Main outcome measure Participants were followed up for 12 months (+/– 4 weeks) after randomisation. The primary outcome measure was a ‘good recovery’ defined as a score of ≤ 2 on the Paediatric Glasgow Outcome Score Extended at 12 months after randomisation. Secondary outcomes The secondary outcomes were clinical, neurological, neuroimaging and neuropsychological results, identification of the proportion of children with immune-mediated encephalitis, and intravenous immunoglobulin safety data. Results We planned to recruit 308 children over a 42-month period. After enrolment of 18 participants (8 male; 44%) over 21 months (from December 2015 to September 2017), funding was withdrawn due to slow recruitment and the study was terminated. Ten participants were randomised to the intravenous immunoglobulin group, and eight to the placebo group, and all 18 participants were included in the analysis. At 12 months after randomisation, 9 participants [50%; intravenous immunoglobulin n = 5 (50%), placebo n = 4 (50%)] made good recovery and 5 participants [28%; intravenous immunoglobulin n = 3 (30%), placebo n = 2 (25%)] made a poor recovery. Three participants in the placebo group (43%) experienced a total of 10 serious adverse events compared with none in the intravenous immunoglobulin group but none of the adverse events were judged to be related to the study treatment. No deaths occurred during the study period. Conclusion ImmunoglobuliN in the Treatment of Encephalitis (IgNiTE) was halted prematurely due to slow recruitment. Given the small sample size, the study was underpowered to evaluate the effect of intravenous immunoglobulin when compared with placebo in childhood encephalitis. The study findings, albeit from a small sample size, support existing evidence that encephalitis results in poor neurological outcomes for many children. Lessons learned from the ImmunoglobuliN in the Treatment of Encephalitis trial would be valuable for the success of future trials set up to address the efficacy of early treatment with intravenous immunoglobulin in all-cause encephalitis in children. Study limitations and future work The study was underpowered to evaluate the efficacy of intravenous immunoglobulin in the treatment of childhood encephalitis due to the small sample size achieved. Future trials should seek to address this important question. Trial registration This trial is registered as Clinical Trials.gov (NCT02308982) and ISRCTN15791925. Funding This award was funded by the National Institute for Health and Care Research (NIHR) Efficacy and Mechanism Evaluation (EME) programme (NIHR award ref: 12/212/15) and is published in full in Efficacy and Mechanism Evaluation; Vol. 11, No. 6. See the NIHR Funding and Awards website for further award information. Plain language summary Encephalitis (inflammation of the brain) is a serious but rare condition affecting approximately 5 in 100,000 children in England. Encephalitis can have a big impact on affected children and their families. Approximately 12 out of 100 affected children will die and half of those that survive experience varying difficulties in the long term; these might include problems with memory, physical disabilities, seizures and changes in how they think and behave. There is some evidence that a treatment called intravenous immunoglobulin may benefit people affected by encephalitis. Intravenous immunoglobulin contains antibodies obtained from blood donations by different people, which is used to treat some types of inflammation. However, there have been no research studies investigating the effect of intravenous immunoglobulin when used in large numbers of children with all types of encephalitis. Furthermore, although intravenous immunoglobulin is sometimes used to treat children with encephalitis, it is often given after other treatments have been unsuccessful. Outcomes from encephalitis are determined largely by the amount of brain inflammation; it would therefore seem logical that giving a treatment early in the illness to limit the inflammation would be beneficial. In the ImmunoglobuliN in the Treatment of Encephalitis study, we aimed to find out whether giving intravenous immunoglobulin to children with encephalitis early in the illness can help them get better more quickly and reduce the difficulties they experience later on. Half of the children in the trial received intravenous immunoglobulin and the other half received an inactive medicine, known as placebo, in addition to the normal care they would receive in a hospital. We aimed to compare the recovery and outcomes between children in these two groups. This trial was stopped early due to withdrawal of funding, as fewer children than expected were enrolled into the study. Too few children were enrolled for us to be sure whether intravenous immunoglobulin benefits children with encephalitis. However, the trial findings highlight the impact of encephalitis on affected children, with around half of children demonstrating ongoing difficulties 1 year after the illness. Scientific summary Background There is significant mortality and morbidity from encephalitis in children despite the current standard of care. Thus, strategies to improve outcomes in patients with encephalitis are urgently required. Theoretical and empirical evidence suggest a beneficial role of intravenous immunoglobulin (IVIG) for viral and auto-immune forms of encephalitis. Therefore, we set up a prospective randomised controlled trial (RCT) to ascertain the efficacy of early IVIG treatment for all-cause encephalitis in children. In the study, children with encephalitis were randomised to receive two doses of either IVIG or placebo within five working days from the suspicion of an encephalitis diagnosis, in addition to normal standard of care. They were then followed up for 12 months after randomisation. We hypothesised that IVIG could have therapeutic benefit for children with encephalitis when administered early in the illness. Objectives The primary objective was to evaluate the efficacy of early IVIG treatment in childhood encephalitis. This was assessed by comparing the proportion of children in the 2 treatment groups who made good recovery, assessed using the paediatric version of the Glasgow outcome score extended, at 12 months after randomisation. The secondary objectives were to: (1) compare clinical, neurological, neuroimaging and neuropsychological outcomes between the treatment groups, (2) evaluate the proportion of participants with autoimmune encephalitis and (3) confirm the safety of IVIG. Methods Trial design This was a phase 3b multicentre, double-blind, randomised, placebo-controlled trial to evaluate the early use IVIG in childhood encephalitis. Setting Participants were recruited from 21 NHS Hospitals in the UK. Participants Inclusion criteria Participants were eligible if they were aged between 6 weeks and 16 years they met the case definition for possible encephalitis based on the International Encephalitis Consensus parents/guardians provided written informed consent, or assent if appropriate. Exclusion criteria The exclusion criteria included: (1) a high clinical suspicion of bacterial meningitis, (2) prior receipt of IVIG during the admission or known contraindication to IVIG, (3) traumatic brain injury, (4) history of metabolic encephalopathy, stroke, toxic or hypertensive encephalopathy, (5) pre-existing demyelinating disorder, (6) significant renal impairment, (7) hypercoagulable state, (8) hyperprolinaemia, (9) participation in another research trial involving an immunomodulatory treatment, (10) known to be pregnant, (11) any significant disease or clinical research that would impact on participation, or interfere with compliance with study requirements. Randomisation Participants were randomly assigned 1 : 1 ratio to receive 2 doses (1 g/kg/dose) of either IVIG or matching placebo, in addition to standard care. Randomisation was stratified by age and receipt of steroid treatment. Interventions Two doses of 1 g/kg/dose of either IVIG or a matching volume of placebo were given 24–36 hours apart, with the first dose administered as soon as possible after enrolment and within five working days from the suspicion of an encephalitis diagnosis. The active treatment (IVIG) used in the study was Privigen (100 mg/ml solution), manufactured and provided by CSL Behring. The placebo was a mixture of 0.9% saline + 0.1% human albumin solution, manufactured at the Royal Broadgreen and Liverpool Aseptic Production Unit, Liverpool, UK, under current good manufacturing practice conditions and its Manufacturer’s Importer’s Authorisation licence. The addition of albumin to saline was necessary to make the placebo visually identical to IVIG. Blinding Participants, treating clinicians, parents/guardians and outcome assessors were blinded to the allocated treatment. Primary outcome The primary outcome was good recovery (i.e. score of ≤2) on the paediatric version of the Glasgow Outcome Score-Extended (GOS-E Peds) at 12 months after randomisation. Secondary outcomes Clinical, neurological and neuropsychological outcomes Multiple clinical and neurological measures were collected during the hospital admission, at 4–8 weeks after discharge from acute care, and at 6 and 12 months after randomisation. A blinded neuropsychologist assessment of cognitive function was carried out at 12 months after randomisation. Radiological outcomes Brain magnetic resonance imaging (MRI) findings at 6 months after randomisation were compared with imaging results obtained during the acute illness. Safety data Safety and adverse events (AEs) data were collected until 12 months after randomisation. Identification of immune-mediated encephalitis Presence of specific auto-antibodies in serum was assessed. Statistical methods The analyses were performed on the intention-to-treat population; this included all participants who were randomised, irrespective of study treatment received. In the analysis of the AEs, all participants who received study treatment were included. Since 20% of participants were recruited before the trial was halted, all analyses are descriptive. Results Recruitment took place between 23 December 2015 and 26 September 2017. Recruitment was paused in September 2017 following withdrawal of funding due to slower than anticipated recruitment. Despite strategies implemented to improve recruitment, funding was not reinstated. Attempts at securing alternative funding were unsuccessful; therefore, the trial was closed on 24 October 2017. Participants and demographics At the time of halting the study, 18 participants had been randomised (IVIG n = 10; placebo n = 8). One participant each from the IVIG and placebo groups were withdrawn from the study before receipt of study treatment and one participant in the placebo arm refused a second dose of study treatment. Therefore 16 participants (IVIG n = 9; placebo n = 7) received at least one dose of the study treatment and 15 participants (IVIG n = 9, placebo n = 6) received two full doses of the study treatment. One participant in the placebo group was lost to follow up after the 6 months visit and 1 participant in the IVIG group withdrew consent prior to the visit 12 months after randomisation. The median age at randomisation was 4.09 years [interquartile range (IQR) = 2.0–11.8], 44% were male and 89% were of white ethnicity. Baseline characteristics were well matched across treatment arms. Primary outcome At 12 months after randomisation, 9 participants [50%; IVIG n = 5 (50%); placebo n = 4 (50%)] made a good recovery (score ≤2 on the GOS-E Peds) and 5 participants [28%; IVIG n = 3 (30%), placebo n = 2 (25%)] made a poor recovery (score > 2). Four participants (22%; IVIG n = 2 (20%), placebo n = 2 (25%)] did not undergo a GOS-E Peds assessment at 12 months after randomisation. Secondary outcomes Inpatient data Ten participants [56%; IVIG n = 5 (50%), placebo = 5 (63%)] were admitted to intensive care, and nine of these [90%; IVIG n = 4 (80%), placebo n = 5(100%)] required invasive ventilation, for a median duration of two days (IQR 2.0–3.0). The median length of stay on intensive care was 4.5 days (3.0–6.8) and the overall median length of hospitalisation for acute care was 11 days (7.8–19.5). Epilepsy diagnosis Three participants [17%; IVIG n = 1 (10%), placebo n = 2 (25%)] had a new diagnosis of epilepsy during the study period. Five participants [28%; IVIG n = 2 (20%), placebo n = 3 (38%)] had incomplete data for this outcome. Glasgow Outcome Score-Extended-Peds at 6 months after randomisation Eight participants [44%; IVIG n = 4 (40%), placebo n = 4 (50%)] made a good recovery at 6 months after randomisation, whereas seven participants [39%; IVIG n = 4 (50%), placebo n = 3 (38%)] made a poor recovery. Three participants [17%; IVIG n = 2 (20%), placebo n = 1 (13%)] did not undergo a GOS-E Peds assessment at 6 months after randomisation. Liverpool outcome score At 4–8 weeks after discharge from acute care, 5 participants [28%; IVIG n = 3 (30%), placebo n = 2 (25%)] made a full recovery [defined as a Liverpool Outcome Score (LOS) of >4], whereas 10 participants [56%; IVIG n = 5 (50%), placebo n = 5 (63%)] had minor to severe sequelae. Three participants [17%; IVIG n = 2 (20%); placebo n = 1 (13%)] did not have LOS data collected at this timepoint. At 12 months after randomisation, six participants [33%; IVIG n = 4 (40%), placebo n = 2 (25%)] made full recovery on the LOS assessment, while 8 participants [44%; IVIG n = 4 (40%), placebo n = 4 (50%)] reported minor to severe sequelae. Four participants [22%; IVIG n = 2 (20%); placebo n = 2 (25%)] did not have LOS data collected at this timepoint. Paediatric quality of life assessment Paediatric quality of life scores were available for seven participants [39%; IVIG n = 5 (50%), placebo n = 2 (25%)] at 4–8 weeks after discharge from acute care and for eight participants [44%; IVIG n = 6 (60%), placebo n = 2 (25%)] at 12 months post randomisation. At 4–8 weeks after discharge from acute care, the mean PedsQL score was 77.9 (standard deviation, SD, 11.10) and 56.5 (SD 7.8) for the IVIG and placebo group, respectively. At 12 months, mean PedsQL scores were 79.9 (SD 21.6) and 63.7 (SD 30.1) for the IVIG and placebo groups, respectively. Gross motor function classification system At 4–8 weeks after discharge from acute care, seven participants [39%; IVIG n = 5 (50%); placebo n = 2 (25%)] had mild impairment of gross motor functioning. These data were not available for 11 participants [61%; IVIG n = 5 (50%), placebo n = 6 (75%)] at this timepoint. At 12 months after randomisation, eight participants [44%; IVIG n = 6 (60%); placebo n = 2 (25%)] experienced mild or severe impairment of gross motor function. These data were not available for ten participants [56%; IVIG n = 4 (40%), placebo n = 6 (75%)] at this timepoint. Strengths and difficulties assessment Strengths and Difficulties Questionnaire (SDQ) results were available for seven participants (IVIG n = 5, placebo n = 2) at 4–8 weeks after discharge from acute care and eight participants (IVIG n = 6, placebo n = 2) at 12 months after randomisation. At 4–8 weeks after discharge from acute care, five participants [28%; IVIG n = 4 (40%); placebo n = 1 (13%)] had a close to average SDQ score, one participant [6%, IVIG n = 1 (10%), placebo n = 0] had a slightly raised SDQ score and one participant [6%, IVIG n = 0, placebo n = 1 (13%)] had a very high SDQ score. At 12 months after randomisation, the same number of participants had a close to average score and slightly raised score, but two participants [11%; IVIG n = 1 (10%), placebo n = 1 (13%)] had a very high SDQ score. Adaptive Behaviors Assessment System – second edition At 4–8 weeks after discharge from acute care, five participants [28%; IVIG n = 4 (40%), placebo n = 1 (13%)] had an Adaptive Behaviors Assessment System-second edition (ABAS-II) score that was either similar or higher than the average score of the normative population. At the same time point, three participants [17%; IVIG n = 2 (20%), placebo n = 1 (13%)] had a score that was lower than the average score. Ten participants [56%; IVIG n = 4 (40%), placebo n = 6 (75%)] did not have ABAS-II assessment at this timepoint. At 12 months after randomisation, the same number of participants had a score that was below the average, but four participants [22%; IVIG n = 3 (30%), placebo n = 1 (13%)] had a score that was either similar or higher than the average score, and 11 participants [61%; IVIG n = 5 (50%), placebo n = 6 (75%)] did not have ABAS-II assessment at this timepoint. Neuropsychology assessment Thirteen participants (72%; IVIG n = 8 (80%); placebo n = 5 (63%)] had blinded neuropsychology assessment at 12 months after randomisation; four [30%; IVIG n = 2 (25%), placebo n = 2 (40%)] of these participants were unable to complete the full battery of assessments due to attention or behavioural needs. Five participants [28%; IVIG n = 2 (20%), placebo n = 3 (38%)] did not undergo neuropsychology assessment. Five participants [28%; IVIG n = 4 (40%), placebo n = 1 (13%)] had a score of ≥85 (indicating normal development) for Full-Scale IQ, six [33%; IVIG n = 4 (40%); placebo n = 2 (25%)] for Verbal Comprehension Index (VCI), five [28%; IVIG n = 4 (40%), placebo n = 1 (13%)] for visual spatial; four [22%; IVIG n = 4 (40%), placebo n = 0 (0%)] for working memory index (WMI); and four [22%; IVIG n = 3 (30%); placebo n = 1 (13%)] for Perceptual Reasoning Index (PRI). Two participants (one in each treatment arm) were assessed using the Bayley scale of infant and toddler development, one participant (IVIG arm) had severe neurodevelopmental impairment while the other (placebo arm) had a normal neurodevelopmental outcome. Neuroimaging Nineteen acute neuroimaging scans were available for 13 participants. Five scans (for five unique participants) had abnormal findings; all of these were MRI scans. Four of the abnormal scans showed bilateral lesions. There were nine follow-up scans for eight unique participants of which six scans (for five unique participants) were normal and unchanged from the acute scan. Three follow-up scans (for three unique participants) had abnormal findings; two of these were unchanged from the acute scans and an acute scan was not available for comparison one participant. Autoantibody testing Twelve participants had autoantibody testing. One participant (placebo n = 1) was positive for LGI1 antibodies, and one participant (placebo n = 1) was positive for myelin oligodendrocyte glycoprotein (MOG) antibodies. Two additional participants (IVIG n = 2) were positive for IgG to live neurons, indicating the presence of IgG antibodies binding to neurons, but negative for antibodies to the specific antigens tested for in the study. Safety reporting One participant in the IVIG group reported an AE of special interest; the participant developed a fever during IVIG infusion; however, this was judged to be unrelated to the study treatment. Ten serious AEs occurred in three participants in the placebo group and none in the IVIG group. None of the serious adverse events (SAEs) were judged to be related to the study treatment. No deaths occurred during the study period. Conclusions ImmunoglobuliN in the Treatment of Encephalitis (IgNiTE) was the first RCT to prospectively evaluate the efficacy of IVIG in childhood encephalitis. It was anticipated that data from IgNiTE would provide definitive evidence on which to base the management of children with encephalitis in the UK and worldwide. However, due to slow recruitment, the trial was terminated early. Therefore, the trial did not reach its pre-determined sample size and was underpowered, making it impossible to reach a conclusion on the role of IVIG in treatment of childhood encephalitis. Nonetheless, IgNiTE has highlighted several key learning points. Firstly, IgNiTE has demonstrated the feasibility of setting up a large multi-centre trial efficacy trial in a cohort of children with a rare condition such as encephalitis. Secondly, data from IgNiTE, albeit derived from a small sample size, suggest the safety of IVIG and provide some insight into the burden that encephalitis places on children and the NHS. Over half of the study participants were admitted to the intensive care unit with 90% of those admitted requiring invasive ventilation, with a prolonged overall length of hospitalisation. In addition, a notable proportion of children experienced some degree of disability at follow-up, highlighting the need to prioritise studies aimed at identifying strategies to alleviate the burden from this rare but debilitating disease. For future studies aimed at addressing the efficacy of early IVIG treatment in childhood encephalitis, consideration should be given to practical challenges with setting up RCTs for rare diseases. These include the necessity to recruit from multiple sites to achieve the sample size, the importance of clinical equipoise amongst treating clinicians, and the trade-off between having a stringent but robust set of entry criteria and the impact of this on recruitment. Trial registration This trial is registered as Clinical Trials.gov (NCT02308982) and ISRCTN15791925. Funding This award was funded by the National Institute for Health and Care Research (NIHR) Efficacy and Mechanism Evaluation (EME) programme (NIHR award ref: 12/212/15) and is published in full in Efficacy and Mechanism Evaluation; Vol. 11, No. 6. See the NIHR Funding and Awards website for further award information.