Autoimmune encephalitis: 2026 so far

More reports looked at step-up options, but most evidence was still uncontrolled.

In the months covered, multiple reports described “salvage” or escalation strategies after steroids/IVIG were not enough. These included efgartigimod and ofatumumab, plus other add-on immune therapies in refractory cases. Most evidence was still case reports or small series. In September, a prospective cohort in relapsing anti-NMDAR encephalitis linked ofatumumab with better seizure and functional outcomes than other second-line approaches, with mostly mild reported safety events. A randomized study protocol (RADIA) was also published to test sequential escalation (ofatumumab followed by daratumumab) in severe autoimmune encephalitis, but it has not reported results. This leaves many people facing step-up decisions with active research underway, but limited head-to-head evidence in autoimmune encephalitis so far.

Research covered in: February 2026, April 2026, May 2026, June 2026, August 2026, September 2026

Signals expanded from single cases to small cohorts and review.

From January 2026 to September 2026, efgartigimod repeatedly came up as an antibody-lowering approach used in antibody-positive autoimmune encephalitis (including anti-LGI1 and anti-NMDAR). Reports generally described improvements in function or cognition, along with reductions in total IgG and/or antibody levels. By May and September, small retrospective and prospective real-world cohorts and a systematic review pulled together many uncontrolled reports. Few serious adverse events were reported, but there were no randomized comparisons. Some reports also described sequential strategies (efgartigimod followed by rituximab) with longer follow-up in very small numbers. This suggests efgartigimod is being used in multiple centers, while the evidence base is still early because most data are uncontrolled and short-term.

Research covered in: February 2026, April 2026, May 2026, June 2026, August 2026, September 2026

More work quantified lingering symptoms and caregiver strain beyond disability scores.

In the months covered, several studies pointed out that broad disability scores can improve even when fatigue, cognition, mood/behavior, sleep, speech, and participation limits continue. In July, a disease-specific patient-reported outcome tool (PROSE) was introduced, and a caregiver study found moderate-to-severe burden at discharge linked to anxiety and financial strain. March added clinical descriptions of sleep-disordered breathing and autonomic symptoms (problems with automatic body functions like heart rate and sweating), including objective overnight testing in referred patients; August also documented persistent sleep difficulties during long-term recovery. August also described a pediatric clinic model that embedded brief cognitive screening to speed access to neuropsychology evaluation. June included a single anti-NMDAR case report where intensive rehabilitation started very late (about 20 months after onset) and coincided with major functional gains, without rehab-related complications reported. These reports keep attention on follow-up needs that can include thinking, sleep, mental health, and caregiver impact, not only motor recovery.

Research covered in: February 2026, March 2026, June 2026, July 2026, August 2026

More cohorts linked relapse and disability to clinical features and MRI changes.

Across the months covered, relapse was repeatedly described as not rare. Meta-analytic and cohort estimates were reported, along with multiple proposed predictors (clinical features, timing of immunotherapy, additional autoantibodies, and CSF immune markers). In July, pediatric and anti-NMDAR relapse models were added, including CSF oligoclonal bands (immune-protein patterns that suggest inflammation) in one cohort, noted to be based on few relapse events. In September, a large multicenter cohort reported that about 1 in 4 people relapsed within 1 year (including antibody-negative patients), and MRI atrophy measures were linked to worse 12-month disability. Multiple studies also tied worse outcomes to early severity markers (e.g., ventilation, reduced consciousness) and to specific MRI findings in LGI1 encephalitis (e.g., hippocampal swelling linked to worse later function). This kind of information is increasingly used to shape follow-up plans, even though prediction tools still vary by setting and subgroup.

Research covered in: February 2026, March 2026, June 2026, July 2026, August 2026, September 2026

Multiple studies showed antibody tests and referrals can mislead without context.

From January 2026 to September 2026, studies repeatedly showed that antibody results and single tests can be misleading. In January, work highlighted test-design and sample-type effects (blood vs CSF) and false-positive patterns in anti-AK5 series. In April, a referral-center review reported that about half ultimately had another diagnosis, including some with positive neural antibodies. July and September added multiple scoring tools and add-on CSF markers aimed at separating autoimmune from infectious encephalitis, with performance varying by setting and antibody type, and several papers emphasized the need for external validation. May and August reinforced the role of add-on imaging when MRI is normal, including PET (a scan that measures tissue activity) approaches that detected abnormalities in subsets and sometimes tracked with severity or change over time. This points toward diagnosis relying on a bundle of evidence (symptoms over time, MRI/EEG, CSF studies, antibody testing, and sometimes PET), and it helps explain why re-evaluation is common when the course does not fit.

Research covered in: January 2026, February 2026, April 2026, May 2026, July 2026, August 2026, September 2026