DexEnceph
(2026)Objective
Test whether adding intravenous dexamethasone to aciclovir improves neurocognitive outcome and is safe in adults with HSV encephalitis.
Study Summary
• No significant difference in any secondary cognitive, functional, imaging, or quality-of-life outcomes, or in all-cause mortality at 78 weeks (13% vs 13%)
• No safety signal: similar adverse event rates (40% vs 38%), no increase in CSF HSV PCR positivity at 2 weeks (11% vs 21%, p=0.25), and no treatment-related deaths
• All 5 clinical relapses occurred in the dexamethasone group; 2 were NMDAR antibody positive
Intervention
Intravenous dexamethasone 10 mg four times daily for 4 days as adjunct to IV aciclovir.
Inclusion Criteria
Adults ≥16 years with suspected encephalitis (febrile illness with new-onset seizure, new focal neurological signs, or alteration in consciousness/cognition/personality/behaviour) and positive HSV type 1 or 2 PCR in CSF, on IV aciclovir 10 mg/kg three times daily.
Study Design
Arms: Dexamethasone 10 mg IV QID x 4 days + IV aciclovir vs IV aciclovir alone
Patients per Arm: 47 randomised per arm (39 dex, 42 control in modified ITT)
Outcome
• Mortality at 78 weeks: 13% vs 13%
• Seizures at 78 weeks: 19% vs 26% (adj OR 0.71, 95% CI 0.22–2.22; p=0.55)
• Time to hospital discharge: adj HR 0.93 (95% CI 0.58–1.48; p=0.75)
• HSV in CSF at 2 weeks: 11% vs 21% (adj OR 0.47, 95% CI 0.13–1.70; p=0.25)
• Serious AEs: 15% vs 9%; all 5 clinical relapses occurred in the dexamethasone group; no treatment-related deaths
Clinical Question
In adults with HSV encephalitis treated with intravenous aciclovir, does adjunct intravenous dexamethasone improve verbal memory at 26 weeks and is it safe?
Bottom Line
In adults with HSV encephalitis, adding IV dexamethasone (10 mg QID x 4 days) to aciclovir did not improve verbal memory at 26 weeks compared with aciclovir alone, but was safe with no increase in CSF HSV PCR positivity at 2 weeks; routine adjunct dexamethasone is not supported, although the safety profile makes empirical use in suspected encephalitis before HSV exclusion reasonable.
Major Points
- First completed phase 3 RCT of corticosteroids in HSV encephalitis: 94 adults from 53 UK NHS hospitals (Sep 2016–Feb 2022).
- Primary outcome (Wechsler Memory Scale-IV Auditory Memory Index at 26 weeks) showed no benefit: 71 vs 69; adjusted difference 1.77 (95% CI –9.57 to 13.12; p=0.76).
- No significant difference in any secondary neuropsychological, functional, imaging, or quality-of-life endpoint, nor in all-cause mortality (13% vs 13% at 78 weeks).
- Dexamethasone did not increase detectable HSV in CSF at 2 weeks (11% vs 21%; adj OR 0.47, 95% CI 0.13–1.70; p=0.25), addressing the long-standing concern of unrestricted viral replication.
- All 5 clinical relapses occurred in the dexamethasone group; 2 were NMDAR antibody-positive but events were generally later than typical post-HSV autoimmune encephalitis.
- Exploratory post-hoc analysis: each day of delay in dexamethasone initiation was associated with a 4.89-point lower verbal memory score (95% CI 2.70–7.08; p<0.0001), suggesting timing may matter.
- Findings do not support routine corticosteroid use in HSV encephalitis but support empirical early steroid use in suspected encephalitis before HSV PCR results return.
Study Design
- Study Type
- Randomized Controlled Trial
- Randomization
- Yes
- Blinding
- Observer-blind (open-label to patients/site clinicians; neuropsychologists, radiologists, statisticians, and lead investigators masked)
- Sample Size
- 94
- Follow-up
- 78 weeks (primary outcome at 26 weeks)
- Centers
- 53
- Countries
- United Kingdom
Primary Outcome
Definition: Verbal memory score at 26 weeks measured by Wechsler Memory Scale-IV Auditory Memory Index
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| 69 (SD 25); n=42 | 71 (SD 26); n=39 | - (–9.57 to 13.12) | 0.76 |
Limitations & Criticisms
- Underpowered for clinically meaningful effects: the 95% CI for primary outcome (–9.6 to 13.1) does not exclude a relevant benefit or harm.
- Open-label, observer-blind design — patients and treating clinicians were not masked, which could bias self-reported and subjective endpoints.
- Long permitted window from admission to randomisation (up to ~3 weeks); median 7 days delay may have placed dexamethasone after the peak inflammatory phase, potentially diluting any benefit.
- Substantial protocol deviations: 17% of control patients received dexamethasone before randomisation as empirical meningitis treatment, and 9% of control patients received dexamethasone after randomisation; one dex-group patient received no dexamethasone.
- Recruited only 94 patients over 5.5 years, far below initial epidemiologic expectations, limiting power for subgroup and imaging analyses.
- Predominantly White UK population (>90%); limited generalisability across racial/ethnic groups and outside high-resource settings.
- Recruitment paused for the COVID-19 pandemic and some follow-up assessments shifted to remote modalities, introducing measurement heterogeneity.
- All 5 clinical relapses occurred in the dexamethasone group — a potential safety signal warranting longer follow-up despite no clear mechanistic explanation.
Citation
Lancet Neurol 2026;25(2):136-146