DexEnceph
(2026)Objective
To evaluate whether IV dexamethasone 10 mg four times daily for 4 days, added to standard IV aciclovir, improves verbal memory at 26 weeks in adults with HSV encephalitis.
Study Summary
• No significant differences were found in any secondary neuropsychological, functional, or imaging outcomes at 26 or 78 weeks.
• Mortality at 78 weeks was 13% in both groups; seizure rates and time to hospital discharge were similar.
• A post-hoc analysis showed earlier dexamethasone initiation was associated with better verbal memory (4.89 points lower per day of delay, p<0.0001), raising the question of whether timing is critical.
Intervention
IV dexamethasone 10 mg four times daily for 4 days plus IV aciclovir vs IV aciclovir alone
Inclusion Criteria
Adults 16 years or older with HSV type 1 or 2 PCR-positive encephalitis, presenting with new-onset seizure or new focal neurological signs or altered consciousness/cognition/personality
Study Design
Arms: Dexamethasone 10 mg IV four times daily x 4 days + aciclovir (n=47) vs Aciclovir alone (n=47)
Patients per Arm: 47 dexamethasone, 47 control (modified ITT: 39 vs 42)
Outcome
• No benefit seen in any secondary cognitive, functional, imaging, or clinical outcome.
• Safety: similar adverse event rates in both groups; no treatment-related deaths; dexamethasone did not increase HSV viral persistence in CSF.
Bottom Line
Adding dexamethasone to standard aciclovir treatment did not improve verbal memory at 26 weeks in adults with HSV encephalitis. The trial was safe with no evidence of harm from dexamethasone and no increase in viral persistence. Routine use of adjunct dexamethasone in HSV encephalitis is not supported; however, a post-hoc signal suggesting earlier treatment may be more effective warrants further study.
Major Points
- Adjunct dexamethasone (10 mg IV four times daily for 4 days) did not significantly improve verbal memory at 26 weeks compared to aciclovir alone (adjusted difference 1.77 points, 95% CI -9.57 to 13.12, p=0.76).
- No significant benefit was found in any secondary neuropsychological domain (visual memory, working memory, processing speed, executive function) or clinical outcomes (mortality, seizure rate, time to discharge) at 26 or 78 weeks.
- Dexamethasone was safe: adverse event rates were similar between groups, there were no treatment-related deaths, and there was no significant increase in HSV PCR positivity in CSF at 2 weeks (11% vs 21%, p=0.25).
- 5 clinical relapses occurred, all in the dexamethasone group -- potentially explained by early corticosteroid suppression of the immune response to HSV before sufficient viral clearance, with subsequent autoimmune activation.
- Important post-hoc finding: each day of delay to dexamethasone initiation was associated with a 4.89-point reduction in verbal memory score (95% CI 2.70-7.08, p<0.0001), suggesting that if corticosteroids are beneficial, timing may be critical.
- The 95% CI for the primary outcome (-9.57 to 13.12) includes the possibility of a clinically meaningful benefit (15.5-point threshold), meaning the trial may have been underpowered to exclude a real benefit.
- Findings are consistent with the hypothesis that early corticosteroid treatment in all suspected encephalitis (before HSV is excluded by CSF PCR) is unlikely to cause harm, and may support empiric use in appropriate patients.
Study Design
- Study Type
- Multicentre, observer-blind, randomised, phase 3, controlled trial
- Randomization
- Yes
- Blinding
- Observer-blind (neuropsychologist assessing primary outcome and statisticians masked; patients, site teams not masked)
- Sample Size
- 94
- Follow-up
- 78 weeks (primary outcome at 26 weeks)
- Centers
- 53
- Countries
- UK
Primary Outcome
Definition: Verbal memory score at 26 weeks, measured by the Wechsler Memory Scale (WMS-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
- Small sample size (n=81 in modified ITT) -- underpowered for secondary outcomes; primary outcome 95% CI includes potential clinically meaningful benefit
- Recruitment took 5.5 years across 53 hospitals, reflecting the rarity of HSV encephalitis and enrollment challenges
- Observer-blind rather than double-blind -- patients and site teams were aware of allocation, introducing potential performance and detection bias
- Delayed enrollment: median 7-8 days from hospital admission to randomisation -- corticosteroids may be most effective in early inflammation; this could have masked benefit
- 5 clinical relapses all occurred in the dexamethasone group, raising concern about post-viral autoimmune phenomena potentially triggered by early immune suppression
- COVID-19 pandemic required remote assessments for some participants, reducing neuropsychological data quality
- All participants were recruited in the UK -- predominantly White population (>90%), limiting generalisability to diverse populations
Citation
Lancet Neurol. 2026;25:136-146