Adjunctive steroids in herpes encephalitis: still no evidence to act on

Design

systematic review and random-effects meta-analysis of randomised and comparative observational studies, searched to July 2026, randomised and observational evidence analysed separately

Population

four comparative studies, 209 patients with herpes simplex virus encephalitis on background aciclovir

Primary outcome

co-primary unfavourable global neurological or functional outcome and all-cause mortality

Effect

unfavourable outcome RR 1.00 (95% CI 0.68-1.47), mortality RR 0.92 (0.35-2.40), serious adverse events 1.14 (0.54-2.44), seizures 0.73 (0.33-1.62)

Corticosteroids are given in herpes simplex virus encephalitis on the reasoning that inflammatory brain injury continues after the virus is controlled. This meta-analysis searched three databases to July 2026 and found four comparative studies with 209 patients between them — two randomised trials and two retrospective series, with aciclovir as background therapy wherever the regimen was specified.

Across the randomised evidence nothing moved. Unfavourable global outcome had a risk ratio of 1.00 (95% CI 0.68-1.47), mortality 0.92 (0.35-2.40), serious adverse events 1.14 (0.54-2.44) and seizures during follow-up 0.73 (0.33-1.62). Barthel Index differences were 3.05 points at six months (-6.99 to 13.09) and 0.30 at discharge. Cognitive findings were not consistently favourable. The retrospective estimates conflicted with each other and were heavily confounded. Two details deserve carrying: in DexEnceph, HSV DNA persisted at day 14 in 4 of 36 dexamethasone recipients against 9 of 43 controls, and five relapses occurred with dexamethasone against none with control — too few events to conclude anything, but not nothing.

The honest position is that 209 patients cannot answer this question, and the authors say so. What follows for practice is that routine adjunctive steroids in unselected patients are not supported, while the situation clinicians actually agonise over — life-threatening cerebral oedema — was never specifically studied and remains a judgement call made without evidence.

• Do not give adjunctive corticosteroids routinely in herpes encephalitis
• The evidence does not address life-threatening cerebral oedema, where the decision stays individual
• Five relapses on dexamethasone against none on control is too few to conclude, and too many to ignore
• Aciclovir, started early, remains the only intervention with established benefit
• If steroids are given for raised intracranial pressure, document the reason and the intended duration

The statistics, in plain English

A risk ratio of exactly 1.00 with an interval from 0.68 to 1.47 is not evidence of equivalence — with 209 patients across four studies, the data are compatible with a third fewer bad outcomes or half as many again. That is the difference between 'no effect found' and 'no effect'. The mortality interval (0.35-2.40) is wider still and effectively uninformative. Pooling only the randomised evidence and reporting the observational studies separately is the right decision, because treatment allocation in retrospective series follows severity.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.