Adjunctive corticosteroids in HSV encephalitis: no difference, on very little data

Design

Systematic review and meta-analysis; two randomised trials pooled, two retrospective cohorts assessed separately

Population

209 patients with herpes simplex virus encephalitis receiving background antiviral therapy, across four comparative studies

Primary outcome

Study-defined unfavourable global neurological or functional outcome, and all-cause mortality

Effect

Unfavourable outcome RR 1.00 (95% CI 0.68 to 1.47); mortality RR 0.92 (0.35 to 2.40)

Four comparative studies, 209 patients in total, were pooled: two randomised trials and two retrospective cohorts, analysed separately because the non-randomised allocation carried obvious confounding. Aciclovir was the background antiviral wherever the regimen was stated. Across the randomised evidence, adding a systemic corticosteroid changed nothing measurable — unfavourable global neurological or functional outcome RR 1.00 (95% CI 0.68 to 1.47), all-cause mortality RR 0.92 (0.35 to 2.40), serious adverse events RR 1.14 (0.54 to 2.44), seizures during follow-up RR 0.73 (0.33 to 1.62). Barthel Index differences at about six months were 3.05 points (−6.99 to 13.09).

The useful reading is not that steroids fail. It is that after two randomised trials the question is still open, and the intervals are wide enough to hold both a worthwhile benefit and a real harm. Two findings deserve watching rather than dismissal: in DexEnceph, HSV DNA persisted in cerebrospinal fluid at about day 14 in 4 of 36 dexamethasone recipients against 9 of 43 controls, and five relapses occurred with dexamethasone against none with control. Five events cannot establish causation, but they are the direction a clinician would worry about.

So routine steroids in unselected patients are not supported. The case that most units actually face — a patient with life-threatening cerebral oedema and mass effect — was not evaluated by any of these studies, and this analysis neither licenses nor forbids steroids there. Treat that as an individual decision made on imaging and intracranial pressure, and document the reasoning, rather than citing a trial that did not enrol such patients.

• Start aciclovir on suspicion; nothing here alters the antiviral, which is the part with a survival benefit.
• Do not add a corticosteroid as routine practice in an unselected patient with HSV encephalitis.
• Where you do use one for mass effect or malignant oedema, record the imaging finding that prompted it.
• If a patient given dexamethasone deteriorates or relapses after apparent recovery, repeat CSF HSV PCR rather than assuming autoimmune encephalitis.
• Count seizures and functional status at discharge — these are the outcomes the trials could not separate.

The statistics, in plain English

A risk ratio of 1.00 with a confidence interval from 0.68 to 1.47 means the pooled result sits exactly on 'no difference', but the data are compatible with anything from a third fewer bad outcomes to half as many again. That is an underpowered answer, not a negative one. The mortality interval, 0.35 to 2.40, is wider still. With 209 patients across four studies and only two of them randomised, no amount of pooling fixes the sample size — which is why the authors call for adequately powered multicentre trials with prespecified severity subgroups rather than declaring the question settled.

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.