BACKGROUND: Herpes simplex virus (HSV) encephalitis is characterised by inflammation and swelling of the brain, resulting in death or severe neurocognitive sequelae. HSV encephalitis is treated with the antiviral aciclovir but still has substantial mortality and morbidity. Adjunct corticosteroids are sometimes used, but whether they improve the outcome is unclear. We aimed to establish the safety and efficacy of adjunct corticosteroids in HSV encephalitis. METHODS: In this multicentre, observer-blind, randomised, phase 3 trial, adults with HSV encephalitis admitted to 53 hospitals in the UK were randomly assigned to receive intravenous dexamethasone (10 mg/kg, four times daily for 4 days) plus intravenous aciclovir (10 mg/kg three times daily for at least 14 days; dexamethasone group), or intravenous aciclovir alone (control group). Eligible patients aged 16 years or older had suspected encephalitis (a febrile illness with new onset seizure, new focal neurological signs or alteration in consciousness, cognition, personality, or behaviour), with positive HSV type 1 or type 2 PCR test in CSF. Participants were randomly assigned by the trial team at the recruiting site, using a secure web-based randomisation programme. The primary outcome was verbal memory score at 26 weeks, measured by the Wechsler Memory Scale (WMS)-IV auditory memory index. Analyses of primary and secondary outcomes were performed according to the modified intention-to-treat principle, and safety analyses were based on whether participants received at least one dose of the study drug. Trial neuropsychologists assessing the primary outcome, and statisticians involved in the study's primary outcome, were masked to treatment group allocation. The trial is registered with the International Standard Randomised Controlled Trial Number Registry, ISRCTN11774734, and EudraCT, EudraCT2016-004835-19, and is completed. FINDINGS: Between Sept 22, 2016, and Feb 2, 2022, 94 patients with HSV encephalitis were recruited, 47 (20 [43%] female and 27 [57%] male) to the dexamethasone group, and 47 (24 [51%] female and 23 [49%] male) to the control group. Dexamethasone was initiated a median 7 (IQR 4-8) days after hospital admission. Seven patients withdrew consent, and six were lost to follow-up. Thus, 81 were included in the modified intention-to-treat analysis (39 in the treatment group and 42 in the control group). The primary outcome, verbal memory score at 26 weeks, did not differ significantly between the groups (71 [SD 26] in the dexamethasone group, and 69 [SD 25] in the control group; adjusted difference 1·77 [95% CI -9·57 to 13·12; p=0·76). There were 27 adverse events involving 18 (38%) participants in the control group, and 25 adverse events involving 18 (40%) participants in the dexamethasone group. The most common serious adverse events were seizures requiring readmission to hospital (affecting one [2%] patient in the dexamethasone group and one [2%] patient in the control group) and thrombotic events, including deep vein thrombosis (affecting one [2%] patient in the dexamethasone group) and pulmonary embolism (affecting one [2%]patient in the dexamethasone group). There were no treatment-related deaths. INTERPRETATION: In adults with HSV encephalitis, dexamethasone plus aciclovir had a satisfactory safety profile but did not improve verbal memory score compared with aciclovir alone. Given the established role of corticosteroids in other inflammatory encephalitides, our findings suggest that their early use in suspected encephalitis is unlikely to be harmful. Future studies should assess more targeted immunomodulatory approaches in HSV encephalitis. FUNDING: National Institute for Health and Care Research.