Dexamethasone in adults with viral meningitis: an observational cohort study

Pelle Trier Petersen,Jacob Bodilsen,Micha Phill Grønholm Jepsen,Lykke Larsen,Merete Storgaard,Birgitte Rønde Hansen,Jannik Helweg-Larsen,Lothar Wiese,Hans Rudolf Lüttichau,Christian Østergaard Andersen,Henrik Nielsen,Christian Thomas Brandt,Danish Study Group of Infections of the Brain (DASGIB)
DOI: https://doi.org/10.1016/j.cmi.2024.08.015
2024-08-23
Abstract:Objectives: To investigate whether there is a dose-dependent association between empiric dexamethasone and outcome in viral meningitis. Methods: Observational cohort study of adults hospitalised for viral meningitis, both with and without a microbiologically confirmed diagnosis, in Denmark between 2015-2020. Dose-dependent associations between dexamethasone (one dose=10 mg) and an unfavourable outcome (Glasgow Outcome Scale score 1-4) at 30 days after discharge were assessed with weighted logistic regression. Entropy balancing was used to compute weights. Results: Of 1,025 included patients, 658 (64%) did not receive dexamethasone, 115 (11%) received 1-2 doses, 131 (13%) received 3-4 doses, and 121 (12%) received ≥5 doses. Among patients treated with dexamethasone, the median number of doses was higher for those without an identified pathogen than for those with a microbiologically confirmed viral aetiology (5 [IQR 3-8] vs 3 [IQR 2-5]; p<0.001). Using no doses of dexamethasone as reference, the weighted odds ratios (OR) for an unfavourable outcome were 0.55 (95% CI 0.29-1.07) for 1-2 doses, 1.13 (95% CI 0.67-1.89) for 3-4 doses, and 1.43 (95% CI 0.77-2.64) for ≥5 doses. In the subgroup of enteroviral meningitis, the weighted OR was 3.08 (95% CI 1.36-6.94) for ≥5 doses but decreased to 2.35 (95% CI 0.65-8.40) when the reference group was restricted to patients treated with antibiotics for suspected bacterial meningitis. Conclusions: This study showed no dose-dependent association between dexamethasone and an unfavourable outcome in patients with viral meningitis. In enteroviral meningitis, ≥5 doses were associated with an increased risk of an unfavourable outcome. However, a sensitivity analysis indicated that the association was affected by unmeasured or residual confounding by severity.
What problem does this paper attempt to address?