Trial of Dexamethasone for Chronic Subdural Haematoma

Hutchinson, PJ, NEJM, 2020: December 16 DOI: 10.1056/NEJMoa202473

Clinical Question

  • In patients with chronic subdural haemorrhage, what effect does dexamethasone compared with placebo have on neurological outcome at 6 months?


  • A Subdural Haematoma forms when bleeding occurs between the dura and arachnoid membranes. In patients with cerebral atrophy the onset of symptoms may be insidious and may be detected some time after a minor head trauma or be associated with anticoagulation therapy
  • The management of chronic SDH includes surgical evacuation of the clot if there is evidence of neurological deterioation or if the clot is >10mm or is causing midline shift of >5mm
  • Systemic reviews of the use of glucocorticoids in CSDH have suggested that they reduce the chance of clot recurrence however the effect on neurological outcome is not known to date


  • Randomised controlled trial: 1:1 randomisation of a 14 day tapering course of dexamethasone OR placebo
  • Permuted blocks (random block sizes of 2 or 4) with stratification according to trial site
  • Allocation concealment achieved via a web-based allocation tool
  • Double-blinding: both the treating clinicians and outcome assessors were unaware of the allocation
  • Powered to allow for a 15% loss to follow-up. Target sample size of 750 patients would allow a 81-92% power to detect a treatment effect of 8 percentage points
  • Consent obtained from the patient or next-of-kin
  • An independent data monitoring committee and ethics committee reviewed the trial at 6-12 monthly intervals


  • 23 sites in the United Kingdom
  • Patients were enrolled from August 2015 to November 2018, with 6-month follow-up completed by mid 2019.


  • Inclusion: Patients >18 years with a symptomatic CSDH seen as a hypodense or isodense crescentic collection on the CT head
  • Exclusion:
    • Glucocorticoids were contraindicated
    • Glucocorticoids were indicated for other medical conditions
    • CSF shunt in situ
    • History of psychotic disorder
    • Patients with acute SDH
    • Patients who were unable to receive the trial drug within 72 hours after admission
  • 2203 were screened for eligibility. 750 were randomised to a trial group (2 were subsequently deemed not eligible). So 748 patients were enrolled with 375 in the dexamethasone group and 373 in the placebo group
  • 45 patients withdrew consent and 23 were lost to follow-up leaving 680 patients in the modified intention to treat analysis
  • Patients were well matched at baseline: they were predominantly elderly males (74 years, 74% male) with the most common presenting symptoms being headache and gait disturbance, with head trauma historically noted in around 70% of cases. 47% were on some form of antithrombotic medication


  • Dexamethasone
    • 2-week course of oral dexamethasone
      • 8mg bd Day 1-3, 6mg bd Day 4-6, 4mg bd on Day 7-9, 2mg bd on Day 10-12 then 2mg daily on Day 13-14. (If the patient could not swallow they had NG delivery of the dexamethasone)


  • Matched placebo

Management common to both groups

  • Other management was at the discretion of the treating clinicians and was not protocolised


  • Primary outcome: 6 month modified rankin score (MRS) – significantly better outcomes in placebo group
    • This scale has 7 categories. 0=no symptoms, 1=no clinically significant disability despite symptoms, 2=slight disability, 3=moderate disability, 4=moderately severe disability, 5=severe disability, 6=death
    • Comparing dexamethasone vs. placebo group
      • A favourable outcome, (MRS of 0-3)
        • 83.9% vs. 90.3%  (-6.4% difference, 95% CI -11.4 to -1.4)
  • Secondary outcomes: Comparing intervention vs. control group
    • Significantly better outcomes in control group
      • MRS 0-3 at 3 months
        • 83.2% vs 91.4% (-8.2%, 95% CI -13.3 to -3.1)
      • Mortality at 30 days
        • 2.1% vs 0.5% (OR 4.08, 95% CI 1.01 to 27.2)
    • No significant difference in
      • MRS 0-3 at discharge from neurosurgical unit
        • 80.2% vs 83.2%, (-3% 95% CI -9.1 to 3)
      • Mortality at 6 months
        • 8.8% vs 5%, (OR 1.83, 95% CI 0.99 to 3.45)
      • Surgical intervention in the index admission
        • 91.7% vs 89.2%
      • Surgical intervention in a subsequent admission
        • 5.1% vs 7.6%
    • Significantly reduced in intervention group
      • Repeat surgery for recurrence of subdural haematoma
        • 1.7% vs 7.1% (-5.4%, 95% CI -8.7 to -2.5)
    • Adverse events – significantly more likely in the dexamethasone group
      • Adverse events of special interest
        • 10.9% vs 3.2%
      • Serious adverse events up to Day 30
        • 16% vs 6.4% (OR 2.49, 95% CI 1.54-4.15)
      • Subdural empyema more common in dexamethasone group

Authors’ Conclusions

  • In patients with symptomatic chronic subdural haematoma, treatment with dexamethasone resulted in fewer favourable outcomes and more adverse events compared with placebo. However, recurrence of the subdural haematoma, requiring repeat surgery was less likely in the dexamethasone group


  • Allocation concealment
  • Blinding
  • Patients were analysed in the group they were allocated to even if they didn’t receive the therapy
  • Analysis was by modified intention to treat, excluding those who withdrew consent or were lost to follow-up
  • Multi-centre study improving external validity of results
  • The study picked a clinically meaningful outcome as the primary outcome. Previous studies of dexamethasone in CSDH had used recurrence of SDH as the outcome which is not as important to the patient as neurological recovery


  • Loss to follow-up or withdrawal of consent in 9% which was accounted for in the power calculation
  • 10 patients assigned to the dexamethasone group did not receive this
  • 12 patients assigned to placebo received dexamethasone

The Bottom Line

  • Patients with chronic subdural haematoma should not be treated with dexamethasone
  • This study is reassuring that my usual practice of not using dexamethasone in this cohort is the correct approach

External Links


Summary author: Celia Bradford @celiabradford
Summary date: 12/1/21
Peer-review editor: @davidslessor



Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.