Robertson et al 2014: EPO, transfusion and TBI

Journal Club 005

Authors: Anna Hart
Chris Nickson

Robertson CS, et al. Effect of erythropoietin and transfusion threshold on neurological recovery after traumatic brain injury: a randomized clinical trial. JAMA. 2014 Jul 2;312(1):36-47. doi: 10.1001/jama.2014.6490. PubMed PMID: 25058216; PubMed Central PMCID: PMC4113910.


Does erythropoietin (EPO), which may have neuroprotective effects, improve neurological recovery in traumatic brain injuries?

Does targeting a haemoglobin threshold of 7  g/dL or 10 g/dL improve neurological recovery in traumatic brain injuries?



  • Randomized clinical trial using a factorial design


  • n = 200 patients (erythropoietin, n = 102; placebo, n = 98) with closed head injury who were unable to follow commands and were enrolled within 6 hours of injury
  • patients were from neurosurgical ICUs in two level I trauma centers in Houston, Texas (USA) between May 2006 and August 2012

Inclusion criteria

  • Closed head injury
  • Unable to follow commands
  • Assessed/enrolled within 6 hours

Exclusion criteria (695 excluded)

  • GCS = 3 with fixed and dilated pupils
  • GCS Motor > 5
  • Penetrating injury
  • Pregnancy
  • Age < 15
  • Severe pre-existing disease
  • Severe systemic injuries


  • erythropoietin (500IU/kg per dose; two regimens were used – see criticisms below) (n=102)
  • saline placebo (n =98)
  • hemoglobin transfusion threshold of 10 g/dL (n=101)
  • hemoglobin transfusion threshold of 7 g/dL (n=99)


  • Glasgow Outcome Scale score at 6 months post-injury
    • dichotomized as favorable (good recovery and moderate disability) or unfavorable (severe disability, vegetative, or dead)
    • no significant differences for EPO
      • placebo favorable outcome rate: 34/89 [38.2%; 95% CI, 28.1% to 49.1%]
      • EPO5 favorable outcome rate: 17/35 [48.6%; 95% CI, 31.4% to 66.0%], P = .13
      • EPO3 favorable outcome rate: 17/57 [29.8%; 95% CI, 18.4% to 43.4%],P < .001
    • no significant differences for EPO
      • 37/87 (42.5%) for the hemoglobin transfusion threshold of 7 g/dL
      • 31/94 (33.0%) for 10 g/dL (95% CI for the difference, −0.06 to 0.25, P = .28)
  • Mortality
    • no significant difference
  • Complications
    • ARDS
      • the transfusion threshold of 10 g/dL was not significantly associated with ARDS, though there was a trend towards this (hazard ratio, 1.79 [95% CI, 0.93- 3.45]; P = .08)
    • Infections (including pneumonia, UTI, bacteraemia and ventriculitis)
      • no significant difference
    • Thromboembolic events
      • patients with the transfusion threshold of 10 g/dL had a significantly greater incidence of 1 or more thromboembolic events (22 patients [21.8%] vs 8 patients [8.1%] with the transfusion threshold of 7 g/dL; odds ratio, 0.32 [95% CI, 0.12 to 0.79], P = .009



  • Intention to treat analysis
  • Baseline balance: similar demographics, similar haemoglobin levels at baseline; some differences in mass lesions, spinal haematomas and need for surgery as a result
  • The Alfred ICU is currently involved in the EPO-TBI trial, which should have greater power, to assess the potential neuroprotective benefits


  • Partially blinded — participants and examiners did not know EPO status yet they needed to know Hb threshold assignment
  • Some of the GOS scores were obtained via telephone interviews from the patient, family or caregivers (reporting bias)
  • External validity (and sample size) is limited by the restrictive exclusion criteria, as only 200/700 patients eligible
  • The study protocol for EPO changed during the study (EPO5 versus EPO3)this was triggered by the results of the ‘EPO Stroke trial’ that found an increased death and complication rate when EPO was used to treat acute ischaemic stroke.
    • EPO5 group ((n = 74): the initial regime was 500IU/kg within 6 hours followed by  doses on day 1 and 2 followed by 1 dose each week for 2 weeks (total  5 doses)
    • EPO3 group (n = 126): after regimen was changed the initial dose was only followed by weekly dose for two weeks (total 3 doses)
    • it is uncertain how important dose dependent effects of EPO may be
  • This study was underpowered
    • small sample size (n=200) combined with factorial design
    • powered to test whether erythropoietin would fail to improve favourable outcomes by 20%
    • Dichotomisation of GOS for outcomes may have missed smaller improvements in outcome


  • Administration of EPO to patients with TBI who are unable to follow commands did not improve mortality or neurological status compared to placebo.
  • Targeting a haemoglobin concentration of >100g/L, compared to >70g/L, in these patients has no survival benefit or improvement in neurological outcomes and increases the increased risk of thromboembolic events and may potentially increase rate of ARDS.
  • The routine use of EPO or a transfusion threshold of >100 g/dL hemoglobin in severe TBI patients in clinical practice is not supported

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.