Carbon monoxide poisoning: normobaric oxygen was more effective and safer than hyperbaric oxygen.

Clinical bottom line (level 1b)

  1. Patients with carbon monoxide poisoning who received normobaric 100% oxygen compared with hyperbaric 100% oxygen were less likely to have continued mental retardation requiring further treatment after three sessions (NNT = 8 at unknown) .
  2. Patients given normobaric oxygen had fewer abnormal neuropsychological tests at the end of treatment.
  3. Patients given normobaric oxygen were less likely to have delayed neurological sequelae (NNT = 21 at 30 days) and were not clearly more likely to die than patients given hyperbaric oxygen.
  4. Patients given normobaric oxygen were less likely to suffer from complications (NNT = 13 at 3 days) .
Scheinkestel et al: Medical Journal of Australia 1999; 170: 203-210
Expires July 2003

The study

Double-blinded concealed randomised trial with intention-to-treat
Setting: hyperbaric chamber, university hospital, Australia

191 patients (aged mean 36 years, 81% male) with any grade of carbon monoxide poisoning (mean CO level 23&; 73% severe) referred hyperbaric oxygen

Excluded if
  • pregnant
  • children
  • burns victim


    • Note:
  • Patients were stratified for type of poisoning (suicide/ accidental) and ventilation (mechanically- ventilated/non-ventilated) before randomisation.


    • Control Group: (n = 87, 87 analysed): 100% oxygen at 1.0 atmosphere by hood, occlusive face mask or mechanical ventilator for 100 minutes daily for three days. The chamber was flushed with air regularly to simulate pressurisation
    Experimental Group: (n = 104, 104 analysed): 100% oxygen by hood, occlusive face mask or mechanical ventilation for 100 minutes daily for 3 days, at 2.8 atmospheres for 60 minutes
    All patients had 14 l/min continuous oxygen by face mask or 100% oxygen by endotracheal tube between treatments.
    100% followed for 3 days
    Outcome notes:
    • delayed neurological sequelae : morbidity found on follow-up not obvious on discharge, or worsening of test scores by >SD
    • chamber-related complications : ear barotrauma, oxygen toxicity, severe claustrophobia

    The evidence

    Outcome Time to outcome CEREERRRR
    (95% CI)    		ARR
    (95% CI)    		NNT
    (95% CI)
    more than three further treatments required 3 days 13
    (15.0%)    		 29
    (27.9%)    		 -87%
    (-236% to -4%)    		 -12.9%
    (-24.4% to -1.52%)    		 -8
    (-66 to -4)
    death unknown 3
    (3.45%)    		 3
    (2.88%)    		 16%
    (-304% to 83%)    		 0.56%
    (-4.44% to 5.57%)    		 180
    (NNT = 18 to infinity;
    NNH = 23 to infinity)
    delayed neurological sequelae 30 days 0
    (0.00%)    		 5
    (4.81%)    		 100%
    (% to %)    		 -4.81%
    (-8.92% to -0.70%)    		 -21
    (-140 to -11)
    chamber-related complications 3 days 1
    (1.15%)    		 8
    (8.65%)    		 -653%
    (-5727% to 3%)    		 -7.50%
    (-13.4% to -1.65%)    		 -13
    (-60 to -7)

    Outcome Control Group
    (SD)    		 Experimental Group
    (SD)    		 Mean Difference
    (95% CI)
    number of abnormal neuropsychological tests 2.7
    ()     		3.4
    ()     		-0.7
    (-1.3 to -0.1)

  • Clinical features (95% CI):
    • suicide attempt- 69% (62% to 75%)
    • coma- 53% (46% to 61%)
    • acidosis- 13% (8.3% to 18%)
    • focal neurological deficits- 7.9% (4.0% to 11.2%)
    • EEG changes- 8.4% (4.4% to 12.3%)
    • hypotension- 2.6% (0.4% to 4.9%)
    • arrhythmias- 4.7% (1.7% to 7.7%)
    • pulmonary oedema- 1.6% ( 0.0% to 3.3%)
    • convulsions- 2.1% (0.0% to 4.1%)
    • cardiac arrest- 1.0% (0.0% to 2.5%)
    • headache- 49% (42% to 56%)
    • fatigue- 45% (38% to 52%)
    • difficulty in thinking- 44% (36% to 51%)
    • dizziness- 31% (24% to 37%)
    • nausea- 33% (26% to 40%)
    • acute confusional state- 16% (11% to 21%)
    • paraesthesia- 8.9% (4.9% to 13%)
    • visual disturbance- 6.8% (3.2% to 10%)
    • palpitations- 5.8% (2.5% to 9.1%)
    • chest pain- 5.2% (2.1% to 8.4%)
    • tinnitus- 3.1% (0.7% to 5.6%)
    • diarrhoea- 2.1% (0.1% to 4.1%)
  • 18% of cases were ventilated.
  • Severe CO poisoning was defined as mini-mental score 24 or less, COHb level >30%, confusion, focal neurological deficits, loss of consciousness, ECG abnormalities, arrhythmias, pulmonary oedema, metabolic acidosis, hypotension, convulsions and cardiac arrest.

    • Comments

    1. The present trial adds significant knowledge to the medical literature not because of evidence of lacking efficacy of hyperbaric oxygen, but rather because of the proven potential harm compared to standard oxygen application.

    Citation

    1. Scheinkestel CD, Bailey M, Myles PS, et al: Hyperbaric or normobaric oxygen for acute carbon monoxide poisoning: a randomised controlled clinical trial. Medical Journal of Australia 1999; 170: 203-210
    Contributor: Chris Ball and Clare Wotton, July 2000
    Reviewer: Dirk Stengel

    Clinical Question.
    Patient CO poisoning
    Intervention or Exposure hyperbaric oxygen
    Comparison normobaric oxygen
    Outcome continued mental retardation, neurological sequale, side effects