No difference in the rate of neurobehavioural cognitive status following the treatment of carbon monoxide poisoning with hyperbaric oxygen at 2.4 atmospheres versus 3.0 atmospheres

1. No evidence for a difference in neurocognitive outcomes between these two protocols
2. Low power study and therefore more investigation required to establish evidence of no benefit with either regimen

Citation/s:
1. Hampson NB , Dunford RG, Ross DE, Wrexford-Brown CE. A prospective, randomised clinical trial comparing two hyperbaric treatment protocols for carbon monoxide poisoning. Undersea Hyperb Med. 2006 Jan-Feb;33(1):27-32
Lead author's name and fax: Neil Hampson neil.hampson@vmmc.org

Three-part Clinical Question: For patients with carbon monoxide poisoning, does treatment with hyperbaric oxygen at 3.0 ATA versus 2.4 ATA result in any difference in neuropsychological outcome?
Search Terms: carbon monoxide poisoning, neuropsychology, hyperbaric dosing

The Study:
Double-blinded randomised controlled trial with intention-to-treat.
The Study Patients: Adults with accidental carbon monoxide poisoning with transient loss of consciousness and presenting to ED within 12 hours. Fully conscious when assessed in ED with normal clinical neurological examination.
Control group (N = 18; 17 analysed): One session of 100% oxygen breathing at 2.4 ATA for 90 minutes.
Experimental group (N = 12; 11 analysed): One session of 100% oxygen at 3.0 ATA for 46 minutes then 2.0 ATA for 50 minutes (USAF CO Table).

The Evidence:

Outcome

Time to Outcome

2.4 ATA group

3.0 ATA group

Relative risk reduction

Absolute risk reduction

NNT

Neurocognitive state abnormal

After treatment

0.22

0.17

25%

0.06

18

95% Confidence Intervals:

-104% to 100%

-0.23 to 0.34

NNT = 3 to INF; NNH = 4 to INF

Neurocognitive state abnormal

14 to 21 days

0.06

0.08

-48%

-0.03

-37

95% Confidence Intervals:

-385% to 100%

-0.22 to 0.16

NNT = 6 to INF; NNH = 5 to INF

Comments:
1. Small study with low power and is described as a pilot to establish the feasibility of randomising such patients

2. Study population is not representative of all CO poisoned patients (transient LOC and accidental only). Those most severely poisoned were excluded.

3. Unequal recruitment into each arm is likely to reflect the inability to recruit the desired number (50). The unequal recruitment is in favour of the shorter, lower pressure treatment regimen and the use of sealed envelope randomisation means there is a possibility of recruitment bias

4. Single treatment session no longer reflects widely practiced standard regimen

5. Clinical significance of the findings is not clear

Appraised by: Mike Bennett, Diving and Hyperbaric Medicine, Prince of Wales Hospital, Sydney ; Saturday, 27 May 2006
Email: m.bennett@unsw.edu.au
Kill or Update By: June 2007