Fukami
Hyperbaric oxygen did not improve the outcome from postoperative adhesive small bowel obstruction
Clinical Bottom Line:
1.Hyperbaric oxygen made no difference to the outcome
Citation/s: 1. Fukami Y, Kobayashi S, Sekoguchi E, Kurumiya Y. Randomized controlled trial of hyperbaric oxygen therapy in adhesive postoperative small bowel obstruction. Langenbeck's archives of surgery. 2018 Aug 1;403(5):555-9.
Lead author's name and fax:Y Fukami yasuyuki490225@yahoo.co.jp
Three-part Clinical Question:For patients with postoperative small bowel obstruction, does the addition of a course of hyperbaric oxygen to the usual care, result in a reduction in the time to recovery or need for operation?
Search Terms:Small bowel obstruction., postoperative compliations
The Study: Non-blinded concealed randomised controlled trial with intention-to-treat.
The Study Patients:Adult patients with a history of abdominal surgery more than four weeks previously and clinical signs and symptoms of mechanical obstruction, with dilated loops of small intestine of CT scan.
Control group(N = 40; 40 analysed): No oral intake, IV fluids and insertion of nasograstric tube.
Experimental group(N = 33; 33 analysed): As above plus 100% osygen breathing at 2 ATA for 90 minutes daily up to seven times or until improvement in symptoms.
The Evidence:
Outcome |
Time to Outcome |
Standard group |
Hyperbaric group |
Relative risk reduction |
Absolute risk reduction |
NNT |
Need for operation |
no defined period |
0.10 |
0.18 |
-82% |
-0.08 |
12 |
95% CIs: |
-243% to 79% |
-0.24 to 0.08 |
NNT = 4 to INF; NNH = 13 to INF | |||
Successful medical treatment |
No defined period |
0.73 |
0.79 |
-9% |
-0.06 |
16 |
95% CIs: |
-36% to 18% |
-0.259 to 0.133 |
NNT = 4 to INF; NNH = 7 to INF |
Measure |
Standard Group |
Hyperbaric group |
Difference |
95% CI | ||
Mean |
SD |
Mean |
SD | |||
Time to oral intake (days) |
4.2 |
3.7 |
4.2 |
3.3 |
0.0 |
-5.8 to 5.8 |
Length of hospital stay (days) |
11.5 |
7.0 |
14.2 |
10.0 |
-2.7 |
-36.4 to 31.0 |
Comments:
1. Small trial, somewhat underpowered (original target 100 individuals randomised), but no indication of likely benefit.
2. Bias remains possible due to unbiased nature of trial, but this seems unlikely to be large given the nature of the endpoints.
Appraised by:Michael Bennett, POWH m.bennett@unsw.edu.au; Monday, 22 July 2019