Upper GI bleed: omeprazole did not prevent rebleeding, need for surgery or death

Clinical bottom line (level 1b)

  1. Patients with an upper GI bleed who had omeprazole compared with placebo were not clearly less likely to rebleed, require surgery or die.
  2. There was no difference in the requirement for transfusion.
Daneshmend et al: British Medical Journal 1992; 304: 143-147
Expires January 2003

The study

Double-blinded ?concealed randomised trial with intention-to-treat
Setting: 2 university hospitals, UK

1147 patients (aged mean 59, 64% male) with overt upper GI bleeding or a history of haematemesis or melaena within the last 24 hours

Excluded if
  • aged <18
  • previous participation in study
  • unable to start medication within 12 hours of admission
  • potential drug interaction: e.g. warfarin or phenytoin
  • pregnant
  • terminal disease or advanced malignancy
  • immediate surgery required
  • trivial bleeding
  • bleeding in patients admitted for other reasons


    • Control Group: (n = 569, 569 analysed): placebo
    Experimental Group: (n = 578, 578 analysed): omeprazole 80 mg iv bolus then 40 mg iv q8h for 1 day, then 40 mg po bd

    100% followed for ? discharge from hospital
    Outcome notes:
    • rebleeding : overt haematemesis, passage of fresh rectal blood, a fall in Hb > 20 g/l within 24 hours, shock in the presence of continuing melaena, or fresh blood in the stomach or duodenum on repeat endoscopy

    The evidence

    Outcome Time to outcome CEREERRRR
    (95% CI)    		ARR
    (95% CI)    		NNT
    (95% CI)
    rebleeding unknown 100
    (17.6%)    		 85
    (14.7%)    		 16%
    (-9% to 36%)    		 2.87%
    (-1.39% to 7.13%)    		 35
    (NNT = 14 to infinity;
    NNH = 72 to infinity)
    transfusion required unknown 302
    (53.1%)    		 298
    (51.6%)    		 3%
    (-8% to 13%)    		 -1.65%
    (-4.41% to 1.12%)    		 -61
    (NNT = 89 to infinity;
    NNH = 23 to infinity)
    operation unknown 63
    (11.1%)    		 62
    (10.7%)    		 3%
    (-35% to 30%)    		 0.35%
    (-3.26% to 3.95%)    		 290
    (NNT = 25 to infinity;
    NNH = 31 to infinity)
    death unknown 30
    (5.27%)    		 40
    (6.92%)    		 -31%
    (-110% to 17%)    		 -1.65%
    (-4.41% to 1.12%)    		 -61
    (NNT = 89 to infinity;
    NNH = 23 to infinity)

    Comments

    1. Patients were randomised in blocks of ten.

    Citation

    1. Daneshmend TK, Hawkey CJ, Langman MJ: Omeprazole versus placebo for acute gastrointestinal bleeding: randomised double-blind controlled trial. British Medical Journal 1992; 304: 143-147
    Contributor: Alan Townsend and Chris Ball, Unknown Month 1999
    Reviewer: .

    Clinical Question.
    Patient upper GI bleed
    Intervention or Exposure omeprazole
    Comparison placebo
    Outcome rebleeding, death, surgery, transfusion