Upper GI bleeds: alcohol added to epinephrine had no clear effect.

Clinical bottom line (level 1b-)

  1. Patients with upper gastrointestinal bleeding who were given epinephrine plus alcohol, had no clear difference in mortality than those given epinephrine alone.
  2. Patients given epinephrine plus alcohol had no clear difference in haemostasis.
Chung et al: Gastrointestinal Endoscopy 1996; 43 (6): 591-595
Expires October 2002

The study

Double-blinded ?concealed randomised trial ?with intention-to-treat
Setting: General hospital, China

160 patients (aged range 17 to 93 years; mean 55, 74% male) acute upper gastrointestinal bleeding

Excluded if
  • bleeding sites not clearly identified
  • non-cooperation with endoscopy
  • protocol violation
  • torrential bleeding that precluded a clear endoscopic view of the ulcers
  • ulcers with contact bleeding only, visible vessels, adherent blood clots or clean bases


    • Note:
  • Apparantly exclusions for 'protocol violation' occurred before randomisation.


    • Control Group: (n = 81, 81 analysed): epinephrine (1/10,000) injection submucosally in 0.5 to 1 ml aliquots around the blleding point until bleeding was controlled
    Experimental Group: (n = 79, 79 analysed): epinephrine , as control plus 0.2-4 ml (median 0.8 ml) of absolute alcohol into the bleeding vessel after the bleeding was controlled by epinephrine injection
    After endoscopy, patients were returned to the surgical gastroenterology ward and given ranitidine intravenously. A blood transfusion was given to maintain the haemoglobin level at 10 gm/dL. All patients were routinely examined with endoscopy 24 hours later and repeat injections of study drugs were repeated if necessary. Emergency surgery was also carried out if necessary. Oral H 2 receptor antagonists or omeprazole were given when each patient was discharged from hospital and endoscopy was carried out 4 weeks later.
    100% followed for 4 weeks

    The evidence

    Outcome Time to outcome CEREERRRR
    (95% CI)    		ARR
    (95% CI)    		NNT
    (95% CI)
    haemostasis unknown 79
    (97.5%)    		 75
    (94.9%)    		 -310%
    (-3490% to 53%)    		 -3.83%
    (-9.23% to 1.57%)    		 -26
    (NNT = 64 to infinity;
    NNH = 11 to infinity)
    mortality unknown 4
    (4.94%)    		 7
    (8.86%)    		 -79.0%
    (-489% to 45.0%)    		 -3.92%
    (-11.8% to 3.92%)    		 -25
    (NNT = 25 to infinity;
    NNH = 8 to infinity)

    Comments

    1. Outcome may have been different if a more effective acid-supression was used (e.g. omeprazole i/v).
    2. The trial was too small to show any clear difference in haemostasis or mortality between the two groups.
    3. The causes of hospital death were terminal malignancy in 2 patients in the epinephrine group and 1 in the combination group; respiratory failure from silicosis (1 in the epinephrine group); duodenal ulcer (1 in the combination group).

    Citation

    1. Chung SCS, Leong H-T, Chan ACW, et al: Epinephrine or epinephrine plus alcohol for injection of bleeding ulcers: a prospective randomized trial. Gastrointestinal Endoscopy 1996; 43 (6): 591-595
    Contributor: Clare Wotton and Musab Hayatli, October 1999
    Reviewer: Lawrence Friedman

    Clinical Question.
    Patient acute upper gastrointestinal bleeding
    Intervention or Exposure epinephrine plus alcohol
    Comparison epinephrine alone
    Outcome haemostasis, mortality