Should we be starting PPIs on undifferentiated upper GI bleed patients in the emergency department prior to endoscopy?
The first question you should ask yourself is what you are hoping to achieve for your patient. What clinically important or patient-oriented outcomes can we improve? When I see somebody vomiting blood, the first thing I am thinking about is saving their life. Next, I am hoping I can get the bleeding stopped (I know: all bleeding stops – eventually) and prevent the morbidity of something like emergency surgery. What does the Cochrane review say about that?
Sreedharan A, Martin J, Leontiadis GI, et al. Proton pump inhibitor treatment initiated prior to endoscopic diagnosis in upper gastrointestinal bleeding. Cochrane Database Syst Rev. 2010;(7)CD005415. [pubmed]
- There are a total of 6 RCTs including 2223 patients
- No change in mortality (odds ratio (OR)1.12; 95% CI 0.72 to 1.73) 6.1% vs 5.5%
- No change in rebleeding (OR 0.81; 95%CI 0.61 to 1.09) 13.9% vs 16.6%
- No change in need for transfusion (OR 0.95; 95% CI 0.78 to 1.16)
- No change in need for surgery (OR 0.96 95% CI 0.68 to 1.35) 9.9% vs 10.2%
- Conclusion: “there is no evidence that PPI treatment affects clinically important outcomes, namely mortality, rebleeding or need for surgery.”
But we use PPIs all the time – they must do something! The Cochrane review did find that there were fewer patients with “stigmata of recent hemorrhage” when treated with a PPI: 37.2% versus 46.5% (OR 0.67; 95% CI 0.54 to 0.84). However, I can’t imagine any of my patients caring very much what the inside of their stomach happens to look like, if they are not going to die, need more blood, rebleed, or have surgery. Also, it is worth noting that the Cochrane review wasn’t even convinced this is real. They state that, “this result was not robust, that is, it became statistically nonsignificant with the exclusion of one of the trials (Daneshmend 1992).”
And it’s even worse than that. “Stigmata of hemorrhage” are classified using the Forrest classification, but unfortunately, endoscopists rarely agree with each other about what they are looking at:
Mondardini A, Barletti C, Rocca G, et al. Non-variceal upper gastrointestinal bleeding and Forrest’s classification: diagnostic agreement between endoscopists from the same area. Endoscopy. 1998;30:(6)508-12. [pubmed]
- Based on video taped endoscopy (so probably better than reality, because everyone is looking at exactly the same images), 47 expert endoscopists (so, again, probably better than reality) reviewed and classified the ulcers displayed
- For type 1a and 1b lesions, there was OK agreement (kappa of 0.76 and 0.61 respectively)
- For all others, kappas were horrible, between 0.44 and 0.49
So although stigmata of hemorrhage may be changed by PPIs, it is not a reliable outcome and it is not a patient centered outcome.
That is probably enough, but there are some interesting notes on some of the individual studies.
Lau JY, Leung WK, Wu JC, et al. Omeprazole before endoscopy in patients with gastrointestinal bleeding. N Engl J Med. 2007;356:(16)1631-40. [pubmed]
- They changed their primary outcome partway through the study when it became obvious that they were not going to be able to demonstrate a mortality benefit (the original primary outcome)
- The trial was registered with clinicaltrials.gov and you can see that the original primary outcome was 30 day mortality, and was changed at the end of the trial to need for endoscopic therapy
- There was no change in amount of blood transfused (1.54 and 1.88 units, respectively; P=0.12) or the number of patients who had recurrent bleeding (11 and 8, P=0.49), who underwent emergency surgery (3 and 4, P=1.00), or who died within 30 days (8 and 7, P=0.78)
- Endoscopic intervention was decreased: 19.1% vs 28.4% p=0.007
Bottom line: No change in the original primary outcome. A secondary outcome, changed to primary part way through the study, showed a decrease in interventions required – but this is a subjective outcome, and also of questionable relevance to patients. Even if it is an important outcome, it was one of many secondary outcomes, and cannot be anything but hypothesis generating.
Daneshmend TK, Hawkey CJ, Langman MJ, Logan RF, Long RG, Walt RP. Omeprazole versus placebo for acute upper gastrointestinal bleeding: randomised double blind controlled trial. BMJ. 1992;304:(6820)143-7. [pubmed]
- 1147 patients randomized to IV omeprazole or placebo
- No significant differences were found between the placebo and omeprazole groups for rates of transfusion (302 (53%) placebo v 298 (52%) omeprazole), rebleeding (100 (18%) v 85 (15%)), operation (63 (11%) v 62 (11%)), and death (30 (5.3%) v 40 (6.9%)). However, there was an unexpected but significant reduction in endoscopic signs of upper gastrointestinal bleeding in patients treated with omeprazole compared with those treated with placebo (236 (45%) placebo 176 (33%) omeprazole; p less than 0.0001).
Bottom line: This is the study that single handedly drives the “stigmata of bleeding” change noted by Cochrane – it was also a secondary outcome.
Hasselgren G, Lind T, Lundell L, et al. Continuous intravenous infusion of omeprazole in elderly patients with peptic ulcer bleeding. Results of a placebo-controlled multicenter study. Scand J Gastroenterol. 1997;32:(4)328-33. [pubmed]
- Patients with KNOWN peptic ulcers
- Stopped early for harm
- 21 day mortality 6.9% in the omeprazole group and 0.6% in the placebo group
- Conclusion of paper: “Three days’ infusion of omeprazole improved overall outcome and reduced need for intervention in PUB patients.” → Seems to be some obvious bias here, in that they put a positive spin on a trial that was stopped early because of harm
Bottom line: This study was stopped early because there was increased mortality in the PPI group!!
Finally, it is worth noting that there is no difference between different dosing regimens for PPIs. I take that to indicate that, if you are giving placebo, or a medication that has no effect, it doesn’t matter how much of it you give. It is also worth including this in case you can’t convince your gastroenterologists to stop using PPIs, at least you can go for the lowest dose and cheapest route.
Wang CH, Ma MH, Chou HC, et al. High-dose vs non-high-dose proton pump inhibitors after endoscopic treatment in patients with bleeding peptic ulcer: a systematic review and meta-analysis of randomized controlled trials. Arch Intern Med. 2010;170:(9)751-8. [pubmed]
- A meta-analysis of 7 RCTs in patients with KNOWN peptic ulcer bleeds (not undifferentiated ED patients)
- High dose = equivalent of 80mg IV bolus then 8mg/hr (192mg/day)
- Low dose = range from 40mg IV daily to BID to 80mg PO BID
- There was no statistically significant difference in the primary outcome of re-bleeding (OR 1.30, 95% CI 0.88 to 1.91; seven RCTs).
- There were also no statistically significant differences in surgical interventions (OR 1.49, 95% CI 0.66 to 3.37; six RCTs), or mortality (RR 0.89, 95% CI 0.37 to 2.13; six RCTs).
I think this information is relatively well summed up in the American College of Gastroenterology’s PUD bleeding guideline (although they give a little too much credence to the non-patient centered “stigmata of hemorrhage”):
Laine L, Jensen DM. Management of patients with ulcer bleeding. Am J Gastroenterol. 2012;107:(3)345-60. PMID: 22310222
“Pre-endoscopic intravenous PPI (e.g., 80 mg bolus followed by 8 mg/h infusion) may be considered to decrease the proportion of patients who have higher risk stigmata of hemorrhage at endoscopy and who receive endoscopic therapy. However, PPIs do not improve clinical outcomes such as further bleeding, surgery, or death (Conditional recommendation).”
Another good summary of this evidence is evident in the UK National Institute for Health and Care Excellence.
Dworzynski K, Pollit V, Kelsey A, Higgins B, Palmer K, . Management of acute upper gastrointestinal bleeding: summary of NICE guidance. BMJ (Clinical research ed.). 344:e3412. 2012. PMID: 22695897 [full text guideline here] [most recent (2016) update here]
“Do not offer acid-suppression drugs (proton pump inhibitors or H2-receptor antagonists) before endoscopy to patients with suspected non-variceal upper gastrointestinal bleeding.”
So my bottom line:
Just stop using PPIs in undifferentiated GI bleeds.
Didn’t get enough from me talking about this?
This has been covered by a few other amazing skeptics:
Ken Milne of the Skeptic’s Guide to Emergency Medicine discusses PPIs in upper GI bleeds in Episode #16: Ho, Ho, Hold the Proton Pump Inhibitor in Upper Gastrointestinal Bleed
If you want more EBM topic reviews, click here.
Morgenstern, J. EBM Lecture Handout #1: PPIs for GI Bleeds, First10EM, June 16, 2015. Available at: