The PEPTIC study is a randomised controlled trial of a Proton Pump Inhibitor (PPI) vs. Histamine-2 Receptor Blockers for ulcer prophylaxis on the ICU. Think Omeprazole vs. Ranitidine as your choice of prophylaxis. Although there are a number of papers and systematic reviews around on this subject, they are far from conclusive1 and practice remains inconsistent1–3. This is despite stress ulcer prophylaxis being one of the most common interventions prescribed to ICU patients4–6.
PPIs are used in many patients, but they have interesting effects, notably that they are immunosuppressants, may increase C.diff infections and perhaps increase pneumonia. To date no large RCTs have been performed on this topic.
You might think that this is a rather small issue, and not that an exciting one, but because so many ICU patients get these drugs, even a small treatment effect could have a large impact.
This study is being presented at the critical care reviews conference in Belfast, January 2020 and we will be updating the blog as the results come out. You can follow the live presentation on the link below.
The full paper is now published on the JAMA website7 and is open access at the moment. As we always say please go read the full paper yourself.
What kind of study is this?
As this is an intervention trial comparing two drugs we really want to see an individual patient randomised controlled trial but that’s not what we’ve got. In this study units used one medication for for 6 months and then flipped over to the other medication. It’s unclear to me why this was done instead of a traditional RCT. This multi-centre cluster randomised trial (the units were randomised into which order they would prescribe) has advantages in terms of convenience and structure of the trial, but I’m unconvinced that it is a superior design. The authors present a rationale for their design in the trial protocol which argues that the intervention would be delivered to a protocol and within a consecutive sample if instituted at department level. Whilst this is a reasonable perspective I don’t think it’s as powerful as an individual patient randomisation method8 and this was taken up by the audience and in the results. Paul Young argues that cluster randomisation has benefits in reducing the ‘noise’ in the data from clustering effects. As it happens, Paul argued that the results they found meant that they had significant power in the trial even if this had been designed as an individual patient randomisation process.
In general cross over trials do not have the same level of blinding/masking as individual patient RCTs. That’s the case here as the trial was an open label design meaning that the clinicians, and perhaps the patients too, knew what they were getting. Cluster randomisation also introduces additional challenges of unknown departmental confounders creeping in to your results. However, the pragmatic characteristics mean that they are easier and cheaper to perform and pretty much guarantee you a consecutive sample of patients, strengthening the generalisability of the result.
Tell me about the patients.
Consideration of stress ulcer prophylaxis is pretty standard ICU treatment here and in most places, although it is worth highlighting that patients on full enteral feed can often avoid it9. Patients aged >18 at risk of stress ulcers or erosions are often prescribed prophylaxis, but regimes and strategies still differ and controversy remains about which patients benefit the most from treatment9. In this trial, the higher risk patients (such as those who already had GI bleeding) were excluded.
What did they do?
Each study ICU used either a PPI or H2RB as routine therapy for a period of six months. At the end of this six month period, the ICU then swapped over to the opposite routine ulcer prophylaxis strategy for another 6 months. The choice, dosage and route of administration of the PPI/H2RB was left to the discretion of the treating physicians/units. Patients were recruited in Australia, New Zealand, Canada, Ireland and the UK. This was an open label trial.
Data was collected from ICU registries and local collection. They also used local audits and pharmacy records to look at whether units were adhering to the randomised therapies.
What were the outcomes?
The primary outcome was in-hospital mortality (censored at 90 days)8. Secondary outcomes looked at evidence for GI bleeds, transfusion requirements and haematological parameters.
Data was reported as confidence intervals rather than p-values which has been a bit of a theme through the conference. This trial is particularly good at illustrating that as depending on the analysis a whole range of p-values were possible.
What did they find?
The PEPTIC study recruited 16982 patients across 50 ICUs in 5 countries. Patients were fairly typical ICU fare, pretty sick and the usual variety of pathologies.
In terms of adherence to protocol was variable. There were quite a few patients who got the other drug and/or both. This is a concern in this study as large numbers of patients/units clearly did not adhere to protocol. We don’t know why this is but may be down clinicians making informed judgements about what to give patients. This could have contaminated the results.
The main result is that 18.3% died in the PPI group and 17.5% died in the H2RB group. That’s a relative risk of 1.05. It does not reach statistical significance (p=0.54 – reported by JAMA but not in the presentation), but when looking at the confidence intervals it seems very unlikely that PPIs are better than H2RBs.
The absolute risk difference is 0.93 with 95% confidence intervals is 0.01% to 1.88%
In terms of the secondary outcomes for GI bleeding the risk ratio was 0.73 (95% CI 0.57 to 0.92) in favour of PPI. Pneumonia was higher in the PPI group 6.5% vs. 5.8%. Risk ration 1.18 (but CIs cross 1).
The authors found (in a reviewer directed sub group analysis) that there was greater treatment effects in favour of H2RBs in patients who were sicker. They also found a reduced level of GI bleeding in the PPI group.
Paul Young’s conclusion is that we should shift to H2RBs, arguing that the results are significant. Interestingly, and perhaps because of the wider disagreement about how we interpret trials, @jama_current does not seem to agree.
It was interesting to see how many people thought that the NNTs (about 125) are really high. I don’t think people understand the treatment effect of many interventions. An NNT of 125 is not at all untypical.
We should also note that these are wquite short term outcomes.
If you want to follow Paul Young’s interpretation, remembering that as lead author he is both better informed than anyone else, but also biased (that’s why we have critical appraisal) then follow the sequence of tweets below.
Clinical Bottom Line.
The PEPTIC study data suggests that we should use H2RBs rather than PPIs. However, there are some concerns in the methodology and the effect size is of questionable. It does look unlikely that PPIs would be found to be better than H2RBs.
My early thoughts are that this trial gives us a weak recommendation to use H2RBs. This is based on the balance of probabilities rather than absolute certainty, despite the results from the largest RCT in the history of critical care.
- 1.Litton E, Eastwood G, Bellomo R, et al. A multicentre feasibility study evaluating stress ulcer prophylaxis using hospital-based registry data. Crit Care Resusc. 2014;16(3):158-163. https://www.ncbi.nlm.nih.gov/pubmed/25161016.
- 2.Eastwood G, Litton E, Bellomo R, et al. Opinions and practice of stress ulcer prophylaxis in Australian and New Zealand intensive care units. Crit Care Resusc. 2014;16(3):170-174. https://www.ncbi.nlm.nih.gov/pubmed/25161018.
- 3.MacLaren R, Reynolds P, Allen R. Histamine-2 receptor antagonists vs proton pump inhibitors on gastrointestinal tract hemorrhage and infectious complications in the intensive care unit. JAMA Intern Med. 2014;174(4):564-574. doi:10.1001/jamainternmed.2013.14673
- 4.Alhazzani W, Alenezi F, Jaeschke RZ, Moayyedi P, Cook DJ. Proton Pump Inhibitors Versus Histamine 2 Receptor Antagonists for Stress Ulcer Prophylaxis in Critically Ill Patients. Critical Care Medicine. March 2013:693-705. doi:10.1097/ccm.0b013e3182758734
- 5.Krag M, Perner A, Wetterslev J, et al. Prevalence and outcome of gastrointestinal bleeding and use of acid suppressants in acutely ill adult intensive care patients. Intensive Care Med. 2015;41(5):833-845. doi:10.1007/s00134-015-3725-1
- 6.Wang Y. Efficacy and safety of gastrointestinal bleeding prophylaxis in critically ill patients: systematic review and network meta-analysis. BMJ. https://www.bmj.com/content/368/bmj.l6744. Published 2020. Accessed 2020.
- 7.Young P. Effect of Stress Ulcer Prophylaxis With Proton Pump Inhibitors vs Histamine-2 Receptor Blockers on In-Hospital Mortality Among ICU Patients Receiving Invasive Mechanical Ventilation. JAMA. https://jamanetwork.com/journals/jama/fullarticle/2759412?guestAccessKey=e3b3c18c-4789-41e8-ab8e-eccab996237f&utm_source=twitter&utm_medium=social_jama&utm_term=3040185957&utm_campaign=article_alert&linkId=80772466. Published 2020. Accessed 2020.
- 8.Young P, Bagshaw S, Forbes A, et al. A cluster randomised, crossover, registry-embedded clinical trial of proton pump inhibitors versus histamine-2 receptor blockers for ulcer prophylaxis therapy in the intensive care unit (PEPTIC study): study protocol. Crit Care Resusc. 2018;20(3):182-189. https://www.ncbi.nlm.nih.gov/pubmed/30153780.
- 9.Nickson C. Stress Ulcer Prophylaxis. LITFL. https://litfl.com/stress-ulcer-prophylaxis/. Published 2019. Accessed 2019.