• 
• 

Three Part Question

• In [adult patients admitted in Intensive care units] are [proton pump inhibitors better than histamine-2 receptor antagonists] at [reducing stress -related upper gastrointestinal bleeding]?

Clinical Scenario

A 36 year old female patient is admitted to the ICU following a significant burn injury, with more than 50% of the body surface area involved, sustained in gas exploding accident. She is being managed in accordance with Burn injury guidelines but in addition, since she had an high risk factor for stress-related UGI bleeding, she was prescribed Esomeprazole 40 mg i.v od, for stress-related upper gastrointestinal bleeding prophylaxis. However, this was changed to Ranitidine 50 mg i.v tid, two days later on the advice of the Gastroenterologist.
Is there is any difference in benefit between the two drugs for the prevention of stress-related upper gastrointestinal bleeding?

Search Strategy

MEDLINE: 1966 - July (Week 3) 2011. Limit to adult Human patients and English articles.
("intensive care"[MeSH Terms] OR ("intensive"[All Fields] AND "care"[All Fields]) OR "intensive care"[All Fields]) AND ("proton pump inhibitors"[MeSH Terms] OR ("proton"[All Fields] AND "pump"[All Fields] AND "inhibitors"[All Fields]) OR "proton pump inhibitors"[All Fields] OR "proton pump inhibitors"[Pharmacological Action]) AND histamine-2[All Fields] AND ("Receptor"[Journal] OR "receptor"[All Fields]) AND ("antagonists and inhibitors"[Subheading] OR ("antagonists"[All Fields] AND "inhibitors"[All Fields]) OR "antagonists and inhibitors"[All Fields] OR "antagonists"[All Fields]) AND ("Stress"[Journal] OR "stress"[All Fields]) AND related[All Fields] AND ("upper gastrointestinal tract"[MeSH Terms] OR ("upper"[All Fields] AND "gastrointestinal"[All Fields] AND "tract"[All Fields]) OR "upper gastrointestinal tract"[All Fields] OR ("upper"[All Fields] AND "gastrointestinal"[All Fields]) OR "upper gastrointestinal"[All Fields]) AND ("hemorrhage"[MeSH Terms] OR "hemorrhage"[All Fields] OR "bleeding"[All Fields]) AND ("prevention and control"[Subheading] OR ("prevention"[All Fields] AND "control"[All Fields]) OR "prevention and control"[All Fields] OR "prophylaxis"[All Fields]).
Cochrane: “intensive care”, “proton pump inhibitors”, “histamine-2 receptor antagonists”, “stress related upper gastrointestinal bleeding prophylaxis”

Search Outcome

Medline - 5 papers found; 2 were considered relevant and critically appraised.
Cochrane – No relevant findings;

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Supot Pongprasobchai 2009 Thailand	Three RCT’s included in the meta-analysis, with a total of 569 adult critically-ill patients admitted in ICU’s, with any of the two risk factors (mechanical ventilation > 48h or coagulopathy);	1a: Meta-analysis of RCTs;	Comparing proton pump inhibitors (PPI- 282 patients) vs. Histamine-2 receptor antagonists (H2RA- 287 patients) for stress related mucosal disease prophylaxis. Outcomes of interest were incidences of clinically important gastrointestinal bleeding and nosocomial pneumonia;	Overall incidence of clinically important bleeding was significant lower in the PPI group (3,5%), as compared to H2RA (8%), odds ratio 0,42 (95% CI 0,20-0,91). The absolute risk reduction of clinically important bleeding of PPI was 4,5%, with a NNT 22.The incidences of Nosocomial pneumonia were not different (10,2% vs. 10,1%, OR 1,02, 95% CI 0,59-1,75) between the two groups;	Low number of patients; The studies included different high risk patients and used different definitions of clinically important bleeding and nosocomial pneumonia; All three studies compared the same type and dosage PPI (omeprazole 40 mg per day) with different H2RA (ranitidine, famotidine and cimetidine), known to have different efficacies on UGI bleeding prophylaxis; One of the studies included, Levy et al, 1997, the one that mainly contributed to the superior efficacy of the PPI group, had important pitfalls in patient randomization: patients were more severe in the H2RA group than those in the PPI one; so the H2RA one, had a very high incidence of clinically important bleeding (31%), too high comparing with previous largest study known;

Comment(s)

Critically ill patients are at increased risk of developing stress-related mucosal lesions, a source of significant morbidity and mortality. Within the first 24 hours of admission to an ICU, 75% - 100% of critically ill patients will demonstrate evidence of stress-related mucosal damage (Ali T, Harty RF, 2009). The pathogenesis of stress-related mucosal disease is not entirely clear, but is often multifactorial, involving splanchnic vasoconstriction and hypoperfusion with altered mucosal blood flow, acid back-diffusion with reduction of bicarbonate secretion, and changes in gastrointestinal motility. Recent observational studies suggest that bleeding from stress ulceration is extremely uncommon in intensive care unit patients, although when it occurs, it is associated with prolonged ICU stay, increased costs and mortality (Carlson RW, et al, 2008). Based on a large, multi-center, prospective trial of critically ill patients (published by Cook et al, 1994), patients with coagulopathy (platelets < 50,0000, International Normalized Ratio < 1.5 or Partial Thromboplastin Time ≥ 2 the control value) or requiring mechanical ventilation for a period exceeding 48 hours, were considered strong independent risk factors for stress ulcer-related bleeding. Other potential risk factors, where the clinician may feel the patient is at higher than usual risk for stress bleeding (e.g. Sepsis, CNS injury, severe multi-system trauma, prior history of gastroduodenal ulcers, burns, etc.), consideration of stress bleeding prophylaxis is also warranted (Cook DJ, Fuller HD, et al, 1994). So currently, stress ulcer prophylaxis (SUP) is regarded a standard of care in patients admitted to the ICU. Estimates indicate that approximately 90% of critically ill patients admitted to the ICU receive some form of SUP (Marik PE, Vasu T, et al, 2010). Many prophylactic regimens have been developed, utilized, and debated. Several medicinal agents such as Sucralfate (do not raise pH and decrease the absorption of concomitantly administered oral medications), Histamine-2 receptor antagonists (raise gastric pH, but develop tolerance, possible drug interactions and neurologic manifestations) and Proton pump inhibitors (PPI – most potent acid-suppressive agents available, that raise gastric pH ≥ 4, up to 24h after a single dose; adverse effects uncommon) have been used, but none has been clearly shown to be superior to the others (Quenot JP, Thiery N, Barbar S, 2009). Guidelines by American Society of Health-system Pharmacists and the Surviving Sepsis Campaign, suggested that high risk patients should receive stress ulcer prophylaxis with histamine-2 receptor antagonists (H2RA), although definite recommendation regarding the choice between agents cannot be made based on available evidence (Lin Pei-Chin, Chang C-H et al, 2010). On the other hand, proton pump inhibitors (PPI’s) have become first-line therapy (as in most acid-related GI disorders) in an increasing percentage of critical care patients, despite it never demonstrated reducing the rate of bleeding from stress ulceration, (Supot Pongprasobchai, et al, 2009). However, we must not forget that if SUP is used indiscriminately, it can be expensive and clinically detrimental, with increased risk of hospital-acquired pneumonia and clostridium difficile infection (Marik PE, Vasu T, et al, 2010). This takes us to the three part question I’ve focused on and to the two meta-analyses that I have critically appraised (level of evidence 1a): In the first one, I found an important clinical question which the reviewers addressed. A comprehensive and systematic literature search was undertaken concerning appropriate data bases and other potentially important sources (reference lists of systematic reviews and identified articles). Randomized and quasi-randomized trials were included, that compared PPI and H2RA in the prophylaxis of stress-related UGI bleeding in adult critically-ill patients. Two independently reviewers evaluated each identified study and abstracted relevant characteristics, including quality of the studies (assessed by Jadad criteria). Differences in opinion between the two reviewers were solved by consensus agreement. It was not appropriate to obtain ethical approval. All of the subjects were accounted for (7 RCT’s suitable for inclusion) and all the appropriate outcomes considered. A statistical analysis was used to estimate the pooled risk differences and 95% confidence intervals (CI) of the efficacy and safety outcomes between PPI and H2RA arms from eligible studies. Heterogeneity was assessed using Cochrane’s Q statistic and quantified using the I² statistic. To further explore the potential source of heterogeneity, a sensitivity analysis (one study removed at a time) and subgroup analysis (took into account the following variables: probability of ventilation need > 48h, presence of balanced patient characteristics at baseline, year of publication-before or after 2000) was also done. A formal test of interaction was performed for each subgroup: results for interactions were considered statistically significant for p< 0,05. The assessment of publication bias was not forgotten, and so funnel plots were examined for asymmetry. The basic data and the differences between studies were adequately described. The results were presented clearly, objectively and in sufficient detail to enable readers to make their own judgment. The results discussed in relation to existing knowledge on the subject. However, there are some study weaknesses to consider, specially: • Limited trial data (only 7 RCT’s with 936 patients); • 3 papers with poor quality (Jadad score 1); • Significant heterogeneity in all seven studies (Chi² value of 17,43, much greater than the number of trials in a meta-analysis, tells us that the trials which contributed to the analysis are different in some important way from one another), but especially in one of them, by Levy et al (1997), the one that mainly contributed to the superior efficacy of the PPI group, with regard to the risk difference of stress-related UGI bleeding (higher for patients in the histamine-2 receptor antagonists group than those in the PPI group: 2,7 vs. 1,9; therefore the stress related clinically significant bleeding occurred at a higher rate, 31%, in those patients receiving ranitidine as compared to omeprazole, 6%). In the sensitivity analysis, once the study by Levy was removed, the heterogeneity reduced significantly, from I² 66 to 26% and shifted the overall risk difference closer to null (pooled risk difference, - 0.02; 95% confidence interval, -0.05– 0.01; p .19); • Only 5 trials reported balanced patient characteristics across patient populations at baseline between two groups after randomization: another reason found for heterogeneity; • Presence of language and publication bias (funnel plot asymmetry); • Nearly 40% patients included in the study were at low risk for stress related UGI bleeding; • ICU mortality was only reported in 3 trials; The author’s conclusion stated: “this meta-analysis did not find strong evidence that proton pump inhibitors were significantly different than histamine-2 receptor antagonists in terms of stress-related important UGI bleeding prophylaxis, pneumonia and mortality in ICU admitted patients”. During my search strategy, I also came across another meta-analysis (level of evidence 1a), by Pongprasobchai S et al (2009), that included Three RCT’s, all mentioned previously in the first meta-analysis discussed: Levy et al (1997); Kantorova et al (2004) and Conrad et al (2005), with a total of 569 adult critically-ill patients admitted in ICU’s, with any of the two following risk factors: mechanical ventilation > 48h or coagulopathy. It Compared proton pump inhibitors (PPI- 282 patients) vs. Histamine-2 receptor antagonists (H2RA- 287 patients) for stress related mucosal disease prophylaxis. Outcomes of interest were incidences of clinically important gastrointestinal bleeding and nosocomial pneumonia. In this paper the overall incidence of clinically important bleeding was significant lower in the PPI group (3,5%), as compared to H2RA (8%), odds ratio 0,42 (95% CI 0,20-0,91); if only another study was excluded (2 RCT’s instead of 3), this statistical difference would no longer be present. The absolute risk reduction of clinically important bleeding of PPI was 4,5%, with a NNT 22. The incidence of nosocomial pneumonia was not different (10,2% vs. 10,1%, OR 1,02, 95% CI 0,59-1,75) between the two groups. Concerning some weaknesses of this study: • All three studies included relatively low number of patients; • Studies included different high risk patients and used different definitions of clinically important bleeding and nosocomial pneumonia, and might somehow affected the reported incidence of outcomes of interest in the present meta-analysis; • All three studies compared the same type and dosage PPI (omeprazole 40 mg per day) with different H2RA (ranitidine, famotidine and cimetidine), known to have different efficacies on UGI bleeding prophylaxis; • PPI more effective than H2RA in UGI bleeding prophylaxis: odds ratio was only 0,42, so clearly under than 1, with an absolute risk reduction of clinically important bleeding of PPI was 4,5%, with a NNT 22; Apart from that, the author’s stated in the study’s conclusion that: “PPI is superior to H2RA in the prevention of clinical important stress–related UGI bleeding, with a similar rate of nosocomial pneumonia”. Although it included 3 RCT´S referred in first paper I clinically appraised, due to the fact it was a recent meta-analysis (2009), with the conclusion stated above different from the first meta-analysis discussed, that I chose to refer this meta-analysis in this Best Bet. In conclusion, there were very few randomized controlled clinical trials, based on a small number of patients, regarding stress ulcer prophylaxis. Therefore a steadfast conclusion cannot presently be reached. More future randomized, multicenter, powerful and well designed clinical trials, with sufficient sample size and emphasized on patient risk stratification and optimal regimen (drug, dosage, mode of administration, timing and duration among patients with different risk levels) of stress ulcer bleeding prevention, are warranted. Until then, we cannot firmly conclude that PPI is superior to H2RA for UGI bleeding prophylaxis. Furthermore, the chosen prophylactic regime, should take into account the risk factors and underlying disease state of individual patients to provide the best therapy to those most likely to benefit.

Clinical Bottom Line

The evidence of these meta-analyses remains inconclusive. Routine prophylaxis against stress ulcers in all ICU patients is not well justified by current evidence; only the ones at risk of stress ulcer-related bleeding are most likely to benefit from prophylaxis. There is no strong evidence suggesting which anti-ulcer regime is preferred, and so in adult patients admitted in Intensive care units, PPI wasn’t better than histamine-2 receptor antagonists, at reducing stress -related upper gastrointestinal bleeding.

References

1. Lin Pei-Chin, Chang C-H et al The efficacy and safety of proton pump inhibitors vs. histamine-2 receptor antagonists for stress ulcer bleeding prophylaxis among critical care patients: A meta-analysis. Critical Care Medicine 38(4): 1197-1205.
2. Supot Pongprasobchai, et al Proton Pump Inhibitors for the prevention of Stress-Related Mucosal Diseases in Critically-Ill patients: A Meta-Analysis. J Med Assoc Thai 92(5): 632-637