Upper gastrointestinal bleeding in critically ill patients: proton-pump inhibitors, histamine-2 receptor antagonists or placebo? Many questions remain unanswered

This article refers to:

Prevention of upper gastrointestinal bleeding in critically ill Chinese patients: a randomized, double-blind study evaluating esomeprazole and cimetidine

Critically ill patients admitted to an intensive care unit (ICU) can suffer stress-related mucosal disease (SRMD) due to several factors such as hypovolemia, hypotension, increased levels of catecholamines and vasoconstriction. These factors produce splanchnic and mucosal hypoperfusion, which in turn causes decreased gastric mucosal blood flow and local ischemia¹.

A high frequency of mucosal gastric lesions is reported in the ICU by endoscopic studies. However, the incidence of bleeding from SRMD is much lower nowadays, occurring in only around 1–6% of patients. Trials from 1980 to 2000 reported an incidence about 15% for stress-ulcer bleeding, although in studies from the past 10 years the incidence of bleeding from SRMD has notably decreased. For example, an international study published in 2015 showed a frequency of clinically important bleeding from SRMD in ICU of 2.6%².

This low frequency is due, among other reasons, to better management of patients, which reduces hypoperfusion states. This includes improved hemodynamic monitoring, better control of mechanical ventilation and probably the wider use of stress-ulcer prophylaxis (SUP), which is now extensively used in patients in the ICU who have risk factors for bleeding from SRMD. The risk factors include mechanical ventilation for more than 48 hours and the presence of a coagulopathy³. Proton-pump inhibitors (PPIs) and histamine-2 receptor antagonists (H2RAs) are the two drugs most widely used for SUP. In the past few years, these drugs have been the subject of controversy over which of the two is better, although, as we note below, the balance is tilting in favor of PPIs.

In an effort to clarify this controversy, Lou and colleagues conducted an interesting and well designed, randomized, double-blind, non-inferiority study across 27 ICUs in China⁴. In their multicenter study, 274 patients requiring mechanical ventilation for an anticipated minimum time of 48 hours, having a gastric tube in place and having at least one additional risk factor for stress-ulcer bleeding were randomized to receive active esomeprazole 40 mg intravenous infusion twice daily and placebo cimetidine, or active cimetidine 50 mg/hour as a continuous infusion following an initial bolus of 300 mg and placebo esomeprazole. The primary endpoint was significant upper gastrointestinal (GI) bleeding determined by evaluating the content of the gastric tube. A total of 2.7% of patients in the esomeprazole group and 4.6% of patients in the cimetidine group had significant upper GI bleeding. The absolute risk difference for patients in the esomeprazole arm compared with those in the cimetidine arm was 1.9% (95% CI: -2.8, 6.5). The authors concluded that esomeprazole is not inferior to cimetidine for the prevention of significant upper GI bleeding.

In the past few years, the evidence shows greater effectiveness with PPIs compared to H2RAs in SUP. In fact, four meta-analyses of available clinical trials^5–8 showed that PPIs are more effective than H2RAs in reducing the risk of clinically significant GI bleeding in the ICU. Therefore, some guidelines^9,10 though not all¹¹, suggest the use of PPIs rather than H2RAs for SUP, and surveys of intensivists show a greater use of PPIs over H2RAs¹². In addition, PPIs have been proven to be superior to H2RAs in preventing upper GI injuries in other contexts, such as in patients taking nonsteroidal anti-inflammatory drugs or low-dose aspirin¹³. Intragastric clotting mechanisms are severely affected by acidic pH below 5.8¹⁴, and PPIs are able to increase the intragastric pH above that limit, while H2RAs do not¹⁵.

On the other hand, esomeprazole has been shown to be more effective in achieving gastric acid suppression than other PPIs¹⁶. Taking into account all of the above, and considering that only 2% of all non-inferiority trials conclude that the intervention drug is inferior¹⁷, would someone think that the result of the Lou and colleagues trial would be that esomeprazole is inferior to cimetidine? We honestly think not, and that it can rather be seen as a self-fulfilling prophecy.

In addition, 90 patients (66%) in the esomeprazole arm and 102 patients (74%) in the cimetidine arm had mechanical ventilation removed after 24 hours and reached the predefined stopping criteria. Therefore, the mean study drug treatment duration for both drugs was 6 days, which makes it difficult to find differences between the treatments.

Two aspects should be highlighted in the Lou and colleagues trial⁴ and other similar studies^18,19. One aspect is the primary endpoint. We believe that the ideal outcome should be survival rate. Lou and colleagues and other authors^4,18,19 have used the definition of persistent blood or positive occult blood as determined by evaluating the content of a gastric tube, but additional conditions such as a decrease in hemoglobin greater than 2 g/dL or presence of melena should be included in this endpoint since there are clear risks of considering clinically irrelevant bleedings as true events.

The study by Lou and colleagues⁴ also shows that patients had a significant number of bleeding events reported as melena or hematochezia that were eventually considered to be upper GI bleeding events. It is unclear whether these events were actually upper (esophageal, gastric or duodenal) or lower (small bowel or colorectum) GI events. Perhaps GI bleeding events as outcomes should be considered for the whole GI tract, since all types (upper and lower) of bleeding can affect mortality and the use of “gastroprotectants” may shift the bleeding towards the lower GI tract²⁰. This reinforces the idea of using survival outcomes as the main endpoint in any new multicenter study if we want to provide solid evidence on this issue.

Another aspect that needs to be outlined here is the arm with which PPIs, and esomeprazole in this case, should be compared. We believe that before considering whether PPIs are better than, equal to or worse than H2RAs, we still need to confirm that PPIs are better than placebo. A recent pilot study, which was a randomized controlled trial (RCT) with pantoprazole and a meta-analysis of five trials with PPIs (604 patients) concluded that these compounds were not better than placebo in preventing upper GI bleeding, infections or mortality²¹.

These results bring forward another relevant aspect regarding the adverse events associated with drug use. In recent years, concerns have been raised on the side effects associated with PPIs in the ICU, mainly with Clostridium difficile infection and pneumonia. In the largest RCTs conducted in critically ill patients and meta-analyses of RCTs, significant differences have not been observed in pneumonia rates between patients treated with PPIs versus H2RAs²², but no RCTs have reported on nosocomial C. difficile infections in the ICU. Only a few observational studies have suggested an increased risk of acquired C. difficile infection with PPIs when compared to H2RAs²³. In the trial by Lou and colleagues⁴, the most commonly reported adverse event term was lung infection, which occurred in 19.0% of patients in the esomeprazole arm and in 20.3% of patients in the cimetidine arm. Adverse events related to study drugs were 12.2% in patients in the esomeprazole group and 15.7% in patients in the cimetidine group. Authors did not report what specifically these side effects were and no data on C. difficile infection was noted.

In summary, despite the key unresolved questions mentioned above, current evidence shows that the risk to benefit ratio of SUP with PPIs seems positive^1,22,24. However, today, even more important than the type of drug used for SUP, the correct and appropriate identification of patients at risk of SRMD who need prophylaxis seems a more urgent task. One study showed that, among patients with no risk factors for SRMD, 68.1% were placed on SUP at ICU admission and 60.4% continued on treatment upon transfer from the ICU²⁵. This is a similar picture to that seen with other PPI indications²⁶. The most widely recognized risk factors for SRMD are respiratory failure requiring mechanical ventilation for more than 48 hours, the presence of a coagulopathy, maximum serum creatinine level and liver injury. Other risk factors are greater than or equal to three coexisting diseases, higher organ failure score, sepsis, major trauma and high-dose corticosteroids, among others^2,3,27. Only patients with these risk factors should receive SUP since, otherwise, the risk to benefit ratio of SUP can be negative. The latest guidelines reflect this recommendation¹¹, and efforts are necessary to raise awareness among clinicians about this problem.

Transparency

Declaration of funding

This work was partially supported by the Official Spanish Agency, Instituto de Salud Carlos III [FIS grant PI08/1301] and the CIBERehd. Dr. Garcia-Rayado is also supported by grant from the Instituto de Salud Carlos III and IIS Aragón.

The contents of the paper and the opinions expressed within are those of the authors, and it was the decision of the authors to submit the manuscript for publication.

Declaration of financial/other relationships

G.G.-R. and A.L. have disclosed that they have no significant relationships with or financial interests in any commercial companies related to this study or article.

CMRO peer reviewers on this manuscript have no relevant financial or other relationships to disclose.