Pharmacist Continuing Education: Stress ulcer prophylaxis: Is it still necessary?
Author: Michael Serlin, PharmD; PGY1 Pharmacy Practice Resident, SSM Health St. Clare Hospital
Preceptor: Christopher Carter, PharmD, BCCCP; Clinical Pharmacy Specialist - Critical Care, SSM Health St. Clare Hospital
Program Number: 2018-07-18Approval Dates: October 1, 2018 to December 31, 2018Approved Contact Hours: One (1) CE(s) per LIVE session.
Acid-reducing medications like proton pump inhibitors (PPIs) and histamine receptor 2 antagonists (H2RAs) are used commonly for stress ulcer prophylaxis in the inpatient setting. Stress ulcer prophylaxis is indicated for certain subgroups within the critically ill population but up to 71% of general medicine patients receive stress ulcer prophylaxis.1 Though it may reduce bleeding due to stress ulcer formation, stress ulcer prophylaxis may also expose patients to an increased risk of Clostridium difficile infections and ventilator-associated pneumonia due to the creation of a higher-pH environment within the stomach and proximal, small bowel.2
The goal of this article is to provide information to assist clinicians in exercising stress ulcer prophylaxis stewardship in the inpatient setting by providing adequate background on the risks for stress ulcer formation and a review of the pertinent literature assessing the safety and efficacy of stress ulcer prophylaxis. By practicing good stewardship, clinicians can optimize the benefit of stress ulcer prophylaxis against the risk of adverse events in the appropriate patient populations.
Stress Ulcer Formation and Enteral NutritionAs previously stated, stress ulcer prophylaxis is only indicated in certain critically ill patients. Critically ill patients have a variety of physiologic changes that can increase risk for ulceration. Compared to non-critically ill patients, critically ill patients have an increased gastric acid secretion and impaired blood flow to the gastric mucosa.3 Impaired blood flow to the mucosa can be caused by hemodynamic changes (i.e. hypotension), an imbalance of endothelin-1 and nitric oxide, as well as positive end-expiratory pressure from mechanical ventilation.3 This impaired blood flow can ultimately decrease production of bicarbonate and mucus that would normally provide a protective barrier in the stomach and increase ulceration risk.
Although the above physiologic changes can increase the risk of ulceration, it is important to note that altered gastric mucosa does not necessarily cause bleeding. Generally, in order for bleeding to occur, there must be a change in the gastric mucosa and an increased gastric acid secretion.3
While medications can be used to modify acid secretion, enteral nutrition has been shown to increase gastric mucosal blood flow, as well as modify the imbalance between endothelin-1 and nitric oxide.3
Types of Upper Gastrointestinal BleedingThere are three types of upper gastrointestinal bleeds: occult, overt, and clinically important. An occult bleed is a type of bleed that isn’t symptomatic or visibly seen. An example of occult bleeding is a positive fecal bleeding test. Overt bleeding is a type of bleed that can be visibly seen, such as hematemesis, bloody nasogastric aspirate, or melena. Clinically important bleeding is defined as overt bleeding plus: a spontaneous decrease of greater than 20 mmHg in systolic or diastolic blood pressure, an orthostatic increase in heart rate by 20 beats per minute and a decrease in systolic blood pressure by 10 mmHg, or a decrease in hemoglobin by at least two grams per deciliter or the need to transfuse at least two units of blood within 24 hours of a bleed.4 Clinically important bleeding is associated with increased mortality and increased intensive care unit (ICU) length of stay and is the main target of stress ulcer prophylaxis.4 The incidence of clinically important bleeding in ICU patients without prophylaxis has historically been from 0.1-4%.4
1999 ASHP Stress Ulcer Prophylaxis GuidelinesThe American Society of Health-System Pharmacists’ stress ulcer prophylaxis guidelines, published in 1999, recommended that stress ulcer prophylaxis be utilized in certain critically ill populations including those that are mechanically ventilated for at least 48 hours and those that have a coagulopathy, including platelet counts less than 50,000 cells/mL3, international normalized ratios (INRs) greater than 1.5, and activated partial thromboplastin time (aPTT) greater than twice the upper limit of normal.5 These recommendations were based on evidence from cohort studies. Of note, there are two other indications for stress ulcer prophylaxis in the ASHP guidelines that were founded upon expert opinion only. These indications include patients that have had a gastrointestinal bleed within the past year and patients that have at least two of the following characteristics: Diagnosis of sepsis, ICU stay longer than six days, and those with occult bleeds lasting longer than five days.5
The ASHP guidelines also provided recommendations for special populations. Based on randomized control trials, stress ulcer prophylaxis is recommended for patients with Glasgow Coma Scale scores less than or equal to 10 and patients with thermal injuries that were greater than 35% of the patient’s body surface area (BSA). Certain populations, including head traumas and spinal cord injuries, had not been well studied when the guidelines were published. Thus, the guidelines stated that prophylaxis might be indicated on the basis of expert opinion.5
Recently Published Stress Ulcer Prophylaxis LiteratureSelvanderan, et al.- 2016 exploratory trial7The article entitled “Pantoprazole or Placebo for Stress Ulcer Prophylaxis (POP-UP): Randomized Double-Blind Exploratory Study” by Selvanderan et al. was a prospective, randomized, parallel-group trial. This exploratory study aimed to establish rates of clinically important bleeding, Clostridium difficile infections, and ventilator-associated pneumonia in critically ill patients with and without prophylaxis of intravenous pantoprazole 40 mg daily. Patients were eligible for the study if their expected time on mechanical ventilation was at least 24 hours and if they were expected to receive enteral nutrition within at least 48 hours of admission. Patients were ineligible for enrollment if they had used acid-altering medication before admission, were admitted with a gastrointestinal bleed, were using at least 100 mg of prednisolone or an equivalent dose of another corticosteroid, were admitted for gastrointestinal or cardiovascular surgery, were pregnant, were Jehovah’s witnesses, were receiving palliative care, or were an ICU readmission. Categorical data were appropriately analyzed via chi-squared or Fisher’s exact test. Parametric continuous data were analyzed via unpaired t-test and nonparametric continuous data were analyzed via Mann-Whitney U.
There were no significant differences in baseline characteristics or processes of care, including inotrope usage or coagulopathies, in the 108 patients that received placebo and the 106 patients that received pantoprazole. There was also no significant difference in the number of patients using enteral nutrition in each group. All of the endpoint studied, including clinically important bleeding, ventilator-associated pneumonia, Clostridium difficile infection, overt bleeding incidence, ICU length of stay, and hospital length of stay were found to be nonsignificant between the groups. There were no incidences of clinically important bleeds observed. The incidence of ventilator-associated pneumonia was 1.9% in the pantoprazole group and 0.9% in the placebo group. The incidence of Clostridium difficile infection was 0.9% in the pantoprazole group and 0% in the placebo group.
These results show that in this population, pantoprazole provided no extra benefit and no harm. These nonsignificant results should be taken in context of the exploratory nature of the trial in that since power was not assessed in the study design, it is possible that the trial was not designed to find a difference if one truly exists.
Alhazzani, et al.- 2017 pilot trial and meta-analysis2The article entitled “Withholding pantoprazole for stress ulcer prophylaxis in critically ill patients: a pilot randomized clinical trial and meta-analysis” by Alhazzani et al. was an international, multicenter, randomized, double-blind, placebo-controlled, parallel-group trial that compared intravenous pantoprazole 40 mg daily to placebo. The study treatment was stopped after death, incidence of gastrointestinal bleeding, or after extubation at the discretion of the treating physician. Patients were included if they were at least 18 years old and were expected to be mechanically ventilated for at least 48 hours. Patients were excluded if they were ventilated before randomization, used PPIs due to an active bleed or for an increased bleeding risk, used dual antiplatelet therapy, were on palliative care, were pregnant, or used PPIs or H2RAs twice daily or more. The primary endpoint was incidence of clinically important gastrointestinal bleeding. The secondary endpoints were similar to those in the trial by Selvanderan et al., including length of stay, incidence of ventilator-associated pneumonia, incidence of Clostridium difficile infection, and morality rates. The trial data was analyzed via Fisher’s exact test and the meta-analysis data was analyzed via DerSimonian and Laird random-effect models.
There were no significant differences noted in baseline characteristics including H2RA or PPI use before the study, median Acute Physiology and Chronic Health Evaluation II (APACHE II) score, and type of anticoagulants used. The only statistically significant difference noted between the 49 patients in the pantoprazole group compared to the 42 patients in the placebo group was the difference in median hospital length of stay, favoring the placebo group (placebo 25 days (interquartile range, IQR 10-42 days) vs. pantoprazole 27 days (IQR 16-38 days), p=0.049). All other differences in pilot trial results, including differences in mortality, clinically important bleeding incidences, and any bleeding incidences were not found to be statistically significant.
The meta-analysis performed by Alhazzani et al. reviewed data on incidence of clinically important bleeding and ventilator-associated pneumonia. For the incidence of clinically important bleeding, the authors found no overall significance in the differences between PPIs vs. placebo in the pooled data from five studies (overall OR 0.96; 95% CI 0.24-3.82). The authors also found no statistically significant difference in the pooled data from four studies for ventilator-associated pneumonia (overall OR 1.32; 95% CI 0.68-2.55).
The results from the pilot trial and the meta-analysis both show no statistically significant differences between PPIs and placebo for clinically important bleeding and infection rates. Due to the pilot nature of the trial aspect of this article, these nonsignificant results should be interpreted with caution, as power was not assessed or calculated.
El-Kersh, et al.- 2017 exploratory trial8The article entitled “Enteral nutrition as stress ulcer prophylaxis in critically ill patients: A randomized controlled exploratory study” by El-Kersh et al. was a prospective, randomized controlled trial that aimed to determine if stress ulcer prophylaxis is needed in combination with enteral feeding. Enteral nutrition, defined as 25-30 kcal per kilogram per day, was given with intravenous pantoprazole 40 mg daily or placebo. To provide appropriate feeding, gastric residual volumes were checked every four hours. If the residual volume was greater than 400 milliliters or if the patient was hypotensive or hypoxic, feeding was paused to decrease the risk of bowel ischemia. Feeding was resumed once the residual volume was less than 400 milliliters or when deemed appropriate by the treating medical teams. The feeding strategy used in the study aligns with the recommendations provided by the Society of Critical Care Medicine and the American Society for Parenteral and Enteral Nutrition.9 Patients were eligible for enrollment if they were at least 18 years old, expected to be ventilated for at least 48 hours, and had no contraindications for enteral nutrition within 24 hours of ICU admission. Patients were ineligible if they had a gastrointestinal bleed or burn injury upon admission, had head trauma or increased intracranial pressure, had a history of gastrectomy, or were pregnant or lactating. As with the previous articles, the primary endpoint was incidence of clinically important bleeding. Secondary endpoints included incidence of Clostridium difficile infection, ICU length of stay, and hospital length of stay.
There were no significant differences found in baseline characteristics, including median Sequential Organ Failure Assessment (SOFA) and Simplified Acute Physiology Score II scores. No differences were found as well in regards to enteral feeding characteristics, including average kcal per kilogram per day given and total enteral nutrition given. Additionally, there were no statistically significant differences in any of the endpoints studied, including the primary endpoint of clinically important bleeding (one incidence in each group; p=0.99) for the 55 patients treated with pantoprazole compared to the 47 patients treated with placebo.
These results show no extra benefit of administering pantoprazole in mechanically ventilated patients receiving enteral nutrition in regards to clinically important bleeding incidence, incidence of Clostridium difficile infection, and length of stay. As with the other newer trials, the exploratory nature of this study should be considered when interpreting these results.
Conclusions and recommendations for practitionersAs mentioned previously, the exploratory and pilot designs of the newly published trials are factors that should not be overlooked when interpreting these results. It is possible that these studies were not powered to find the difference between PPIs and placebo that truly existed. With that said, these new trials offer hope that there may evidence to show a reduced need for stress ulcer prophylaxis. All three studies found no statistical difference when examining stress ulcer prophylaxis in a population in which prophylaxis was originally recommended for in the ASHP guidelines. Until larger, powered studies provide evidence suggesting that stress ulcer prophylaxis is not needed, it is reasonable to still utilize it in critically ill patients that have coagulopathies or that have been ventilated for at least 48 hours. Despite these populations not being a focus of recently published literature, it may be reasonable to utilize prophylaxis in patients with low Glasgow Coma Scale scores and with significant thermal injury. It is difficult to recommend prophylaxis be utilized in the populations that were previously recommended solely based on expert opinion due to the limited evidence for such recommendations. For reasons stated previously, it is also unlikely that stress ulcer prophylaxis is beneficial in many other populations than the critically ill, particularly those admitted to a general unit. Therefore, it is also reasonable that stress ulcer prophylaxis should generally be discontinued in non-critically ill patients.