752

Abstract

Introduction: Stress ulcers are superficial lesions of the gastrointestinal tract induced by physiologic stress. In a 1994 study, Cook et al. identified mechanical ventilation ≥ 48 hours and coagulopathy as two major risk factors which significantly predict stress-induced gastrointestinal bleeding (SiGIB). The current American Society of Health-System Pharmacists (ASHP) guidelines recommend administering stress ulcer prophylaxis (SUP) using histamine H2 receptor antagonists (H2RA) or proton pump inhibitors (PPI) to all mechanically ventilated or coagulopathic patients. The ASHP guidelines also recommend SUP for intensive care unit (ICU) patients with either an inability to obey simple commands or a Glasgow Coma Scale (GCS) ≤ 10. The results of two randomized controlled trials are cited which found a substantially increased incidence of clinically important bleeding in patients with severe head injury. Halloran, et al. defined severe head injuries as patients unable to obey simple commands while Burgess, et al. included patients with GCS ≤10. To our knowledge, there are no previous studies which quantify the association between GCS and SiGIB. Despite the higher level of recommendation for SUP in patients with severe head injuries, many institutions do not administer prophylaxis unless patients are also intubated or coagulopathic. We therefore conduct a retrospective data analysis to evaluate the Glasgow Coma Scale as a risk factor for SiGIB. Methods: This retrospective data analysis was performed at Intermountain Medical Center, a 465-bed Level 1 Trauma Center located in Murray, Utah. Patient information was obtained from the Electronic Data Warehouse (EDW), a data repository which includes ICD-9 codes, laboratory values, and vital signs as well as through electronic chart review. All adult patients (age ≥ 18) admitted to the Shock-Trauma, Neuroscience, and Respiratory Intensive Care Units between 01/01/2009 and 09/31/2012 were included.. Patients meeting any of the following criteria were excluded: diagnosis of an upper GIB within 48 hours before or 24 hours after admission, a history of Zollinger-Ellison syndrome, variceal bleeding, death, discharge, or transfer from the ICU within 24 hours of admission. Cases of GIB were initially extracted using ICD-9 codes. Unfortunately this generated an erroneously low incidence likely due to improper recording within the EDW. We therefore extracted data from patients who received a PPI infusion, the current standard of care at Intermountain Medical Center for suspected or confirmed GIB. Manual electronic chart review of each case was then performed to assess for inclusion and exclusion criteria. The primary outcome attempted to identify the GCS cutoff below which SUP should be indicated due to risk of overt or clinically important bleeding. Secondary outcomes compared the risk of mechanical ventilation, coagulopathy, and GCS scores on the development of overt bleeding or clinically important bleeding. A test of proportions assessed the risk of bleeding at each GCS and logistical regression was used to determine the association between stress ulceration, mechanical ventilation, coagulopathy, and GCS. Results: 11,224 total admissions were included in the final analysis. Stress-induced overt bleeding occurred in a total of 27 (0.24%) cases, of which 22 (0.20%) were clinically important. 44.5% of overt bleeds occurred in patients with GCS of 3. A test of proportions and receiver operator characteristic failed to determine a GCS cutoff at which high rates of SiGIB would warrant SUP. Logistical regression found mechanical ventilation (p=0.020) and coagulopathy (p=0.018) to be significantly associated with risk of overt bleeding. Conclusions: No significant association was found between low GCS and SiGIB. We are unable to support guideline recommendations to administer SUP for GCS ≤ 10. Clinicals should continue to administer SUP for mechanically ventilated or coagulopathy patients.