Case
A 69-year-old man with Type 2 diabetes mellitus and chronic obstructive pulmonary disease is admitted to the ICU with respiratory compromise related to community-acquired pneumonia (CAP), accompanied by delirium, hyperglycemia, and hypovolemia. He responds well to supportive, noninvasive ventilatory therapy, but develops positive stool occult blood testing during the second day in the ICU. Upon clinical improvement, you transfer him to the general medical floor. What is the best strategy for preventing clinically significant gastrointestinal (GI) bleeding during his hospitalization?
Background
Stress-related mucosal disease (SRMD) refers to superficial erosions or focal ulceration of the proximal gastrointestinal mucosa resulting from physiologic demand in acute illness. Multiple factors contribute to its development, including disruption of the protective mucosal barrier, splanchnic vasculature hypoperfusion, and release of inflammatory mediators.1,2 Increasing severity and number of lesions are associated with the propensity for stress-related mucosal bleeding (SRMB). Based on severity, GI hemorrhage can be defined as occult (detected on chemical testing), overt (grossly evident), or clinically important (overt with compromised hemodynamics or requiring transfusion).3
The majority of clinically significant GI bleeding events occur in critically ill patients. Although more than 75% of patients have endoscopic evidence of SRMD within 24 hours of ICU admission, lesions often resolve spontaneously as patients stabilize, and the average frequency of significant bleeding is only 6%. However, when present, SRMB in ICU patients increases the length of hospitalization, cost, and mortality rates.1,3 By contrast, significant GI bleeding occurs in less than 1% of inpatients without critical illness.4
While preventing clinically important bleeding in hospitalized patients is a crucial objective, current practice reflects significant stress ulcer phophylaxis (SUP) overutilization, with substantial economic impact and potential for harm. One in three patients takes antisecretory therapy (AST) upon admission.5 Additionally, SUP is prescribed in 32% to 54% of general medical inpatients, despite the low risk for SRMB. Importantly, these prophylactic agents are continued on discharge in more than half of these patients.6-9 Clinician prescribing practices potentially can set an unfounded standard of care for obligatory prophylaxis among inpatients.
Data for Clinical Decision-Making
Several studies report the risks for gastrointestinal hemorrhage related to acute illness. In a prospective study of 2,252 ICU patients, two independent predictors of clinically important, new-onset SRMB were identified: mechanical ventilation for more than 48 hours and coagulopathy (see Table 1). Of these risk factors, respiratory failure was present in virtually all patients with GI hemorrhage; only one patient had coagulopathy alone. Mechanical ventilation or coagulopathy was associated with a 4% risk of clinically important GI bleeding, whereas patients with neither symptom had a 0.1% risk.
Though GI bleeding was uncommon, mortality associated with bleeding was 49%, compared with 9% in the nonbleeding group. In the absence of one of these two risk factors, 900 ICU patients would need to be treated to prevent one clinically important GI bleeding event.3 Other studies identify an increased risk of GI bleeding in subsets of patients with trauma, thermal injury, and organ transplantation. Additional possible risk factors might include septic shock, glucocorticoid or NSAID use, renal or hepatic failure, and prior GI bleeding or ulcer.10 The likelihood of GI bleeding increases proportionate to the number of risk factors present.