Authors: Aaron Woofter, MD; Richard Goodgame, MD Series Editor: Richard Goodgame, MD


August 22, 2006

Upper Gastrointestinal Bleeding in the Critically Ill Patient

Injury vs Clinically Significant Bleeding

In the case of critically ill patients admitted to the ICU, some form of stress-related mucosal disease occurs in most individuals.[6,7,8] However, the lesions are initially asymptomatic. With correction of the hemodynamic deficits, ventilatory insufficiency, and systemic response to tissue injury, the lesions remain asymptomatic and resolve. But in 5% to 25% of patients, occult or overt bleeding may occur.[9] In their classic studies, Cook and colleagues[9] differentiated occult or overt bleeding (positive nasogastric blood, mild fall in hemoglobin over several days, melena) from clinically significant bleeding (overt bleeding followed [within 24 hours] by a decrease in systolic blood pressure, rise in pulse rate, a 2-g/dL fall in hemoglobin, or a definite transfusion requirement of 2 units of blood).[9] Clinically significant bleeding (occurred in 1.5% of ICU patients in the study by Cook and colleagues) is much less common than occult or overt bleeding alone (5% to 25% of patients).[5,9,10] In a survey of 500 intensivists, Daley and colleagues[11] found that the perceived incidence of clinically significant stress-related bleeding among ICU admissions was 2%.

Low-Risk vs High-Risk ICU Patients

The frequency of clinically significant stress-related mucosal bleeding in critically ill patients is related to certain risk factors. In 1994, Cook and colleagues[9] reported that respiratory failure and coagulopathy (INR > 1.5 or platelet count < 50 per 109/L) are independent risk factors for clinically important bleeding. Bleeding was encountered in 3.7% of 847 patients who had one or both of these risk factors, vs in 0.1% of those who had neither complication. The same investigators later studied ventilated patients and found the following additional risk factors: high creatinine, not receiving enteral nutrition, and not receiving stress ulcer prophylaxis with ranitidine.[12] Recent algorithms have included other risk factors for which the evidence is weaker: severe burns (> 35% of body surface), neurologic trauma, multiple trauma, liver failure with associated coagulopathy, multiple organ failure, posttransplant (de novo), septicemia, and active gastrointestinal disease.[13]

The 1999 study by Cook and colleagues[12] found that enteral nutrition was associated with significantly lower bleeding rates, presumably due to its effect on raising intragastric pH. However, the subjects in this study were not randomized. Patients who had an inherently lower risk of bleeding could have been more likely to tolerate tube feeds. Modern ICU care emphasizes the beneficial effects of enteral feeding, but the effect of enteral nutrition on the risk of bleeding is not clear. The survey by Daley and colleagues[11] showed that current intensivists often stop stress bleeding prophylaxis when enteral feeding is initiated. In fact, enteral access for nutrition and medication is not possible in many ICU patients with bleeding, obstruction, ileus, or mesenteric ischemia (vasopressors). Intolerance to enteral feeding occurs in 15% to 50% of ICU patients, probably due to splanchnic hypoperfusion and poor motility.[4] Enteral feeding is not recommended as the sole agent for prophylaxis.[14]

The influence of H pylori on the development of stress-related injury is uncertain, but it is unlikely to have the same major etiologic role as in peptic ulcer disease.[15,16,17,18]

Four Decades Ago vs Present Time

The frequency of stress-ulcer and clinically significant ICU bleeding appears to have declined over the past 4 decades.[19] Major gastrointestinal bleeding used to occur in 30% of all severely burned patients.[6] Allen and colleagues[19] have documented that since the writing and wide acceptance of the major guidelines advocating stress-ulcer prophylaxis, most studies have found 0% to 2.8% bleeding rates in the critical care setting. In an observational study aimed at improving the appropriateness of stress-bleeding prophylaxis, Coursol and Sanzari[13] noted only a 1% bleeding rate among both patients with and without prophylaxis. Faisy and colleagues[20] collected evidence from the literature and personal observations about the low incidence of stress-related bleeding and the lack of impact of prophylaxis. No study in the new millennium has compared stress-related bleeding with placebo treatment. Faisy and colleagues[20] performed a retrospective observational study in a University Hospital ICU. They looked at 2 periods (phases) in the same ICU: 15 months with standard prophylaxis (736 patients) followed by 15 months without prophylaxis (737 patients).They recorded acute blood loss, the number of upper endoscopies, and overall mortality. Overt gastrointestinal bleeding and clinically significant gastrointestinal bleeding were defined according to the criteria used by Cook and colleagues[9] in their studies. Comparing phases 1 and 2, rates of overt gastrointestinal bleeding (14/736 and 12/737, respectively), and clinically significant gastrointestinal bleeding (10/736 and 8/737, respectively), did not differ significantly. The frequency of clinically significant bleeding was less than 1.5% throughout the 2 phases. These investigators believe that improvements in ICU care -- such as oxygenation, fluid resuscitation, nutrition, antibiotics, and cardiac care -- have led to improved microcirculation and tissue oxygenation in critically ill patients. This may in turn have led to the declining incidence of clinically significant gastrointestinal bleeding.

Given the clinical features of this case, what is the significance and likely outcome of this bleeding event?

View the correct answer.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.