Clinically significant gastrointestinal bleeding occurred in < 2% of the 2,252 patient enrolled in the study at 4 university-affiliated medical-surgical ICUs.
Patient selection: > 16 years of age
Exclusions:
(1) upper gastrointestinal bleeding within 48 hours before or 24 after admission
(2) previous total gastrectomy
(3) facial trauma
(4) epistaxis
(5) brain death
(6) hopeless prognosis
Parameters significantly associated with stress ulceration and significant gastrointestinal bleeding based on multiple regression analysis:
(1) mechanical ventilation > 48 hours (respiratory failure)
(2) coagulopathy
|
Parameters
|
odds ratio
|
p value
|
|
mechanical ventilation > 48 hours
|
15.6
|
< 0.001
|
|
coagulopathy
|
4.3
|
< 0.001
|
from Table 4
where:
• Receiving anticoagulant therapy was identified in the simple regression model but not in the multiple regression analysis (odds ratio 1.1, p value 0.88).
If neither of these risk factors were present, the risk of clinically important gastrointestinal bleeding was very low 0.1% (95% confidence interval 0.02 - 0.5%). Stress ulcer prophylaxis may be withheld in this group of patients. This can help reduce cost and the risk for complications associated with prophylactic therapy.
If one or both of these risk factors were present, the risk of clinically important gastrointestinal bleeding was 3.7% (despite some form of stress ulcer prophylaxis in > 50%).
