Description

Fiddian-Green et al developed models for predicting massive bleeding from stress ulcerations. This can help identify patients who may require more aggressive monitoring and management. The authors are from the University of Michigan in Ann Arbor.


 

Risk factors for bleeding from stress ulcerations:

(1) admission to the ICU for post-operative care, burn, major trauma, or myocardial infarction

(2) established liver disease with total bilirubin > 0.9 mg/dL

(3) renal disease with a serum creatinine > 1.3 mg/dL

(4) respiratory failure requiring mechanical ventilation

(5) definite evidence of sepsis

(6) systemic hypotension with systolic blood pressure < 80 mm Hg

 

Models:

(1) based on intramural pH (measured using a pH probe) and number of risk factors

(2) gastric juice pCO2 in mm Hg

(3) arterial bicarbonate in mEq/L

 

Parameters defined for each model (Table 2, page 616):

(1) µ expression

(2) critical value

 

In general, the models consisted of:

 

X =

= (critical value) - (µ expression)

 

Y = probability of not massively bleeding =

= 1 / (1 + EXP((-1) * X))

 

probability of massive bleeding =

= 1 - Y

Model

Critical Value

µ Expression

pH and risk factors

-9.93

(1.05 * ((number of risk factors) + 1)) - (2.27 * (intramural pH))

gastric juice pCO2

3.69

0.0146 * pCO2

arterial bicarbonate

0.058

-0.099 * (arterial bicarbonate)

from Table 2, Figure 3 and Figure 4

 


To read more or access our algorithms and calculators, please log in or register.