Moore et al used the tissue hemoglobin oxygen saturation to predict outcomes for a trauma patient receiving massive transfusion. The authors are from multiple institutions in the United States.
Tissue hemoglobin oxygen saturation was measured using a thenar eminence StO2 sensor placed on intact skin.
Massive transfusion was defined as >= 10 units of PRBCs (>=3,000 mL) transfused in the first 24 hours after trauma center admission. The authors later recommended that the definition be changed from first 24 hours to first 6 hours.
Poor outcome: multi-organ dysfunction or death
Parameters:
(1) minimum tissue hemoglobin oxygen saturation (StO2) at 1 hour in percent
(2) minimum tissue hemoglobin oxygen saturation (StO2) at 2 hours in percent
(3) minimum tissue hemoglobin oxygen saturation (StO2) at 3 hours in percent
(4) age in years
(5) minimum arterial pH at 2 hours
(6) minimum arterial pH at 3 hours
value of X at 1 hour =
= 1.6407 - (0.0284 * (min StO2 at 1 hour))
value of X at 2 hours using StO2 only =
= 1.3202 - (0.0273 * (min StO2 at 2 hours))
value of X at 2 hours using multivariates =
= 25.8889 + (0.0347 * (age)) - (0.0347 * (min StO2 at 2 hours)) - (3.5683 * (min pH at 2 hours))
value of X at 3 hours using StO2 only =
= 1.2165 - (0.0286 * (min StO2 at 3 hours))
value of X at 3 hours using multivariates
= 32.9769 + (0.0308 * (age)) - (0.0338 * (min StO2 at 3 hours)) - (4.5862 * (min pH at 3 hours))
probability of poor outcome =
= 1 / (1 + EXP((-1) * X))
Performance:
• The area under the ROC curves was 0.77 for multivariate 3 hours and 0.75 for multivariate 2 hours.
• The area under the ROC curves for the other models was 0.65 to 0.67.
