Transfusion of packed red blood cells (PRBCs) increases morbidity and mortality in
select surgical specialty patients. The impact of low-volume, leukoreduced RBC transfusion
on general surgery patients is less well understood.
The American College of Surgeons National Surgical Quality Improvement Program participant
use file was queried for general surgery patients recorded in 2005 to 2006 (n = 125,223).
Thirty-day morbidity (21 uniformly defined complications) and mortality, demographic,
preoperative, and intraoperative risk variables were obtained. Infectious complications
and composite morbidity and mortality were stratified across intraoperative PRBCs
units received. Multivariable logistic regression was used to assess influence of
transfusion on outcomes, while adjusting for transfusion propensity, procedure type,
wound class, operative duration, and 30+ patient risk factors.
After adjustment for transfusion propensity, procedure group, wound class, operative
duration, and all other important risk variables, 1 U PRBCs significantly (p < 0.05)
increased risk of 30-day mortality (odds ratio [OR] = 1.32), composite morbidity (OR
= 1.23), pneumonia (OR = 1.24), and sepsis/shock (OR = 1.29). Transfusion of 2 U additionally
increased risk for these outcomes (OR = 1.38, 1.40, 1.25, 1.53, respectively; p <or=
0.05) plus surgical-site infection (OR = 1.25; p < 0.05). A risk index for calculating
transfusion likelihood demonstrated very good discrimination (c-index = 0.844).
Intraoperative transfusion of PRBCs increases risk for mortality and several morbidities
in general surgery patients. These risks, substantial for even 1 U, remain after adjustment
for transfusion propensity and numerous risk factors available in the American College
of Surgeons National Surgical Quality Improvement Program. Transfusion for mildly
hypovolemic or anemic patients should be discouraged in light of these risks.