BUR n s 4 2 ( 2 016) 1 7 21–1 7 2 7

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.elsevier.com/locate/burns

Blood transfusions in severe burn patients:

Epidemiology and predictive factors

Guosheng Wu a,1, Mingzhu Zhuang b,1, Xiaoming Fan a,1, Xudong Hong a,
Kangan Wang a, He Wang a, Zhengli Chen a, Yu Sun a, Zhaofan Xia a,*
a
Department of Burn Surgery, Changhai Hospital, The Second Military Medical University, Shanghai, People’s
Republic of China
b
Department of Blood Transfusion, Changhai Hospital, The Second Military Medical University, Shanghai, People’s
Republic of China

A R TI CLE I NF O
A BS TR A C T

Article history: Background: Blood is a vital resource commonly used in burn patients; however,
Accepted 2 June 2016 description of blood transfusions in severe burns is limited. The purpose of this study was
to describe the epidemiology of blood transfusions and determine factors associated with
Keywords: increased transfusion quantity.
Methods: This is a retrospective study of total 133 patients with >40% total body surface
Burn
Blood transfusions area
Epidemiology (TBSA) burns admitted to the burn center of Changhai hospital from January 2008 to
December 2013. The study characterized blood transfusions in severe burn patients. Uni-
variate and Multivariate regression analyses were used to evaluate the association of
clinical variables with blood transfusions.
Results: The overall transfusion rate was 97.7% (130 of 133). The median amount of total
blood (RBC and plasma), RBC and plasma transfusions was 54 units (Interquartile range
(IQR), 20–84), 19 units (IQR, 4–37.8) and 28.5 units (IQR, 14.8–51.8), respectively. The
number of
RBC transfusion in and outside operation room was 7 (0, 14) and 11 (2, 20) units, and the
number of plasma was 6 (0.5, 12) and 21 (11.5, 39.3) units. A median of one unit of blood
was transfused per TBSA and an average of 4 units per operation was given in the series.
The consumption of plasma is higher than that of RBC. On multivariate regression analysis,
age, full-thickness TBSA and number of operations were significant independent predictors
associated with the number of RBC transfusion, and coagulopathy and ICU length showed a
trend toward RBC consumption. Predictors for increased plasma transfusion were female,
high full-thickness TBSA burn and more operations.
Conclusions: Severe burn patients received an ample volume of blood transfusions.
Fully understanding of predictors of blood transfusions will allow physicians to better
optimize burn patients during hospitalization in an effort to use blood appropriately.
Ⓒ 2016 Elsevier Ltd and ISBI. All rights reserved.

* Corresponding author. Tel.: +86 21 31161821; fax: +86 21 65589829.

E-mail addresses: drwuguosheng@sina.cn (G. Wu), zmz_smmu@sina.com (M. Zhuang), fxm_smmu@sina.com (X. Fan),
h x d _ s m m u @ s i n a . c o m (X. Hong), w k a s m m u @ 1 6 3 . c o m (K. Wang), w h _ s m m u @ s i n a . c o m (H. Wang),
c z l _ s m m u @ s i n a . c o m (Z. Chen), sy_smmu@sina.com (Y. Sun), xiazhaofan_smmu@163.com (Z. Xia).
1
These authors contributed equally to the article.
http://dx.doi.org/10.1016/j.burns.2016.06.002
0305-4179/Ⓒ 2016 Elsevier Ltd and ISBI. All rights reserved.

1. Background
Severe burn patients usually experience anemia throughout
the entire duration of burn care and multiple factors contribute
to this complication including surgical management of
wounds, red blood cells (RBC) sequestration, direct
erythrocyte damage, nutritional deficiencies, bone marrow
dysfunction, and iatrogenic factors such as blood loss from
dressing changes and laboratory draws. Thus, to date, large
amounts of blood transfusions became the major and
indispensable method to combat the blood loss and
impaired erythropoiesis [1]. However, blood transfusions
were not a benign treatment. The use of blood products is
associated with various adverse events such as infections,
immunosuppression, transfusion- related acute lung injury and
transfusion errors [2]. Moreover, previous studies reported
that the amount of transfused blood was independently
associated with increased mortality [3–5]. Concerning the
risk/benefit ratio of blood transfusions, physicians propose
restricting the blood utilization in critical illness. As a special
critical illness population, the strategies to restrict blood use
should be based on not only hemoglobin level but also a
good knowledge of blood transfusions in burn
patients. Because of limited literature systematically analyz-
ing blood utilization in severe burn (>40% total body surface
area (TBSA)) patients, we aimed to describe blood
transfusions
in this population at a single burn center in Shanghai to be
able to improve the clinical appropriateness of blood
utilization in burn patients.
2. Methods
This is a retrospective study conducted at the burn center of
Changhai Hospital from January 2008 to December 2013.
Changhai hospital is a tertiary general hospital and its burn
center is equipped with 67 beds and mainly responsible for
the treatment of burns in Shanghai and other areas of east
China. 165 adult severe burn patients (aged 18 and over) with
burn area over 40% TBSA were admitted to the center during
this period. Those patients with incomplete medical records
and those who presented more than 72 h after injury were
excluded from this study. One woman who was pregnant
was also excluded. Eventually, 133 patients were included in
this study. Patient demographics, etiology of burn, burn
extent, comorbidities, concomitant injuries, blood transfu-
sions and outcomes were reviewed.
Treatment strategies commonly used for severe burn
patients in China were used in this center, including rapid
establishment of airway and ventilator support, sufficient
fluid resuscitation during the shock stage, early extensive
escharectomy and skin grafting used for coverage of deep
wounds, infection control, active management of
complica- tions, routine use of antacids and gastric
mucosal protective agents, and early administration of
enteric nutrients [6].
Blood transfusions were performed based on the clinical
judgment of attending physician or the medical consultant in
this center. The indications for administration of RBC
included anemia (hemoglobin <7 g/dL) and clinical
symptoms such as hypotension, hypoxia, fatigue, and low
urine output, while
indications for the transfusion of plasma were hypoproteine-
mia and coagulation disorders. The diagnosis of coagulopathy
is based on laboratory abnormalities such as prothrombin
time (PT) ≥16.3 s or activated partial thromboplastin time
(aPTT) ≥45 s or an international normalized ratio (INR) ≥1.5,
which represent our local laboratory’s definition of coagulo-
pathy. Three patients with cardiac history transfused for a
hemoglobin between 7 and 10 g/l were not considered
separately. One unit of any kind of component (RBC and
plasma) equals to 200 ml in this study.
All of the statistical analyses were made by IBM SPSS
Statistics 21. All variables across groups were compared using
the independent-sample t test, Fisher test, and the Mann–
Whitney test, when appropriate. Univariate and Multivariate
regression analyses were used to evaluate the association of
clinical variables with blood transfusions. p Values <0.05
were considered statistically significant. Variables
showing a p < 0.05 on univariate analysis were included in
the multivari- ate model. In addition, Stepwise regression
was applied to the multivariate regression analysis in order
to adjust the confounding factors. The inclusion criteria
was 0.1 and the exclusion criteria was 0.15.
3. Results
3.1. General characteristics
The characteristics of the enrolled 133 patients are shown in
Table 1. The average age for all patients was 41.4 14.1 years
(range 18–86 years). 99 (74.4%) were males and 34 (25.6%) were
females. The overall affected burn size was 70 (50, 87.5)%
TBSA (range 40–98.5%) and the full-thickness burn size was
30 (13, 60)% TBSA (range 0–98%). The Flame (n = 108) was the
most frequent cause of burns, followed by scald (n = 14),
chemical (n = 10), and electrical (n = 1). At the time of
admission, 35.3% of patients had pre-existing diseases or
concomitant injuries, and the most common ones were
pulmonary blast injury (16.5%) and hypertention (9.8%).
There were more females with cardiac disease compared with
males (8.8% vs. 0), and more males than females had brain
injury (5.1% vs. 0). No patients had pre- existing cogulopathy,
but 67 patients developed coagulopathy. In all, 55 (41.4%)
patients presented to the burn center from the scene, and
22.6% patients transferred from another hospital within 24 h
after injury and 36.1% within 72 h. Inhalation injury was
found in 99 (74.4%) patients. Tracheostomies were performed
in 82 (61.7%) patients, and 32.3% of patients had
escharotomies during their first 48 h. More males (66.7%)
underwent tracheostomy than females (47.1%). Median length
of hospital stay (LOS) was 46 days (IQR, 24–80.5), with a
median of 26 days (IQR, 12.5–46.5) in intensive care unit
(ICU). 29 of the 133 patients died (21.8%) during their hospital
stay.
Of all 133 patients, 115 (86.5%) received RBC transfusion
and 130 (97.7%) received plasma transfusion. Three patients
did not receive any kind of blood. Fifteen patients received
plasma transfusion only. As shown in Fig. 1, with the
number of severe burn patients fluctuated from 29 (2008) to
25 (2013), the total number of RBC consumption ranged from
689.5 units (2008) to 602 units (2013) and the total plasma
amount decreased from 1326.5 units (2008) to 798 units
(2013). Among all patients, the
 BUR n s 4 2 ( 2016) 1 7 21– 1 7 2 1723
7

Table 1 – General characteristics of patients.

Parameter Total (n = 133) Male (N = 99) Female (n = 34) p Value

Age (yrs) 41.4 14.1 40.0 13.0 45.3 16.5 >0.05
TBSA total (%) 70 (50, 87.5) 70 (50, 90) 60 (50, 81.3) >0.05
TBSA full thickness (%) 30 (13, 60) 28 (11, 60) 30 (20, 62.5) >0.05
Etiology (n) >0.05
Flame 108 (81.2%) 76 (76.8%) 32 (94.1%)
Scald 14 (10.5%) 12 (12.1%) 2 (5.9%)
Chemical 10 (7.5%) 10 (10.1%) 0
Electrical 1 (0.8%) 1 (1.0%) 0
Inhalation injury (n) 99 (74.4%) 71 (71.7%) 28 (82.4%) >0.05
Comorbidities (n)
Hypertension 13 (9.8%) 9 (9.1%) 4 (11.8%) >0.05
Diabetes 3 (2.3%) 2 (2.0%) 1 (2.9%) >0.05
Anemia 3 (2.3%) 1 (1.0%) 2 (5.9%) >0.05
Chronic lung disease 5 (3.8%) 4 (4.0%) 1 (2.9%) >0.05
Hepatic disease 5 (3.8%) 3 (3.0%) 2 (5.9%) >0.05
Kidney disease 2 (1.5%) 1 (1.0%) 1 (2.9%) >0.05
Cardiac disease 3 (2.3%) 0 3 (8.8%) <0.05
Brain injury 5 (3.8%) 5 (5.1%) 0 <0.05
Fracture 2 (1.5%) 1 (1.0%) 1 (2.9%) >0.05
Pulmonary blast injury 22 (16.5) 17 (17.2%) 5 (14.7%) >0.05
Coagulopathy 67 (50.4%) 52 (52.5%) 20 (58.8%) >0.05
Transfer (n) >0.05
Non 55 (41.4%) 38 (38.4%) 17 (50.0%)
<24 h 30 (22.6%) 22 (22.2%) 8 (23.5%)
24–72 h 48 (36.1%) 39 (39.4%) 9 (26.5%)
Tracheostomy (n) 82 (61.7%) 66 (66.7%) 16 (47.1%) <0.05
Escharotomy during first 48 h (n) 43 (32.3%) 34 (34.3%) 9 (26.5%) >0.05
Time to 1st operation (days)a 4 (2, 6) 4 (3, 6) 4 (3, 6) >0.05
Number of operations 4 (2, 7) 4 (1, 7) 4 (2, 7.3) >0.05
Time to 1st receive transfusion (days)b 1 (0, 3) 1 (0, 4) 0.5 (0, 2) >0.05
Time to last receive transfusion (days)b 19 (7, 38.5) 20.5 (7, 37.8) 19.5 (9, 44.8) >0.05
Total blood (units)c 59 (26.5, 110.5) 54 (17, 84.3) 58.5 (28.5, 108.5) >0.05
RBC (units) 19 (4, 37.8) 19 (3, 38) 19 (10, 35) >0.05
RBC during first 48 h (units) 0 (0, 0) 0 (0, 0) 0 (0, 0) >0.05
Plasma (units) 28.5 (14.8, 51.8) 27 (14, 49) 31 (17.5, 63) >0.05
Plasma during first 48 h (units) 4 (0, 13) 2 (0, 13) 7 (0, 13) >0.05
Total perioperative RBC use (units) 11 (2, 20) 11 (2, 20) 11.3 (5.8, 21.3) >0.05
Total operative RBC use (units) 7 (0, 14) 7 (0, 14) 6.5 (4.4, 14.8) >0.05
Total perioperative plasma use (units) 21 (11.5, 39.3) 20 (11, 38) 23 (12, 47.9) >0.05
Total operative plasma use (units) 6 (0.5, 12) 5 (0, 12) 6 (4.8, 15.3) >0.05
PLT 0 (0, 1.88) 0 (0, 3) 0 (0, 0) <0.05
Whole blood 0 (0, 0) 0 (0, 0) 0 (0, 0) >0.05
Blood per TBSAc 1.0 (0.5, 1.9) 1 (0.4, 1.8) 1.2 (0.8, 2.5) >0.05
Blood per operationa,d 4.0 3.6 3.9 3.8 4.5 2.9 >0.05
ICU length (days) 26 (12.5, 46.5) 25 (10, 45) 31 (16.3, 52.5) >0.05
Length of stay (days) 46 (24, 80.5) 44 (19, 81) 55 (29.8, 79) >0.05
Mortality 29 (21.8%) 22 (22.2%) 7 (20.6) >0.05
TBSA, total body surface area; RBC, red blood cell; PLT, platelet; ICU, intensive care unit.
a
16 patients without surgery were excluded from the analysis.
b
Three patients without transfusions were excluded from the analysis.
c
All four components (RBC, plasma, PLT and whole blood).
d
Blood includes RBC and plasma.
Bold numbers were statistically significant ( p < 0.05)

median number of total blood, RBC and plasma utilization was
54 units (IQR, 20–84), 19 units (IQR, 4–37.8) and 28.5 units
(IQR, 14.8–51.8), respectively (Table 1). The number of RBC
trans- fused in and outside operation room was 7 units (IQR,
0–14) and 11 units (IQR, 2–20). The corresponding number
of units of plasma was 6 units (IQR, 0.5–12) and 21 units (IQR,
11.5–39.3). A median of 0.8 unit of blood was transfused per
%TBSA and an average of 4.0 units per operation. Patients
received their first blood transfusion a median of one day after
injury, and the last
transfusion was administered 19 (IQR, 7–38.5) days after
injury. There was no significant difference between males
and females in blood transfusions (Table 1).
3.2. Predictors for the number of RBC and
plasma transfusions in severe burn patients
Factors predicting the number of RBC and plasma
transfusions in survivors are shown in Tables 2 and 3.
Univariate regression

Fig. 1 – Number of severe burn patients and total
transfusion quantity (2008–2013).
analyses showed that age (coefficient = 0.504), total
TBSA (coefficient = 0.7), full-thickness TBSA (coefficient =
0.751), coagulopathy (coefficient = 21.393), transfer
(coeffi- cient = 21.477), transfer within 24 h (coefficient =
16.552), tracheostomy (coefficient = 20.79), escharotomy
(coeffi- cient = 23.801), number of operations (coefficient =
4.914) and ICU length (coefficient = 0.74) were associated
with a significantly increased consumption of RBC (Table 2).
Howev- er, multivariate regression analysis after adjusting
evaluated factors showed that age (coefficient = 0.598), full-
thickness TBSA (coefficient = 0.279) and number of operations
(coeffi- cient = 2.529) were significant predictors for the
number of RBC transfusion, and coagulopathy
(coefficient = 8.183, p = 0.07) and ICU length (coefficient =
0.233, p = 0.058) showed a trend toward RBC consumption
(Table 2).
Univariate regression analyses showed that female (coeffi-
cient = 16.677), total TBSA (coefficient = 0.742), full-thickness
TBSA (coefficient = 0.772), inhalation injury (coeffi-
cient = 15.332), coagulopathy (coefficient = 16.223), transfer
within 24 h (coefficient = 22.227), tracheostomy (coeffi-
cient = 17.848), escharotomy (coefficient = 24.55), number of
operations (coefficient = 4.345), ICU length (coefficient = 0.577)
and LOS (coefficient = 0.145) were significant predictors for
increased amount of plasma transfusion (Table 3). In a
multivariate regression analysis, the predictors of plasma
consumption were found to be only female (coeffi-
cient = 14.497), full-thickness TBSA (coefficient = 0.499) and
number of operations (coefficient = 2.571).
4. Discussion
Much has written about blood transfusions and associated
complications in burn patients, but little has been concerning
the magnitude of blood utilization in severe burns [7,8]. To the
best of our knowledge, this is the first study that reported the
epidemiology and predictors for RBC and plasma transfusions
in patients with over 40% TBSA burns.
Blood is a vital resource commonly used in burn patients
and the rate of transfusion varies greatly in reports from
different authors. In a study including 109 pediatric patients,
Birdsell et al. [9] reported that 100% of children with ≥30%
TBSA burn received a blood transfusion, but no
transfusions
were given for TBSA of <5%. Yogore et al. [10] reported a
5.7% rate in 1282 patients with <10% TBSA burn, 21% with
11–20% TBSA burn, 39% with 21–30% TBSA burn, and 62%
of patients with >30% TBSA burn. Palmieri et al. [5] reported
a 74.7% transfusion rate in 620 patients with over 20% TBSA
burns from 21 centers. Posluszny et al. [11] reported
88.7% of 71 patients with ≥20% TBSA burn received a blood
transfusion. In our population studied, the transfusion rate was
97.7%. In a study conducted by Lu et al. [8], 71.9% of 89
patients with 15– 65% TBSA burns received PRBC and 44.9%
received plasma transfusion. In our series, 88.5% patients
received RBC and 97.7% received plasma transfusion.
In addition, severe burns usually need large amounts of
transfusion due to their sustained metabolic response to
injury and multiple surgical managements. Graves et al. [12]
reported an average of 19.7 units was transfused in their
patients with >10%TBSA. Vasko et al. [13] reported patients
with >10% TBSA burn received an average of 8.94 units and for
patients with over 30% TBSA, 17 units were transfused. In
Palmieri et al. [5] study, patients with ≥20% and ≥50%TBSA
burn had 13.7 1.1 and over 30 units of RBC transfused,
respectively. In a retrospective review, patients with over 40%
TBSA burn received an average of 20 units [14]. In the
present study, a median of 59 units of blood with a range
of 0–
288.5 units (mean: 68.2) were transfused per patient, which
was significantly higher than previous studies.
Our studies showed that, from 2008 to 2013, the total
number of plasma consumption was consistently higher than
RBC. Furthermore, the plasma transfusion per patient was
higher than RBC, especially during the first 48 h after injury
and perioperative period, while RBC transfusion was predom-
inant in the operation room. One reason is that plasma is
used as a kind of colloid fluid for resuscitation to increase
circulating blood volume in burn patients. In the past years,
colloids were forbidden to use, but now are starting to
increase their presence in resuscitation practices. The
advantages of colloid use for fluid resuscitation in burn
patients include avoiding ‘fluid creep’, reduced edema, and
protective against pulmonary complications [15]. The
second reason is that aggressive use of plasma can
decrease overall RBC quantity and overall cost of plasma and
RBC. Palmieri et al. [16], in a prospective randomized trial,
compared the impact of a 4:1 versus a 1:1 packed red blood
cell-fresh frozen plasma transfusion strategy on outcomes
in children with >20% TBSA burns. The 1:1 group received
less RBC than the 4:1 group
(40.7 0.02 U in 1:1 group vs. 73.1 0.02 U in 4:1 group) and
resulted in higher antithrombin and protein C levels postop-
eratively without a difference in coagulation function.
Furthermore, the other reasons include improving immunity,
adding up to nutrition and supplying blood albumin [17].
In the present study, we found multiple predictors for
increased number of RBC transfusion were significant on
univariate analysis, but only age, full-thickness TBSA and
number of operations maintained significance in multivariate
analysis. In contrast with the present study, age was not
reported to be significantly associated with increased RBC
transfusion in Lu et al’s [8] study, but their data showed a
trend toward statistical significance. The difference may be
due to their small sample size and moderate severity burn
patients. That a larger full-thickness TBSA burned predicted
increasing
 Table 2 – Predictors for RBC transfusion in survivors.
Univariate Multivariatea
Coefficient Standard error p Value Coefficient Standard error p Value
Age, yrs 0.504 0.246 0.043 0.598 0.186 0.002
Gender, male —5.007 6.713 0.457
TBSA total 0.700 0.147 <0.001
TBSA full thickness 0.751 0.094 <0.001 0.279 0.105 0.010
Etiology
Flame 0.326 7.488 0.965
Scald —13.070 9.926 0.191
Chemical 14.661 10.395 0.161
Electrical —7.738 30.230 0.798
Inhalation injury 10.086 6.437 0.121
Comorbidities
Hypertension 4.770 15.338 0.756
27.119 —
Diabetes 17.425
8.927
Anemia 17.609 0.123
—2.971
Chronic lung disease 20.029 21.485 0.613
Hepatic disease 4.674 30.175 0.890
Kidney disease —3.735 — 0.508
Cardiac disease —1.320 12.959 —
Brain injury 21.484 0.719
Fracture 7.979 0.862
Pulmonary blast injury 21.393 0.869
Coagulopathy 5.567 <0.001 8.183 4.454 0.070
Transfer —21.477
Non 16.552 5.582 <0.001
<24 h 10.180 6.810 0.017
24–72 h 20.790 6.120 0.099
Tracheostomy 23.801 5.550 0.003
Escharotomy during first 48 h 4.914 6.144 0.001
Number of operations 1.034 0.535 <0.001 2.529 0.774 0.002
Time to 1st receive transfusion (days)a 0.740 0.669 0.125
ICU length (days) 0.067 0.088 <0.001 0.233 0.121 0.058
Length of stay (days) 0.046 0.147
TBSA, total body surface area; RBC, red blood cell; ICU, intensive care unit.
a
Twelve patients without RBC transfusion were excluded from the analysis.
Bold numbers were statistically significant ( p < 0.05). The inclusion criteria was 0.1 and the exclusion criteria was 0.15.

transfusion is not surprising, the effect of operation on the
amount of transfusions has not been described. An explana-
tion for the prediction of number of operations could be
repeated extensive excision and grafting conducted in burn
wounds lead to increased surgical blood loss [18]. Notably,
our data showed a trend for increased RBC transfusion in
patients with coagulopathy. In trauma patients, coagulopathy
was asscociated with increased RBC transfusion [19].
Similarly, Hofstra et al. [20] reported that developed
coagulopathy contributed to increased transfusion needs in
burn patients. Although numerous reports exist indicating that
burn patients with inhalation injury had a higher likelihood of
receiving more blood transfusions, most studies fail to show a
significant impact on the number of RBC or plasma transfu-
sions, similar to our findings [8,21,22].
The study that focusing on the predictors of number of
plasma transfusions is limited. Lu et al. [8] demonstrated
that TBSA and argatroban use were significant predictors of
plasma transfusions. A multivariate regression analysis on
the signifi- cant univariate variables from our population
confirmed full- thickness TBSA was an independent
predictor for increased amount of plasma transfusion. In
addition, we found female and number of operations were
independent predictors for
plasma transfusion. Therefore, according to the above
results, reducing the number of operations or intraoperative
blood loss could decrease blood transfusions. Surprisingly,
we did not find coagulopathy to be significantly associated
with plasma transfusion. We do not have a clear explanation
for this result, but it is possible that cryoprecipitate is more
often than plasma used to supply coagulopathy factors in
China. Further study is needed to identify predictors of
plasma transfusion to help physicians to use plasma
rationally.
Some limitations of the present work have to be mentioned.
First, this is a single center retrospective study which includes
a relatively small number of patients. Second, we did not
analyze the association between the number of transfusion
and use of anticoagulants, which has been reported as a
predictor of increased transfusion [8]. In addition, potentially
relevant information regarding body weights, presence of
infections and mechanical ventilation were not determined in
our study. In summary, the overall transfusion rate in our
series of severe burn patients is 97.7%. We showed severe
burn patients received an ample volume of blood
transfusions. The overall transfusion quantity is driven by
the number of patients, and the consumption of plasma is
higher than that of RBC. We identified significant
predictors for the number of RBC and
 Table 3 – Predictors for plasma transfusion in survivors.
Univariate Multivariate #
Coefficient Standard error p Value Coefficient Standard error p Value
Age, yrs 0.229 0.294 0.438
Gender, male S16.677 7.718 0.033 S14.497 6.377 0.025
TBSA total 0.742 0.176 <0.001
TBSA full thickness 0.772 0.118 <0.001 0.499 0.138 <0.001
Etiology
Flame 5.405 15.024 0.720
Scald S14.810 11.646 0.206
Chemical 7.740 12.284 0.530
Electrical 2.034 35.461 0.954
Inhalation injury 15.322 7.486 0.043
Comorbidities
Hypertension S9.085 13.972 0.614
Diabetes 12.543 20.638 S
Anemia S8.223 20.659 0.545
Chronic lung disease S17.833 20.600 0.691
Hepatic disease S11.092 35.444 0.389
Kidney disease S10.946 – 0.755
Cardiac disease S2.279 25.174 S
Brain injury 4.236 25.196 0.665
Fracture 9.349 0.928
Pulmonary blast injury 0.651
Coagulopathy 16.223 6.797 0.019
Transfer
Non S12.916 6.887 0.064
<24 h 22.227 7.913 0.006
24–72 h S3.504 7.266 0.631
Tracheostomy 17.848 6.713 0.009
Escharotomy during first 48 h 24.550 7.324 0.001
Number of operations 4.345 0.731 <0.001 2.571 0.835 0.003
Time to 1st receive transfusion (days)a 0.169 0.789 0.831
ICU length (days) 0.577 0.121 <0.001
Length of stay (days) 0.145 0.052 0.007
TBSA, total body surface area; ICU, intensive care unit.
a
Three patients without plasma transfusion were excluded from the analysis.
Bold numbers were statistically significant ( p < 0.05). The inclusion criteria was 0.1 and the exclusion criteria was 0.15.

plasma transfusions. We believe that fully understanding of

predictors of transfusion will allow physicians to better
optimize burn patients during hospitalization in an effort to
reduce the need for transfusion and its associated complica-
tions.
Conflict of interest
The authors declare that they have no conflict of interest.
Author contribution
Acknowledgements
Special thanks to Dr. Pengfei Luo who works in Department
of Burn Surgery, Changhai Hospital, Second Military Medical
University for his statistical advice. This work was funded
by Young Talents Training Program of Shanghai Health
System (XYQ2013079 and XYQ2013075), by ‘‘1255’’ Discipline
Construction Program of Changhai Hospital (CH125510200).
This work was also supported by Shanghai ‘‘priority’’
for clinical key discipline project and Joint Research
Program of important diseases of Shanghai Health System
(2013ZYJB0008).

Substantial contributions to the design of the work: GSW, re FE R E N C E S

MZZ, XMF.
Acquisition, analysis and interpretation of data for the
work: GSW, MZZ, XMF, XDH, KAW, HW, ZLC, YS and ZFX.
Drafting the work: GSW, MZZ, XMF.
Revising the work critically for important intellectual
content: GSW, MZZ, XMF, ZFX.
Final approval of the version submitted for publication:
ZFX.
[1] Wallner SF, Vautrin R. The anemia of thermal injury:
mechanism of inhibition of erythropoiesis. Proc Soc Exp
Biol Med 1986;181(1):144–50.
[2] Curinga G, Jain A, Feldman M, Prosciak M, Phillips B, Milner
S. Red blood cell transfusion following burn. Burns
2011;37(5):742–52.
 [3] Vincent JL, Baron JF, Reinhart K, Gattinoni L, Thijs L,
Webb A, et al. Anemia and blood transfusion in critically ill
patients. JAMA 2002 25;288(12):1499–507.
[4] Corwin HL, Gettinger A, Pearl RG, Fink MP, Levy MM,
Abraham E, et al. The CRIT Study: anemia and blood
transfusion in the critically ill—current clinical practice in
the United States. Crit Care Med 2004;32(1):39–52.
[5] Palmieri TL, Caruso DM, Foster KN, Cairns BA, Peck MD,
Gamelli RL, et al. Effect of blood transfusion on outcome
after major burn injury: a multicenter study. Crit Care Med
2006;34(6):1602–7.
[6] Xie B, Xiao SC, Peng XD, Zhu SH, Lv KY, Li HY, et al.
Epidemiology and outcome analysis of severe extensive
burns: a 12-year summary of 103 cases in a burn center in
China. J Burn Care Res 2012;33(3):e127–32.
[7] Boral L, Kowal-Vern A, Yogore 3rd M, Patel H, Latenser BA.
Transfusions in burn patients with/without comorbidities.
J Burn Care Res 2009;30(2):268–73.
[8] Lu RP, Lin FC, Ortiz-Pujols SM, Adams SD, Whinna HC,
Cairns BA, et al. Blood utilization in patients with burn
injury and association with clinical outcomes (CME).
Transfusion 2013;53(10):2212–21.
[9] Birdsell DC, Birch JR. Anemia following thermal burns:
a survey of 109 children. Can J Surg 1971;14(5):
345–50.
[10] Yogore 3rd MG, Boral L, Kowal-Vern A, Patel H, Brown S,
Latenser BA. Use of blood bank services in a burn unit.
J Burn Care Res 2006;27(6):835–41.
[11] Posluszny Jr JA, Conrad P, Halerz M, Shankar R, Gamelli RL.
Classifying transfusions related to the anemia of critical
illness in burn patients. J Trauma 2011;71(1):
26–31.
[12] Graves TA, Cioffi WG, Mason AD, McMANUS WF, Pruitt Jr
BA. Relationship of transfusion and infection in a burn
population. J Trauma 1989;29:948–52.
[13]
Vasko SD, Burdge JJ, Ruberg RL, Verghese AS. Evaluation of
erythropoietin levels in the anemia of thermal injury.
J Burn Care Rehabil 1991;12:437–41.
[14] Posluszny JA, Gamelli RL. Anemia of thermal injury:
combined acute blood loss anemia and anemia of critical
illness. J Burn Care Res 2010;31:229–42.
[15] Greenhalgh DG. Burn resuscitation: the results of the ISBI/
ABA survey. Burns 2010;36(2):176–82.
[16] Palmieri TL, Greenhalgh DG, Sen S. Prospective comparison
of packed red blood cell-to-fresh frozen plasma transfusion
ratio of 4:1 versus 1:1 during acute massive burn excision.
J Trauma Acute Care Surg 2013;74(1):76–83.
[17] Zhu C, Gao Y, Li Z, Li Q, Gao Z, Liao Y, et al. A
systematic review and meta-analysis of the clinical
appropriateness of blood transfusion in China. Medicine
2015;94(50):e2164.
[18] Pidcoke HF, Isbell CL, Herzig MC, Fedyk CG, Schaffer BS,
Chung KK, et al. Acute blood loss during burn and soft
tissue excisions: an observational study of blood product
resuscitation practices and focused review. J Trauma Acute
Care Surg 2015;78:S39–47.
[19] Smith SA, Livingston MH, Merritt NH. Early coagulopathy
and metabolic acidosis predict transfusion of packed red
blood cells in pediatric trauma patients. J Pediatr Surg 2016.
PMID: 26960738.
[20] Hofstra JJ, Vlaar AP, Knape P, Mackie DP, Determann RM,
Choi G, et al. Pulmonary activation of coagulation and
inhibition of fibrinolysis after burn injuries and inhalation
trauma. J Trauma 2011;70(6):1389–97.
[21] Jeschke MG, Chinkes DL, Finnerty CC, Przkora R, Pereira
CT, Herndon DN. Blood transfusions are associated with
increased risk for development of sepsis in severely burned
pediatric patients. Crit Care Med 2007;35(2):579–83.
[22] Koljonen V, Tuimala J, Haglund C, Tukiainen E, Vuola J,
Juvonen E, et al. The use of blood products in adult patients
with burns. Scand J Surg 2016. PMID: 26929291.