Eur J Trauma Emerg Surg

DOI 10.1007/s00068-014-0478-4

ORIGINAL ARTICLE

Comparison of recombinant human thrombomodulin

and gabexate mesylate for treatment of disseminated
intravascular coagulation (DIC) with sepsis following emergent
gastrointestinal surgery: a retrospective study
T. Akahoshi · H. Sugimori · N. Kaku · K. Tokuda ·
T. Nagata · E. Noda · M. Morita · M. Hashizume ·
Y. Maehara

Received: 28 July 2014 / Accepted: 6 November 2014

© Springer-Verlag Berlin Heidelberg 2014

Abstract
Purpose Recombinant thrombomodulin (rTM) has been
available in Japan since 2008, but there is concern about
its association with postoperative hemorrhage. The efficacy
and safety of rTM were examined in patients with dissemi-
nated intravascular coagulation (DIC) caused by a septic
condition after gastrointestinal surgery.
Methods Forty-two patients were emergently admitted
to the intensive care unit after emergent gastrointestinal
surgery in Kyushu University Hospital from May 2008
to April 2013. Of these patients, 22 had DIC (defined as
an acute DIC score ≥4). All but three patients received
treatment with gabexate mesylate (GM) (n = 9) or rTM
(n = 10). The causes of sepsis were peritonitis with colo-
rectal perforation, anastomotic leakage, and intestinal
necrosis. Acute DIC score, sepsis-related organ failure
assessment score, platelet count, and a variety of biochemi-
cal parameters were compared between rTM and GM
recipients after treatment administration.
Results There were no significant differences between the
groups for any parameter except C-reactive protein levels.
The CRP level tended to be lower in the rTM group than in
T. Akahoshi · T. Nagata · M. Hashizume
Department of Disaster and Emergency Medicine, Graduate
School of Medical Sciences, Kyushu University, Fukuoka, Japan
T. Akahoshi (*) · M. Morita · Y. Maehara
Emergency and Critical Care Center, Kyushu University,
Fukuoka, Japan
e-mail: a_tom411@yahoo.co.jp
T. Akahoshi · H. Sugimori · N. Kaku · K. Tokuda · E. Noda ·
Y. Maehara
Department of Surgery and Science, Graduate School of Medical
Sciences, Kyushu University, 3‑1‑1 Maidashi, Higashi‑ku,
Fukuoka 812‑8582, Japan
the GM group. Acute DIC score in the rTM group resolved
significantly earlier than that in the GM group. No patient
stopped the administration of rTM because of postoperative
bleeding.
Conclusion rTM may be an effective therapeutic drug for
the treatment of septic patients with DIC following emer-
gent gastrointestinal surgery.
Keywords DIC · Gastrointestinal surgery · rTM · Sepsis ·
Panperitonitis
Introduction
Sepsis following gastrointestinal surgery has a poor
prognosis; it mainly occurs in patients with colorectal
perforation-associated peritonitis [1, 2]. Furthermore,
when postoperative sepsis is complicated by dissemi-
nated intravascular coagulation (DIC), the prognosis
worsens [2]. Recombinant human soluble thrombomodu-
lin (rTM) has been available in Japan since May 2008,
and is covered by the national health insurance system
for the treatment of DIC. According to several reports,
systemic inflammatory response (SIRS) and sepsis-
related organ failure assessment (SOFA) scores and
prognoses were significantly improved in patients with
infection-based DIC treated with rTM [3, 4], although
no studies have reported its use after gastrointestinal
surgery. The administration of rTM after surgery may
aggravate postoperative hemorrhage, and the advan-
tages and disadvantages of its use have not been clearly
elucidated.
This study compared the efficacy and safety of rTM with
those of gabexate mesylate (GM) in patients with septic
DIC after gastrointestinal surgery.

13
 T. Akahoshi et al.

Methods in four patients after digestive tract surgery, all of whom

Between May 2008 and March 2012, postoperative inten-
sive care unit (ICU) treatment was administered in 42
patients with septic condition with one of the following
conditions: digestive tract perforation, anastomotic failure
after gastrointestinal surgery, traumatic intestinal injury, or
strangulation intestinal necrosis. Of these patients, 22 had
an acute DIC score ≥4 in the ICU. The 22 patients with
DIC were treated with GM, an antithrombin III (AT-III), or
rTM; no patient was treated with heparin or low-molecular-
weight heparin. Prior to the present study, rTM had been
used for postoperative DIC in 10 patients (Table 1). While
rTM became clinically available in 2008, it was first used
in the ICU of our hospital mainly for medical patients with
DIC following severe infection. In 2008, DIC was observed
were treated with GM. Subsequently, rTM was adminis-
tered in only one of four patients with postoperative DIC
(25 %) in 2009, three of five patients (60 %) in 2010, and
two of four patients (50 %) in 2011. In 2012, all 4 patients
with postoperative DIC were treated with rTM (Table 1).
Of 22 postoperative septic DIC patients, 2 were treated
only with AT-III, and one patient was not treated at all for
DIC because of prompt improvement of DIC score after
endotoxin absorption therapy (PMX). Excluding these 3
patients, 19 were administered rTM or GM for DIC treat-
ment as shown in Table 1.
The outcomes in the ten patients treated with rTM (the
rTM group) and the nine patients treated with GM (the GM
group) were compared. Platelet (Plt), fibrinogen degrada-
tion product (FDP), d-dimer, and fibrinogen levels, and

Table 1  Characteristics of patients with severe sepsis and DIC following Gl surgery
Cause of panperitonitis Operation Drug for DIC

2008 Intestinal necrosis after traumatic mesenteric vein injury Resection of necrotic intestine GM + AT-III
2008 #1. Massive bleeding from mesenteric vessel injury Ligation of mesenteric vessels, resection of left adrenal and GM
#2. Traumatic kidney injury kidney, cholecystectomy
#3. Cholecystic injury
2008 Perforated ascending colon diverticulum Ileocecal resection, abdominal drainage, colostomy GM
2008 Panperitonitis due to perforated sigmoid colon diverticu- Sigmoidectomy, abdominal drainage, colostomy GM
lum
2008 #1. Panperitonitis due to sigmoid colon perforation Closure of perforation, abdominal drainage, colostomy No treatment
#2. Ileus due to rectal cancer
2009 Colon necrosis due to ischemic colitis Abdominal drainage, transverse colectomy, sigmoidectomy, rTM + AT-III
colostomy
2009 Intestinal perforation Repair of perforated intestine and abdominal drainage GM
2009 Perforated ascending colon due to ileus with cecum cancer Abdominal drainage, ileostomy and colostomy GM
2009 Necrotic intestine due to superior mesenteric artery Resection of necrotic intestine, right hemi colectomy, ileos- AT-III
tomy
2010 #1. Traumatic intestinal multiple perforation Abdominal drainage, closure of perforated intestine, partial rTM + AT-III
#2. Sigmoid colon perforation resection of small intestine, sigmoid colectomy, colostomy
2010 Intestinal necrosis due to superior mesenteric vein throm- Abdominal drainage, massive small bowel resection, ileostomy rTM + AT-III
bosis
2010 Anastomotic leakage after ascendectomy Abdominal drainage, colostomy GM
2010 #1. Anastomotic leakage after total gastrectomy Thoracic cavity, mediastinal and abdominal drainage GM + AT-III
#2. Mediastinitis
2010 Traumatic perforated ileum Partial resection of ileum, abdominal drainage, ileostomy rTM
2011 Sigmoid colon perforation with diverticulitis Abdominal drainage, closure of perforation, colostomy rTM
2011 Duodenal perforation Abdominal drainage, closure of perforation rTM + AT-III
2011 Intestinal perforation due to ileus, carcinomatous perito- Abdominal drainage, colostomy GM + AT-III
nitis
2011 Ascending colon perforation due to strangulated ileus Right hemicolectomy AT-III
2012 Sigmoid colon perforation with diverticulitis Sigmoidectomy, colostomy rTM + AT-III
2012 Ascending colon perforation due to ileus colon cancer Right hemicolectomy, colostomy, abdominal drainage rTM + AT-III
2012 Ascending colon perforation with diverticulitis Abdominal drainage, ileostomy rTM + AT-III
2012 Traumatic perforation of small intestine Abdominal drainage, ileostomy rTM

13
Management of DIC in digestive surgical patients

acute DIC score and SOFA score were examined in both
group, from the day of treatment initiation through day 7.
Furthermore, we studied daily changes in aspartate ami-
notransferase (AST) and alanine aminotransferase (ALT)
levels as hepatic impairment parameters, creatinine (Cr) as
a renal function parameter, and C-reactive protein (CRP)
as an inflammation parameter. Table 2 shows the com-
parison between the rTM and GM groups concerning sex,
age, causal disease for surgery, concurrent drugs for DIC
treatment, PMX application, acute DIC score, and SOFA
score. No significant differences were observed between
the groups for any items, including concurrent use of AT-III
and PMX.
Similarly, no significant differences were observed for
the parameters of Plt, FDP, d-dimer, prothrombin time-
international normalized ratio (PT-INR), activated partial
thromboplastin time (APTT), CRP, serum Cr, AST, and
ALT at the time of starting drug administration (Table 3).
This study was retrospectively performed and conducted in
accordance with the ethical guidelines of the Declaration of
Helsinki and the International Conference on Harmoniza-
tion guidelines for good clinical practice. This study was
performed after approval from the ethical committee at
Kyushu University.
Statistics
Comparisons associated with the evolution of clinical
course were analyzed by the Wilcoxon signed-rank test;
comparisons between the groups were analyzed by the
Mann–Whitney U test. Comparisons of ratios between the
two groups were analyzed by Fisher’s exact probability
test. Acute DIC remission rate was analyzed by Kaplan–
Meier method and the difference between groups were ana-
lyzed with log-rank test. For all the analyses, p < 0.05 was
defined as significant.

Table 2  Characteristics of Factor rTM group (n = 10) GM group (n = 9) p value

patients treated with rTM or
GM Sex (male/female) 7/3 7/2 ns
Age (mean, range) 52.2 (22–83) 64.6 (28–85) ns
Cause of sepsis
Upper gastrointestinal perforation 1 1
Colon perforation 5 5
Colon necrosis 1 1
Intestinal necrosis 1 2
Anastomotic leakage 1 1
Duration of rTM or GM (days) 2–10 (ave. 5.3) 3–14 (ave. 6.5) ns
Co-therapy with AT-III 5 (50 %) 4 (44 %) ns
Co-therapy with PMX + CHDF 2 (20 %) 3 (33 %) ns
Use of vasopressor (CIa > 10) 5 (50 %) 5 (55 %) ns
a
CI catecholamine Average amount of infusion within 3 days (mL/day) 5,010 (3,850–6,510) 5,110 (4,210–7,230) ns
index = noradrenaline Acute DIC score 5.6 ± 1.3 6.7 ± 1.1 ns
10(γ) + dobutamine SOFA score 12.0 ± 3.2 11.4 ± 3.1 ns
(γ) + dopamine (γ)

Table 3  Laboratory data in rTM group (n = 10) GM group (n = 9) p value

patients treated with rTM or
GM Mean ± SD Range Mean ± SD Range

PLT (×104/µL) 6.7 ± 1.9 4.3–10.2 9.5 ± 4.9 2.9–11.6 0.3442

FDP (µg/mL) 25.1 ± 23.0 4.3–63.7 25.8 ± 16.7 8.1–57.9 0.7751
FIB (mg/dL) 295.2 ± 106.3 116–408 222.7 ± 154.8 116–511 0.3776
d-Dimer (µg/mL) 13.3 ± 10.9 2.0–26.3 23.7 ± 7.8 15.4–33.0 0.2207
PT–INR 1.42 ± 0.29 1.18–1.97 1.45 ± 0.24 1.0–1.75 0.4772
APTT-C (s) 33.1 ± 1.0 33.1–34.7 34.2 ± 0.9 32.8–35.4 0.7982
CRP (mg/dL) 16.72 ± 6.33 9.12–24.85 15.66 ± 11.87 1.11–33.83 0.8973
s-Cr (mg/dL) 1.33 ± 1.07 0.14–3.21 1.65 ± 1.17 0.70–3.73 0.6510
AST (IU/L) 125.3 ± 209.3 25–552 171.3 ± 236.6 16–714 0.5181
ALT (IU/L) 93.0 ± 174.7 14–449 138.6 ± 218.1 10–631 0.8973
Mann–Whitney U test

13
 T. Akahoshi et al.

Results in the GM group (16.6 ± 9.4 × 104/μL). No significant

Compared with day 0, the mean Plt counts on days 1–3
after surgery remained low in the GM group, but were sig-
nificantly increased in the rTM group (p = 0.04). Although
not statistically significant, the mean platelet count on day
7 was higher in the rTM group (18.4 ± 6.4 × 104/μL) than
differences were observed at any time point (Fig. 1a). PT–
INR (prothrombin time–international normalized ratio) and
fibrinogen values significantly changed after treatment ini-
tiation (Figs. 1b, 2b). Mean FDP level and d-dimer level
remained high (FDP >20 µg/mL and d-dimer >10 µg/mL)
until day 7 (Fig. 2b, c). A significant reduction in acute

Fig. 1  Effects of recombinant thrombomodulin (filled circle rTM, n = 10) or gabexate mesylate (open circle GM, n = 9) on platelet count and
PT (prothrombin time)–INR (results are given as mean ± SD). *p < 0.05 vs. 0 day

Fig. 2  Change in parameters related to coagulation following the administration of recombinant thrombomodulin (filled circle rTM) or gabexate
mesylate (open circle GM). There was no significant difference in either group

13
Management of DIC in digestive surgical patients

Fig. 3  Changes in the acute

disseminated intravascular
coagulation (DIC) score and
DIC remission rate. a Sig-
nificant reduction in acute DIC
score on day 3 in the recom-
binant thrombomodulin group
and on day 7 in the gabexate
mesylate (GM) group compared
with the levels on day 0. There
was no significant difference in
the scores between the groups.
b Cumulative DIC remis-
sion rate showed that the DIC
remission rate was significantly
earlier in the rTM group than in
the GM group (p = 0.02, log-
rank test)

DIC score was observed on day 3 in the rTM group and on
day 7 in the GM group, compared with the levels on day
0 (Fig. 3a). However, no differences were observed in the
scores between the groups. When acute DIC scores of ≤3
points were considered to indicate DIC remission, the rate
of early DIC remission was significantly earlier in the rTM
group (Fig. 3b).
A significant difference in the CRP level was observed
between the groups on days 2 and 3 (Fig. 4a). The CRP
level tended to be lower after surgery in the rTM group
than in the GM group. Cr levels did not significantly
decrease compared with baseline (day 0) levels in the GM
group, whereas Cr levels were significantly decreased in
the rTM group after day 3 (Fig. 4b). However, significant
differences were not observed between the groups at any
time point. No significant changes in the AST or ALT lev-
els were noted, and there were no significant differences
between the groups for these parameters (Fig. 4c, d).
SOFA score decreased significantly at days 3 and 7 com-
pared with that at day 0 in the rTM and GM groups. There
were no significant differences between the rTM and GM
groups at any evaluated day (Fig. 5).
No cases of hemorrhage resulting from indwelling cath-
eter drainage or surgical wound during administration of
GM and rTM were observed. Therefore, rTM administra-
tion was not discontinued because of hemorrhage in any
patient.
Days of ICU stay and prognosis
The mean duration of ICU stay was 5.5 ± 1.5 days in the
rTM group and 6.2 ± 1.8 days in the GM group (p > 0.05).
There were no cases of mortality within 28 days in the
present study. One patient in the rTM group died of res-
piratory dysfunction due to interstitial pneumonia 3 months
after discharge from the ICU. Another patient in the GM
group died of advanced colon cancer 2 months after dis-
charge from the ICU. Except for these two patients, all
patients were confirmed alive.
Discussion
Recombinant thrombomodulin binds to thrombin to inac-
tivate coagulation. The thrombin–rTM complex activates
protein C and is known to produce activated protein (APC),
which inactivates factors VIIIa and Va in the presence of
protein C. Gabexate mesilate (GM) is a synthetic serine
protease inhibitor, which inactivates coagulation [5, 6].
Gabexate mesilate is effective in patients with DIC associ-
ated with sepsis [7]. Before rTM became available in criti-
cal care units, gabexate mesilate was a popular drug for
treatment of DIC with sepsis in Japan.
rTM has been available in Japan since 2008, and is
covered by the national health insurance system for the
treatment of DIC. rTM has been used as a first-choice
drug in our ICU since 2008, for non-surgical patients
with DIC following severe infectious disease. However, in
2008, DIC was observed in four patients after gastroin-
testinal surgery, all of whom were treated with GM. rTM
was not used because of surgeons’ concerns over rTM-
associated postoperative hemorrhage. Subsequently, rTM
was administered to one of four patients (25 %) in 2009,
three of five patients (60 %) in 2010, and two of four
patients (50 %) in 2011. In 2012, all four patients with
postoperative DIC were treated with rTM. In the present

13

Fig. 4  Change in C-reactive protein (CRP), serum creatinine (s-Cr),
aspartate aminotransferase (AST), and alanine aminotransferase
(ALT) levels after treatment with recombinant thrombomodulin (filled
study, GM or rTM was started in all patients within 3 days
of surgery. The criterion for starting treatment was an
acute DIC score of >4. While there was no significant dif-
ference in Plt levels at the start of treatment, levels tended
to be lower in the rTM group. As for FDP and d-dimer
levels, which are associated with DIC, no significant dif-
ferences were observed between the groups. Plt counts
increased significantly earlier in the rTM group. Reflect-
ing this, the rate of early acute DIC score remission was
significantly better in the rTM group. Prompt DIC score
remission in the rTM group could be ascribed to improve-
ments in platelet counts.
Postoperative hemorrhage was not observed in any
patient treated with rTM; therefore, the administration
of rTM can be considered safe. The findings of particu-
lar interest were the significantly lower CRP values in the
rTM group on days 2 and 3 after administration, and the
reductions in serum Cr levels after administration of rTM.
It could be that CRP elevation is suppressed by the anti-
inflammatory activity of rTM in its direct inhibition of
HGMB-1 (a cell death mediator), as previously reported
[8, 9]. Furthermore, significant improvement was observed
for Cr levels compared with day 0 in the rTM group,
13
T. Akahoshi et al.
circle rTM) or gabexate mesylate (open circle GM) There was a sig-
nificant difference in CRP levels at days 2 and 3 between the rTM
and GM groups. *p < 0.05 vs. day 0, #p < 0.05 between groups
suggesting a possible renoprotective effect for rTM. Fur-
ther investigations are required on the therapeutic activities
of rTM, besides its efficacy in DIC.
In some instances in Japan, rTM was administered after
AT-III failed to improve patients’ condition [10–12]. Recent
experimental [13] and clinical [14] investigations have
demonstrated that co-therapy from the beginning would be
better in severe cases. Co-therapy with AT-III was used in
our hospital because of reports of good results with con-
current use, and also because of the severity of disease in
our ICU patients (as revealed by SOFA scores). Due to this
policy, many patients were treated with co-therapy.
In the present study, SOFA scores were used to deter-
mine severity. The SOFA score, introduced by Vincent
et al. [15], is simpler to calculate than the acute physiology
and chronic health evaluation (APACHE)-II score, and is
useful for predicting prognosis. SOFA score can also assess
the effects of artificial respirators or vasopressors over time.
Because the duration of a high SOFA score, in particular, is
strongly correlated with prognosis, medical intervention is
required to promptly improve SOFA scores [16, 17]. Based
on these considerations, we used SOFA score as the indica-
tor for therapeutic efficacy.
Management of DIC in digestive surgical patients
Fig. 5  Sepsis-related organ failure assessment (SOFA) score
decreased significantly on days 3 and 7 compared with that on day 0
in the recombinant thrombomodulin (filled circle rTM) and gabexate
mesylate (open circle GM) groups. There were no significant differ-
ences between the rTM and GM groups on any evaluated day
One of the advantages of rTM lies in its administration
method. GM and heparin, which have been used for DIC
treatment, require 24-h continuous intravenous infusion;
however, drip infusion of rTM can be completed within 1 h.
It is therefore easier to secure the drip infusion route, which
is usually provided for catecholamine or various other
drugs in ICU. This relative ease of administration provides
a manageability advantage for rTM. However, no clear sig-
nificant advantage of rTM has been confirmed regarding
prognosis, compared with the existing DIC drugs, as shown
in an earlier report [4]. Further efforts are needed to accu-
mulate relevant case data, and basic science and clinical
investigations are needed into rTM’s anti-inflammatory and
renoprotective effects, if its superiority is to be confirmed.
At the same time, economic aspects of treatment should be
studied.
Our retrospective study only had a small number of
cases, therefore, we cannot conclude that rTM is a first-
choice drug for DIC. However, we demonstrated that rTM
is effective against DIC following emergency gastrointes-
tinal surgery, because of its ability to improve promptly
DIC parameters without accelerating hemorrhage that may
occur soon after surgery.
Acknowledgments All authors contributed significantly to this
study and agree with content of the manuscript.
Conflict of interest Tomohiko Akahoshi, Hiroshi Sugimori, Nori-
yuki Kaku, Kentaro Tokuda, Eiichiro Noda, Takashi Nagata, Masaru
Morita, Makoto Hashizume, and Yoshihiko Maehara, have no con-
flicts of interest and received no financial support for this study.
Compliance with ethical requirements The authors comply with
the ethical guidelines for authorship and publishing in the European
Journal of Trauma and Emergency Surgery.
References
1. Kosugi C, Yasuda H, Koda K, et al. Evaluation of the usefulness
of APACHE II, SOFA and POSSUM scores for the postoperative
risk assessment in patients with colorectal perforation. JAEM.
2009;29(5):697–702.
2. Chin S, Fugita R, Kawa Y, et al. Prognostic factors and postop-
erative complications in patients with colorectal perforation. Jpn
J Gastroenterol Surg. 2010;43:1007–13.
3. Saito H, Maruyama I, Shimazaki S, et al. Efficacy and safety
of recombinant human soluble thrombomodulin (ART-123) in
disseminated intravascular coagulation: results of a phase III,
randomized, double-blind clinical trial. J Thromb Haemost.
2007;5:31–41.
4. Aikawa N, Shimazaki S, Yamamoto Y, et al. Thrombomodulin
alfa in the treatment of infectious patients complicated by dis-
seminated intravascular coagulation: subanalysis from the phase
3 trial. Shock. 2011;35:349–54.
5. Van de Wouwer M, Collen D. Conway EM: Thrombomodulin-
protein CEPCR system: integrated to regulate coagulation and
inflammation. Arterioscler Thromb Vasc Biol. 2004;24:1374–83.
6. Tamura Y, Hirado M, Okamura K, Minato Y, Fujii S. Synthetic
inhibitors of trypsin, plasmin, kallikrein, thrombin, C1r- and C1
esterase. Biochim Biophys Acta. 1977;484:417–22.
7. Taenaka N, Shimada Y, Hirata T, Nishijima MK, Takezawa J,
Yoshiya I, Kambayashi J. Gabexate mesilate (FOY) therapy of
disseminated intravascular coagulation due to sepsis. Crit Care
Med. 1983;11:735–8.
8. Ito T, Kawahara K, Okamoto K, et al. Proteolytic cleavage of high
mobility group box 1 protein by thrombin-thrombomodulin com-
plexes. Arterioscler Thromb Vasc Biol. 2008;28(10):1825–30.
9. Abeyama K, Stern DM, Ito Y, et al. The N-terminal domain
of thrombomodulin sequesters high-mobility group-B1 pro-
tein, a novel antiinflammatory mechanism. J Clin Investig.
2005;115(5):1267–74.
10. Eguchi Y. Efficacy of recombinant thrombomodulin in antithrom-
bin substitution-resistant septic disseminated intravascular
coagulation: results of a pilot study. J Jpn Soc Surg Infect.
2010;7(2):143–8.
11. Yada N, Nishio K, Seki T, et al. Therapeutic effect of recom-
binant thrombomodulin on the hemostatic and inflammatory
response in patients with septic DIC. J Jpn Assoc Acute Med.
2011;22:749–57.
12. Nakahara K, Okuse C, Adachi S, et al. Use of antithrombin and
thrombomodulin in the management of disseminated intravas-
cular coagulation in patients with acute cholangitis. Gut Liver.
2013;7(3):363–70.
13. Iba T, Nakarai E, Takayama T, et al. Combination effect of
antithrombin and recombinant human soluble thrombomodu-
lin in a lipopolysaccharide induced rat sepsis model. Crit Care.
2009;13:R203.
14. Sawano H, Shigematsu K, Yoshinaga Y, et al. Combined
therapy with antithrombin and recombinant human soluble
13
 T. Akahoshi et al.

thrombomodulin in patients with severe sepsis and disseminated 17. Cabre L, Mancebo J, Solsona JF, et al. Multicenter study of the
intravascular coagulation. JJAAM. 2013;24:119–31. multiple organ dysfunction syndrome in intensive care units: the
15. Vincent JI, Moreno R, Takala J, et al. The SOFA (sepsis-related usefulness of sequential organ failure assessment scores in deci-
organ failure assessment) score to describe organ dysfunction/ sion making. Intensive Care Med. 2005;31:927–33.
failure. Intensive Care Med. 1996;22:707–10.
16. Ferreira FL, Bota DP, Bross A, et al. Serial evaluation of the
SOFA score to predict outcome in critically ill patients. JAMA.
2001;286:1754–8.

13