Soluble form of the receptor for advanced glycation end products

is a marker of acute lung injury but not of severe sepsis in

critically ill patients*
Matthieu Jabaudon, MD; Emmanuel Futier, MD; Laurence Roszyk, PharmD; Elodie Chalus, PharmD;
Renaud Guerin, MD; Antoine Petit, MD; Segolene Mrozek, MD; Sebastien Perbet, MD; Sophie Cayot-Constantin, MD;
Christian Chartier, MD; Vincent Sapin, PharmD, PhD; Jean-Etienne Bazin, MD, PhD;
Jean-Michel Constantin, MD, PhD

Objectives: Levels of the soluble form of the receptor for ad-
vanced glycation end products (sRAGE) are elevated during acute
lung injury. However, it is not known whether this increase is linked
to its involvement in alveolar epithelium injury or in systemic inflam-
mation. Whether sRAGE is a marker of acute lung injury and acute
respiratory distress syndrome, regardless of associated severe sepsis or
septic shock, remains unknown in the intensive care unit setting.
Design: Prospective, observational, clinical study.
Setting: Intensive care unit of an academic medical center.
Patients: A total of 64 consecutive subjects, divided into four
groups: acute lung injury/acute respiratory distress syndrome
(n ⴝ 15); acute lung injury/acute respiratory distress syndrome
plus severe sepsis/septic shock (n ⴝ 18); severe sepsis/septic
shock (n ⴝ 16); and mechanically ventilated controls (n ⴝ 15).
Interventions: None.
Measurements and Main Results: Plasma sRAGE levels were
measured at baseline and on days 3, 6, and 28 (or at intensive care
unit discharge, whichever occurred first). Baseline plasma levels of
sRAGE were significantly higher in patients with acute lung injury/
acute respiratory distress syndrome, with (median, 2951 pg/mL) or
without (median, 3761 pg/mL) severe sepsis, than in patients with
severe sepsis (median, 488 pg/mL) only and in mechanically venti-
lated controls (median, 525 pg/mL). Levels of sRAGE were correlated
with acute lung injury/acute respiratory distress syndrome severity
and decreased over time but were not associated with outcome.
Lower baseline plasma sRAGE was associated with focal loss of
aeration based on computed tomography lung morphology.
Conclusions: sRAGE levels were elevated during acute lung
injury/acute respiratory distress syndrome, regardless of the
presence or absence of severe sepsis. The plasma level of sRAGE
was correlated with clinical and radiographic severity in acute
respiratory distress syndrome patients and decreased over time,
suggesting resolution of the injury to the alveolar epithelium.
Further study is warranted to test the clinical utility of this
biomarker in managing such patients and to better understand its
relationship with lung morphology during acute lung injury/acute
respiratory distress syndrome. (Crit Care Med 2011; 39:480 – 488)
KEY WORDS: receptor for advanced glycation end products;
acute lung injury; acute respiratory distress syndrome; biomarker;
intensive care unit; severe sepsis

A cute lung injury (ALI) and acute
respiratory distress syndrome
(ARDS) remain important clini-
cal and economic challenges in
the intensive care unit (ICU), with severe
complications such as multiorgan failure
and high mortality rates (1). The lack of spe-
cific diagnostic tools continues to hamper the
development of improved therapies for this
syndrome. The internationally accepted diag-
*See also p. 589.
From the Department of Anesthesiology and
Critical Care Medicine (MJ, EF, RG, AP, SM, SP, SCC,
CC, JEB, JMC), Intensive Care Unit, and Dep-
artment of Medical Biochemistry and Molecular Bi-
ology (LR, EC, VS), Estaing University Hospi-
tal, CHU Clermont-Ferrand, Clermont-Ferrand,
France.
Supported, in part, by grants from the Delegation à
la Recherche Clinique et à l’Innovation d’Auvergne,
Ministry of Health, France.
nostic criteria for ALI and ARDS (2) are non-
specific to the point of including patients with
relatively mild hypoxia and those with lung
pathology that might be different from the
typical diffuse alveolar damage.
The receptor for advanced glycation
end products (RAGE) is a member of the
immunoglobulin superfamily that acts as
a multiligand receptor and is involved in
propagating the inflammatory response (3,
Trial Registration: clinicaltrials.gov Identifier:
NCT00811629.
The authors have not disclosed any potential con-
flicts of interest.
For information regarding this article, E-mail:
mjabaudon@chu-clermontferrand.fr
Copyright © 2011 by the Society of Critical Care
Medicine and Lippincott Williams & Wilkins
DOI: 10.1097/CCM.0b013e318206b3ca
4). Even in the physiologic state, pulmo-
nary tissue expresses relatively high levels
of RAGE. RAGE expression in the lung is
primarily located on the basal surface of
alveolar type 1 cells (5, 6). Consistent with
these findings, the ligand–RAGE axis has
been suggested to play a relevant role in
pulmonary pathophysiology (7). Soluble
RAGE (sRAGE), consisting of the extracel-
lular domain but lacking the transmem-
brane and cytoplasmic domains of RAGE,
has been described in human plasma. Cir-
culating sRAGE either is produced by re-
ceptor ectodomain shedding (8, 9) or rep-
resents splice variants of RAGE, such as
endogenous secreted RAGE (10).
A growing body of evidence suggests
that local and circulating levels of sRAGE
may be considered as biomarkers of pul-
monary tissue damage. sRAGE levels
have been reported to be significantly

480 Crit Care Med 2011 Vol. 39, No. 3

higher in pulmonary edema fluid and
plasma from patients with ARDS com-
pared to that from healthy volunteers and
patients with hydrostatic pulmonary
edema (11). Furthermore, in a large, ran-
domized, controlled trial of lower tidal
volume ventilation in ARDS (12, 13),
baseline plasma sRAGE levels were asso-
ciated with clinical outcome in patients
randomized to higher tidal volumes (12
mL/kg predicted body weight).
Although the precise function of RAGE
remains unclear, it is important to empha-
size that nonpulmonary sources of RAGE
may contribute to plasma levels in the set-
ting of ARDS and that evidence is mount-
ing that RAGE plays an important role in
regulating acute and chronic inflammation
(14, 15). Thus, elevated sRAGE levels have
been described in the plasma of surgical
ICU patients with severe sepsis or septic
shock compared with healthy volunteers
(16). However, it cannot be excluded that
the increase in sRAGE levels observed dur-
ing ARDS derives in part from the role of
sRAGE in the inflammatory cascade rather
than from its release from alveolar type 1
cells. This clinical study was performed to
determine whether plasma levels of sRAGE
are elevated in ICU patients with ARDS,
regardless of associated severe sepsis or
septic shock.
MATERIALS AND METHODS
Patients. Sixty-four consecutive mechani-
cally ventilated patients admitted to the general
ICU of Clermont-Ferrand University Hospital,
France, between January and July 2009 were
prospectively enrolled in the study within 24 hrs
of disease onset. Patients with ALI or ARDS were
identified based on the American European Con-
sensus Conference definitions (2), and patients
with severe sepsis or septic shock were identified
based on the American College of Chest Physi-
cians/Society of Critical Care Medicine Consen-
sus Conference (17). Fifteen mechanically ven-
tilated patients served as controls and were
recruited from patients who did not meet the
severe sepsis, septic shock, or ALI/ARDS criteria
of our study but who were sufficiently ill to
require mechanical ventilation. Patients were
excluded if: they were ⬍18 yrs old or pregnant;
they had a history of acute exacerbation of dia-
betes, dialysis for end-stage kidney disease, Alz-
heimer’s disease, amyloidosis, or evolutive solid
neoplasm; or their attending physicians were
against aggressive treatment at the time of en-
rollment. The Institutional Review Board of the
hospital and the French Ministry of Health ap-
proved the study protocol. Informed written
consent was obtained from all participants.
Study Design. Plasma levels of sRAGE in
ALI/ARDS patients were compared with those
in patients with severe sepsis or septic shock,
Crit Care Med 2011 Vol. 39, No. 3
associated with ALI/ARDS or not, and in me-
chanically ventilated patients on days 0, 3, 6,
and 28 (or at discharge from the ICU, which-
ever occurred first) after inclusion. Pulmonary
edema fluid levels of sRAGE were measured in
patients with ALI/ARDS, associated with se-
vere sepsis or not, on days 0, 3, and 6. Inten-
sive care management of patients included in
the study was conducted using our ICU-based
standard protocols. Thus, mechanical ventila-
tion strategy (including weaning strategy),
sepsis management, and the use of sedative
agents were based on currently available
guidelines (13, 18). A lung-protective ventila-
tion strategy was applied; a tidal volume of 6
mL/kg (predicted) body weight and a pressure
plateau of ⬍30 cmH2O were targeted in all
mechanically ventilated patients (13).
Clinical Data. After inclusion in the study,
clinical data and biological characteristics
were obtained for all living patients at all time
points. According to our institutional proto-
col, a lung computed tomography (CT) scan
was performed on day 0 in ALI/ARDS patients,
as previously described (19). Two independent
radiologists performed the qualitative CT anal-
ysis according to the “CT scan ARDS study
group” criteria (20 –22). Three patterns of loss
of aeration distribution were identified: focal
or lobar, diffuse, and patchy. Nonfocal lung
morphology was noted for patients with dif-
fuse or patchy loss of aeration (23). Clinical
outcome was recorded until day 28 or dis-
charge from the ICU, whichever occurred first.
Laboratory Data. Blood samples were col-
lected from all patients at all time points and
undiluted pulmonary edema fluid was collected
simultaneously in intubated ALI/ARDS patients
only on days 0, 3, and 6, as described previously
(24). sRAGE concentrations were measured in
duplicate using a commercially available sand-
wich enzyme-linked immunosorbent assay kit
(R&D Systems, Minneapolis, MN).
Statistical Analyses. Variables were tested
for normal distribution by the one-sample Kol-
mogorov-Smirnov test. The Mann-Whitney U
test (comparing two independent samples of ob-
servations) and the Kruskal-Wallis test, with
Dunn’s multiple-comparison test for compari-
son among groups, were used to determine the
significance of differences between subject
groups concerning nonparametric data (sRAGE
levels, gravity scores, number of ventilator-free
days, and other data). Bonferroni correction was
applied after analysis of variance for multiple
comparison concerning parametric data (age,
PaO2/FIO2 ratio, tidal volume, plateau pressure,
PaCO2, respiratory rate, arterial pH, and other
data). An overall p value for multiple group tests
is reported, as well as corrected p values for
intergroup comparisons. Proportions were com-
pared among groups using semi-parametric test
(chi-square test), and categorical data that result
from classifying objects in two different ways
were analyzed by Fisher’s exact test. Receiver-
operating characteristic curve was computed
and area under the curve was used to evaluate
how well the model distinguished ALI/ARDS pa-
tients with focal and nonfocal loss of aeration
based on CT scan lung morphology. Confidence
intervals (CIs) for areas under receiver-operating
characteristic curves were calculated using non-
parametric assumptions. To assess the relation-
ship between two variables, Pearson’s correla-
tion analysis was used for variables that followed
a normal distribution, and Spearman’s rank-
correlation coefficient analysis was used for
those that did not. All analyses were performed
using MedCalc version 10.1.2.0 (Frank
Schoonjans, Mariakerke, Belgium). A p ⬍ .05
was considered statistically significant.
RESULTS
Baseline Patient Characteristics
A total of 64 consecutive mechanically
ventilated patients were enrolled in the
study. Patients were classified into four
groups according to the ALI/ARDS and
Sepsis Consensus Conference definitions:
patients with ALI/ARDS (ALI/ARDS group,
n ⫽ 15); patients with ALI/ARDS associated
with severe sepsis or septic shock (ALI/
ARDS plus severe sepsis group, n ⫽ 18);
patients with severe sepsis or septic shock
without associated ALI/ARDS (severe sepsis
group, n ⫽ 16); and patients without ALI/
ARDS or severe sepsis criteria (control
group, n ⫽ 15). Table 1 summarizes the
basic demographics and clinical character-
istics of these patients. Patients with severe
sepsis, associated with ALI/ARDS or not,
had higher baseline severity-of-illness
scores (Acute Physiology and Chronic
Health Evaluation II, Sequential Organ
Failure Assessment) than patients without
severe sepsis. None of the patients from the
control group had systemic inflammatory
response syndrome criteria at the time of
inclusion. Crude mortality rates at 28 days
were lower in the control group than in the
severe sepsis group and ALI/ARDS plus se-
vere sepsis group, but no difference in
other baseline demographics, clinical char-
acteristics, comorbidities known to possibly
influence sRAGE levels, or other outcome
variables were observed between the four
groups. The groups were similar with re-
spect to disease etiology, except for pneu-
monia, trauma, and neurovascular disease.
Baseline respiratory and ventilator param-
eters are summarized in Table 2.
sRAGE Levels Are High in
Plasma From Patients With ALI/
ARDS Regardless of Associated
Severe Sepsis or Septic Shock
Plasma sRAGE levels were significantly
higher on day 0 in the ALI/ARDS group
481
Table 1. Baseline demographics, clinical characteristics, comorbidities, and main outcome variables

Acute Lung Injury/ Acute Lung Injury/Acute

Acute Respiratory Respiratory Distress
Demographic or Distress Syndrome Syndrome ⫹ Severe Severe Sepsis Control Group Overall Multiple
Characteristic Group (n ⫽ 15) Sepsis Group (n ⫽ 18) Group (n ⫽ 16) (n ⫽ 15) Group Tests (p)

Age (yrs) 58.7 (12.7) 59.2 (16.0) 65.3 (9.6) 52.7 (19.3) .1
Male, n (%) 8 (53.3) 14 (77.8) 10 (62.5) 9 (60) .5
Acute Physiology and 28 关26–32兴 32 关29–36兴 30 关27–34兴 22 关19–25兴 ⬍.001
Chronic Health d
p ⫽ .0003 vs. C a
p ⫽ .02 vs. A d
p ⫽ .0003 vs. C
Evaluation II score a
p ⬍ .0001 vs. C
Sequential Organ Failure 11 关8.5–11兴 12 关11–13兴 9.5 关9–12.5兴 8 关4–9兴 ⬍.0001
Assessment score d
p ⫽ .001 vs. C a
p ⫽ .013 vs. A d
p ⫽ .001 vs. C
d
p ⬍ .0001 vs. C
Coexisting conditions, n (%)
Diabetes 1 (6.7) 0 (0) 4 (25) 1 (6.7) .08
Hypertension and 5 (33.3) 6 (33.3) 6 (37.5) 2 (13.3) .5
vascular disorder
Chronic renal failure 0 (0) 0 (0) 2 (12.5) 0 (0) .1
Hematologic neoplasms 1 (6.7) 1 (6.7) 3 (18.8) 0 (0) .3
Respiratory diseases 4 (26.7) 2 (11.1) 4 (25) 2 (13.3) .6
Tobacco smoking 4 (26.7) 6 (33.3) 6 (37.5) 2 (13.3) .6
Main outcome variables
28-day mortality (%) 20 55.6 50 6.7 .02
d
p ⫽ .0344 vs. C c
p ⫽ .04 vs. severe sepsis
group
N of ventilator-free 14 关1–15兴 0 关0–0兴 3 关0–18兴 12 关0–22兴 .08
days at day 28
N of intensive care unit-free 8 关0–14兴 0 关0–0兴 0 关0–14兴 0 关0–12) .2
days at day 28
N of proven or probable 7 (46.7) 4 (22.2) 3 (18.8) 2 (13.3) .2
cases of ventilator-
associated pneumonia
within 28 days (%)
Etiology, n (%)
Pneumonia 8 (53) 6 (33) 3 (19) 0 (0) .008
d
p ⫽ 0.01 vs. C
Aspiration 3 (20) 2 (11) 0 (0) 1 (7) .3
Trauma 1 (7) 1 (6) 0 (0) 6 (40) .003
d
p ⬍ .006 vs. C
Peritonitis/pancreatitis 3 (20) 7 (39) 7 (44) 1 (7) .2
Catheter-related 0 (0) 2 (11) 3 (19) 0 (0) .1
bloodstream infection
Necrotizing fasciitis 0 (0) 0 (0) 3 (19) 0 (0) .2
Neurovascular disease 0 (0) 0 (0) 0 (0) 7 (47) ⬍.0001
a
p ⫽ .03 vs. A
b
p ⫽ .03 vs. acute lung
injury/acute respiratory
distress syndrome ⫹ severe
sepsis group
c
p ⫽ .03 vs. severe sepsis
group

A, acute lung injury/acute respiratory distress syndrome group; C, control group.

a,b,c,d
The p values are written corrected. The Mann-Whitney U test and the Kruskal-Wallis test with Dunn’s multiple-comparison test, when needed, were
used to determine the significance of differences between subject groups concerning nonparametric data. Bonferroni correction was applied after analysis
of variance for multiple comparisons concerning parametric data. Data are presented as mean ⫾ SD and are analyzed with one-way analysis of variance
(parametric data) or median 关interquartile range兴 and are analyzed with Kruskal-Wallis test (nonparametric data). Proportions were compared among
groups using semiparametric test (␹2 test), and categorical data that result from classifying objects in two different ways were analyzed by Fisher’s exact
test. The number of ventilator-free days is the median number 关interquartile range兴 of days between day 0 and day 28 on which the patient had breathed
without assistance for at least 48 consecutive hours. Percentages may not exactly total 100% because of rounding.

(median [interquartile range (IQR)], 3761
pg/mL [3317– 4892]) and in the ALI/ARDS
plus severe sepsis group (2951 pg/mL
[1335–5285]) than in the severe sepsis
group (488 pg/mL [366 – 696]; p ⬍ .0001
for both ALI/ARDS groups) and control
group (525 [387– 671] pg/mL; p ⬍ .0001
and p ⫽ .0001,respectively) (Fig. 1). Base-
line plasma sRAGE levels were similar in
the ALI/ARDS groups, associated with se-
vere sepsis or septic shock or not (p ⫽ .4),
and in the groups without ALI/ARDS (p ⫽
.9). The overall p was ⬍.0001 when com-
paring baseline plasma sRAGE levels be-
tween the four different groups. Baseline

482 Crit Care Med 2011 Vol. 39, No. 3

Table 2. Baseline respiratory and ventilator parameters

Acute Lung Injury/Acute

Acute Lung Injury/Acute Respiratory Distress
Respiratory Distress Syndrome ⫹ Severe Sepsis Severe Sepsis Control Group Overall Multiple
Parameter Syndrome Group (n ⫽ 15) Group (n ⫽ 18) Group (n ⫽ 16) (n ⫽ 15) Group Tests (p)

PaO2/FIO2 (torr) 124 (62) 144 (64) 310 (99) 321 (102) ⬍.001
a
p ⬍ .001 vs. S a
p ⬍ .0001 vs. S
b
p ⬍ .0001 vs. C b
p ⬍ .0001 vs. C
Tidal volume (mL/kg of 7.5 (1.3) 6.9 (1.4) 7.5 (1.4) 7.9 (1.6) .3
ideal body weight)
Positive end-expiratory 10 关7.3–12.8兴 12 关8–15兴 9 关7–10兴 bp ⫽ .007 vs. C 7 关5–7兴 ⬍.0001
pressure (cm of water) b
p ⫽ .0005 vs. C a
p ⫽ .015 vs. S
b
p ⬍ .0001 vs. C
Lung injury score 3 关2.8–3.3兴 3 关2.5–3.3兴 1.3 关1–1.5兴 1 关0.6–1.3兴 ⬍.0001
a
p ⬍ .0001 vs. S a
p ⬍ .0001 vs. S
b
p ⬍ .0001 vs. C b
p ⬍ .0001 vs. C
Plateau pressure 26.4 (3) 27.9 (3.8) 25 (5) 24.6 (4) .09
(cm of water)
PaCO2 (torr) 45.1 (12.7) 49.8 (12.2) 37.2 (8.2) 37.4 (5.9) .001
b
p ⫽ .001 vs. C a
p ⫽ .001 vs. S
b
p ⫽ .002 vs. C
Respiratory rate 24 关19–30兴 26 关20–30兴 20 关16–24兴 18 关15–20兴 .004
(breaths/min) b
p ⫽ .04 vs. C a
p ⫽ .01 vs. S
b
p ⫽ .001 vs. C
Arterial pH 7.37 (0.08) 7.25 (0.12) 7.35 (0.10) 7.42 (0.06) ⬍.001
b
p ⫽ .002 vs. acute lung a
p ⫽ .002 vs. S
injury/acute respiratory b
p ⫽ .0001 vs. C
distress syndrome ⫹ severe
sepsis group

C, control group; S, severe sepsis group.

a,b
The p values are written corrected. The Mann-Whitney U test and the Kruskal-Wallis test with Dunn’s multiple-comparison test, when needed, were
used to determine the significance of differences between subject groups concerning nonparametric data. Bonferroni correction was applied after analysis
of variance for multiple comparisons concerning parametric data. Data are presented as mean ⫾ SD and analyzed with one-way analysis of variance
(parametric data) or median 关interquartile range兴 and analyzed with Kruskal-Wallis test (nonparametric data). Lung injury scores range from 0 to 4, with
higher scores indicating more severe lung injury.

plasma sRAGE levels in patients from the

ALI/ARDS plus severe sepsis group with
direct ARDS (n ⫽ 8) were similar to those
in patients from the ALI/ARDS plus se-
vere sepsis group with indirect ARDS
(n ⫽ 10) (3082 [1115– 6253] vs. 2654
pg/mL [2066 –5041]; p ⫽ .9). To verify
that the differences in sRAGE levels ob-
served between ALI/ARDS and non-ALI/
ARDS patients did not reflect the differ-
ing prevalence of direct lung injury
(pneumonia or aspiration) in these two
groups, we compared baseline plasma
sRAGE levels in indirect ALI/ARDS pa-
tients (without pneumonia or aspiration)
to those in severe sepsis patients without
pneumonia or aspiration using nonpara-
metric statistical test. Plasma sRAGE lev-
Figure 1. Box and whisker plots of baseline form of the receptor for advanced glycation end products els were higher on day 0 in the ALI/ARDS
(sRAGE) levels in patients with acute lung injury/acute respiratory distress syndrome (ALI/ARDS patients without pneumonia or aspiration
group), ALI/ARDS and severe sepsis or septic shock (ALI/ARDS plus severe sepsis group), severe sepsis (n ⫽ 14) than in the patients without
or septic shock (severe sepsis group), and mechanically ventilated patients without ALI/ARDS or severe
pneumonia or aspiration from the severe
sepsis/septic shock (control group). Boxes show medians with interquartile ranges, error bars indicate
10th–90th percentiles, and ⌬ represents “extreme” outliers (*p ⬍ .05 vs. severe sepsis group and sepsis group (n ⫽ 13; median [IQR], 3996
control group). Lower plasma sRAGE levels have been measured in a neutropenic patient from the [2070 –5369] vs. 581 pg/mL [345–741];
ALI/ARDS plus severe sepsis group and in the plasma from six ALI/ARDS patients (with and without p ⫽ .0002). The area under the receiver-
severe sepsis) with focal damage based on computed tomography scan findings. operating characteristic curve when plasma

Crit Care Med 2011 Vol. 39, No. 3 483

 Baseline sRAGE Levels Are
High in Plasma From Patients
With “Nonfocal” ALI/ARDS
Based on CT Scan Lung
Morphology
CT scan lung morphology was re-
corded on day 0 in all 33 ALI/ARDS pa-
tients. Plasma sRAGE levels were signifi-
cantly higher in the 27 ALI/ARDS
patients with “nonfocal” loss of aeration
(median [IQR], 4306 pg/mL [3238 –
5238]) than in the six patients with “fo-
cal” loss of aeration (1115 pg/mL [1041–
1339]; p ⫽ .0007) (Fig. 3). Pulmonary
edema fluid levels of sRAGE were higher
in patients with nonfocal damage (3.4 ⫻
105 pg/mL [0.4 ⫻ 105 to 14.3 ⫻ 105])
than in those with “focal” damage (1.5 ⫻
105 pg/mL [0.2 ⫻ 105 to 3.4 ⫻ 105]), but
this difference did not reach significance
(p ⫽ .3). The edema fluid-to-plasma ratio
of sRAGE was similar in the focal sub-
group (104 [16 –323]) and in the nonfocal
Figure 2. Form of the receptor for advanced glycation end products (sRAGE) levels in patients with acute subgroup (94 [10 – 447]; p ⫽ .89). The
lung injury/acute respiratory distress syndrome (ALI/ARDS group), ALI/ARDS and severe sepsis or septic ability of baseline plasma sRAGE levels
shock (ALI/ARDS plus severe sepsis group), severe sepsis or septic shock (severe sepsis group), and
from patients with ALI/ARDS to discrim-
mechanically ventilated patients without ALI/ARDS or severe sepsis/septic shock (control group) on days 0,
3, 6, and 28 (or at discharge from the intensive care unit, whichever occurred first). Values are expressed
inate between the types of loss of aeration
as median (interquartile) (*p ⬍ .05 vs. severe sepsis group and control group). based on CT scan lung morphology was
determined. The area under the receiver-
Table 3. Levels of soluble receptor for advanced glycation end products in pulmonary edema fluid
operating characteristic curve when
plasma sRAGE was used to differentiate
Acute Lung Injury/Acute Acute Lung Injury/Acute Respiratory the presence from absence of nonfocal
Respiratory Distress Distress Syndrome ⫹ Severe loss of aeration was 0.951 (95% CI,
Day Syndrome Group Sepsis Group p 0.813– 0.993; p ⫽ .0001) (Fig. 4). A cut-
off value of 2066 pg/mL had a sensitivity
0 3.9 关0.8–10.6兴 (n ⫽ 15) 1.3 关0.2–6.8兴 (n ⫽ 18) .1 of 93% (95% CI, 75.7–98.9) and a speci-
3 0.1 关0.08–0.5兴 (n ⫽ 14) 0.07 关0.02–0.3兴 (n ⫽ 15) .2
6 0.09 关0.02–0.8兴 (n ⫽ 11) 0.01 关0.008–0.05兴 (n ⫽ 10) .03
ficity of 100% (95% CI, 54.1–100).

Results are presented as median 关interquartile range兴. Numbers of patients tested at each time Relation Between SRAGE Levels
point are listed (n). Levels of soluble form of the receptor for advanced glycation end products are and ALI/ARDS Severity
expressed in ⫻105 pg/mL.
The Kruskal-Wallis test was used to determine the significance of differences between subject The correlation between sRAGE levels
groups concerning nonparametric data. and several available indicators of lung
injury was determined in all ALI/ARDS
sRAGE was used to differentiate the pres- (p ⫽ .001), day 6 (p ⫽ .0001), and day patients, with and without associated se-
ence from absence of ALI/ARDS in patients 28 (p ⫽ .0002), but no difference was vere sepsis or septic shock. Higher base-
with severe sepsis or septic shock was 0.944 found between other time points. There line levels of plasma sRAGE were associ-
(95% CI, 0.807– 0.992; p ⫽ .0001). A cut-off was no difference in plasma sRAGE lev- ated with more severe ALI/ARDS, as
value of 980 pg/mL had a sensitivity of els over time in the severe sepsis group reflected by measures of pulmonary phys-
94.4% (95% CI, 72.6 –99.1) and a specificity (p ⫽ .2) or control group (p ⫽ .7). iology (Fig. 5). There was no difference in
of 100% (95% CI, 79.2–100). baseline plasma and alveolar sRAGE lev-
Repeated-measures analysis of variance re-
els with respect to the cause of lung in-
vealed that both the within-subjects factor
Evolution of sRAGE Levels Over jury, such as aspiration or pneumonia,
(time) and the dependent variable (group) and to coexisting conditions, such as di-
Time significantly influenced plasma sRAGE lev- abetes, end-stage renal disease, essential
In ALI/ARDS patients, plasma sRAGE els (p ⬍ .001 for both factors). hypertension, or other vascular disease.
levels decreased in a significant manner The edema fluid levels of sRAGE were Baseline sRAGE levels in the pulmonary
from baseline to day 28 (Fig. 2). In similar in both ALI/ARDS groups on day 0 edema fluid from patients with indirect
patients with ALI/ARDS and severe sep- and on day 3, but they were higher in and direct ALI/ARDS were similar (me-
sis, plasma sRAGE levels were signifi- patients without associated severe sepsis dian [IQR], 1.5 ⫻ 105 pg/mL [0.3–7.1] vs.
cantly higher on day 0 than on day 3 on day 6 (Table 3). 3.4 ⫻ 105 pg/mL [0.6 –10.6]; p ⫽ .3).

484 Crit Care Med 2011 Vol. 39, No. 3


Figure 3. Box and whisker plots of baseline form of the receptor for advanced glycation end products
(sRAGE) levels in acute lung injury/acute respiratory distress syndrome (ALI/ARDS) patients with focal
and nonfocal loss of aeration, as revealed by computed tomography scan lung morphology on day 0.
Boxes show medians with interquartile ranges; error bars indicate 10th–90th percentiles.
groups except for a lower mortality rate
in the control group. In ALI/ARDS pa-
tients overall, with and without associ-
ated severe sepsis, baseline levels of
plasma sRAGE were similar in 28-day
survivors and nonsurvivors (median
[IQR], 3504 pg/mL [1887– 4801] vs. 3618
pg/mL [2377–5774], respectively; p ⫽ .6).
Baseline plasma and alveolar sRAGE were
not associated with 28-day mortality rate,
number of ventilator free-days at day 28,
number of ICU-free days at day 28, rate of
recurrence or onset of septic shock, or
rate of ventilator-associated pneumonia
in ALI/ARDS patients.
DISCUSSION
Figure 4. Receiver-operating characteristic curve
of baseline plasma soluble form of the receptor The results of the present prospective
for advanced glycation end product levels in dif- observational study show that: 1) sRAGE
ferentiating between the presence and absence of
levels were high in plasma from patients
nonfocal loss of aeration during computed to-
mography scan morphology studies on day 0 in
with ALI/ARDS, regardless of the pres-
acute lung injury/acute respiratory distress syn- ence or absence of severe sepsis, and 2)
drome patients. The area under the receiver- levels of sRAGE were positively correlated
operating characteristic curve was 0.951 (95% with increased severity of pulmonary tissue
CI, 0.813– 0.993; p ⫽ .0001) for a cut-off value of injury and declined over time in patients
2066 pg/mL, with a sensitivity of 92.6% and a with ALI/ARDS, thus reflecting resolution
specificity of 100%. Results are expressed as area of the injury to the alveolar epithelium. In
under the curve (AUC) (95% confidence interval). addition, baseline plasma sRAGE may serve
as a reliable marker of the presence of non-
focal loss of aeration in ALI/ARDS based on
sRAGE Levels and Clinical CT scan lung morphology.
Outcome Although sRAGE is not specific for al-
veolar epithelial injury, previous studies
The main 28-day clinical outcomes in provide supportive evidence that the pri-
all groups are shown in Table 1. There mary source of plasma sRAGE in ALI/
was no significant difference between the ARDS patients is alveolar type 1 cells (6,
Crit Care Med 2011 Vol. 39, No. 3
11). Consistent with these findings,
sRAGE was found in the pulmonary
edema fluid at much higher levels than in
plasma in our study, suggesting a pulmo-
nary source for plasma sRAGE in ALI/
ARDS patients. Likewise, in a recently
published study using isolated perfused
human lung preparations in which the
lung was the only possible source of
sRAGE, perfusate levels of sRAGE were
inversely associated with alveolar fluid
clearance. The authors suggested that
sRAGE quantitatively reflects alveolar ep-
ithelial damage (25, 26). In the current
study, plasma sRAGE levels decreased
over time and were associated with ALI/
ARDS severity (PaO2-to-FIO2 ratio, lung
injury score). These data support an alve-
olar epithelial source of sRAGE. Further-
more, the decrease in sRAGE levels over
time, along with the improvement in
clinical and radiographic parameters of
ALI/ARDS, may reflect resolution of the
injury to the alveolar epithelium.
Despite this evidence, it is still un-
known whether nonpulmonary sources of
sRAGE can contribute to plasma levels in
the setting of ALI/ARDS. RAGE was first
identified in lung tissue (27, 28). Whereas
RAGE is constitutively expressed during
embryonal development, its expression is
down-regulated in adult life. Known ex-
ceptions are skin and lung, which consti-
tutively express RAGE throughout life.
Elevated plasma sRAGE levels have been
reported in 29 patients with severe sepsis
or septic shock when compared to
healthy volunteers and were associated
with patient outcome in a surgical ICU
setting (16). Our results challenge these
findings because sRAGE levels were sim-
ilar in patients with severe sepsis and in
mechanically ventilated controls in our
study. Interestingly, in the study by Bopp
et al (16), PaO2-to-FIO2 ratio was 139 ⫾ 81
mm Hg in survivors and 116 ⫾ 52 mm
Hg in nonsurvivors, suggesting that mea-
sured sRAGE levels probably reflected se-
vere lung injuries associated with sepsis
in this study. Interestingly, a very-low-
plasma sRAGE level (94 pg/mL) was mea-
sured in a neutropenic patient from the
ALI/ARDS plus severe sepsis group. Be-
cause RAGE is implicated in leukocyte
recruitment (29), investigations are
needed to better-understand interactions
between RAGE axis and leukocytes.
To our knowledge, this study is the
first to demonstrate that severe sepsis
alone, in the absence of ALI/ARDS, does
not result in increased sRAGE levels, sug-
gesting that the elevation of sRAGE in
485
Figure 5. A, PaO2/FIO2 ratio vs. form of the receptor for advanced glycation end products (sRAGE) level (pg/mL) in acute lung injury/acute respiratory
distress syndrome patients with and without severe sepsis or septic shock (Spearman’s ␳ ⫽ ⫺0.566; 95% CI, ⫺0.681 to ⫺0.423; p ⬍ .0001). B, Lung injury
score vs. plasma sRAGE levels (pg/mL) in acute lung injury/acute respiratory distress syndrome patients with and without severe sepsis or septic shock
(Spearman’s ␳ ⫽ 0.528; 95% CI, 0.377– 0.651; p ⬍ .0001).
patients with ALI/ARDS associated with
severe sepsis is specific for ALI/ARDS.
These findings suggest that sRAGE could
be used as a reliable quantitative bio-
marker of alveolar epithelial damage dur-
ing ALI/ARDS. In our study, sRAGE levels
were not influenced by the presence or
absence of RAGE-related diseases such as
diabetes mellitus, end-stage renal disease,
coronary artery disease, rheumatoid ar-
thritis, Alzheimer’s disease, and essential
hypertension, reinforcing the role of
sRAGE as a biomarker of ALI/ARDS in the
ICU setting. In the present study, baseline
levels of sRAGE did not influence the
main clinical outcomes in ALI/ARDS pa-
tients, in contrast with recent findings
(12). In this analysis of data from a large,
randomized, controlled trial, baseline lev-
els of plasma sRAGE were independently
associated with clinical outcome in ALI/
ARDS patients randomized to higher tidal
volumes (12 mL/kg predicted body
weight). Lack of statistical power in our
study, with respect to secondary objec-
tives such as clinical outcome, could ex-
plain this “negative” finding. Further-
more, relatively low tidal volumes were
used in our study (7.2 ⫾ 1.4 mL/kg pre-
dicted body weight), and the lack of asso-
ciation between sRAGE levels and 28-day
mortality could be explained in part by a
beneficial effect of lower tidal volume
ventilation in decreasing injury to the
alveolar epithelium.
sRAGE levels were also correlated with
CT scan lung morphology in our study.
As shown previously, focal and nonfocal
ALI/ARDS have distinct physiopathologic
patterns (19, 30, 31). Whereas focal ALI/
ARDS is characterized by a decrease in
gas volume and moderate inflammatory
486
edema, nonfocal ALI/ARDS is associated
with significant inflammatory edema and
impaired alveolar fluid clearance. Thus,
we speculate that sRAGE may represent a
reliable marker of diffuse alveolar damage
and impaired alveolar fluid clearance dur-
ing ALI/ARDS. However, further studies
are needed to confirm this finding.
There are some limitations to this
study. Variability of alveolar sRAGE levels
could reflect technical difficulties attrib-
utable to specific sampling method (24).
The aim of our study was to determine
whether the presence or absence of se-
vere sepsis influences the elevation of
sRAGE in ALI/ARDS in the ICU setting.
Nevertheless, we recommend that sRAGE
levels should be interpreted with caution
in the clinical setting. The cut-off value of
sRAGE to diagnose ALI/ARDS in critically
ill patients in our study is relatively low
when compared to the median level pre-
viously reported in healthy volunteers
(median [IQR], 1400 pg/mL [1100 –
1700]) (11). In our study population,
higher cut-off values resulted in lower
sensibility values, but no gain in specific-
ity was obtained. As with any study of a
diagnostic biomarker, a derivation cohort
will always demonstrate better biomarker
test characteristics than a validation co-
hort. We acknowledge that the lack of a
validation cohort and the discrepancy
with previous results are important lim-
itations, because the cut-off in our pa-
tients likely will be different in other pop-
ulations. Frequently, the biology of RAGE
coincides with settings in which ligands
of the receptor accumulate, especially in
a proinflammatory environment. More
work is needed for us to understand the
mechanisms by which RAGE is regulated,
especially with regard to the expression of
its soluble forms, and probably will result
in various levels of RAGE isoforms, both
“in space” and “in time.” Because our
study was designed to compare sRAGE
levels between ICU patients with ALI/
ARDS, those with severe sepsis, and me-
chanically ventilated controls, it does not
provide further information about the in-
volvement of RAGE in pulmonary health
and pathophysiology, or about the nature
of lung epithelial injury that results in
the release of sRAGE into the air spaces.
To date, the full extent of RAGE expres-
sion and the molecular mechanisms that
control the receptor in the context of
various stimuli have not been adequately
evaluated. Understanding the role of
RAGE signaling could provide an insight
in the reduction of lung injury and dis-
ease associated with abnormal expression
of RAGE or its soluble form, sRAGE (12,
32–36). Interestingly, at least in human
subjects but not in mice, circulating lev-
els of endogenous secretory RAGE have
been detected in plasma and tissue. The
enzyme-linked immunosorbent assay kit
used in our study quantifies concentra-
tions of total sRAGE and could not differ-
entiate between endogenous secreted
RAGE and total sRAGE. The exact biolog-
ical role of sRAGE is only partly under-
stood. Spanning the ligand-binding do-
main, sRAGE might act as an endogenous
competitive inhibitor of RAGE, thus play-
ing a critical role within the modulatory
network of the ligand–RAGE axis (37, 38).
Similarly, sRAGE may affect the activity
of other RAGE–ligand receptors, such as
Toll-like receptors and scavenger recep-
tors (39). It is conceivable that modula-
tion of total sRAGE (cleaved sRAGE plus
Crit Care Med 2011 Vol. 39, No. 3
endogenous secreted RAGE) or its indi-
vidual components may prevent, in part,
aggressive accumulation and action of li-
gands (40). Levels of sRAGE/endogenous
secreted RAGE therefore may be innate
markers of vulnerability to disease or its
severity or both, as suggested by the
sRAGE levels in ALI/ARDS patients in our
study. Further research is needed to com-
pare sRAGE-mediated mechanisms that
may compensate for increased proinflam-
matory cytokine expression and full-
length RAGE-mediated mechanisms that
enhance proinflammatory signaling.
CONCLUSIONS
Our data suggest that plasma levels of
sRAGE are elevated during ALI/ARDS in
the ICU setting, regardless of the pres-
ence or absence of severe sepsis/septic
shock. Elevated baseline levels of plasma
sRAGE were associated with clinical se-
verity and nonfocal loss of aeration based
on CT scan in ALI/ARDS patients. The
plasma level of sRAGE decreased over
time during ALI/ARDS, suggesting reso-
lution of the injury to the alveolar epi-
thelium. sRAGE may be a useful diagnos-
tic biomarker for ALI/ARDS, but further
validation is needed before it can be rec-
ommended as such. Thus, measurement
of sRAGE may stimulate more interest in
the relationship between injury to alveo-
lar type 1 epithelial cells and the patho-
physiology of ALI/ARDS. Larger clinical
studies are warranted to determine the
exact role of RAGE and its soluble iso-
forms in the pathogenesis and resolution
of ALI/ARDS. Understanding the basic
signal transduction events triggered by
this multi-ligand receptor may offer new
diagnostic and therapeutic options in pa-
tients with ALI/ARDS.
REFERENCES
1. Rubenfeld GD, Caldwell E, Peabody E, et al:
Incidence and outcomes of acute lung injury.
N Engl J Med 2005; 353:1685–1693
2. Bernard GR, Artigas A, Brigham KL, et al:
The American-European Consensus Confer-
ence on ARDS. Definitions, mechanisms, rel-
evant outcomes, and clinical trial coordina-
tion. Am J Respir Crit Care Med 1994; 149(3
Pt 1):818 – 824
3. Bierhaus A, Humpert PM, Morcos M, et al:
Understanding RAGE, the receptor for ad-
vanced glycation end products. J Mol Med
2005; 83:876 – 886
4. Schmidt AM, Yan SD, Yan SF, et al: The
Crit Care Med 2011 Vol. 39, No. 3
multiligand receptor RAGE as a progression
factor amplifying immune and inflammatory
responses. J Clin Invest 2001; 108:949 –955
5. Fehrenbach H, Kasper M, Tschernig T, et al:
Receptor for advanced glycation endproducts
(RAGE) exhibits highly differential cellular
and subcellular localisation in rat and hu-
man lung. Cell Mol Biol (Noisy-le-grand)
1998; 44:1147–1157
6. Shirasawa M, Fujiwara N, Hirabayashi S, et
al: Receptor for advanced glycation end-
products is a marker of type I lung alveolar
cells. Genes Cells 2004; 9:165–174
7. Mukherjee TK, Mukhopadhyay S, Hoidal JR:
Implication of receptor for advanced glyca-
tion end product (RAGE) in pulmonary
health and pathophysiology. Respir Physiol
Neurobiol 2008; 162:210 –215
8. Raucci A, Cugusi S, Antonelli A, et al: A
soluble form of the receptor for advanced
glycation endproducts (RAGE) is produced
by proteolytic cleavage of the membrane-
bound form by the sheddase a disintegrin
and metalloprotease 10 (ADAM10). FASEB J
2008; 22:3716 –3727
9. Zhang L, Bukulin M, Kojro E, et al: Receptor
for advanced glycation end products is sub-
jected to protein ectodomain shedding by
metalloproteinases. J Biol Chem 2008; 283:
35507–35516
10. Cheng C, Tsuneyama K, Kominami R, et al:
Expression profiling of endogenous secretory
receptor for advanced glycation end products
in human organs. Mod Pathol 2005; 18:
1385–1396
11. Uchida T, Shirasawa M, Ware LB, et al: Re-
ceptor for advanced glycation end-products
is a marker of type I cell injury in acute lung
injury. Am J Respir Crit Care Med 2006;
173:1008 –1015
12. Calfee CS, Ware LB, Eisner MD, et al: Plasma
receptor for advanced glycation end-products
and clinical outcomes in acute lung injury.
Thorax 2008; 63:1083–1089
13. Ventilation with lower tidal volumes as com-
pared with traditional tidal volumes for acute
lung injury and the acute respiratory distress
syndrome. The Acute Respiratory Distress
Syndrome Network. N Engl J Med 2000; 342:
1301–1308
14. Chavakis T, Bierhaus A, Nawroth PP: RAGE
(receptor for advanced glycation end prod-
ucts): A central player in the inflammatory
response. Microbes Infect 2004; 6:1219 –1225
15. van Zoelen MA, Achouiti A, Schmidt AM, et
al: Ligands of the receptor for advanced gly-
cation end products, including high-mobility
group box 1, limit bacterial dissemination
during Escherichia coli peritonitis. Crit Care
Med 2010; 38:1414 –1422
16. Bopp C, Hofer S, Weitz J, et al: sRAGE is
elevated in septic patients and associated
with patients outcome. J Surg Res 2008; 147:
79 – 83
17. American College of Chest Physicians/
Society of Critical Care Medicine Consensus
Conference: Definitions for sepsis and organ
failure and guidelines for the use of innova-
tive therapies in sepsis. Crit Care Med 1992;
20:864 – 874
18. Dellinger RP, Levy MM, Carlet JM, et al:
Surviving Sepsis Campaign: international
guidelines for management of severe sepsis
and septic shock: 2008. Crit Care Med 2008;
36:296 –327
19. Malbouisson LM, Muller JC, Constantin JM,
et al: Computed tomography assessment of
positive end-expiratory pressure-induced al-
veolar recruitment in patients with acute re-
spiratory distress syndrome. Am J Respir Crit
Care Med 2001; 163:1444 –1450
20. Puybasset L, Cluzel P, Gusman P, et al: Re-
gional distribution of gas and tissue in acute
respiratory distress syndrome. I. Conse-
quences for lung morphology. CT Scan
ARDS Study Group. Intensive Care Med
2000; 26:857– 869
21. Rouby JJ, Puybasset L, Cluzel P, et al: Re-
gional distribution of gas and tissue in acute
respiratory distress syndrome. II. Physiolog-
ical correlations and definition of an ARDS
Severity Score. CT Scan ARDS Study Group.
Intensive Care Med 2000; 26:1046 –1056
22. Puybasset L, Gusman P, Muller JC, et al:
Regional distribution of gas and tissue in
acute respiratory distress syndrome. III. Con-
sequences for the effects of positive end-
expiratory pressure. CT Scan ARDS Study
Group. Adult Respiratory Distress Syndrome.
Intensive Care Med 2000; 26:1215–1227
23. Constantin JM, Cayot-Constantin S, Roszyk
L, et al: Response to recruitment maneuver
influences net alveolar fluid clearance in
acute respiratory distress syndrome. Anes-
thesiology 2007; 106:944 –951
24. Matthay MA, Wiener-Kronish JP: Intact epi-
thelial barrier function is critical for the res-
olution of alveolar edema in humans. Am
Rev Respir Dis 1990; 142(6 Pt 1):1250 –1257
25. Frank JA, Briot R, Lee JW, et al: Physiological
and biochemical markers of alveolar epithe-
lial barrier dysfunction in perfused human
lungs. Am J Physiol Lung Cell Mol Physiol
2007; 293:L52–L59
26. Briot R, Frank JA, Uchida T, et al: Elevated
levels of the receptor for advanced glycation
end products, a marker of alveolar epithelial
type I cell injury, predict impaired alveolar
fluid clearance in isolated perfused human
lungs. Chest 2009; 135:269 –275
27. Neeper M, Schmidt AM, Brett J, et al: Clon-
ing and expression of a cell surface receptor
for advanced glycosylation end products of
proteins. J Biol Chem 1992; 267:14998 –
15004
28. Schmidt AM, Vianna M, Gerlach M, et al:
Isolation and characterization of two binding
proteins for advanced glycosylation end
products from bovine lung which are present
on the endothelial cell surface. J Biol Chem
1992; 267:14987–14997
29. Chavakis T, Bierhaus A, Al-Fakhri N, et al:
The pattern recognition receptor (RAGE) is a
counterreceptor for leukocyte integrins: A
novel pathway for inflammatory cell recruit-
ment. J Exp Med 2003; 198:1507–1515
487
30. Puybasset L, Cluzel P, Chao N, et al: A
computed tomography scan assessment of
regional lung volume in acute lung injury.
The CT Scan ARDS Study Group. Am J
Respir Crit Care Med 1998; 158(5 Pt 1):
1644 –1655
31. Lu Q, Constantin JM, Nieszkowska A, et al:
Measurement of alveolar derecruitment in
patients with acute lung injury: Computer-
ized tomography versus pressure-volume
curve. Crit Care 2006; 10:R95
32. Sternberg DI, Gowda R, Mehra D, et al:
Blockade of receptor for advanced glycation
end product attenuates pulmonary reperfu-
sion injury in mice. J Thorac Cardiovasc
Surg 2008; 136:1576 –1585
33. Zhang H, Tasaka S, Shiraishi Y, et al: Role
488
of soluble receptor for advanced glycation
end products on endotoxin-induced lung
injury. Am J Respir Crit Care Med 2008;
178:356 –362
34. Queisser MA, Kouri FM, Konigshoff M, et al:
Loss of RAGE in pulmonary fibrosis: Molec-
ular relations to functional changes in pul-
monary cell types. Am J Respir Cell Mol Biol
2008; 39:337–345
35. Englert JM, Hanford LE, Kaminski N, et al: A
role for the receptor for advanced glycation
end products in idiopathic pulmonary fibro-
sis. Am J Pathol 2008; 172:583–591
36. Reynolds PR, Schmitt RE, Kasteler SD, et
al: Receptors for advanced glycation end-
products (RAGE) targeting protect against
hyperoxia-induced lung injury in mice.
Am J Respir Cell Mol Biol 2010; 42:
545–551
37. Santilli F, Vazzana N, Bucciarelli LG, et al:
Soluble forms of RAGE in human diseases:
Clinical and therapeutical implications. Curr
Med Chem 2009; 16:940 –952
38. Creagh-Brown BC, Quinlan GJ, Evans TW:
RAGE inhibition: Healthy or harmful? Crit
Care Med 2010; 38:1487–1490
39. Schmidt AM, Yan SD, Yan SF, et al: The
biology of the receptor for advanced glyca-
tion end products and its ligands. Biochim
Biophys Acta 2000; 1498(2–3):99 –111
40. Ramasamy R, Yan SF, Schmidt AM: RAGE:
Therapeutic target and biomarker of the in-
flammatory response–The evidence mounts.
J Leukoc Biol 2009; 86:505–512
Crit Care Med 2011 Vol. 39, No. 3