Journal of Critical Care 29 (2014) 185.e1185.

e7

Contents lists available at ScienceDirect

Journal of Critical Care

journal homepage: www.jccjournal.org

Fluid resuscitation with hydroxyethyl starches in patients with sepsis is associated

with an increased incidence of acute kidney injury and use of renal replacement
therapy: A systematic review and meta-analysis of the literature
Ary Serpa Neto, MD, MSc a, b, c,, Denise P. Veelo, MD a, Victor Galvo Moura Peireira, MD a,
Murillo Santucci Cesar de Assuno, MD, MSc b, Jos Antnio Manetta, MD a, Daniel Crepaldi Espsito, MD a,
Marcus J. Schultz, MD, PhD c
a
Medical Intensive Care Unit, ABC Medical School (FMABC), Santo Andr, Brazil
b
Department of Critical Care Medicine, Hospital Israelita Albert Einstein, So Paulo, Brazil
c
Department of Intensive Care Medicine & Laboratory of Experimental Intensive Care and Anesthesiology, Academic Medical Center, University of Amsterdam, The Netherlands

a r t i c l e i n f o a b s t r a c t

Keywords: Purpose: Fluid resuscitation is a key intervention in sepsis, but the type of uids used varies widely. The aim of
Sepsis this meta-analysis is to determine whether resuscitation with hydroxyethyl starches (HES) compared with
Colloid crystalloids affects outcomes in patients with sepsis.
Crystalloid Materials and Methods: Search of MEDLINE, EMBASE, and Cochrane Central Register of Controlled Trials up to
Hydroxyethyl starch
February 2013. Studies that compared resuscitation with HES versus crystalloids in septic patients, and
Meta-analysis
reported incidence of acute kidney injury (AKI), renal replacement therapy (RRT), transfusion of red blood
cell (RBC) or fresh frozen plasma and/or mortality. Three investigators independently extracted data into
uniform risk ratio measures. The Grading of Recommendations Assessment, Development and Evaluation
framework was used to determine the quality of the evidence.
Results: Ten trials (4624 patients) were included. An increased incidence of AKI (risk ratio [RR], 1.24 [95%
Condence Interval {CI}, 1.13-1.36], and need of RRT (RR, 1.36 [95% CI, 1.17-1.57]) was found in patients who
received resuscitation with HES. Resuscitation with HES was also associated with increased transfusion of
RBC (RR, 1.14 [95% CI, 1.01-1.93]), but not fresh frozen plasma (RR, 1.47 [95% CI, 0.97-2.24]). Furthermore,
while intensive care unit mortality (RR, 0.74 [95% CI, 0.43-1.26]), and 28-day mortality (RR, 1.11 [95% CI,
0.96-1.28]) was not different, resuscitation with HES was associated with higher 90-day mortality (RR, 1.14
[95% CI, 1.04-1.26]).
Conclusions: Fluid resuscitation practice with HES as in the meta-analyzed studies is associated with increased
an increase in AKI incidence, need of RRT, RBC transfusion, and 90-day mortality in patients with sepsis.
Therefore, we favor the use of crystalloids over HES for resuscitation in patients with sepsis.
2014 Elsevier Inc. All rights reserved.

1. Introduction
Fluid resuscitation is essential in the management of patients with
sepsis. The Surviving Sepsis Campaign guidelines strongly recom-
mends early and aggressive uid resuscitation in these patients to
maintain adequate mean arterial pressures and to improve and secure
blood ow [1]. Surprisingly, there is a wide variation in the type of
uid used for resuscitation in patients with sepsis worldwide, which is
primarily the result of personal preferences, local availability, and
probably also marketing [2,3]. Indeed, in Australia, more than half of
all uid resuscitations are with colloids, mainly albumin, and in
No nancial support. The authors declare that they have no competing interests.
Corresponding author. Av. Lauro Gomes, 2000Santo Andr, Brazil. Tel./fax: +55 11
4993 5400.
E-mail address: aryserpa@terra.com.br (A. Serpa Neto).
European countries, nearly half of all uid resuscitations are with
colloids, mainly hydroxyethyl starches (HES) [3]. By contrast, in the
United States crystalloids are preferred [3]. Colloid solutions,
including HES, are suggested to result in a more rapid and lasting
circulatory stabilization than crystalloids, but there are limited studies
that support these potentially benecial effects [4].
Use of colloids, especially HES, may be not without risks.
Randomized controlled trials suggest use of HES to be associated
with coagulation alterations, allergic reactions, increased incidence of
acute kidney injury (AKI) and need of renal replacement therapy
(RRT), and higher overall mortality [57]. While it was believed that
harm was only found with use of HES solutions with higher molecular
weights and higher substitution ratios [8], recent trials suggest harm
from HES solutions with a lower molecular weight and a lower
substitution ratio as well [815]. One recent meta-analysis comparing
resuscitation with HES and resuscitation with albumin in patients

0883-9441/$ see front matter 2014 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcrc.2013.09.031
185.e2 A. Serpa Neto et al. / Journal of Critical Care 29 (2014) 185.e1185.e7
with sepsis conrms earlier ndings of harm from HES [16], as do
other meta-analyses comparing resuscitation with HES with resusci-
tation with other uids in critically ill patients [2,17]. Notably, these
meta-analyses did not focus on patients with sepsis, but included
studies of unselected critically ill patients [2,17].
We aimed to determine whether there is a difference in the
incidence of AKI, need of RRT, red blood cell (RBC) and fresh frozen
plasma (FFP) transfusion, and overall mortality in patients with sepsis
receiving HES for uid resuscitation compared with uid resuscitation
with crystalloids.
2. Methods
2.1. Search methods to identify studies
Studies were identied by 2 authors through a computerized
blinded search of Medline (1966-2013), EMBASE, and Cochrane
Central Register of Controlled Trials (CENTRAL) using a sensitive
search strategy combining the following Medical Subject Headings
and keywords: (colloid [MeSH Terms] OR albumin [MeSH Terms] OR
starch [MeSH Terms] OR dextran [MeSH Terms] OR gelatin [MeSH
Terms] OR plasma [MeSH Terms]) AND (sepsis [MeSH Terms] OR
septic shock [MeSH Terms]). Then, reviewed articles and cross-
referenced studies from retrieved articles were screened for
pertinent information.
2.2. Selection of studies
Articles were selected for inclusion in the systematic review if they
evaluated HES exclusively in patients with sepsis. In one arm of the
study (named starch), uid resuscitation with HES was performed.
Then, this group was compared with another arm of study (named
crystalloid), in which uid resuscitation only with crystalloid
infusion was performed. We included randomized trials as well as
observational studies (cohort, before/after, and cross-sectional), with
no restrictions on language or scenario (intensive care unit [ICU] and/
or emergency room). We excluded revisions, studies that used any
type of colloid (gelatin, HES, and albumin) in the crystalloid arm,
studies that compared HES with another articial colloid and/or
albumin, and studies that did not report the outcomes of interest
(dened below). When we found duplicate reports of the same study
in preliminary abstracts and articles, we analyzed data from the most
complete data set. When necessary we contact the authors for
additional unpublished data.
2.3. Data extraction
Data were independently extracted from each report by three
authors, using a data recording form developed for the purpose of this
meta-analysis. After extraction, data were reviewed and compared by
the rst author. Instances of disagreement between the two other
extractors were solved by a consensus among the investigators.
Whenever needed, we obtained additional information about a
specic study by directly questioning the principal investigator.
2.4. Validity Assessment
In randomized trials, we assessed allocation concealment,
blinding process, use of intention-to-treat analysis, lost to follow-
up, early stopping, and the baseline similarity of groups. We used
the Grading of Recommendations Assessment, Development and
Evaluation (GRADE) approach to summarize the quality of evidence
for each outcome [18]. In this approach, randomized trials begin as
high quality evidence but can be rated down for study limitations,
imprecision, inconsistency, indirectness, or suspicion of a publica-
tion bias.
2.5. Denition of endpoints
The primary endpoint was the development of AKI in each arm of
the study. Secondary endpoints included: RRT, transfusion of RBC and
FFP, and ICU-, 28-day, and 90-day mortality. Additional endpoints
were ICU- and hospital-length of stay (LOS), and total amount of study
uid (HES in the starch arm and crystalloid in the crystalloid arm)
infused on the rst day and during follow-up. Severity of AKI was
measured by Acute Kidney Injury Network (AKIN) classication
(when RIFLE-scoring was used we considered RIFLE R as AKIN I, RIFLE
I as AKIN II, and RIFLE F as AKIN III).
2.6. Statistical analysis
We extracted data regarding the study design, patient charac-
teristics, type of HES used, and amount of uid infused. For the
analysis of AKI development, RRT, transfusion, and mortality, we
used the most protracted follow-up in each trial up to hospital
discharge. We calculated a pooled estimate of risk ratio (RR) in the
individual studies using a random-effect model according to Mantel
and Haenszel and graphically represented these results using forest
plot graphs.
We explored the following variables as potential modiers in
subgroup analyses: (1) type of HES used (10% HES 200/0.5 vs 6% HES
130/0.4); (2) study design (randomized trial vs observational study);
and (3) amount of colloid infused (on the rst day and during follow-
up). We reasoned that each of these might inuence the benecial or
harmful effect of HES on outcome.
To explore whether these variables modied the outcome, we
compared pooled effects among studies with and without them. For
continuous variables, we used the standardized mean difference
(SMD) which is the difference in means divided by a standard
deviation. The homogeneity assumption was checked by a 2 test
with a df equal to the number of analyzed studies minus 1. Also, the
heterogeneity was measured by the I 2 which describes the
percentage of total variation across studies that is due to
heterogeneity rather than chance. I 2 was calculated from basic
results obtained from a typical meta-analysis as I 2 = 100% (Q
df)/Q, where Q is Cochran's heterogeneity statistic and df is the
degrees of freedom. A value of 0% indicates no observed
heterogeneity, and larger values show increasing heterogeneity.
When heterogeneity was found we tried to identify and describe
the reason. A sensitivity analysis was carried out by recalculating
pooled RR estimates for different subgroups of studies based on
relevant clinical features. This analysis demonstrates whether the
overall result has been affected by a change in the meta-analysis
selection criteria. A potential publication bias was assessed
graphically with funnel plots, as well as by Begg and Mazumdars
rank correlation and Eggers regression. Interrater reliability was
determined by comparing the number of studies included by author
1 with author 2 in each stage of the search using the coefcients.
Parametric variables were presented as the mean SD and
non-parametric variables were presented as the median (inter-
quartile range). All analyses were conducted with Review Manager
v.5.1.1 and SPSS v.16.0.1. For all analyses, a 2-sided P b .05 was
considered signicant.
3. Results
3.1. Literature search
The search strategy retrieved 2423 unique citations. Of these
citations, 2368 were excluded after the rst screening based on the
abstracts or titles, leaving 55 articles for a full-text review (Fig. 1).
Forty ve articles were further excluded for the following reasons:
no data on outcome of interest (n = 19); comparison with other
 A. Serpa Neto et al. / Journal of Critical Care 29 (2014) 185.e1185.e7 185.e3

Table 2
Characteristics of the patients in each arm analyzed

Variables Starch Crystalloid P

(n = 2355) (n = 2269)

Age, years 65.27 2.19 65.21 2.79 0.902

SAPS II 50.00 0.01 52.33 1.15 0.100
APACHE II 19.40 2.13 19.16 1.87 0.981
Type of HES
6% HES 130/0.4, n (%) 2072 (87.9)
10% HES 200/0.5, n (%) 283 (12.1)

SAPS, simplied acute physiology score; APACHE, acute physiology and chronic health
evaluation.
Kruskal-Wallis test.

3.3. Outcomes

Fig. 1. Literature search strategy.
type of colloid (n = 12); retracted articles (n = 6); data on septic
patients not available (n = 4); same cohort previously analyzed (n
= 2); and other reason (n = 2). Thus, 10 articles (4624 patients)
were included in the meta-analysis [5,715]. Description of the
patients, interventions, comparisons, outcomes, and study design is
presented in the online data. For all of the comparisons of interrater
reliability in each stage of the search, the kappa coefcient ranged
from 0.79 to 0.96.
3.2. Characteristics of the studies and patients
Table 1 summarizes studies characteristics. Four studies evaluated
patients with severe sepsis or septic shock [5,11,12,15], 3 evaluated
only patients with severe sepsis [8,10,14], and 3 evaluated only septic
shock patients [7,9,13]. Two studies used 10% HES 200/0.5 (283
patients, 12% of the total analyzed in the meta-analysis) [5,7]. Eight
studies evaluated 6% HES 130/0.4 (2072 patients, 88% of the total
analyzed in the meta-analysis) [815]. In 6 studies HES was compared
with normal saline [7,9,10,12,14,15], in two with Ringer lactate [5,13],
one with Ringer acetate [8], and one with balanced saline [11]. All
but one study were randomized controlled trials [11]. Studies quality
assessment is show in Table E1 and outcomes assessed in each study
were show in Table E2. Table 2 summarizes characteristics of the
patients at baseline. Age and prognostic scores were comparable
between the two groups of patients. Characteristics and outcomes
stratied by type of colloid administered are described in Table E3.
Four hundred and ninety eight of the 1120 patients analyzed in the
starch group (44.4%) and 384 of the 1105 patients analyzed in the
crystalloid group (34.7%) developed AKI during the follow-up (RR,
1.24 [95% CI, 1.13-1.36]; P b .001; I 2 = 0%; P = 0.430). The difference
observed in AKI incidence between the two groups was mainly due to
a higher incidence of AKIN III AKI in the starch group (RR, 1.27 [95% CI,
1.09-1.48]; P = 0.002; I 2 = 0%; P = .990) (Table 3, Fig. 2).
Three hundred and twenty two of the 1158 patients in the starch
group (27.8%) and 233 of the 1152 patients in the crystalloid group
(20.2%) were treated with any type of RRT during the follow-up (RR,
1.36 [95% CI, 1.17-1.57]; P b .001; I 2 = 0%; P = .570) (Table 3, Fig. 2).
More patients in the starch group received RBC transfusion
compared with patients in the crystalloid group (RR, 1.14 [95% CI,
1.01-1.21]; P = 0.04). FFP transfusion was comparable in the two
groups (RR, 1.47 [95% CI, 0.97-2.24]). Both analyses, though, showed
heterogeneity between the studies analyzed (I 2 = 69%; P = .020, and
I 2 = 87%; P = .005, respectively) (Table 3, Fig. 3).
Ninety-day mortality was higher in patients managed with HES
infusion (RR, 1.14 [95% CI, 1.04-1.26]; P = .005; I 2 = 0%; P = .780).
ICU-, and 28-day mortality were comparable in the two groups
analyzed (RR, 0.74 [95% CI, 0.43-1.26]; I 2 = 53%; P = .070; RR, 1.11
[95% CI, 0.96-1.28]; I 2 = 0%; P = .920, respectively). Heterogeneity
was due to the study of Bayer et al [11], probably due to its
retrospective nature. Exclusion of this study changed overall results
and decreased the heterogeneity (RR, 0.56 [95% CI, 0.34-0.94]; I 2 =
0%, P = .570) (Table 3, Fig. 4).
ICU- and hospital LOS was similar in the two groups. There was no
difference in the total amounts of uid on the rst day of admission
(1978.80 563.58 vs 3011.80 vs 1972.34 mL; SMD, 1.42 [95% CI,
3.00 to 0.16]; I 2 = 97%; P b 0.0001). Heterogeneity was due to the
study of Dubin et al [9], probably due to the low number of patients
included in this study and the high amount of uid infused in the

Table 1
Characteristics of the included studies

Study, year Design Type of uid (n) Population Primary outcome

Control Intervention

Starch vs crystalloid
Brunkhorst, 2008 [5]
McIntyre, 2008 [7]
Dubin, 2010 [9]
Zhu, 2011 [14]
Guidet, 2012 [10]
Bayer, 2012 [11]
Myburgh, 2012 [12]
Perner, 2012 [8]
Lv, 2012 [13]
Siegemund, 2013 [15]
RCT Ringer lactate (275) 10% HES 200/0.5 (262) Severe sepsis or septic shock 28-day mortality and mean SOFA
RCT NaCl 0.9% (19) 10% HES 200/0.5 (21) Septic shock Feasibility measures for the RCT
RCT NaCl 0.9% (11) 6% HES 130/0.4 (9) Septic shock Microcirculatory parameters
RCT NaCl 0.9% (96) 6% HES 130/0.4 (100) Severe sepsis Physiological parameters
RCT Ringer lactate (45) 6% HES 130/0.4 (90) Severe sepsis Amount of uid to achieve HDS
PSA Balanced saline (334) 6% HES 130/0.4 (360) Severe sepsis or septic shock Time to shock reversal
RCT NaCl 0.9% (945) 6% HES 130/0.4 (976) Severe sepsis or septic shock 90-day mortality
RCT Ringer acetate (400) 6% HES 130/0.4 (398) Severe sepsis 90-day mortality or ESKF
RCT Ringer lactate (20) 6% HES 130/0.4 (22) Septic shock Coagulation parameters
RCT NaCl 0.9% (124) 6% HES 130.04 (117) Severe sepsis or septic shock Mortality, ICU length of stay

RCT, randomized controlled trial; NaCl, sodium chloride; SOFA, sequential organ failure assessment; HDS, hemodynamic stabilization; PSA, prospective sequential analysis; ESKF,
end-stage kidney failure (dependence on dialysis); B/F, before and after analysis; OBS, observational.
Sepsis comprises septic shock or severe sepsis.
185.e4 A. Serpa Neto et al. / Journal of Critical Care 29 (2014) 185.e1185.e7

Table 3
Outcomes of the patients

Variables Starch Crystalloid Risk ratio I2 P

(n = 2355) (n = 2269) (95% CI)

ICU mortality, n/total n (%) 140/502 (27.9) 125/429 (29.1) 0.74 (0.43-1.26) 53% .270
28-day mortality, n/total n (%) 264/781 (33.8) 240/790 (30.4) 1.11 (0.96-1.28) 0% .160
90-day mortality, n/total n (%) 596/1736 (34.3) 521/1716 (30.3) 1.14 (1.04-1.26) 0% .005
Renal replacement therapy, n/total n (%) 322/1158 (27.8) 233/1152 (20.2) 1.36 (1.17-1.57) 0% b.0001
Acute kidney injury, n/total n (%) 498/1120 (44.4) 384/1105 (34.7) 1.24 (1.13-1.36) 0% b.0001
AKIN I, n/total n (%) 121/858 (14.1) 134/830 (16.1) 0.89 (0.67-1.19) 33% .430
AKIN II, n/total n (%) 120/858 (14.0) 103/830 (12.4) 1.13 (0.88-1.44) 0% .330
AKIN III, n/total n (%) 259/858 (30.2) 195/830 (23.5) 1.27 (1.09-1.48) 0% .002
Red blood cell transfusion
N/Total (%) 754/1120 (67.3) 655/1105 (59.3) 1.14 (1.01-1.28) 69% .040
Volume received, mL 1091.80 684.87 766.25 403.59 .486
Fresh frozen plasma transfusion
N/Total (%) 318/758 (42.0) 201/734 (27.4) 1.47 (0.97-2.24) 87% .070
Volume received, mL 1790.00 1004.09 1475.00 742.46 .333
Length of stay, days
ICU 13.88 4.07 13.26 5.96 .686
Hospital 29.06 9.74 30.56 11.43 .964
Study uid intake
Day 1, mL 1978.80 563.58 3011.80 1972.34 1.42 (3.00 to 0.16) 97% .080
During follow-up, mL 4089.42 1470.14 5424.42 2189.62 1.24 (2.95 to 0.47) 98% .160
Random-effect model (starch vs crystalloid).

Standardized mean difference (95% CI).

crystalloid group. The exclusion of this study, however, did not change
the result but decreased the heterogeneity (SMD, 0.32 [95% CI,
0.58 to 0.06]; I 2 = 0%, P = .980). The amount of uid infused during
complete follow-up was similar in the two groups (4089.42
1470.14 vs 5424.42 2189.62 mL; SMD, 1.24 [95% CI, 2.95 to
0.47]; I 2 = 980%; P b .0001) (Table 3, Fig. E1).
3.4. Subgroup analyses
Regarding mortality, 90-day mortality was signicantly higher in
starch group in studies that evaluated 6% HES (RR, 1.12 [95%CI,
1.00-1.24]), in those in whom colloid infusion on the rst day was
less than 1500 mL and total amount of uids was below 4000 mL
(RR, 1.18 [95% CI, 1.02-1.35] and RR, 1.17 [95% CI, 1.01-1.36],
respectively). ICU mortality was lower in the starch group when
analyzing only randomized controlled trials, however this result
was due to the inclusion of the studies by Dubin et al [9], and Lv
et al [13], that were both of low quality (RR, 0.42 [95% CI, 0.24-
0.73]) [13,14] (Fig. E2).
AKI was higher in the group of HES in all subgroup analyses with
the exception in studies where the colloid infusion on the rst day
was less than 1500 mL (RR, 1.20 [95% CI, 0.98-1.47]). RRT and
transfusion of RBC was lower in the group of crystalloid in all
subgroup analyses (Figs. E3, E4).
GRADE prole for all outcomes is show in Table E4. There was no
signicant publication bias in the assessments. Funnel plots were
show in Figs. E5, E6, and E7.
4. Discussion
We found an association between uid resuscitation with HES and
increased incidence of AKI, RRT, RBC transfusions, and 90-day mortality
compared with resuscitation with any type of crystalloid in patients
with sepsis. Also, the meta-analysis suggests neither a difference in the
amounts of uids used on the rst day nor during follow-up between
patients resuscitated with HES and patients resuscitated with saline.
Recently, 2 meta-analyses of studies of uid resuscitation with
HES in critically ill patients were published. Haase et al [16] showed
that uid resuscitation with HES was associated with increased use
of RRT, and RBC transfusion when compared with uid resuscitation
with crystalloids or albumin in patients with sepsis. Zarychanski et
al [17], reported on a meta-analysis of studies of uid resuscitation
with HES in unselected critically ill patients, i.e., not restricting the
analysis to studies of patients with sepsis. This meta-analysis
showed uid resuscitation with HES to be associated with
development of AKI and mortality. The present meta-analysis
resiliently conrms the results from these two meta-analyses,
showing resuscitation with HES to be associated with increased
morbidity and mortality in patients with sepsis.
Our meta-analysis adds to the existing literature in several ways.
First, one difference of our meta-analysis from that previous published
in the literature is that our meta-analysis is the rst to analyze HES
against crystalloid exclusively. We excluded trials that compared HES
with albumin, because albumin is another colloid and could
contaminate the overall results. We also excluded trials that have
been retracted due inconsistencies and fraudulency [17]. Sepsis is a
very specic and special condition and a certain treatment that is safe
and without risk in general critically ill patients can be of higher risk in
septic patients. Thus, our analysis is the rst to focus exclusively in
patients with sepsis, managed with only crystalloids in the control
arm. Also, our analysis is the only meta-analysis that provides more
extensive and specic analysis of mortality, AKI, AKI severity, and
transfusion. Finally, we conducted an extensive literature search, an
independent screening of articles and data extraction by 2 authors,
and assessed the risk of bias, which increases the strength of our
meta-analysis. We used a very pragmatic and conservative approach
to calculate the outcomes, using a random-effect model, associated
with subgroup analyses dened a priori to explain clinical and
methodological heterogeneity, and we try to elucidate all type of
statistical heterogeneity.
It is important to consider that all studies dealing with uid
resuscitation have points of inconsistencies. First of all, the trigger for
uid infusion is an important concern, since uid resuscitation in
patients who do not need uids and/or are not uid responsiveness
could cause harm and contribute to AKI development and mortality
[19,20]. The scheme of uid infusion, as well as endpoint of
resuscitation chosen are important concerns that needs to be assessed
in these studies. All this converges to the point that it may not be the
type of uid chosen that is causing harm, but the way in which it was
used in the studies. So, we cannot exclude benet from using HES
according to different protocols, triggers, and endpoints from those
chosen in the studies analyzed in this meta-analysis.
One other important point is that some studies included patients
with renal impairment, a formal contraindication for HES infusion.
 A. Serpa Neto et al. / Journal of Critical Care 29 (2014) 185.e1185.e7 185.e5

P
P

P
P

P
P

P
P

Fig. 2. Acute kidney injury and its severity and need of renal replacement therapy in patients managed with HES vs crystalloid. AKIN, acute kidney injury network.

Unfortunately the data available was insufcient to conduct a stratied
analysis according to the presence or not of renal impairment. Also,
while HES is only indicated for initial hemodynamic stabilization, some
studies analyzed in our meta-analysis suggest incorrect use, since
patients could receive HES after hemodynamic stabilization and for
longer periods. Several studies are now evaluating more correct use of
HES with other strategies of uid resuscitation.
Despite the abovementioned concerns this is not the only meta-
analysis that suggests harm with the use of HES [16,17]. In addition,
well designed trials with high power showed worse outcomes in ICU
patients managed with HES, with increased incidences of AKI
mandating increased need for RRT, as well as increased overall
mortality [8,12]. Thus, we have a high quality level of evidence that
suggests harm of HES infusion in general critically ill patients that
need uid resuscitation.
Moreover, there are several reasons for why HES may cause harm in
critically ill patients in general, and sepsis patients in particular. Recent
evidences suggest that the limiting factor of glomerular ltration rate
(GFR) is the renal plasma ow and the plasma protein concentration.
An increase of capillary hydraulic pressure will cause the ultraltrate to
185.e6 A. Serpa Neto et al. / Journal of Critical Care 29 (2014) 185.e1185.e7

Fig. 3. Red blood cell and fresh frozen plasma transfusion in patients managed with HES vs crystalloid.

Fig. 4. Mortality in patients managed with HES vs crystalloid.

 A. Serpa Neto et al. / Journal of Critical Care 29 (2014) 185.e1185.e7
be mainly generated on the rst portion of the afferent side of the
capillary network and to cease when hydraulic and oncotic pressures
became equal [21]. Therefore, the oncotic pressure becomes the
limiting factor for GFR and the use of colloid will increase this pressure
and consequently decrease GFR [22]. This is one of the mechanisms
that elucidate the nephrotoxicity of HES found in this meta-analysis.
A high fraction of HES is deposited in the tissues where it cannot be
metabolized and may act as a foreign body with toxic effects in kidney,
liver, and bone marrow [16,23]. The development of AKI and the
subsequent need of RRT can also explain the increase in 90-day
mortality in our meta-analysis, since RRT has been repeatedly associated
with death [24]. Finally, results of a subgroup analysis of the SAFE trial
suggest that albumin compared to saline do not impair renal or other
function in sepsis patients and may decrease the risk of death [25].
Limitations of our study include the risk of bias which may
exaggerate the studys conclusion if publication is related to the
strength of the results. Also, there are some trials that do not report all
the outcomes of interest, which increases the bias of these studies.
Analyzing 6% HES together with 10% HES could be another source of
bias; however, we conducted a subgroup analysis stratifying the
results in these 2 groups.
5. Conclusions
In conclusion, resuscitation practice with HES as used in meta-
analyzed studies in patients with sepsis is associated with an increase
in AKI incidence, need of RRT, RBC transfusion, and 90-day mortality.
Taken together with the results of recently published well-powered
randomized controlled trials, HES as used in the uid resuscitation
protocols in the meta-analyzed studies of patients with sepsis can no
longer be recommended.
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.jcrc.2013.09.031.
References
[1] Dellinger RP, Levy MM, Rhodes A, et al. Surviving sepsis campaign: international
guidelines for management of severe sepsis and septic shock, 2012. Intensive Care
Med 2013;39:165228.
[2] Gattas DJ, Dan A, Myburgh J, Billot L, Lo S, Finfer S. Fluid resuscitation with 6%
hydroxyethyl starch (130/0.4) in acutely ill patients: an updated systematic
review and meta-analysis. Anesth Analg 2012;114:15969.
[3] Finfer S, Liu B, Taylor C, et al. Resuscitation uid use in critically ill adults: an
international cross-sectional study in 391 intensive care units. Crit Care 2010;14:
R185.
[4] Perel P, Roberts I. Colloids versus crystalloids for uid resuscitation in critically ill
patients. Cochrane Database Syst Rev 2011;3:CD000567.
185.e7
[5] Brunkhorst FM, Engel C, Bloos F, et al. Intensive insulin therapy and pentastarch
resuscitation in severe sepsis. N Engl J Med 2008;358:12539.
[6] Schortgen F, Lacherade JC, Bruneel F, et al. Effects of hydroxyethylstarch and
gelatin on renal function in severe sepsis: a multicentre randomised study. Lancet
2001;357:9116.
[7] McIntyre LA, Fergusson D, Cook DJ, et al. Fluid resuscitation in the septic shock
(FINESS): a feasibility trial. Can J Anesth 2008;55:81926.
[8] Perner A, Haase N, Guttormsen AB, et al. Hydroxyethyl starch 130/0.42 versus
ringers acetate in severe sepsis. N Engl J Med 2012;367:12434.
[9] Dubin A, Pozo MO, Casabella CA, et al. Comparison of 6% hydroxyethyl starch
130/0.4 and saline solution for resuscitation of the microcirculation during
the early goal-directed therapy of septic patients. J Crit Care 2010;25:
65966.
[10] Guidet B, Martinet O, Boulain T, et al. Assessment of hemodynamic efcacy and
safety of 6% hydroxyethylstarch 130/0.4 vs. 0.9% NaCl uid replacement in
patients with severe sepsis: The CRYSTMAS study. Crit Care 2012;16:R94-103.
[11] Bayer O, Reinhart K, Kohl M, et al. Effects of uid resuscitation with synthetic
colloids or crystalloids alone on shock reversal, uid balance, and patient
outcomes in patients with severe sepsis: a prospective sequential analysis. Crit
Care Med 2012;40:254351.
[12] Myburgh JA, Finfer S, Bellomo R, et al. Hydroxyethyl starch or saline for uid
resuscitation in intensive care. N Engl J Med 2012;367:190111.
[13] Lv J, Zhao HY, Liu F, An YZ. The inuence of lactate Ringer solution versus
hydroxyethyl starch on coagulation and brinolytic system in patients with septic
shock. Zhongguo Wei Zhong Bing Ji Jiu Yi Xue 2012;24:3841.
[14] Zhu GC, Quan ZY, Shao YS, Zhao JG, Zhang YT. The study of hypertonic saline and
hydroxyethyl starch treating severe sepsis. Zhongguo Wei Zhong Bing Ji Jiu Yi Xue
2011;23:1503.
[15] Basel Starch Evaluation in Sepsis (BaSES). Last veried January 2013 http://clinical
trials.gov/ct2/show/NCT00273728; 2012.
[16] Haase N, Perner A, Hennings LI, et al. Hydroxyethyl starch 130/0.38-0.45 versus
crystalloid or albumin in patients with sepsis: systematic review with meta-
analysis and trial sequential analysis. BMJ 2013;346:f839Q50.
[17] Zarychanski R, Abou-Setta AM, Turgeon AF, et al. Association of hydroxyethyl
starch administration with mortality and acute kidney injury in critically ill
patients requiring volume resuscitation: a systematic review and meta-analysis.
JAMA 2013;309:67888.
[18] Guyatt GH, Oxman AD, Kunz R, et al. What is quality of evidence and why is it
important to clinicians? BMJ 2008;336:9958.
[19] Boyd JH, Forbes J, Nakada T, Walley KR, Russell JA. Fluid resuscitation in septic
shock: a positive uid balance and elevated central venous pressure are associated
with increased mortality. Crit Care Med 2011;39:25965.
[20] Bouchard J, Soroko SB, Chertow GM, et al. Fluid accumulation, survival and
recovery of kidney function in critically ill patients with acute kidney injury.
Kidney Int 2009;76:4227.
[21] Legrand M, Payen D. Understanding urine output in critically ill patients. Ann
Intensive Care 2011;1:1320.
[22] Moran M, Kapsner C. Acute renal failure associated with elevated plasma oncotic
pressure. N Engl J Med 1987;317:1503.
[23] Bellmann R, Feistritzer C, Wiedermann CJ. Effect of molecular weight and
substitution on tissue uptake of hydroxyethyl starch: a meta-analysis of clinical
studies. Clin Pharmacokinet 2012;51:22536.
[24] Bagshaw SM, Uchino S, Bellomo R, et al. Septic acute kidney injury in critically ill
patients: clinical characteristics and outcomes. Clin J Am Soc Nephrol 2007;2:
4319.
[25] SAFE Study Investigators, Finfer S, McEvoy S, Bellomo R, McArthur C, Myburgh J,
et al. Impact of albumin compared to saline on organ function and mortality of
patients with severe sepsis. Intensive Care Med 2011;37:8696.