Renal effects of synthetic colloids and crystalloids in patients with

severe sepsis: A prospective sequential comparison*

Ole Bayer, MD; Konrad Reinhart, MD; Yasser Sakr, MD, PhD; Bjoern Kabisch, PhD; Matthias Kohl, PhD;
Niels C. Riedemann, MD; Michael Bauer, MD; Utz Settmacher, MD; Khosro Hekmat, MD; Christiane S. Hartog, MD

Objectives: Hydroxyethyl starch 200 is associated with renal
impairment in sepsis, but hydroxyethyl starch 130/0.4 and gelatin
are considered to be less harmful. We hypothesized that fluid
therapy with only crystalloids would decrease the incidence of
acute kidney injury.
Design: Prospective sequential comparison during intensive
care unit stay.
Setting: Surgical intensive care unit.
Patients: Patients with severe sepsis.
Interventions: Changes in standard fluid therapy, with predom-
inantly 6% hydroxyethyl starch from January 2005 to June 2005,
4% gelatin from January 2006 to June 2006, and only crystalloids
from September 2008 to June 2009.
Measurements and Main Results: Acute kidney injury was
defined by the presence of at least one RIFLE class; 118 patients
received hydroxyethyl starch, 87 patients received gelatin, 141
patients received only crystalloids. Baseline serum creatinine
values were similar. Patients received median cumulative doses
of 46 (interquartile range, 18 –92) mL/kg hydroxyethyl starch and
43 (interquartile range, 18 –76) mL/kg gelatin. Total median fluid
amounts were 649 (interquartile range, 275–1098) mL/kg in the
hydroxyethyl starch group, 525 (237– 868) mL/kg in the gelatin
group, and 355 (173–911) mL/kg in the crystalloid group. The
difference was statistically significant for hydroxyethyl starch
after adjustment for multiple testing. Mean daily fluid intake and
fluid balance were higher on days 0 and 1 in the crystalloid group.
Acute kidney injury occurred in 70% of patients receiving hy-
droxyethyl starch (adjusted p ⴝ .002) and in 68% of patients
receiving gelatin (adjusted p ⴝ .025) vs. 47% patients receiving
crystalloids. Need for renal replacement therapy tended to be
higher in the hydroxyethyl starch group (34%; adjusted p ⴝ .086)
and in the gelatin group (34%; adjusted p ⴝ .162) in comparison
to the crystalloid group (20%). Intensive care unit and hospital
mortality were similar in each group (hydroxyethyl starch: 35%
and 43%; gelatin: 26% and 31%; crystalloids: 30% and 37%).
Conclusion: Fluid resuscitation with only crystalloids was
equally effective, resulted in a more positive fluid balance only on
the first 2 days, and was associated with a lesser incidence of
acute kidney injury. (Crit Care Med 2011; 39:1335–1342)
KEY WORDS: plasma substitutes; severe sepsis; hydroxyethyl
starch; gelatin; crystalloids; acute renal failure

I n recent years, meta-analyses and
clinical trials have found that col-
loid or crystalloid resuscitation
lead to similar survival (1, 2).
Fluid choice therefore is deemed to be of
secondary importance and sepsis man-
agement guidelines leave the choice of
fluid to the user (3). Hydroxyethyl starch
(HES) is now the most commonly used
colloid, followed by gelatins, human al-
bumin, and dextrans (4 – 6).
*See also p. 1565.
From the Department of Anesthesiology and Inten-
sive Care Medicine (OB, KR, YS, BK, MK, NCR, MB,
CSH), Jena University Hospital, Jena, Germany; De-
partment of General, Visceral, and Vascular Surgery
(US), Jena University Hospital, Jena Germany; Depart-
ment of Cardiothoracic Surgery (KH), Jena University
Hospital, Jena, Germany.
Supplemental digital content is available for this ar-
ticle. Direct URL citations appear in the printed text and
are provided in the HTML and PDF versions of this article
on the journal’s Web site (www.ccmjournal.com).
Supported, in part, by an unrestricted grant of
the Thuringian Ministry of Cultural Affairs (Lande-
Although there is growing evidence
that the use of some starch solutions
may increase renal failure and mortal-
ity (7–10), “modern third-generation”
starches with low molecular weight
(HES 130/0.4) are claimed to be safe for
use in the intensive care unit (ICU) (11,
12). In 2007, the Food and Drug Admin-
istration approved use of 6% HES 130/
0.4 for the treatment and prophylaxis of
hypovolemia.
sprogramm ProExzellenz; PE 108-2), the Foundation
of Technology, Innovation, and Research Thuringia
(STIFT), and the German Sepsis Society (all
to OB).
Dr. Reinhart received honoraria/speaking fees
from B. Brann in the past. The remaining authors have
not disclosed any potential conflicts of interest.
For information regarding this article, E-mail:
konrad.reinhart@med.uni-jena.de
Copyright © 2011 by the Society of Critical Care
Medicine and Lippincott Williams & Wilkins
DOI: 10.1097/CCM.0b013e318212096a
Recent systematic reviews, however,
found that published evidence was insuf-
ficient to uphold the claim that low-
molecular-weight starches have fewer
side effects than older starches (9) and
identified only one exploratory trial with
HES 130/0.4 in the setting of ICU or
sepsis patients (10).
We therefore replaced 6% HES 130/
0.4 with 4% gelatin in the ICU, as well as
in the emergency department and in the
operating rooms. To our surprise, the
need for renal replacement therapy (RRT)
remained unchanged at 36%. Further-
more, both synthetic colloids were asso-
ciated with a dose-related increase of re-
nal failure (13).
Gelatins are thought to be less neph-
rotoxic, but this notion is derived from
trials in severe sepsis and kidney trans-
plantation patients in whom gelatins
served as control to older HES solutions
(8, 14). Randomized controlled trials that
address gelatin efficacy and safety are
scarce.

Crit Care Med 2011 Vol. 39, No. 6 1335

 Here, we report on the clinical out-
comes of severe sepsis patients from
three sequential treatment periods. Dur-
ing the first two periods, patients were
sequentially treated with the synthetic
colloids HES and gelatin and additional
crystalloids. In the third period, patients
received only crystalloids. Apart from
that, treatment protocols for severe sep-
sis patients remained unchanged
throughout.
MATERIALS AND METHODS
Patients
The study population consisted of all pa-
tients with severe sepsis or septic shock
treated in a 50-bed interdisciplinary surgical
ICU of a university hospital. Patients were
excluded from analysis if they had chronic
renal failure requiring hemodialysis before
ICU admission or had been enrolled in a ran-
domized HES trial (7). The local ethics board
waived the need for informed consent because
of the observational nature of the study.
Procedures
Before July 2006, standard fluid treatment
of hypovolemia in our ICU, emergency depart-
ment, and the operating room consisted of a
combination of crystalloids and synthetic col-
loids. In the first study phase from January 1,
2005 until June 30, 2005, patients were
treated with predominantly 6% HES (6% HES
130/0.4%; Voluven; Fresenius-Kabi Bad Hom-
burg) and crystalloids (HES group). Because
of concerns about HES-related nephrotoxicity,
HES solutions were replaced by 4% gelatin.
From January 1, 2006 to June 30, 2006, all
patients with severe sepsis or septic shock
were treated with 4% modified gelatin (Gela-
fusal; Serumwerk Bernburg AG, Bernburg,
Germany) and crystalloids (gelatin group). We
prospectively planned to compare the effects
of the different fluid regimens in a before-and-
after study with sequential treatment periods
of 6 months that were separated by a washout
period of 6 months to allow for a complete
implementation of the new standard fluid. The
analysis of the first two treatment periods was
published previously (13) and revealed that
the need for RRT was not different between
the two cohorts of ICU patients receiving HES
or gelatin. We therefore extended our study.
Gelatin was abolished and replaced by crystal-
loids starting in July 2006. Until now, only
balanced crystalloid solution (Jonosteril; Fre-
senius Kabi, Bad Homburg, Germany) has
been used in the ICU, emergency department,
and operating room, except in patients with
hyperkalemia who receive normal saline (NaCl
0.9%; Fresenius Kabi, Bad Homburg, Ger-
1336
many). Study phase III lasted from September
2008 until June 2009. All patients in phase III
were treated only with crystalloids (crystalloid
group).
Therapy for sepsis followed institutional
standard operating procedures based on the
guidelines of the Surviving Sepsis Campaign
(3) and the German Sepsis Society (15). Dur-
ing the HES and gelatin phase, the indication
for colloid administration was left to the dis-
cretion of the attending critical care physi-
cian. Standard hemodynamic management in-
cluded vasopressors with noradrenaline to
maintain a mean arterial pressure ⬎65 mm
Hg, in addition to repeated fluid challenges
with crystalloids and synthetic colloids (gela-
tin or hydroxyethyl starch) to allow the lowest
possible dose of vasopressors. If needed, do-
butamine (up to a dose of 10 ␮g/kg/min) or
epinephrine (up to 0.15 ␮g/kg/min) was
added. Extended hemodynamic monitoring
with pulmonary artery catheter (Swan-Ganz
CCOmbo; Edwards Life Sciences, Unter-
schleissheim, Germany) or PiCCO (Pulsiocath
5-Fr Thermodilution Catheter; Pulsion Medi-
cal Systems München, Germany) was consid-
ered if the required vasopressor dose increased
⬎0.4 ␮g/kg/min of noradrenaline. Central ve-
nous oxygen saturation was measured at least
two times per day from blood samples drawn
from the central venous catheter.
Human albumin was not used for volume
replacement. Its use was restricted to treat-
ment of severe hypoalbuminemia ⬍15
mmol/mL by 20% human albumin (Albunorm
20%; Octapharma, Langenfeld, Germany).
Continuous veno-venous hemodialysis was
the single modality for RRT in our ICU during
all study periods. Indications for RRT in our
ICU included diuretic-resistant oliguria (urine
output ⬍0.5 mL/kg body weight) persistent
for 6 hrs or anuria persistent for 3 hrs de-
spite adequate volume therapy or associated
with volume overload with threatened or
established pulmonary edema, the presence
of hyperkalemia, or severe metabolic acido-
sis (pH ⬍7.1).
Primary and Secondary
Outcomes: Definitions and Data
Collection Methods
The primary end point for our analysis was
the development of acute kidney injury (AKI).
New occurrence of RRT was a co-primary end
point. We defined AKI according to the stan-
dardized and validated RIFLE criteria, includ-
ing separate criteria for creatinine and urine
output, and using the criteria that led to the
worst outcome. Urine output was measured
over 24 hrs and calculated back to 6- or 12-
hourly values (16). Accordingly, RIFLE “risk”
was fulfilled by a 1.5-fold increase in serum
creatinine levels or urine output ⬍0.5 mL/
kg/hr (or both) for ⱖ24 hrs, RIFLE “injury” by
a two-fold increase in serum creatinine levels
or urine output ⬍0.3 mL/kg/hr (or both) for
ⱖ24 hrs, and RIFLE “failure” was defined by a
three-fold increase in serum creatinine levels
and new RRT or serum creatinine ⱖ354
␮mol/L with an acute rise of at least 44
␮mol/L or urine output ⬍0.3 mL/kg/hr ⱖ24
hrs (or both) or anuria ⱖ12 hrs for ⱖ24 hrs.
AKI was defined by one or more RIFLE criteria
or new occurrence of RRT. Groups included
only patients for whom the entry criterion was
the most severe classification, i.e., patients in
the RIFLE risk group had no more severe
injury than that defined by RIFLE risk.
Secondary end points were cumulative
fluid doses, total fluid balance, need for me-
chanical ventilation, maximum degree of or-
gan failure as assessed by Sequential Organ
Failure Assessment (SOFA) score, vasopressor
use, need for blood products, ICU length of
stay, and ICU and hospital mortality. Total
fluids were defined as all intravenous fluids
administered as maintenance or bolus fluids
or for intravenous drug administrations, in-
cluding blood products and human albumin,
parenteral and enteral feeding solutions, and
any oral intake. Creatinine clearance was de-
termined from serum creatinine using the
Cockcroft-Gault formula (17). Study period
was length of stay in the ICU.
Data recorded on admission included age,
gender, body weight, referring facility, and
surgical procedures preceding admission.
Screening for the presence of severe sepsis was
performed routinely by the senior physicians
in our ICU and documented by trained re-
search nurses. The main source of sepsis was
also documented. The Simplified Acute Phys-
iology Score II and SOFA scores were calcu-
lated within 24 hrs of admission by the attend-
ing physician who was in charge of the patient.
Data were collected prospectively from
vital sign monitors, ventilators, and infusion
pumps and automatically recorded by a clin-
ical information system. The clinical infor-
mation system (Copra System GmbH,
Sasbachwalden, Germany) provided staff
with complete electronic documentation,
order entry (e.g., medication, fluid dose, and
duration), and direct access to laboratory
and microbiology results. We recorded the
use of nephrotoxic drugs, i.e., nonsteroidal
anti-inflammatory drugs, diuretics, angio-
tensin-converting enzyme inhibitors, anti-
microbials and antimycotics, and iodinated
contrast media. Data on ICU and hospital
length of stay and ICU and hospital mortal-
ity were collected for all patients.
Severe sepsis was defined as the presence
of a defined focus on infection and at least two
of the four Systemic Inflammatory Response
Syndrome criteria and infection-related or-
gan dysfunction (18). A defined focus on
Crit Care Med 2011 Vol. 39, No. 6
Table 1. Patient baseline characteristics. RESULTS
Hydroxyethyl Gelatin Crystalloid
Starch Group Group Group Participants
(n ⫽ 118) (n ⫽ 87) (n ⫽ 141)
A total number of 400 patients were
Age (yr) mean ⫾ SD 65.0 ⫾ 16.2 63.0 ⫾ 14.2 66.4 ⫾ 13.9 treated for severe sepsis in our ICU dur-
Male, n (%) 79 (56) 57 (66) 91 (65) ing the three sequential study periods.
Comorbidities, n (%)
Hypertension 59 (42) 44 (51) 69 (49) Fifty-four patients fulfilled exclusion
Diabetes mellitus 32 (23) 19 (22) 36 (26) criteria (chronic renal failure requiring
Cancer 16 (11)a 19 (22) 34 (24) hemodialysis before ICU admission, n ⫽
Chronic renal failure 19 (13) 10 (11) 16 (11) 45; participation in a randomized con-
Liver cirrhosis 8 (6) 8 (9) 4 (3)
Surgical procedures, n (%)
trolled HES trial described previously,
Abdominal 48 (41) 25 (29)b 68 (48) n ⫽ 9) (7).
Cardiac/thoracic surgery 36 (31) 36 (41)a 38 (27) One hundred eighteen patients re-
Trauma 6 (5) 5 (6) 9 (6) ceived HES 6%, 87 received gelatin 4%,
Othersd 22 (19) 13 (15) 16 (11) and 141 patients received crystalloids.
No surgery 6 (5) 8 (9) 10 (7)
Source of sepsis, n (%) The baseline characteristics of patients
Respiratory 52 (44) 44 (51) 52 (37) were similar except for fewer patients
Abdominal 29 (25)c 25 (29)b 66 (47) with cancer in the HES group and fewer
Urogential 7 (6) 3 (3) 5 (4) patients with an abdominal source of sep-
Othere 30 (25)a 15 (17) 18 (13)
Hemodynamic and laboratory values,
sis in the HES and gelatin groups, and
median (interquartile range) fewer patients with abdominal surgery
Mean blood pressure, mm Hg 75 (68–83) 74 (68–81) 74 (68–80) and more patients with cardiac/thoracic
Heart rate, beats/min 96 (81–106) 95 (85–111) 92 (81–109) surgery in the gelatin group. SOFA score
Central venous pressure, mm Hg 8 (5–11)a 10 (7–12) 9 (7–13) was lower in the HES group, but creati-
Vasopressor use, n (%) 103 (87) 77 (89) 129 (91)
C-reactive protein, mg/dL 183 (88–268) 179 (91–241) 182 (100–251)
nine, creatinine clearance, and Simplified
Platelet count, ⫻103/mL 158 (99–217) 160 (104–246) 156 (93–235) Acute Physiology Score II scores did not
Serum creatinine, ␮mol/L 99 (83–159) 103 (71–150) 98 (74–142) differ between groups (Table 1).
Creatinine clearance, mL/min 54 (41–79) 68 (44–98) 59 (40–89)
Simplified Acute Physiology Score II 53 (41–63) 51 (41–64) 57 (41–70)
score Hemodynamics and Fluid
Sequential Organ Failure Assessment 9 (7–11)a 9 (7–13) 10 (7–12)
Balance
score

The p values were calculated with the Student’s t test or the Mann-Whitney test and Fisher’s exact
Hemodynamics and fluid balances
test, as appropriate. were analyzed during the first 14 days of
a
.01 ⬍ p ⬍ .05; b.001 ⬍ p ⬍ .01; cp ⬍ .001; dincludes neurosurgical, metabolic, renal urinary tract, the treatment period in the ICU. Mean
and gynecologic procedures; eincludes catheter-related, wound, central nervous system, and blood- arterial pressures were significantly
stream infections and endocarditis. Hydroxyethyl starch and gelatin groups were compared with the higher in the HES group after day 2 (Fig.
crystalloid group. 1). Although mean heart rates did not
differ, mean values for central venous
pressure and central venous oxygen sat-
infection was defined by either an organism variables, respectively. In case of primary uration were higher in the crystalloid
cultured from blood or a sterile site or an and secondary outcomes, resulting p values group on single study days (Fig. 1). Nor-
abscess or volume of infected tissue (e.g., were adjusted by the Bonferroni-Holm adrenaline use was similar in the first 3
pneumonia, peritonitis, skin or soft tissue method. Otherwise, the unadjusted p values days and significantly higher in the crys-
infection). Infection-related organ dysfunc- are reported. talloid group on days 4 – 8 and days 10
tion was considered to be present if at least In addition, forward and backward stepwise and 11 (Fig. 1). Daily fluid intake and
one of the following applied: respiratory, multiple logistic regression analysis based on mean daily fluid balance were signifi-
hematologic, or hepatic SOFA score ⬎1 and Akaike information criterion was performed to cantly higher in the crystalloid group at
cardiovascular SOFA score of 1, 3, or 4 or identify risk factors for AKI and RRT. The baseline (day 0) and on day 1 (Fig. 1).
renal SOFA score ⬎2. Organ dysfunction resulting model was checked by the Hosmer- Fluid volumes are given in Table 2. The
was defined previously (7). cumulative fluid volumes administered
Lemeshow goodness-of-fit test as well as the
area under the receiver-operating character- were significantly higher in both the HES
Statistical Analysis istic curve. In a second step, we added period and the gelatin groups than in the crys-
All analyses were performed in SPSS ver- effects of HES and gelatin as categorical talloid group on day 0 and 1 (Fig. 1). The
sion 17.0 and R version 2.11.1 (R Development variables with crystalloids as reference cat- total amount of fluids was significantly
Core Team 2010; R Foundation for Statistical egory. Thereafter, we added variables for the high in the HES group (Table 2). The
Computing, Vienna, Austria). All reported p total fluid amount when the three periods ratios of total fluid amounts were 1.47:1
values are two-sided. were analyzed separately. For the dose– (crystalloid compared to HES group) and
For univariate analyses, we applied the t response relationship between total fluid 1.44:1 (crystalloid compared to gelatin
test or the Mann-Whitney test and Fisher amount and AKI and RRT, we computed group) on days 0 –1 and 1.17:1 and 1.21:1
exact test for continuous and categorical linear and quadratic fits. for the first 4 days, respectively.

Crit Care Med 2011 Vol. 39, No. 6 1337

Figure 1. Means of daily heart rate, daily arterial pressure, daily central venous pressure, daily central venous saturation concentrations, daily Sequential
Organ Failure Assessment scores, cumulative daily dose of noradrenaline, daily fluid intake, and daily fluid balance for days 0 –14. Median and interquartile
range are presented. ***/ⴙⴙⴙp ⬍ .001, **/ ⫹⫹ p ⬍ .01, */⫹ p ⬍ .05 (*comparisons between hydroxyethyl starch [HES] and crystalloids groups; ⫹
comparisons between gelatin and crystalloid groups). ScvO2, central venous oxygen saturation.

1338 Crit Care Med 2011 Vol. 39, No. 6

Table 2. Total fluids and blood products administered during intensive care unit stay
Total fluid amount, median (mL/kg)
BWa (IQR)
Total amount of study fluids, median
(mL/kg) BW (IQR)
Total amount of crystalloids, median
(mL/kg) BW (IQR)
Blood products, median units (IQR)
Fresh-frozen plasma
Platelet concentrates
Red blood cells
Human albumin 20%, median (mL/kg)
BW IQR
BW, body weight; IQR, interquartile range.
a
Total fluid amount includes oral and parenteral fluids, including nutrition, drug administration, blood
products, albumin, and intravenous volume replacement; b.01 ⬍ p ⬍ .05; c.001 ⬍ p ⬍ .01; dp ⬍ .001. The
p values were calculated with the Mann-Whitney test. Hydroxyethyl starch and gelatin groups were
compared with the crystalloid group.
Morbidity and Mortality: AKI
and RRT
Patient exposure to a range of nephrotoxic
drugs did not differ between groups except for
a higher use of antimycotics in the HES
group and angiotensin-converting enzyme
inhibitors in the gelatin group (Supplemen-
tary Table I [Supplemental Digital Content 1,
http://links.lww.com/CCM/A232]). There was
also no difference in serum creatinine at on-
set of renal replacement therapy (Supple-
mentary Table II [Supplemental Digital Con-
tent 1, http://links.lww.com/CCM/A232]).
AKI occurred in 47% of patients in the
crystalloid group, in 70% of patients in
the HES group (adjusted p ⫽ .002), and
in 68% of patients in the gelatin group
(adjusted p ⫽ .025; Table 3). Patients
were allocated to RIFLE classes according
to their highest risk. The RIFLE risk and
RIFLE injury classes contained few pa-
tients who were evenly distributed. The
majority of patients were allocated to the
RIFLE failure class, 47% of patients after
administration of HES (adjusted p ⫽
.002), 40% after gelatin (adjusted p ⫽
.162), and 25% of patients after crystal-
loids. RRT tended to occur more fre-
quently in patients who received HES
(34%) compared to crystalloids (20%; ad-
justed p ⫽ .086). In the gelatin group, the
differences were only significant before p
value adjustment (Table 3).
Severity scores, ICU and hospital mor-
tality, and ICU length of stay did not
differ significantly between groups after
adjustment of p values, but there was a
trend toward shorter ICU length of stay in
the crystalloid group (Table 3).
Crit Care Med 2011 Vol. 39, No. 6
regarding age, hemodynamics, and se-
rum creatinine. However, patients receiv-
Hydroxyethyl Gelatin Crystalloid
ing synthetic colloids less often had an
Starch Group Group Group
abdominal source of sepsis.
(n ⫽ 118) (n ⫽ 87) (n ⫽ 141)
Colloids were associated with higher
649 (275–1098)b 525 (237–868) 355 (173–911) mean arterial pressure on most study
days, whereas on some days central ve-
46 (18–92)c 43 (18–76)d 105 (41–270) nous oxygen saturation and central ve-
nous pressure were higher in the crystal-
359 (206–619) 345 (216–649) 325 (180–636) loid group and mean SOFA scores were
lower in the colloid groups.
8 (4–16) 6 (4–12) 6 (4–12) The use of nephrotoxic drugs was not
3 (2–5)b 2 (1–3) 2 (1–3) different between study groups except
2 (2–4) 2 (2–2) 2 (1–4) that patients receiving HES also received
12 (5–24) 8 (5–27) 10 (4–19) more antifungals (amphotericin B and
fluconazole) and patients receiving gela-
tin also received more angiotensin-
converting enzyme inhibitors. Multivari-
ate analysis suggests that besides
administration of HES and gelatin, anti-
fungal medication as well as 20% human
albumin, cardiothoracic surgery, and io-
Multiple logistic regression analysis dinated contrast media are independent
with AKI as a dependent binary variable risk factors for AKI.
showed that baseline creatinine, cardio-
thoracic surgery, antimycotics, human Meaning of the Study in
albumin 20%, iodinated contrast media, Relation to Other Studies
HES 6%, and gelatin 4% were indepen-
dent risk factors. A dose-dependency of Adverse effects of starch solutions on
adverse effects (per mL/kg body weight) renal outcomes have been demonstrated
was only apparent for crystalloids with a repeatedly in sepsis and other critically ill
linear or maybe even quadratic risk in- patients (7–10, 14, 19, 20). Our findings
crease (Table 4). HES and gelatin were do not support the hypothesis that mod-
also independent risk factors after mul- ern third-generation starches are devoid
tiple logistic regression with RRT as of negative effects on kidney function (11,
dependent variable without an apparent 12) because 6% HES 130/0.4 comprised
dose– effect relationship for any colloid ⬎90% of the starches in this study (13).
or crystalloid (Supplementary Table III Experimental data also suggest that im-
[Supplemental Digital Content 1, pairment of kidney function after 6%
http://links.lww.com/CCM/A232]). HES 130/0.4 and 10% HES 200/0.5 only
differs by degree (21), and a recent exper-
imental study suggests that both 6% HES
DISCUSSION 130/0.4 and gelatin may induce signifi-
cantly more impairment of renal function
Principal Findings and damage to kidney morphology than
crystalloid (22).
In severe sepsis patients, a change of Other investigators also have found
fluid replacement regimen from HES and that adverse effects of starches or gelatins
gelatin to only crystalloids was associated were related to the cumulative dose (7,
with a significantly lower incidence of 20, 23, 24). In retrospective observational
AKI and RRT. ICU and hospital mortality studies, HES administration in low cu-
were similar between the groups. Inter- mulative doses in the range of 15–20 mL/
estingly, patients in the crystalloid group kg, which is less than half of the recom-
received more fluid volume and had a mended maximum daily dose, was not
more positive fluid balance only on the associated with an increased incidence of
first two days of treatment in the ICU. AKI or acute renal failure (6, 25).
The total amount of fluid administered The need of RRT that was observed in
was significantly higher in the HES than our study after synthetic colloids was
in the crystalloid group during the whole 34% after both HES and gelatin. This rate
ICU stay. is considerably higher than the rates ob-
The sequential study cohorts of severe served elsewhere after the use of crystal-
sepsis patients were similar at baseline loids or albumin. In a recent multicenter
1339
Table 3. Primary and secondary outcomes

Hydroxyethyl Gelatin Crystalloid

Starch Group Group Group
(n ⫽ 118) p Adjusted p ((n ⫽ 87) p Adjusted p ((n ⫽ 141)

RIFLE risk, n (%)a 15 (13) .698 1.000 10 (11) .831 1.000 15 (11)
RIFLE injury, n (%)b 12 (10) .842 1.000 14 (16) .319 1.000 16 (11)
RIFLE failure, n (%)c 56 (47) ⬍.001 0.002 35 (40) .018 .162 35 (25)
AKI, n (%)d 83 (70) ⬍.001 0.002 59 (68) .003 .025 66 (47)
Renal replacement therapy, n (%) 40 (34) .011 0.086 30 (34) .019 .162 28 (20)
Sequential Organ Failure score maximum, 11 (9–14) .355 1.000 13 (10–15) .332 1.000 12 (9–14)
median (IQR)e
Sequential Organ Failure score mean, 7 (6–10) .032 .227 8 (6–10) .122 .853 8 (6–11)
median (IQR)e
Intensive care unit mortality, n (%) 41 (35) .506 1.000 23 (26) .550 1.000 43 (30)
Hospital mortality, n (%) 51 (43) .311 1.000 27 (31) .393 1.000 52 (37)
Intensive care unit length of stay, days, 14 (6–28) .070 .421 13 (6–26) .167 1.000 10 (5–20)
median (IQR)

IQR, interquartile range.

a
Five-fold increase in serum creatinine levels and/or urine output ⬍0.5 mL/kg/hr for ⱖ24 hrs; btwo-fold increase in serum creatinine levels and/or urine
output ⬍0.3 mL/kg/hr for ⱖ24 hrs; cthree-fold increase in serum creatinine levels and/or renal replacement therapy, serum creatinine ⱖ354 ␮mol/L with
an acute increase of at least 44 ␮mol/L, and/or urine output ⬍0.3 mL/kg/hr ⱖ24 hrs or anuria ⱖ12 hrs for ⱖ24 hrs; ddefined by any RIFLE category; ewithin
28 days of admission to the intensive care unit. The p values were calculated with the Mann-Whitney test and Fisher’s exact test, as appropriate. For p value
adjustment, the Bonferroni-Holm method was used.

Table 4. Multiple logistic regression analysis with acute kidney injury as dependent binary variable
trial with 537 severe sepsis patients, need
Adjusted Odds Ratio for RRT was 18.6% in the Ringer’s lactate
n (95% Confidence Interval) p group and 31% in the HES group; 28-day
mortality rates were 24.1% and 26.7%,
Age (per yr) 333 1.02 (1.00–1.03) .101 respectively (7). In the severe sepsis sub-
Simplified Acute Physiology Score II 333 1.02 (1.00–1.03) .064
(per point)
group of the SAFE trial (n ⫽ 1218),
Baseline creatinine (per ␮mol/L) 333 1.01 (1.01–1.02) ⬍.001 which compared 4% albumin to 0.9%
Cardiac/thoracic surgery, yes vs. no 333 1.94 (1.09–3.46) .024 saline, RRT occurred in 18.7% of patients
Antimycotics, yes vs. no 333 2.54 (1.15–5.61) .021 assigned to albumin compared to 18.2%
Vancomycin, yes vs. no 333 1.85 (0.99–3.45) .053
Ionidated contrast media, yes vs. no 333 2.02 (1.16–3.53) .013 of patients assigned normal saline. The
Human albumin 20%, yes vs. no 333 2.19 (1.24–3.87) .007 28-day mortality rates were 30.7% and
Added after model selection 35.3%, respectively (26).
Period effects A high need for RRT may coincide
Period, reference ⫽ crystalloids
Gelatin 4% 333 3.65 (1.81–7.35) ⬍.001 with a widespread use of synthetic col-
HES 6% 130/0.4 333 4.52 (2.27–8.99) ⬍.001 loids. Synthetic colloids and especially
Dose–effect relationship HES solutions are preferred volume ex-
Crystalloids (per mL/kg BW) 136 1.005 (1.002–1.008) .002
Crystalloids, quadratic (per mL/kg BW 136 1.00001 (1.00000–1.00001) .009
panders in Germany (4, 27). A represen-
squared) tative survey conducted in German ICUs
HES 6% 130/0.4 (per mL/kg BW) 112 1.010 (0.998–1.022) .115 in 2003–2004 found an incidence of acute
HES 6% 130/0.4, quadratic (per mL/kg BW 112 1.0001 (1.0000–1.0001) .116 renal failure in these patients of 42.4%
squared) (28). The survey also documented that
Gelatin 4% (per mL/kg BW) 85 1.001 (0.990–1.012) .865
Gelatin 4%, quadratic (per mL/kg BW 85 1.00000 (0.99997–1.00002) .785 35.2% of patients with severe sepsis and
squared) septic shock received synthetic colloids
mostly in form of HES 6%, HES 10%, or
BW, body weight; HES, hydroxyethyl starch. gelatin (29).
Variables considered in the analysis: age, liver cirrhosis, diabetes, baseline creatinine, baseline
An international sepsis register docu-
Simplified Acute Physiology Score II, cardiac/thoracic surgery, nonsteroidal anti-inflammatory drugs,
angiotensin-converting enzyme inhibitors, aminoglycosides, antimycotics, vancomycin, ionidated
mented an overall need for RRT of 21.3%
contrast media, diuretics, human albumin 20%, period effects of HES 6% 130/0.4 and gelatin 4% as in ⬎12,000 patients from eight countries.
categorical variables with crystalloids as reference category, crystalloids (linear and quadratic, per The only European country was Ger-
mL/kg BW), HES 6% (130/0.4; linear and quadratic, per mL/kg BW), and gelatin 4% (linear and many, where HES is widely used. Need
quadratic, per mL/kg BW). Forward and backward stepwise multiple logistic regression analysis based for RRT in German patients was 33.4%,
on Akaike information criterion) was used to derive a multivariable model in which the Hosmer- whereas the RRT rates in non-European
Lemeshow goodness of fit test (C-test: p ⫽ .712, H-test: p ⫽ .464) and receiver-operator curve analysis
countries ranged between 11.7% and
(area under the curve ⫽ 0.62) indicate an acceptable fit and discrimination, respectively. The p values
were obtained by Wald’s test. After model selection, a period effect was studied, then the model was 25.9%. Hospital mortality rates for severe
applied to the subcohorts, and the variables for synthetic colloids and crystalloids were added sepsis patients were 43.4% in Germany
correspondingly to investigate the dose– effect relationship. and 49.6% overall (30).

1340 Crit Care Med 2011 Vol. 39, No. 6

 AKI occurred in 47% of patients in the
crystalloid group, and significantly more
frequently after gelatin 4% (68%) and
HES 6% (70%). The need for RRT was
also higher in both the HES and the
gelatin groups, but the difference was
nonsignificant after adjustment for mul-
tiple testing. There may be several rea-
sons for the failure to achieve signifi-
cance. The Bonferroni-Holm adjustment
for multiple testing may have been too
conservative or the sample size may have
been too small, particularly in the gelatin
group. Gelatin has a lower molecular
weight (30 kDa) than HES (130 kDa) and
was used as 4% solution compared to the
6% HES solution. Little is known about
the mechanisms of colloid-associated
renal impairment (31). Previous studies
(8, 14) suggest that gelatin may have a
lesser effect on renal function com-
pared to HES. The lower incidence of
acute renal failure with gelatin com-
pared to HES observed in one random-
ized controlled trial in severe sepsis pa-
tients may relate to the fact that in this
trial a 3% gelatin solution was com-
pared with a 6% HES solution (8).
Gelatin-associated side effects may
have escaped notice in trials in which gel-
atin was used in comparison to HES (8, 14,
32). However, a first report appeared re-
cently that demonstrated a similar degree
of vesicle formation in the proximal tubule
in septic rats with an increase in serum
creatinine and urea concentrations after
administration of 10 mL/kg 4% gelatin and
6% HES 130/0.4 (22).
Our study adds to the findings from
prospective studies of critically ill pa-
tients, which have found that crystalloid-
to-colloid volume ratios of only 2:1 or
less (1, 7, 33–35) are necessary to achieve
and maintain similar hemodynamic end
points. It seems that over longer periods
of time, the extent to which colloids are
thought to remain in the vasculature is
not so much greater than that of crystal-
loids. Within the first hours after fluid
administration, hematocrit decreased
more rapidly in the colloid groups, as was
shown by Wills et al (35) in children with
Dengue shock (HES 200/0.5 or dextran
70), or in smaller studies on preoperative
hemodilution (36), or in healthy volun-
teers (37). However, over the following
hours, hematocrit values increased more
in the crystalloid than in the colloid
groups (35) and were no longer different
at 24 hrs (36, 37).
Crit Care Med 2011 Vol. 39, No. 6
Limitations of This Study and
Future Research
This study has several limitations,
namely that it was a single-center un-
blinded study with small sample sizes in
all groups, with changes of fluid regimen
with possibly extended inception period,
post hoc adjusted analyses, and multivar-
iate modeling. Conclusions therefore
should be drawn with circumspection.
The lack of baseline differences does not
preclude the presence of unidentified
group differences because of the nonran-
domized nature of the study. It remains
unclear whether the observed differences
in mean arterial pressures are attribut-
able to different patient management.
The crystalloid group included more pa-
tients with an abdominal source of sepsis.
We do not believe this constitutes a con-
founder, because the multivariate analy-
sis identified no association between
source of sepsis and AKI, and the need for
RRT was found to be significantly higher
in patients with abdominal than pulmo-
nary sepsis in a large European survey
(38). That crystalloids, but not HES or
gelatin, were found to have a significant
dose-dependent effect on AKI perhaps
may be attributable to the smaller num-
ber of observations compared to crystal-
loids. Furthermore, the dose– effect rela-
tionship of HES 6% or gelatin 4% might
be more complex than a linear or qua-
dratic relation.
Some observations strengthen our
findings. The similarity between disease
severity not only at baseline but also dur-
ing the course of the disease, as well as
the lack of differences in ICU and hospital
mortality between the synthetic colloid
and the crystalloid groups, support the
assumption that the most likely cause
for the increased incidence of AKI must
be attributed to synthetic colloid use.
Thus, our study generates the hypoth-
esis that not only 6% HES 130/0.4 but
also 4% gelatin may be associated with
AKI in patients with severe sepsis. This
study provides hypothesis generation
and calls for high-quality, randomized,
controlled trials to provide evidence on
the safety of HES 130/0.4 and gelatin.
Fortunately, such trials are now under-
way for HES 130/0.4, such as the
CHEST trial (www.clinicaltrials.gov;
NCT00935168) comparing HES to nor-
mal saline in planned 7000 ICU patients
and the 6S trial (www.clinicaltrials.gov;
NCT00962156) comparing HES 130/0.4
to Ringer’s acetate in planned 800 ICU
patients. Study designs can be found on
the respective Web sites.
CONCLUSIONS
The results of this study question the
superiority of synthetic colloids over
crystalloids with respect to their volume-
saving potential in severe sepsis. Syn-
thetic colloid use did not lead to a less
positive total fluid balance. This chal-
lenges a common rationale for their use
in patients with severe sepsis. The results
from our study place serious doubt on the
assumption that third-generation HES
130/0.4 and low-molecular-weight gelatin
are innocuous regarding renal function. To
date, there is a lack of evidence from well-
designed studies in intensive care to sup-
port the widespread use of synthetic col-
loids. Fortunately, such trials are now
underway for 6% HES 130/0.4.
REFERENCES
1. Finfer S, Bellomo R, Boyce N, et al: A com-
parison of albumin and saline for fluid resus-
citation in the intensive care unit. N Engl
J Med 2004; 350:2247–2256
2. Perel P, Roberts I: Colloids versus crystal-
loids for fluid resuscitation in critically ill
patients. Cochrane Database Syst Rev 2009:
CD000567
3. 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
4. Schortgen F, Deye N, Brochard L, et al: Pre-
ferred plasma volume expanders for critically
ill patients: Results of an international sur-
vey. Intensive Care Med 2004; 30:2222–2229
5. Basora M, Moral V, Llau JV, et al: [Perioper-
ative colloid administration: A survey of
Spanish anesthesiologists’ attitudes]. Rev
Esp Anestesiol Reanim 2007; 54:162–168
6. Sakr Y, Payen D, Reinhart K, et al: Effects of
hydroxyethyl starch administration on renal
function in critically ill patients. Br J An-
aesth 2007; 98:216 –224
7. Brunkhorst FM, Engel C, Bloos F, et al: In-
tensive insulin therapy and pentastarch re-
suscitation in severe sepsis. N Engl J Med
2008; 358:125–139
8. 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:
911–916
9. Dart AB, Mutter TC, Ruth CA, et al: Hydroxy-
ethyl starch (HES) versus other fluid thera-
pies: Effects on kidney function. Cochrane
Database Syst Rev 2010:CD007594
10. Zarychanski R, Turgeon AF, Fergusson DA,
et al: Renal outcomes and mortality follow-
ing hydroxyethyl starch resuscitation of crit-
1341
 ically ill patients: Systematic review and
meta-analysis of randomized trials. Open
Med 2009; 3:E196 –E209
11. Boldt J: PRO: Hydroxyethylstarch can be
safely used in the intensive care patient–the
renal debate. Intensive Care Med 2009; 35:
1331–1336
12. Westphal M, James MF, Kozek-Langenecker
S, et al: Hydroxyethyl starches: Different
products—Different effects. Anesthesiology
2009; 111:187–202
13. Schabinski F, Oishi J, Tuche F, et al: Effects
of a predominantly hydroxyethyl starch
(HES)-based and a predominantly nonHES-
based fluid therapy on renal function in sur-
gical ICU patients. Intensive Care Med 2009;
35:1539 –1547
14. Cittanova ML, Leblanc I, Legendre C, et al:
Effect of hydroxyethylstarch in brain-dead
kidney donors on renal function in kidney-
transplant recipients. Lancet 1996; 348:
1620 –1622
15. Reinhart K, Brunkhorst FM, Bone HG, et al:
[Diagnosis and Therapy of Sepsis: Guidelines
of the German Sepsis Society Inc. and the
German Interdisciplinary Society for Inten-
sive and Emergency Medicine]. Internist
(Berl) 2006; 47:356 –373
16. Bagshaw SM, George C, Bellomo R, et al: A
comparison of the RIFLE and AKIN criteria
for acute kidney injury in critically ill pa-
tients. Nephrol Dial Transplant 2008; 23:
1569 –1574
17. Cockcroft DW, Gault MH: Prediction of cre-
atinine clearance from serum creatinine.
Nephron 1976; 16:31– 41
18. 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
19. Dehne MG, Mühling J, Sablotzki A, et al:
[Effect of hydroxyethyl starch solution on
kidney function in surgical intensive care
patients]. Anasthesiol Intensivmed Not-
fallmed Schmerzther 1997; 32:348 –354
1342
20. Rioux JP, Lessard M, De Bortoli B, et al:
Pentastarch 10% (250 kDa/0.45) is an inde-
pendent risk factor of acute kidney injury
following cardiac surgery. Crit Care Med
2009; 37:1293–1298
21. Huter L, Simon TP, Weinmann L, et al: Hy-
droxyethylstarch impairs renal function and
induces interstitial proliferation, macro-
phage infiltration and tubular damage in an
isolated renal perfusion model. Crit Care
2009; 13:R23
22. Schick MA, Isbary TJ, Schlegel N, et al: The
impact of crystalloid and colloid infusion on
the kidney in rodent sepsis. Intensive Care
Med 2010; 36:541–548
23. Bork K: Pruritus precipitated by hydroxy-
ethyl starch: A review. Br J Dermatol 2005;
152:3–12
24. Tseng MY, Hutchinson PJ, Kirkpatrick PJ:
Effects of fluid therapy following aneurysmal
subarachnoid haemorrhage: A prospective
clinical study. Br J Neurosurg 2008; 22:
257–268
25. Boussekey N, Darmon R, Langlois J, et al:
Resuscitation with low volume hydroxyeth-
ylstarch 130 kDa/0.4 is not associated with
acute kidney injury. Crit Care 2010; 14:R40
26. Finfer S, McEvoy S, Bellomo R, et al: Impact
of albumin compared to saline or organ func-
tion and mortality of patients with severe
sepsis. Intensive Care Med 2011; 37:86 –96
27. Kastrup M, Markewitz A, Spies C, et al: Cur-
rent practice of hemodynamic monitoring
and vasopressor and inotropic therapy in
post-operative cardiac surgery patients in
Germany: Results from a postal survey. Acta
Anaesthesiol Scand 2007; 51:347–358
28. Engel C, Brunkhorst FM, Bone HG, et al:
Epidemiology of sepsis in Germany: Results
from a national prospective multicenter
study. Intensive Care Med 2007; 33:606 – 618
29. Hartog CS, Brunkhorst FM, Bloos F, et al:
Practice of volume therapy in patients with
severe sepsis: Results from a nationwide sep-
sis prevalence study. Intensive Care Med
2010; 36:553–554
30. Beale R, Reinhart K, Brunkhorst FM, et al:
Promoting Global Research Excellence in Se-
vere Sepsis (PROGRESS): Lessons from an
international sepsis registry. Infection 2009;
37:222–232
31. Schortgen F, Brochard L: Colloid-induced
kidney injury: Experimental evidence may
help to understand mechanisms. Crit Care
2009; 13:130
32. Mahmood A, Gosling P, Vohra RK: Random-
ized clinical trial comparing the effects on
renal function of hydroxyethyl starch or ge-
latine during aortic aneurysm surgery. Br J
Surg 2007; 94:427– 433
33. van der Heijden M, Verheij J, van Nieuw
Amerongen GP, et al: Crystalloid or colloid
fluid loading and pulmonary permeability,
edema, and injury in septic and nonseptic
critically ill patients with hypovolemia. Crit
Care Med 2009; 37:1275–1281
34. Verheij J, van Lingen A, Raijmakers PG, et al:
Effect of fluid loading with saline or colloids
on pulmonary permeability, oedema and
lung injury score after cardiac and major
vascular surgery. Br J Anaesth 2006; 96:
21–30
35. Wills BA, Nguyen MD, Ha TL, et al: Compar-
ison of three fluid solutions for resuscitation
in dengue shock syndrome. N Engl J Med
2005; 353:877– 889
36. Fenger-Eriksen C, Hartig Rasmussen C, Kap-
pel Jensen T, et al: Renal effects of hypoten-
sive anaesthesia in combination with acute
normovolaemic haemodilution with hy-
droxyethyl starch 130/0.4 or isotonic saline.
Acta Anaesthesiol Scand 2005; 49:969 –974
37. Lobo DN, Stanga Z, Aloysius MM, et al: Effect
of volume loading with 1 liter intravenous
infusions of 0.9% saline, 4% succinylated
gelatine (Gelofusine) and 6% hydroxyethyl
starch (Voluven) on blood volume and endo-
crine responses: A randomized, three-way
crossover study in healthy volunteers. Crit
Care Med 2010; 38:464 – 470
38. Vincent JL, Sakr Y, Sprung CL, et al: Sepsis
in European intensive care units: results of
the SOAP study. Crit Care Med 2006; 34:
344 –353
Crit Care Med 2011 Vol. 39, No. 6