Professional Documents
Culture Documents
Volume of fluids administered during resuscitation for severe sepsis and septic
shock and the development of the acute respiratory distress syndrome☆
a, a a a
Dong W. Chang, MD , Richard Huynh, MD , Eric Sandoval, MD , Neung Han, MD , Clinton
c b
J. Coil, MD, MPH , Brad J. Spellberg, MD
a Divisions of Respiratory and Critical Care Physiology and Medicine, Los Angeles County Department of Health Services and Los Angeles Biomedical Research Institute at Harbor-University of
California Los Angeles Medical Center, Torrance, CA, USA
b General Internal Medicine, Los Angeles County Department of Health Services and Los Angeles Biomedical Research Institute at Harbor-University of California Los Angeles Medical Center,
Torrance, CA, USA
c Department of Medicine and the Department of Emergency Medicine, Los Angeles County Department of Health Services and Los Angeles Biomedical Research Institute at Harbor-University of
California Los Angeles Medical Center, Torrance, CA, USA
article info abstract
Keywords: Purpose: The purpose of this study was to examine the association between the volume of intravenous (IV) fluids administered
Sepsis in the resuscitative phase of severe sepsis and septic shock and the development of the acute respiratory distress syndrome
Acute respiratory distress syndrome (ARDS).
Resuscitation Materials and methods: This was a retrospective cohort study of adult patients admitted with severe sepsis and septic shock at
Fluid balance
a large academic public hospital. The relationship between the volume of IV fluids administered and the development of
ARDS was examined using multivariable logistic regression analysis. Results: Among 296 patients hospitalized for severe
sepsis and septic shock, 75 (25.3%) developed ARDS. After controlling for confounding variables, there was no signi ficant
association between the volume of IV fluids administered in the first 24 hours of hospitalization and the development of ARDS
(odds ratio [OR], 1.05; 95% confidence interval [CI], 0.95-1.18). Serum albumin (OR, 0.52; 95% CI, 0.31-0.87) and Acute
Physiology and Chronic Health Evaluation II score (OR, 1.08; 95% CI, 1.04-1.13) on admission were the most informative
covariates for the development of ARDS in the regression model.
Conclusions: For patients hospitalized for severe sepsis and septic shock, fluid administration to improve end-organ perfusion
should remain the top priority in early resuscitation despite the potential risk of inducing ARDS.
© 2014 Elsevier Inc. All rights reserved.
1. Introduction that, in the resuscitative phase of
severe sepsis and septic shock,
Severe sepsis and septic shock restoring intravascular volume and
are the most severe manifestations maintaining end-organ perfusion are
of the sepsis syndrome and the top priorities. However, one
characterized by end-organ concern regarding aggressive
hypoperfu-sion and hypotension volume resuscitation is that it may
due to infection [1]. Previous increase the risk for complications
studies have shown that early goal- due to volume overload, such as the
directed resuscitation of patients acute respiratory distress syndrome
with severe sepsis and septic shock (ARDS) [3–5].
improves mortality [1,2]. One of the
key interventions in early goal- Acute respiratory distress
directed therapy is aggressive syndrome is a devastating
administra-tion of intravenous (IV) complication of sepsis that affects
fluids using physiologic targets to its clinical management and
assess for improvements in end- outcomes [6–8]. Previous studies
organ perfusion [2]. For patients have shown that interventions that
with severe sepsis or septic shock, minimize the administration of IV
the Surviving Sepsis Guidelines fluids in hemodynamically stable
recommend an initial bolus of 30 patients with ARDS decrease the
mL/kg of fluid followed by repeated duration of mechanical ventilation
fluid administration as long as there and intensive care unit (ICU) stay
are continued responses in hemody- without compromising end-organ
namic parameters [1]. The rationale perfusion [9]. As such, the fluid
for these recommendations is management strategies for ARDS
can become discordant with those of
sepsis when pulmonary edema
complicates the early resuscitative
☆ Research support: none.
phase of the sepsis syndrome.
Corresponding author. Department of
Medicine, Harbor-UCLA Medical Center, Given the increased propensity
Box 405, 1000 W. Carson St, Torrance, CA of the lungs to develop pulmonary
90509. Tel.: +1 310 222 3803; fax: +1 310 edema during sepsis, it is possible
328 9849.
that the effects of positive fluid
E-mail address: dchang@labiomed.org
(D.W. Chang). balance during fluid resuscitation in
patients admitted with severe sepsis
http:/ and septic shock may increase the
/dx.d risk for developing ARDS.
oi.org
/10.1
016/j.
jcrc.2
014.0
6.005
0883-
9441/
©
2014
Elsev
ier
Inc.
All
rights
reser
ved.
1012 D.W. Chang et al. / Journal of Critical Care 29 (2014) 1011–1015
confounding factors on the
The objective of this study was We examined the association All data were collected by relationship between primary and
explore this risk by examining the between the total volume of IV retrospective chart review by 2 of secondary predictor variables and
association between the volume of fluids administered in the first 24 the investigators (RH and NH) using the development of ARDS. The
IV fluids administered in the first 24 hours of hospitalization and the a structured data abstraction form. volume of IV fluids administered in
hours of hospitalization for severe development of ARDS within 72 Patient identifiers were removed the first 24 hours (primary variable
sepsis or septic shock and the hours of hospital admission in from aggregated data and were of interest) and the administration of
development of ARDS. We patients admitted with severe sepsis coded using a numeric identifier. an initial fluid bolus (secondary
hypothesized that increased IV fluid and septic shock. The total volume The list of numeric codes and variable) were included in all
administration is associated with a of IV fluids administered was corresponding patient identifiers was models. The administration of an
greater incidence of ARDS despite determined from nursing flow sheets maintained in a password-protected initial fluid bolus was included in
controlling for known predisposing that documented the total fluid computer by one of the investigators the multivariable models despite the
factors. intake and output for each patient. (DWC). The study was approved as lack of association in univariate
The development of ARDS was analysis because this variable was of
an exempt protocol by the John G
identified by chart review using the clinical interest as a potential risk
2. Materials and methods Wolf Institutional Review Board at
Berlin definition of ARDS [11,12]. factor for ARDS. The volume of IV
the Los Angeles Biomedical
Specifically, patients were identified fluids administered in the first 6
2.1. Study design and patients Research Institute. The need for
as having ARDS if they had bilateral hours of hospitalization (secondary
informed consent was waived.
opacities on chest radiograph, acute variable) was not included in the
This was a retrospective
onset of respiratory failure not fully multivariable models because it was
observational study of patients with
explained by cardiac failure, and
severe sepsis or septic shock 2.3. Statistical analysis collinear with the volume of IV
admitted to the emergency PaO2/FIO2 ratio less than 300 mm Hg fluids administered in the first 24
department of a large academic within 72 hours of admission for The primary analysis compared hours. We evaluated the models for
county hospital (Harbor-UCLA severe sepsis or septic shock. the volume of IV fluids adminis- multicollinearity using a correlation
Medical Center, Torrance, CA) tered in the first 24 hours of coefficient matrix of the covariates
between December 2011 and hospitalization in patients who [15]. All covariates included in the
January 2013. The cohort of patients 2.2. Risk factors for the models showed a correlation of less
developed ARDS vs those who did
development of ARDS than 0.60. The values of the
with severe sepsis and septic shock not. The Student t test was used to
was identified by retrospective chart compare the volume of fluids variables in the models are reported
The primary predictor variable as adjusted odds ratios (ORs) with
review of all patients seen in the administered to each group at 6 and
for the development of ARDS was 95% confidence intervals (CIs). A P
emergency department using the 24 hours. The secondary predictor
the volume of IV fluids (liters) value b .05 was used for statistical
clinical definition from the variables, volume of fluids
administered in the first 24 hours of significance in the model variables.
Surviving Sepsis Guidelines and the administered in the first 6 hours of
hospitalization. Secondary variables The data analysis was performed
Society of Critical Care hospitalization and the
included the administration of an using JMP version 11.0 (SAS
Medicine/European Society of administration of an initial fluid
initial fluid bolus of greater than 20 Institute, Cary, NC).
Intensive Care Medicine/American bolus, were analyzed using a t test
mL/kg and the volume of IV fluids
College of Chest Physicians 2
or χ test, as appropriate. A P value
given in the first 6 hours of
International Sepsis Definitions b .05 was used to define statistical
Conference [1,10]. Specifically, hospitalization. Additional risk
factors that were examined included significance. To control for factors
sepsis was defined as a probable and that may confound the relationship
suspected infection as documented patient demographics (age, sex,
race/ ethnicity), medical between the volume of fluids
by a physician in the medical record 3. Results
comorbidities (Charlson administered in the first 24 hours
with at least 2 manifestations of a (primary predictor variable) and the
systemic inflammatory response Comorbidity Index), presenting vital 3.1. Baseline characteristics between
signs and laboratory data (serum development of ARDS (primary
(temperature N 38.3°C or b35.6°C, the ARDS and control groups
sodium, bicarbon-ate, glomerular outcome variable), multivariable
heart rate N 90 beats per minute, logistic regression analysis was
respiratory rate N 20/min, or white filtration rate, albumin, lactate), The study cohort consisted of
medications, use of blood products, performed [13,14]. Prior to
blood cell count N 12.0 × 10 or
3 296 patients who were admitted
composite scores for severity of multivar-iable analysis, univariate
3 with severe sepsis or septic shock.
b4.0 × 10 , or normal white blood illness (Acute Physiology and analysis was performed on all
Of these, 75 patients (25.3%)
cell count with N 10% immature Chronic Health Evaluation covariates to identify those with the
developed ARDS within 72 hours of
band forms). Severe sepsis was [APACHE] II, Sequential Organ strongest association with the
hospital admission. There were
defined as sepsis with organ Failure Assessment [SOFA]), and development of ARDS. All
multiple differences in the baseline
dysfunction or tissue hypoperfusion suspected source of sepsis. covariates that were significant at a
characteristics between patients that
(systolic blood pressure b 90 mm Additional clinical outcomes that level of P b .10 were considered for
developed ARDS and those that did
Hg, mean arterial pressure b 70 mm were examined include hospital and multivariable analysis [15]. For
multivariable regres-sion analysis, 2 not develop ARDS (Table 1). The
Hg, lactate above the upper limit of ICU mortality. Laboratory studies
models were generated. First, all ARDS group was older (mean age,
normal, urine output b 30 mL/h that were not normally distributed
covariates of interest were included 71.5 vs 62.9 years; P = .004). In
despite initial fluid resuscitation, were log transformed for statistical
in a comprehensive model that addition, the ARDS group had a
creatinine N 2.0 mg/dL, bilirubin N analysis. The APACHE II and SOFA
contained any factors that showed lower proportion of Hispanic
2 mg/dL, platelet count b 100,000, scores were calculated from the
an association in the univariate patients compared to the no-ARDS
international normalized ratio N clinical data available on admission.
1.5). Septic shock was defined as analysis. Next, the same covariates group (17.3% vs 38.9%, P b .001).
sepsis-induced hypotension were included in stepwise logistic The serum
resulting in a systolic blood pressure regression analysis to generate a
less than 90 mm Hg or mean arterial parsimonious model that contained
pressure less than 70 mm Hg only the most informative
requiring vasopressor support covariates. These 2 models were
despite IV fluid administration used as complementary approaches
[1,10]. to control for the influence of
D.W. Chang et al. / Journal of Critical Care 29 (2014) 1011–1015 1013
Table 1
Baseline characteristics