Original Research

Journal of Intensive Care Medicine

1-9
The Effect of Fluid Resuscitation Timing in © The Author(s) 2023
Article reuse guidelines:
sagepub.com/journals-permissions
Early Sepsis Resuscitation DOI: 10.1177/08850666231180530
journals.sagepub.com/home/jic

Hani I. Kuttab, MD1, Chad G. Evans, PhD1, Joseph D. Lykins, MD2,

Michelle D. Hughes, MD1, Jason A. Kopec, MD3,
Michael A. Hernandez, MD4, and Michael A. Ward, MD1

Abstract
Purpose: The dose and timing of early fluid resuscitation in sepsis remains a debated topic. The objective of this study is to
evaluate the effect of fluid timing in early sepsis management on mortality and other clinical outcomes.
Methods: Single-center, retrospective cohort study of emergency-department-treated adults (>18 years, n = 1032) presenting with
severe sepsis or septic shock. Logistic regression evaluating the impact of 30 mL/kg crystalloids timing and mortality-versus-time plot
controlling for mortality in emergency department sepsis score, lactate, antibiotic timing, obesity, sex, systemic inflammatory response
syndrome criteria, hypotension, and heart and renal failures. This study is a subanalysis of a previously published investigation.
Results: Mortality was 17.1% (n = 176) overall and 20.4% (n = 133 of 653) among those in septic shock. 30 mL/kg was given
to 16.9%, 32.2%, 16.2%, 14.5%, and 20.3% of patients within ≤1, 1 ≤ 3, 3 ≤ 6, 6 ≤ 24, and not reached within 24 h, respectively.
A 24-h plot of adjusted mortality versus time did not reach significance, but within the first 12 h, the linear function showed a
per-hour mortality increase (odds ratio [OR] 1.29, 95% confidence interval [CI] 1.02-1.67) which peaks around 5h, although the
quadratic function does not reach significance (P = .09). When compared to patients receiving 30 mL/kg within 1 h, increased
mortality was observed when not reached within 24 h (OR 2.69, 95% CI 1.37-5.37) but no difference when receiving this volume
between 1 and 3 (OR 1.11, 95% CI 0.62-2.01), 3 and 6 (OR 1.83, 95% CI 0.97-3.52), or 6 and 24 h (OR 1.51, 95% CI 0.75-3.06).
Receiving 30 mL/kg between 1 and 3 versus <1 h increased the incidence of delayed hypotension (OR 1.83, 95% CI 1.23-2.72) but
did not impact need for intubation, intensive care unit admission, or vasopressors.
Conclusions: We observed weak evidence that supports that earlier is better for survival when reaching fluid goals of 30 mL/kg,
but benefits may wane at later time points. These findings should be viewed as hypothesis generating.

Keywords
sepsis, shock, septic, fluids, timing, early resuscitation, emergency department

Introduction
Fluid resuscitation is recognized as an important factor for early
sepsis management, however, the mandate on specific dosing,
timing, and consideration of comorbidities has come under
recent scrutiny.1‐6 In 2018, the Surviving Sepsis Campaign
(SSC) released a 1h treatment bundle recommending to begin
rapid administration of 30 mL/kg crystalloid for sepsis-induced
hypoperfusion with little to no supporting evidence specific to
very early fluids.7 Recently, the SSC updated their guidelines
that no longer includes fluid resuscitation as part of a 1h
bundle but still recommends initiating 30 mL/kg immediately
upon recognition and considers this “best practice” with
“no new data suggesting that a change is needed.”8 However,
in a separate policy statement, the American College of
Emergency Physicians (ACEP) have deviated from these guide-
lines and do not support a prespecified volume of fluid, noting
that “an initial volume of 500 to 1000 mL of crystalloid is a
common and reasonable practice” with no further mention of
timing.9
Many, like ACEP, cite evidence that suggests positive fluid
balances and large volume resuscitation may have deleterious
impacts. However, it is important to clarify that these findings
come from largely intensive care unit (ICU)-based studies
1
Department of Emergency Medicine, University of Wisconsin-Madison,
Madison, WI, USA
2
Department of Emergency Medicine & Internal Medicine, Virginia
Commonwealth University Health System, Richmond, VA, USA
3
Division of Emergency Medicine, Carle Foundation Hospital, Urbana, IL, USA
4
Division of Pulmonary, Critical Care and Sleep Medicine, University of
Washington, Seattle, WA, USA
Received January 25, 2023. Received revised April 24, 2023. Accepted
May 22, 2023.
Corresponding Author:
Michael A. Ward, Department of Emergency Medicine, University of
Wisconsin-Madison, 800 University Bay Drive, Suite 300-56, Mail Code 9123,
Madison, WI 53705, USA.
Email: maward@medicine.wisc.edu
2 Journal of Intensive Care Medicine 0(0)

starting at least 12 h from the onset of sepsis with most evalu- Data Abstraction and Accuracy
ating fluid effects days out from initial presentation.10‐14
Computational algorithms identified cases and demographic
In contrast, a collection of studies examining the effect of
data, effectively blinding the selection process to variables of
guideline-based fluids early in sepsis resuscitation (within 3h)
sepsis management and clinical endpoints. Notably, the physi-
observe a survival benefit.15‐21 Taken together, a treatment
cian initially identifying eligible cases was not blinded,
window may exist early in sepsis care by which fluids are ben-
although these cases were prospectively selected for quality
eficial but may wane with care downstream or even cause harm.
improvement purposes. All sepsis-relevant interventions and
The study herein is a subanalysis of our original findings
clinical factors were manually collected. Abstractors (HIK,
observing a survival benefit with administration of 30 mL/kg
JDL, MDH, and JAK) were trained using strict definitions
within 3h of sepsis onset and included a dosing analysis sug-
and protocols using electronic data-abstraction forms and
gesting 30 mL/kg is an important benchmark for resuscita-
were closely monitored by the principal investigator (MAW)
tion.15 However, we did not conduct any further evaluation of
to ensure data accuracy. A randomly selected subset of 35
timing and, thus, the objective of this study is to evaluate the
charts were reviewed, with raw agreement of 95.1% and inter-
impact of timing (from sepsis onset out to 24 h) of 30 mL/kg
rater reliability, κ = 0.93. Abstractors were not blinded to the
crystalloid on mortality and other clinical outcomes. This
purpose of the study.
study is meant to be merely hypothesis generating as an analysis
Fluid volumes were calculated through electronic chart
over this time scale is especially prone to endogeneity or unob-
abstraction (Epic, Verona, WI) using the Medication
served variable bias in that it will be difficult to disentangle the
Administration Records (MAR) and based on documentation
causal relationship between fluid timing and sepsis outcomes.
of completion time. When completion time was not available,
For example, there are factors that may impact fluid administra-
fluid boluses were assumed to be completed at a rate of 1L/h.
tion and factors that may impact outcomes independent of
Fluids were defined as only crystalloids which comprised of
current sepsis interventions that will be challenging to fully
mostly normal saline or lactated ringers, although this was
account for.
not specifically tracked. Fluids administered along with medica-
tion administration (ie, antibiotics) were not included. The
Materials and Methods timing of fluid administration was determined based on when
criteria for severe sepsis or septic shock were met per sepsis
This study is a novel analysis from a previously published study
II definitions. Therefore, as just a couple of examples, onset
cohort and a complete description of methods is available in the
would be time of first documented hypotension or resulted ele-
original report.15 This was a retrospective analysis using a conve-
vated lactate (whichever came first), assuming 2 systemic
nience sample from January 1, 2014 to April 30, 2015 and
inflammatory response syndrome (SIRS) criteria had already
February 1, 2016 to May 31, 2017. Data identifying positive
been met. As noted in our adjusted mortality versus time
sepsis cases were originally collected as part of a quality
plots, this would allow for 30 mL/kg of crystalloids to be
improvement initiative and were not available between these
administered before time 0, accounting for the negative time
dates. The cohort includes adult patients (≥18 years old) present-
points in a few circumstances.
ing to the emergency department (ED) with severe sepsis or
Initial hypotension was defined as systolic blood pressure
septic shock at an urban, tertiary-care center (∼70 000 ED
<90 mm Hg or mean arterial pressure <65 mm Hg within 3h
visits per year). The majority of patients treated were African
of sepsis onset. Delayed hypotension included the same blood
American (∼73% of all visits). The study was approved by the
pressure cutoffs but between 3 and 12 h after sepsis onset.
Institutional Review Board (approval #17-1444). We aimed to
The medical history of renal failure was defined as peritoneal
adhere to the Strengthening the Reporting of Observational
or hemodialysis prior to hospitalization and heart failure as doc-
Studies in Epidemiology (STROBE) guidelines.22
umented prior to hospitalization or echocardiogram report
We evaluated the study population for the presence of
within 1 year of hospitalization showing an ejection fraction
severe sepsis or septic shock within 8h of ED arrival.
of <45%.
Cases were identified by International Classification of
Disease-Revision 9 codes (ICD-9) or ICD-10 codes related
to sepsis and are the same as those used for case selection
by the Centers for Medicare & Medicaid Services (CMS). Statistical Analysis
Next, to capture instances where there was clinical concern A multivariable logistic regression model was used to assess the
for infection, we flagged orders for blood cultures, urine cul- relationship between fluid goal timing and mortality. Variables
tures, or antibiotics within 12 h of ED arrival. A physician were selected a priori. The model was estimated using R
(MAW) reviewed each chart using a standardized process 4.0.3.23 The key model input in this study was time to reaching
to determine whether infection-related organ dysfunction fluid goals, defined as hours between meeting severe sepsis or
occurred within 8h of ED arrival. Each case was reviewed septic shock criteria and administration of 30 mL/kg of fluids.
using a standardized process and was confirmed with treating The model included both a linear and quadratic term for this
providers and a multidisciplinary quality group to improve key input, thus allowing a test of both the linear and nonlinear
quality control. effects of fluid goal timing on mortality, incidence of delayed
Kuttab et al
hypotension, need for intubation, need for vasopressors, and
need for ICU admission. The initial models included the full
24 h time range, but wide variance was noted approximately
after the 12 h time mark. Therefore, in a post hoc fashion, we
had decided to further examine these relationships within a
12 h time window. Covariables were included in the model to
control for background characteristics and other features with
the potential to confound key study relationships.
Specifically, the model adjusted for sex, obesity, end-stage
renal disease, congestive heart failure, initial lactate value,
number of SIRS criteria present, presence of early hypotension
(before 3h of sepsis onset), mortality in emergency department
sepsis (MEDS) score, and time to antibiotics. Model covariates
were chosen a priori. Statistical significance was defined as P
value of <.05. Time to 30 mL/kg of fluid that was earlier than
−2 h were excluded from analysis to better assess the impact
of acute fluid administration around the time of sepsis onset.
Lastly, we excluded cases from the analyses where 30 mL/kg
of fluid volume was not reached in the time range examined.
In addition to measuring and analyzing fluid timing as a con-
tinuous variable, this study also investigated the treatment of
time as a discrete variable. In some cases, such treatment of
timing can generate insights by allowing for better expression
of nonlinear relationships. To this end, study participants
were separated into 5 groups based on the time interval in
which their fluid goals were reached (<1 h, 1 to ≤ 3 h, 3 to ≤
6 h, 6 to ≤ 24 h, and > 24 h). This discrete group/timing variable
was then substituted for the continuous time variable, and
models were re-estimated for each of the clinical outcomes.
The coefficients in each model were converted to odds ratios
allowing for a comparison of risk among the 5 fluid groups (ref-
erence category = “<1 h”).
Results
Demographics, Sepsis Measures, and Outcomes
Table 1 describes the study cohort, categorized by the time in
which the 30 mL/kg volume was reached. Of note, 174 (16.9%)
of patients achieved 30 mL/kg in <1 h, 332 (32.2%) achieved
from 1 to ≤ 3 h, 167 (16.2%) achieved from 3 to ≤ 6 h, 150
(14.5%) achieved from 6 to ≤ 24 h, and 209 (20.3%) achieved
this > 24 h or never at all. The group where 30 mL/kg was
achieved in <1 h included 56 instances (5.4% of total cases)
where 30 mL/kg was reached before sepsis onset. Median
volumes of intravenous (IV) fluids given at discrete time
points—1, 3, 6, and 24 h—are outlined. Overall mortality in
the study cohort was 17.1%, with 12.7% mortality in patients
with severe sepsis and 20.4% in patients with septic shock.
There were 22 deaths within the first 24 h and 3 cases (0.3%
of total cases) where patient death occurred within 24 h and
30 mL/kg was not reached.
Supplemental S1 compares the study cohort based on date of
presentation. In summary, the latter group (2016, 2017) had less
severe disease compared to the earlier group: MEDS score 10.2
± 4.6 versus 11.4 ± 4.6; lactate 3.3 ± 2.4 versus 3.8 ± 3.1 mmol/
3
L; septic shock 60.6% versus 68.7%; mortality 14.6% versus
22.1%. No major differences were noted in management vari-
ables, including median fluid volumes at all measured time
points and time to 30 mL/kg.
Evaluation of Continuous Time Variable in Meeting the
30 mL/kg Fluid Goal for Mortality
Supplemental Figure 1 describes the multivariable logistic
regression model, including a quadratic term, for the timing
of 30 mL/kg as a continuous variable, between 0 and 24 h
(Figure 1A) and 0 and 12 h (Figure 1B) of severe sepsis/
septic shock onset, in relation to in-hospital mortality. The
timing of 30 mL/kg from 0 to 24 h of severe sepsis/septic
shock onset in relation to mortality did not demonstrate signifi-
cance. For the 0–12 hour time window, odds of mortality is
associated with an increase of 1.29 for each 1h delay from
severe sepsis/septic shock onset (95% confidence interval [CI]
1.02-1.67, P = .04). However, this effect appears to peak at
about the 5 to 6 h mark; the quadratic term is otherwise not
significant (odds ratio [OR] 0.98, 95% CI 0.95-1.00, P = .09).
The complete output of associations between all covariates
and mortality are listed in the Supplemental section (S.2).
A separate multivariable logistic regression model evaluat-
ing the impact of 30 mL/kg timing on in-hospital mortality
over a 12 h range included septic shock as an additional covar-
iate with minimal change noted (OR 1.28, 95% CI 1.00-1.65, P
= .051 for the linear term; OR 0.98, 95% CI 0.95-1.01, P = .09).
These results are listed in Supplemental section S3 and include
analyses over this 12 h range for both the severe sepsis and
septic shock groups separately.
Evaluation of Continuous Time Variable in Meeting the
30 mL/kg Fluid Goal for Other Outcomes
Supplemental Figure 2 describes the multivariable logistic
regression model, including a quadratic term, for the timing
of 30 mL/kg as a continuous variable, between 0 and 12 h of
severe sepsis/septic shock onset, in relation to delayed hypoten-
sion and need for vasopressors, intubation, and ICU admission.
Of note, timing of 30 mL/kg in these models demonstrated no
significance for any of these clinical variables (Figure 2).
Evaluation of Discrete Time Points to Meet the 30 mL/kg
Fluid Goal
Table 2 describes a multivariable regression analysis compar-
ing patients receiving 30 mL/kg within 1 h to those receiving
this fluid volume at discrete time points. No differences were
observed when receiving this volume between 1 and 3 h (OR
1.11, 95% CI 0.62-2.01), 3 and 6 h (OR 1.83, 95% CI
0.97-3.52), or 6 and 24 h (OR 1.51, 95% CI 0.75-3.06).
However, not receiving the recommended 30 mL/kg volume
by 24 h was associated with increased odds of in-hospital mor-
tality (OR 2.69, 95% CI 1.37-5.37, P = .004).
Table 1. Demographics, Sepsis Measures, and Outcomes.

Total ≤1 h 1≤3 h 3≤6h 6 ≤ 24 h Not by 24 h*

Total patients 1032 174 (16.9%) 332 (32.2%) 167 (16.2%) 150 (14.5%) 209 (20.3%)

Demographics and baseline characteristics

Age (years), mean ± SD 63.4 ± 17.0 60.0 ± 17.5 64.1 ± 17.6 66.3 ± 18.5 64.1 ± 13.9 62.1 ± 16.0
Male sex, n(%) 510 (49.4) 76 (43.7) 158 (47.6) 90 (53.9) 75 (50.0) 111 (53.1)
BMI (kg/m2), mean ± SD 27.2 ± 8.8 23.4 ± 6.6 24.1 ± 5.9 27.3 ± 7.6 30.8 ± 9.7 32.8 ± 10.6
MEDS score, mean ± SD 10.6 ± 4.7 10.4 ± 4.6 10.9 ± 4.5 11.3 ± 4.9 10.4 ± 4.9 9.7 ± 4.6
Time to SS/SS diagnosis (h), mean ± SD 1.7 ± 1.5 2.1 ± 1.6 1.5 ± 1.5 1.4 ± 1.5 1.7 ± 1.4 1.8 ± 1.6
ESRD, n(%) 101 (9.8) 8 (4.6) 11 (3.3) 12 (7.2) 17 (11.3) 53 (25.4)
HF, n(%) 245 (23.7) 18 (10.3) 49 (14.8) 36 (21.6) 40 (26.7) 102 (48.8)
Hospital metrics
ED LOS (h), mean ± SD 9.4 ± 4.7 9.0 ± 4.6 9.2 ± 4.3 10.2 ± 5.3 9.7 ± 5.1 9.3 ± 4.6
ICU admission, n(%) 606 (58.7) 101 (58.0) 211 (63.6) 98 (58.7) 90 (60.0) 106 (50.7)
ICU LOS (days), mean ± SD 2.6 ± 4.3 2.0 ± 3.3 2.6 ± 4.1 2.7 ± 4.4 3.0 ± 4.6 2.9 ± 5.0
Hospital LOS (days), mean ± SD 9.9 ± 9.4 8.8 ± 7.6 9.5 ± 8.0 9.4 ± 11.4 10.8 ± 10.9 11.4 ± 9.4
Sepsis measures
≥ 2 SIRS criteria, n(%) 880 (85.3) 152 (87.4) 290 (87.3) 150 (89.8) 123 (82.0) 165 (78.9)
≥ 2 qSOFA criteria, n(%) 345 (33.4) 61 (35.1) 126 (38.0) 63 (37.7) 41 (27.3) 54 (25.8)
Lactate mmol/L, mean ± SD 3.5 ± 2.7 3.6 ± 2.5 3.8 ± 2.9 3.5 ± 2.7 3.0 ± 2.0 3.1 ± 2.6
Lactate >2.0 mmol/L, n(%) 766 (74.2) 138 (79.3) 260 (78.3) 126 (75.4) 102 (68.0) 140 (67.0)
Lactate >4.0 mmol/L, n(%) 275 (26.6) 57 (32.8) 106 (31.9) 43 (25.7) 26 (17.3) 43 (20.6)
Time to antibiotics (h), mean ± SD 1.3 ± 3.3 0.1 ± 1.9 1.1 ± 3.9 1.8 ± 3.1 1.9 ± 2.9 1.7 ± 3.2
Initial hypotension, n (%) 590 (57.2) 83 (47.7) 201 (60.5) 105 (62.9) 87 (58.0) 114 (54.5)
Volume IVF at 1 h (L), median (IQR)+ 1.0 (0.0-1.7) 2.0 (1.7-3.0) 1.0 (0.7-1.6) 0.6 (0.0-1.0) 0.5 (0.0-1.0) 0.0 (0.0-1.0)
Volume IVF at 3 h (L), median (IQR)+ 2.0 (1.0-3.0) 3.0 (2.0-4.0) 2.8 (2.0-3.2) 1.9 (1.0-2.0) 1.0 (1.0-2.0) 0.6 (0.0-1.0)
Volume IVF at 6 h (L), median (IQR)+ 2.7 (2.0-4.0) 3.5 (3.0-4.6) 3.4 (2.6-4.0) 3.0 (2.5-3.5) 2.0 (1.0-2.0) 1.0 (0.3-1.7)
Volume IVF at 24 h (L), median (IQR)+ 4.0 (2.0-6.0) 5.0 (4.0-7.0) 5.0 (4.0-6.7) 4.5 (3.7-6.0) 4.0 (3.0-5.0) 1.5 (0.5-2.0)
Steroids, n(%) 65 (6.3) 15 (8.6) 6 (1.8) 9 (5.4) 9 (6.0) 14 (6.7)
Central line placed, n(%) 323 (31.3) 42 (24.1) 111 (33.4) 57 (34.1) 52 (34.7) 61 (29.2)
Septic shock, n(%) 653 (63.3) 116 (66.7) 240 (72.3) 103 (61.7) 83 (55.3) 111 (53.1)
Sepsis source, n (%)
Pneumonia 313 (30.3) 49 (28.2) 113 (34.0) 48 (28.7) 46 (30.7) 57 (27.3)
Genitourinary 248 (24.0) 50 (28.7) 80 (24.1) 46 (27.5) 29 (19.3) 43 (20.6)
Abdominal 151 (14.6) 19 (10.9) 56 (16.9) 26 (15.6) 27 (18.0) 23 (11.0)
Skin/soft tissue 143 (13.9) 19 (10.9) 34 (10.2) 20 (12.0) 25 (16.7) 45 (21.5)
Indwelling IV line 53 (5.1) 8 (4.6) 14 (4.2) 6 (3.6) 6 (4.0) 19 (9.1)
Other/unknown 124 (12.0) 29 (16.7) 35 (10.5) 21 (12.6) 17 (11.3) 22 (10.5)
Outcomes
Delayed hypotension, n(%) 668 (64.7) 94 (54.0) 228 (68.7) 111 (66.5) 103 (68.7) 132 (63.2)

(continued)
Kuttab et al 5

Supplemental S4 lists the remaining multivariable regression

range; IVF, intravenous fluids; LOS, length of stay; MEDS, mortality in emergency department sepsis; qSOFA, quick sequential organ failure assessment (defined in Supplemental 1b); SIRS, systemic inflammatory
Not by 24 h*

Total volume IVF measured up to 24 h from SS/SS onset (only fluid boluses were considered between 0 and 6 h. Fluid boluses and maintenance fluids were considered between 6 and 24 h. Volumes were not
The table describes the study cohort, categorized by time in which the 30 mL/kg fluid volume was achieved. HF defined as echocardiogram with ejection fraction <45% within 1 year or if documented in the medical

Abbreviations: BMI, body mass index; ED, emergency department; ESRD, end-stage renal disease (on peritoneal or hemodialysis prior to hospitalization); HF, heart failure; ICU, intensive care unit; IQR, interquartile
209 (20.3%)
analyses comparing 30 mL/kg within 1 h to those receiving this

28 (13.4)
40 (19.1)
56 (26.8)
44 (21.1)
26 (26.5)
32 (28.8)
volume at other time points for the secondary clinical outcomes.
Receiving 30 mL/kg <1 h was observed to decrease the inci-
dence of delayed hypotension (OR 1.83, 95% CI 1.23-2.72)

*Twenty cases were found where a time for completion of 30 mL/kg of >24 h was documented. Therefore, 189 cases remain where 30 mL/kg is categorized as “not given” for these analyses.
when compared to the 1 to 3 h group and increase the need
for ICU admission (OR 0.58, 95% CI 0.34-0.96) when com-
pared to the >24 h group. No impact was observed for group
150 (14.5%)

comparisons when the need for intubation or vasopressors

31 (20.7)
27 (18.0)
45 (30.0)
23 (15.3)

16 (19.3)
6 ≤ 24 h

7 (10.4)
was examined.

Discussion
A 12 h plot evaluating adjusted mortality over time suggest that
earlier administration of 30 mL/kg of crystalloids are beneficial,
167 (16.2%)

although there is a plateau effect around 5 h that does not reach

23 (13.8)
40 (24.0)
60 (35.9)
36 (21.6)

28 (27.2)
8 (12.5)
3≤6h

significance. A group comparison using patients receiving

30 mL/kg within 1h as a reference also suggests a similar
trend but does not reach significance until the <1 h group is
compared to the group not receiving 30 mL/kg within 24 h of
sepsis onset. Taken together, evidence weakly suggests that
there are benefits of early fluid administration.
332 (32.2%)

113 (34.0)
50 (15.1)
76 (22.9)

51 (15.4)

39 (16.3)

Our initial plot examining adjusted mortality versus time

1≤3 h

3 (3.3)

spanned over 24 h and did not reach significance but appeared

to have wide-ranging error beyond 12 h, likely related to limited
n-size given the larger time ranges (Supplemental Figure 1A).
Additionally, it is possible that fluid measurements are less reli-
able at later time points when the care of the patient transitions
response syndrome (defined in Supplemental 1b); SD, standard deviation; SS/SS, severe sepsis/septic shock.
174 (16.9%)

from the emergency department period to the inpatient team.

26 (14.9)
32 (18.4)
51 (29.3)
22 (12.6)

18 (15.5)

The average ED length of stay in this study cohort was mea-

4 (6.9)

sured at 9.4 h, which is a measurement of total time in the

≤1 h

ED prior to the patient departing for their inpatient bed. Boyd

et al10 had showed a mortality risk for patients with more pos-
itive fluid balances days out from sepsis onset but a survival
benefit at 12 h when central venous pressure (CVP) <8 mm
Hg. Thus, we decided to perform a post hoc analysis of adjusted
158 (15.3)
215 (20.8)
325 (31.5)
176 (17.1)

133 (20.4)
48 (12.7)

mortality versus time over 12 h and found a linear mortality risk

(OR 1.29, 95% CI 1.02-1.67) per hour of 30 mL/kg treatment
1032
Total

delay. The trend line peaks at approximately the 5h mark;

however, the quadratic function does not reach significance
(P = .09). Taken together, this may support the theory of a treat-
chart (including those with preserved ejection fraction).

ment window by which fluids offer benefit early in resuscitation

but may wane over time. Evaluating the discrete time points
demonstrates an increased odds of in-hospital mortality if the
Severe sepsis (% of severe group)

30 mL/kg volume was not reached within 24 h (OR 2.69,

95% CI 1.37-5.37, P = .004), but not at the other discrete
time points. This may be due to underpowering due to small
Positive blood cultures, n(%)

Shock (% of shock group)

Mechanical ventilation, n(%)

n-sizes.
There is a collection of observational studies that observe a
survival benefit with early fluid resuscitation within 3h of sepsis
Table 1. (continued)

Vasopressor, n(%)

onset.15‐21 However, a set of studies evaluating fluid adminis-

tration within 6h of sepsis onset show more mixed results
collected >24 h).
Total patients

with about half showing survival benefit and half showing no

Mortality

mortality impact.24‐34 Notably, Lee et al had showed improved

survival with fluids given over 20 mL/kg over the first 3 h and
no mortality impact when measured at 6h.35 Therefore, we
+
6 Journal of Intensive Care Medicine 0(0)

Figure 1. Timing of 30 mL/mg (as a continuous variable) from 0 to 24 h (A) and 0 to 12 h (B) of sepsis onset versus mortality. (A) Time to
30 mL/kg versus mortality (0-24 h). (B) Time to 30 mL/kg versus mortality (0-12 h). The figure estimates from the multivariable logistic
regression models, including a quadratic term, for timing of IV fluids as a continuous variable (0-24 h [A] and 0-12 h [B] of severe sepsis/septic
shock onset) in relation to in-hospital mortality. Gray shaded area indicates 95% confidence interval. (A) Ranging over 24 h does not reach
significance. This analysis included 812 patient cases; 1032 total cases excluding 11 cases that reached 30 mL/kg earlier than before 2 h of sepsis
onset and excluding 209 cases did not reach 30 ml/kg within 24 h. (B) Over 12 h observes an increased odds of mortality by 1.29 for each 1h
increment in the time to 30 mL/kg. The linear portion of the curve is significant (P = .04), while the quadratic term is not (P = .09). This analysis
included 758 patient cases as an additional 54 cases reached 30 mL/kg between 12 and 24 h and thus were excluded. Abbreviation: IV,
intravenous.

theorized that an important effect may exist between 3 and 6 h
as evidenced by our results.
Per the above adjusted mortality versus time analyses, we
aimed to determine if group differences could be found using
the group reaching 30 mL/kg by 1h as a reference. It should
be noted that the original analysis from this subanalyzed
cohort had observed a survival benefit when reaching 30 mL/
kg within 3h of sepsis onset (OR 0.60, 95% CI 0.40-0.88).
We theorized, based on the above literature referenced, that a
difference may be found when fluid was given between 3 to
6 h compared to 1h, however, this did not quite reach signifi-
cance for mortality risk (OR 1.83, 95% CI 0.97-3.52).
Further, the odds for mortality relatively decreased between 6
and 24 h compared to the first hour (OR 1.51, 95% CI
0.75-3.06), although this did not reach significance. Next, as
previously noted, significance was not reached until the refer-
ence group was compared to the group not receiving 30 mL/
kg within 24-h. Taken together, group comparisons did show
some interesting trends suggesting earlier fluids reaching
30 mL/kg are beneficial, but clear or more convincing patterns
could not be established.
We had observed little impact for the timing of fluids in the
evaluation of secondary clinical outcomes, including need for
intubation, vasopressors, ICU admission, and delayed onset
of hypotension. We had found in our original analysis that
reaching 30 mL/kg within the first 3h of sepsis onset to
reduce the incidence of later hypotension.15 Leisman et al36
had previously observed that a delay in fluid initiation >2h
approximately doubled the odds of refractory hypotension.
This current analysis had a similar finding with greater odds
of delayed hypotension when 30 mL/kg was given between 1
and 3 h and <1 h (OR 1.83, 95% CI 1.23-2.72). However, com-
parisons with patient groups at later time points did not show
any discernible pattern and the adjusted delayed hypotension
versus time plot did not reach significance. It is difficult to
determine whether the later time points represent the passage
of a treatment window or whether there are other factors not con-
trolled for or well understood that may impact who may be more
likely to receive fluids and who may be more likely to experience
clinical outcomes. This may highlight the challenge in attempting
to examine these relationships observationally.
Establishing clear standards for a treatment window and
dosing of fluids in sepsis poses a major challenge and, more
likely, the medical community will be relegated to piecing
together a patchwork of clinical trends and physiological
theory. First, sepsis remains a highly heterogeneous disease
and the timing of patient presentations will vary. For
example, we establish time 0 as the moment a low blood pres-
sure is documented or a lab results, however, the onset of
serious infection and initial stages of organ dysfunction may
precede these measurements by hours and even days.
Secondly, the dosing of fluids may also significantly change
Kuttab et al 7

Figure 2. Timing of 30 ml/mg (as a continuous variable) from 0 to 12 h of SS/SS onset versus various clinical outcomes. The figure estimates
from the multivariable logistic regression models for timing of IV fluids as a continuous variable (0-12 h of onset) in relation to other various
outcomes. Gray shaded area indicates 95% confidence interval. No significance was observed between timing of IV fluids and these outcomes.
These analyses each included 758 cases. Abbreviations: IV, intravenous; SS/SS, severe sepsis/septic shock.

the effect of timing. We had previously observed that patients
receiving little to no fluids are at the highest risk of mortality,
thus, there may be considerable differences comparing the
fluid administration times at volumes of 0 versus 45 and 25
versus 32 mL/kg.15 A collection of studies specifically evalu-
ated time to fluid initiation to control for some of the dosing
issues and have observed a survival benefit with earlier fluid
administration ranging from prehospital to 120 min out from
sepsis onset.18,37‐40 Ideally, a practical and reliable measure-
ment of volume status and fluid responsiveness will be devel-
oped that will remove the need for dosing guidelines and
guesswork. Strides have been made in this regard; however, a
standard has been sought for many years and a true consensus
has yet to be reached.
8 Journal of Intensive Care Medicine 0(0)

Table 2. Multivariable Regression Analysis for Discrete Time Points. clear pattern was unable to be distinguished. Thus, further
observational studies will be required as the question of
Variable OR 95% CI P
timing may be difficult to examine prospectively.
30 mL/kg by 1-3 h 1.11 0.62-2.01 .74
30 mL/kg by 3-6 h 1.83 0.97-3.52 .06 Acknowledgment
30 mL/kg by 6-24 h 1.51 0.75-3.06 .25
30 mL/kg not by 24 h 2.69 1.37-5.37 .004 We would like to thank the passionate work of David P. Liedke in the
MEDS score 1.16 1.12-1.21 <.001 support of this and other work.
Time to antibiotics (h) 1.02 0.97-1.07 .34
BMI >30 kg/m2 0.66 0.42-1.04 .08 Declaration of Conflicting Interests
Male sex 0.60 0.42-0.86 .01
The author(s) declared no potential conflicts of interest with respect to
ESRD 1.15 0.62-2.06 .64
the research, authorship, and/or publication of this article.
HF 1.25 0.82-1.89 .28
Lactic acid (mmol/L) 1.17 1.11-1.24 <.001
# of SIRS criteria met 1.17 0.96-1.42 .13 Funding
Initial hypotension (0-3h) 1.37 0.94-2.01 .11 The author(s) disclosed receipt of the following financial support for
The table describes a multivariable regression analysis for in-hospital mortality, the research, authorship, and/or publication of this article: This work
using discrete time points for fluid goals (ie, 3 h, 6 h, 24 h, or not reached by was supported by the National Center for Advancing Translational
24 h) and compared to patients receiving 30 mL/kg within 1 h. Patients not Sciences (grant number 5UL1TR002389-02).
receiving 30 mL/kg by 24 h of severe sepsis/septic shock onset (or never), had
an increased odds of in-hospital morality (OR 2.69, 95% CI 1.37-5.37, P .004).
Abbreviations: BMI, body mass index; CI, confidence interval; ESRD, end-stage ORCID iD
renal disease; HF, heart failure; MEDS, mortality in emergency department Michael A. Ward https://orcid.org/0000-0002-5760-5890
sepsis; OR, odds ratio; SIRS, systemic inflammatory response syndrome.

Many limitations exist beyond the timing and dosing
issues noted above that will impact the interpretations of
this study. First, this study is single center and results may
not be generalizable. Further, this is a retrospective analysis
and is prone to endogeneity. Therefore, full control for con-
founding is not possible and we believe may be evidenced by
lack of true pattern with respect to comparison of fluid
administration over different time points, especially when
looking at secondary clinical outcomes. This point may be
further evidenced by the contradictory observation of
increased need for ICU admission with the rapid initiation
of fluids. Next, given we had attempted to compare differ-
ences across a longer time frame, we have exacerbated the
issue of limited sample size. Also, study abstractors were
not blinded to the purpose of this study which may put
these results at higher risk of bias. Further, there were
instances where fluid administration completion times were
not fully documented and, thus, a 1h/1L administration
time was approximated based on a common nursing practice
of infusing fluids on a pump at maximum rate. This was not a
common occurrence; however, it should be noted that the fre-
quency of this occurrence was not tabulated. Lastly, prospec-
tive trials are best to determine causation, however, the
question of timing may be more difficult to assess and obser-
vational data, such as this study, may be required.
Conclusion
We have observed weak evidence that supports that earlier is
likely better for reaching 30 mL/kg crystalloid administration
in adult patient with severe sepsis or septic shock. However,
our findings did not appear to be consistently robust, and a
Supplemental Material
Supplemental material for this article is available online.
References
1. Rivers E, Nguyen B, Havstad S, et al. Early goal-directed therapy
in the treatment of severe sepsis and septic shock. N Engl J Med.
2001;345(19):1368‐1377. doi:10.1056/NEJMoa010307
2. ARISE Investigators and the ANZICS Clinical Trials Group.
Goal-directed resuscitation for patients with early septic shock.
N Engl J Med. 2014;371(16):1496‐1506.
3. ProCESS Investigators. A randomized trial of protocol-based care
for early septic shock. N Engl J Med. 2014;370(18):1683‐1693.
4. Mouncey PR, Osborn TM, Power GS, et al. Trial of early, goal-
directed resuscitation for septic shock. N Engl J Med.
2015;372(14):1301‐1311. doi:10.1056/NEJMoa1500896
5. Taylor SP, Karvetski CH, Templin MA, Heffner AC, Taylor BT.
Initial fluid resuscitation following adjusted body weight dosing is
associated with improved mortality in obese patients with sus-
pected septic shock. J Crit Care. 2018;43:7‐12. doi:10.1016/j.
jcrc.2017.08.025
6. Marik PE, Malbrain M. The SEP-1 quality mandate may be
harmful: how to drown a patient with 30 mL per kg fluid!
Anaesthesiol Intensive Ther. 2017;49(5):323‐328. doi:10.5603/
AIT.a2017.0056
7. Levy MM, Evans LE, Rhodes A. The surviving sepsis campaign
bundle: 2018 update. Crit Care Med. 2018;46(6):997‐1000.
doi:10.1097/CCM.0000000000003119
8. Evans L, Rhodes A, Alhazzani W, et al. Surviving sepsis cam-
paign: International guidelines for management of sepsis and
septic shock 2021. Crit Care Med. 2021;4:e1063‐e1143. doi:10.
1097/CCM.0000000000005337
9. Yealy DM, Mohr NM, Shapiro NI, Venkatesh A, Jones AE, Self
WH. Early care of adults with suspected sepsis in the emergency
department and out-of-hospital environment: a consensus-based
task force report. Ann Emerg Med. 2021;78(1):1‐19. doi:10.
1016/j.annemergmed.2021.02.006
Kuttab et al
10. Boyd JH, Forbes J, Nakada TA, Walley KR, Russell JA. Fluid
resuscitation in septic shock: a positive fluid balance and elevated
central venous pressure are associated with increased mortality.
Crit Care Med. 2011;39(2):259‐265. doi:10.1097/CCM.0b013
e3181feeb15
11. Marik PE, Linde-Zwirble WT, Bittner EA, Sahatjian J, Hansell D.
Fluid administration in severe sepsis and septic shock, patterns
and outcomes: an analysis of a large national database. Intensive
Care Med. 2017;43(5):625‐632. doi:10.1007/s00134-016-4675-y
12. Micek ST, McEvoy C, McKenzie M, Hampton N, Doherty JA,
Kollef MH. Fluid balance and cardiac function in septic shock
as predictors of hospital mortality. Crit Care. 2013;17(5):R246.
doi:10.1186/cc13072
13. Acheampong A, Vincent JL. A positive fluid balance is an inde-
pendent prognostic factor in patients with sepsis. Crit Care.
2015;19:251. doi:10.1186/s13054-015-0970-1
14. Sakr Y, Rubatto Birri PN, Kotfis K, et al. Higher fluid balance
increases the risk of death from sepsis: results from a large inter-
national audit∗. Crit Care Med. 2017;45(3):386‐394. doi:10.1097/
CCM.0000000000002189
15. Kuttab HI, Lykins JD, Hughes MD, et al. Evaluation and predic-
tors of fluid resuscitation in patients with severe sepsis and septic
shock. Crit Care Med. 2019;47(11):1582‐1590. doi:10.1097/
CCM.0000000000003960
16. Hu B, Chen JCY, Dong Y, et al. Effect of initial infusion rates of
fluid resuscitation on outcomes in patients with septic shock: a his-
torical cohort study. Crit Care. 2020;24(1):137. doi:10.1186/
s13054-020-2819-5
17. Liu VX, Morehouse JW, Marelich GP, et al. Multicenter imple-
mentation of a treatment bundle for patients with sepsis and inter-
mediate lactate values. Am J Respir Crit Care Med.
2016;193(11):1264‐1270. doi:10.1164/rccm.201507-1489OC
18. Leisman D, Wie B, Doerfler M, et al. Association of fluid resusci-
tation initiation within 30 minutes of severe sepsis and septic
shock recognition with reduced mortality and length of stay.
Ann Emerg Med. 2016;68(3):298‐311. doi:10.1016/j.annemergmed.
2016.02.044
19. Boccio E, Haimovich A, Jacob V, Maciejewski KR, Wira CR,
Belsky J. Sepsis fluid metric compliance and its impact on out-
comes of patients with congestive heart failure, end-stage renal
disease or obesity. J Emerg Med. 2021;61(5):466‐480. doi:10.
1016/j.jemermed.2021.03.004
20. Acharya R, Patel A, Schultz E, et al. Fluid resuscitation and out-
comes in heart failure patients with severe sepsis or septic shock: a
retrospective case-control study. PLoS One. 2021;16(8):
e0256368. doi:10.1371/journal.pone.0256368
21. Jeong WI, Kim J-S, Yu J, et al. The prognostic value of 1-hour
bundle completion in septic shock patients. J Korean Soc Emerg
Med. 2019;30(6):537‐544.
22. von Elm E, Altman DG, Egger M, et al. The Strengthening the
Reporting of Observational Studies in Epidemiology (STROBE)
statement: guidelines for reporting observational studies. Lancet.
2007;370(9596):1453‐1457. doi:10.1016/S0140-6736(07)61602-X
23. R: A language and environment for statistical computing. R Core
Team; 2021. https://www.R-project.org/
24. Truong TTN, Dunn AS, McCardle K, et al. Adherence to fluid
resuscitation guidelines and outcomes in patients with septic
shock: reassessing the “one-size-fits-all” approach. J Crit Care.
2019;51:94‐98. doi:10.1016/j.jcrc.2019.02.006
25. Khan RA, Khan NA, Bauer SR, et al. Association between volume
of fluid resuscitation and intubation in high-risk patients with
9
sepsis, heart failure, end-stage renal disease, and cirrhosis.
Chest. 2020;157(2):286‐292. doi:10.1016/j.chest.2019.09.029
26. Rajdev K, Leifer L, Sandhu G, et al. Fluid resuscitation in patients
with end-stage renal disease on hemodialysis presenting with
severe sepsis or septic shock: a case control study. J Crit Care.
2020;55:157‐162. doi:10.1016/j.jcrc.2019.10.008
27. Miller RR III, Dong L, Nelson NC, et al. Multicenter implemen-
tation of a severe sepsis and septic shock treatment bundle. Am
J Respir Crit Care Med. 2013;188(1):77‐82.
28. Ferrer R, Artigas A, Suarez D, et al. Effectiveness of treatments for
severe sepsis: a prospective, multicenter, observational study. Am
J Respir Crit Care Med. 2009;180(9):861‐866. doi:10.1164/rccm.
200812-1912OC
29. Seymour CW, Gesten F, Prescott HC, et al. Time to treatment and
mortality during mandated emergency care for sepsis. N Engl J
Med. 2017;376(23):2235‐2244. doi:10.1056/NEJMoa1703058
30. Waechter J, Kumar A, Lapinsky SE, et al. Interaction between
fluids and vasoactive agents on mortality in septic shock: a multi-
center, observational study. Crit Care Med. 2014;42(10):2158‐
2168. doi:10.1097/ccm.0000000000000520
31. Noritomi DT, Ranzani OT, Monteiro MB, et al. Implementation of
a multifaceted sepsis education program in an emerging country
setting: clinical outcomes and cost-effectiveness in a long-term
follow-up study. Intensive Care Med. 2014;40(2):182‐191.
doi:10.1007/s00134-013-3131-5
32. Scheer CS, Fuchs C, Kuhn SO, et al. Quality improvement initia-
tive for severe sepsis and septic shock reduces 90-day mortality: a
7.5-year observational study. Crit Care Med. 2017;45(2):241‐252.
doi:10.1097/CCM.0000000000002069
33. Murphy CV, Schramm GE, Doherty JA, et al. The importance of
fluid management in acute lung injury secondary to septic shock.
Chest. 2009;136(1):102‐109. doi:10.1378/chest.08-2706
34. Levy MM, Rhodes A, Phillips GS, et al. Surviving sepsis cam-
paign: association between performance metrics and outcomes
in a 7.5-year study. Crit Care Med. 2015;43(1):3‐12. doi:10.
1097/CCM.0000000000000723
35. Lee SJ, Ramar K, Park JG, Gajic O, Li G, Kashyap R. Increased
fluid administration in the first three hours of sepsis resuscitation is
associated with reduced mortality: a retrospective cohort study.
Chest. 2014;146(4):908‐915. doi:10.1378/chest.13-2702
36. Leisman DE, Doerfler ME, Schneider SM, Masick KD, D’Amore
JA, D’Angelo JK. Predictors, prevalence, and outcomes of early
crystalloid responsiveness among initially hypotensive patients
with sepsis and septic shock. Crit Care Med. 2018;46(2):189‐
198. doi:10.1097/CCM.0000000000002834
37. Lane DJ, Wunsch H, Saskin R, et al. Association between early
intravenous fluids provided by paramedics and subsequent
in-hospital mortality among patients with sepsis. JAMA Network
Open. 2018;1(8):e185845. doi:10.1001/jamanetworkopen.2018.
5845
38. Seymour CW, Band RA, Cooke CR, et al. Out-of-hospital charac-
teristics and care of patients with severe sepsis: a cohort study. J
Crit Care. 2010;25(4):553‐562. doi:10.1016/j.jcrc.2010.02.010
39. Leisman DE, Goldman C, Doerfler ME, et al. Patterns and out-
comes associated with timeliness of initial crystalloid resuscitation
in a prospective sepsis and septic shock cohort. Crit Care Med.
2017;45(10):1596‐1606. doi:10.1097/CCM.0000000000002574
40. Pruinelli L, Westra BL, Yadav P, et al. Delay within the 3-hour
surviving sepsis campaign guideline on mortality for patients
with severe sepsis and septic shock. Crit Care Med.
2018;46(4):500‐505. doi:10.1097/ccm.0000000000002949