YAJEM-56650; No of Pages 6

American Journal of Emergency Medicine xxx (2017) xxxxxx

Contents lists available at ScienceDirect

American Journal of Emergency Medicine

journal homepage: www.elsevier.com/locate/ajem

Emergency department septic shock patient mortality with refractory hypotension vs

hyperlactatemia: A retrospective cohort study,,
Michael D. April, MD, DPhil, MSc a,, Chase Donaldson, MD b, Lloyd I. Tannenbaum, MD a, Tyler Moore, MD a,
Jose Aguirre, MD a, Alexander Pingree, MD a, James H. Lantry, MD a
a
Department of Emergency Medicine, San Antonio Uniformed Services Health Education Consortium, Fort Sam Houston, TX, United States
b
Department of Emergency Medicine, William Beaumont Army Medical Center, Fort Bliss, TX, United States

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

Article history: Background: Our objective was to compare in-hospital mortality among emergency department (ED) patients
Received 26 January 2017 meeting trial-based criteria for septic shock based upon whether presenting with refractory hypotension (systol-
Received in revised form 16 April 2017 ic blood pressure b 90 mm Hg after 1 L intravenous uid bolus) versus hyperlactatemia (initial lactate
Accepted 26 April 2017 4 mmol/L).
Available online xxxx
Methods: We conducted a retrospective cohort analysis by chart review of ED patients admitted to an intensive
care unit with suspected infection during 1 August 201228 February 2015. We included all patients with
Keywords:
Sepsis
body uid cultures sampled either during their ED stay without antibiotic administration or within 24 h of anti-
Mortality biotic administration in the ED. We excluded patients not meeting criteria for either refractory hypotension or
Emergency department hyperlactatemia. Trained chart abstractors blinded to the study hypothesis double entered data from each
Critical care patient's record including demographics, clinical data, treatments, and in-hospital mortality. We compared in-
Intravenous uid hospital mortality among patients with isolated refractory hypotension, isolated hyperlactatemia, or both. We
also calculated odds ratios (ORs) via logistic regression for in-hospital mortality based on presence of refractory
hypotension or hyperlactatemia.
Results: Of 202 patients included in the analysis, 38 (18.8%) died during hospitalization. Mortality was 10.9%
among 101 patients with isolated refractory hypotension, 24.4% among 41 patients with isolated
hyperlactatemia, and 28.3% among 60 patients with both (p = 0.01). Logistic regression analyses yielded in-hos-
pital mortality OR for refractory hypotension of 1.3 (95% CI 0.53.8) versus OR for hyperlactatemia of 2.9 (95% CI
1.27.4).
Conclusions: Hyperlactatemia appears associated with higher in-hospital mortality compared to refractory hypo-
tension among ED patients with septic shock.
Published by Elsevier Inc.

1. Introduction
1.1. Background
Sepsis is a clinical syndrome resulting from a systemic host response
to infection [1,2]. A spectrum of severity exists in which patients with
hypo-perfusion or refractory hypotension experience higher mortality
rates [3]. These patients require early recognition, careful resuscitation,
Disclaimers: The view(s) expressed herein are those of the author(s) and do not
reect the ofcial policy or position of Brooke Army Medical Center, the U.S. Army
Medical Department, the U.S. Army Ofce of the Surgeon General, the Department of the
Army and Department of Defense or the U.S. Government.
Funding: None
Conicts of interest: None
Corresponding author at: MCHE-EMR, 3551 Roger Brooke Dr., Fort Sam Houston, TX
78234, United States.
E-mail address: Michael.D.April@post.harvard.edu (M.D. April).
and often admission to an intensive care unit (ICU) setting to maximize
their chances of survival [4].
Clinical denitions of sepsis remain an ongoing challenge in the eld
of sepsis research, particularly among this sickest subset of patients.
Specic inclusion criteria vary slightly across studies, but most of the
landmark randomized controlled trials of the treatment of patients
with septic shock required that subjects meet at least one of two types
of inclusion criteria. The rst type is refractory hypotension, or a systolic
blood pressure b 90 mm Hg after an intravenous uid bolus. The second
type is hyperlactatemia, or a lactate concentration equal to 4 mmol/L or
greater [5-9]. While not aligned exactly with these trial-based criteria,
international consensus statements and the Surviving Sepsis Campaign
Guidelines alike emphasize both hypotension and lactate levels in their
various sepsis denitions [2,10,11].
Despite this consensus that both of these clinical features are impor-
tant indicators of sepsis patient acuity and need for resuscitation, the
comparative prognostic value of these two criteria is uncertain. There

http://dx.doi.org/10.1016/j.ajem.2017.04.055
0735-6757/Published by Elsevier Inc.

Please cite this article as: April MD, et al, Emergency department septic shock patient mortality with refractory hypotension vs hyperlactatemia: A
retrospective cohort study, American Journal of Emergency Medicine (2017), http://dx.doi.org/10.1016/j.ajem.2017.04.055
2 M.D. April et al. / American Journal of Emergency Medicine xxx (2017) xxxxxx
is some limited data to suggest sepsis patients may have very different
characteristics and outcomes based upon whether they exhibit refracto-
ry hypotension, hyperlactatemia, or both. A pair of ICU-based observa-
tional studies did compare clinical parameters and survival among
septic shock patients with and without hyperlactatemia [12,13]. These
studies both concluded that the concomitant nding of hyperlactatemia
with refractory hypotension appears to represent a very distinct patho-
physiological prole. Indeed, patients with lactate levels exceeding
2.4 mmol/L had a signicantly higher mortality (42.9%) compared to pa-
tients without elevated lactate levels (7.7%) [13].
Mortality risk with refractory hypotension versus hyperlactatemia
remains unclear. Emergency physicians would benet from clarication
of the different implications of refractory hypotension versus
hyperlactatemia for eventual hospital outcomes.
1.2. Study purpose
This study comprises a retrospective cohort analysis by chart review
measuring patient characteristics, interventions, and outcomes among
patients admitted from the emergency department (ED) to an ICU. All
patients met at least one of two trial-based criteria for septic shock: re-
fractory hypotension (systolic blood pressure b 90 mm Hg after a 1 L in-
travenous uid bolus) or hyperlactatemia (initial lactate 4 mmol/L).
Our objective was to compare in-hospital mortality between patients
with isolated refractory hypotension, isolated hyperlactatemia, or both.
2. Methods
2.1. Study design
We conducted a retrospective cohort analysis by chart review of
adults presenting to an urban tertiary care hospital ED with an annual
patient volume exceeding 90,000. Providers included Emergency Med-
icine attending physicians, residents in training, and physician assis-
tants. The hospital has burn, cardiac, medical, surgical, and trauma
ICUs. Our hospital's institutional review board approved the study. We
adhered to the Strengthening the Reporting of Observational Studies
in Epidemiology (STROBE) statement in our research design, reporting,
and analysis [14].
2.2. Inclusion and exclusion criteria
We dened all inclusion and exclusion criteria a priori. The study in-
cluded adult (18 years) ED patients admitted to an ICU with suspected
infection during 1 August 2012 to 28 February 2015. We searched the
hospital's medical records to identify eligible patient encounters. We re-
trieved the medical charts for all potentially eligible patients. Regarding
inclusion criteria, we inferred suspected infection if during the ED stay
providers drew any non-wound body uid cultures (e.g., blood, urine)
regardless of antibiotic administration. Alternatively, we inferred infec-
tion if providers administered antibiotics during the ED stay and subse-
quently sampled body cultures within 24 h of antibiotic administration
regardless of whether sampling occurred during the ED stay [15,16].
Exclusion criteria included patients b 18 years of age and pregnant
women. We also excluded patients whose primary diagnoses com-
prised non-sepsis etiologies potentially causing hypo-perfusion. These
diagnoses included acute coronary syndrome, cardiac dysrhythmias
other than chronic atrial brillation, active hemorrhage, trauma,
burns, and overdoses [7-9,17]. In accordance with the landmark sepsis
trial exclusion criteria, we excluded patients with contraindications to
standard resuscitation measures such as advanced directives or reli-
gious objections to blood product transfusion. Finally, we excluded pa-
tients not meeting the trial-based criteria for septic shock as dened
by recent RCTs of sepsis management: systolic blood pressure
b 90 mm Hg after a 1 L intravenous uid bolus or initial lactate
4 mmol/L [6-8].
Please cite this article as: April MD, et al, Emergency department septic shock patient mortality with refractory hypotension vs hyperlactatemia: A
retrospective cohort study, American Journal of Emergency Medicine (2017), http://dx.doi.org/10.1016/j.ajem.2017.04.055
2.3. Selection of subjects
We identied eligible subjects by reviewing ED nursing notes for
every patient admitted to any ICU in our hospital during the study peri-
od. Two investigators independently reviewed these records to deter-
mine patient eligibility. We then cross-referenced this list against a
computer-generated list of patients admitted to any ICU in our hospital
during the study period with diagnoses related to sepsis per Interna-
tional Classication of Disease code to maximize capture of eligible pa-
tients. We then reviewed the records of these patients not already
captured by our previous algorithms for possible inclusion into the
study. Investigators resolved discordant eligibility decisions based on
mutual agreement between the two investigators.
2.4. Data abstraction methodology
We used standard chart review methodology to abstract data from
the hospital records for all study subjects [18]. All data abstractors
were resident physicians blinded to the study objective. After comple-
tion of all data collection, we conrmed that abstractor blinding to the
objective was effective by questioning each abstractor. Demographic
data collected from each patient chart included age, sex, residence
(nursing home versus other) and means of arrival to the ED (ambulance
versus private vehicle). Clinical data included initial vital signs,
suspected infection source, laboratory data, and past medical history.
Therapeutic data included times and quantities of administered crystal-
loid uid, antibiotics, vasopressors, blood products throughout the en-
tirety of the ED course and then during the rst 6 h of the ICU course.
Other data collected included the primary outcome of in-hospital mor-
tality as determined by documentation of a death note in each subject's
hospital record. Secondary outcomes collected included dialysis and
lengths of vasopressor therapy, mechanical ventilation, and
hospitalization.
Our data abstraction quality assurance measures included hour-long
training sessions for each data abstractor prior to study start. This ses-
sion entailed data abstraction from a practice patient chart with entry
into a template of the study database. During the data collection period,
all data abstractors had access to the study manual comprising abstrac-
tion rules for all study variables. Investigators abstracted all data into a
secure password-protected Excel database (version 14, Microsoft, Red-
mond, WA) with data validation for all study variables to minimize data
entry errors. Two separate investigators abstracted all study data. Data
collection spanned four months. Halfway through data completion,
the principal investigator calculated interim kappa coefcients for all
categorical variables and intra-class correlation coefcients for all con-
tinuous variables. The investigator provided e-mail reminders of data
collection rules for those variables with coefcients demonstrating
less than excellent inter-rater agreement (b0.81). The principal investi-
gator acted as nal arbiter to resolve all data entry discrepancies be-
tween abstractors.
2.5. Data analysis
We performed statistical analyses using SPSS (version 22, IBM,
Armonk, NY). We used descriptive statistics to describe our data:
means or medians for continuous variables and proportions for categor-
ical variables with 95% condence intervals (CIs) or interquartile ranges
(IQRs) as applicable. We stratied variables by three groups of patients
based upon septic shock inclusion criteria: (1) refractory hypotension;
(2) hyperlactatemia; or (3) both.
As we included all eligible study subjects during the study period, we
did not base our sample size on a sample size estimate. We rst com-
pared in-hospital mortality between the three groups of patients using
a chi-squared test. We compared secondary outcomes using chi-
squared tests or independent sample student t-tests as applicable. We
then calculated odds ratios via logistic regression models for in-hospital
 M.D. April et al. / American Journal of Emergency Medicine xxx (2017) xxxxxx 3

mortality based on presence of refractory hypotension or
hyperlactatemia. These models included variables for patient demo-
graphics, patient clinical characteristics, other ED therapeutic interven-
tions (e.g., uid administration). The models excluded cases with
missing data by list-wise deletion.
3. Results
3.1. Subject characteristics
We identied 321 patients admitted to an ICU in our hospital with
diagnoses potentially consistent with septic shock during the study pe-
riod. We excluded 119 patients. This resulted in 202 subjects included in
the study. Of these subjects, 101 (50.0%) met only the trial-based crite-
rion for septic shock of refractory hypotension. Forty-one (20.2%) of pa-
tients met only the criterion of hyperlactatemia. The remaining 60
(29.7%) patients met both criteria (Fig. 1).
Aside from systolic blood pressure and lactate levels, there were no
marked differences in baseline patient characteristics between the
three groups (Table 1). Patients with refractory hypotension were
more likely than patients with isolated hyperlactatemia to receive
30 mL/kg of intravenous uid within 30 min of meeting criteria for
septic shock. There was similarly a trend towards more vasopressor ad-
ministration in this group. ED interventions were otherwise comparable
across the three groups (Table 2).
3.2. Main results
Of the 202 patients, 164 (81.2%) survived to discharge while 38
(18.9%) died in hospital. In-hospital mortality was 10.9% among the
101 patients with isolated refractory hypotension, 24.4% among pa-
tients with isolated hyperlactatemia, and 28.3% among patients meeting
both criteria (p = 0.01). Patients with refractory hypotension
underwent longer courses of vasopressor therapy during their hospital-
izations. Other outcomes were generally comparable across the three
groups (Table 3).
For the logistic regression analysis, there was a signicant associa-
tion between the presence of hyperlactatemia and in-hospital mortali-
ty: odds ratio (OR) 2.9, 95% condence interval (CI) 1.27.4.
Conversely, there was no such association between the presence of re-
fractory hypotension and in-hospital mortality: OR 1.3, 95% CI 0.53.8.
There was similarly no association observed between other model var-
iables and patient survival (Table 4).
4. Discussion
4.1. Overview of ndings
Despite an extensive and growing literature on the therapeutic man-
agement of septic shock [6-8], much remains unknown regarding the
prognosis for various subsets of these patients. The literature reports
both hyperlactatemia [19] and distributive shock as hallmarks of the
sepsis syndrome [20]. Most of the landmark trials forming the founda-
tion of our contemporary understanding of the therapeutic manage-
ment of sepsis patients included subjects meeting criteria consistent
with either or both of these physiologic derangements [6-9]. Yet pa-
tients with refractory hypotension, hyperlactatemia, or both may all
represent very different patient populations with distinct physiologies
and outcomes. Our study offers an EDfocused comparison of these pa-
tient subsets using an observational design among a critically ill cohort
of patients with septic shock ultimately admitted to an ICU.
Our results indicate that patients with hyperlactatemia may repre-
sent individuals with more severe forms of septic shock. Our data
show that these patients are at higher risk of in-hospital mortality

Fig. 1. Flow diagram of study subject inclusion.

Please cite this article as: April MD, et al, Emergency department septic shock patient mortality with refractory hypotension vs hyperlactatemia: A
retrospective cohort study, American Journal of Emergency Medicine (2017), http://dx.doi.org/10.1016/j.ajem.2017.04.055
4 M.D. April et al. / American Journal of Emergency Medicine xxx (2017) xxxxxx

Table 1
Characteristics of patients meeting-trial based criteria for septic shock (n = 202).a

Variable Refractory hypotension only Hyperlactatemia only Refractory hypotension and hyperlactatemia All patients
(n = 101)a (n = 41)a (n = 60)a (n = 202)

Patient characteristics
Mean age, years (95% CI) 67.1 (63.970.3) 68.9 (63.674.4) 69.3 (64.674.1) 68.1 (65.870.4)
Male sex, % (95% CI) 58.6 (49.668.2) 51.4 (37.468.1) 60.0 (48.172.9) 57.9 (51.064.5)
Nursing home resident, % (95% CI) 10.9 (5.917.8) 12.2 (2.422.0) 10.0 (3.318.3) 10.9 (7.216.0)
Arrival by ambulance, % (95% CI) 47.5 (37.657.4) 63.4 (48.878.0) 51.7 (38.365.0) 52.0 (45.158.8)
Comorbidities
Chronic kidney disease, % (95% CI) 14.9 (7.920.8) 17.1 (7.329.3) 6.7 (1.713.3) 12.9 (8.918.2)
Congestive heart failure, % (95% CI) 13.9 (7.920.8) 12.2 (2.422.0) 13.3 (5.021.7) 13.4 (9.318.8)
Coronary artery disease, % (95% CI) 23.8 (15.832.7) 17.1 (7.329.3) 31.7 (20.043.3) 24.8 (19.331.1)
Hypertension, % (95% CI) 46.5 (36.656.4) 68.3 (53.782.9) 68.3 (56.780.0) 57.4 (50.564.0)
Presumed infection source
Gastrointestinal, % (95% CI) 13.9 (7.920.8) 7.3 (0.014.6) 11.7 (3.320.0) 11.9 (8.117.1)
Respiratory, % (95% CI) 43.6 (33.753.5) 51.2 (36.665.9) 38.3 (25.050.0) 43.6 (36.950.5)
Soft tissue, % (95% CI) 9.9 (5.015.8) 0.0 1.7 (3.318.3) 5.5 (3.09.6)
Urinary, % (95% CI) 17.8 (10.925.7) 22.0 (9.834.1) 23.3 (13.335.0) 20.3 (15.326.4)
Other, % (95% CI) 14.9 (6.924.8) 19.5 (4.924.3) 25.0 (10.041.6) 18.8 (14.024.8)
Initial vital signs
Median systolic blood pressure, mm Hg (IQR) 108 (90132) 128 (102154) 100 (81112) 107 (90131)
Median heart rate, beats per minute (IQR) 102 (88126) 114 (93132) 104 (87104) 106 (88124)
Median respiratory rate, breaths per minute (IQR) 21 (2026) 22 (2030) 23 (2030) 22 (2028)
Median temperature, degree Fahrenheit (IQR) 98.9 (98.1101.2) 99.0 (98.1101.2) 98.3 (97.699.9) 98.8 (97.9101.0)
Median oxygen saturation (IQR) 95 (9098) 95 (8798) 95 (9198) 95 (9098)
Initial laboratory values
Median bilirubin, mg/dL (IQR) 0.6 (0.41.1) 0.8 (0.52.2) 0.8 (041.5) 0.7 (0.41.4)
Median creatinine, mg/dL (IQR) 1.2 (0.91.8) 1.1 (0.81.6) 1.6 (1.02.2) 1.2 (0.91.9)
Median hemoglobin, g/dL (IQR) 11.7 (9.913.4) 11.6 (9.014.7) 11.4 (9.611.4) 11.6 (9.713.4)
Median lactate, mmol/L (IQR) 1.7 (1.32.5) 5.9 (4.87.9) 4.2 (4.06.3) 2.8 (1.65.0)
Median platelets, 10^3/L (IQR) 207 (130306) 213 (133273) 165 (123256) 186 (129282)
Median white blood cell count, 10^3/L (IQR) 11.9 (8.116.3) 13.5 (8.918.0) 15.1 (9.520.9) 13.0 (9.019.0)
Median white blood cell count bands, % (IQR) 15.0 (8.025.0) 25.0 (10.537.0) 16.0 (3.026.0) 16.0 (8.030.0)

Abbreviations: CI-condence intervals; IQR-interquartile range.

a
Septic shock is here dened as the presence of either refractory hypotension (systolic blood pressure b 90 mm Hg after 1 L of intravenous uid) or hyperlactatemia (lactate N 4 mmol/L).

than patients presenting with isolated refractory hypotension. These
main ndings held in our logistic regression analyses controlling for pa-
tient severity of illness and therapeutic interventions.
Our results are relevant to the ongoing debates about how best to
prognosticate patients with sepsis. The Third International Consensus
Denitions for Sepsis and Septic Shock (Sepsis-3) recognize the clinical
value of lactate in their revised denition of septic shock as a
vasopressor requirement to maintain a mean arterial pressure of
65 mm Hg or greater and serum lactate level greater than 2 mmol/L in
the absence of hypovolemia [2]. Yet, their recommended decision aid
for sepsis prognostication, quick sequential organ failure assessment
(qSOFA) does not incorporate lactate [2,16,21].
The literature is still in the process of establishing the applicability of
these recommendations to the undifferentiated ED population [22,23].

Table 2
Emergency department therapeutic interventions among emergency department patients with septic shock.a

Variable Refractory hypotension Hyperlactatemia Refractory hypotension and All patients

only (n = 101)a only (n = 41)a hyperlactatemia (n = 60)a (n = 202)

Intravenous crystalloid uid

Intravenous crystalloid uid administered, % (95% CI) 98.0 (95.0100.0) 92.7 (82.9100.0) 93.3 (85.098.3) 95.5
(91.697.8)
Mean door-to-uid bolus time, minutes (95% CI) 64.7 (46.890.6) 65.8 (33.5105.8) 68.3 (36.1119.2) 66.0
(47.384.6)
Intravenous crystalloid uid 30 mL/kg administered within 30 min 40.1 (31.149.0) 16.0 (5.028.7) 33.8 (21.145.9) 32.2
of meeting septic shock criteriaa, % (95% CI) (26.138.9)
Antibiotics
Antibiotics administered, % (95% CI) 98.9 (96.8100.0) 97.1 (91.4100.0) 96.3 (90.7100.0) 98.5
(95.599.7)
Mean door-to-antibiotic time, minutes (95% CI) 139.5 (118.0167.4) 124.8 (72.4182.2) 105.5 (86.7128.2) 127.4
(110.0144.8)
Vasopressors
Vasopressors administered, % (95% CI) 27.7 (19.837.6) 12.2 (2.422.0) 31.7 (20.045.0) 25.7
(20.232.2)
Mean door-to-vasopressor time, minutes (95% CI) 140.5 (90.8181.8) 165.7 120.2 (85.2164-5) 139.8
(130.2200.8) (105.8173.8)

Abbreviations: CI-condence intervals; IQR-interquartile range.

a
Septic shock is here dened as the presence of either refractory hypotension (systolic blood pressure b 90 mm Hg after 1 L of intravenous uid) or hyperlactatemia (lactate N 4 mmol/L).

Please cite this article as: April MD, et al, Emergency department septic shock patient mortality with refractory hypotension vs hyperlactatemia: A
retrospective cohort study, American Journal of Emergency Medicine (2017), http://dx.doi.org/10.1016/j.ajem.2017.04.055
 M.D. April et al. / American Journal of Emergency Medicine xxx (2017) xxxxxx 5

Table 3
Patient outcomes among emergency department patients with septic shock.a

Variable Refractory hypotension only Hyperlactatemia only Refractory hypotension and hyperlactatemia All patients
(n = 101)a (n = 41)a (n = 60)a (n = 202)

Mortality, % (95% CI) 10.9 (5.016.8) 24.4 (12.236.6) 28.3 (16.741.6) 18.8 (13.924.3)
In-hospital dialysis, % (95% CI) 9.8 (2.419.5) 5.9 (2.010.9) 8.3 (1.715.0) 7.4 (4.011.4)
Mean timeb receiving vasopressors, hours (95% CI) 153.8 (109.1199.4) 30.7 (5.367.2) 130.2 (96.2174.2) 132.2 (102.0166.7)
Mean timeb on mechanical ventilation, days (95% CI) 6.7 (3.010.9) 3.7 (1.86.0) 5.2 (2.28.7) 5.8 (3.48.6)
Mean hospitalization lengthb, days (95% CI) 15.0 (10.520.0) 5.5 (2.211.0) 12.0 (7.316.6) 12.9 (9.716.8)

Abbreviations: CI-condence intervals; IQR-interquartile range.

a
Septic shock is here dened as the presence of either refractory hypotension (systolic blood pressure b 90 mm Hg after 1 L of intravenous uid) or hyperlactatemia (lactate
N 4 mmol/L).
b
All time-based outcomes measured over entirety of hospitalization.

The Sepsis-3 validation cohorts reported qSOFA prognostic accuracy data entry discrepancies across data abstractors. We minimized selec-
values for ED encounters as part of pooled analyses including all non- tion bias in our study population through the systematic use of multiple
ICU encounters. The implication was that utility of qSOFA among ED pa- data sources to identify all potentially eligible patients during the study
tients is equivalent to the high prognostic accuracy values reported period.
among all non-ICU encounters [16]. In fact, the ED population is extraor- Another important limitation is minimal data regarding the specic
dinarily heterogeneous. Emerging data suggests that qSOFA's prognos- etiologies for our primary outcome of in-hospital mortality. In particu-
tic accuracy for in-hospital mortality among ED sepsis patients lar, we did not identify those patients who died as a result of withdrawal
ultimately requiring ICU-level care may more closely resemble the of support. That said, we did exclude all patients with documentation of
more modest values reported among sicker ICU patients [15,24]. advanced directives at the time of initial presentation.
The present study suggests that among these more acutely ill ED
sepsis patients, lactate offers an important additional prognostic tool. 4.3. Future research
Indeed, at least one prospective study of patients admitted to an ICU
with suspected infection noted increased prognostic accuracy of Given that this was a relatively small single center study, the litera-
qSOFA for in-hospital mortality when coupled with lactate values [25]. ture would benet from replication of our study in other settings to es-
Pending future research on optimal prognostication tools for ED pa- tablish the external validity of our ndings. Such studies would be
tients with sepsis, we believe emergency medicine providers should particularly powerful if gathering data in prospective fashion to mini-
routinely evaluate lactate values for all of their patients in planning mize the potential for data abstraction errors. Such studies would also
therapeutic and disposition plans. ideally achieve larger sample sizes, so enabling more robust matching
of patient groups prior to mortality comparisons via statistical methods
such as regression or propensity matching. Future research should also
4.2. Study limitations
build upon the growing body of literature seeking to synthesize newer
prognostication tools with established laboratory markers such as lac-
The main limitation of this study lies in its retrospective chart review
tate to further optimize the ability to accurately prognosticate undiffer-
design [18]. We sought to minimize data abstraction errors through for-
entiated ED patients with suspected infection [25].
mal chart abstractor training. Furthermore, we double entered all data
points used in this study. We also sought to minimize investigator bias
4.4. Conclusions
through the a priori explication of methods for collection of study vari-
ables, systematic training using mock charts, and dissemination of a
Hyperlactatemia appears associated with higher in-hospital mortal-
study manual with regular quality assurance checks to evaluate for
ity compared to refractory hypotension among ED patients with septic
shock. These results highlight the prognostic value of incorporating lac-
Table 4 tate measurements into the assessment of undifferentiated ED patients
Logistic regression analysis. with suspected infection. Emergency physicians should be particularly
OR for in-hospital 95% condence
vigilant in formulating treatment and disposition plans for patients
mortality interval with septic shock exhibiting hyperlactatemia.
Lower Upper
References
Age, years 1.02 0.99 1.05
Female sex 0.54 0.23 1.2 [1] Angus DC, van der Poll T. Severe sepsis and septic shock. N Engl J Med 2013;369:
HR, beats per minute 1.00 0.99 1.02 84051.
RR, breaths per minute 1.03 0.97 1.09 [2] Singer M, Deutschman CS, Seymour CW, Shankar-Hari M, Annane D, Bauer M, et al.
O2 Saturation, % 1.02 0.96 1.10 The Third International Consensus Denitions for Sepsis and Septic Shock (Sepsis-
Arrival by ambulance 0.49 0.20 1.18 3). JAMA 2016;315:80110.
Nursing home resident 2.09 0.63 6.94 [3] Rangel-Frausto MS, Pittet D, Costigan M, Hwang T, Davis CS, Wenzel RP. The natural
Congestive heart failure 1.48 0.64 2.31 history of the systemic inammatory response syndrome (SIRS). A prospective
Coronary artery disease 1.10 0.38 3.18 study. JAMA 1995;273:11723.
[4] Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J, Pinsky MR. Epidemi-
Hypertension 1.30 0.52 3.30
ology of severe sepsis in the United States: analysis of incidence, outcome, and asso-
Chronic kidney disease 0.89 0.26 3.06
ciated costs of care. Crit Care Med 2001;29:130310.
Bolus 30 mL/kg within 30 min of meeting 1.50 0.63 3.55
[5] Angus DC, Barnato AE, Bell D, Bellomo R, Chong CR, Coats TJ, et al. A systematic re-
septic shock criteriaa view and meta-analysis of early goal-directed therapy for septic shock: the ARISE,
Refractory hypotensiona 1.33 0.47 3.78 ProCESS and ProMISe investigators. Intensive Care Med 2015;41:154960.
Hyperlactatemiaa 2.92 1.15 7.40 [6] Investigators ARISE. Goal-directed resuscitation for patients with early septic shock.
N Engl J Med 2014;371:1496506.
Abbreviations: ED-emergency department, HR-heart rate, ICU-intensive care unit, O2-ox-
[7] Mouncey PR, Osborn TM, Power GS, Harrison DA, Sadique MZ, Grieve RD, et al. Trial
ygen, OR-odds ratio, RR-respiratory rate, SBP-systolic blood pressure.
a of early, goal-directed resuscitation for septic shock. N Engl J Med 2015;372:
Septic shock criteria include the presence of either refractory hypotension 130111.
(systolic blood pressure b 90 mm Hg after 1 L of intravenous uid) or hyperlactatemia [8] Investigators PROCESS. A randomized trial of protocol-based care for early septic
(lactate N 4 mmol/L). shock. N Engl J Med 2014;370:168393.

Please cite this article as: April MD, et al, Emergency department septic shock patient mortality with refractory hypotension vs hyperlactatemia: A
retrospective cohort study, American Journal of Emergency Medicine (2017), http://dx.doi.org/10.1016/j.ajem.2017.04.055
6 M.D. April et al. / American Journal of Emergency Medicine xxx (2017) xxxxxx

[9] Rivers E, Nguyen B, Havstad S, Ressler J, Muzzin A, Knoblich B, et al. Early goal-direct-
ed therapy in the treatment of severe sepsis and septic shock. N Engl J Med 2001;
345:136877.
[10] Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach H, Opal SM, et al. Surviving
sepsis campaign: international guidelines for management of severe sepsis and sep-
tic shock: 2012. Crit Care Med 2013;41:580637.
[11] Levy MM, Fink MP, Marshall JC, Abraham E, Angus D, Cook D, et al. 2001 SCCM/
ESICM/ACCP/ATS/SIS international sepsis denitions conference. Crit Care Med
2003;31:12506.
[12] Hernandez G, Bruhn A, Castro R, Pedreros C, Rovegno M, Kattan E, et al. Persistent
sepsis-induced hypotension without hyperlactatemia: a distinct clinical and physio-
logical prole within the spectrum of septic shock. Crit Care Res Prac 2012;2012:
536852.
[13] Hernandez G, Castro R, Romero C, de la Hoz C, Angulo D, Aranguiz I, et al. Persistent
sepsis-induced hypotension without hyperlactatemia: is it really septic shock? J Crit
Care 2011;26(435):e914.
[14] von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vandenbroucke JP, et al. The
strengthening the reporting of observational studies in epidemiology (STROBE)
statement: guidelines for reporting observational studies. PLoS Med 2007;4:e296.
[15] April MD, Aguirre J, Tannenbaum LI, Moore T, Pingree A, Thaxton RE, et al. Sepsis
clinical criteria in emergency department patients admitted to an intensive care
unit: an external validation study of quick sequential organ failure assessment. J
Emerg Med 2016 [epub ahead of print].
[16] Seymour CW, Liu VX, Iwashyna TJ, Brunkhorst FM, Rea TD, Scherag A, et al. Assess-
ment of clinical criteria for sepsis: for the Third International Consensus Denitions
for Sepsis and Septic Shock (Sepsis-3). JAMA 2016;315:76274.
[17] Peake SL, Delaney A, Bailey M, Bellomo R, Cameron PA, Cooper DJ, et al. Goal-direct-
ed resuscitation for patients with early septic shock. N Engl J Med 2014;371:
1496506.
[18] Kaji AH, Schriger D, Green S. Looking through the retrospectoscope: reducing bias in
emergency medicine chart review studies. Ann Emerg Med 2014;64:2928.
[19] Suetrong B, Walley KR. Lactic acidosis in sepsis: It's not all anaerobic: implications
for diagnosis and management. Chest 2016;149:25261.
[20] Quezado ZM, Natanson C. Systemic hemodynamic abnormalities and vasopressor
therapy in sepsis and septic shock. Am J Kidney Dis 1992;20:21422.
[21] Shankar-Hari M, Phillips GS, Levy ML, Seymour CW, Liu VX, Deutschman CS, et al.
Developing a new denition and assessing new clinical criteria for septic shock:
for the Third International Consensus Denitions for Sepsis and Septic Shock (Sep-
sis-3). JAMA 2016;315:77587.
[22] Faust JS. No SIRS; quick SOFA instead: SCCM redenes sepsis without emergency
medicine input. Ann Emerg Med 2016;67:15A9A.
[23] Zhou X, Tang G. Quick sepsis-related organ failure assessment (qSOFA) predicting
outcomes in patients with infection, some lingering doubts. Am J Emerg Med
2017;35:649.
[24] Wang JY, Chen YX, Guo SB, Mei X, Yang P. Predictive performance of quick sepsis-re-
lated organ failure assessment for mortality and ICU admission in patients with in-
fection at the ED. Am J Emerg Med 2016;34:178893.
[25] Ho KM, Lan NS. Combining quick sequential organ failure assessment with plasma
lactate concentration is comparable to standard sequential organ failure assessment
score in predicting mortality of patients with and without suspected infection. J Crit
Care 2017;38:15.

Please cite this article as: April MD, et al, Emergency department septic shock patient mortality with refractory hypotension vs hyperlactatemia: A
retrospective cohort study, American Journal of Emergency Medicine (2017), http://dx.doi.org/10.1016/j.ajem.2017.04.055