Minejima et al.

BMC Infectious Diseases (2019) 19:149

https://doi.org/10.1186/s12879-019-3770-4

RESEARCH ARTICLE Open Access

Utility of qSOFA score in identifying

patients at risk for poor outcome in
Staphylococcus aureus bacteremia
Emi Minejima1, Vanessa Delayo1, Mimi Lou1, Pamela Ny2, Paul Nieberg3, Rosemary C. She4 and
Annie Wong-Beringer1,2*

Abstract
Background: The prognostic capability of the quick Sequential Organ Failure Assessment (qSOFA) bedside scoring
tool is uncertain in non-ICU patients with sepsis due to bacteremia given the low number of patients previously
evaluated.
Methods: We performed a retrospective cohort study of adult hospitalized patients with Staphylococcus aureus
bacteremia (SAB). Medical charts were reviewed to determine qSOFA score, systemic inflammatory response
syndrome (SIRS) criteria, and Pitt bacteremia score (PBS) at initial presentation; their predictive values were
compared for ICU admission within 48 h, ICU stay duration > 72 h, and 30-day mortality.
Results: Four hundred twenty-two patients were included; 22% had qSOFA score ≥2. Overall, mean age was 56y
and 75% were male. More patients with qSOFA ≥2 had altered mentation (23% vs 5%, p < 0.0001), were infected
with MRSA (42% vs 30%, p = 0.03), had endocarditis or pneumonia (29% vs 15%, p = 0.0028), and bacterial
persistence ≥4d (34% vs 20%, p = 0.0039) compared to qSOFA <2 patients. Predictive performance based on
AUROC was better (p < 0.0001) with qSOFA than SIRS criteria for all three outcomes, but similar to PBS ≥2. qSOFA≥2
was the strongest predictor for poor outcome by multivariable analysis and showed improved specificity but lower
sensitivity than SIRS ≥2.
Conclusions: qSOFA is a simple 3-variable bedside tool for use at the time of sepsis presentation that is more
specific than SIRS and simpler to calculate than PBS in identifying septic patients at high risk for poor outcomes
later confirmed to have S. aureus bacteremia.
Keywords: S. aureus, Bacteremia, Mortality, Sepsis-3, qSOFA

Background Sepsis is currently defined as a life-threatening acute

Sepsis is a significant cause of critical illness worldwide organ dysfunction secondary to a dysregulated host re-
with an increasing incidence afflicting over 500,000 sponse to infection (Sepsis-3) [6]. Previously, sepsis was
cases/year [1] and mortality rates up to 80% [2, 3]. defined as meeting ≥2 systemic inflammatory response
Improved understanding of the pathobiology of sepsis syndrome (SIRS) criteria plus suspected infection. How-
has highlighted the heterogeneity in the host immune re- ever, the SIRS definition was criticized for its lack of sen-
sponse in sepsis, prompting a need to redefine and update sitivity; a recent study showed that 1 in 8 patients
the clinical criteria to characterize the syndrome [4–7]. admitted to the intensive care unit (ICU) with severe
sepsis did not meet the requisite minimum of 2 SIRS
criteria to fulfill the sepsis definition [8]. Thus, to im-
* Correspondence: anniew@usc.edu
1
prove the identification of patients at risk for clinical
Department of Clinical Pharmacy, University of Southern California School of
Pharmacy, 1985 Zonal Ave, Los Angeles, CA 90089, USA
deterioration from infection, Sepsis-3 recommends a
2
Department of Pharmacy, Huntington Hospital, 100 W. California Blvd, quick scoring system, quick sequential organ failure as-
Pasadena 91105, USA sessment score (qSOFA), comprised of 3 elements
Full list of author information is available at the end of the article

© The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Minejima et al. BMC Infectious Diseases (2019) 19:149 Page 2 of 9

assessed at the bedside (altered mental status, respiratory Extracted data were entered into a secure database, the
rate, and systolic blood pressure) and without the need Research Electronic Data Capture software hosted by
for laboratory tests. Patients who show evidence of 2 out University of Southern California.
of 3 elements are considered positive for risk of clinical
deterioration and therefore should prompt clinicians for Study definitions
investigation of organ dysfunction, initiation or escal- qSOFA score was calculated for each patient using the
ation of therapy, and consideration of transfer of care to worst value documented in the chart within the 24 h
the ICU. period of when the first positive blood culture for S. aur-
Staphylococcus aureus is the most common bacterial eus was drawn for three clinical variables worth 1 point
cause of sepsis in ICU patients [9]. S. aureus bacteremia each: systolic blood pressure ≤ 100 mmHg, respiratory
(SAB) affects an estimated 50/100,000 population, with rate ≥ 22 breaths/min, and Glasgow coma score < 15 or
an overall mortality rate of 19–57% in adults [10]. Despite altered mental status noted by the treating physician [6].
receipt of the standard treatment, persistent bacteremia qSOFA score of ≥2 points was used as the prognostic
occurs in one-third of patients beyond 7 days [11, 12], cutoff value in predicting clinical deterioration (ICU
which can lead to metastatic complications, relapse, pro- admission within 48 h and ICU stay > 72 h) and death
longed hospitalization, and increased mortality [11, 13–16]. within 30 days. PBS was calculated for each patient at
Since Sepsis-3, the prognostic utility of qSOFA for the onset of bacteremia to characterize the severity of ill-
in-hospital mortality has been prospectively validated in ness based on 5 variables which included temperature,
patients presenting to the emergency department with blood pressure, need for mechanical ventilation, evidence
suspected infection [17], as well as retrospectively evalu- of cardiac arrest, and mental status [14]. The SIRS criteria
ated [18–26] but only a small number of evaluated and severity of sepsis were also evaluated according to the
patients had documented bacteremia. Thus, our study 2012 surviving sepsis campaign definitions [27].
aims to evaluate non-ICU patients presenting with sepsis The source of bacteremia was divided relative to risk
who later were confirmed to have S. aureus bacteremia of mortality: low (< 10%), intermediate (10–20%), and
for 1) differentiating clinical characteristics based on high (> 20%) as previously defined [28]. Antibiotic ther-
qSOFA score and 2) comparison of the predictive per- apy was considered effective if antimicrobial sensitivity
formance of qSOFA to SIRS criteria and Pitt bacteremia was documented. Early clinical response was evaluated
score (PBS) as a prognostic tool to identify those at high on day 4 (72–96 h after initiation of effective antibiotic
risk for poor outcome. therapy) and determined as success or failure. Success
was complete or partial resolution (objective signs of
Methods improvement without complete resolution) of fever,
This was a 4-year retrospective cohort study conducted leukocytosis, local signs of infection, and clearance of
at three university-affiliated medical centers (600-bed blood cultures. Failure was defined as persistent growth
community teaching, 600-bed county teaching, and of blood cultures and/or worsening in objective signs
400-bed academic) in Los Angeles County, California. and symptoms of infection, including lack of resolution
The study protocol was approved by the institutional of fever, worsening or no improvement of leukocytosis,
review boards at each study site (University of Southern and/or lack of improvement of local signs of infection as
California and Quorum Review) and as the study was documented by the treating physician. Mortality was
retrospective, informed consent was waived. Microbiol- defined as death occurring within 30 days from date of
ogy reports were screened for all patients with at least when the first positive blood culture was drawn.
one positive blood culture for S. aureus between 2012
and 2016. Inclusion criteria were: 1) age ≥ 18 years, 2) Data analysis
receipt of ≥48 h of effective anti-staphylococcal therapy, Patients were grouped by high (≥ 2 points) or low (< 2
3) initiation of effective anti-staphylococcal therapy points) qSOFA scores as defined by Sepsis-3 criteria [6]
within 48 h from the time the first positive blood culture and compared for demographics, clinical presentation
was drawn, 4) monomicrobial growth in the blood and management, and outcomes. Endpoints for evalu-
culture, and 5) not admitted to the ICU prior to or at ation of prognostic capability were ICU admission
onset of bacteremia. Patients with receipt of ≤48 h of ef- within 48 h of SAB presentation, length of ICU stay > 72
fective therapy were excluded as assessment of therapy h during the course of SAB, and 30-day mortality. The
on mortality outcome was unlikely to be contributory to prognostic value of meeting at least two SIRS criteria
the antibiotic therapy given. [27] was compared to qSOFA score of ≥2. The positive
Medical charts were reviewed for relevant demographics predictive value (PPV), negative predictive value (NPV),
including comorbidities, laboratory, radiographic data, sensitivity, and specificity of qSOFA score of ≥2, SIRS
surgical and antimicrobial management, and clinical data. score ≥2, and PBS score ≥2 to predict each study
Minejima et al. BMC Infectious Diseases (2019) 19:149 Page 3 of 9

endpoint were calculated. PBS score ≥ 2 was chosen infected with MRSA (42% vs 30%, OR 1.72, 95% CI
based on the study by Hill et al. which showed a doub- 1.06–2.79, p = 0.03) and twice as likely to have a
ling of relative risk for mortality when PBS scores were 2 high-risk source of bacteremia (29% vs 15%, OR 2.39,
compared to patients with scores of < 2 in SAB [14]. A 95% CI 1.37–4.15, p = 0.0028). The high-risk sources
receiver operating characteristic (ROC) curve was cre- were primarily endocarditis/endovascular source (16%)
ated to calculate the corresponding area under the ROC or pneumonia (11%) in the high qSOFA group. Skin and
(AUROC). Secondary endpoints included SAB outcomes soft tissue (20%) and osteoarticular (20%) were the most
of day 4 clinical failure and persistence of bacteremia common sources in the low qSOFA group.
past 4 days.
Descriptive analyses were conducted using Wilcoxon Clinical presentation and management
rank sum tests or student t test for continuous data and The high qSOFA group had a sicker presentation at on-
chi-square or Fisher’s exact test for categorical data. set of bacteremia with higher proportion of patients with
Univariate followed by multivariable logistic regression PBS ≥2 points (69% vs 12%, p < 0.0001). (Table 2) Nearly
analysis were performed to determine the predictors of all (98%) patients in the high qSOFA group met SIRS
each outcome after controlling for age, gender, and criteria for sepsis compared to 74% in the low qSOFA
source risk category. Only variables found to be signifi- group (p < 0.0001). In the high qSOFA group, respiratory
cantly different between groups by univariate analysis rate ≥ 22 was the most frequent qSOFA criteria met
were included in the multivariable logistic regression (85% vs low qSOFA 24%, p < 0.0001), followed by sys-
analysis for each primary endpoint. A p value < 0.05 was tolic blood pressure ≤ 100 mmHg (68% vs low qSOFA
considered significant. Statistical analyses were per- 18%, p < 0.0001).
formed using GraphPad Prism v4.0 (San Diego, CA, Overall, combination of vancomycin with a
USA) or SAS version 9.4 (SAS Institute, Cary, NC). beta-lactam antibiotic for empiric therapy was the most
common therapy administered (high 63% vs low 53%)
Results though vancomycin monotherapy was more frequent in
Study population the low qSOFA group (29%) compared to the high
A total of 623 hospitalized patients with growth of S. qSOFA group (12%). Of the patients infected with
aureus in a blood culture were screened; 402 patients MSSA, 76% (205/271) were treated with an anti-
met inclusion criteria. Two hundred and twenty-one pa- staphylococcal beta-lactam agent for definitive therapy,
tients were excluded for the following reasons: 3 patients while MRSA bacteremia patients were treated with
were < 18 years old, 60 patients received < 48 h of effect- vancomycin in 59% (77/131) and daptomycin in 27%
ive antimicrobial therapy, 9 patients were initiated on (36/131) as definitive therapy. There was a trend towards
antibiotics > 48 h from the first positive blood culture, more patients in the high qSOFA group (78%) receiving
70 patients had polymicrobial blood cultures, 17 patients effective therapy on or before the first day of positive
had incomplete medical charts, and 62 patients were ad- blood culture (vs low 68%, p = 0.088). Duration of effect-
mitted in the ICU > 24 h prior to the onset of ive therapy during hospitalization was significantly lon-
bacteremia. Overall, demographics showed the mean age ger in the high qSOFA group compared to the low
of 56 years, 75% were male, and 87% had community- qSOFA group (high 14 days vs low 8.5 days, p = 0.0011).
onset bacteremia (Table 1). Nearly one quarter (22%, 90/ Similar rates of patients in both groups received Infec-
402) of included patients had qSOFA scores ≥2 and were tious Disease (ID) consultation (high 52% vs low 58%, p
considered the high qSOFA group. Among those who = 0.33) with the same median of 2 days for time to re-
acquired nosocomial SAB (n = 53), high qSOFA patients ceive consultation between the groups. Source control
had longer duration of hospitalization prior to onset of procedure for the SAB management was also similar in
bacteremia compared to the low qSOFA group (median rates in both groups (high 40% vs low 49%, p = 0.12)
20 vs 9 days, p = 0.018). Regardless of qSOFA score, with a median of 2 days to perform the procedure (high
about half the patients (high 50% vs low 53%, p = 0.63) 2 days (IQR 1, 4.75) vs low 2 days (IQR 1, 4), p = 0.41).
had history of three or more comorbidities as docu-
mented by the treating physician in the medical chart. Prognosis and outcome
Notably, a higher proportion of low qSOFA patients Patients with high qSOFA score had significantly worse
(10% vs 2%, p = 0.023) had no pre-existing comorbid prognosis compared to those in low qSOFA group: ICU
condition per physician documentation in the medical admission at any point after the onset of SAB was more
chart. One third of patients had presence of hardware at frequent (64% vs 21%, p < 0.0001), more patients were
the onset of SAB (high 33% vs low 30%, p = 0.53). transferred into the ICU within 48 h of onset of SAB
Overall, MSSA was the predominant (67%) infecting (57% vs 16%, p < 0.0001), overall duration of ICU stay
pathogen. The high qSOFA group were more likely to be was prolonged by 2 days (median, p = 0.011), and a
Minejima et al. BMC Infectious Diseases (2019) 19:149 Page 4 of 9

Table 1 Baseline Demographics comparing patients with high qSOFA score vs low qSOFA scores
Characteristics High qSOFA n = 90 (%) Low qSOFA n = 312 (%) p value
a
Age, yr. 56.5 ± 13.99 55.2 ± 15.1 0.66
Male 60 (67) 241 (77) 0.053
Residence Prior to Admission
Home 66 (73) 240 (77) 0.81
Skilled Nursing Facility 5 (6) 11 (4)
Other hospital/rehab center 9 (10) 30 (10)
Homeless 10 (11) 31 (10)
Comorbid conditions
None 2 (2) 30 (10) 0.02
Diabetes Mellitus 35 (39) 142 (46) 0.28
End stage renal disease on dialysis 16 (18) 51 (16) 0.75
Cirrhosis 11 (13) 32 (11) 0.52
b
Cardiovascular disease 40 (44) 170 (54) 0.096
Immunosuppressed c 15 (17) 37 (12) 0.28
≥ 3 comorbid conditions 45 (50) 165 (53) 0.63
Race/Ethnicity 0.18
Caucasian 28 (33) 92 (30)
Asian 14 (16) 26 (8)
African American 10 (12) 34 (11)
Hispanic 32 (37) 137 (44)
Other 2 (2) 21 (7)
History of Intravenous Drug Use 13 (15) 32 (10) 0.26
History of S. aureus infection 14 (16) 57 (18) 0.08
History of IV vancomycin therapy 12 (13%) 45 (15%) 0.79
Community-onset SAB 77 (86) 272 (87) 0.72
Microbial characteristics 0.03
MSSA 52 (58) 219 (70)
MRSA 38 (42) 93 (30)
d
Source Risk Category 0.0036
Low risk 20 (22) 64 (21)
Intermediate risk 43 (48) 202 (65)
High risk 26 (29) 46 (15)
Study site 0.43
County teaching hospital 64 (71) 205 (66)
Academic hospital 11 (12) 35 (11)
Community teaching hospital 15 (17) 72 (23)
a
mean ± standard deviation; Cardiovascular disease includes hypertension, dyslipidemia, congestive heart failure, coronary artery disease; c Immunosuppressed:
b

malignancy, recent chemotherapy, chronic steroid use (prednisone ≥20 mg/day or equivalent); SAB = S. aureus bacteremia; d Sources of infection considered low
risk were intravascular (IV) catheters, urinary tract infection, ear-nose-larynx, gynecologic, and several manipulation-related sources; intermediate risk were
osteoarticular, soft-tissue, and unknown sources; and high risk were endovascular, lower respiratory tract, intra-abdominal, and central nervous system foci

significantly higher rate of ICU stays of > 72 h (48% vs stay after the onset of SAB by 5 days (median) compared
12%, p < 0.0001). (Table 3) In addition, the high qSOFA to the low qSOFA group (p = 0.001). Among patients
group had higher rate of persistently positive blood cul- with initial high qSOFA scores, 49% experienced clinical
tures for S. aureus despite receipt of ≥4 days of effective failure at day 4 and were 11 times more likely to die
therapy (34% vs 20%, p = 0.0039), higher rate of 30-day than those showing early clinical response (34%, 15/44
mortality (19% vs 3%, p < 0.0001), and a longer length of vs 4%, 2/45, p = 0.0004; OR 11.12, 95% CI: 2.36–52.35).
Minejima et al. BMC Infectious Diseases (2019) 19:149 Page 5 of 9

Table 2 Clinical Presentation at onset of S. aureus bacteremia Table 3 Comparison of Clinical Outcomes of S. aureus
Clinical Presentation High qSOFA Low qSOFA p value bacteremia by qSOFA score
n = 90 (%) n = 312 (%) Clinical Outcomes High qSOFA Low qSOFA p value
Pitt Bacteremia Score a 2 (0, 3) 0 (0, 1) < 0.0001 n = 90 (%) n = 312 (%)

Score ≥ 2 57 (69) 37 (12) < 0.0001 ICU admission 58 (64) 67 (21) < 0.0001

Score ≥ 4 19 (23) 7 (2) < 0.0001 ICU admission within 48 h 51 (57) 50 (16) < 0.0001
of first positive culture
SIRS criteria
Need for vasopressors 31 (35) 17 (6) < 0.0001
Sepsis 88 (98) 232 (74) < 0.0001 a
Duration of ICU stay, days 5 (2, 12.3) 3 (1, 7) 0.01
Severe sepsis 59 (66) 86 (28) < 0.0001
Duration of ICU stay > 72 h 43 (48) 38 (12) < 0.0001
Septic shock 25 (28) 3 (1) < 0.0001
Day 4 Success 45 (50) 228 (73) < 0.0001
No sepsis 2 (2) 80 (26)
Died 2 (4) 2 (1) 0.07
Presenting symptoms
Day 4 Failure 44 (49) 83 (27) < 0.0001
Fever 35 (39%) 100 (32%) 0.25
Died 15 (34) 7 (8) 0.0003
Pain 28 (31%) 155 (50%) 0.0018
Microbial Persistence 31 (34) 62 (20) 0.0039
Altered mental status 21 (23%) 16 (5%) < 0.0001 on day 4
Shortness of breath 10 (11%) 19 (6%) 0.11 Died 10 (32) 3 (5) 0.0007
a
median (IQR), data available for 83 patients in High qSOFA group and 301 Initial GCS < 15 54/89 (61) 31/306 (10) < 0.0001
patients in the Low qSOFA group
Died 13 (24) 2 (6) 0.04
Initial SBP ≤ 100 mmHg 57/84 (68) 52/297 (18) < 0.0001
On the contrary, the mortality rate was only 8% (7/83)
in the low qSOFA group who experienced early failure. Died 12 (21) 3 (6%) 0.026
Among those who died (n = 26), median time to death Initial RR ≥ 22 breaths 72/85 (85) 71/298 (24) < 0.0001
per minute
from initial positive blood culture was 10 days and 18
days for the high and low qSOFA group, respectively (p Died 14 (19) 4 (6) 0.021
= 0.26). In a sub-analysis of patients in the high qSOFA 30-day mortality 17 (19) 9 (3) < 0.0001
group, those who died (n = 17) vs survived (n = 73) were Total Hospital LOS, days a
16 (8, 32) 10 (6, 19) < 0.0001
older (mean age: 61.7y ± 3 vs 55.6y ± 1.7, p = 0.11), had LOS after first positive 14 (7.8, 23) 9 (6.25, 17) 0.001
more comorbid conditions (3 or more: 71% vs 45%, p = culture, days a
0.1), received ID consultation (65% vs 49%, p = 0.29), Disposition of survivors N = 73 N = 303 0.92
and infection with MRSA (53% vs 40%, p = 0.42) but the Home 48 (66) 198 (65)
differences observed did not reach statistical signifi-
Skilled nursing facility 12 (16) 47 (16)
cance. However, rate of patients with ICU admission
(88% vs 45%, p = 0.002) and need for vasopressors (88% Outside Hospital 4 (5) 23 (8)
vs 22%, p < 0.0001) were significantly more frequent in Rehab center 7 (10) 23 (8)
those who died vs survived. Homeless/jail 2 (3) 12 (4)
ICU intensive care unit, LOS length of stay, GCS glasgow coma scale score, SBP
Predictive performance of qSOFA, SIRS, and PBS clinical systolic blood pressure, RR respiratory rate; a median (IQR)

criteria
The predictive performance of qSOFA was compared to (95% CI 0.65–0.75) vs 0.7 (95% CI 0.65–0.75), p = 1],
SIRS and PBS criteria and is shown in Fig. 1. qSOFA ICU stay > 72 h [0.7 (95% CI 0.64–0.76) vs 0.67 (95% CI
had a high specificity and negative predictive value with 0.67 (0.61–0.73), p = 0.34)], and 30-day mortality [0.76
moderate to poor sensitivity and positive predictive (95% CI 0.67–0.86) vs 0.72 (95% CI 0.62–0.82), p = 0.46].
values for all three endpoints, while the reverse was The most significant factors identified from the univari-
shown for SIRS criteria. The AUROC was significantly ate analysis were evaluated in a multivariable model and
higher for qSOFA ≥2 compared to SIRS ≥2 for ICU ad- controlled for age, gender, and source risk category.
mission within 48 h of onset of bacteremia [0.7 (95% CI qSOFA scores ≥2 was the most significant predictor of
0.65–0.75) vs 0.58 (95% CI 0.54–0.62)], ICU stay longer each outcome in the model, with more than 4 times
than 72 h [0.7 (95% CI 0.64–0.76) vs 0.56 (95% CI 0.52– greater risk for poor outcome compared to those with
0.60)], and 30-day mortality [0.76 (95% CI 0.67–0.86) vs qSOFA < 2 (Table 4). SIRS ≥2 was removed from the
0.54 (95% CI 0.47–0.62) (all comparisons p < 0.0001). model as it was not significant. PBS ≥2 was a significant
The AUROC for qSOFA ≥2 was similar to PBS ≥2 for predictor for all three outcomes but the risk associated
ICU admission within 48 h of onset of bacteremia [0.7 with qSOFA ≥2 was approximately two-fold greater for
Minejima et al. BMC Infectious Diseases (2019) 19:149 Page 6 of 9

the outcome of duration of ICU stay > 72 h and 30-day

mortality.

Discussion
Sepsis-3 introduces qSOFA as a simple bedside tool for
screening patients with suspicion of infection who are at
increased risk for clinical deterioration [6]. Our study
aimed to retrospectively evaluate the predictive perform-
ance of qSOFA scoring system in the setting of S. aureus
bacteremia which has not been adequately studied
previously.
Our findings are consistent with those from prior
studies involving other infectious syndromes. Freund et
al. prospectively evaluated the prognostic accuracy of
qSOFA to predict poor outcomes in patients presenting to
the emergency department with suspicion of infection.
Similar to our study, they reported 24% of their study
population had qSOFA scores ≥2 and that qSOFA had
improved predictive performance compared to SIRS with
AUROC of 0.73 (95% CI 0.68–0.77) for ICU admission,
0.71 (95% CI 0.66–0.76) for duration in ICU > 72 h, and
0.8 (95% CI 0.74–0.85) for in-hospital mortality [17]. In
our study, those with initial qSOFA score of ≥2 were 8
times more likely to have prolonged ICU length of stay (>
72 h) and a fatal outcome when compared to patients with
a qSOFA score < 2. We used another severity of illness
score, PBS previously validated in SAB to predict mortality
[14]. PBS scoring system uses 5 variables, 3 of which
overlap with qSOFA (hypotension, mental status, and re-
spiratory status) while the other 2 are temperature and
evidence of cardiac arrest; the latter requires laboratory
testing. Although similar variables are used in the PBS,
parameters such as fever and altered mental status are
Fig. 1 Sensitivity, specificity, positive predictive value, and negative stratified such that more points are assigned as the
predictive value of scoring systems to predict outcomes. Error bars
represent 95% confidence intervals; PBS = Pitt Bacteremia Score
measured value is increasingly abnormal, whereas with
qSOFA the points are assigned with a simple yes or no.
The AUROC was comparable between qSOFA and PBS
for each of the predefined outcomes. Considering that
both PBS ≥2 and qSOFA ≥2 were shown to significantly
predict the three clinical outcomes by multivariable
logistic regression analysis, qSOFA would be a more

Table 4 Predictors of Outcome by Multivariable Logistic Regression

ICU admission within 48 h of onset of SAB Duration of ICU stay longer than 72 h 30-day mortality
Variable OR (95% CI) p-value OR (95% CI) p-value OR (95% CI) p-value
Age 0.99 (0.97–1.00) 0.10 0.99 (0.97–1.00) 0.13 1.03 (1.00–1.06) 0.08
Gender 1.14 (0.62–2.11) 0.68 1.04 (0.55–1.97) 0.90 0.90 (0.34–2.42) 0.84
Source Risk Category: High vs. 2.42 (1.27–4.59) 0.007 2.01 (1.05–3.85) 0.04 0.78 (0.26–2.29) 0.65
Intermediate & Low
qSOFA ≥ 2 4.40 (2.40–8.09) < 0.0001 4.67 (2.51–8.68) < 0.0001 6.94 (2.49–19.31) 0.0002
PBS ≥ 2 4.69 (2.54–8.67) < 0.0001 2.89 (1.53–5.48) 0.001 3.38 (1.25–9.16) 0.02
ICU intensive care unit, qSOFA quick sequential organ failure assessment, PBS pitts bacteremia score, OR odds ratio
Minejima et al. BMC Infectious Diseases (2019) 19:149 Page 7 of 9

practical tool for use at the bedside given its ease of use death was observed among those who continued on the
compared to PBS. same treatment despite lack of early response compared
The goal of the qSOFA scoring system is to identify to those who switched to alternative therapy (mortality
those at increased risk for clinical deterioration [29], and 38% vs 10%, p = 0.13). In addition, every additional day
qSOFA score ≥ 2 had good negative predictive value and of persistently positive blood culture with S. aureus,
specificity in predicting prolonged ICU stay and 30-day significantly increases the patient’s risk for mortality
all-cause mortality in patients with SAB, similar to prior [32]. Thus, early recognition of high-risk patients using
studies [17, 21, 26]. Inversely, SIRS criteria has high qSOFA could prompt clinicians to make a timely change
sensitivity but was associated with high rate of false in management including performing source control
positives. Although qSOFA may lack sensitivity com- procedures earlier in the course to avoid negative out-
pared to SIRS in capturing patients who died, there were comes. As the availability of rapid diagnostic technology
only 9 deaths in the low qSOFA group, of which 7 has significantly shortened time to organism identifica-
patients had non-SAB related deaths, including with- tion for bloodstream infection cases, early recognition of
drawal of care. As SIRS criteria was criticized for lacking patients at high risk of poor outcome could allow anti-
specificity which resulted in over-prescribing of antimi- microbial stewardship teams to increase their vigilance
crobials, the increased specificity with qSOFA criteria of aiding clinicians to obtain early source control, screen
supports its use to be a favored screening system to for metastatic complications, obtain ID consultation,
identify patients most likely to have poor outcome and and initiate optimal antimicrobial therapy as soon as
therefore needing higher level of care. resistance information is known.
MSSA was the predominant pathogen overall. None- Our study has several limitations. As qSOFA scores
theless, a significantly higher proportion of patients in were not used to prospectively guide therapeutic man-
the high qSOFA group (42%) was infected with MRSA. agement in this study, we could not control for all
Notably, among patients with high qSOFA score, clinical confounding variables that may have affected the clinical
outcomes were worse in those infected with MRSA outcomes. It is notable that similar proportion of
compared to those with MSSA bacteremia: longer length patients between the high and low qSOFA groups re-
of hospital stay and microbial persistence beyond 4 days ceived infectious disease consultation and source control
of effective therapy. This is consistent with prior litera- procedure. While this study was conducted in the same
ture showing an association of worse outcomes with geographic region, the number of patients included in
MRSA compared to MSSA bacteremia [30]. Despite the this study was relatively large and included diverse
outcome difference observed between MSSA and patient populations such as the elderly and the medically
MRSA-infected patients in our study, there was no sig- underserved younger populations (age range 19–89
nificant difference seen in MRSA infected patients with years). Our study population had a higher proportion of
high qSOFA scores in terms of initial PBS score, the male over female patients which is consistent with prior
proportions with high risk sources of SAB (MSSA 25% epidemiologic studies, which found SAB to occur more
vs MRSA 34%, p = 0.48), and time to initiation of effect- frequently in male patients [33, 34]. We evaluated
ive therapy compared to MSSA infected patients with mortality within 30-days from onset of SAB, which is a
high qSOFA scores. It is possible that inherent difference common endpoint used in SAB studies while others
in antimicrobial efficacy and/or indirect immunomodu- employed in-hospital mortality as an endpoint in the
latory effects between agents used to treat MRSA and sepsis literature. Only one patient in our study died past
those (beta-lactams) available to treat MSSA bacteremia 30 days (died on day 46) after the onset of SAB. Thus,
contributed to the outcome difference [30]. Taken using an alternative definition of mortality did not affect
together, results from our study adds to the existing data our main analysis. Although at initial presentation of
[17, 18, 20] in support of the use of qSOFA as a bedside sepsis, the diagnosis of S. aureus bacteremia would not
tool in identifying non-ICU patients presenting with sep- have been known, our study demonstrated that a high
sis (including those later confirmed to have bacteremia qSOFA score measured at initial presentation of sepsis
due to S. aureus) for early and aggressive management. predicted poor outcomes in this large cohort of patients
Importantly, we found that in patients with persistent who had confirmed S. aureus bacteremia later. Thus, if
bacteremia, a high qSOFA score at initial presentation prospectively applied, qSOFA would likely demonstrate
was associated with a 9 times higher risk of death than prognostic capability in patients with sepsis due to a
those with a low qSOFA. This finding is consistent with variety of infection types, including bacteremia.
our previously published study in that a lack of early
response in S. aureus bacteremia was the strongest pre- Conclusion
dictor of treatment failure in a multivariable logistic re- qSOFA is a simple 3-variable bedside tool that is more
gression model [31]. A trend towards higher risk of specific than SIRS and simpler to calculate than PBS in
Minejima et al. BMC Infectious Diseases (2019) 19:149 Page 8 of 9

identifying septic patients at initial presentation (later 3

Department of Medicine – Infectious Diseases, Huntington Hospital, 100 W.
confirmed to have S. aureus bacteremia) who are at high California Blvd, Pasadena 91105, USA. 4Department of Pathology, Keck School
of Medicine, Los Angeles 90089, USA.
risk for poor outcomes. Patients identified with high
qSOFA score should receive aggressive management for Received: 22 October 2018 Accepted: 31 January 2019
infection including possible transfer to higher-level of
care in the ICU. Future studies should include prospective
evaluation of the utility of qSOFA scoring system in References
1. Martin GS, Mannino DM, Eaton S, Moss M. The epidemiology of sepsis in
guiding antibiotic selection (e.g. need for MRSA coverage) the United States from 1979 through 2000. N Engl J Med. 2003;348(16):
particularly when combining with use of rapid diagnostic 1546–54.
platform to confirm diagnosis of S. aureus bacteremia. 2. Jawad I, Lukšić I, Rafnsson SB. Assessing available information on the
burden of sepsis: global estimates of incidence, prevalence and mortality. J
Abbreviations Glob Health. 2012;2(1):010404.
AUROC: area under the receiver operating characteristic; ICU: Intensive care 3. Vincent JL, Rello J, Marshall J, Silva E, Anzueto A, Martin CD, Moreno R, Lipman
unit; MRSA: Methicillin-resistant S. aureus; MSSA: Methicillin-sensitive S. aureus; J, Gomersall C, Sakr Y, et al. International study of the prevalence and
NPV: Negative predictive value; PBS: Pitt bacteremia score; PPV: Positive outcomes of infection in intensive care units. JAMA. 2009;302(21):2323–9.
predictive value; qSOFA: Quick sequential organ failure assessment; 4. Horan TC, Andrus M, Dudeck MA. CDC/NHSN surveillance definition of
ROC: Receiver operating characteristic; SAB: Staphylococcus aureus health care-associated infection and criteria for specific types of infections
bacteremia; SIRS: Systemic inflammatory response syndrome in the acute care setting. Am J Infect Control. 2008;36(5):309–32.
5. Jones RN. Microbial etiologies of hospital-acquired bacterial pneumonia and
Acknowledgements ventilator-associated bacterial pneumonia. Clin Infect Dis. 2010;51(Suppl 1):
We thank Anne Au and the lab personnel at Huntington Hospital, Los S81–7.
Angeles County-University of Southern California Medical Center, and Keck 6. Singer M, Deutschman CS, Seymour CW, Shankar-Hari M, Annane D, Bauer
Hospital of USC for assistance with laboratory specimen collection; and M, Bellomo R, Bernard GR, Chiche JD, Coopersmith CM, et al. The third
Nancy Bui, Joshua Wang, Daniel Salas, Nicole Pepe, and Stephanie Mac for international consensus definitions for Sepsis and septic shock (Sepsis-3).
assisting with data collection. JAMA. 2016;315(8):801–10.
7. Hotchkiss RS, Monneret G, Payen D. Immunosuppression in sepsis: a novel
Funding understanding of the disorder and a new therapeutic approach. Lancet
This research did not receive any specific grant from funding agencies in the Infect Dis. 2013;13(3):260–8.
public, commercial, or not-for-profit sectors. 8. Kaukonen KM, Bailey M, Pilcher D, Cooper DJ, Bellomo R. Systemic
inflammatory response syndrome criteria in defining severe sepsis. N Engl J
Availability of data and materials Med. 2015;372(17):1629–38.
The datasets generated and/or analyzed during the current study are not 9. Vincent JL, Marshall JC, Namendys-Silva SA, François B, Martin-Loeches I,
publicly available as data was collected through manual chart review from Lipman J, Reinhart K, Antonelli M, Pickkers P, Njimi H, et al. Assessment of
each institution’s electronic medical record and IRB approval to publish the the worldwide burden of critical illness: the intensive care over nations
raw data was not requested by the authors. Datasets are available from the (ICON) audit. Lancet Respir Med. 2014;2(5):380–6.
corresponding author on reasonable request. 10. van Hal SJ, Jensen SO, Vaska VL, Espedido BA, Paterson DL, Gosbell IB.
Predictors of mortality in Staphylococcus aureus bacteremia. Clin Microbiol
Authors’ contributions Rev. 2012;25(2):362–86.
EM, VD, and AWB collected and analyzed data, and were major contributors 11. Hawkins C, Huang J, Jin N, Noskin GA, Zembower TR, Bolon M. Persistent
in writing of the manuscript. ML, PN2, and RS was involved with analyzing Staphylococcus aureus bacteremia: an analysis of risk factors and outcomes.
and interpreting the data and editing of the manuscript. PN1 was involved Arch Intern Med. 2007;167(17):1861–7.
with data collection and editing of the manuscript. All authors read and 12. Chua T, Moore CL, Perri MB, Donabedian SM, Masch W, Vager D, Davis SL,
approved the final manuscript. Lulek K, Zimnicki B, Zervos MJ. Molecular epidemiology of methicillin-
resistant Staphylococcus aureus (MRSA) bloodstream isolates in urban
Ethics approval and consent to participate Detroit. J Clin Microbiol. 2008;46(7):2345–52.
The study protocol was approved by the institutional review boards (IRB) at 13. Soriano A, Martínez JA, Mensa J, Marco F, Almela M, Moreno-Martínez A,
each respective study site (University of Southern California IRB for both Los Sánchez F, Muñoz I, Jiménez de Anta MT, Soriano E. Pathogenic significance
Angeles County-University of Southern California Medical Center and Keck Hos- of methicillin resistance for patients with Staphylococcus aureus bacteremia.
pital of University of Southern California; Quorum Review IRB for Huntington Clin Infect Dis. 2000;30(2):368–73.
Hospital). As the study was non-interventional and there was no contact made 14. Hill PC, Birch M, Chambers S, Drinkovic D, Ellis-Pegler RB, Everts R, Murdoch
with enrolled patients, the University of Southern California IRB and Quorum Re- D, Pottumarthy S, Roberts SA, Swager C, et al. Prospective study of 424
view IRB allowed for informed consent to be waived. The study did not include cases of Staphylococcus aureus bacteraemia: determination of factors
vulnerable groups or where coercion of the patient could take place. affecting incidence and mortality. Intern Med J. 2001;31(2):97–103.
15. Das I, O'Connell N, Lambert P. Epidemiology, clinical and laboratory
Consent for publication characteristics of Staphylococcus aureus bacteraemia in a university hospital
Not applicable. in UK. J Hosp Infect. 2007;65(2):117–23.
16. Chang FYMB, Peacock JE Jr, Musher DM, Triplett P, Mylotte JM, O'Donnell A,
Competing interests Wagener MM, Yu VL. A prospective multicenter study of Staphylococcus
The authors declare that they have no competing interests. aureus bacteremia: incidence of endocarditis, risk factors for mortality, and
clinical impact of methicillin resistance. Medicine (Baltimore). 2003;82(5):
322–32.
Publisher’s Note 17. Freund Y, Lemachatti N, Krastinova E, Van Laer M, Claessens YE, Avondo A,
Springer Nature remains neutral with regard to jurisdictional claims in Occelli C, Feral-Pierssens AL, Truchot J, Ortega M, et al. Prognostic accuracy of
published maps and institutional affiliations. Sepsis-3 criteria for in-hospital mortality among patients with suspected
infection presenting to the emergency department. JAMA. 2017;317(3):301–8.
Author details 18. Chen YX, Wang JY, Guo SB. Use of CRB-65 and quick Sepsis-related organ
1
Department of Clinical Pharmacy, University of Southern California School of failure assessment to predict site of care and mortality in pneumonia
Pharmacy, 1985 Zonal Ave, Los Angeles, CA 90089, USA. 2Department of patients in the emergency department: a retrospective study. Crit Care.
Pharmacy, Huntington Hospital, 100 W. California Blvd, Pasadena 91105, USA. 2016;20(1):167.
Minejima et al. BMC Infectious Diseases (2019) 19:149 Page 9 of 9

19. Wang JY, Chen YX, Guo SB, Mei X, Yang P. Predictive performance of quick
Sepsis-related organ failure assessment for mortality and ICU admission in
patients with infection at the ED. Am J Emerg Med. 2016;34(9):1788–93.
20. Besen BA, Romano TG, Nassar AP, Taniguchi LU, Azevedo LC, Mendes PV,
Zampieri FG, Park M. Sepsis-3 definitions predict ICU mortality in a low-
middle-income country. Ann Intensive Care. 2016;6(1):107.
21. Kim M, Ahn S, Kim WY, Sohn CH, Seo DW, Lee YS, Lim KS. Predictive
performance of the quick sequential organ failure assessment score as a
screening tool for sepsis, mortality, and intensive care unit admission in
patients with febrile neutropenia. Support Care Cancer. 2017;25(5):1557–62.
22. April MD, Aguirre J, Tannenbaum LI, Moore T, Pingree A, Thaxton RE,
Sessions DJ, Lantry JH. 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. 2017;52(5):622–31.
23. Giamarellos-Bourboulis EJ, Tsaganos T, Tsangaris I, Lada M, Routsi C,
Sinapidis D, Koupetori M, Bristianou M, Adamis G, Mandragos K, et al.
Validation of the new Sepsis-3 definitions: proposal for improvement in
early risk identification. Clin Microbiol Infect. 2017;23(2):104–9.
24. Williams JM, Greenslade JH, McKenzie JV, Chu K, Brown AF, Lipman J.
Systemic inflammatory response syndrome, quick sequential organ function
assessment, and organ dysfunction: insights from a prospective database of
ED patients with infection. Chest. 2017;151(3):586–96.
25. Raith EP, Udy AA, Bailey M, McGloughlin S, MacIsaac C, Bellomo R, Pilcher
DV. (CORE) AaNZICSACfOaRE: prognostic accuracy of the SOFA score, SIRS
criteria, and qSOFA score for in-hospital mortality among adults with
suspected infection admitted to the intensive care unit. JAMA. 2017;317(3):
290–300.
26. Churpek MM, Snyder A, Han X, Sokol S, Pettit N, Howell MD, Edelson DP.
Quick Sepsis-related organ failure assessment, systemic inflammatory
response syndrome, and early warning scores for detecting clinical
deterioration in infected patients outside the intensive care unit. Am J
Respir Crit Care Med. 2017;195(7):906–11.
27. Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach H, Opal SM, Sevransky
JE, Sprung CL, Douglas IS, Jaeschke R, et al. Surviving sepsis campaign:
international guidelines for management of severe sepsis and septic shock:
2012. Crit Care Med. 2013;41(2):580–637.
28. Soriano A, Marco F, Martínez JA, Pisos E, Almela M, Dimova VP, Alamo D,
Ortega M, Lopez J, Mensa J. Influence of vancomycin minimum inhibitory
concentration on the treatment of methicillin-resistant Staphylococcus
aureus bacteremia. Clin Infect Dis. 2008;46(2):193–200.
29. Vincent JL, Martin GS, Levy MM. qSOFA does not replace SIRS in the
definition of sepsis. Crit Care. 2016;20(1):210.
30. Cosgrove SE, Sakoulas G, Perencevich EN, Schwaber MJ, Karchmer AW,
Carmeli Y. Comparison of mortality associated with methicillin-resistant and
methicillin-susceptible Staphylococcus aureus bacteremia: a meta-analysis.
Clin Infect Dis. 2003;36(1):53–9.
31. Joo J, Yamaki J, Lou M, Hshieh S, Chu T, Shriner KA, Wong-Beringer A. Early
response assessment to guide management of methicillin-resistant
Staphylococcus aureus bloodstream infections with vancomycin therapy.
Clin Ther. 2013;35(7):995–1004.
32. Rose WE, Eickhoff JC, Shukla SK, Pantrangi M, Rooijakkers S, Cosgrove SE,
Nizet V, Sakoulas G. Elevated serum interleukin-10 at time of hospital
admission is predictive of mortality in patients with Staphylococcus aureus
bacteremia. J Infect Dis. 2012;206(10):1604–11.
33. Klevens RM, Morrison MA, Nadle J, Petit S, Gershman K, Ray S, Harrison LH,
Lynfield R, Dumyati G, Townes JM, et al. Invasive methicillin-resistant
Staphylococcus aureus infections in the United States. JAMA. 2007;298(15):
1763–71.
34. Laupland KB, Ross T, Gregson DB. Staphylococcus aureus bloodstream
infections: risk factors, outcomes, and the influence of methicillin resistance
in Calgary, Canada, 2000-2006. J Infect Dis. 2008;198(3):336–43.