T i m i n g a n d Sp e c t r u m o f

A n t i b i o t i c Tre a t m e n t f o r
S u s p e c t e d Se p s i s a n d S e p t i c S h o c k
Why so Controversial?

Theodore R. Pak, MD, PhDa,b, Chanu Rhee, MD, MPH

b,c
,
Michael Klompas, MD, MPHb,c,*

KEYWORDS
 Sepsis  Evidence-based medicine  Retrospective studies  time-to-antibiotics
 antibiotic stewardship  quality improvement

KEY POINTS
 A third or more of patients treated for possible sepsis or septic shock have noninfectious
conditions or nonbacterial infections.
 The literature on the association between time-to-antibiotics and mortality is almost exclu-
sively observational and at high risk of bias. Sources of bias include failure to differentiate
between sepsis and septic shock, insufficient adjustment for potential confounders, and
blending together the high increases in mortality associated with very long delays until an-
tibiotics with the small or absent effects associated with short delays until antibiotics, thus
generating a misleading impression that every hour until antibiotics increases mortality
and does so equally.
 Choosing appropriate empiric antibiotics for patients with possible sepsis or septic shock
is a balancing act: both failure to cover the active pathogen and treating with unnecessar-
ily broad regimens are associated with increased mortality.
 Clinicians are advised to tailor the timing and breadth of antibiotics for each patient with
possible sepsis and septic shock to each individual’s likelihood of infection, risk factors for
resistant pathogens, and severity of illness.

a
Division of Infectious Diseases, Massachusetts General Hospital, 55 Fruit Street, Cox Building,
Suite #515, Boston, MA 02114, USA; b Department of Population Medicine, Harvard Medical
School/Harvard Pilgrim Health Care Institute, 401 Park Drive, Suite 401 East, Boston, MA 02215,
USA; c Division of Infectious Diseases, Brigham and Women’s Hospital, 75 Francis Street, Bos-
ton, MA 02115, USA
* Corresponding author. 401 Park Drive, Suite 401 East, Boston, MA 02215.
E-mail address: mklompas@bwh.harvard.edu

Infect Dis Clin N Am 36 (2022) 719–733

https://doi.org/10.1016/j.idc.2022.08.001 id.theclinics.com
0891-5520/22/ª 2022 Elsevier Inc. All rights reserved.

Descargado para Biblioteca Medica Hospital México (bibliomexico@gmail.com) en National Library of Health
and Social Security de ClinicalKey.es por Elsevier en diciembre 02, 2022. Para uso personal exclusivamente. No
se permiten otros usos sin autorización. Copyright ©2022. Elsevier Inc. Todos los derechos reservados.
720 Pak et al

BACKGROUND

Sepsis, defined as a dysregulated host response to infection leading to acute organ

dysfunction, affects millions of patients per year. It is associated with 850,000 visits
to US emergency departments (EDs) annually1 and one-third of US hospitalizations
that end in death or discharge to hospice.2 Guidelines on sepsis management have
long emphasized that early recognition and treatment are key to lowering mortality.
Many observational studies have attempted to quantify the association between
delays in antibiotics and mortality risk. One of the earliest studies analyzed 2,731
cases of septic shock and estimated that mortality increased by 7.6% for every
hour until effective antimicrobials were administered after the onset of recurrent
or persistent hypotension.3 Subsequent studies have not replicated as dramatic
a result,4 but associations of about 1% absolute increase in mortality risk per
hour have been reported by several large studies that included tens of thousands
of patients.5–7
Based in part upon these data, the Surviving Sepsis Campaign (SSC), the Centers
for Medicaid and Medicare Services (CMS), the New York State Department of Health,
and other state regulators have promoted 1-h or 3-h bundles of care for sepsis that
include requirements to give empiric broad-spectrum antibiotics within brief time win-
dows.8–11 In 2016, SSC guidelines reaffirmed a target of less than 1 h to deliver anti-
biotics for all patients with sepsis (per Sepsis-3 definitions) or septic shock.10,12,13 The
Severe Sepsis and Septic Shock Management (SEP-1) bundle by CMS was imple-
mented in 2015 and modeled on SSC guidelines.14 It requires lactate measurements,
blood cultures, and broad-spectrum antibiotics within 3 h of meeting specific physio-
logic criteria for severe sepsis with or without shock (per Sepsis-2 definitions).
These recommendations stirred considerable controversy. The Infectious Diseases
Society of America (IDSA), for example, withheld their endorsement from the 2016
SSC guidelines.15 One of their primary concerns was the SSC guidelines’ failure to
acknowledge the high levels of uncertainty inherent to the diagnosis of sepsis. Up
to 40% of patients treated with antibacterials for possible sepsis turn out to have
noninfectious conditions or nonbacterial infections.16,17 These patients risk suffering
the potential harms of antibacterial treatments without their potential benefits. These
potential harms are not insignificant; a growing body of literature associates unneces-
sary antibiotics and unnecessarily broad antibiotics with higher rates of Clostridioides
difficile infections, acute kidney injury, selection for resistant pathogens, and in some
studies higher mortality rates.18–20 IDSA feared that 1-h or even 3-h time-to-treatment
requirements would perpetuate or exacerbate the problem of overtreatment given the
difficulty of establishing a clear diagnosis within this timeframe, and thus inadvertently
cause more harm than benefit to patients.
Moreover, IDSA noted substantial weaknesses in the data being used to support
sepsis management bundles and their strict time-to-antibiotic deadlines, finding
that nearly all studies on the association between time-to-antibiotics and mortality
were at substantial risk of bias.4,15 As such, the premise driving strict time-to-
treatment mandates was potentially flawed. Finally, the 2016 SSC guidelines equated
patients with suspected sepsis and suspected septic shock for the purposes of recom-
mending the 1 h time-to-treatment, even though these groups have very different as-
sociations between time-to-antibiotics and mortality.7,9 Studies published after the
2016 guidelines were released have reported that the association between time-to-
antibiotics and mortality for sepsis without shock is much weaker than that for septic
shock, and when present, the signal suggesting increased mortality in sepsis without
shock only becomes apparent after delays of three to 5 h, not 1 h.21,22

Descargado para Biblioteca Medica Hospital México (bibliomexico@gmail.com) en National Library of Health
and Social Security de ClinicalKey.es por Elsevier en diciembre 02, 2022. Para uso personal exclusivamente. No
se permiten otros usos sin autorización. Copyright ©2022. Elsevier Inc. Todos los derechos reservados.
 Timing and breadth of antibiotics for possible sepsis 721

Challenges in Interpreting Evidence on Antibiotic Delays and Associated Outcomes

How has the evidence regarding timing of antibiotics for sepsis patients led to such a
wide range of conclusions? First, most of the evidence available for this question is
from retrospective observational studies, given the difficulty of performing prospective
trials that randomize patients to different antibiotic strategies during emergency care.
We are only aware of two randomized controlled trials (RCTs) specifically examining
the impact of faster time-to-antibiotics for sepsis on patient outcomes. One was a ran-
domized trial of antibiotics in the prehospital setting versus the ED for patients with
suspected sepsis.23 Despite a 96-min difference in median time to antibiotics between
groups there were no significant differences in mortality. The second study was a ran-
domized trial comparing a multifaceted set of interventions focusing on reducing time-
to-antibiotics versus conventional continuous medical education.24 Unfortunately, this
study did not achieve a significant difference in time-to-antibiotics between groups
and thus was uninformative on this specific question. All the other evidence for the ex-
istence of an association between time-to-antibiotics and increased mortality is from
observational studies, which carry inherent risks of confounding and are more difficult
to interpret as causal evidence.
Second, most early observational studies (particularly those cited by SSC in 2016)
were limited to intensive care unit (ICU) populations and thus enriched with a high pro-
portion of patients with septic shock, potentially introducing unintended bias when
formulating guidelines for all sepsis patients, including those without shock. Most of
these studies did not perform subgroup analyses comparing sepsis with and without
shock, and the few studies that included subgroup analyses found smaller associa-
tions or no significant association for sepsis without shock.21 For example, in
35,000 presentations of sepsis to California EDs, there was an absolute increase in
mortality of 1.8% (95% confidence interval [CI], 0.8%–3.0%; adjusted odds ratio
[aOR], 1.14; aOR 95% CI, 1.06 to 1.23) associated with each hour of delay in antibi-
otics for patients with septic shock, whereas for severe sepsis patients without shock,
this association shrank to 0.4% (95% CI, 0.1%–0.8%; aOR, 1.07; aOR 95% CI, 1.01–
1.24).7 Similarly, a study of 49,331 patients triggering sepsis protocols in New York
EDs found a significant association between in-hospital death and hourly delays in
completing a bundle that included antibiotics among patients that required vasopres-
sors (aOR, 1.07; 95% CI, 1.05–1.09), but the association was not significant for pa-
tients not requiring vasopressors (aOR, 1.01; 95% CI, 0.99–1.04).9 So too, in a
study of 10,811 adult ED patients in Utah with sepsis predominantly without shock
(92% did not have septic shock) there was no impact on 30-day mortality until 5 h
had elapsed from ED triage until antibiotics were given.5 And in a study of 74,114 adult
ED patients with suspected infection without shock, there was an association between
time-to-antibiotics and progression to shock (aOR 1.03 per hour, 95% CI 1.02–1.04)
but no association between time-to-antibiotics and hospital mortality (aOR 1.02,
95% CI 0.99–1.04).25 Most recently, a prospective non-randomized study of 178 Jap-
anese ICU patients with sepsis found an association between decreased in-hospital
mortality and completion of a 1-h bundle including antibiotics (aOR, 1.28; 95% CI,
1.04–1.57).26 But once more, in risk-adjusted subgroup analyses comparing patients
with and without shock, the association between in-hospital mortality and hourly
bundle delays was only significant for the 55% of patients with septic shock (aOR,
1.25; 95% CI, 1.03–1.52), there was no significant association for the 45% of patients
without shock (aOR, 1.17; 95% CI, 0.73–1.87).26
Third, most studies have linearized the relationship between time-to-antibiotics and
mortality and reported a single blended estimate of the effect of each hour interval until

Descargado para Biblioteca Medica Hospital México (bibliomexico@gmail.com) en National Library of Health
and Social Security de ClinicalKey.es por Elsevier en diciembre 02, 2022. Para uso personal exclusivamente. No
se permiten otros usos sin autorización. Copyright ©2022. Elsevier Inc. Todos los derechos reservados.
722 Pak et al

antibiotics. This is highly problematic because inspection of the data underlying these
studies shows clear nonlinear relationships.21 Generally, very long intervals until anti-
biotics (typically >5 or 6 h) are associated with increased mortality but shorter intervals
are not. Blending the effects of long intervals with the effects of shorter intervals gives
the misimpression that each hour until treatment is associated with a clear and consis-
tent increase in mortality. A very recent analysis of the impact of time-to-antibiotics in
4,792 patients with sepsis, for example, found that each additional hour until time-to-
antibiotics was associated with a statistically significant 0.42% increase in 28-day
mortality when combining data from all patients with time-to-antibiotic intervals
ranging from <1 h to 48 h.27 When the investigators compared intervals of 1 to 3 h
versus <1 h and intervals of 3 to 6 h versus <1 h, however, there was no significant
difference in 28-day mortality rates; only intervals of >6 h were associated with signif-
icantly higher mortality rates.27
Fourth, the timing of antibiotic administration in real-world practice is not random.
Sicker patients with more obvious presentations of sepsis tend to be treated sooner.
Confounders must be carefully controlled to disentangle the effects of time-to-
antibiotics from patients’ underlying mortality risk attributable to their preexisting
comorbidities, presenting signs and symptoms, and severity of illness, all of which
may influence the timing of antibiotics. Clinicians tend to order antibiotics more imme-
diately for more ill-appearing patients and for those with higher baseline mortality risk,
whether because of their demographics, medical history, or other data available at
presentation, and for patients with more obvious clinical signs of infection (eg, fe-
ver).28,29 Consequently, even in the largest observational studies, results change dras-
tically when including different confounders in the model, going so far as to shift
negative associations (earlier antibiotics paradoxically associated with greater mortal-
ity) to positive associations (earlier antibiotics associated with lower mortality).7
Despite the clear importance of rigorously adjusting for potential confounders, studies
vary widely in the number, breadth and granularity of the variables they use for this
process; some do not even include age or comorbid diseases.21 When medical history
is included as a confounder, it is via simplifying rubrics such as Charlson or Elixhauser
indices,30,31 which conflate conditions with vastly different severities (eg, “mild inter-
mittent asthma, uncomplicated” and “chronic obstructive pulmonary disease” incur
equivalent risk in the Elixhauser Comborbidity Index).32 Furthermore, these indices
are not comprehensive; they omit key comorbidities (such as cystic fibrosis and
some congenital immunodeficiencies) that may be uncommon but nonetheless clearly
increase the risk of sepsis and death.33 Few of the studies incorporate patients’ prior
infection-related data (culture results, recent antibiotic exposures, and recent infec-
tions) as confounders, even though these are often important determinants of clini-
cians’ prescribing behavior and likely correlate with patients’ likelihood of true
sepsis and risk of death.

2016 to 2020: Conflicting Guidelines on a Time-To-Antibiotics Target

The net effect of these weaknesses and inconsistencies in source evidence as well
as the controversy stirred by various sepsis guidelines and mandates is that guide-
lines on expected antibiotic timing for sepsis have changed frequently in the last
6 years, with contradictory recommendations between different professional soci-
eties. As described above, IDSA did not endorse the SSC 2016 guidelines, citing
concerns regarding the promotion or perpetuation of antibiotic overuse, overdiagno-
sis of sepsis, weaknesses of the data supporting time-to-antibiotic goals, conflation
of suspected sepsis and septic shock, and disagreement on specific recommenda-
tions regarding antimicrobial selection, blood cultures, procalcitonin, and treatment

Descargado para Biblioteca Medica Hospital México (bibliomexico@gmail.com) en National Library of Health
and Social Security de ClinicalKey.es por Elsevier en diciembre 02, 2022. Para uso personal exclusivamente. No
se permiten otros usos sin autorización. Copyright ©2022. Elsevier Inc. Todos los derechos reservados.
 Timing and breadth of antibiotics for possible sepsis 723

duration.15 The concurrent release of the Third International Consensus definitions

for sepsis (“Sepsis-3”) sowed further confusion, partly because of terminology
changes (redefining severe sepsis as sepsis)34 and partly because different
groups variably adopted the new definitions. SSC 2016 did adopt Sepsis-3
terminology but cited data using the old definitions to support their recommenda-
tions. Meanwhile, CMS continues to use the old definitions for their SEP-1 reporting
requirement.
In 2018, the SSC released a “bundle update” combining its 3 h and 6 h bundles into
a single 1 h bundle, which included lactate levels, blood cultures, volume resuscita-
tion, vasopressors if fluid-refractory, and broad-spectrum antibiotics, all within 1 h
of ED triage.35 As more than 80% of sepsis is diagnosed on admission and initially
managed in the ED,1,2 concerns over the lack of evidence to support such a significant
change to emergency medicine workflow were quickly raised.36–38 In response to
these concerns, the Society of Critical Care Medicine (SCCM) and American College
of Emergency Physicians (ACEP) published a joint statement saying: “We recommend
that hospitals not implement the Hour-1 Bundle in its present form in the United States
at this time,” whereas the authors of the SSC bundle update acknowledged that it was
intended as a facilitative tool and not as a potential quality indicator.39 “Time-zero” for
bundle initiation was also revised from ED triage to time of sepsis recognition. Collec-
tively, the net result was substantial confusion and/or disagreement among frontline
providers, hospitals, and regulators on which components of the 2016 guidelines
and the 2018 bundle update were in fact appropriate for implementation. Finally, in
2020, IDSA recommended revisions to the SEP-1 bundle created by CMS, again citing
the lack of evidence supporting early antimicrobials for suspected sepsis without
shock, concerns over the potential to drive antibiotic overuse, and the high complexity
of SEP-1’s “time-zero” definition.22

The Optimal Breadth of Broad-Spectrum Antibiotics

A parallel controversy has been determining the appropriate spectrum of empiric anti-
biotic coverage for sepsis, given similar mismatches between evolving guidelines and
the level of evidence supporting them. In practice, just as important as when antibi-
otics are needed for sepsis is the corollary question of which antibiotics must be given.
Antibiotic coverage is ideally tailored to causative pathogens and their susceptibility
profiles, but these data are rarely available when antibiotics are started.
In 2016, SSC specified a number of overlapping terms for antimicrobial selection in
sepsis (“empiric”, “targeted/definitive”, “broad-spectrum”, “multidrug”, and “combi-
nation”).10 IDSA felt that these terms were confusing, because differences between
most of them would not be obvious even to experts in the field, and the terms were
ultimately used inconsistently within the guidelines.15 For example, SSC 2016 recom-
mended “combination” therapy for septic shock—which they defined as the use of
two or more antibiotics active against the patient’s causative pathogen—“for several
days.” This recommendation was controversial for two reasons. First, many clinicians
use “combination therapy” to refer to the empiric use of two agents active against
gram-negative bacteria in order to increase the probability that the patient receives
at least one agent active against their causative pathogen(s), that is, to broaden
coverage. On the contrary, SSC defined it as two agents active against the same path-
ogen “to accelerate pathogen clearance rather than broaden antimicrobial
coverage.”15 Second, the evidence supporting sustained use of two active agents
for septic shock is very weak. Two large and rigorous randomized trials fail to support
this approach, one focused on patients with sepsis in general40 and one focused on
patients with VAP.41 Similarly, a meta-analysis of 13 randomized trials found no

Descargado para Biblioteca Medica Hospital México (bibliomexico@gmail.com) en National Library of Health
and Social Security de ClinicalKey.es por Elsevier en diciembre 02, 2022. Para uso personal exclusivamente. No
se permiten otros usos sin autorización. Copyright ©2022. Elsevier Inc. Todos los derechos reservados.
724 Pak et al

difference between empiric mono-versus combination antibiotic therapy on length-of-

stay, mortality, or any other objective outcome among patients with sepsis.42
There is increasing evidence that the use of overly broad empiric antibiotics is not
only unnecessary but may in fact cause harm.43 A retrospective, propensity-
weighted study of methicillin-resistant Staphylococcus aureus (MRSA) coverage in
88,605 Veterans’ Affairs (VA) patients treated for community-acquired pneumonia
(CAP) showed that the addition of anti-MRSA therapy was associated with increased
adjusted risk of 30-day mortality and secondary infections including Clostridioides
difficile, vancomycin-resistant Enterococcus (VRE), and gram-negative bacilli.44
Counterintuitively, even when restricting to patients with MRSA found on cultures,
there was no mortality benefit to adding anti-MRSA therapy, belying the complexity
of diagnosing certain infections (did these patients have true pneumonias or
mimicking conditions?) and the inconsistent relationships between antibiotic utiliza-
tion and outcomes. Similar retrospective studies of 2,198 CAP patients in four US
EDs, and a second VA cohort of 15,071 nonsevere pneumonia patients, showed
increased mortality associated with broad-spectrum antibiotics.45,46 A prospective
cohort study of 303 ICU patients at risk for multidrug-resistant (MDR) pneumonia
noted that guideline-directed (broader) therapy was associated with increased
mortality.47
Regarding infections other than pneumonia, a 2-year quasi-experimental before-
and-after observational cohort study of surgical ICU admissions at one hospital found
that implementation of a conservative strategy for antimicrobial treatment, where an-
tibiotics were started only after cultures or gram stain supported an infection, was
associated with reduced mortality when compared with an aggressive strategy where
broad empiric antimicrobials were started upon clinical suspicion and then stopped if
cultures and gram stains were negative.48 A retrospective study of 17,430 patients
admitted to 104 US hospitals with sepsis and positive cultures (primarily blood, urine,
or respiratory) found that unnecessarily broad empiric antibiotics, defined as including
coverage of MRSA, VRE, Pseudomonas, or other MDR organisms when none of these
were ultimately isolated on cultures, were independently associated with increased in-
hospital mortality after adjusting for baseline characteristics and illness severity.18
Identifying causes of these increases in mortality was outside the scope of these
studies, but antibiotics have many known risks, including acute kidney injury, liver
toxicity, cytopenias, selection for drug-resistant flora, mitochondrial toxicity, and
disruption of the gut microbiome (an important modulator of the immune system).
On the contrary, empiric treatment of sepsis with a regimen that fails to cover the
causative pathogen is also independently associated with increased mortality. Studies
on this are limited to those with positive cultures so that antibiotic appropriateness can
be determined, and therefore may not be representative of all suspected sepsis, for
which a substantial fraction have negative cultures. Acknowledging this limitation,
the aforementioned study of 17,430 patients with positive cultures did also show an
association between inadequate empiric coverage and higher in-hospital mortality.18
Similarly, a retrospective study of 21,608 patients from 131 US hospitals with blood-
stream infections (about 20% of all sepsis cases)2 found that among the 19% that
received an inadequate initial regimen (failure to cover the bloodstream isolate), there
was an independent association with increased mortality that was not affected by the
presence or absence of drug resistance, sepsis, or septic shock.49
Therefore, the optimal “breadth” of broad-spectrum antibiotics for sepsis faces a
Goldilocks problem50: evidence supports harms of both overly broad and overly nar-
row therapy, so in the face of diagnostic uncertainty, practical compromises have to
be struck. As with time-to-antibiotics, the high rate of overdiagnosis of sepsis needs

Descargado para Biblioteca Medica Hospital México (bibliomexico@gmail.com) en National Library of Health
and Social Security de ClinicalKey.es por Elsevier en diciembre 02, 2022. Para uso personal exclusivamente. No
se permiten otros usos sin autorización. Copyright ©2022. Elsevier Inc. Todos los derechos reservados.
 Timing and breadth of antibiotics for possible sepsis 725

to be taken into account, as well as the variability in the signal between time-to-
appropriate-antibiotics and mortality depending on patients’ clinical syndrome and
severity-of-illness. For conditions such as septic shock with less room for error, it is
appropriate to use broader antibiotics up front, whereas ideally a stable patient with
lower probability of infection should receive narrower or no antibiotics until a causative
pathogen is confirmed.

2021 and Beyond: Toward Consensus in Sepsis Antimicrobial Guidelines

In July 2021, ACEP released its own set of guidelines on early sepsis care in the ED,
which were endorsed by IDSA, SCCM, and 10 other professional societies. Taking a
different tone from SSC 2016, ACEP’s guidelines acknowledged the “inherent diffi-
culty in establishing the early diagnosis of sepsis” in the ED.51 Although prompt anti-
biotic administration is encouraged once sepsis is diagnosed, the authors concluded
“there are insufficient data to recommend a specific time threshold for administration
of antibiotics.”51 Regarding antibiotic spectrum, the guidelines avoided endorsing
two-agent or “combination” therapy, instead recommending “initiation of broad-
spectrum antibiotics with activity against gram-negative and gram-positive bacteria
according to local susceptibility patterns.”51
In November 2021, SSC published a new version of their guidelines that responded
to much of the feedback from IDSA, ACEP, and other professional societies on anti-
microbial management recommendations.52 This update was endorsed by IDSA,
signaling a shift back toward consensus. Although the general outline of
antimicrobial-related recommendations remained similar, the evidence for many rec-
ommendations was downgraded from “moderate” quality evidence to “low” or “very
low”, and several crucial changes were made.
First, the new guidelines emphasize “the challenge of diagnostic uncertainty early in
a patient’s presentation,” and now advise clinicians to “stratify antimicrobial timing
recommendations based on the likelihood of sepsis and presence of shock.”53 Spe-
cifically, the cohort targeted for antimicrobials within 1 h has been narrowed from all
patients with suspected sepsis or septic shock to just those patients with septic shock
and possible sepsis “where likelihood of infection is high.” For patients with an inter-
mediate likelihood of infection, the guidelines now suggest a time-limited course of
rapid investigation to determine a diagnosis, concluding with the administration of an-
timicrobials within 3 h only if concern for infection persists; the guidelines categorize
evidence for this timeframe as weak instead of strong. Finally, the 2021 guidelines
define an entirely new group of patients, “adults with low likelihood of infection and
without shock,” who explicitly do not need antimicrobials unless and until they are
proven to have an infection. This recommendation will hopefully reinforce that many
patients initially suspected of sepsis do not have a bacterial infection, and therefore
it can be appropriate to do a diagnostic workup before starting antibiotics.
Second, a subtle but important change was added to the language of the SSC 2021
guidelines regarding time to antibiotics. The 2016 guidelines used durations of “within
X hours” with the starting point left unnamed, leaving open an implication that this
referred to the onset of sepsis. As the time of sepsis onset is not known for many pa-
tients entering the hospital, and could have occurred hours to days before admission,
this is difficult to incorporate into metrics for hospital performance or observational
studies, setting aside the minority case of hospital-onset sepsis.2 Therefore, when
creating bundles and cohort definitions, SSC, CMS, and sepsis researchers have all
chosen different “time-zero” definitions—ranging from ED triage time in SSC’s 2018
bundle update,35 a combination of documentation and physiologic data within a
sliding 6 h window in SEP-1,8 and variations on these themes among the time-to-

Descargado para Biblioteca Medica Hospital México (bibliomexico@gmail.com) en National Library of Health
and Social Security de ClinicalKey.es por Elsevier en diciembre 02, 2022. Para uso personal exclusivamente. No
se permiten otros usos sin autorización. Copyright ©2022. Elsevier Inc. Todos los derechos reservados.
726 Pak et al

antibiotics studies.21 By contrast, the SSC 2021 guidelines now explicitly refer to the
clock on antibiotics starting “from the time when sepsis was first recognized,”
acknowledging the reality of a clinician needing time to evaluate the patient and
consider other diagnoses before management can begin. Some time-to-antibiotics
studies have also tried to use clinical recognition as their time-zero, although it is a
labor-intensive and subjective concept to abstract from charts. Even time-zero defini-
tions based on vital signs and laboratory measures are subject to high interobserver
variability. Manual abstractors of SEP-1’s time-zero, for example, disagree over half
of the time on when time-zero occurred.54 Nevertheless, by including a conceptual
time-zero (sepsis recognition) for the first time, the 2021 guidelines helped narrow
down potential interpretations of time-to-antibiotics and may encourage future re-
searchers to build consensus on an operational definition.
Finally, the SSC 2021 guidelines made significant changes to recommendations on
spectrum of coverage. The artificial terminology of “empiric”, “targeted/definitive,”
“combination,” “multidrug,” etc., was eliminated. Instead, the new guidelines frame
spectrum of coverage around specific categories of pathogens that should be consid-
ered when starting antimicrobials. For instance, there is a new best practice statement
on including empiric MRSA coverage for sepsis only when the patient is judged to
have “high risk of MRSA infection,” and a similar “weak” suggestion for antifungal
coverage only if the patient is at “high risk for fungal infection.” There still is a
“weak” suggestion for using two antimicrobial agents (previously called “combination
therapy”) to cover gram-negative bacteria, but instead of applying to all patients with
suspected sepsis, it is aimed at patients specifically with high risk for MDR organisms.
There is a parallel recommendation to narrow this regimen to a single agent once the
causative pathogen(s) and susceptibilities are known, and the evidence categorization
for these suggestions has been downgraded to “very low quality.” Overall, although
evidence ratings for the new antimicrobial spectrum recommendations are modest,
by framing them around categories with specific microbiologic correlates such as
“MRSA coverage,” they have increased clarity and adaptability to local data, such
as the use of community prevalence of MRSA and MDR organisms to help determine
the appropriateness of broader empiric coverage.

DISCUSSION

Overall, the 2021 SSC and ACEP guidelines have made significant progress in build-
ing consensus recommendations for antimicrobial timing and breadth, balancing the
importance of rapid treatment for potentially septic patients with bona fide infections
against the potential harms associated with antibiotics for those who are not
infected. By casting a more critical light upon the limitations of existing evidence,
they also forecast several areas where higher quality evidence is needed. For
instance, although the 2021 SSC guidelines now emphasize concepts intrinsic to
the early clinical evaluation and management of sepsis, including “sepsis where
the likelihood of infection is high,” “the time when sepsis was first recognized,”
“high risk for MDRs,” and “high risk for fungal infection,” the onus is on researchers
to characterize these probabilities and timepoints more concretely so that clinicians
may leverage them for decision-making.
The fundamental challenge in sepsis management remains: how might we better
identify which patients along the spectra of sepsis risk and severity benefit most
from earlier antibiotics, and which merit broader versus narrower coverage? One
approach might be to raise the standard for large observational studies on time-to-
antibiotics, which are increasingly robust in terms of sample size but have been less

Descargado para Biblioteca Medica Hospital México (bibliomexico@gmail.com) en National Library of Health
and Social Security de ClinicalKey.es por Elsevier en diciembre 02, 2022. Para uso personal exclusivamente. No
se permiten otros usos sin autorización. Copyright ©2022. Elsevier Inc. Todos los derechos reservados.
 Timing and breadth of antibiotics for possible sepsis 727

rigorous about the methods used to control for confounders. Future studies can and
should expand their input variables to incorporate more of the richness and breadth of
the data available in electronic medical records (EMRs). For example, studies rarely
use data from prior encounters to formulate confounders and the likelihood for a pa-
tient’s subsequent presentation of suspected sepsis. Incorporating more detailed
measures of current and prior clinical markers of organ dysfunction as well as their tra-
jectories could improve differentiation and quantification of acute changes catalyzed
by infection vs chronic underlying conditions—something that clinicians do routinely
when evaluating patients, but has been left out of cohort definitions and adjustment
models underlying nearly all time-to-antibiotics studies. In addition, machine-
learning techniques may disentangle the vast heterogeneity in EMR data from sepsis
populations by identifying distinct sub-phenotypes of sepsis using clustering and
similar techniques.55–59 Future work could integrate machine-learning sub-phenotyp-
ing with data on antibiotic timing, breadth, and outcomes to generate predictive
models of the sub-phenotypes that most likely benefit from earlier antibiotics or
broader-spectrum coverage.
Translational research on rapid diagnostics for sepsis may also permit a more
tailored approach. After decades of research, likely due to the biological heterogeneity
of the syndrome,60 there still are no gold-standard diagnostic tests or biomarkers for
sepsis, which remains a clinical diagnosis based on the same kinds of data that were
first used to describe the “systemic inflammatory response syndrome” in 1992.61
Nonetheless, earlier accurate identification of sepsis caused by bacteria and their sus-
ceptibility profiles could both accelerate the time-to-antibiotics and mitigate antibiotic
overuse. Over 200 biomarkers for sepsis have been investigated,62 with C-reactive
protein and procalcitonin among the most popular, but as these two markers are
both upregulated in noninfectious inflammation, neither is sensitive nor specific
enough to reliably guide initial sepsis treatment. Procalcitonin has shown more prom-
ise as a longitudinal marker of antibiotic response in serious infections, permitting
shortened antibiotic courses that are associated with lower in-hospital mortality in
meta-analyses.63,64 Of the other biomarkers, only 26 have been investigated in sam-
ples of at least 300 patients, and it seems increasingly unlikely that any single host
biomarker will “break through” and significantly alter early sepsis diagnosis or
management.62
Other researchers have focused on molecular diagnostics targeting bacterial nucleic
acids, such as polymerase chain reaction (PCR) of 16S ribosomal RNA and microar-
rays. In principle, these may yield equally accurate and slightly faster results compared
with blood cultures,65,66 but their utility is less clear for non-bloodstream infections.
Similarly, mass spectrometry is increasingly integrated into culturing workflows,
providing the greatest speed benefit when identifying slow growing and less common
organisms.67 Recent methods have attempted to decrease turnaround time for mass
spectrometry on blood culture isolates by replacing subculturing with centrifugation
and protein extraction, although this still requires waiting for growth in the initial blood
culture bottle.68,69 Whole-genome sequencing of pathogens provides high-resolution
data useful for genotyping and molecular epidemiology, but is unlikely to usurp
culturing and mass spectrometry on turnaround time for species identification, except
for pathogens that are difficult to culture.70 Researchers have also directed omics as-
says toward the host immune response, finding potential sepsis sub-phenotypes in
meta-analysis of microarray data from 600 patients,71 and more recently, a unique
CD141 monocyte state distinguishing 29 sepsis patients from 36 controls using
single-cell RNA sequencing.72 New diagnostics leveraging these insights are still
undergoing development and validation for potential clinical applications.

Descargado para Biblioteca Medica Hospital México (bibliomexico@gmail.com) en National Library of Health
and Social Security de ClinicalKey.es por Elsevier en diciembre 02, 2022. Para uso personal exclusivamente. No
se permiten otros usos sin autorización. Copyright ©2022. Elsevier Inc. Todos los derechos reservados.
728 Pak et al

SUMMARY

From 2016 to 2020, sepsis guidelines and regulatory mandates encouraged increas-
ingly brief targets—as short as 1 h—for initiating broad-spectrum antimicrobials for
patients with suspected sepsis or septic shock. This sparked considerable contro-
versy due to weaknesses in the underlying evidence and concern that strict antibiotic
deadlines cause inadvertent harm by perpetuating or accelerating overtreatment at
the expense of diagnostic inquiry. A third or more of patients treated for sepsis and
septic shock have noninfectious or nonbacterial conditions. These patients risk harm
from unnecessary antibiotics. New guidelines from both ACEP and SSC in 2021 now
emphasize the importance of tailoring treatment to each patient’s likelihood of infec-
tion, risk for drug-resistant pathogens, and severity of illness. These guidelines will
benefit from future research that raises the standards for evidence derived from
observational studies of time-to-antibiotics and associated outcomes. New diag-
nostics that rapidly quantify the likelihood of infection by bacteria, MDR organisms,
fungi, and other pathogen groups relevant to antimicrobial selection could also have
a major clinical impact, given the emphasis on these concepts in the latest
guidelines.

CLINICS CARE POINTS

 A third or more of patients treated for sepsis turn out to have noninfectious conditions or
nonbacterial infections.
 When evaluating a patient for possible sepsis, tailor the urgency of antibiotics to the
patient’s likelihood of infection and severity of illness.
 In patients with possible septic shock, administer antibiotics immediately. In patients with
less severe illness, undertake a rapid course of diagnostics (eg, lab studies, imaging,
microbiological assays, etc.) and treat for noninfectious possibilities if present (eg, fluids,
diuretics, bronchodilators, heart rate control, etc.). Continually reevaluate the likelihood of
infection based on the results of diagnostic studies and response to treatments for
noninfectious conditions. If high concern for infection persists at 3 h from when sepsis was
first suspected, then administer antimicrobials.
 Choose the spectrum of antibiotics according to severity of illness (as critically ill patients
have less margin for error) and the local distribution of organisms and their resistance
patterns as well as each patient’s individual risk factors for antibiotic-resistant organisms.

DISCLOSURE

T.R. Pak, author, reports grant funding from the National Institute of Allergy and Infec-
tious Diseases (5T32AI007061). C. Rhee and M. Klompas report grant funding from the
Centers for Disease Control and Prevention and Agency for Healthcare Research and
Quality to conduct research related to sepsis. C. Rhee reports royalties from UpTo-
Date, Inc. for chapters related to procalcitonin use. MK reports royalties from UpTo-
Date, Inc. for chapters related to hospital-acquired pneumonia.

REFERENCES

1. Wang HE, Jones AR, Donnelly JP. Revised National Estimates of Emergency
Department Visits for Sepsis in the United States. Crit Care Med 2017;45(9):
1443–9.

Descargado para Biblioteca Medica Hospital México (bibliomexico@gmail.com) en National Library of Health
and Social Security de ClinicalKey.es por Elsevier en diciembre 02, 2022. Para uso personal exclusivamente. No
se permiten otros usos sin autorización. Copyright ©2022. Elsevier Inc. Todos los derechos reservados.
 Timing and breadth of antibiotics for possible sepsis 729

2. Rhee C, Dantes R, Epstein L, et al. Incidence and trends of sepsis in US hospitals

using clinical vs claims data, 2009-2014. JAMA 2017;318(13):1241–9.
3. Kumar A, Roberts D, Wood KE, et al. Duration of hypotension before initiation of
effective antimicrobial therapy is the critical determinant of survival in human sep-
tic shock. Crit Care Med 2006;34(6):1589–96.
4. Sterling SA, Miller WR, Pryor J, et al. The Impact of Timing of Antibiotics on Out-
comes in Severe Sepsis and Septic Shock: A Systematic Review and Meta-Anal-
ysis. Crit Care Med 2015;43(9):1907–15.
5. Door-to-Antibiotic Time and Long-term Mortality in Sepsis. In: Peltan ID,
Brown SM, Bledsoe JR, et al, editors. Chest 2019;155(5):938–46.
6. Ferrer R, Martin-Loeches I, Phillips G, et al. Empiric antibiotic treatment reduces
mortality in severe sepsis and septic shock from the first hour: Results from a
guideline-based performance improvement program. Crit Care Med 2014;
42(8):1749–55.
7. Liu VX, Fielding-Singh V, Greene JD, et al. The Timing of Early Antibiotics and
Hospital Mortality in Sepsis. Am J Respir Crit Care Med 2017;196(7):856–63.
8. Centers for Medicare & Medicaid Services. Hospital Inpatient Specifications
Manuals. 2020. Available at: https://qualitynet.cms.gov/inpatient/specifications-
manuals. Accessed February 8, 2022.
9. 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):2235–44.
10. Rhodes A, Evans LE, Alhazzani W, et al. Surviving Sepsis Campaign: Interna-
tional Guidelines for Management of Sepsis and Septic Shock: 2016. Crit Care
Med 2017;45(3):486–552.
11. Dwyer J. One boy’s death moves state to action to prevent others. The New York
Times; 2012. Available at: https://www.nytimes.com/2012/12/21/nyregion/one-
boys-death-moves-state-to-action-to-prevent-others.html. Accessed February
13, 2022.
12. Seymour CW, Liu VX, Iwashyna TJ, et al. Assessment of Clinical Criteria for
Sepsis: For the Third International Consensus Definitions for Sepsis and Septic
Shock (Sepsis-3). JAMA 2016;315(8):762–74.
13. Dellinger RP, Levy MM, Rhodes A, et al. Surviving sepsis campaign: International
guidelines for management of severe sepsis and septic shock, 2012. Intensive
Care Med 2013;39(2):165–228.
14. Pepper DJ, Jaswal D, Sun J, et al. Evidence underpinning the centers for medi-
care & medicaid services’ severe sepsis and septic shock management bundle
(SEP-1) a systematic review. Ann Intern Med 2018;168(8):1–11.
15. IDSA Sepsis Task Force, Kalil AC, Gilbert DN, Winslow DL, et al. Infectious Dis-
eases Society of America (IDSA) POSITION STATEMENT: Why IDSA Did Not
Endorse the Surviving Sepsis Campaign Guidelines. Clin Infect Dis 2018;
66(10):1631–5.
16. Klein Klouwenberg PMC, Cremer OL, van Vught LA, et al. Likelihood of infection
in patients with presumed sepsis at the time of intensive care unit admission: A
cohort study. Crit Care 2015;19(1):1–8.
17. Shappell CN, Klompas M, Ochoa A, et al. Likelihood of Bacterial Infection in Pa-
tients Treated With Broad-Spectrum IV Antibiotics in the Emergency Department.
Crit Care Med 2021;49(11):e1144–50.
18. Rhee C, Kadri SS, Dekker JP, et al. Prevalence of Antibiotic-Resistant Pathogens
in Culture-Proven Sepsis and Outcomes Associated With Inadequate and Broad-
Spectrum Empiric Antibiotic Use. JAMA Netw Open 2020;3(4):e202899.

Descargado para Biblioteca Medica Hospital México (bibliomexico@gmail.com) en National Library of Health
and Social Security de ClinicalKey.es por Elsevier en diciembre 02, 2022. Para uso personal exclusivamente. No
se permiten otros usos sin autorización. Copyright ©2022. Elsevier Inc. Todos los derechos reservados.
730 Pak et al

19. Tamma PD, Avdic E, Li DX, et al. Association of Adverse Events With Antibiotic
Use in Hospitalized Patients. JAMA Intern Med 2017;177(9):1308–15.
20. Ong DSY, Frencken JF, Klein Klouwenberg PMC, et al. Short-Course Adjunctive
Gentamicin as Empirical Therapy in Patients With Severe Sepsis and Septic
Shock: A Prospective Observational Cohort Study. Clin Infect Dis 2017;64(12):
1731–6.
21. Weinberger J, Rhee C, Klompas M. A critical analysis of the literature on time-to-
antibiotics in suspected sepsis. J Infect Dis 2021;222:S110–8.
22. Rhee C, Chiotos K, Cosgrove SE, et al. Infectious Diseases Society of America
Position Paper: Recommended Revisions to the National Severe Sepsis and Sep-
tic Shock Early Management Bundle (SEP-1) Sepsis Quality Measure. Clin Infect
Dis 2021;72(4):541–52.
23. Alam N, Oskam E, Stassen PM, et al. Prehospital antibiotics in the ambulance for
sepsis: a multicentre, open label, randomised trial. Lancet Respir Med 2018;6(1):
40–50.
24. Bloos F, Rüddel H, Thomas-Rüddel D, et al. Effect of a multifaceted educational
intervention for anti-infectious measures on sepsis mortality: a cluster randomized
trial. Intensive Care Med 2017;43(11):1602–12.
25. Bisarya R, Song X, Salle J, et al. Antibiotic Timing and Progression to Septic
Shock Among Patients in the ED With Suspected Infection. Chest 2022;161(1):
112–20.
26. Umemura Y, Abe T, Ogura H, et al. Hour-1 bundle adherence was associated with
reduction of in-hospital mortality among patients with sepsis in Japan. PLoS One
2022;17(2):1–12.
27. Rüddel H, Thomas-Rüddel DO, Reinhart K, et al. Adverse effects of delayed anti-
microbial treatment and surgical source control in adults with sepsis: results of a
planned secondary analysis of a cluster-randomized controlled trial. Crit Care
2022;26(1):51.
28. Filbin MR, Lynch J, Gillingham TD, et al. Presenting symptoms independently
predict mortality in septic shock: Importance of a previously unmeasured
confounder. Crit Care Med 2018;46(10):1592–9.
29. Henning DJ, Carey JR, Oedorf K, et al. The absence of fever is associated with
higher mortality and decreased antibiotic and IV fluid administration in emer-
gency department patients with suspected septic shock. Crit Care Med 2017;
45(6):e575–82.
30. Van Walraven C, Austin PC, Jennings A, et al. A modification of the elixhauser co-
morbidity measures into a point system for hospital death using administrative
data. Med Care 2009;47(6):626–33.
31. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic
comorbidity in longitudinal studies: development and validation. J Chronic Dis
1987;40(5):373–83.
32. Agency for Healthcare Research and Quality. Elixhauser comorbidity software
refined for ICD-10-CM reference file, v2022.1. Available at: https://www.hcup-
us.ahrq.gov/toolssoftware/comorbidityicd10/CMR-Reference-File-v2022-1.xlsx.
Accessed March 13, 2022.
33. Alrawashdeh M, Klompas M, Simpson SQQ, et al. Prevalence and Outcomes of
Previously Healthy Adults Among Patients Hospitalized With Community-Onset
Sepsis. Chest 2022;1–10. https://doi.org/10.1016/j.chest.2022.01.016.
34. Singer M, Deutschman CS, Seymour CW, et al. The Third International Consensus
Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA 2016;315(8):801–10.

Descargado para Biblioteca Medica Hospital México (bibliomexico@gmail.com) en National Library of Health
and Social Security de ClinicalKey.es por Elsevier en diciembre 02, 2022. Para uso personal exclusivamente. No
se permiten otros usos sin autorización. Copyright ©2022. Elsevier Inc. Todos los derechos reservados.
 Timing and breadth of antibiotics for possible sepsis 731

35. Levy MM, Evans LE, Rhodes A. The Surviving Sepsis Campaign Bundle: 2018
update. Intensive Care Med 2018;44(6):925–8.
36. Marik PE, Farkas JD, Spiegel R, et al. POINT: Should the Surviving Sepsis
Campaign Guidelines Be Retired? Yes Chest 2019;155(1):12–4.
37. Kalantari A, Rezaie SR. Challenging the one-hour sepsis bundle. West J Emerg
Med 2019;20(2):185–90.
38. Talan DA, Yealy DM. Challenging the One-Hour Bundle Goal for Sepsis Antibi-
otics. Ann Emerg Med 2019;73(4):359–62.
39. Levy MM, Rhodes A, Evans LE. Steering and Executive Committee of the Surviv-
ing Sepsis Campaign. COUNTERPOINT: Should the Surviving Sepsis Campaign
Guidelines Be Retired? No. Chest 2019;155(1):14–7.
40. Brunkhorst FM, Oppert M, Marx G, et al. Effect of empirical treatment with moxi-
floxacin and meropenem vs meropenem on sepsis-related organ dysfunction in
patients with severe sepsis: A randomized trial. JAMA 2012;307(22):2390–9.
41. Heyland DK, Dodek P, Muscedere J, et al. Randomized trial of combination
versus monotherapy for the empiric treatment of suspected ventilator-
associated pneumonia. Crit Care Med 2008;36(3):737–44.
42. Sjövall F, Perner A, Hylander Møller M. Empirical mono- versus combination anti-
biotic therapy in adult intensive care patients with severe sepsis – A systematic
review with meta-analysis and trial sequential analysis. J Infect 2017;74(4):
331–44.
43. Arulkumaran N, Routledge M, Schlebusch S, et al. Antimicrobial-associated harm
in critical care: a narrative review. Intensive Care Med 2020;46(2):225–35.
44. Jones BE, Ying J, Stevens V, et al. Empirical Anti-MRSA vs Standard Antibiotic
Therapy and Risk of 30-Day Mortality in Patients Hospitalized for Pneumonia.
JAMA Intern Med 2020;180(4):552–60.
45. Webb BJ, Sorensen J, Jephson A, et al. Broad-spectrum antibiotic use and poor
outcomes in community-onset pneumonia: A cohort study. Eur Respir J 2019;
54(1):1–9.
46. Attridge RT, Frei CR, Restrepo MI, et al. Guideline-concordant therapy and out-
comes in healthcare-associated pneumonia. Eur Respir J 2011;38(4):878–87.
47. Kett DH, Cano E, Quartin AA, et al. Implementation of guidelines for management
of possible multidrug-resistant pneumonia in intensive care: An observational,
multicentre cohort study. Lancet Infect Dis 2011;11(3):181–9.
48. Hranjec T, Rosenberger LH, Swenson B, et al. Aggressive versus conservative
initiation of antimicrobial treatment in critically ill surgical patients with suspected
intensive-care-unit-acquired infection: A quasi-experimental, before and after
observational cohort study. Lancet Infect Dis 2012;12(10):774–80.
49. Kadri SS, Lai YL, Warner S, et al. Inappropriate empirical antibiotic therapy for
bloodstream infections based on discordant in-vitro susceptibilities: a retrospec-
tive cohort analysis of prevalence, predictors, and mortality risk in US hospitals.
Lancet Infect Dis 2021;21(2):241–51.
50. Leisman DE. The Goldilocks Effect in the ICU—When the Data Speak, but Not the
Truth. Crit Care Med 2020;48(12):1887–9.
51. Yealy DM, Mohr NM, Shapiro NI, et al. 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.
52. Evans L, Rhodes A, Alhazzani W, et al. Surviving Sepsis Campaign: International
Guidelines for Management of Sepsis and Septic Shock 2021. Crit Care Med
2021;49(11):e1063–143.

Descargado para Biblioteca Medica Hospital México (bibliomexico@gmail.com) en National Library of Health
and Social Security de ClinicalKey.es por Elsevier en diciembre 02, 2022. Para uso personal exclusivamente. No
se permiten otros usos sin autorización. Copyright ©2022. Elsevier Inc. Todos los derechos reservados.
732 Pak et al

53. Evans L, Rhodes A, Alhazzani W, et al. Executive Summary: Surviving Sepsis

Campaign: International Guidelines for the Management of Sepsis and Septic
Shock 2021. Crit Care Med 2021;49(11):1974–82.
54. Rhee C, Brown SR, Jones TM, et al. Variability in determining sepsis time zero
and bundle compliance rates for the centers for medicare and medicaid services
SEP-1 measure. Infect Control Hosp Epidemiol 2018;39(8):994–6.
55. Zhang Z, Zhang G, Goyal H, et al. Identification of subclasses of sepsis that
showed different clinical outcomes and responses to amount of fluid resuscita-
tion: a latent profile analysis. Crit Care 2018;22(1):347.
56. Gårdlund B, Dmitrieva NO, Pieper CF, et al. Six subphenotypes in septic shock:
Latent class analysis of the PROWESS Shock study. J Crit Care 2018;47:70–9.
57. Han X, Spicer A, Carey KA, et al. Identifying High-Risk Subphenotypes and Asso-
ciated Harms From Delayed Antibiotic Orders and Delivery. Crit Care Med 2021;
49(10):1694–705.
58. Seymour CW, Kennedy JN, Wang S, et al. Derivation, Validation, and Potential
Treatment Implications of Novel Clinical Phenotypes for Sepsis. JAMA 2019;
321(20):2003–17.
59. Ibrahim ZM, Wu H, Hamoud A, et al. On classifying sepsis heterogeneity in the
ICU: Insight using machine learning. J Am Med Inform Assoc 2020;27(3):437–43.
60. Stanski NL, Wong HR. Prognostic and predictive enrichment in sepsis. Nat Rev
Nephrol 2020;16(1):20–31.
61. Bone RC, Balk RA, Cerra FB, et al. Definitions for sepsis and organ failure and
guidelines for the use of innovative therapies in sepsis. Chest 1992;101(6):
1644–55.
62. Pierrakos C, Velissaris D, Bisdorff M, et al. Biomarkers of sepsis: Time for a reap-
praisal. Crit Care 2020;24(1):1–15.
63. Pepper DJ, Sun J, Rhee C, et al. Procalcitonin-Guided Antibiotic Discontinuation
and Mortality in Critically Ill Adults: A Systematic Review and Meta-analysis.
Chest 2019;155(6):1109–18.
64. Elnajdy D, El-Dahiyat F. Antibiotics duration guided by biomarkers in hospitalized
adult patients; a systematic review and meta-analysis. Infect Dis (Auckl) 2022;
0(0):1–16.
65. Järvinen AKK, Laakso S, Piiparinen P, et al. Rapid identification of bacterial path-
ogens using a PCR- and microarray-based assay. BMC Microbiol 2009;9. https://
doi.org/10.1186/1471-2180-9-161.
66. Shang S, Chen G, Wu Y, et al. Rapid diagnosis of bacterial sepsis with PCR
amplification and microarray hybridization in 16S rRNA gene. Pediatr Res
2005;58(1):143–8.
67. Altun O, Botero-Kleiven S, Carlsson S, et al. Rapid identification of bacteria from
positive blood culture bottles by MALDI-TOF MS following short-term incubation
on solid media. J Med Microbiol 2015;64(11):1346–52.
68. Watanabe N, Koyama S, Taji Y, et al. Direct microorganism species identification
and antimicrobial susceptibility tests from positive blood culture bottles using
rapid Sepsityper Kit. J Infect Chemother 2022;2021(2). https://doi.org/10.1016/
j.jiac.2021.12.030.
69. Lin JFF, Ge MCC, Liu TPP, et al. A simple method for rapid microbial identification
from positive monomicrobial blood culture bottles through matrix-assisted laser
desorption ionization time-of-flight mass spectrometry. J Microbiol Immunol Infect
2018;51(5):659–65.

Descargado para Biblioteca Medica Hospital México (bibliomexico@gmail.com) en National Library of Health
and Social Security de ClinicalKey.es por Elsevier en diciembre 02, 2022. Para uso personal exclusivamente. No
se permiten otros usos sin autorización. Copyright ©2022. Elsevier Inc. Todos los derechos reservados.
 Timing and breadth of antibiotics for possible sepsis 733

70. Pak TR, Kasarskis A. How Next-Generation Sequencing and Multiscale Data
Analysis Will Transform Infectious Disease Management. Clin Infect Dis 2015;
61(11):1695–702.
71. Sweeney TE, Azad TD, Donato M, et al. Unsupervised analysis of transcriptomics
in bacterial sepsis across multiple datasets reveals three robust clusters. Crit
Care Med 2018;46(6):915–25.
72. Reyes M, Filbin MR, Bhattacharyya RP, et al. An immune-cell signature of bacte-
rial sepsis. Nat Med 2020;26(3):333–40.

Descargado para Biblioteca Medica Hospital México (bibliomexico@gmail.com) en National Library of Health
and Social Security de ClinicalKey.es por Elsevier en diciembre 02, 2022. Para uso personal exclusivamente. No
se permiten otros usos sin autorización. Copyright ©2022. Elsevier Inc. Todos los derechos reservados.