Towards novel BIOmarkers to diagnose SEPsis on the emergency room

Despite a global decrease in sepsis burden, sepsis still causes almost 20% of
all deaths worldwide [1, 2] and contributes significantly to in-hospital
mortality [3]. The surviving sepsis campaign pointed out diagnosis as a
fundamental challenge [4]. Early diagnosis and treatment are pivotal and
associated with decreased mortality rates [5]. However, in everyday clinical
practice taking action upon suspicion of severe infection, e.g. by
administering antibiotics, often takes more than 4 hours after admission [6].
In early stages of the septic response, the source of the infection may be
unclear and the clinical signs indistinguishable from non-infectious diseases,
leading to missed or delayed diagnoses [7].

For over 25 years, patients with an infection who met two or more of the
systemic inflammatory response syndrome (SIRS) criteria, were diagnosed with
sepsis [8]. The limited specificity of the SIRS criteria led to a redefinition
of sepsis in 2016 in which the presence of organ failure, as measured by the
sequential organ failure assessment (SOFA) score, in a patient with a
(suspected) infection became key [5]. Furthermore, since the SOFA score
requires laboratory testing, the quick SOFA (qSOFA) score which is calculated
according to three parameters (systolic blood pressure, respiratory rate and
Glascow Coma Scale), was introduced to enable early sepsis recognition at
bedside. However, the qSOFA unfortunately lacks sensitivity [9]. Early Warning
Scores (EWS) also have a poor prognostic value in predicting sepsis mortality
[10]. The most recent Surviving Sepsis Campaign guidelines recommend use of
SIRS, National Early Warning Score (NEWS) or Modified Early Warning Score
(MEWS) as a screening tool for sepsis [4]. The most commonly used are the MEWS
and NEWS [11]. To overcome the limitations of clinical scores, several
biomarkers for the diagnosis of sepsis have been studied, but so far, none of
them have sufficient specificity or sensitivity to be routinely employed in
clinical practice [12, 13]. Furthermore, microbiological cultures are
time-consuming and only a small percentage of culture tests will yield a
positive result for existence of microbes, thus an ideal diagnostic tool for
sepsis is not available at this moment [14].

Data from the United States (US) indicate that sepsis is a common presentation
in the emergency department (ED) and represents roughly one third of all
hospital admissions that culminate in death [15]. In addition, one Finnish
study from 2011 has estimated that up to 50% of ED presentations for a
suspected infection (defined by the physician*s decision to take samples for
blood cultures) have sepsis with a 28-days mortality rate of up to 30% [16].
Patients with hospital-acquired sepsis are even three times more at risk for
in-hospital mortality compared to patients with community-acquired sepsis [14].

After hospital discharge patients who were initially admitted for sepsis still
have an increased risk of death [17]. Data from the US showed that one-third of
sepsis survivors die during the following year, of which half of the deaths
were related to complications of sepsis. Furthermore, one-sixth of sepsis
survivors did experience severe persistent physical disability or cognitive
impairment [18]. As a result, the Surviving Sepsis Campaign guidelines now
recommend assessment and follow-up for physical, cognitive, and emotional
problems of patients after hospital discharge for sepsis [4]. Although the
challenges of critical illness survivorship are now increasingly well
documented, there are relatively few studies on enhancing recovery and how to
identify patients at increased risk [17].

There is a high need to rapidly and effectively identify patients on the ED who
are at risk of progressing along the infection-sepsis spectrum. The objectives
of the present study are 1) to compare the immune response of patients with or
without sepsis presenting to the ED with (suspected) infection, 2) to determine
immune response aberrations that are associated with an increased risk to
develop sepsis in patients presenting to the ED with (suspected) infection
without sepsis and 3) to determine the long term cognitive and physical
sequelae of sepsis after admission. We will investigate the immune response
using Raman spectroscopy [19, 20] and genomic, transcriptomic, metabolomic and
proteomic data in order to better stratify sepsis patients and gain insight
into the pathophysiology of sepsis. The hypothesis is that Raman spectroscopy
can be used as a classifier of immunological profiles that could allow to
identify a fraction of the population with a high likelihood of developing
sepsis. Raman spectroscopy is a non-destructive analytical technique that uses
the inelastic scattering of light to provide information on chemical
composition. Raman spectroscopy in unbiased in terms of detection and allows to
capture the combined response of samples. Raman spectroscopy can determine
changes or fluctuations in elements of the immune system, alerting about a
possible infection.

Objectives:
1) To compare the immune response of patients with or without sepsis presenting
to the ED with a(n) (suspected) infection.
2) To determine immune response aberrations that are associated with an
increased risk of developing sepsis in patients presenting to the ED with a(n)
(suspected) infection without sepsis.
3) To determine the long term cognitive and physical sequelae of sepsis after
admission.

Sepsis will be defined in accordance with the current Sepsis 3.0 criteria as
a(n) (suspected) infection with evidence of organ failure, as reflected by a
SOFA (Sequential Organ Failure Assessment) score of >=2 [5]. Notably, a
molecular definition of sepsis does not exist and there is no pathological gold
standard; therefore, in accordance with the current international consensus
[5], we consider the commonly used clinical organ failure (SOFA) criteria the
best option. The SOFA score is composed of six organ dysfunctions
(cardiovascular, pulmonary, renal, hepatic, coagulation and neurological) [21].
The SOFA score was developed for ICU patients, but its components can be easily
scored in an ED (and hospital ward) setting with the exception of the pulmonary
component; this pulmonary dysfunction score is based on the PaO2/FiO2 (PF)
ratio, wherein PaO2 is the partial pressure of oxygen in arterial blood and
FiO2 the fraction of inspired oxygen. Measurement of the PaO2 requires an
arterial blood puncture, which is not routinely done on the ER or hospital
ward. Therefore, we will use an alternative method to determine the respiratory
SOFA by determining the SpO2/FiO2 (SF) ratio, wherein SpO2 is peripheral oxygen
saturation [22]. SpO2 is routinely measured by finger pulse oximeter in
patients with suspected infection; FiO2 is 21% when breathing room temperature
and increases by 4% with each liter of oxygen provided per minute to a patient
via a nasal cannula. Cut-off values for SF ratios correlating with SOFA
pulmonary scores based on PF ratios have been validated in large data sets
[22].

Prospective observational cohort study conducted at the Amsterdam UMC, location
AMC and location VUmc. We expect to include a maximum of 3120 patients in 48
months.
Adults presenting at ED of the Amsterdam UMC with a(n) (suspected) infection
and a Modified Early Warning Scores (MEWS) of >= 2 will be screened for
eligibility for this study.

Patients have no benefit of participation in the study. Patients receive their
regular treatment as was intended and this will not be influenced by the study.
The burden of the blood samples taken is negligible. There is no additional
risk in participating in the study. Patients participating in this study will
be two times subjected to a blood withdrawal of a maximum of 67,5 ml, one time
to a rectal swab and two times subjected to a questionnaire (60-day and 1-year
follow-up). However, participation will contribute to the improvement of the
knowledge about sepsis.