Septic pathology is an extremely frequent reason for consultation in our emergency services, with an annual incidence of severe forms between 50 to 95 cases per 100,000 inhabitants and a constant increase estimated at 9% per year. Diagnosing these patients early and precisely is a major challenge for the clinician, as this diagnosis will lead to more or less aggressive medical management.

The criteria of S.I.R.S, used to define and to sort patients in sepsis according to the old definition, were completely abandoned in the last recommendations for lack of specificity but also of sensitivity. The latest recommendations suggest using another score, the "Quick Sepsis Related Organ Failure Assesment (qSOFA) score", in order to early detect septic patients at risk of poor progress. However, the recent literature highlights a very low sensitivity of the qSOFA score for the screening of septic patients, ranging from 30 to 60% according to the studies. In addition to qSOFA, other scores are described in the literature with apparently higher sensitivity, and thus seem more suitable for our daily practice. Among them is the NEWS score or the RETTS score.

Each of these scores is again based upon the values of vital signs recorded as soon as the patient arrives in the emergency department. To date, very few studies have been interested, in a prospective way, in the sensitivity and the specificity of these different scores to diagnose the "infected" patients in the emergency departments.

Therefore a non-interventional, prospective, multicenter cohort study is carried out here, in order to be able to compare, on the same cohort of patients admitted into emergency services, the diagnostic performance of these different scores with respect to the presence or absence of an infection.

The aim of this study is to define the best clinical score to use in emergency medicine to quickly diagnose the infected patients, and offer them the best medical care.

Septic pathology is an extremely frequent reason for consultation in our emergency services, with an annual incidence of severe forms between 50 to 95 cases per 100,000 inhabitants and a constant increase estimated at 9% per year. Diagnosing these patients early and precisely is a major challenge for the clinician, as this diagnosis will lead to more or less aggressive medical management with direct repercussions on the patient's morbidity and mortality.

The latest publications of March 2017 have completely changed the definition of the septic patient, in particular by eliminating the concept of "severe sepsis". Today the patient can be classified as either an "infected" or "sepsis" patient when there is associated organ failure, or as "septic shock" when it is necessary to use a cathecholaminergic support to maintain a sufficient blood pressure. There is a real nosological continuum between these three diagnostic entities and each of them is tainted by a growing intra-hospital mortality. One in ten patients will progress from infection to sepsis during their hospitalization and 3% of sepsis will progress to septic shock, making it easy to understand the value of screening these patients quickly in order to stop the progression of the infection. The triage of these patients, as soon as they arrive in an emergency service is therefore essential.

The criteria of S.I.R.S, used to define and to sort patients in sepsis according to the old definition, were completely abandoned in the last recommendations for lack of specificity but also of sensitivity. Indeed, a large number of non-septic patients presenting to the emergency department, has at least two SIRS criteria from the initial phase of their care, making it difficult to use this score as a reliable diagnostic tool.

The latest recommendations suggest using another score, the "Quick Sepsis Related Organ Failure Assesment (qSOFA) score", in order to early detect septic patients at risk of poor progress. This score is easily usable because it takes into account only three criteria that are: (i) the respiratory rate ≥ 22 / min, (ii) the systolic blood pressure ≤ 100 mmHg, (iii) and the level of vigilance of the patient to evaluate by the Glasgow score <15. Each criterion is weighted by 1 point if it is present. A score of qSOFA ≥2 is considered positive, and in this case the patient must be managed quickly with optimization of its volemia and implementation of a suitable antibiotic treatment within one hour.

However, the recent literature highlights a very low sensitivity of the qSOFA score for the screening of septic patients, ranging from 30 to 60% according to the studies. Such sensitivity values seem insufficient to properly screen septic patients in an emergency department. Some authors even report similar AUCs for qSOFA and SIRS at 0.72 and 0.73, respectively, to screen for the severe septic patient, concluding that the decision to abandon SIRS for qSOFA may be a little premature.

In addition to qSOFA, other scores are described in the literature with apparently higher sensitivity, and thus seem more suitable for our daily practice. Among them is the NEWS score or the RETTS score.

Each of these scores is again based upon the values of vital signs recorded as soon as the patient arrives in the emergency department, such as the respiratory rate, the heart rate, the oxygen saturation, the temperature, the blood pressure or the level of patient's vigilance.

A prospective cohort study of 1535 patients in emergency departments have compared the RETTS score to the qSOFA score showing a much better sensitivity of the RETTS score with a value of 85% against only 32% for qSOFA, for the screening of patients with severe sepsis according to the "SEPSIS-2" definition.

A recent study published in April 2017, compared qSOFA, SIRS, and NEWS to predict mortality or hospitalizations into intensive care units for septic patients in emergencies. Once again the qSOFA score showed a low sensitivity at 54% versus 67% for a NEWS score ≥ 8.

To date, several studies have compared the prognostic contribution of these different scores and their ability to predict mortality or even hospitalization in intensive care units, but very few studies have been interested, in a prospective way, in the sensitivity and the specificity of these different scores to diagnose the "infected" patients in the emergency departments. Whereas this first step seems essential in order to implement the most appropriate therapeutic strategy, today, if we stick to the recommendations, the diagnosis of the infected patient is based solely on the physician clinical feeling.

Therefore it seemed interesting to carry out a prospective, multicenter cohort study, in order to be able to compare, on the same cohort of patients admitted into emergency services, the diagnostic performance of these different scores (qSOFA, SIRS, NEWS, and RETTS ) with respect to the presence or absence of an infection.

Inclusion Criteria:

• Patient of 18 years old or more
• Patients consulting in the emergency department
• Verbal agrement to participate

Exclusion Criteria:

• Patients consulting for an isolated traumatologic reason
• Patients under legal protection

-Patients meeting the eligibility criteria from all participating investigational centers during the study period