Sepsis screening in the emergency department
Independently of advances in diagnostics and therapeutics, the sepsis rate, as in hospital diagnosis, is increasing worldwide.1–3 Mortality has decreased, specifically in the intensive care unit (ICU) setting.3,4 The reasons for this are probably multifactorial, including modifications on the demography with ageing, concentration of risk patients, increased number of interventions, and ‘superbugs’; all of which have been recognised as contributors to the problem. A better knowledge, that raises the awareness of this condition, is also a factor of the increasing rate.
Reduction in mortality is a result of the Surviving Sepsis Campaign (SSC), providing guidelines, clear definitions, and advice on management based in early gold-directed therapies (EGDT). This approach has reduced mortality and modified incidence, decreased mortality in relation to the SSC bundles,5 and improved compliance.6
In the paediatric population during the last decades, an increase in incidence of sepsis has been associated with the better survival of preterm and low birth-weight infants and children with severe chronic conditions.7 The estimated number of paediatric patients with sepsis in the EU is >18/100,000 patients per year visiting the emergency department (ED).8 In adult patients the estimated number of cases is 400 cases per 100,000/year.3
Developing countries are quickly incorporating sepsis as a health treat9 and although most of the information comes from ICUs, an estimation of 3–4-fold increase during the last ten years are the actual figure, with limitations on the information.10
The profile of the septic patient is relevant; male gender has 1.28-fold increases in risk of sepsis, non-white races have greater incidence and also more severe cases.11 The effect of race is still a matter of debate and the association of non-white race and delays in administration of antibiotics has not been confirmed in recent publications.12 Actually there is more information regarding the genomic peculiarities of this ‘inadequate’ inflammatory response,13 identifying groups with more mortality that will help future screening and personalise management.
Extreme age and male gender are the predominant characteristics of septic patients. Older individuals (>65 years) have a five-fold increased likelihood of severe sepsis than younger adults; individuals living in nursing homes have higher mortality compared with patients of the same age who live at home.14
Difficulties in clinical identification of the septic patients have been established as one of the reasons for delays on treatment.15 A number of predictive scores for sepsis based on demographic, clinical and analytical variables, have demonstrate different performances when compared with clinical judgment.16 The use of education sessions, regular audits, personal feedback, and check lists have all resulted an increase in compliance with recommendations.17
Sepsis secondary to surgical interventions also has increased in incidence during the last decades, although the mortality has decreased.18
An important limitation of the information gathered and published is the focus on ICU observational studies19; considering that only 50% of sepsis admissions go straight to an ICU, this orientation only reflects part of the problem. Although with the inherent difficulties due to the weak diagnostic criteria, the estimation of septic patients in urban EDs is close to 4%20,21 of all visits, reflecting the burden of the problem in this services.
The ED population has clear differences to the ICU population. The definition of sepsis based on systemic inflammatory response syndrome (SIRS) was not specific enough; large epidemiological studies in the ED setting are needed to provide a clear picture of the sepsis problem.22 A clinical picture of sepsis patients has been drawn mainly from ICU studies, which have created a picture away from other settings such as the ED or the community services.23 One relevant aspect for the ED is the progression of the infected patients in the time following sepsis, septic shock (SS) or the need of ICU, which are crucial aspects for ED management; some efforts have been done to predict evolution.24
Modifications on definitions and the continuous search for an adequate diagnostic test demonstrate the heterogeneity of the disease; international coding systems (ICD 9-10) have little utility in identification.25,26 Administrative calculations are usually based on infection codes plus one organ dysfunction code, although there are specific ICD 9–10 codes for sepsis. There are still problems in accurately identifying severe sepsis and septic shock using this source of information.27
Initial efforts to reach an agreement about sepsis were carried out in the early 1990s, with a clear focus on ICU patients.28 The use of SIRS to define sepsis, and organ failure to define severe sepsis was assumed.
Sepsis was defined as suspected or demonstrated infection plus two or more SIRS criteria (Table 1). Sepsis complicated by organ dysfunction was termed severe sepsis, which could progress to septic shock, defined as ‘sepsis-induced hypotension persisting despite adequate fluid resuscitation.’
One decade later,29 no modifications to the definitions were made by the Second International Consensus. Although the concept and the utility of SIRS was discussed, confirming lack of sensitivity, specificity and the need for staging of the septic patients were the main contributions. Previous concepts and definitions of sepsis, severe sepsis and septic shock remained. The PIRO concept (Predisposition, Infection, Response, and Organ Dysfunction) was introduced to help understand the evolution, proposed as a template for the future, with further development needed.30
The need of more specific characterisation of patients generated interest in tools (scores) and new biomarkers,31 and focus on the concept of sepsis as a severe condition generated by inflammatory response to infection that leads to organ dysfunction. Non-infected ‘sterile’ patients with SIRS and organ dysfunction need a clear different definition for specific treatments.
The increasing recognition of the low sensitivity and specificity of SIRS has promoted new definitions in the Third International Consensus (Sepsis-3) that was published in 2016. The new recommendations have expanded the orientation to settings including out-of-hospital, emergency department, and hospital ward settings, pinpointing the importance of early recognition.
Sepsis is defined as a life-threatening organ dysfunction caused by infection through an inappropriate host response. The concept of sepsis was limited to patients with infection and at least one affected organ.32 Severe sepsis is not an accepted concept; the severity is estimated using the SOFA score based on organ dysfunction. Patients with a SOFA (Table 2) score ≥2 are defined as septic; this in-hospital mortality is close to 10%. Patients with septic shock can be clinically identified by a vasopressor requirement to maintain a mean arterial pressure (MAP) of 65mmHg or greater, and/or serum lactate level >2mmol/l (>18mg/dl) in the absence of hypovolaemia. Patients with septic shock have greater mortality than septic patients.
In order to facilitate the identification of the septic patient, a reduced version of a validated score – the qSOFA (Table 3), with only three variables (respiratory rate, mental status, systolic blood pressure) – was proposed. The qSOFA can be of great use in the prehospital setting and ED to evaluate organ dysfunction. Patients with qSOFA≥2 require further evaluation to establish sepsis criteria. The new sepsis definition has good correlation with mortality not just in the ICU setting; qSOFA has a low sensitivity (48%) to detect organ failure in other settings.33
Precise sepsis definitions are not only needed to improve homogeneity of patient populations included in clinical trials, but clinical work is also of great interest for identification and implementation of the recommendations because some publications still express concerns on the daily application of the new definitions.15
It is universally accepted that timely recognition of sepsis is a main objective of management. Definitions have played a crucial role in the awareness and application of EGDT in the ED with nearly universal benefits after the implementation.34 The search for a lab test that identifies septic patients in the early phase has become the objective of researchers. The recognition of the septic patient is not enough; risk estimation, risk of progression, need of ICU or death are the natural steps for an adequate management. As the septic patient is in a continuous evolution, progression to more severe states is a relevant aspect; one quarter of the patients initially evaluated in the ED will progress to more severe situations, with more or extend organ dysfunction.35 This progression is difficult to predict using initial physiologic parameters; SOFA has demonstrated more adequacy for this task in the ICU.36 The new Sepsis-3 definition has been supported by publications in which the redefinition of sepsis predicts hospital mortality better than SIRS and the previous definition of severe sepsis definition. qSOFA is a reliable tool in ED37 for prediciting in-hospital mortality.
Prehospital sepsis identification
Results on identification of the septic patient by the EMS have been poor,38 and more analysis is needed considering the different EMS models and the variability of the findings.39 The feasibility of sepsis identification by paramedics and ED application in order to reduce time to administering antibiotics has been determined.40 Reduced management time and lack of resources jeopardised the identification of sepsis by the EMS; identification of septic patients reduces mortality, demonstrating the importance this phase of management.41
Identification of septic patients in the triage station enables the possibility of implementing specific management paths, such as those established for stroke, acute coronary syndrome, and severe trauma, that have proven efficacies. The feasibility of sepsis identification in the triage station through the use of a check list in all patients with hypo- or hyperthermia and including altered mental status, glycaemia, SatO2, heart rate (HR), respiratory rate (RR) and mottled skin has been investigated; if more than three criteria are present, an alert is set for further evaluation. The results show a reduction in time to fluids or antibiotics.42 In EDs with a National Early Warning System (NEWS) implemented in the triage station, the use of this risk score has demonstrated a high sensitivity (92.6%; 95% CI 74.2–98.7) for aggregated levels of 3 and 4, and has been a good tool in the identification of sepsis or septic shock.43 Triage score and way of access to the ED are variables related to severity and associated to in hospital mortality and ICU requirement in septic patients.44
Classically the identification of sepsis has been based on the clinical impression and deterioration of the vital signs, using biomarkers for confirmation. The SIRS parameters are founded on the vital signs HR, RR, temperature plus white blood cell count (WBC); blood pressure is not included as reduction in blood pressure is considered a marker of cardiovascular system deterioration.45
Elevated lactate is classically a sign associated with tissue hypoperfusion due to anaerobic metabolism with reduction of mitochondrial lactate clearance. Elevated lactate in septic shock is mostly due to stimulation of beta-2 adrenergic receptors generated by internal catecholamines. Lactate in septic patients is a good predictor of 28 days mortality, Lactate ≥4mmol/l is associated with higher in-hospital mortality, if associated hypotension reaches 45%. Mortality is greater even if the patients are not hypotensive.46,47 Lactate clearance, defined as the reduction of the initial lactate after treatment, supports the value of lactate as a risk indicator, patients that clear lactate have lower mortality and less therapeutic effort is required (vasopressor, mechanical ventilation).48 Lactate is part of the recommendations for the first hour.
The early sepsis concepts of inflammation, cell dysfunction, coagulation, pathogen identification, and metabolopathies are the fundaments for biomarker design. In an extended review, 178 potential biomarkers where identified,49 reflecting the difficulty in selecting the adequate biomarkers that can accomplish early identification of a septic patient in the population of patients with infection, with diagnostic and prognostic proficiencies, and therapeutic orientation capabilities. Unfortunately six years since publication, the statement is still valid,50 the difficulties coming from the imprecise definition of sepsis and the multiple different populations included in this syndrome. A recent extended review identified 60 different types of markers with publications of enough quality for pooling the information, and only seven biomarkers with more than four publications: procalcitonin (PCT); C-reactive protein (CRP); interleukin 6 (IL-6); soluble triggering receptor expressed on myeloid cells-1 (sTREM-1); presepsin (sCD14-ST); lipopolysaccharide binding protein (LBP); and CD64.
PCT, CRP, sTREM-1, presepsin, LBP and CD64, have validity as diagnostic tests for sepsis; after pooling the selected publications the Area Under the ROC Curve (AUC) values are 0.85, 0.77, 0.79, 0.88, 0.71 and 0.85, respectively. Other biomarkers have AUC close to 0.9 but sample size and other methodological limitations affect this value.50
The biomarkers have been used for different purposes; as diagnostic tests for sepsis identification in the population of SIRS or infected patients; for risk stratification predicting need of ICU admission or 28 days mortality; progression to a more severe situation; and guidance for therapeutic management, specifically on the use of antibiotics. It is relevant, when biomarkers are evaluated, to consider the clinical setting in the publication, which should be the same in which the recommendation is going to carried. PCT has been used extensively as a diagnostic test in the ED to differentiate bacterial or viral infection from sepsis; AUCs for WBC, PCT and CRP are 0.72, 0.79, and 0.53, respectively, also demonstrating a prognostic value for inhospital mortality (AUC of 0.72).51
In a review of 3224 patients,52 in different settings and disease severities, PCT as a diagnostic tool had an AUC of 0.85. As a tool to guide therapy, a recent review has demonstrated the utility in lower respiratory infections, in the ED in adult patients and in reducing the use of antibiotics, but these results have not been reproduced in paediatric populations or in other types of infections.53
PCT was effective in predicting ICU admission in an elderly ED population,54 and similar results were found in cancer patients in the ED.55 The presence of low PCT (≤0.25ng/ml) in septic patients in ED population has a prevalence of 12.7%; this group has a lover 28 days’ mortality, and a number of factors including obesity, type of infection and level of CRP, are associated with this finding, reflecting the need for further studies on specific groups to evaluate the utility as diagnostic, risk stratification, or therapy guidelines.
PCT has demonstrated utility alone or in combination with other biomarkers in predicting bacteraemia.56 In ED patients, prediction of bacteraemia for management is not useful; 40% of the patients with severe sepsis have negative blood cultures, and no significant differences in ICU mortality (35.8%/44%).57
Therapeutic guidance of PCT from a review published in 2012 that supports the use of antibiotics based on procalcitonin levels58 has received criticism due to the limitation of the studies on specific groups, and the difficulties in adherence in prospective studies.53 Procalcitonin has demonstrate utility as a cost-effective diagnostic and guidance tool in the ED, at least in respiratory infections.59
Risk stratification to identify the more severe cases in the ED has been an early goal, not only to provide adequate care but also for uniform clinical trials or as a benchmark. The first score specifically designed for septic patients in the ED was the Mortality in Emergency Department Sepsis (MEDS) score,60 that showed good correlation with 28 days’ mortality, (AUC 0.8). Limitations of the MEDS are concentrated in the most severe cases, although the utility is superior to biomarkers such as PCR or PCT.61
MEDS, APACHE, CURB-65, and REMS were compared in a population of septic patients in the ED. MEDS demonstrate superiority in predicting 28 days mortality, with AUCs 0.82, 0.71, 0.78, for MEDS, APACHE and CURB-65, respectively.62 One important consideration is the feasibility to calculate the scores, MEDS, CURB-65, and REMS was calculated in more than 90% of the ED patients.
In sepsis, as in other conditions, management is based on risk estimations; a SOFA score with ≥2 points identifies a group of patients with in-hospital mortality of 10%, and has been proposed by Sepsis-3 as a risk estimation score. In-hospital or 28 days mortality is the predicted risk although others can be use such as ICU admission or need of vasopressors.
SOFA has been demonstrated as a superior risk estimator to qSOFA and SIRS for septic patients or in-hospital mortality in a extended population of ICU patients.36 This result has been confirmed in the ED setting, showing good performance in predicting in-hospital mortality; qSOFA shows a low performance in the ED setting,33 limiting the actual recommendation.
The definition of sepsis is still in development. Definitions are important for reporting and for determining inclusion criteria for clinical trials. Administrative codings (ICD9-10) need adaptation to accommodate recent developments. Further studies on biomarkers used for early sepsis diagnosis with prognostic power, and enough specificity to differentiate SIRS from sepsis are required. In future, the identification of different sepsis patterns based on biomarkers, pathophysiology or the responsible infection will allow the exploration of personalised treatments.
1 Kempker JA, Martin GS. The changing epidemiology and definitions of sepsis. Clin Chest Med 2016;37(2):165–79.
2 Perner A et al. Sepsis: frontiers in diagnosis, resuscitation and antibiotic therapy. Intensive Care Med 2016;42(12):1958–69.
3 Stoller J et al. Epidemiology of severe sepsis: 2008–2012. J Crit Care 2016;31(1):58–62.
4 Shankar-Hari M et al. Differences in impact of definitional elements on mortality precludes international comparisons of sepsis epidemiology –
A cohort study illustrating the need for standardized reporting. Crit Care Med 2016;44(12):2223–30.
5 Levy MM et al. Surviving Sepsis Campaign: association between performance metrics and outcomes in a 7.5-year study. Crit Care Med 2015;43(1):3–12.
6 Romero B, Fry M, Roche M. The impact of evidence based sepsis guidelines on emergency department clinical practice: a pre-post medical record audit.
J Clin Nurs 2017; Jan 10 [Epub ahead of print].
7 de Souza DC, Barreira ER, Faria LS. The epidemiology of sepsis in childhood. Shock 2017;47(1S Suppl 1):2–5.
8 Singhal S et al. National estimates of emergency department visits for pediatric severe sepsis in the United States. Peer J 2013;1:e79.
9 Liao X et al. Current epidemiology of sepsis in mainland China. Ann Transl Med 2016;4(17):324.
10 Fleischmann C et al. Assessment of global incidence and mortality of hospital-treated sepsis. Current estimates and limitations. Am J Respir Crit Care Med 2016;193(3):259–72.
11 Moss M. Epidemiology of sepsis: race, sex, and chronic alcohol abuse. Clin Infect Dis 2005;41 Suppl 7:S490–7.
12 Madsen TE, Napoli AM. Analysis of race and time to antibiotics among patients with severe sepsis or septic shock. J Racial Ethn Health Disparities 2016;Aug 23 [Epub ahead of print].
13 Rautanen A, Mills TC, Gordon AC, et al. Genome-wide association study of survival from sepsis due to pneumonia: an observational cohort study. Lancet Respir Med 2015;3(1):53–60.
14 Ginde AA et al. Impact of older age and nursing home residence on clinical outcomes of US emergency department visits for severe sepsis. J Crit Care 2013;28(5):606–11.
15 Brown T et al. Comparative effectiveness of physician diagnosis and guideline definitions in identifying sepsis patients in the emergency department. J Crit Care 2015;30(1):71–7.
16 Balamuth F et al. Comparison of two sepsis recognition methods in a pediatric emergency department. Acad Emerg Med 2015;22(11):1298–306.
17 Bentley J et al. Seeking sepsis in the emergency department – Identifying barriers to delivery of the Sepsis 6. BMJ Qual Improv Rep 2016;5(1).
18 Bateman BT et al. Temporal trends in the epidemiology of severe postoperative sepsis after elective surgery: a large, nationwide sample. Anesthesiology 2010;112(4):917–25.
19 Whittaker SA et al. Epidemiology and outcomes in patients with severe sepsis admitted to the hospital wards. J Crit Care 2015;30(1):78–84.
20 Cowan SL et al. The burden of sepsis in the Emergency Department: an observational snapshot. Eur J Emerg Med 2015;22(5):363–5.
21 McNevin C et al. What proportion of patients meet the criteria for uncomplicated sepsis in an Irish Emergency Department? Ir Med J 2016;109(7):435.
22 Gille-Johnson P, Hansson KE, Gardlund B. Severe sepsis and systemic inflammatory response syndrome in emergency department patients with suspected severe infection. Scand J Infect Dis 2013;45(3):186–93.
23 Rohde JM et al. The epidemiology of acute organ system dysfunction from severe sepsis outside of the intensive care unit. J Hosp Med 2013;8(5):243–7.
24 Capp R et al. Predictors of patients who present to the emergency department with sepsis and progress to septic shock between 4 and 48 hours of emergency department arrival. Crit Care Med 2015;43(5):983–8.
25 Ibrahim I et al. Accuracy of International classification of diseases, 10th revision codes for identifying severe sepsis in patients admitted from the emergency department. Crit Care Resusc 2012;14(2):112–18.
26 Bouza C, Lopez-Cuadrado T, Amate-Blanco JM. Use of explicit ICD9-CM codes to identify adult severe sepsis: impacts on epidemiological estimates. Crit Care 2016;20(1):313.
27 Shahraz S et al. Use of systematic methods to improve disease identification in administrative data: The case of severe sepsis. Med Care 2017;55(3):
28 American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference: definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Crit Care Med 1992;20(6):864–74.
29 Vincent JL. Sepsis definitions. Lancet Infect Dis 2002;2(3):135.
30 Levy MM et al. 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference. Intensive Care Med 2003;29(4):530–8.
31 Vincent JL, Martinez EO, Silva E. Evolving concepts in sepsis definitions. Crit Care Nurs Clin North Am 2011;23(1):29–39.
32 Townsend SR, Rivers E, Tefera L. Definitions for sepsis and septic shock. JAMA 2016;316(4):457–8.
33 Giamarellos-Bourboulis EJ et al. Validation of the new Sepsis-3 definitions: proposal for improvement in early risk identification. Clin Microbiol Infect 2017;23(2):104–9.
34 Wira CR et al. Meta-analysis of protocolized goal-directed hemodynamic optimization for the management of severe sepsis and septic shock in the emergency department. West J Emerg Med 2014;15(1):51–9.
35 Arnold RC et al. Multicenter observational study of the development of progressive organ dysfunction and therapeutic interventions in normotensive sepsis patients in the emergency department. Acad Emerg Med 2013;20(5):433–40.
36 Raith EP et al. 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.
37 Freund Y et al. Prognostic accuracy of Sepsis-3 criteria for in-hospital mortality among patients with suspected infection presenting to the emergency department. JAMA 2017;317(3):301–8.
38 Smyth MA et al. Identification of adults with sepsis in the prehospital environment: a systematic review. BMJ Open 2016;6(8):e011218.
39 Lane D et al. Prehospital management and identification of sepsis by emergency medical services: a systematic review. Emerg Med J 2016;33(6):408–13.
40 Carberry M, Harden J. A collaborative improvement project by an NHS Emergency Department and Scottish Ambulance Paramedics to improve the identification and delivery of sepsis 6. BMJ Qual Improv Rep 2016;5(1).
41 Roest AA et al. Ambulance patients with nondocumented sepsis have a high mortality risk:
a retrospective study. Eur J Emerg Med 2017;24(1): 36–43.
42 Patocka C et al. Evaluation of an emergency department triage screening tool for suspected severe sepsis and septic shock. J Healthcare Qual 2014;36(1):52–61; quiz 59–61.
43 Keep JW et al. National early warning score at emergency department triage may allow earlier identification of patients with severe sepsis and septic shock: a retrospective observational study. Emerg Med J 2016;33(1):37–41.
44 Ibrahim I, Jacobs IG. Can the characteristics of emergency department attendances predict poor hospital outcomes in patients with sepsis? Singapore Med J 2013;54(11):634–8.
45 Knaus WA et al. Evaluation of definitions for sepsis. Chest 1992;101(6):1656–62.
46 Casserly B et al. Lactate measurements in sepsis-induced tissue hypoperfusion: results from the Surviving Sepsis Campaign database. Crit Care Med 2015;43(3):567–73.
47 Howell MD et al. Occult hypoperfusion and mortality in patients with suspected infection. Intensive Care Med 2007;33(11):1892–9.
48 Bhat SR et al. Lactate clearance predicts survival among patients in the emergency department with severe sepsis. West J Emerg Med 2015;16(7):1118–26.
49 Pierrakos C, Vincent J-L. Sepsis biomarkers: a review. Crit Care 2010;14(1):R15.
50 Liu Y et al. Biomarkers for diagnosis of sepsis in patients with systemic inflammatory response syndrome: a systematic review and meta-analysis. Springerplus 2016;5(1):2091.
51 Magrini L et al. Comparison between white blood cell count, procalcitonin and C reactive protein as diagnostic and prognostic biomarkers of infection or sepsis in patients presenting to emergency department. Clin Chem Lab Med 2014;52(10):
52 Wacker C et al. Procalcitonin as a diagnostic marker for sepsis: a systematic review and meta-analysis. Lancet Infect Dis 2013;13(5):426–35.
53 van der Does Y et al. Procalcitonin-guided therapy for the initiation of antibiotics in the ED: a systematic review. Am J Emerg Med 2016;34(7):1286–93.
54 Lee WJ et al. Are prognostic scores and biomarkers such as procalcitonin the appropriate prognostic precursors for elderly patients with sepsis in the emergency department? Aging Clin Exp Res 2016;28(5):917–24.
55 Kece E et al. Comparison of diagnostic and prognostic utility of lactate and procalcitonin for sepsis in adult cancer patients presenting to emergency department with systemic inflammatory response syndrome. Turk J Emerg Med 2016;16(1):
56 Kim S-Y et al. Procalcitonin in the assessment of bacteraemia in emergency department patients: results of a large retrospective study. Ann Clin Biochem 2015;52(6):654–59.
57 Phua J et al. Characteristics and outcomes of culture-negative versus culture-positive severe sepsis. Crit Care 2013;17(5):R202.
58 Schuetz P et al. Procalcitonin to initiate or discontinue antibiotics in acute respiratory tract infections. Cochrane Database Syst Rev 2012;(9):CD007498.
59 Schuetz P et al. Economic evaluation of procalcitonin-guided antibiotic therapy in acute respiratory infections: a US health system perspective. Clin Chem Lab Med 2015;53(4):583–92.
60 Shapiro NI et al. Mortality in Emergency Department Sepsis (MEDS) score: a prospectively derived and validated clinical prediction rule. Crit Care Med 2003;31(3):670–5.
61 Carpenter CR et al. Risk stratification of the potentially septic patient in the emergency department: the Mortality in the Emergency Department Sepsis (MEDS) score. J Emerg Med 2009;37(3):319–27.
62 Hilderink MJ et al. Predictive accuracy and feasibility of risk stratification scores for 28-day mortality of patients with sepsis in an emergency department. Eur J Emerg Med 2015;22(5):331–7.