Recognition and treatment of severe sepsis in the emergency department: retrospective study in two French teaching hospitals

Severe sepsis and septic shock represent the most severe forms of infection. Two consecutive epidemiological surveys demonstrated an increasing incidence for severe sepsis [1, 2] from 2000 to 2012; it accounted for 500,000 annual emergency visits in USA in 2006 [3]. During the same period, attributable mortality rate declined from 40% in 2000 to 18% in 2012 [2].

Severe sepsis was defined in 2001 by a consensus conference as a presumed infection with a systemic inflammatory response syndrome (SIRS) and almost one organ dysfunction [4]. Although this definition had some weaknesses, it was used during almost two decades. More recently, a new consensus conference (SEPSIS-3) modified sepsis definition with introduction of screening by the quick SOFA (qSOFA) [5].

In their famous paper, Rivers et al. demonstrated that an intensive treatment during the first 6 hours after triage decreased mortality [6]. Since 2004, several recommendations from the Surviving Sepsis campaign (SSC) have been released [7] with objectives at three and 6 hours after triage. Objectives at 3 hours are: obtain lactate level and blood culture, administer broad spectrum antibiotics and administer 30 ml/kg crystalloid in case of hypotension or lactate ≥4 mmol/l. The implementation of theses guidelines has proved to decrease mortality from 37 to 30% [8].

However, the implementation of the SSC bundles in overcrowded Emergency Departments (ED) - in particular volume and delays of fluid loading and first antibiotic administration - remains difficult [9]. The early recognition of severe sepsis by the triage nurse is a real challenge because the clinical presentation might be confusing. This low compliance was demonstrated in Australia [10].

These facts motivated us to conduct a retrospective study of the compliance with the SSC 3 h-bundle and of the delay of recognition of severe sepsis in two EDs of French teaching hospitals. This study was realized before SEPSIS-3 release, thus the previous severe sepsis definition was used. Primary output was compliance with 3 h-bundle, secondary were analysis of the delay of severe sepsis recognition and description of the population.

This study was designed in accordance with STROBE statement [11].

Few real-life studies have explored the compliance to the SSC bundles in European EDs. Our main result was the low proportion (19%) of patients who were treated in accordance with the 3-h sepsis bundle [7]. This proportion was similar to one previously reported in Scottish EDs (between 29 and 43%) [14]. Outside Europe, compliance to SSC bundles was comparable to our study, 29% in Australia [10] and 28% in USA [15].

The low compliance with the 3-h SSC bundle was mainly due to low volume of fluid loading (18 ± 11 ml/kg) when compared to the recommendations (30 ml/kg) [7]. This volume was comparable to that of the control group in an ED trial [16]. The advanced age and the presence of cardiac comorbities might explain this result because physicians might fear fluid overloading. Our patients were older than those included in the ARISE trial (71 vs 63 years) and the incidence of cardiac comorbity was higher (24 vs 10%) [17]. Similar results were reported in a European ICU severe sepsis population [18]. A non invasive approach using bedside cardiac and lung ultrasound has demonstrated its usefulness and is recommended by a consensus conference [19]. Firstly, a Point-of-Care Ultrasound could affirm the diagnosis of severe sepsis when faced to an undifferentiated shock [20]. Secondly, index of hypovolemia like left ventricular obliteration [21] and/or collapsed inferior vena cava [22] could secure initial fluid loading. Thirdly, after fist infusion, dynamic indices of fluid responsiveness such as passive leg rising [23] could be used along with lung ultrasound for tolerance [24].

The poor compliance to the 3-h SSC bundle is partly explained by the delay between admission and severe sepsis recognition (more than 3 h). Although priorities after triage indicated a need for first medical contact within 120 min and 180 min, this delay was respected in only 32%. Although not constant, overcrowding of these two EDs was frequent. This fact has been associated with a poor compliance with SSC bundles [9]. A systematic review has explored the efficacy of electronic systems for the diagnosis of severe sepsis and did not find any improvement of survival [25]. A real-time electronic surveillance in the ED decreased time to blood culture but not to antibiotic administration [26]. A complex quality improvement project increased the adherence to sepsis bundles from 28% to 71% in a US ED [15]. A multicenter improvement program in 97 Australian EDs increased overall survival through the implementation of a screening tool and sepsis bundles [10]. All these studies included an enhanced triage process and simple guidelines (e.g. antibiotic administration in less than 1 hour after sepsis recognition). The implication of all health-care workers (nurses, physicians) was mandatory along with the implementation of electronic alerts [10, 16]. Finally, the quick SOFA, a simple bedside clinical score recommended by the SEPSIS-3 consensus conference, may improve the screening of sepsis in the ED [5].

There is a poor compliance with the 3-h SSC bundle in the ED and delay between admission and severe sepsis diagnosis has to be improved. This French situation is comparable to that observed elsewhere before implementation of new sepsis protocols. If confirmed by other studies, this situation should motivate the emergency physicians to improve their practice. A complete process involving nurses and physicians education, electronic sepsis alert, enhanced triage process and diffusion of simple sepsis bundle could be deployed. Societies of Emergency Medicine could play an incentive role in this process.