Prehospital characteristics among patients with sepsis: a comparison between patients with or without adverse outcome

The early identification of patients with sepsis in the prehospital setting is crucial for reducing the adverse outcome risks due to inadequate assessments or delayed medical interventions [1]. Therefore, it is necessary for clinicians in the emergency medical services (EMS) to correctly assess and interpret the prehospital characteristics of these patients.

Sepsis is a common and serious worldwide medical condition [2‐4]. The incidence in Sweden of severe sepsis according to Sepsis-2 criteria is 687/100,000 inhabitants/year, and the incidence of the newly developed definition Sepsis-3 is 780/100,000 inhabitants/year [4]. Approximately 50 to 75% of the patients with sepsis are cared for and transported to emergency departments (EDs) by the EMS [5‐7]. Unfortunately, these patients have a high in-hospital mortality rate [1, 5‐7] and relatively few patients (only 6 to 36%) are suspected of having sepsis by EMS clinicians in the prehospital field [1, 6, 8, 9]. However, the clinical presentation of sepsis is often nonspecific [10], making early identification difficult [11], which may lead to delayed treatment and a worse prognosis.

The definition of sepsis recently changed. Previously, sepsis (Sepsis-2) was defined as a suspected infection in combination with the systemic inflammatory response syndrome (SIRS) [12, 13]. Sepsis with organ dysfunction was termed severe sepsis or septic shock if also hypotensive and not responding to fluid challenge [12, 13]. Since 2016 sepsis is defined as ‘life-threatening organ dysfunction caused by a dysregulated host response to infection’ commonly known as Sepsis-3 [14].

Early identification of sepsis by the EMS clinicians is crucial for these patients [15, 16]. Previous studies have indicated that a documented suspicion of sepsis in the EMS electronic health records shortened the time until the administration of antibiotics [1, 6] and a delayed time to the start of antibiotics is associated with increased progression of severe sepsis to septic shock and increased mortality [17]. These results indicate that a documented suspicion by the EMS are important for time to treatment in patients with sepsis. Prehospital screening tools for the identification of patients with sepsis and/or severe sepsis include the PREhospital Severe Sepsis (PRESS) score, Robson screening tool, Sepsis Alert protocol, quick Sequential (Sepsis-related) Organ Failure Assessment (qSOFA) and BAS 90–30-90 (based on the oxygen saturation, respiratory rate and systolic blood pressure) [8, 14, 18‐20]. Despite the prehospital screening tools and treatment guidelines that are available, there is still extensive patient morbidity and mortality due to sepsis [14]. A previous study indicate that hospital treated sepsis is the leading cause of mortality worldwide [3] and patients with sepsis arriving to the ED trough EMS were more likely to have severe grades of sepsis, and more often admitted to the ICU than those arriving by other means [5, 7].

Symptoms and vital signs tend to vary among patients with sepsis. For example, recent Swedish studies have indicated that symptoms as respiratory difficulty, an altered mental status, nausea, diarrhoea and/or vomiting, severe localized pain, muscle weakness, a lack of energy, fever and/or chills were common among patients with severe sepsis [21, 22]. However, in one previous study, only a small percentage of the patients with severe sepsis and septic shock were hypotensive according to the EMS records (14% with a systolic blood pressure of < 90 mmHg and 10% with a mean arterial pressure of < 65 mmHg) [23]. Another confounding vital sign is the body temperature, which does not rise in all patients with sepsis, making the diagnosis even more difficult [24]. Normal vital signs tend to be associated with a lower level of monitoring while the patient is being cared for in the ED, which may increase the risk of unnoticed deterioration and nonspecific symptoms tend to be associated with less favourable outcomes [25]. Vital signs and symptoms tend not to be specific to identify patients with sepsis and further research is required in the field.

Today, there is limited knowledge regarding the prehospital characteristics of patients with sepsis. To our knowledge, there is no previous study that has compared the prehospital characteristics of patients with sepsis with and without adverse outcomes during the hospital stay. Further research into the differences in the characteristics observed during the prehospital phase of sepsis is required, particularly in those patients at risk for poorer prognoses. Therefore, the objectives of the present study were to investigate the prehospital characteristics of patients with sepsis while they were being transported by the EMS, and to compare these values to those of the patients with and without adverse outcomes during their hospital stays.

In the present study, the sepsis identification rate of the EMS clinicians, as documented in the patients’ electronic health records, was found to be low. This result corresponds with those of previous studies regarding the EMS clinician assessments of patients with sepsis and severe sepsis [1, 6, 8, 20, 31]. Additionally, our study also compared the EMS sepsis identification rate in relation to the patient outcomes. Unfortunately, challenges still remain with regard to the early recognition of patients with sepsis, without or with adverse outcomes. The prehospital identification of patients with sepsis can be difficult to establish [32]. The reason for these difficulties can only be speculated. One reason is limited medical knowledge of sepsis among EMS clinicians which can undermine the ability to correctly identify conditions in patients with sepsis [31, 33, 34]. Another reason is that EMS clinicians have suspected other conditions [32] . Finally, the difficulties could be related to the working conditions and the fact that EMS clinicians prioritize care of the patients rather than the documentation of observation and measurements carried out in the patient assessment [6, 7, 35]. However, research has shown that sepsis recognition training combined with the use of a screening tool could improve the ability of EMS clinicians to look for sepsis and identify these patients [19]. This underscores the value of education and clinical training in identifying relevant conditions when caring for patients with sepsis. It could also be an opportunity for EMS clinicians to increase the survival and decrease the morbidity of patients with sepsis in much the same manner as they do with other time-critical, life-threatening conditions, such as acute myocardial infarctions [36] and strokes [37].

EMS clinicians play important roles in the early assessment and identification of patients with sepsis, especially those patients at risk for adverse outcomes. In this study, oxygen saturation, serum glucose level, body temperature, shivering and mental status were found in the analyses to be abnormal in patients with sepsis. These prehospital characteristics may be important in the prehospital setting, where access to laboratory testing is limited. For example, lower oxygen saturations were related to adverse outcomes in patients with sepsis in this study. A lower oxygen saturation is known to be an important vital sign in patients developing severe sepsis [8, 11, 21], and it has been used as a comarker in several screening tools, such as BAS 90–30-90 and PRESS score [8, 11]. Still, when investigating whether a oxygen saturation < 90% could predict a development of adverse outcome, there was no clear likelihood of this. This may indicate that low oxygen saturation as a vital sign could identify patients with sepsis who are at risk for adverse outcomes, however which limit value that could predict adverse outcome requires further research. Another prehospital characteristic that was associated with an adverse outcome was an increased serum glucose level. However, when investigating whether a serum glucose > 11 mmol/l could predict a development of adverse outcome, there was no clear likelihood of this. Hyperglycaemia in critical illness cases, such as severe sepsis, has been shown to be a comarker of the illness severity and a predictor of a poor outcome in the ICU [30, 38‐40]. A higher serum glucose level may be the result of the metabolic changes in the body caused by sepsis [41]. There have been no previous studies indicating that the serum glucose could be used during the prehospital assessment in patients suspected of having sepsis; therefore, further research is required in this area. The decrease in the oxygen saturation nor the increase in the serum glucose level seemed to be a result of a patient’s previous medical condition or comorbidity. When comparing the groups, there were no significant differences between them with regard to the prevalences of chronic obstructive pulmonary disease and diabetes mellitus, see Table 1.

An elevated body temperature has been associated with a lower in-hospital mortality rate in patients with severe sepsis. As the body temperature rises, so does the suspicion of sepsis, contributing to a faster recognition and treatment of this condition [42]. In the current study, those patients with adverse outcomes were found to have lower body temperatures when compared to those without adverse outcomes. Neither shows a temperature > 38 °C positive likelihood for predicting adverse outcome. A previous study indicating that many critically ill patients with sepsis do not have elevated body temperatures, [24] and hypothermia has been associated with sepsis-related in-hospital mortality [22]. However, further research is required on the body temperature in patients with sepsis and its impact on patients’ outcome.

In the present study patients with adverse outcome were found to have lower prevalence of shivering compared to those without adverse outcome and showed also a lower likelihood for predicting shivering in patients with adverse outcome. A previous study showed that a lower prevalence of shivering during the EMS period indicated a higher sepsis-related in-hospital mortality rate [22].

In the present study, an altered mental status was the only recognizable early symptom that was found to be associated with adverse patient outcomes in the analysis and indicated a positive likelihood to predict adverse outcome. These results are comparable to those of another study showing that an altered mental status during the EMS period indicated a higher sepsis-related in-hospital mortality rate [22]. Another study by Edman-Waller and colleagues [21] indicated that an altered mental status combined with a suspected infection should be taken as a warning sign that the patient may have or may be developing severe sepsis. Previous studies have advocated for acknowledging and integrating the symptom presentation into the sepsis identification screening tools [21, 22]. However, with the exception of an altered mental status, the results of the present study indicated that there may be some difficulty when using symptoms to identify patients with sepsis who are at risk for adverse outcomes.

This study was limited by the retrospective and convenience data collection from patients with sepsis in an epidemiological study database and symptoms and signs documented in patients electronic health records. This includes the inevitable uncertainty of assessing subjective symptoms based on the retrospective registration of data from these records. Another limitation was the small sample size of patients with adverse outcomes. This small sample size could lead to low statistical power and increases the risk that the null-hypothesis is rejected as false even though it is true.

Additionally, this study was conducted at a single centre, which could limit the generalizability of our findings. Another limitation was that data was collected from 2012. However, this was an explorative study in which the prehospital characteristics were described on the basis of a disease. Even though the presentation of sepsis may have been nonspecific [10], the specific characteristics have not changed over time. In several previous studies, the International Classification of Diseases (ICD) codes were used to define the diagnoses [20‐22, 43]; however, this is known to consistently underestimate the prevalence [27]. By retrospectively determining whether the patients fulfilled the sepsis diagnosis in the electronic health records according to a suspect infection and ≥ 2 SIRS criteria, a larger number of patients with sepsis may have been included. If only the ICD codes were used to identify the sepsis patients, there may have been a higher percentage of severe sepsis patients and patients with symptoms more typical of the common picture of sepsis, because these patients are more readily identified in the clinical setting. Therefore, our methods may have increased the generalizability of the results of this study.

More prospective studies are needed to determine whether the prehospital characteristics that were identified are representative of patients with sepsis with adverse outcomes in the prehospital setting, and to understand their predictive value.

The results of this study suggested that a low oxygen saturation and body temperature, high serum glucose level and altered mental status may be early prehospital characteristics that are related to poorer prognoses and adverse outcomes in patients with sepsis. The early recognition of prehospital characteristics by the EMS clinicians may affect the identification of these patients, and in the long run, the outcomes of the patients diagnosed with sepsis. However, it is not possible to predict limit values of saturation, body temperature and serum glucose to indicate adverse outcome in patients with sepsis. Further research is required to validate the use of these prehospital characteristics in the prehospital setting.