Introduction

Sepsis is a common and life-threatening medical emergency and immediate medical attention is needed [1]. Nevertheless, health care personnel often fail to identify septic patients [2]. This may be explained, in part, by the non-specific presentation of sepsis [3,4,5]. Most of the sepsis screening tools in use today are based on vital signs, which is a limitation in the identification of the septic patient as approximately one third of the patients with serious infections do not present with fever and one third present with near normal vital signs [6, 7]. Hence, additional information, e.g., that of symptom presentation, can be assumed to be needed.

Identification of sepsis among older people is especially challenging as older patients often have comorbidities which may mask the underlying infection and sepsis [8]. Moreover, non-specific presentations of sepsis are more common among older people [8, 9]. There are also no prior studies of the effect of sex on symptom presentation in septic patients. Prior literature focuses mainly on pathophysiological mechanisms and differences in outcome between men and women while knowledge of symptom presentation in relation to sex is limited [10, 11].

Hence, we hypothesize that a better understanding of the presentation of the septic patient in relation to age and sex could add to the identification of the septic patient in the ED. Therefore, the aim of the current study was to analyze the prevalence of keywords describing the presentation of septic patients to the ED with respect to age and sex.

Methods

Study design

The current study is a retrospective, cross-sectional study. The prevalence of the 90 keywords, identified in a prior study [12] using a sequential mixed methods approach [13,14,15], were quantified among a cohort of adult patients presenting to the ED and discharged with an ICD-code compatible with sepsis. The patients were stratified with respect to age and sex. The overall study design used in the current study had a similar approach used in a previous study where sepsis presentation based on mode of arrival to the ED was explored [16].

Study setting

The study setting was the ED of Södersjukhuset, Stockholm, Sweden. Södersjukhuset was at the time the largest ED in all of the Nordic countries with 110,000 -120,000 ED visits annually [17].

Study population

Inclusion criteria

Patients ≥ 18 years of age and admitted to in-hospital care at Södersjukhuset via the ED between January 1st 2013 and December 31st 2013 and subsequently discharged with an ICD-10-code (International Classification of Disease) consistent with sepsis (A02.1, A22.7, A26.7, A32.7, A39.2, A39.4, A40.0 – A40.3, A48—A49, A41.0—A41.5, A41.8—A41.9, A42.7, B37.7, R57.2, R65.0–65.1) were included.

Exclusion criteria

Exclusion criteria were 1) interfacility transports of patients already treated for sepsis, 2) lack of ED records, 3) patients with healthcare-associated infections (HCAI), defined as a sepsis-onset ≥ 48 h after admission to in-hospital care. HCAI was determined by chart review of in-hospital records for all patients admitted to in-hospital care without signs or symptoms of infection or sepsis in the ED, but later developed signs and symptoms of infection or sepsis > 48 h after admission to in-hospital care and were subsequently discharged with an ICD-10 code compatible with sepsis.

Definitions

The current study used the term “keywords” as a hypernym for signs, symptoms and other factors describing the presentation of the septic patient to the ED. Examples of primary keywords are “abdominal pain”, “back pain”, “extremity pain” and “vomiting”, “diarrhea”, “obstipation” which together formed the combined keywords “pain” and “gastrointestinal symptoms” respectively.

Data collection and analysis

90 primary and 14 combined keywords in total, were quantified. Quantification of keywords was done in the following sections of the ED records documented by the admitting physician (IT based and documented after dictation): “chief complaint” (reflecting the patient’s own wording upon ED arrival), “current medical history” (reflecting onset of current symptomology) and “preliminary assessment” (final part of the ED record including assessment of reason for the patient to attend the ED including a plan forward). Symptoms with an onset within 3 weeks preceding the index ED visit were defined as “new” and considered relevant. Chart review of free text to assign keywords to patient cases was performed manually by two abstractors (AS and JS) under the supervision of (UW) who had participated in the prior study where the keywords were derived [12]. Disagreements regarding keyword categorization was discussed between the two abstractors and the supervising researcher until agreement was achieved.

Patients were stratified based on age and sex. Three age categories were created: < 65, 65–74 and ≥ 75 years of age based on cut-off values from Angus et al. [18]. Only keywords with a prevalence of 20% or more were presented in the manuscript. The keywords in their entirety were presented in supplemental tables.

Statistics

SPSS® (Statistical Package for Social Sciences SPSS®, IBM®, student version 22.0) was used for statistical analysis. The analysis of categorical values was performed using Chi-Square tests. Fisher’s exact test was used when expected cell count was < 5. P-values are presented in the tables without adjustment for multiple comparisons. The Bonferroni-adjusted significance level is presented in the footnote of each table, calculated by dividing the significance level 0.05 with the number of performed comparisons. Only two tailed P-values that remained significant according to the Bonferroni-adjusted significance level were considered significant in the current study.

Ethical approval

Stockholm Regional Ethical Review Board approval was obtained for this study (reference number 2012/1288–31/3). The study was performed according to the ethical standards of the 1964 Declaration of Helsinki and its later amendments. However, no written informed consent was obtained from the study participants as this was a register-based study and thus not needed according to Swedish practice [19].

Results

Characteristics

A total of 479 patients were included, see Fig. 1 for flow chart of inclusion and exclusion of patients. The median age was 75 years of age (IQR 61–85). 224 (46,8%) patients were women. 137 (28.6% were < 65 years of age, 102 (21.3%) were 65–74 years of age and 240 (50.1%) were ≥ 75 years of age. 99 (20.7%) patients died during in-hospital care. See Table 1 for patient characteristics.

Fig. 1
figure 1

Flow chart illustrating patient inclusion and exclusion. Flow chart for inclusion and exclusion of adult patients arriving to the ED of Södersjukhuset during 2013 and discharges with an ICD code compatible with sepsis. ED = Emergency Department, ICD = International Classification of Disease, HCAI = Health Care Associated Infection

Full size image
Table 1 Demographic data of patients with sepsis admitted to the ED of Södersjukhuset in 2013
Full size table

Keywords with a prevalence exceeding 20% in the entire group of ED patients

Twelve keywords had a prevalence of ≥ 20% in the entire cohort (n = 479): “abnormal, or suspected abnormal temperature” (n = 319, 66.6%), “pain” (n = 230, 48.0%), “abnormal breathing” (n = 210, 43.8%), “risk factors for sepsis” (n = 172, 36.0%), “abnormal circulation” (n = 163, 34.0%), “temporal deterioration” (n = 144, 30.1%), “gastrointestinal symptoms” (n = 137, 28.6%), “acute altered mental status” (n = 127, 26.5%), “abnormal skin” (n = 125, 26.1%), “abnormal urination” (n = 118, 24.6%), “loss of energy” (n = 113, 23.6%) and “decreased mobility” (n = 106, 22.1%), see Tables 2 and 3.

Table 2 Prevalence of keywords* exceeding 20% among ED patients discharged with ICD-10 code sepsis based on age
Full size table
Table 3 Prevalence of keywords* exceeding 20% among ED patients discharged with ICD-10 code sepsis based on sex
Full size table

99.6% of all septic patients had ≥ 1 of these twelve keywords, 96.9% had ≥ 2, 83.7% had ≥ 3 and 64.1% had ≥ 4 keywords.

Comparisons between age categories

For prevalence of all keywords, based on age, see Supplementary Table 1 (primary keywords) and Supplementary Table 2 (combined keywords).

“Pain” and “risk factors for sepsis” were significantly more common among patients < 65 years of age as compared with those ≥ 75 years of age: (n = 87/137; 63.5% vs n = 99/240; 41.3%, P-value < 0.000) and (n = 74/137; 54.0% vs 55/240; 22.9%, P-value < 0.000) respectively. “Risk factors for sepsis” was also significantly more common among patients 65–74 years of age as compared with those ≥ 75 years of age: (n = 43/102; 42.2% vs 55/240; 22.9%, P-value < 0.000), see Table 2.

Comparisons between the sexes

For prevalence of all keywords, based on sex, see Supplementary Table 3 (primary keywords) and Supplementary Table 4 (combined keywords).

“Pain” and “gastrointestinal symptoms” were significantly more common among women as compared with men: (n = 128/224; 57.1% vs n = 102/255; 40.0%, P-value < 0.000) and (n = 82/244; 36.6% vs n = 55/255; 21.6%, P-value < 0.000) respectively, see Table 3.

Discussion

The keywords “pain” and “risk factors for sepsis” were more common among younger patients and “pain” and “gastrointestinal symptoms” were more common among women. However, most keywords had a similar prevalence irrespective of age and sex.

Presentation in relation to age

Pain was more common among septic patients below the age of 65 as compared to those ≥ 75 years of age. Prior literature indicates that age does not affect pain tolerance per se, but that aging may decrease the sensitivity for some types of pain [20]. Also, older patients have, a greater prevalence of chronic conditions, e.g. dementia and post stroke, that might affect the ability to communicate symptoms [21]. Hence, we speculate that the older septic patients may have difficulties in expressing pain, as compared with the younger patients.

Patients below the age of 75 had a higher prevalence of the keyword “risk factors for sepsis”. The combined keyword “risk factors for sepsis” includes “external” risk factors for sepsis such as recent invasive procedures, substance abuse and ongoing antibiotic treatment. Older people are predisposed to infections due to a general decline in their physiological, anatomical and immunological defense against microbes [22], a.k.a. immunosenescence, which is likely to contribute to the overall higher prevalence of sepsis among older people [9]. We speculate that younger people may, so to speak, require an external risk factor for sepsis, which is not observed among older people. Approximately half of the patients included in the current study were above 75 years of age and had an in-hospital mortality of nearly one in three, which was substantially higher than that of the other age groups.

Presentation in relation to sex

Women had a higher prevalence of the keywords “pain” and “gastrointestinal symptoms”. The keyword “pain” involves all types of pain, without specific focus. However, abdominal pain was the most common type of pain. Previous literature indicates that women as a group may experience pain more often [23]. Also, previous literature indicate that women generally report more intense and more frequent bodily symptoms, while men have been shown to be less prone to admit feeling ill or experiencing pain [24]. The findings in the current study are in line with other literature that has investigated sex-related differences in the perception of pain in other acute medical conditions, which points to the need to include this information when assessing patient in the ED [25,26,27].

Women more frequently reported gastrointestinal symptoms which is supported by prior studies [28, 29]. Estrogen has previously been connected to nausea, vomiting and reduced appetite [30] and we speculate that it may play a role also in sepsis. However, the exact mechanisms explaining the difference in prevalence between the sexes remains unclear.

No keyword with a prevalence exceeding 20% was more common among men than women. It is possible that men underreport the presence of symptoms as compared to women (as discussed above), which may affect the current results.

Strengths and limitations

This is the first study, to our knowledge, to analyze differences in the presentation of sepsis to the ED with respect to age and sex. The understanding of the presentation including the symptomatology of sepsis in the ED may lead to an increased identification and enable timely treatment of the septic patient.

There are several limitations to the current study. The retrospective study design has inherent limitations e.g. missing data. Only predefined sections of the ED medical records were analyzed for keywords and are written by the admitting doctor. Due to interindividual variations in the competence and experience of the doctors working in the ED as well as a variations in work load the detail of the records may vary. Also, factors such as baseline co-morbidities, laboratory results and information as to which ward the patients were admitted to (e.g. ICU) were not retrieved at the time of data extraction which is a limitation of the current study. However, this was not the aim of the study.

Multiple comparisons were performed. Bonferroni corrections were therefore applied to adjust the levels of significance. However, the Bonferroni correction is overly strict and may infer a type II error, which may have resulted in true differences being regarded as non-significant.

Also, the generalizability of the results may be limited since this is a single center study. However, the ED of Södersjukhuset was at the time the largest in Scandinavia. Hence it is likely that the results are generalizable to other ED settings.

The inclusion criteria in current study were based on ICD-10 codes consistent with sepsis and ICD-10 coding is a crude method for identification of the septic patient, however, the only method applicable for registry studies.

Assignment of keywords is inherently in part subjective which can be viewed as a limitation. However, assignment of keywords as in the current study is in line with method [12,13,14,15,16].

Conclusion

The results show that “pain” is more common among septic patients below the age of 65 and “risk factors for sepsis” is more common among patients below the age of 75. Women with sepsis more often presented with pain and gastrointestinal symptoms than men. However, most keywords related to the presentation of septic patients to the ED had a similar distribution. The results could potentially be used to augment sepsis screening tools or clinical decision tools. Other tools e.g. machine learning may be a method with which to incorporate large amounts of data, including symptoms, to aid in the identification of sepsis.