Infectious Diseases

Unexplained hypothermia is associated with bacterial infection in the Emergency Department

a b s t r a c t

Background: Early recognition and antibiotic therapy improve the prognosis of bacterial infections. Triage tem- perature in the Emergency department (ED) constitutes a diagnostic and prognostic marker of infection. The ob- jective of this study was to assess the prevalence of community-acquired bacterial infections and the Diagnostic ability of conventional Biological markers in patients presenting to the ED with hypothermia.

Methods: We conducted a retrospective single-center study over a 1-year period before the COVID-19 pandemic. Consecutive adult patients admitted to the ED with hypothermia (body temperature < 36.0 ?C) were eligible. Pa- tients with evident cause of hypothermia and patients with viral infections were excluded. Diagnosis of infection was based on the presence of at least two among the three following pre-defined criteria: (i) the presence of a potential source of infection, (ii) microbiology data, and (iii) patient outcome under antibiotic therapy. The asso- ciation between traditional biomarkers (white blood cells, lymphocytes, C-reactive protein [CRP], Neutrophil to lymphocyte count Ratio [NLCR]) and underlying bacterial infections was evaluated using a univariate and a mul- tivariate (logistic regression) analysis. Receiver operating characteristic curves were built to determine threshold values yielding the best sensitivity and specificity for each biomarker.

Results: Of 490 patients admitted to the ED with hypothermia during the study period, 281 were excluded for

circumstantial or viral origin, and 209 were finally studied (108 men; mean age: 73 +- 17 years). A bacterial infection was diagnosed in 59 patients (28%) and was mostly related to Gram-negative microorganisms (68%). The area under the curve (AUC) for the CRP level was 0.82 with a confidence interval (CI) ranging from 0.75 to 0.89. The AUC for the leukocyte, neutrophil and lymphocyte counts were 0.54 (CI: 0.45-0.64), 0.58 (CI: 0.48-0.68) and 0.74 (CI: 0.66-0.82), respectively. The AUC of NLCR and Quick Sequential Organ Failure Assessment reached 0.70 (CI: 0.61-0.79) and 0.61 (CI: 0.52-0.70), re- spectively. In the multivariate analysis, CRP >= 50 mg/L (OR: 9.39; 95% CI: 3.91-24.14; p < 0.01) and a NLCR >=10 (OR: 2.73; 95% CI: 1.20-6.12; p = 0.02) were identified as independent variables associated with the diagnosis of underlying bacterial infection.

Conclusion: Community-acquired bacterial infections represent one third of diagnoses in an Unselected population presenting to the ED with unexplained hypothermia. CRP level and NLCR appear useful for the diagnosis of causative bacterial infection.

(C) 2023

  1. Introduction

* Corresponding author at: Service d’Urgences, Centre Hospitalier Universitaire Dupuytren, 2, avenue Martin Luther King, 87042 Limoges cedex, France.

E-mail address: [email protected] (T. Lafon).

1 Present address: Emergency Department, Albi Hospital Center, F-81000 Albi, France.

Bacterial infections can be life-threatening conditions that constitute a frequent reason for admission to the Emergency Department (ED). Although early and accurate recognition is crucial to improve the progno- sis [1], diagnosis frequently remains challenging due to highly variable clinical presentations.

https://doi.org/10.1016/j.ajem.2023.06.037

0735-6757/(C) 2023

Body temperature is a vital parameter routinely measured upon ad- mission, which is used as a diagnostic and prognostic marker of infec- tion. The Systemic Inflammatory Response Syndrome and National Early Warning Score use body temperature to help identify potentially infected patients and to stratify their risk of deterio- ration in the ED [2]. However, body temperature is influenced by envi- ronmental parameters and varies widely during the pathogenic evolution of infection [3,4].

Hypothermia is reported in approximately 10% of bacterial infec- tions and can occur regardless of the severity of the inflammatory re- sponse [4]. The prevalence of bacterial infections in hypothermic patients presenting to the ED has been scarcely assessed. Although the Diagnostic and prognostic value of biomarkers such as C-reactive pro- tein (CRP), leukocyte and lymphocyte counts, or neutrophil to lympho- cyte count ratio (NLCR) have been extensively studied in septic patients [5], none of these biomarkers or scores has been evaluated in patients admitted to the ED with hypothermia.

Accordingly, in a representative cohort of patients admitted to the ED with unexplained hypothermia, we assessed the prevalence of bac- terial infections and the diagnostic performance of the biomarkers rou- tinely used to predict infection prior to the identification of the infectious source.

  1. Patients and methods

We conducted a retrospective single-center study over one year (January to December 2018) in the ED of a French tertiary-care Univer- sity Hospital Center (with an annual census of approximately 45,000 visits in the ED). All consecutive adult patients (18 years or older) with unexplained hypothermia (body temperature < 36.0 ?C) were el- igible for analysis. Patients with evident cause of hypothermia (external or accidental hypothermia) and viral infection were excluded. The diag- nosis of infection was confirmed by an adjudication committee (3 phy- sicians) and was based on the presence of at least two among the three following pre-defined criteria: (i) the presence of a potential source of infection, (ii) microbiology data, and (iii) patient outcome under antibi- otic therapy (if favorable evolution of patients was noted after antibiotic treatment, in the absence of a confounding factor, in particular an uncertain differential diagnosis) [6].

The following data were collected with the first blood sample ob- tained on ED admission: demographic criteria, vital signs (respiratory rate, body temperature, heart rate and blood pressure), quick Sequential Organ Failure Assessment score (qSOFA) (change in mental status, re- spiratory rate >= 22 breaths/min, systolic blood pressure <= 100 mmHg) [7], site of infection, identification of micro-organisms, and biological parameters (leukocyte count, lymphocyte count, CRP, lactates) [1,8]. Comorbidities including diabetes, chronic renal failure (creatinine clear- ance <=30 mL/min), chronic respiratory failure (long-term oxygen ther- apy), severe heart failure (left ventricular ejection fraction <=30%), occlusive arterial disease, stroke, and all-cause hospital mortality were also recorded.

Descriptive statistics, including mean (+- standard deviation [SD]) and frequency distributions were used to describe the cohort. Quantita- tive variables were compared using the Student’s t-test and proportions using the Chi-square test. The thresholds for biological markers to de- tect underlying bacterial infection considered for univariate analysis were: a CRP >= 50 mg/L [9]; a WBC count <4.0 G/L or >12.0 G/L [10]; a Neutrophil count >=10 G/L, a lymphocyte count <=1.0 G/L, or a NLCR >=10 [11]. receiver operating characteristic curves were built to deter- mine the best compromise between sensitivity and specificity of the CRP level, the leukocyte, neutrophil, and lymphocyte counts, the qSOFA score and the NLCR score, for the diagnosis of bacterial infection. For non-linear relation, Box-Tidwell regressions were performed; CRP and N/L were discretized in order to linearize them. After univariate lo- gistic regression analyses of the diagnostic predictive factors, variables with a p-value lower than 0.30 were included in a multivariate logistic

model. The initial multivariate model was simplified by a stepwise back- ward elimination method. Model calibration was assessed using Pear- son residual tests. A ROC curve analysis of the data was performed. A p-value <0.05 was considered significant. All calculations were per- formed using the R software, version 4.2.1 (R foundation for Statistical Computing, Vienna, Austria).

  1. Results

Body temperature was recorded in the medical charts of 7081 pa- tients who were admitted to the ED during the study period. Of these, 490 patients had hypothermia. An evident cause or circumstantial hy- pothermia was present in 274 patients, and seven patients were diag- nosed with a viral infection (Fig. 1). Finally, 209 patients were analyzed (108 men [52%]; mean age: 73 +- 17 years). Bacterial infection was diagnosed in 59 (28%) patients (35 men [59%]; mean age: 76 +- 15 years) (Fig. 1). Microbiological identification was obtained in 31 pa- tients (53%) with a predominance of Gram-negative microorganisms (68%) (Supplementary Table). Infection sites were predominantly the urinary tract (44%), lungs (27%), digestive tract (23%) and Soft tissues (7%). The overall hospital mortality rate was 19%.

Baseline characteristics were not statistically different between in- fected and non-infected patients, including age (76 +- 15 vs 72 +- 18 years old: p = 0.14), proportion of qSOFA score >= 2 (8% vs 7%: p = 0.64) and comorbidities. In-hospital mortality was significantly higher in the infected patients (29% vs 15%: p = 0.03) (Table 1). When compared to non-infected patients, patients with a bacterial infection exhibited a significantly higher CRP level (95 +- 107 vs 14 +- 28 mg/L: p < 0.01), a lower lymphocyte count (1.1 +- 0.7 vs 1.9 +- 1.6 G/L: p < 0.01), and a

greater NLCR (15.5 +- 21.7 vs. 6.1 +- 2.71: p < 0.01) (Table 2).

In univariate analysis, CRP >= 50 mg/L (OR: 13.5; 95% CI: 5.9-30.7; p < 0.01), neutrophil count >=10 G/L (OR: 2.4; 95% CI: 1.2-4.6; p < 0.01), lymphocyte count <=1.0 G/L (OR: 6.0; 95% CI: 3.1-11.6; p < 0.01), and NLCR >=10 (OR: 4.6; 95% CI: 2.3-9.0; p < 0.01) were significantly associated with a bacterial infection. In contrast, a qSOFA score >= 2 (OR: 1.3; 95% CI: 0.4-3.9; p = 0.72) was not associated with the diagnosis of bacterial infection (Table 3).

In the study cohort, 29 (49%) infected patients had a CRP level >= 50 mg/L (sensitivity 49%) against 10 (6%) in the non-infected group (specificity 93%). The positive predictive value (PPV) of CRP in di- agnosing infection was 74% against a negative predictive value (NPV) of 82%. Twenty-two (37%) infected patients had a leukocyte count above 12 G/L (sensitivity 37%) versus 41 (27%) patients in the non-infected group (specificity 72%). Using SIRS criteria as the cut-off defining nor- mal versus abnormal, the PPV of leukocyte count in diagnosing infection was 34% against a NPV of 75%. Moreover, 22 (37%) infected patients had a neutrophil count above an arbitrarily set cut-off of 10 G/L (sensitivity 37%) against 30 (20%) patients in the non-infected group (specificity 80%). Lymphopenia had a sensitivity of 61% and a specificity of 79%. PPV of neutrophil and lymphocyte counts were 42% and 53%, respec- tively, against a NPV of 76% and 83%, respectively (Table 3).

The AUC for the CRP level was 0.82 (CI: 0.75-0.89). The AUC for the leukocyte, neutrophil and lymphocyte counts were 0.54 (CI: 0.45-0.64),

0.58 (CI: 0.48-0.68) and 0.74 (CI: 0.66-0.82), respectively. The NLCR

had an AUC of 0.70 (CI: 0.61-0.79) and the qSOFA of 0.61 (CI: 0.52-0.70) (Fig. 2).

In the multivariate analysis, CRP >= 50 mg/L (OR: 9.39; 95% CI: 3.91-24.14; p < 0.01) and NLCR >=10 (OR: 2.73; 95% CI: 1.20-6.12;

p = 0.02) were identified as independent variables associated with the diagnosis of underlying bacterial infection in these patients with hy- pothermia on ED admission.

  1. Discussion

In this cohort of patients presenting to the ED with unexplained hy- pothermia, almost one third had a causative bacterial infection,

Image of Fig. 1

Fig. 1. Flowchart of the study.

* External cause: n = 112, massive hemorrhage: n = 49, anaphylaxis: n = 5, cardiac arrest: n = 33, drug poisoning: n = 52, status epilepticus: n = 15, miscellaneous: n = 8.

** Diagnosis of infection was based on the presence of two of the three following pre-defined criteria: (i) the presence of a potential source of infection, (ii) microbiology data (positive culture with identified germs for a suspected infection), and (iii) patient outcome under antibiotic therapy.

Table 1

Baseline clinical and Biological characteristics of the study population.

Total cohort

Infected patients

Non-infected patients

p-value

n = 209 (%)

n = 59 (%)

n = 150 (%)

Age (year)

73 +- 17

76 +- 15

72 +- 18

0.14

Male (%)

Co-morbidities:

108 (52)

35 (60)

73 (49)

0.17

predominantly due to Gram-negative bacilli. CRP level and NLCR ap- peared significantly associated with an underlying infection of bacterial origin.

Hypothermia identified in the ED may result from various parame- ters (e.g., method of measurement, exposure to cold environment) [12]. Normothermia and hypothermia in patients admitted to the ED with suspected sepsis are known to be significantly associated with an increased risk of in-hospital mortality and a lower rate of compliance with the Sepsis Bundles [12-14]. The mortality rate of the infected group in our study is in line with this higher mortality for septic pa- tients. This reflects the challenge of accurate recognition of bacterial in- fection in the ED, since the diagnosis is based on heterogeneous, non-

Diabetes

65 (31)

21 (36)

44 (29)

0.38

specific clinical and biological signs, especially in the presence of con-

Chronic renal failure Chronic respiratory failure

3 (1)

2 (1)

2 (3)

1 (2)

1 (1)

1 (1)

0.14

0.49

founding factors [8,15]. Our results represent the significant proportion of hypothermic patients whose hypothermia might be due to bacterial

Table 2

Chronic heart failure

10 (5)

3 (5)

7 (5)

0.90

Peripheral arterial disease

6 (3)

3 (5)

3 (2)

0.23

Cerebrovascular disease

17 (8)

4 (7)

13 (9)

0.65

Active blood disease

9 (4)

2 (3)

7 (5)

0.68

Active cancer

42 (20)

18 (31)

24 (16)

0.06

Clinical signs:

102 (49)

30 (51)

72 (48)

0.71

Biological characteristics of study population.

Systolic blood

Heart rate >= 90 bpm

67 (32)

15 (25)

52 (35)

0.20

Total cohort

Infected

Non-infected

p value

Respiratory rate >= 20 cpm

16 (8)

5 (8)

11 (7)

0.78

patients

patients

qSOFA >=2

15 (7)

5 (8)

10 (7)

0.64

n = 209 (%)

n = 59 (%)

n = 150 (%)

pressure <= 100 mmHg

23 (11)

10 (17)

13 (9)

0.08

Leukocytes (G/L)

10.8 +- 6.0

12.2 +- 8.4

10.3 +- 4.7

0.11

Respiratory rate >= 22 cpm

17 (8)

5 (8)

12 (8)

0.22

Lymphocytes

1.7 +- 1.4

1.1 +- 0.7

1.9 +- 1.6

< 0.01

Altered mental status

31 (15)

15 (25)

16 (11)

< 0.01

(G/L)

(GCS < 15)

Neutrophils (G/L)

8.5 +- 5.8

10.3 +- 8.4

7.8 +- 4.2

0.04

Mortality (n)

40 (19)

17 (29)

23 (15)

0.03

Platelets (103/uL)

243 +- 97

244 +- 113

243 +- 90

0.99

ICU admission (n)

37 (18)

7 (12)

30 (20)

0.16

CRP (mg/L)

36 +- 71

95 +- 107

14 +- 28

< 0.01

Lactates (mmol/L)

3.0 +- 2.8

3.5 +- 3.0

2.0 +- 2.4

0.07

Abbreviations: SIRS, Systemic Inflammatory Response Syndrome; CRP, C-Reactive Protein; qSOFA, quick Sepsis related Organ Failure Assessment; ICU, intensive care unit; GCS, Glas- gow coma scale; NLCR, Neutrophil to Lymphocyte Count Ratio.

Numbers in parentheses denote percentages.

NLCR 8.9 +- 9.2 15.5 +- 21.7 6.1 +- 2.7 < 0.01

NLCR, Neutrophil to Lymphocyte Count Ratio, CRP, C-Reactive Protein. Numbers in parentheses denote percentages.

Table 3 Diagnostic performance of conventional biological markers and usual scores used to iden- tify bacterial infection in patients with hypothermia on ED admission.

Univariate analysis Sensitivity Specificity PPV NPV OR p 95% CI (%) (%) (%) (%)

hypothermia [8,21]. In our cohort, a CRP level >= 50 mg/L was associated with bacterial infection and appeared to be the most effective and spe- cific Diagnostic biomarker. To a lesser extent, NLCR >=10 was also identi- fied as an independent factor of bacterial infection. Although the relation between these markers and infection has been previously de-

Lymphocyte count

6.0

< 0.01

3.1-11.6

61

79

53

83

NLCR

4.6

< 0.01

2.3-9.0

44

85

54

79

Score

qSOFA >=2

1.3

0.70

0.4-3.9

8

93

33

72

Abbreviations: SIRS, Systemic Inflammatory Response Syndrome. CRP, C-reactive protein. NLCR, Neutrophil to Lymphocyte Count Ratio. qSOFA, quick Sepsis related Organ Failure Assessment. OR, Odds ratio. CI, Confidence intervals, PPV, Positive Predictive Value. NPV, Negative Predictive Value.

Numbers in parentheses denote percentages.

* Sensitivity, specificity, positive predictive value and negative predictive value of the C-reactive protein level (>=50 mg/L), leukocyte count (>12 G/L), neutrophil count (>10 G/L), lymphocyte count (<1.0 G/L) and the neutrophil to lymphocyte count ratio (>10) in diagnosing infection.

infection (28%). Both SOFA and qSOFA scores are not accurate for the di- agnosis of infection [1], and they also appeared of limited interest for the diagnosis of a potential underlying infection in our patients with unex- plained hypothermia. SIRS criteria, known to have greater diagnostic ability than qSOFA for infection [16], also showed limited value in our patients with hypothermia.

Early and accurate recognition of infection, which diagnosis is often only presumptive in ED, is complex [17]. Minderhoud et al. [18] reported that a third of confirmed bacterial infection were underestimated by using SIRS criteria and clinical signs in a cohort of 269 patients with suspected sepsis in the ED. In the present study, the rate of positive culture was similar to the literature [19] as only approxi- mately 50% of infected patients had a microbiological identification of bacterial infection.

In our population, assessment of the performance of usual biomark- ers associated with infection seemed relevant, since microbiology testing does not provide adequate information within a suitable timeframe [20]. The use of conventional biomarkers (i.e., leukocyte and neutrophil counts) has shown limited diagnostic ability, with a low specificity for the identification of infection in patients with

lular destruction. Accordingly, the significance of an increased neutro- phil count in patients with hypothermia is challenging to interpret. Wyllie et al. [23] highlighted the clinical usefulness of lymphocytopenia as a diagnostic marker of bacteremia. Accordingly, the use of a compos- ite score analyzing white cell population on ED admission, appears clin- ically relevant: NLCR, which was previously explored for the prognosis of cancer and cardiovascular disease, showed even higher value in predicting bacteremia [11]. This marker is simple, easily obtained and included in the initial clinical evaluation of patients presenting to the ED with hypothermia.

In our study, the definition of hypothermia was based on SIRS criteria with a cut-off temperature of 36.0 ?C. Although the recruitment of the present cohort was exhaustive, we could not exclude unreported confounding factors predisposing to hypothermia (e.g., unreported cold exposure during pre-hospital transportation) because of the retrospec- tive design. Secondly, not all diagnoses of bacterial infection were based on a microbiological documentation, even though 50% of bacterial infec- tion are not confirmed by culture in practice [18,19]. Nevertheless, un- derlying infections were confirmed after hospital discharge depending on the microbiological results and patients’ outcome under adapted an- tibiotic therapy. Finally, this pragmatic study has been performed in the challenging clinical setting of patients presenting to the ED with unex- plained hypothermia.

Biology? CRP level

13.5

< 0.01

5.9-30.7

49

93

74

82

scribed, this association has been poorly studied in the ED settings, es-

pecially in patients with hypothermia [22]. Neutrophils are well

Leukocyte count

1.58

0.15

0.8-2.9

37

72

34

75

known to undergo wide variations in response to psychological stress,

Neutrophil count

2.38

< 0.01

1.2-4.6

37

80

42

76

such as infection. Hypothermia reduces basal metabolism, including cel-

  1. Conclusions

The prevalence of bacterial infections reached 28% in patients admit- ted to the ED with unexplained hypothermia and appeared as a leading cause. Irrespective of the reason for ED admission and clinical symp- toms, CRP level and NLCR yielded the highest diagnostic ability to iden- tify underlying bacterial infection in patients with hypothermia.

Funding / support

This research did not receive any specific grant from funding agen- cies in the public, commercial, or not-for-profit sectors.

100

80

Sensitivity%

60

40

20

0

0 20 40 60 80 100

100% – Specificity%

CRP

NLCR

Neutrophiles

Lymphocytes

Leucocytes

CRediT authorship contribution statement

Arthur Baisse: Writing – original draft, Investigation, Conceptualiza- tion. Simon Parreau: Writing – review & editing, Investigation. Stephanie Dumonteil: Writing – review & editing, Investigation. Alexandre Organista: Writing – review & editing, Investigation. Mathilde Alais: Writing – review & editing, Investigation. Vincent Ouradou: Writing – review & editing, Investigation. Rafaela Piras: Writing – review & editing, Investigation. Philippe Vignon: Writing – review & editing, Supervision, Investigation. Thomas Lafon: Writing – original draft, Investigation, Conceptualization.

Declaration of Competing Interest

the authors have no conflict of interest to declare.

Fig. 2. Receiver Operating Characteristic curves of infections biomarkers.

The AUC for the CRP level was 0.82 (CI = 0.75-0.89). The AUC for the leukocyte, neutro- phil and lymphocyte counts were 0.54 (CI = 0.45-0.64), 0.58 (CI = 0.48-0.68) and 0.74 (CI = 0.66-0.82) respectively. The NLCR had an AUC of 0.70 (CI = 0.61-0.79).

Abbreviations: AUC: Area Under the Curve; CRP: C-reactive protein; CI: Confidence Inter- val; NLCR: Neutrophil to Lymphocyte Count Ratio.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi. org/10.1016/j.ajem.2023.06.037.

References

  1. Evans L, Rhodes A, Alhazzani W, Antonelli M, Coopersmith CM, French C, et al. Sur- viving sepsis campaign: international guidelines for management of sepsis and sep- tic shock 2021. Crit Care Med. 2021;49:e1063-143. https://doi.org/10.1097/CCM. 0000000000005337.
  2. Schuttevaer R, Brink A, Alsma J, de Steenwinkel JEM, Verbon A, Schuit SCE, et al. The association of body temperature with antibiotic therapy and mortality in patients at- tending the emergency department with suspected infection. Eur J Emerg Med. 2021;28:440-7. https://doi.org/10.1097/MEJ.0000000000000817.
  3. Bhavani SV, Wolfe KS, Hrusch CL, Greenberg JA, Krishack PA, Lin J, et al. Temperature trajectory subphenotypes correlate with immune responses in patients with sepsis. Crit Care Med. 2020;48:1645-53. https://doi.org/10.1097/CCM.0000000000004610.
  4. Peiris AN, Jaroudi S, Gavin M. Hypothermia. JAMA. 2018;319:1290. https://doi.org/ 10.1001/jama.2018.0749.
  5. Ljungstrom L, Pernestig A-K, Jacobsson G, Andersson R, Usener B, Tilevik D. Diagnos- tic accuracy of procalcitonin, neutrophil-lymphocyte count ratio, C-reactive protein, and lactate in patients with suspected bacterial sepsis. PloS One. 2017;12:e0181704. https://doi.org/10.1371/journal.pone.0181704.
  6. Miller JM, Binnicker MJ, Campbell S, Carroll KC, Chapin KC, Gilligan PH, et al. A guide to utilization of the microbiology laboratory for diagnosis of infectious diseases: 2018 update by the infectious diseases society of America and the American Society for Microbiology. Clin Infect Dis. 2018;67:e1-94. https://doi.org/10.1093/cid/ciy381.
  7. Singer M, Deutschman CS, Seymour CW, Shankar-Hari M, Annane D, Bauer M, et al. The third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA. 2016;315:801-10. https://doi.org/10.1001/jama.2016.0287.
  8. Povoa P, Coelho L, Dal-Pizzol F, Ferrer R, Huttner A, Conway Morris A, et al. How to use biomarkers of infection or sepsis at the bedside: guide to clinicians. Intensive Care Med. 2023;49:142-53. https://doi.org/10.1007/s00134-022-06956-y.
  9. Reny J-L, Vuagnat A, Ract C, Benoit M-O, Safar M, Fagon J-Y. Diagnosis and follow-up of infections in intensive care patients: value of C-reactive protein compared with other clinical and biological variables. Crit Care Med. 2002;30:529-35. https://doi. org/10.1097/00003246-200203000-00006.
  10. Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, Knaus WA, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM consensus conference committee. American College of Chest Physi- cians/Society of Critical Care Medicine. Chest. 1992;101:1644-55. https://doi.org/10. 1378/chest.101.6.1644.
  11. de Jager CPC, van Wijk PTL, Mathoera RB, de Jongh-Leuvenink J, van der Poll T, Wever PC. Lymphocytopenia and neutrophil-lymphocyte count ratio predict bacter- emia better than conventional infection markers in an emergency care unit. Crit Care. 2010;14:R192. https://doi.org/10.1186/cc9309.
  12. Ramgopal S, Horvat CM, Adler MD. Association of triage hypothermia with in- hospital mortality among patients in the emergency department with suspected sepsis. J Crit Care. 2020;60:27-31. https://doi.org/10.1016/j.jcrc.2020.07.011.
  13. Park S, Jeon K, Oh DK, Choi EY, Seong GM, Heo J, et al. Normothermia in patients with sepsis who present to emergency departments is associated with low compliance with sepsis bundles and increased in-hospital mortality rate. Crit Care Med. 2020; 48:1462-70. https://doi.org/10.1097/CCM.0000000000004493.
  14. Yamaga S, Kawabata A, Hosokawa K, Shime N. Hypothermia, poorly recognized by clinicians, is associated with higher mortality among critically ill patients with infec- tions: a retrospective observational study. J Infect Chemother. 2021;27:540-3. https://doi.org/10.1016/j.jiac.2020.12.016.
  15. Filbin MR, Lynch J, Gillingham TD, Thorsen JE, Pasakarnis CL, Nepal S, et al. Present- ing symptoms independently predict mortality in septic shock: importance of a pre- viously unmeasured confounder. Crit Care Med. 2018;46:1592-9. https://doi.org/10. 1097/CCM.0000000000003260.
  16. Garcia O, Alvarez T, Granados S, Garzon V, Gonzalez S. Comparison of quick SOFA and SIRS scales at the bedside of patients with Staphylococcus aureus bacteremia. Biomedica. 2020;40:125-31. https://doi.org/10.7705/biomedica.4943.
  17. Brun-Buisson C, Meshaka P, Pinton P, Vallet B, EPISEPSIS Study Group. EPISEPSIS: a reappraisal of the epidemiology and outcome of severe sepsis in French intensive care units. Intensive Care Med. 2004;30:580-8. https://doi.org/10.1007/s00134- 003-2121-4.
  18. Minderhoud TC, Spruyt C, Huisman S, Oskam E, Schuit SCE, Levin MD. Microbiolog- ical outcomes and antibiotic overuse in emergency department patients with suspected sepsis. Neth J Med. 2017;75:196-203.
  19. Heffner AC, Horton JM, Marchick MR, Jones AE. Etiology of illness in patients with se- vere sepsis admitted to the hospital from the emergency department. Clin Infect Dis. 2010;50:814-20. https://doi.org/10.1086/650580.
  20. Shapiro NI, Wolfe RE, Moore RB, Smith E, Burdick E, Bates DW. Mortality in Emer- gency Department Sepsis (MEDS) score: a prospectively derived and validated clin- ical prediction rule. Crit Care Med. 2003;31:670-5. https://doi.org/10.1097/01.CCM. 0000054867.01688.D1.
  21. Barichello T, Generoso JS, Singer M, Dal-Pizzol F. Biomarkers for sepsis: more than just fever and leukocytosis-a narrative review. Crit Care. 2022;26:14. https://doi. org/10.1186/s13054-021-03862-5.
  22. Baisse A, Daix T, Hernandez Padilla AC, Jeannet R, Barraud O, Dalmay F, et al. High prevalence of infections in non-COVID-19 patients admitted to the Emergency De- partment with severe lymphopenia. BMC Infect Dis. 2022;22:295. https://doi.org/ 10.1186/s12879-022-07295-5.
  23. Wyllie DH, Bowler ICJW, Peto TEA. Relation between lymphopenia and bacteraemia in UK adults with Medical emergencies. J Clin Pathol. 2004;57:950-5. https://doi. org/10.1136/jcp.2004.017335.