A preliminary report of clinical experience in managing patients with sepsis and septic shock in emergency medicine ward

Abstract

Background:

Sepsis and septic shock are common causes of hospital admission, morbidity, and mortality, posing a significant burden on the health-care systems.

Objective:

The objective of this study was to report the clinical experience of management and outcomes of sepsis patients in the emergency medical ward of a community hospital. The risk factors associated with adverse outcomes of sepsis patients were also analyzed.

Methods:

This was a retrospective cohort study of patients with sepsis or septic shock managed in the emergency medical ward of North Lantau Hospital from 1 March 2015 to 31 March 2017. Their characteristics, clinical outcomes, risk factors associated with in-hospital mortality, 28-day mortality, and prolonged hospital stay (>14 days) were analyzed.

Results:

A total of 68 eligible patients met the inclusion criteria during the study period. The mean age of the patients was 73 (standard deviation, 16.7; range, 34–100) years. The mean Sequential Organ Failure Assessment score of all the cases was 4.5 (standard deviation, 2.4); range, 2–11). The most common source of infection was pneumonia (50%). During the stay in the emergency medical ward, 35 cases (49%) required vasopressor support for management of septic shock, and 12 cases required non-invasive ventilation (NIV) support. Five patients were eventually transferred to tertiary hospital (Princess Margaret Hospital) for further management. There were five in-hospital mortality cases and two 28-day mortality cases. From the univariate analysis, factors associated with in-hospital mortality included Sequential Organ Failure Assessment score >6 (p < 0.000), increasing number of organ dysfunction (p < 0.000), presence of chronic liver disease (p = 0.025), respiratory dysfunction during admission (p = 0.028); factors associated with 28-day mortality were advanced age (p < 0.000), increasing number of organ dysfunction (p = 0.033), presence of congestive heart failure (p = 0.004), and the presence of cancer (p = 0.034); factors associated with prolonged hospital stay were advanced age, presence of chronic obstructive airway disease (p = 0.003), advanced age (p = 0.041), and the use of NIV support (p = 0.001). In multivariate analysis, weak associations between in-hospital mortality and Sequential Organ Failure Assessment score >6 (p = 0.226) and increased number of organ dysfunction (p = 0.108) were demonstrated; there was a trend of prolonged length of stay with increased age (p = 0.139).

Conclusion:

Our experience and knowledge in managing sepsis patients in the emergency medical ward with implementation of critical care bed services increased significantly. These preliminary results demonstrated that, with appropriate patient selection, sepsis patients can be safely managed in the emergency medical ward. Further study with larger sample size is needed to identify risk factors of adverse outcomes in this group of patients managed in the emergency medical ward.

Keywords

Emergency medicine ward critical care sepsis outcomes risk

Introduction

Owing to the population aging, high prevalence of chronic medical diseases and increased expectation of health-care service from the public, the demand for critical care services is increasing. Management of critically ill patients is one the fundamental and essential roles of the Accident and Emergency Department (AED) in the health-care system. Emergency physicians are often responsible for providing resuscitation and critical care for patients in their early phase of critical illnesses. The current dual-fellowship training pathway of emergency medicine (EM) (Fellowship of Hong Kong College of Emergency Medicine) and intensive care medicine (ICM) (Fellowship of Hong Kong College of Anaesthesiologists (Intensive Care)) formalizes the qualified training of “emergency intensivists in Hong Kong and strengthens critical care delivery in the local AEDs.

North Lantau Hospital (NLTH) is a community hospital located in central Tung Chung of Lantau Island, which provides a 24-h accident and emergency service and various out-patient services serving the population of the Lantau Island and also the floating population in the Hong Kong International Airport at Chek Lap Kok. The AED of NLTH began services on 24 September 2013 and the Emergency Medicine Ward (EMW) began its service on 24 September 2014. The EMW of NLTH consists of 20 beds and shares the same ward with the Extended Care (EC) beds of the Medicine and Geriatrics Department of NLTH. The EMW serves all the acute in-patient service of the NLTH, while the EC beds mainly provide care for the convalescent patients from Princess Margaret Hospital (PMH). With the rapid development of Tung Chung and increasing population, the health-service need in the district is also increasing. Increasing number of complicated critical cases is encountered together with the increase in the number of patients attending the AED. As there is no acute medical ward and intensive care unit (ICU) service in NLTH, the responsibility of resuscitation and initial stabilization of critically ill patients entirely falls to the emergency physicians. After initial stabilization, critically ill patients are transported to the nearest tertiary hospital, PMH. The transport time is about 40 min and is sometime prolonged when an incident occurs in the highway. The long transport time always poses high risk to the critically ill patients and also a great challenge to the doctors and the nursing staff.

An “Emergency Critical Care Management” program was created in response to this challenge, with the objectives of extending the spectrum of patient care in the EMW, enhancing utilization of existing resource in NLTH, minimizing risk of patient transport, and ultimately improving the quality of patient care. In the EMW of NLTH, up to two beds are used for managing critically ill patients. The patients are under the care by the EMW team, which includes EM specialists with specialized training in intensive care and the EMW nursing staff who are further supervised by an emergency intensivist. Under this program, extended resuscitation is provided for critically ill patients whose conditions only need a short period of intensive care. The EMW also serves as a step down unit to minimize the risk of long-distant transport and also relieve the workload of PMH. These patients are directly transferred to the EC beds for convalescent care and rehabilitation after their acute phase of critical illnesses.¹

One of the main target groups of critically ill patients to be managed under the “Emergency Critical Care Management” program is patients with sepsis.

Exclusion criteria for selecting sepsis patients to be managed in the EMW included:

Patients requiring cardiac care unit and ICU care, defined as patients classified as priority 1 or 2 in the ICU Admission, Discharge, and Triage guidelines by the Society of Critical Care Medicine.²

Patients indicated for renal replacement therapy.

Patients requiring assessment by medical subspecialists, such as including cardiologists, hematologists, neurologists, rheumatologists and so on.

Patients with suspected surgical causes of infection.

Immunocompromised patients including cancer patients receiving chemotherapy, transplant recipients on immunosuppressant, patients on immunosuppressant for other indications and human immunodeficiency virus (HIV) patients.

Patients with decompensated liver cirrhosis.

Patient with high infectious risk, such as including (suspected) avian flu, tuberculosis, and chickenpox.

Pregnant patients.

Pediatric patients with age <18 years.

Training programs were provided to both emergency physicians and nurses including teaching programs about sepsis management, central venous catheterization (CVC), and basic ICU care. Standardized guideline for sepsis management (both in AED and EMW) has also been set up. This study aims to evaluate the characteristics of patients with severe sepsis, as evidenced by the presence of organ dysfunctions or septic shock, managed in our EMW under the “Emergency Critical Care Management” program and to identify the independent risk factors associated with those adverse clinical outcomes.

Methods

This study is designed as a retrospective cohort study. All patients who were admitted to EMW from 1 March 2015 to 31 March 2017, with sepsis or septic shock, together with known source of infection, as defined as positive microbiological workup, clinical or radiological evidence of infection were recruited. The diagnostic criteria for sepsis and septic shock would follow the recommendations made in the Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3),³ which stated that sepsis is a life-threatening organ dysfunction due to a dysregulated host response to infection, with organ dysfunctions defined as an increase of two or more points in the Sequential Organ Failure Assessment (SOFA) score. The previous term of “severe sepsis” has been replaced by the new definition of the SOFA score, and the term of “sepsis” used in our study already implies the presence of organ dysfunctions. Septic shock is a subset of sepsis in which underlying circulatory and cellular/metabolic abnormalities are profound enough to substantially increase mortality, its clinical criteria includes sepsis and persistent hypotension requiring vasopressors to maintain mean arterial pressure greater than or equal to 65 mmHg, which is equivalent to the previous definition.⁴ Only patients with severe sepsis or septic shock were included in the study. Sepsis patients with infection but no evidence of organ dysfunctions were not included in this study. Primary outcome(s) in this study would be in-hospital and 28-day mortality, with secondary outcome(s) being prolonged hospital stay >2 weeks, and rate of transfer-out to other hospital.

The AED and hospital records of each admission was reviewed, with the following data collected and compared between patients with measured outcome and those who did not have measured outcome: Age, gender, underlying chronic illnesses (e.g. coronary artery disease, congestive heart failure, chronic liver disease, chronic obstructive pulmonary disease, diabetes mellitus, hypertension), SOFA score, site of infection including bacteremia, respiratory, genitourinary, gastrointestinal, soft tissue, causative organisms, appropriateness of empirical antibiotics, time of antibiotic administration, type of organ dysfunctions, requirement of inotropic support, NIV support, or other ventilation supports.

Statistical analysis

The Statistical Package for the Social Sciences (Windows, version 19.0; SPSS, Inc, Chicago, IL, US) was used for all analyses. Descriptive statistics were used to summarize patient demographics. Univariate analysis was performed with the chi-square test or Fisher’s exact test for comparison of proportions and Student’s t-test for comparison of means. Independent risk factors for adverse outcomes, including in-hospital mortality, 28-day mortality, and prolonged hospital stay (>2 weeks), were evaluated by multivariate analysis and a p-value of <0.05 was considered statistically significant.

Ethical considerations

This study was approved by the clinical research ethics committees of Kowloon West Cluster of Hospital Authority.

Statement on informed consent: written informed consent was not necessary because no patient data have been included in the manuscript.

Results

Totally, 68 patients with sepsis, as evidenced by the presence of organ dysfunctions or septic shock, were managed in the EMW under the “Emergency Critical Care Management” program and were recruited in this study. The mean age of the patients was 73 ((standard deviation (SD), 16.7; range, 34–100) years, There were 40 males and 28 females. The mean SOFA score of the cohort of these cases was 4.5 (SD, 2.4; range, 2–11). During the stay in the EMW, 35 cases (49%) required vasopressor support for management of septic shock, and 12 cases required NIV support.

The most common source of infection was pneumonia (34 cases) (Table 1). Among all the cases with evidence of pneumonia (by clinical presentation and radiological findings), 17 of them had positive results in respiratory tract specimens. The most common organism identified in respiratory tract specimens was Pseudomonas aeruginosa (29%), followed by Klebsiella pneumoniae and Proteus mirabilis. Two patients had Influenza A polymerase chain reaction (PCR) test positive of their sputum. One patient was confirmed to suffer from Legionnaires’ disease by urine Legionella antigen test and sputum PCR test of Legionella pneumophila DNA. Totally, 23 patients developed bacteremia, and the most common organism identified in blood culture was Escherichia coli (57%), followed by K. pneumoniae (17%). One patient had positive serology test for scrub typhus. Except two female patients with E. coli bacteremia secondary to pelvis inflammatory disease, the sources of E. coli bacteremia of all the other cases were urinary tract infection. A total of 24 patients had evidences of urinary tract infection by clinical presentation, urine white blood cells or nitrite test, or radiological findings, and 15 of them had positive results in urine culture. The most common organisms identified in urine were E. coli (40%), followed by Proteus species (20%) and Enterococcus species (20%). Two patients had soft tissues infection, one of them had mixed growth withf P. aeruginosa, methicillin-resistant Staphylococcus aureus (MRSA), Bacteroides and yeast in sacral sore swab culture. The positive microbiology results in blood cultures, respiratory tract samples, and urine cultures were shown in Table 2.

Table 1.

Source of infection.

Source of infection	Number of cases	Percentage (%)
Lung	34	50
Genitourinary tract	26	38.2
GI	2	2.9
Soft tissue	2	2.9
Unknown	2	2.9
Hepatobiliary	1	1.5
Other	1	1.5
Total	68	100

GI: gastro intestinal.

Table 2.

Positive microbiology results in blood cultures, sputum, and urine samples.

Micro-organisms	Number
Blood cultures
Escherichia coli	13
Klebsiella pneumoniae	4
Proteus mirabilis	1
MRSA	1
Pseudomonas aeruginosa	1
Pseudomonas cepacia	1
MRSA + Escherichia coli	1
Enterococcus species	1
Total	23
Sputum samples
Pseudomonas aeruginosa	5
Klebsiella pneumoniae	3
Proteus mirabilis	2
Influenza A (sputum PCR)	2
Streptococcus agalactiae	1
MRSA	1
Group G Streptococcus	1
Legionella pneumophila (urine antigen + sputum PCR)	1
Haemophilus influenzae	1
Total	17
Urine cultures
Escherichia coli	6
Proteus species	3
Enterococcus species	3
Klebsiella pneumoniae	1
Pseudomonas aeruginosa	1
Candida albicans	1
Total	15

MRSA: methicillin-resistant Staphylococcus aureus; PCR: polymerase chain reaction.

All bacteremic patients without an apparent source of infection would have ultrasound assessment of abdomen to identify the intraabdominal source of infection. Around 10 of them had computer tomography (CT) of abdomen and pelvis done in NLTH after initial ultrasound screening. Nine of them had CT findings of sources of infection (Table 3). CT scans were arranged in four patients with hydronephrosis in the initial ultrasound assessments in order to identify the presence of obstructing stone. One elderly patient with poor premorbid state was confirmed to have an obstructing stone at his right kidney. He declined for surgical intervention, and he was treated with antibiotic therapy in the EMW. He eventually recovered and was discharged from the EMW. Two patients with Klebsiella bacteremia and septic shock were transferred to PMH for assessment by the ICU team. Both patients had CT scans of abdomen and pelvis performed in PMH, and one of them was confirmed to have liver abscess. One patient had CT scan of thorax performed in NLTH to delineate the nature of loculated parapneumonic effusion.

Table 3.

Ultrasound and CT scan findings of cases with CT scans of abdomen and pelvis performed in NLTH.

Admission condition	Ultrasound findings	CT scan findings
Pyelonephritis/Escherichia coli bacteremia	Mild right hydronephrosis	Right renal stone (nonobstructive) with features of pyelonephritis
Escherichia coli bacteremia	Unremarkable	Bilateral tubo-ovarian abscesses
Sepsis with acute kidney injury	Right hydronephrosis with bladder distension	Early emphysematous cystitis; mild right hydroureteronephrosis; no obstructing stone
Escherichia coli bacteremia	Unremarkable	Left acute pyelonephritis with infected or inflamed left ureter
Urosepsis/septic shock	Left hydronephrosis	Left pyelonephritis with hydronephrosis; no obstructing stone
Epigastric pain with septic shock	Unremarkable	Newly developed patchy consolidation over bilateral lungs suggestive of infection; no evidence of pulmonary embolism; no intraabdominal sepsis source, note: confirmed legionnaires’ disease
Klebsiella pneumoniae bacteremia	Gallbladder stones and sludge; common bile duct and intrahepatic duct not dilated	2.8-cm size abscess at splenorenal recess
Septic shock	Unremarkable	Bilateral pyelonephritis
Escherichia coli bacteremia	Unremarkable	No significant abnormality detected
Enterococcus bacteremia	Right side hydronephrosis	Right upper ureteric stone with right hydroureteronephrosis

In our cohort, 42 patients (62%) had received antibiotics in the AED before being admitted to the EMW. For those who had not received antibiotics in the EMW, most of them were elderly patients presented with non-specific symptoms initially including hyperglycemia, syncope, dyspnea, or decreased general condition. It highlighted that the variable presentation of sepsis, particularly in elderly patients and the importance of early sepsis recognition by attending clinicians could not be overstated.

Concerning the adverse outcomes, 5 patients died during in-hospital stay, 2 patients died within 28 days after hospital discharge, and 15 had prolonged hospital stay (>14 days). Excluding the five transfer-out cases, factors and predictors associated with those adverse outcomes were analyzed by univariate and multivariate analyses.

From the univariate analysis, factors associated with in-hospital mortality included SOFA score >6 (p < 0.000), increasing number of organ dysfunction (p < 0.000), presence of chronic liver disease (p = 0.025), respiratory dysfunction during admission (p = 0.028), and the presence of chronic liver disease (p < 0.05); factors associated with 28-day mortality were advanced age (p < 0.000), increasing number of organ dysfunction (p = 0.033), presence of congestive heart failure (p = 0.004), and the presence of cancer (p = 0.034). Although there are no statistically significant predictors for in-hospital mortality and 28 day mortality in multivariate analysis, weak associations between in-hospital mortality and SOFA score >6 (p = 0.226) and increased number of organ dysfunction (p = 0.108) were observed (Tables 4 to 6).

Table 4.

Univariate and multivariate analyses of factors associated with in-hospital mortality.

			In-hospital mortality		Univariate analysis (p-value)	Multivariate analysis (p-value)
			Yes	No	Univariate analysis (p-value)	Multivariate analysis (p-value)
Past health	DM	Yes	2	18	0.679	/
	DM	No	3	40	0.679	/
	HT	Yes	4	28	0.173	/
	HT	No	1	30	0.173	/
	Hyperlipidemia	Yes	1	15	0.773	/
	Hyperlipidemia	No	4	43	0.773	/
	CVA	Yes	2	25	0.893	/
	CVA	No	3	33	0.893	/
	IHD	Yes	1	10	0.876	/
	IHD	No	4	48	0.876	/
	CHF	Yes	0	4	1.000*	/
	CHF	No	5	54	1.000*	/
	AF	Yes	0	7	1.000*	/
	AF	No	5	51	1.000*	/
	Renal impairment	Yes	1	15	0.773	/
	Renal impairment	No	4	43	0.773	/
	COAD	Yes	0	7	0.609	/
	COAD	No	4	51	0.609	/
	Chronic liver disease	Yes	1	1	0.025	0.996
	Chronic liver disease	No	4	57	0.025	0.996
	Rheumatological diseases	Yes	1	2	0.095	0.997
	Rheumatological diseases	No	4	56	0.095	0.997
	Dementia	Yes	1	13	0.901	/
	Dementia	No	4	45	0.901	/
	Parkinson’s disease	Yes	0	4	1.000*	/
	Parkinson’s disease	No	5	54	1.000*	/
	Cancer	Yes	0	5	1.000*	/
	Cancer	No	5	53	1.000*	/
	Adrenal insufficiency	Yes	0	3	1.000*	/
	Adrenal insufficiency	No	5	55	1.000*	/
SOFA score >6		Yes	4	7	0.000	0.226
SOFA score >6		No	1	51	0.000	0.226
Age (mean (SD))			84.0 (10.6)	72.7 (17.1)	0.151^	/
No. of organ dysfunction (mean (SD))			3.8 (0.8)	2.1 (1.0)	0.000^	0.108
			In-hospital mortality		Univariate analysis (p-value)	Multivariate analysis (p-value)
			Yes	No	Univariate analysis (p-value)	Multivariate analysis (p-value)
Source of infection	Lung		4	28	0.143	/
	HBS		0	0
	GI		1	1
	GU		0	24
	Soft tissue		0	2
	Other		0	1
	Unknown		0	2
Renal dysfunction	Yes		4	24	0.095	0.565
Renal dysfunction	No		1	34	0.095	0.565
Liver dysfunction	Yes		3	15	0.105	/
Liver dysfunction	No		2	43	0.105	/
Respiratory dysfunction	Yes		4	18	0.028	0.996
Respiratory dysfunction	No		1	40	0.028	0.996
CVS dysfunction	Yes		4	38	0.510	/
CVS dysfunction	No		1	20	0.510	/
			In-hospital mortality		Univariate analysis (p-value)	Multivariate analysis (p-value)
			Yes	No	Univariate analysis (p-value)	Multivariate analysis (p-value)
Platelet dysfunction	Yes		3	21	0.293	/
Platelet dysfunction	No		2	37	0.293	/
CNS dysfunction	Yes		1	4	0.298	/
CNS dysfunction	No		4	54	0.298	/
Bacteremia	Yes		3	17	0.157	/
Bacteremia	No		2	41	0.157	/
Vasopressor use	Yes		4	29	0.197	/
Vasopressor use	No		1	29	0.197	/
NIV use	Yes		2	8	0.124	/
NIV use	No		3	50	0.124	/
First antibiotic in AED	Yes		3	38	0.804	/
First antibiotic in AED	No		2	20	0.804	/

AED: Accident and Emergency Department; CNS: central nervous system; AF: atrial fibrillation; CHF: congestive heart failure; CVA: cerebrovascular accident; CVS: cardiovascular system; DM: diabetes mellitus; IHD: ischaemic heart disease; GI: gastrointestinal; SOFA: Sequential Organ Failure Assessment; GU: genitourinary; CAOD: chronic obstructive airways disease; NIV: non-invasive ventilation.

Fisher’s exact test; excluding five cases being transferred to PMH.

^Independent t-test.

Table 5.

Univariate and multivariate analysis of factors associated with 28-day mortality.

			28-day mortality		Univariate analysis (p-value)	Multivariate analysis (p-value)
			Yes	No	Univariate analysis (p-value)	Multivariate analysis (p-value)
Past health	DM	Yes	0	18	1.000*
	DM	No	2	38	1.000*
	HT	Yes	2	26	0.229*
	HT	No	0	30	0.229*
	hyperlipidemia	Yes	1	14	0.428
	hyperlipidemia	No	1	42	0.428
	CVA	Yes	2	23	0.181*
	CVA	No	0	33	0.181*
	IHD	Yes	1	9	0.212
	IHD	No	1	47	0.212
	CHF	Yes	2	2	0.004*	0.997
	CHF	No	0	54		0.997
	AF	Yes	0	7	1.000*
	AF	No	2	49	1.000*
	Renal impairment	Yes	2	13	0.064*	0.997
	Renal impairment	No	0	43	0.064*	0.997
	COAD	Yes	0	7	1.000*
	COAD	No	2	49	1.000*
	Chronic liver disease	Yes	0	1	1.000*
	Chronic liver disease	No	2	55	1.000*
	Rheumatological diseases	Yes	0	2	1.000*
	Rheumatological diseases	No	2	54	1.000*
	Dementia	Yes	1	12	0.341
	Dementia	No	1	4	0.341
	Parkinson’s disease	Yes	0	4	1.000*
	Parkinson’s disease	No	2	52	1.000*
	Cancer	Yes	1	4	0.034	0.998
	Cancer	No	1	52		0.998
	Adrenal insufficiency	Yes	0	3	1.000*
	Adrenal insufficiency	No	2	53	1.000*
SOFA score >6		Yes	0	7	1.000*
SOFA score >6		No	2	49	1.000*
Age (mean (SD))			97.5 (0.7)	71.8(16.7)	0.000^	0.998
No. organ dysfunction (mean (SD))			3.5 (0.7)	2.0 (0.9)	0.033^	0.998
		28-day mortality			Univariate analysis (p-value)	Multivariate analysis (p-value)
		Yes	No		Univariate analysis (p-value)	Multivariate analysis (p-value)
Source of infection	Lung	2	26		0.818	/
	HBS	0	0
	GI	0	1
	GU	0	24
	Soft tissue	0	2
	Other	0	1
	Unknown	0	2
Renal dysfunction	Yes	2	22		0.167**	/
Renal dysfunction	No	0	34		0.167**	/
Liver dysfunction	Yes	0	15		1.000**	/
Liver dysfunction	No	2	41		1.000**	/
Respiratory dysfunction	Yes	2	16		0.115**	/
Respiratory dysfunction	No	0	40		0.115**	/
CVS dysfunction	Yes	1	37		0.638	/
CVS dysfunction	No	1	19		0.638	/
Platelet dysfunction	Yes	2	19		0.127**	/
Platelet dysfunction	No	0	37		0.127**	/
CNS dysfunction	Yes	0	4		1.000**	/
CNS dysfunction	No	2	52		1.000**	/
Bacteremia	Yes	0	17		1.000**	/
Bacteremia	No	2	39		1.000**	/
Vasopressor use	Yes	1	28		1.000	/
Vasopressor use	No	1	28		1.000	/
NIV use	Yes	0	8		1.000**	/
NIV use	No	2	48		1.000**	/
First antibiotic in AED	Yes	2	36		0.522**	/
First antibiotic in AED	No	0	20		0.522**	/

Fisher’s exact test, ^Independent t-test; excluding five cases being transferred out to PMH and five in-hospital mortality cases.

Fisher’s exact test; excluding five cases being transferred out to PMH and five in-hospital mortality cases.

Table 6.

Univariate and multivariate analyses of factors associated with prolonged hospital stay.

			Prolonged hospital stay		Univariate analysis (p-value)	Multivariate analysis (p-value)
			Yes	No	Univariate analysis (p-value)	Multivariate analysis (p-value)
Past health	DM	Yes	2	16	0.085	0.214
	DM	No	13	27	0.085	0.214
	HT	Yes	7	21	0.885	/
	HT	No	8	22	0.885	/
	Hyperlipidemia	Yes	2	13	0.198	/
	Hyperlipidemia	No	13	30	0.198	/
	CVA	Yes	7	18	0.746	/
	CVA	No	8	25	0.746	/
	IHD	Yes	3	7	0.743	/
	IHD	No	12	36	0.743	/
	CHF	Yes	0	4	0.564*	/
	CHF	No	15	39	0.564*	/
	AF	Yes	2	5	0.861	/
	AF	No	13	38	0.861	/
	Renal impairment	Yes	2	13	0.198	/
	Renal impairment	No	13	30	0.198	/
	COAD	Yes	5	2	0.003	0.451
	COAD	No	10	41	0.003	0.451
	Chronic liver disease	Yes	0	1	1.000*	/
	Chronic liver disease	No	15	42	1.000*	/
	Rheumatological diseases	Yes	0	2	0.395*	/
	Rheumatological diseases	No	15	41	0.395*	/
	Dementia	Yes	4	9	0.646	/
	Dementia	No	11	34	0.646	/
	Parkinson’s disease	Yes	2	2	0.253	/
	Parkinson’s disease	No	13	41	0.253	/
	Cancer	Yes	0	5	0.313*	/
	Cancer	No	15	38	0.313*	/
	Adrenal insufficiency	Yes	1	2	0.762	/
	Adrenal insufficiency	No	14	41	0.762	/
SOFA score >6		Yes	3	4	0.273	/
SOFA score >6		No	12	39	0.273	/
Age (mean (SD))			78.9 (11.1)	70.5 (18.3)	0.041^	0.139
No. of organ dysfunction			2.0 (1.1)	2.1 (0.9)	0.752^
			Prolonged hospital stay		Univariate analysis (p-value)	Multivariate analysis (p-value)
			Yes	No	Univariate analysis (p-value)	Multivariate analysis (p-value)
Source of infection	Lung		10	18	0.483	/
	HBS		0	0
	GI		0	1
	GU		4	20
	Soft tissue		0	2
	Other		0	1
	Unknown		1	1
Renal dysfunction	Yes		4	20	0.179	/
Renal dysfunction	No		11	23	0.179	/
Liver dysfunction	Yes		3	12	0.547	/
Liver dysfunction	No		12	31	0.547	/
Respiratory dysfunction	Yes		7	11	0.129	/
Respiratory dysfunction	No		8	32	0.129	/
CVS dysfunction	Yes		8	30	0.249	/
CVS dysfunction	No		7	13	0.249	/
Platelet dysfunction	Yes		6	15	0.723	/
Platelet dysfunction	No		9	28	0.723	/
CNS dysfunction	Yes		2	2	0.253	/
CNS dysfunction	No		13	41	0.253	/
			Prolonged hospital stay		Univariate analysis (p-value)	Multivariate analysis (p-value)
			Yes	No	Univariate analysis (p-value)	Multivariate analysis (p-value)
Bacteremia	Yes		4	13	0.794	/
Bacteremia	No		11	30	0.794	/
Vasopressor use	Yes		5	24	0.134	/
Vasopressor use	No		10	19	0.134	/
NIV use	Yes		6	2	0.001	0.204
NIV use	No		9	41	0.001	0.204
First antibiotic in AED	Yes		8	30	0.249	/
First antibiotic in AED	No		7	13	0.249	/

Fisher’s exact test, excluding three cases being transferred out to PMH and five in-hospital mortality cases.

^Independent t-test.

The five in-hospital mortality cases were reviewed. All were elderly patients with mean age of 84 (range, 72–94) years. All of them had marginal premorbid status prior to admission and had deterioration of their condition despite optimal treatment for sepsis. Consensus to carry out end-of-life care was made after discussing with the families. The hospital authority policy on “Do Not Attempt Cardio-Pulmonary Resuscitation” was adopted. For the two cases of 28-day mortality, both were in their advanced age. One of them died of recurrent pneumonia on subsequent admission, and another died of gastrointestinal bleeding from underlying carcinoma of stomach.

The average length of stay in EMW was 11.3 (range, 4–24) days. Around 10 cases were transferred to the EC beds, and the average length of stay in the EC beds was 12.1 (range, 5–24) days. In univariate analysis, factors associated with prolonged hospital stay were advanced age (p = 0.041), presence of chronic obstructive airway disease (p = 0.003), and the use of NIV support (p = 0.001). No significant predictor for prolonged hospital stay was identified in the multivariate analysis, but there was a trend of prolonged length of stay with increased age (p = 0.139). The length of stay for sepsis patients was longer than the usual EMW-targeted patients (<48 h), which may be accounted for by the need of completion of a course of antimicrobial therapy and rehabilitation after episode of acute illnesses, particularly for elderly patients. Indeed, all of them had reversal of septic shock within 2–3 days and weaned off NIV support within 3–5 days in the EMW. The prolonged length of stay for the sepsis patients in the EMW was mostly related to the shortage of EC beds for the convalescent patients. Nevertheless, the medical team together with the physiotherapy team and occupational therapy team were involved in the rehabilitation program for the deconditioned patients even when the patients could not be transferred to the EC beds due the shortage of resources. With better availability of rehabilitation resources, the length of stay of sepsis patients in the EMW could definitely be shortened.

Five cases were transferred to the tertiary hospital (PMH) for further management including one case of liver abscess requiring ultrasound-guided drainage; one case of emphysematous cystitis with acute retention of urine ( for urological intervention); one case of bilateral tubo-ovarian abscess (for gynecologist assessment); one cases of Klebsiella bacteremia complicated by diabetic ketoacidosis (DKA), septic shock, acute respiratory distress syndrome (ARDS), and multi-organ dysfunction; and one case of respiratory failure who failed NIV support and required intubation and mechanical ventilation.

Discussion

Sepsis poses a significant burden on health-care systems worldwide. In the United State, it was estimated that around 750,000 cases of sepsis patients were admitted to hospitals every year, and the incidence was projected to increase by 1.5% per annum. The average cost of healthcare per case was US$22,100.⁵ Severe sepsis (sepsis with organ dysfunction) and septic shock carry particularly high mortality rate (>50%).⁶ The key to successful management of patients with sepsis are early recognition, aggressive intervention, prompt antimicrobial therapy, and source control. Rivers et al.⁷ showed that early goal-directed hemodynamic optimization starting in the AED for patients with severe sepsis and septic shock significantly decreased the morbidity and mortality. Delayed in administration of appropriate antimicrobial therapy had been also shown to be associated with a measurable increase in mortality.^8–10

The in-hospital mortality in this cohort was about 8%. Under the selection criteria of sepsis patients managed in the EMW, a relatively less severe group of patients was selected as indicated by the mean SOFA score of 4.5. The reported mortality rate of SOFA score <6 was less than 10% which was compatible to our result.^11,12 About half of the patients in this cohort developed septic shock and almost all of them had CVC placement for the administration of vasopressor. With the strict implementation of CVC care bundle in the EMW, none of the patients developed complications from CVC placement, including catheter associated bloodstream infection. Our preliminary result demonstrated the feasibility of implementing critical care services for sepsis patients in the EMW staffed by emergency physicians with specialized training in critical care.

In this study, the most common site of infection was respiratory tract which was similar to other epidemiology studies on sepsis patients in the ICUs.^13,14

About one-third of patients had bacteremia, and the most common organism found in the blood cultures was E. coli. Similar results had been reported in other study in both in-patient setting¹⁵ and out-patient setting.¹⁶ All the cases with E. coli bacteremia responded well with early antibiotic therapy and aggressive resuscitation for hemodynamic stabilization. An elderly patient with underlying liver cirrhosis suffering from spontaneous bacterial peritonitis complicated by E. coli bacteremia and septic shock recovered with treatment in the EMW. However, he subsequently deteriorated and died of decompensated liver failure. There was no in-hospital mortality in all the other cases of E. coli bacteremia. Four patients had Klebsiella bacteremia, and three of them had severe complications including severe DKA, liver abscess, and lung empyema. The case with Klebsiella bacteremia complicated by lung empyema was an elderly patient with marginal premorbid state. His condition initially improved with drainage of empyema and antibiotic therapy. However, he eventually died of sputum retention and hospital acquired pneumonia. Another patient with Klebsiella bacteremia was found to have an abscess in the splenorenal recess. In view of the small size of the abscess, she was treated conservatively with prolonged course of antibiotic therapy, and she eventually recovered. Klebsiella bacteremia is often associated with metastatic infection and carries a much higher mortality.¹⁷ This group of sepsis patients often requires more resources in the management including imaging, surgical assessment and intervention, and even organ supportive therapies in the ICU.

Source control is one of the essential components in sepsis management and radiological investigations are often required. Ultrasound is a very important screening tool to identify source of infection in sepsis patients. However, it alone is sometimes not sufficient as shown in this study. Thus, the availability of urgent CT scan is very important in the management of sepsis patients.

There are several limitations in our study. First, our study was a retrospective cohort study, and the data were retrieved from the clinical notes and records retrospectively. Incomplete data entry was encountered, and the most common missing data were the value of blood gases which was one of the variables in calculating the SOFA score. In our study, missing data for the SOFA score were assumed to be zero point. Data about PaO2 were missed in about 30% of the cases. This may underestimate the true severity of disease in the patient group. In addition, most of the patients in this study were elderly and had pre-existing cognitive impairment. This could cause difficulty in assessing their Glasgow Coma Scale (GCS) accurately leading to error in data collection. Second, some of the patients meeting the admission criteria were transferred to PMH due to the limitation in the bed availability and manpower. This could cause selection bias in the results. Another major limitation for the statistical analysis was the small sample size in this study. Although, the multivariate analyses failed to identify any significant factors associated with adverse outcomes, the results should be interpreted with caution. Further study should be conducted when larger sample size is available.

Conclusion

Sepsis is a significant health-care burden worldwide. AED is often the first point of contact for patients presenting with sepsis, and early screening in the AED often plays a key role in successful management of sepsis patients. The implementation of critical care service in the EMW enhanced the knowledge and experience in managing sepsis patients by the emergency physicians. From our limited data, we have shown that with appropriate screening and patient selection, sepsis patients could be successfully managed by the EMW medical team with proper training in critical care.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Availability of data and materials

All patients’ identifications were removed before storage. The master data set was kept by the investigators in computer, and a password was required to open the file. Access to the data would be limited to the investigators and clinical research ethics committees of Kowloon West Cluster, Hong Kong Hospital Authority. Please contact principal investigator should further information of the data is required.

Informed consent

Informed consent was waived by the clinical research ethics committees of Kowloon West Cluster, Hong Kong Hospital Authority; Written informed consent was not necessary because no patient data have been included in the manuscript.

Ethical approval

This study was approved by the clinical research ethics committees of Kowloon West Cluster, Hong Kong Hospital Authority.

Human rights

The study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki.

References

Wong

Cheung

et al . Emergency critical care service in emergency medicine ward: the experience in North Lantau Hospital. Hong Kong J Emerg Med 2017; 24(1): 25–34.

Nates

Nunnally

Kleinpell

et al . ICU admission, discharge, and triage guidelines: a framework to enhance clinical operations, development of institutional policies, and further research. Crit Care Med 2016; 44(8): 1553–1602.

Singer

Deutschman

Seymour

et al . The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA 2016; 315(8): 801–810.

Levy

Fink

Marshall

et al . 2001 SCCM/ESICM/ACCP/ATS/SIS international sepsis definitions conference. Intensive Care Med 2003; 29(4): 530–538.

Angus

Linde-Zwirble

Lidicker

et al . Epidemiology of severe sepsis in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med 2001; 29(7): 1303–1310.

Angus

Wax

RS.

Epidemiology of sepsis: an update. Crit Care Med 2001; 29(7 suppl.): S109–S116.

Rivers

Nguyen

Havstad

et al . Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med 2001; 345(19): 1368–1377.

Kumar

Roberts

Wood

et al . Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock. Crit Care Med 2006; 34(6): 1589–1596.

Ferrer

Martin-Loeches

Phillips

et al . Empiric antibiotic treatment reduces mortality in severe sepsis and septic shock from the first hour: results from a guideline-based performance improvement program. Crit Care Med 2014; 42(8): 1749–1755.

10.

Garnacho-Montero

Gutiérrez-Pizarraya

Escoresca-Ortega

et al . Adequate antibiotic therapy prior to ICU admission in patients with severe sepsis and septic shock reduces hospital mortality. Crit Care 2015; 19: 302.

11.

Vincent

De Mendonça

Cantraine

et al . Use of the SOFA score to assess the incidence of organ dysfunction/failure in intensive care units: results of a multicenter, prospective study. Working group on “sepsis-related problems” of the European Society of Intensive Care Medicine. Crit Care Med 1998; 26(11): 1793–1800.

12.

Ferreira

Bota

Bross

et al . Serial evaluation of the SOFA score to predict outcome in critically ill patients. JAMA 2001; 286(14): 1754–1758.

13.

Sands

Bates

Lanken

et al . Epidemiology of sepsis syndrome in 8 academic medical centers. JAMA 1997; 278(3): 234–240.

14.

Brun-Buisson

Doyon

Carlet

et al . Incidence, risk factors, and outcome of severe sepsis and septic shock in adults. A multicenter prospective study in intensive care units. French ICU Group for Severe Sepsis. JAMA 1995; 274(12): 968–974.

15.

French

Cheng

Duthie

et al . Septicaemia in Hong Kong. J Antimicrob Chemother 1990; 25(suppl. C): 115–125.

16.

Ling

Lam

Chan

EW.

What have we learnt from community-acquired infections in Hong Kong?

J Antimicrob Chemother 2003; 51(4): 895–904.

17.

Watanakunakorn

Jura

Klebsiella bacteremia: a review of 196 episodes during a decade (1980–1989). Scand J Infect Dis 1991; 23(4): 399–405.