Observational Study

Assessment of a Cellular Host Response Test as a Sepsis Diagnostic for Those With Suspected Infection in the Emergency Department

O’Neal, Hollis R. Jr , MD, MSc^1–3; Sheybani, Roya PhD⁴; Caffery, Terrell S. MD^1,2; Musso, Mandi W. PhD^1,2; Hamer, Diana PhD^1,2; Alwood, Shannon M. MD^1,2; Berlinger, Matthew S. MD^1,2; Jagneaux, Tonya MD^1–3; LaVie, Katherine W. MD^1,2; O’Neal, Catherine S. MD^1,2; Sanchez, Michael A. MD^1–3; Walker, Morgan K. MD^1,2; Shah, Ajay M. PhD⁴; Tse, Henry T. K. PhD⁴; Thomas, Christopher B. MD^1–3

Author Information

¹ Louisiana State University Health Sciences Center, Baton Rouge, LA.

² Our Lady of the Lake Regional Medical Center, Baton Rouge, LA.

³ Baton Rouge General Medical Center, Baton Rouge, LA.

⁴ Cytovale, Inc., San Francisco, CA.

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (http://journals.lww.com/ccejournal).

Cytovale, Inc. funded this research, in part through Federal funds from the Department of Health and Human Services; Office of the Assistant Secretary for Preparedness and Response; Biomedical Advanced Research and Development Authority, under Contract No. HSO100201900010C.

Drs. Sheybani, Shah, and Tse, are employees of Cytovale and played a role in the design, analysis, interpretation, writing, and decision to publish this study. Patient consent and medical direction for the study was conducted independently at the study facilities under the direction of Drs. O’Neal, Caffery, and Thomas who have no financial interest in Cytovale. The authors affiliated with Cytovale have an equity interest in the company whose device is the subject of this study. The other authors declare that they have no competing financial interests. The system that is the subject of this study is for investigational use only (not yet cleared for commercial use) and is covered by the following patents: 8,361,415; 8,935,098; 2,619,545; and 9,897,532.

This work was performed at Our Lady of the Lake Regional Medical Center, Baton Rouge, LA, and Baton Rouge General Medical Center, Baton Rouge, LA.

This is an open-access article distributed under the terms of the Creative Commons Attribution-Non Commercial-No Derivatives License 4.0 (CCBY-NC-ND), where it is permissible to download and share the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

Critical Care Explorations 3(6):p e0460, June 2021. | DOI: 10.1097/CCE.0000000000000460

Open
SDC

Abstract

Sepsis is a common emergency medical condition (¹^,²), and delays in treatment are detrimental to patients and costly to healthcare systems (^3–5). Most cases of sepsis arise in the community (⁶), so the initial diagnosis and treatment frequently occurs in the emergency department (ED) where time is short and information is limited. In addition, the definition of sepsis is dynamic, evolving from the 1992 consensus definition (⁷), to the current Sepsis-3 definition of “life-threatening organ dysfunction caused by a dysregulated host response to infection” (⁸). Despite the evolving definition, sepsis remains a clinical syndrome, identified by a constellation of signs and symptoms, without a gold-standard test for diagnosis.

Infection by a pathogen and activation of the host innate immunity are fundamental processes of sepsis pathogenesis. During activation, the cells of innate immunity—monocytes and neutrophils—undergo biochemical and biophysical changes that allow them to fulfill their function (⁹^,¹⁰). Although progress has been made in the rapid identification of pathogens (¹¹), there are limited options available for the quantification of the biophysical changes characteristic of the host response (¹²). Historically, the quantification of this response required laborious and time-consuming assays, which have only been assessed in the critically ill (¹³^,¹⁴). Our early work demonstrated that deformability cytometry that quantifies innate immune activation may be used to distinguish septic from nonseptic patients (¹⁵). More recently, we have expanded this work to develop the IntelliSep Index (ISI), a quantitative measure of the changes in cellular structure exhibited during the activation of the host immunity, which distinguishes sepsis from nonsepsis with an area under the curve of greater than or equal to 0.9 (¹⁶).

This study evaluates the utility of the ISI for diagnosing sepsis in the ED. We performed a prospective, observational study on a cohort of patients presenting to the ED with signs or suspicion of infection for assessment of the ISI as a diagnostic and prognostic marker for sepsis. We hypothesized that patients with a higher ISI would be more likely to have sepsis and associated increases in severity of illness.

METHODS

Study Population

Between April 22, 2019 and September 13, 2019, we enrolled 289 adult patients in a prospective, observational study at two hospitals in Baton Rouge, LA. Inclusion criteria were signs or suspicion of infection, with signs of infection defined as two or more modified (systemic inflammatory response syndrome [SIRS]) criteria (with at least one being aberration of temperature or WBC count) and suspicion of infection defined as a clinician order for culture of a body fluid (e.g., blood, urine, sputum). We excluded patients with an expected palliative course; history of hematologic disorders, receipt of cytotoxic chemotherapy within 3 months of the ED encounter, prisoners, patients transferred from other acute care facilities, and patients unwilling or unable to consent. We included 266 subjects in the final analysis (Fig. S1, Supplemental Digital Content 1, https://links.lww.com/CCX/A678). The study protocol was approved by the Louisiana State University Health Sciences Center Institutional Review Board (IRB; LSUHSC-NO number 19-019) as well as by local, hospital-specific IRBs (Franciscan Missionaries of Our Lady University IRB: number 2019-012 and Baton Rouge General IRB: number 2018-017). The study team obtained written informed consent from all subjects.

ISI Performance and Results

Upon consent, coordinators obtained EDTA-anticoagulated peripheral blood samples for the IntelliSep test. A subset of patients consented for optional storage of a specimen for later measurement of alternative biomarkers, including procalcitonin. Clinical management proceeded via standard care. The IntelliSep test (Cytovale, San Francisco, CA) requires 100 microliters of whole blood (Fig. S2, Supplemental Digital Content 1, https://links.lww.com/CCX/A678). It assesses intracellular and nuclear changes that occur during leukocyte activation and provides a single score, the ISI, within 10 minutes. The ISI ranges from 0.1 to 10.0 (inclusive) and is stratified into three discrete interpretation bands of risk for sepsis: green (low risk, 0.1–5.4), yellow (intermediate risk, 5.5–6.7), and red (high risk, 6.8–10.0). The algorithm for calculating the ISI and limits for interpretation bands were derived from prior investigation (¹⁶).

Data Collection

Coordinators collected radiographic, historical, physical, and laboratory data for assignment of the presence of infection based on predefined criteria derived from the Centers for Disease Control and Prevention’s National Healthcare Safety Network criteria (¹⁷) with minor modifications. As the purpose of these definitions is to define infections for surveillance, the definitions were modified to remove surveillance-specific language and be applicable to the ED setting. Coordinators also calculated baseline and daily Sequential Organ Failure Assessment (SOFA) scores (¹⁸), for each of the first 3 days of hospitalization.

Confirmation of the Diagnosis

Because there is no reference standard for the diagnosis of sepsis (¹⁹), we relied on a structured adjudication process for determining the presence or absence of the diagnosis by Sepsis-3 criteria. Because of known clinician variability in the diagnosis of sepsis (²⁰), each case underwent a rigorous, three-tiered process of adjudication: objective review (predetermined criteria for infection and SOFA calculations) and site-level clinical review were performed by on-site investigators. These data were compiled and electronically and transmitted to an external, independent adjudication committee for final determination using the Research Electronic Data Capture electronic data capturing tool (²¹). Study personnel in every tier of the process were blinded to the ISI results. The adjudication committee determined each case to be “Sepsis” or “Not sepsis” with “unanimous,” “consensus,” or “forced” determinations. Analyses of the adjudicated endpoint included each population. Details of the patients requiring a forced adjudication are in Figure S3 (Supplemental Digital Content 1, https://links.lww.com/CCX/A678).

Statistical Analysis

Unless otherwise stated, p values are derived from an unpaired two-sample Welch’s t test, where the null hypothesis is that the mean of the two samples is equal. Descriptive statistics are presented as means, sds, medians, and interquartile ranges (Q1–Q3) for the continuous variables and as counts and percentages for categorical variables. An alpha level of 5% is used for all analyses, unless otherwise stated. Two-sided CIs for proportions are provided using the Clopper-Pearson method, where appropriate. Finally, we constructed receiver operating characteristic (ROC) curves to illustrate the performance of the ISI for classification of patients as septic or not septic.

RESULTS

Table 1 contains the baseline characteristics of all 266 patients analyzed. The median age was 57 years (Q1–Q3, 40–72 yr), with 115 male (43.2%) and 151 female (56.8%). Of the total, 144 were White (54.1%), 115 were Black (43.2%), and seven were members of other races (2.6%). Regarding inclusion criteria, 109 (41%) were enrolled with the SIRS criteria, 110 (41.4%) with culture criteria, and 47 (17.6%) with both.

TABLE 1. - Baseline Characteristics of All Patients in Total as Well as All Patients Stratified by Interpretation Band

Characteristics		Total, N = 266	Green Band, N = 143	Yellow Band, N = 68	Red Band, N = 55	p
Age, median (Q1–Q3)		57 (40–72)	55 (38–72)	59 (41–69)	61 (47–77)	NS
Age (≥ 65), n (%)		103 (38.7)	51 (35.7)	25 (36.8)	27 (49.1)	NS
Gender (female), n (%)		151 (56.8)	86 (60.1)	38 (55.9)	27 (49.1)	NS
Race, n (%)	White	144 (54.1)	80 (55.9)	31 (45.6)	33 (60.0)	NS
	African American	115 (43.2)	60 (42.0)	33 (48.5)	22 (40.0)
	Other	7 (2.6)	3 (2.1)	4 (5.9)	0 (0.0)
Comorbidities, n (%)	Hypertension	155 (58.3)	72 (50.3)^a	43 (63.2)	40 (72.7)^a	< 0.01^a
	Diabetes	72 (27.1)	36 (25.2)	20 (29.4)	16 (29.1)	NS
	Obesity (body mass index ≥ 40)	39 (14.7)	19 (13.3)	13 (19.1)	7 (12.7)	NS
	Cancer	37 (13.9)	18 (12.6)	11 (16.2)	8 (14.5)	NS
	Chronic kidney disease	31 (11.7)	15 (10.5)	9 (13.2)	7 (12.7)	NS
	Autoimmune disease	13 (4.9)	5 (3.5)	4 (5.9)	4 (7.3)	NS
Infected—meeting objective criteria by organ system, n (%)	Respiratory	30 (11.3)	9 (6.3)^a	9 (13.2)	12 (21.8)^a	< 0.05^a
	Gastrointestinal	16 (6.0)	4 (2.8)	7 (10.3)	5 (9.1)	NS
	Urinary	34 (12.8)	18 (12.6)	5 (7.4)^a	11 (20.0)^a	< 0.05^a
	Skin	24 (9.0)	8 (5.6)	10 (14.7)	6 (10.9)	NS
	Other	12 (4.5)	1 (0.7)^a,b	5 (7.3)^b	6 (10.9)^a	< 0.05^a; < 0.05^b
	Not Infected	161 (60.5)	105 (73.4)^a,b	34 (50.0)^a	22 (40.0)^b	< 0.01^a; < 0.001^b
Infected by adjudication, n (%)		112 (42.1)	38 (26.6)^a,b	37 (54.4)^a	37 (67.3)^b	< 0.001^a; < 0.001^b
Organ dysfunction by adjudication, n (%)		119 (44.7)	53 (37.1)^a	32 (47.1)	34 (61.8)^a	< 0.01^a
Infected with organ dysfunction by adjudication, n (%)		55 (20.7)	11 (7.7)^a,b	19 (27.9)^a,b	25 (45.5)^b,c	< 0.001^a; < 0.001^b; < 0.05^c
Septic by Sepsis-3 definition, n (%)		48 (18.0)	6 (4.2)^a,b	17 (25.0)^a,c	25 (45.5)^b,c	< 0.001^a; < 0.001^b; < 0.05^c
Adjudication, n (%)	Unanimous	215 (80.8)	121 (84.6)	50 (73.5)	44 (80.0)	NS
	Consensus	40 (15.0)	16 (11.1)	16 (23.5)	8 (14.5)	NS
	Forced	11 (4.1)	6 (4.2)	2 (2.9)	3 (5.5)	NS
Blood culture, n (%)	Number tested	113 (42.5)	37 (25.9)^a,b	33 (48.5)^a,c	43 (78.2)^b,c	< 0.01^a; < 0.001^b; < 0.001^c
Blood culture, n (%)	Number positive (of tested)	24 (21.2)	4 (10.8)^a,b	8 (24.2)^a	12 (27.9)^b	< 0.05^a; < 0.01^b
Lactate measured, n (%)		126 (47.4)	37 (25.9)^b,c	44 (64.7)^a,b	45 (81.8)^a,c	< 0.05^a; < 0.001^b; < 0.001^c
Lactate, median (Q1–Q3)		1.5 (1.1–2.4)	1.3 (1.0–1.9)^a	1.5 (1.1–2.5)	1.9 (1.3–3.0)^a	< 0.05^a
Admitted to hospital, n (%)		163 (61.3)	69 (48.3)^b,c	48 (70.6)^b	46 (83.6)^c	< 0.01^b; < 0.001^c
Admitted to ICU, n (%)		14 (5.3)	2 (1.4)^a	5 (7.4)	7 (12.7)^a	< 0.05^a
Hospital-free days (admitted subjects), median (Q1–Q3)		25 (22–26)	25 (22–26)^a	25 (23–26)	24 (21–25)^a	< 0.05^a
In-hospital mortality, n (%)		6 (2.3)	1 (0.7)	1 (1.5)	4 (7.3)	NS
Sequential Organ Failure Assessment, 3 d maximum, median (Q1–Q3)		2 (0–3)	1 (0–2)^a,c	2 (1–3)^a,b	3 (1–5)^b,c	< 0.05^a; < 0.01^b; < 0.001^c
Antibiotics Administered in the emergency department, n (%)	Antipseudomonal (aztreonam, cefepime, levofloxacin, meropenem, piperacillin/tazobactam)	54 (20.3)	13 (9.1)^a,b	19 (27.9)^a	22 (40.0)^b	< 0.01^a; < 0.001^b
	Anti-MRSA (daptomycin, vancomycin, linezolid)	40 (15.0)	7 (4.9)^a,b	18 (26.5)^a	15 (27.3)^b	< 0.001^a; < 0.001^b
	Other (excluding antipseudomonal and anti-MRSA)	59 (22.2)	31 (21.7)	14 (20.6)	14 (25.5)	NS
Acute Physiology and Chronic Health Evaluation-II, median (Q1–Q3)		5 (3–10)	5 (2–9)^a	6.5 (3–10)	7 (3.5–11.5)^a	< 0.05^a
WBC (10³ cells/μL), median (Q1–Q3)		12.1 (7.5–15.2)	9.0 (6.2–12.9 ^b,c	13.3 (9.0–17.2)^b	15.3 (11.2–18.8)^b,c	< 0.001^b; < 0.001^c
Platelets (10³ cells/μL), median (Q1–Q3)		249 (192–313)	251 (196–311)^a	275 (211–354)^b	215 (155–270)^a,b	< 0.05^a; < 0.01^b
Creatinine (mg/dL), median (Q1–Q3)		0.97 (0.8–1.5)	0.92 (0.8–1.3)	1.00 (0.8–1.4)	1.16 (0.8–1.9)	NS
Triage temperature (F), median (Q1–Q3)		98.3 (97.9–98.8)	98.1 (97.8–98.4)^a,b	98.5 (98.0–99.6)^a	98.9 (98.0–100.1)^b	< 0.001^a; < 0.001^b
IntelliSep index, median (Q1–Q3)		5.2 (4.0–6.5)	4.1 (3.1–4.7)^a,b	6.0 (5.7–6.4)^a,c	7.4 (7.2–8.1)^b,c	< 0.05^a; < 0.01^b; < 0.001^c

MRSA = methicillin-resistant Staphylococcus aureus, NS = not significant, Q1–Q3 = interquartile range.

p values were obtained from an unpaired two-sample Welch’s t test (except for hospital-free days, where the Mann-–Whitney U was due to the nonnormal distribution), with the null hypothesis that the mean of the two samples are equal.

Adjudication of Sepsis

Adjudicators determined that 112 of the subjects (42.1%) had an infection, 119 (44.7%) had or developed at least one organ failure, and 55 (20.7%) had both infection and organ dysfunction. Of these 55 patients, adjudicators determined 48 (18.0%) met the Sepsis-3 standard of dysregulated immune response to infection as the cause of the organ dysfunction. Adjudicators reached unanimous conclusion in 215 patients (80.8%) and consensus conclusion in 40 of the patients for a total of 255 patients (95.9%) with either unanimous or consensus determination. The remaining 11 required a forced determination.

Characteristics of Patients with Sepsis

Characteristics of patients by adjudicated outcome are presented in Table 2. When compared with patients adjudicated as not septic, those with sepsis were older. There was no difference in gender or race. Patients with sepsis were more likely to have hypertension and/or autoimmune disease; however, there was no difference in the presence of diabetes, obesity, malignancy, or chronic kidney disease between those with sepsis and those without. In septic patients, the most common sources of infection, by objective criteria, were respiratory (25% of cases) and urinary (27.1% of cases).

TABLE 2. - Characteristics of Patients by Adjudication Result

Characteristics		Sepsis-3 Definition		p
Characteristics		Sepsis, N = 48	No Sepsis, N = 218	p
Age, median (Q1–Q3)		67 (54–77)	55 (39–71)	< 0.001
Age (≥ 65), n (%)		27 (56.2)	76 (34.9)	0.01
Gender (female), n (%)		26 (54.2)	125 (57.3)	NS
Race, n (%)	White	29 (60.4)	115 (52.8)	N
	African American	19 (39.6)	96 (44.0)
	Other	0 (0.0)	7 (3.2)
Comorbidities, n (%)	Hypertension	36 (75.0)	119 (54.6)	0.01
	Diabetes	15 (31.3)	57 (26.1)	NS
	Obesity (body mass index ≥ 40)	8 (16.7)	31 (14.2)	NS
	Cancer	8 (16.7)	29 (13.3)	NS
	Chronic kidney disease	9 (18.8)	22 (10.1)	NS
	Autoimmune disease	7 (14.6)	6 (2.8)	0.05
Infected—meeting objective criteria by organ system, n (%)	Respiratory	12 (25.0)	18 (8.3)	0.05
	Gastrointestinal	8 (16.7)	8 (3.7)	0.05
	Urinary	13 (27.1)	21 (9.6)	0.05
	Skin	8 (16.7)	16 (7.3)	NS
	Other	7 (14.6)	5 (2.3)	0.05
	Not infected	7 (14.6)	154 (70.6)	0.001
Infected by adjudication, n (%)		48 (100.0)	64 (29.4)	0.001
Organ dysfunction by adjudication, n (%)		48 (100.0)	71 (32.6)	0.001
Infected with organ dysfunction by adjudication, n (%)		48 (100.0)	7 (3.2)	0.001
Adjudication, n (%)	Unanimous	31 (64.6)	184 (84.4)	0.01
	Consensus	12 (25.0)	28 (12.8)	NS
	Forced	5 (10.4)	6 (2.8)	NS
Blood culture, n (%)	Number tested	39 (81.3)	74 (33.9)	0.001
Blood culture, n (%)	Number positive (of tested)	16 (33.3)	8 (3.7)	0.001
Lactate measured, n (%)		41 (85.4)	85 (39.0)	0.001
Lactate, median (Q1–Q3)		1.90 (1.25–3.05)	1.50 (1.00–2.25)	NS
Admitted to hospital, n (%)		46 (95.8)	117 (53.7)	0.001
Admitted to ICU, n (%)		8 (16.7)	6 (2.8)	0.05
Hospital-free days (admitted subjects), median (Q1–Q3)		22 (20–25)	25 (23–26)	0.001
In-hospital mortality, n (%)		3 (6.3)	3 (1.4)	NS
SOFA, 3 d maximum, median (Q1–Q3)		3 (2–5)	1 (0–2)	0.001
Maximum SOFA change from enrollment for those admitted, median (Q1–Q3)		0 (–1.0 to 1.0)	0 (–2.0 to 1.0)	NS
Antibiotics Administered in the emergency department, n (%)	Antipseudomonal (aztreonam, cefepime, levofloxacin, meropenem, piperacillin/tazobactam)	25 (52.1)	29 (13.3)	0.001
	Anti-MRSA (daptomycin, vancomycin, linezolid)	17 (35.4)	23 (10.6)	0.01
	Other (excluding antipseudomonal and anti-MRSA)	15 (13.6)	44 (20.2)	NS
Acute Physiology and Chronic Health Evaluation-II, median (Q–Q3)		9 (6–12)	5 (2–9)	0.001
WBC (10³ cells/μL), median (Q1–Q3)		15.3 (8.5–17.6)	11.2 (7.4–14.1)	0.01
Platelets (10³ cells/μL), median (Q1–Q3)		213 (150–280)	255 (196–319)	0.05
Creatinine (mg/dL), median (Q1–Q3)		1.42 (0.86–2.10)	0.92 (0.75–1.26)	NS
Triage temperature (F), median (Q1–Q3)		98.6 (97.8–100.4)	98.3 (97.9–98.7)	0.05
IntelliSep index, median (Q1–Q3)		6.9 (6.1–7.6)	4.7 (3.7–5.9)	0.001

MRSA = methicillin-resistant Staphylococcus aureus, NS = not significant, Q1–Q3 = interquartile range, SOFA = Sequential Organ Failure Assessment.

p values were obtained from an unpaired two-sample Welch’s t test (except for hospital-free days, where the Mann-Whitney U was due to the nonnormal distribution), with the null hypothesis that the mean of the two samples are equal.

ISI Values

Table 1 includes baseline data of patients by ISI interpretation band. The median ISI for all patients was 5.2. Using the prespecified cut offs (¹⁶), 143 (53.8%) were in the green band, 66 (24.8%) in the yellow, and 55 (20.7%) in the red. There were no significant differences in baseline demographics (age, sex, and race) or comorbidities across interpretation bands, except for hypertension, which was more common in the red band patients (40 [72.7%]) than in green band patients (72 [50.3%]; p < 0.01).

Sepsis-3 Diagnosis

Patients adjudicated to meet Sepsis-3 criteria had significantly higher ISI values (median = 6.9; 95% CI, 6.5–7.3) than those adjudicated as not septic (median = 4.7; 95% CI, 4.5–4.9; p < 0.001) (Table 2). We assessed the capacity of the ISI to differentiate patients with signs or suspicion of infection who met Sepsis-3 criteria from those who did not through ROC curve analysis. Test performance was impacted by confidence in the adjudicated Sepsis-3 label: when adjudication was unanimous, the area under the ROC curve (AUC) for the ISI was 0.89, whereas inclusion of consensus and forced cases resulted in an AUC for the ISI was 0.84 and 0.83, respectively (Fig. 1A). Of the entire 266 patients, 48 patients met Sepsis-3 criteria with six of these (12.5%) in the green band, 17 (35.4%) in yellow, and 25 (52.1%) in red (Table 1). These findings result in a 4.2% probability of Sepsis-3 in the green band, 25% probability of Sepsis-3 in the yellow band, and 45.5% probability of Sepsis-3 in the red band (Fig. 1B). Patients in the red band, in addition to having a higher probability of sepsis, had higher probabilities of both infection and organ dysfunction (Fig. 2A). Figure 2B delineates clinical outcomes of patients adjudicated as septic in each interpretation band.

Figure 1.:
Diagnostic performance of the IntelliSep Index (ISI) in the differentiating of patients with sepsis as defined by the Sepsis-3 standard. A, Receiving operating characteristic curve for the ISI for various levels of confidence in the adjudicated endpoint. B, Diagnostic performance, including sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for each level of adjudication (unanimous, consensus, and forced). C, Distribution of patients and prevalence of sepsis, as well as likelihood ratio for sepsis, within each band, stratified by level of adjudication. AUC = area under the curve, DOR = diagnostic odds ratio.

Figure 2.:
Clinical characteristics of patients within each interpretation band. A, Graphical representation of patients within each interpretation band for each component of sepsis: infection, organ dysfunction, and causation by dysregulated immunity (i.e., sepsis). B, Clinical metrics for severity of illness for those adjudicated as septic, including admission status, mortality, length of stay, and 3 d maximum (max) Sequential Organ Failure Assessment (SOFA) score.

Severity of Illness Prognostication

Independent of adjudicated endpoints, red band patients were more likely to have higher severity of illness scores, to be admitted to the hospital, and to have adverse outcomes (Table 1). As depicted in Figure S4A (Supplemental Digital Content 1, https://links.lww.com/CCX/A678), the 3-day maximum SOFA score increased with interpretation bands: 1 (Q1–Q3, 0–2) for green band patients, 2 (Q1–Q3, 1–3) for yellow band patients, and 3 (Q1–Q3, 1–5) for red band patients. Also, patients in the red band were more likely to have an increase in 3-day SOFA scores as compared to those in green (Fig. S4B, Supplemental Digital Content 1, https://links.lww.com/CCX/A678). Acute Physiology and Chronic Health Evaluation-II scores had similar increases across interpretation bands, with a median of 5 (Q1–Q3, 2–9) for greens band patients, 6.5 (Q1–Q3, 3–10) for yellow, and 7 (Q1–Q3, 3.5–11.5) for red (Table 1) (Fig. S4C, Supplemental Digital Content 1, https://links.lww.com/CCX/A678).

Hospitalization and Antibiotic Use

Sixty-nine green band patients (48.3%) required hospital admission, whereas 48 yellow band patients (70.6%) and 46 red band patients (83.6%) required admission (Table 1) (Fig. S4E, Supplemental Digital Content 1, https://links.lww.com/CCX/A678). Red band patients were more likely to require ICU admission (7; 12.7%), as opposed to patients in the green band (2; 1.4%; p < 0.05) (Table 1) (Fig. S4F, Supplemental Digital Content 1, https://links.lww.com/CCX/A678). Once admitted, green band patients were more likely to have a higher number of hospital-free days than those in the red band (Table 1) (Fig. S4G, Supplemental Digital Content 1, https://links.lww.com/CCX/A678). Red band patients were more likely to have blood cultures ordered and returned positive and more likely to receive antibiotics (Fig. S4, Supplemental Digital Content 1, https://links.lww.com/CCX/A678). Furthermore, the ISI effectively risk stratified patients adjudicated as having infection, independent of being adjudicated as having sepsis.

As noted above, patients in the red band were more likely to have infection, independent of being adjudicated as having sepsis. In total, 112 patients were adjudicated as having an infection, with 38 in the green band (26.6% of green band patients), 37 in the yellow band (54.4% of yellow band patients), and 37 in the red band (67.3% of red band patients). Among those adjudicated as infected, those in the red band had higher SOFA scores, were more likely to be admitted to the hospital and the ICU, have longer hospital length of stay, and were more likely to have positive blood cultures (Fig. S5, Supplemental Digital Content 1, https://links.lww.com/CCX/A678).

Comparison With Procalcitonin and Common Assessments of Organ Dysfunction

Of the 266 total patients enrolled, 186 (71.4%) consented for specimen storage for comparison of the ISI to alternative biomarkers including procalcitonin. In this subgroup, 33 (17.4%) were adjudicated as septic. Figure S6 (Supplemental Digital Content 1, https://links.lww.com/CCX/A678) includes a comparison of the ISI and procalcitonin performance as depicted by ROC curves. Because there are no widely accepted cut off values for procalcitonin in the diagnosis of sepsis in the ED, in this analysis, both the ISI and procalcitonin were considered as continuous variables. The procalcitonin AUC was 0.79 as opposed to the ISI AUC of 0.84, for the diagnosis of sepsis by Sepsis-3 criteria. Table S1 (Supplemental Digital Content 1, https://links.lww.com/CCX/A678) includes a comparison of the ISI with a selection of other commonly assessed indicators of sepsis (²².

DISCUSSION

In this study of patients presenting to the ED with signs or suspicion of infection, we found the ISI, a quantification of innate immune activation, to be a reliable diagnostic indicator for Sepsis-3 (⁸^,²³). Also, the ISI can assist in the risk stratification of these patients, as it correlates with severity of illness scores, the need for hospitalization, and ICU care. The ISI’s rapid turnaround time (< 10 min) and collection in a standard, “purple-top” tube make it a potentially valuable tool for assisting ED clinicians in the efficient and effective identification and risk stratification of sepsis.

Achieving best outcomes in critical sepsis patients requires rapid identification and intervention with both appropriate antibiotic therapy (³^,⁵^,²⁴) and effective cardiopulmonary support. Unfortunately, recognizing sepsis can be challenging in the ED (²⁵), where these undifferentiated patients often present with a paucity of data. The Surviving Sepsis Campaign (²⁶) and Centers for Medicare & Medicaid Services sepsis core measure were designed to help guide and improve sepsis care; however, this guidance may encourage clinicians to use broad-spectrum antibiotics and resource-intensive interventions in patients with low acuity, a population in whom these measures are unlikely beneficial and may be potentially harmful (^27–32). The identification of a low-risk population that clinicians can safely treat more judiciously is equally as important as the timely identification of high-risk patients who require the recommended, aggressive interventions (²⁹).

With negative predictive values of 97% for Sepsis-3 in green band patients, the ISI can identify, early in the course of an ED visit, a subset of patients with signs or suspicion of infection in whom sepsis is unlikely, and the risk of adverse outcomes due to infection is low. This information can facilitate antibiotic stewardship and reduce the consumption of valuable resources. For example, green band patients had a low probability of positive blood cultures; however, one in three of these patients who received antibiotics received an antipseudomonal and/or anti–methicillin-resistant Staphylococcus aureus agent (Fig. S4D, Supplemental Digital Content 1, https://links.lww.com/CCX/A678 and Fig. S4H, Supplemental Digital Content 1, https://links.lww.com/CCX/A678). Green band may signal clinicians to approach these patients more deliberately and conservatively, encouraging evaluation for alternative etiologies prior to the initiation of sepsis-specific therapy and supporting the use of narrower-spectrum antimicrobials. Furthermore, once admitted, the length of stay was lower for those in the green band (Fig. S4 E–G, Supplemental Digital Content 1, https://links.lww.com/CCX/A678), further indicating a lower severity of illness for this population, a finding which may support clinicians in the decision to discharge low-risk patients when the clinical course is uncertain.

In addition to identifying a low-risk subset of patients, the ISI also identifies a subset of patients with a markedly higher probability of sepsis, elevated severity of illness, and the associated risk of adverse outcomes. Over two thirds of patients in the red band were infected, as opposed to nearly one in four of green band patients. As opposed to those in the green band who often had localized infection, infected patients in the red band were more often systemically ill, as they had substantially higher severity of illness scores, received more antibiotics, and required higher rates of admission to the hospital and ICU (Fig. S5, Supplemental Digital Content 1, https://links.lww.com/CCX/A678). These findings indicate that red band patients, independent of adjudicated outcome, are more likely to have a severe infection that may benefit from aggressive, early interventions as defined by current guidelines (²⁶). Red band patients who are infected but not adjudicated as septic may indicate patients with systemic activation or dysregulation of the immune system who have not yet developed manifestations of organ failure, because many clinical and laboratory indicators lag behind the causative biologic processes (^33–35). Red band patients who are not infected may have alternate causes of dysregulated immunity, such as severe exacerbations of autoimmune or other inflammatory disease. In patients with infection, a leading indicator of organ dysfunction caused by dysregulated immunity, such as the ISI, may allow clinicians to act on the basis of the dysregulated immunity itself, prior to clinically evident organ damage, potentially reducing sepsis-related morbidity for these patients.

In this study, our preliminary work also shows that the ISI compares favorably with procalcitonin as well as other biomarkers and more complex scoring systems for the risk stratification of patients with infection. The AUC for the ISI is at least as good as that of procalcitonin; however, the small sample size makes definitive comparison difficult. Additionally, there are no accepted cut off values for procalcitonin in the diagnosis of sepsis in the ED, which makes a direct comparison of positive and negative percent agreement with the adjudicated standard difficult. Despite the small sample size of our study, our assessment of the performance of procalcitonin is similar to other studies evaluating its performance for similar application (³⁶^,³⁷), and the performance of the ISI is also similar to our previous findings (¹⁶).

Our study has significant limitations. Although performed in two hospitals with different ED and inpatient medical staffs, both are in the same city. Also, despite enrolling a significant number of White and African-American patients, other races were not well represented. Furthermore, the prevalence of the primary endpoint (18%) of Sepsis-3 was relatively low in the study population, as was the acuity; however, we feel this reflects the daily reality of the ED setting and the current state of sepsis management in an undifferentiated patient population.

Finally, no reference standard exists for the diagnosis of sepsis, so we relied rigorous and structured adjudication processes with the flexibility of determining sepsis by either of the currently accepted definitions. Despite these efforts, the diagnosis of sepsis remains subjective (³⁸). Similar to previous observations (¹⁴), confidence in the adjudicated endpoint impacts the perceived performance of the diagnostic test and may define upper (unanimous) and lower (forced) performance boundaries for the assay given our current understanding of the condition.

CONCLUSIONS

The ISI provides a sensitive indicator of host immune activation, a fundamental process in the development of sepsis. In this study of a rapid diagnostic for sepsis on a specimen collected within hours of ED presentation, the green, yellow, and red bands of the ISI correlate with retrospective adjudication of sepsis, Severity Of Illness scores, and clinical metrics. This tool may enable ED physicians to focus resources upon the most critical patients while treating less severe patients more conservatively. In addition, the test may help to define populations of patients with innate immune activation for future research and the administration of therapeutics. Further investigation will evaluate the role that quantification of host response and the ISI will play in clinical practice.

ACKNOWLEDGMENTS

We would like to express their gratitude to the Office of Research at Our Lady of the Lake Regional Medical Center and the personnel who worked tirelessly to screen and enroll patients: Jennifer Daigle, BSN; Brittany Rishell, BSN; Katie Vance, PhD; JoAnne Hollingsworth, BSN, Allie Piehet, BS, and Emily Perdue, RN. We would also like to recognize the contributions of the team at Cytovale for their work in the set-up and completion of this study: Anne Jensen, Audun Johnson, Bliss Lambert, Chris Dahlberg, Hon Seng, Kylie Ortego, Lionel Guillou, Maddy Hem, and Nicholas Martinez. Finally, we thank the adjudication committee members, Drs. Clark Files, Daniel Henning, Dave Janz, Michael Noto, Steven Opal, Todd Rice, and Robert Scoggins, for their insight and efforts in the execution of this study.

REFERENCES

1. Rhee C, Dantes R, Epstein L, et al.; CDC Prevention Epicenter Program. incidence and trends of sepsis in US hospitals using clinical vs claims data, 2009-2014. JAMA. 2017; 318:1241–1249

Cited Here |
View Full Text | PubMed | CrossRef

2. Paoli CJ, Reynolds MA, Sinha M, et al. Epidemiology and costs of sepsis in the United States-an analysis based on timing of diagnosis and severity level. Crit Care Med. 2018; 46:1889–1897

Cited Here |
View Full Text | PubMed | CrossRef

3. Kumar A, Roberts D, Wood KE, 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; 34:1589–1596

Cited Here |
View Full Text | PubMed | CrossRef

4. Jones SL, Ashton CM, Kiehne LB, et al. Outcomes and resource use of sepsis-associated stays by presence on admission, severity, and hospital type. Med Care. 2016; 54:303–310

Cited Here |
View Full Text | PubMed | CrossRef

5. Ferrer R, Martin-Loeches I, Phillips G, 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; 42:1749–1755

Cited Here |
View Full Text | PubMed | CrossRef

6. Page DB, Donnelly JP, Wang HE. Community-, Healthcare-, and hospital-acquired severe sepsis hospitalizations in the University HealthSystem Consortium. Crit Care Med. 2015; 43:1945–1951

Cited Here |
View Full Text | PubMed | CrossRef

7. Bone RC, Balk RA, Cerra FB et al: American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference: Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Crit Care Med. 1992; 20: 864–874

Cited Here |
PubMed

8. Singer M, Deutschman CS, Seymour CW, et al. The third international consensus definitions for sepsis and septic Shock (Sepsis-3). JAMA. 2016; 315:801–810

Cited Here |
View Full Text | PubMed | CrossRef

9. Gossett DR, Tse HT, Lee SA, et al. Hydrodynamic stretching of single cells for large population mechanical phenotyping. Proc Natl Acad Sci U S A. 2012; 109:7630–7635

Cited Here |
PubMed | CrossRef

10. Zeming KK, Vernekar R, Chua MT, et al. Label-free biophysical markers from whole blood microfluidic immune profiling reveal severe immune response signatures. Small. 2021; 17:e2006123

Cited Here |
PubMed

11. Eubank TA, Long SW, Perez KK. Role of rapid diagnostics in diagnosis and management of patients with sepsis. J Infect Dis. 2020; 222:S103–S109

Cited Here |
PubMed | CrossRef

12. Crouser ED, Parrillo JE, Seymour C, et al. Improved early detection of sepsis in the ED with a novel monocyte distribution width biomarker. Chest. 2017; 152:518–526

Cited Here |
View Full Text | PubMed | CrossRef

13. Tawfik DM, Vachot L, Bocquet A, et al. Immune profiling panel: A proof-of-concept study of a new multiplex molecular tool to assess the immune status of critically ill patients. J Infect Dis. 2020; 222:S84–S95

Cited Here |
PubMed | CrossRef

14. Miller RR 3rd, Lopansri BK, Burke JP, et al. Validation of a host response assay, SeptiCyte LAB, for discriminating sepsis from systemic inflammatory response syndrome in the ICU. Am J Respir Crit Care Med. 2018; 198:903–913

Cited Here |
View Full Text | PubMed | CrossRef

15. Crawford K, DeWitt A, Brierre S, et al. Rapid biophysical analysis of host immune cell variations associated with sepsis. Am J Respir Crit Care Med. 2018; 198:280–282

Cited Here |
View Full Text | PubMed | CrossRef

16. Guillou L, Sheybani R, Jensen AE, et al. Development and validation of a cellular host response test as an early diagnostic for sepsis. PLoS One. 2021; 16:e0246980

Cited Here |
PubMed

17. Centers for Disease Control and Prevention, National Healthcare Safety Network. CDC/NHSN Surveillance Definitions for Specific Types of Infections. Atlanta, GA, Centers for Disease Control and Prevention, 2020

Cited Here

18. Jones AE, Trzeciak S, Kline JA. The sequential organ failure assessment score for predicting outcome in patients with severe sepsis and evidence of hypoperfusion at the time of emergency department presentation. Crit Care Med. 2009; 37:1649–1654

Cited Here |
View Full Text | PubMed | CrossRef

19. Rhee C, Klompas M. Sepsis trends: Increasing incidence and decreasing mortality, or changing denominator? J Thorac Dis. 2020; 12:S89–S100

Cited Here |
PubMed

20. Lopansri BK, Miller Iii RR, Burke JP, et al. Physician agreement on the diagnosis of sepsis in the intensive care unit: Estimation of concordance and analysis of underlying factors in a multicenter cohort. J Intensive Care. 2019; 7:13

Cited Here |
PubMed

21. Harris PA, Taylor R, Thielke R, et al. Research electronic data capture (REDCap)–a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009; 42:377–381

Cited Here |
PubMed | CrossRef

22. Jekarl DW, Lee S, Kim M, et al. Procalcitonin as a prognostic marker for sepsis based on SEPSIS-3. J Clin Lab Anal. 2019; 33:e22996

Cited Here |
PubMed | CrossRef

23. Levy MM, Fink MP, Marshall JC, et al.; SCCM/ESICM/ACCP/ATS/SIS. 2001 SCCM/ESICM/ACCP/ATS/SIS international sepsis definitions conference. Crit Care Med. 2003; 31:1250–1256

Cited Here |
View Full Text | PubMed | CrossRef

24. Kalil AC, Johnson DW, Lisco SJ, et al. Early goal-directed therapy for sepsis: A novel solution for discordant survival outcomes in clinical trials. Crit Care Med. 2017; 45:607–614

Cited Here |
View Full Text | PubMed | CrossRef

25. Morr M, Lukasz A, Rübig E, et al. Sepsis recognition in the emergency department - impact on quality of care and outcome? BMC Emerg Med. 2017; 17:11

Cited Here |
PubMed

26. Rhodes A, Evans LE, Alhazzani W, et al. Surviving sepsis campaign: International guidelines for management of sepsis and septic shock: 2016. Crit Care Med. 2017; 45:486–552

Cited Here |
View Full Text | PubMed | CrossRef

27. Force IST. Infectious Diseases Society of America (IDSA) position statement: Why IDSA did not endorse the surviving sepsis campaign guidelines. Clin Infect Dis. 2018; 66:1631–1635

Cited Here |
View Full Text | PubMed | CrossRef

28. Septimus EJ. Sepsis perspective 2020. J Infect Dis. 2020; 222:S71–S73

Cited Here |
PubMed | CrossRef

29. Rhee C, Chiotos K, Cosgrove SE, et al. Infectious Diseases Society of America position paper: Recommended revisions to the national severe sepsis and septic shock early management bundle (SEP-1) sepsis quality measure. Clin Infect Dis. 2021; 72:541–552

Cited Here |
PubMed | CrossRef

30. Pepper DJ, Jaswal D, Sun J, et al. Evidence underpinning the centers for medicare & medicaid services’ severe sepsis and septic shock management bundle (SEP-1): A systematic review. Ann Intern Med. 2018; 168:558–568

Cited Here |
View Full Text | PubMed | CrossRef

31. Pepper DJ, Natanson C, Eichacker PQ. Evidence underpinning the centers for medicare & medicaid services’ severe sepsis and septic shock management bundle (SEP-1). Ann Intern Med. 2018; 168:610–612

Cited Here |
View Full Text | PubMed | CrossRef

32. Swenson KE, Winslow DL. Impact of sepsis mandates on sepsis care: Unintended consequences. J Infect Dis. 2020; 222:S166–S173

Cited Here |
PubMed | CrossRef

33. Aydemir H, Piskin N, Akduman D, et al. Platelet and mean platelet volume kinetics in adult patients with sepsis. Platelets. 2015; 26:331–335

Cited Here |
PubMed | CrossRef

34. Nesseler N, Launey Y, Aninat C, et al. Clinical review: The liver in sepsis. Crit Care. 2012; 16:235

Cited Here |
PubMed | CrossRef

35. Waikar SS, Bonventre JV. Creatinine kinetics and the definition of acute kidney injury. J Am Soc Nephrol. 2009; 20:672–679

Cited Here |
PubMed | CrossRef

36. Kim SJ, Hwang SO, Kim YW, et al. Procalcitonin as a diagnostic marker for sepsis/septic shock in the emergency department; a study based on Sepsis-3 definition. Am J Emerg Med. 2019; 37:272–276

Cited Here |
View Full Text | PubMed | CrossRef

37. Ljungström L, Pernestig AK, Jacobsson G, et al. Diagnostic accuracy of procalcitonin, neutrophil-lymphocyte count ratio, C-reactive protein, and lactate in patients with suspected bacterial sepsis. PLoS One. 2017; 12:e0181704

Cited Here |
PubMed | CrossRef

38. Rhee C, Kadri SS, Danner RL, et al. Diagnosing sepsis is subjective and highly variable: A survey of intensivists using case vignettes. Crit Care. 2016; 20:89

Cited Here |
PubMed

Keywords:

cellular viscoelastic properties; emergency department; immune dysregulation; leukocyte biomechanical properties; microfluidics; sepsis diagnosis

Supplemental Digital Content

CCX_1_1_2021_05_26_TSE_21-00165_SDC1.pdf; [PDF] (706 KB)

Article Keywords

Keyword Highlighting
Highlight selected keywords in the article text.

	cellular viscoelastic properties
	emergency department
	immune dysregulation
	leukocyte biomechanical properties
	microfluidics
	sepsis diagnosis

Secondary Logo

Journal Logo

Assessment of a Cellular Host Response Test as a Sepsis Diagnostic for Those With Suspected Infection in the Emergency Department

Abstract

Objectives:

Design, Setting, SUBJECTS:

MEASUREMENTS AND MAIN RESULTS:

CONCLUSIONS: