Host Immune Markers Distinguish Clostridioides difficile Infection From Asymptomatic Carriage and Non–C. difficile Diarrhea

Abstract

Background

Recent data indicate that Clostridioides difficile toxin concentrations in stool do not differentiate between C. difficile infection (CDI) and asymptomatic carriage. Thus, we lack a method to distinguish a symptomatic patient with CDI from a colonized patient with diarrhea from another cause. To address this, we evaluated markers of innate and adaptive immunity in adult inpatients with CDI (diagnosed per US guidelines), asymptomatic carriage, or non-CDI diarrhea.

Methods

CDI-NAAT patients had clinically significant diarrhea and positive nucleic acid amplification testing (NAAT) and received CDI treatment. Carrier-NAAT patients had positive stool NAAT but no diarrhea. NAAT-negative patients (with and without diarrhea) were also enrolled. A panel of cytokines and anti–toxin A and B immunoglobulin (Ig) were measured in serum; calprotectin and anti–toxin B Ig A/G were measured in stool. NAAT-positive stool samples were tested by an ultrasensitive toxin assay (clinical cutoff, 20 pg/mL).

Results

Median values for interleukin (IL)-4, IL-6, IL-8, IL-10, IL-15, granulocyte colony-stimulating factor (GCSF), MCP-1, tumor necrosis factor α (TNF-α), and IgG anti–toxin A in blood and IgA/G anti–toxin B in stool were significantly higher in CDI patients compared with all other groups (P < .05). Concentration distributions for IL-6, GCSF, TNF-α, and IgG anti–toxin A in blood, as well as IgA and IgG anti–toxin B in stool, separated CDI patients from all other groups.

Conclusions

Specific markers of innate and adaptive immunity distinguish CDI from all other groups, suggesting potential clinical utility for identifying which NAAT- and toxin-positive patients with diarrhea truly have CDI.

We compared markers of innate and acquired immunity in adults with symptomatic Clostridioides difficile infection (CDI), asymptomatic carriage, or non-CDI diarrhea. We found markers that distinguish CDI from all other groups, suggesting utility for identifying patients with true clinical CDI.

(See the Editorial Commentary by Pardi and Khanna on pages 1094–5.)

Optimal strategies for the diagnosis of Clostridioides (formerly Clostridium) difficile infection (CDI) remain unclear [1]. Exposure to C. difficile can lead to asymptomatic carriage (presence of toxinogenic C. difficile in the colon, but no symptoms of CDI) or to clinical CDI with a variety of clinical presentations (ranging from mild diarrhea to severe colitis and/or death) [2, 3]. Current approaches to the diagnosis of CDI include tests capable of detecting toxinogenic C. difficile bacteria (nucleic acid amplification testing [NAAT] or culture) or the toxins themselves (enzyme immunoassay or tissue culture cytotoxicity assay) [2], as well as algorithmic test combinations [1, 4]. Given the high sensitivity of NAAT, over the past decade many laboratories shifted to NAAT for CDI diagnosis only to encounter concern that this method was insufficiently specific for the diagnosis of clinical CDI (vs colonization) [3, 5]. Most recently, the field has reverted towards utilization of toxin detection for making treatment decisions [4] based on data suggesting that patients with toxin-positive stool (compared with patients with NAAT-positive but toxin-negative stool) are at higher risk of poor outcomes (eg, [5, 6]). However, we [7] recently compared stool toxin concentrations in patients with CDI versus asymptomatic carriage, using an ultrasensitive and quantitative toxin immunoassay [7–9], and found that toxin concentration alone could not distinguish a patient with CDI (diagnosed by either NAAT or toxin detection plus new-onset diarrhea) from an asymptomatic carrier because concentration distributions in both cohorts overlapped substantially. Those findings indicated that neither stool toxin concentration nor NAAT cycle threshold (Ct) value can reliably distinguish a symptomatic patient with true CDI from a symptomatic, colonized patient whose diarrhea has another cause.

Infection with C. difficile leads to both adaptive and innate immune responses, with an impact on clinical outcomes including disease expression, disease severity, and recurrence risk [10]. Adaptive, humoral immune responses to C. difficile toxins have been associated with symptomless carriage and protection against recurrence [11–13]. Clostridioides difficile and its toxins are potent activators of innate immune responses in vitro and in vivo, including activation of nuclear transcription factor κB (NF-κB) and mitogen-activated protein (MAP) kinase pathways [14–17]. This has been associated with increases in many markers of the host inflammatory response, including cytokines, chemokines, and other mediators of neutrophil activation and intestinal mucosal injury [18–22]. These factors have the potential to act as clinically useful markers of active CDI and colitis.

In search of novel indicators to distinguish true CDI from the combination of C. difficile colonization and diarrhea, we measured a panel of markers of innate and adaptive immunity in serum and stool from our CDI and asymptomatic carrier cohorts, in parallel with samples from cohorts of subjects with NAAT-negative stools (with and without diarrhea). The purpose of this study was to determine whether measurement of these analytes might separate our CDI cohort from all other cohorts, thus enabling future studies to confirm the analytes’ diagnostic utility and prognostic value.

METHODS

Patient Cohorts

Eligibility and enrollment for CDI-NAAT and Carrier-NAAT cohorts have been described in detail previously [7]. Briefly, CDI-NAAT patients were inpatients aged 18 years or older with new-onset diarrhea, positive clinical stool NAAT result, and decision to treat. The diagnostic clinical stool sample was captured as a discarded sample; a discarded serum sample collected within 1 day of that stool sample was also captured. Patients were excluded if the diagnostic stool specimen was more than 72 hours old, if they had received CDI treatment for more than 24 hours prior to stool collection, or if they had a colostomy. Carrier-NAAT patients were inpatients aged 18 years or older, admitted for at least 72 hours, who had received at least 1 dose of an antibiotic within the past 7 days, and did not have diarrhea in the 48 hours prior to stool sample submission. Patients with 2 or more loose stools within a 24-hour period were excluded; patients with 1 loose stool were included only if providers had recently administered a laxative. Patients were excluded if they had a colostomy; received oral or intravenous metronidazole, oral vancomycin, oral rifaximin, and/or oral fidaxomicin for more than 24 hours within the prior 7 days; had been diagnosed with CDI in the past 6 months; or had tested negative for C. difficile within the past 7 days. Stool samples were collected prospectively under verbal informed consent. A discarded serum sample from within 1 day of the stool sample was also captured. NAAT (Xpert C. difficile/Epi, Cepheid, Sunnyvale, CA) was performed on all samples, and positive samples were retained as the Carrier-NAAT cohort.

Patients in the “Diarrhea, NAAT-neg” cohort had diarrhea (confirmed using the same definition used for the CDI-NAAT cohort) but had tested NAAT negative on clinical C. difficile testing; stool samples were captured prior to discarding. Patients in the “No diarrhea, NAAT-neg” cohort had screened as eligible for the Carrier-NAAT group, but the submitted stool sample tested negative on research NAAT. Discarded serum samples for both cohorts were captured within 1 day of the stool sample.

For all cohorts, peak white blood cell count/creatinine and nadir albumin values within 5 days before to 2 days after stool collection were recorded.

Serum Cytokine Measurement

Human serum cytokines (granulocyte colony-stimulating factor [GCSF], interleukin [IL]-1β, IL-2, IL-4, IL-6, IL-8, IL-10, IL-13, IL-15, monocyte chemoattractant protein-1 (MCP-1), vascular endothelial growth factor A (VEGF-A), and tumor necrosis factor α [TNF-α]) were measured (pg/mL serum) using a Milliplex magnetic bead kit and Luminex analyzer (MAGPIX; Millipore Sigma, Inc), following the manufacturer’s instruction manual.

Antitoxin Antibody Enzyme-linked Immunosorbent Assay

Serum concentrations of immunoglobulin (Ig) A, IgG, and IgM antibodies to C. difficile toxins A and B were measured by semiquantitative enzyme-linked immunosorbent assay (ELISA) as previously described [11, 12, 23, 24]. Toxin A and toxin B were separately purified from the culture supernatant of strain VPI 10463 (American Type Culture Collection; 43255-FZ) as previously described [16, 25, 26]. After dilution (1:200 for IgG, 1:50 for IgA, 1:20 for IgM) using phosphate-buffered saline–Tween 20, sera were added to microtiter plates that had been coated with 50 µg/mL toxin A or toxin B in the presence of blocking buffer (1% bovine serum albumin) and incubated overnight. After washing, wells were incubated with horseradish peroxidase–conjugated anti–human IgA, IgG, or IgM; washed; and exposed to substrate. The absorbance was determined at 450 nm in a standard plate reader. Sera from a separate group of patients with very high antitoxin antibody measurements for each assay were pooled to form standards that were assigned an arbitrary antitoxin antibody content of 100 ELISA units. ELISA-unit readings for each sample/assay were expressed in relation to a dilution curve of the standard assayed on the same plate. For several samples, individual assays yielded absorbances above the readable range of the standard curve; these were assigned values by ranking of their absorbance readings, consistent with our planned use of nonparametric statistical analyses.

Calprotectin ELISA

Calprotectin fCAL ELISA (Bühlmann Laboratories) was performed according to the manufacturer’s instructions using the extended-range ELISA procedure (30–1800 µg/mL). Stool samples were processed according to the manufacturer’s instructions to achieve a final dilution of 1:7500.

VIDAS IgA and IgG Anti–toxin B Assays

Stool samples were diluted 1:4 (proprietary diluent: bioMerieux) and centrifuged at 3000 × g for 10 minutes. IgA and IgG anti–toxin B assays were performed on the supernatant using a VIDAS (bioMerieux) platform, utilizing a 2-step enzyme immunoassay sandwich method. Clostridioides difficile toxin B (Native Antigen Company) coated on a Solid Phase Receptacle was used to capture anti–toxin B Ig present in the sample. Detection of IgA or IgG anti–toxin B antibodies utilized anti–human IgA or IgG conjugated with recombinant alkaline phosphatase (bioMerieux); substrate (4-methyl-umbelliferyl phosphate) hydrolysis generated a fluorescent product measured at 450 nm. Results were reported in relative fluorescent values without adjustment.

Single Molecule Array

Stool toxin concentrations were measured by Single Molecule Array for the CDI-NAAT and Carrier-NAAT cohorts as previously described [7].

Statistical Analysis

Categorical measures were expressed as counts with percentages, and continuous measures were expressed as medians (25th percentiles, 75th percentiles). Log-normal transformations were used to reduce variation in data sets. Comparisons between groups were tested using Fisher’s exact tests for categorical measures, Wilcoxon rank-sum tests for comparison of 2 category continuous variables, and Kruskal-Wallis procedure for comparison of more than 2 categories. All tests were 2-sided, and a P value <.05 was considered statistically significant. Receiver operating characteristic analysis allowed assessment of the predictive accuracy of GCSF concentrations for CDI classification. All statistical analysis was performed using SAS (version 9.4; SAS Institute) software.

RESULTS

In our previous analysis [7], 122 stool samples/subjects met all inclusion/exclusion criteria for the CDI-NAAT cohort and 44 stool samples/subjects formed the Carrier-NAAT cohort. Aliquots of 121 of 122 CDI-NAAT stool samples and 44 of 44 Carrier-NAAT stool samples were available for measurement of calprotectin and/or IgA and IgG anti–toxin B levels. Aliquots from stool samples from the first 44 subjects meeting criteria for the Diarrhea, NAAT-neg cohort and from 50 subjects meeting criteria for the No diarrhea, NAAT-neg cohort (matched by approximate date of study sample collection to the 44 Carrier-NAAT samples) were also selected for analysis. All available serum samples (collected within 1 day of the corresponding stool sample) from members of these 4 cohorts (n = 102 CDI-NAAT; n = 35 Carrier-NAAT; n = 44 Diarrhea, NAAT-neg; and n = 45 No diarrhea, NAAT-neg serum samples) were analyzed for blood markers (see Methods). Given that toxin detection is considered by many to be a more specific CDI diagnostic tool than NAAT, stool toxin concentrations previously measured by Single Molecule Array were used to classify the subsets of the CDI-NAAT and Carrier-NAAT cohorts who had stool toxin A+B concentrations of 20 pg/mL or greater (CDI-Tox20 and Carrier-Tox20 cohorts, n = 79 and 34, respectively) [7]. All 6 cohorts were demographically similar and did not show any significant differences in baseline white blood cell, creatinine, and albumin values (Table 1, Supplementary Table 1) other than significantly lower median albumin levels observed in the CDI-NAAT and CDI-Tox20 cohorts (compared with all other cohorts) (Supplementary Table 1).

Table 1.

Comparison of Demographics and Baseline Laboratory Values for Study Cohorts

Variable	CDI-NAAT (n = 122)	Carrier-NAAT (n = 44)	P Value	Diarrhea, NAAT-neg (n = 44)	No Diarrhea, NAAT-neg (n = 50)	CDI-Tox20 (n = 79)	Carrier-Tox20 (n = 34)
Age, median (IQR), y [N]	67.0 (55.0, 76.0) [122]	64.5 (49.0, 73.5) [44]	.365	61.5 (53.0, 71.5) [44]	63.0 (55.0, 70.0) [50]	68.0 (57.0, 77.0) [79]	65.5 (52.0, 77.0) [34]
Sex			.860
Female, n (%) [N]	67 (54.9) [122]	23 (52.3) [44]		22 (50.0) [44]	17 (34.0) [50]	44 (55.7) [79]	17 (50.0) [34]
Male, n (%)	55 (45.1)	21 (47.7)		22 (50.0)	33 (66.0)	35 (44.3)	17 (50.0)
WBC, median (IQR), × 109/L [N]	10.6 (6.9, 17.4) [120]	11.2 (6.9, 14.8) [44]	.725	10.4 (8.1, 17.2) [44]	12.5 (7.4, 14.7) [50]	10.9 (8.1, 17.4) [78]	10.6 (6.6, 14.9) [34]
WBC ≥15, n (%) [N]	41 (34.2) [120]	10 (22.7) [44]	.186	15 (34.1) [44]	12 (24.0) [50]	28 (35.9) [78]	8 (23.5) [34]
Creatinine, median (IQR), mg/dL [N]	1.0 (0.8, 1.7) [120]	1.2 (0.8, 1.9) [44]	.481	1.4 (0.9, 2.2) [44]	1.2 (0.8, 1.7) [50]	1.0 (0.8, 1.7) [78]	1.1 (0.8, 1.8) [34]
Creatinine ≥1.5, n (%) [N]	38 (31.7) [120]	14 (31.8) [44]	1.0	19 (43.2) [44]	19 (38.0) [50]	22 (28.2) [78]	9 (26.5) [34]
Albumin, median (IQR), mg/dL [N]	2.9 (2.5, 3.6) [108]	3.4 (3.1, 3.8) [21]	.078	3.4 (3.0, 3.8) [33]	3.3 (3.0, 3.6) [24]	2.9 (2.6, 3.5) [71]	3.4 (3.2, 3.8) [16]
027-NAP1-BI, n (%) [N]	15 (12.3) [122]	2 (4.6) [44]	.244	N/A	N/A	15 (19.0) [79]	2 (5.9) [34]

P values are for the comparison between CDI-NAAT and Carrier-NAAT cohort medians. CDI-NAAT and Carrier-NAAT cohorts were defined by positive results on stool NAAT, while CDI-Tox20 and Carrier-Tox20 cohorts included only the subjects who were positive by both NAAT and Single Molecule Array (toxin A+B ≥20 pg/mL). Abbreviations: CDI, Clostridioides difficile infection; IQR, interquartile range; NAAT, nucleic acid amplification testing; neg, negative; Tox20, CDI-NAAT or Carrier-NAAT cohorts who had stool toxin A+B concentrations of ≥20 pg/mL; WBC, white blood cell count.

Measurements for a panel of blood and stool analytes in samples from each of the 6 cohorts are presented in Table 2. Median values for 11 analytes (GCSF, IL-10, IL-15, IL-4, IL-6, IL-8, MCP-1, TNF-α, and IgG anti–toxin A in blood and IgA and IgG anti–toxin B in stool) were significantly higher (P < .05) in the CDI-NAAT versus the Carrier-NAAT cohorts (Table 2). Boxplots demonstrating the distributions of values in each of the 6 cohorts indicated that GCSF distributions (Figure 1) clearly distinguished CDI-NAAT/CDI-Tox20 cohorts from all other cohorts (Carrier-NAAT; Carrier-Tox20; Diarrhea, NAAT-neg; and No diarrhea, NAAT-neg), with highly significant P values for all comparisons of medians between CDI-NAAT/CDI-Tox20 cohorts and other cohorts (Figure 1, inset table; P < .0001). Conversely, there were no significant differences in median GCSF values when the non-CDI cohorts were compared with each other (Figure 1, inset table). A similar, although less dramatic, distinction between CDI-NAAT/CDI-Tox20 and the other 4 cohorts was also observed for distributions of IL-6 (Figure 2), TNF-α (Figure 3), and IgG anti–toxin A in blood (Figure 4) and for IgA (Figure 5) and IgG (Figure 6) anti–toxin B in stool. Receiver operating curve analysis for GCSF concentrations in CDI-NAAT (case) versus Carrier-NAAT (control) cohorts demonstrated an area under the curve of 0.842 (P < .001) (Figure 7). Analysis using CDI-NAAT (case) compared with the combination of Carrier-NAAT; Diarrhea, NAAT-neg; and No diarrhea, NAAT neg cohorts as the control group demonstrated a similar area under the curve (0.844). Boxplots demonstrating the distributions of values for the remaining 5 analytes with significantly different medians (CDI-NAAT vs Carrier-NAAT) (IL-4, IL-8, IL-10, IL-15, and MCP-1) are shown in Supplementary Figures 1–5. Plots of distributions for the 10 analytes without significantly different medians are shown in Supplementary Figure 6.

Table 2.

Comparison of Serum and Stool Analyte Measurements in Study Cohorts

Variable	CDI-NAAT (n = 122)	Carrier-NAAT (n = 44)	P Value	Diarrhea, NAAT-neg (n = 44)	No Diarrhea, NAAT-neg (n = 50)	CDI-Tox20 (n = 79)	Carrier-Tox20 (n = 34)
Blood
G CSF, pg/mL	47.2 (22.6, 160.7) [102]	9.8 (2.2, 21.2) [35]	<.001	9.8 (2.2, 32.4) [44]	5.3 (0.5, 14.2) [45]	47.2 (22.6, 157.7) [66]	9.8 (2.2, 29.4) [27]
IL-10, pg/mL	10.8 (1.6, 29.5) [102]	1.8 (0.0, 10.8) [35]	.002	3.4 (0.0, 14.1) [44]	0.0 (0.0, 3.1) [45]	9.5 (1.2, 28.1) [66]	5.4 (0.0, 16.3) [27]
IL-13, pg/mL	0.0 (0.0, 1.8) [102]	0.0 (0.0, 0.0) [35]	.447	0.0 (0.0, 0.3) [44]	0.0 (0.0, 0.0) [45]	0.0 (0.0, 4.7) [66]	0.0 (0.0, 0.0) [27]
IL-15, pg/mL	1.7 (0.1, 5.3) [102]	0.3 (0.0, 1.6) [35]	.007	0.3 (0.0, 3.3) [44]	0.0 (0.0, 0.2) [45]	1.6 (0.2, 5.7) [66]	0.3 (0.0, 1.4) [27]
IL-1β, pg/mL	0.0 (0.0, 0.0) [102]	0.0 (0.0, 0.0) [35]	.873	0.0 (0.0, 0.0) [44]	0.0 (0.0, 0.0) [45]	0.0 (0.0, 0.0) [66]	0.0 (0.0, 0.0) [27]
IL-2, pg/mL	0.0 (0.0, 0.0) [102]	0.0 (0.0, 0.0) [35]	.666	0.0 (0.0, 0.0) [44]	0.0 (0.0, 0.0) [45]	0.0 (0.0, 0.0) [66]	0.0 (0.0, 0.0) [27]
IL-4, pg/mL	2.6 (0.0, 9.1) [102]	0.0 (0.0, 0.0) [35]	.002	9.1 (2.6, 12.4) [44]	0.0 (0.0, 0.0) [45]	1.0 (0.0, 5.8) [66]	0.0 (0.0, 1.0) [27]
IL-6, pg/mL	12.5 (2.6, 38.1) [102]	0.5 (0.0, 5.8) [35]	<.001	2.7 (0.0, 10.7) [44]	0.6 (0.0, 3.8) [45]	12.7 (2.6, 35.3) [66]	0.4 (0.0, 5.1) [27]
IL-8, pg/mL	60.9 (27.2, 127.9) [102]	21.0 (9.0, 49.3) [35]	<.001	30.9 (14.7, 76.0) [44]	24.9 (12.4, 59.2) [45]	58.2 (27.2, 127.9) [66]	20.7 (9.0, 40.2) [27]
MCP-1, pg/mL	688.2 (523.4, 990.1) [102]	579.5 (426.9, 802.1) [35]	.045	579.5 (387.0, 791.3) [44]	478.3 (397.2, 545.0) [45]	674.7 (518.2, 862.9) [66]	585.5 (434.8, 884.0) [27]
TNF-ɑ, pg/mL	18.3 (13.9, 29.7) [102]	9.7 (4.5, 15.3) [35]	<.001	13.0 (8.7, 22.7) [44]	7.4 (6.2, 10.9) [45]	18.1 (12.8, 28.5) [66]	11.6 (5.6, 14.4) [27]
VEGFA, pg/mL	63.0 (18.3, 134.6) [102]	62.2 (27.5, 194.6) [35]	.448	69.3 (34.3, 127.7) [44]	77.7 (35.8, 118.3) [45]	75.4 (25.4, 150.8) [66]	73.2 (27.5, 194.6) [27]
IgA toxin A, ELISA units	20.6 (12.2, 104.0) [102]	24.6 (9.8, 103.0) [35]	.485	12.9 (8.2, 17.5) [44]	12.7 (7.2, 62.5) [45]	19.3 (9.7, 104.0) [66]	25.3 (9.7, 103.0) [27]
IgG toxin A, ELISA units	27.2 (14.8, 57.9) [102]	18.6 (11.0, 24.9) [35]	.008	16.7 (11.1, 27.8) [44]	13.4 (9.7, 22.4) [45]	26.3 (14.1, 51.0) [66]	18.6 (11.0, 24.4) [27]
IgM toxin A, ELISA units	1.3 (0.0, 2.9) [102]	0.4 (0.0, 3.2) [35]	.276	0.8 (0.0, 3.4) [44]	0.4 (0.0, 2.9) [45]	1.3 (0.0, 3.0) [66]	0.4 (0.0, 2.8) [27]
IgA toxin B, ELISA units	12.9 (4.8, 68.7) [102]	13.8 (5.0, 20.2) [35]	.359	8.8 (4.1, 19.2) [44]	5.0 (2.7, 8.6) [45]	9.0 (4.3, 27.6) [66]	9.7 (3.6, 20.1) [27]
IgG toxin B, ELISA units	8.1 (5.1, 18.2) [102]	10.6 (3.9, 16.3) [35]	.920	8.1 (5.9, 13.5) [44]	6.2 (4.3, 9.8) [45]	6.8 (4.3, 12.3) [66]	10.8 (4.1, 15.7) [27]
IgM toxin B, ELISA units	3.8 (1.9, 10.9) [102]	5.7 (2.5, 9.2) [35]	.329	5.1 (2.7, 12.8) [44]	3.5 (2.3, 9.2) [45]	3.6 (1.7, 9.8) [66]	5.1 (2.5, 7.8) [27]
Stool
Calprotectin, ELISA units	290.8 (64.6, 888.3) [120]	174.9 (75.3, 409.2) [43]	.088	74.9 (16.2, 174.8) [44]	171.0 (75.5, 400.5) [28]	395.16 (163.0, 1059.9) [77]	150.0 (60.9, 415.2) [34]
Vidas IgA, RFVs	110.0 (29.0, 413.0) [115]	11.0 (6.0, 33.0) [43]	<.001	55.5 (18.0, 133.0) [44]	11.5 (4.0, 64.0) [50]	133.00 (28.0, 539.00) [73]	11.0 (7.0, 51.0) [33]
Vidas IgG, RFVs	15.0, (0.0, 295.0) [115]	0.0 (0.0, 1.0) [43]	<.001	0.0 (0.0, 9.0) [44]	0.0 (0.0, 0.0) [50]	73.00 (0.00, 555.0) [73]	0.0 (0.0, 1.0) [33]

All data shown are medians (IQR); n in brackets. P values are for the comparison between CDI-NAAT and Carrier-NAAT cohort medians (for analytes with significant P values [<.05], additional median comparisons are presented in Figures 1–6 and Figures S1–S5]. CDI-NAAT and Carrier-NAAT cohorts were defined by positive results on stool NAAT, while CDI-Tox20 and Carrier-Tox20 cohorts included only the subjects who were positive by both NAAT and Single Molecule Array (toxin A+B ≥20 pg/mL). Abbreviations: CDI, Clostridioides difficile infection; ELISA, enzyme-linked immunosorbent assay; GCSF, granulocyte colony-stimulating factor; Ig, immunoglobulin; IL, interleukin; IQR, interquartile range; MCP-1, monocyte chemoattractant protein-1; NAAT, nucleic acid amplification testing; neg, negative; TNF-ɑ, tumor necrosis factor ɑ; Tox20, CDI-NAAT or Carrier-NAAT cohorts who had stool toxin A+B concentrations of ≥20 pg/mL; RFV, relative fluorescent unit; VEGF-A, vascular endothelial growth factor A.

Figure 1.

Dot plots showing the distribution of GCSF concentrations (pg/mL) in 6 cohorts: patients with symptomatic CDI (CDI-NAAT, defined by positive stool NAAT result and new-onset diarrhea), asymptomatic carriers (Carrier-NAAT, defined by positive stool NAAT result and absence of diarrhea), NAAT-negative patients with diarrhea (Diarrhea, NAAT-neg) and without diarrhea (No diarrhea, NAAT-neg), and symptomatic (CDI-Tox20) versus asymptomatic (Carrier-Tox20) cohorts (subsets of CDI-NAAT and Carrier-NAAT, defined by positive stool NAAT result and stool toxin A+B ≥20 pg/mL by Simoa). The bottom and top edges of the boxes for each cohort indicate the IQR, the horizontal line bisecting the box indicates the median value, and the whiskers represent values up to 1.5 IQR. The inset table indicates P values for comparison of median values for paired cohorts as indicated. Abbreviations: CDI, Clostridioides difficile infection; GCSF, granulocyte colony-stimulating factor; IQR, interquartile range; n/a, not applicable; NAAT, nucleic acid amplification test; neg, negative; Simoa, Single Molecule Array; Tox20, CDI-NAAT or Carrier-NAAT cohorts who had stool toxin A+B concentrations of ≥20 pg/mL.

Figure 2.

Dot plots showing the distribution of IL-6 concentrations (pg/mL) in the same 6 cohorts defined in Figure 1. Abbreviations: CDI, Clostridioides difficile infection; IL, interleukin; NAAT, nucleic acid amplification test; neg, negative; Tox20, CDI-NAAT or Carrier-NAAT cohorts who had stool toxin A+B concentrations of ≥20 pg/mL.

Figure 3.

Dot plots showing the distribution of TNF-α concentrations (pg/mL) in the same 6 cohorts defined in Figure 1. Abbreviations: CDI, Clostridioides difficile infection; NAAT, nucleic acid amplification test; neg, negative; TNF-ɑ, tumor necrosis factor ɑ; Tox20, CDI-NAAT or Carrier-NAAT cohorts who had stool toxin A+B concentrations of ≥20 pg/mL.

Figure 4.

Dot plots showing the distribution of serum IgG anti–toxin A concentrations (ELISA units) in the same 6 cohorts defined in Figure 1. Abbreviations: CDI, Clostridioides difficile infection; ELISA, enzyme-linked immunosorbent assay; IgG, immunoglobulin G; NAAT, nucleic acid amplification test; neg, negative; toxA, toxin A; Tox20, CDI-NAAT or Carrier-NAAT cohorts who had stool toxin A+B concentrations of ≥20 pg/mL.

Figure 5.

Dot plots showing the distribution of the relative fluorescent values for stool IgA anti–toxin B detection, as measured on a VIDAS platform, in the same 6 cohorts defined in Figure 1. Abbreviations: CDI, Clostridioides difficile infection; IgA, immunoglobulin A; NAAT, nucleic acid amplification test; neg, negative; Tox20, CDI-NAAT or Carrier-NAAT cohorts who had stool toxin A+B concentrations of ≥20 pg/mL.

Figure 6.

Dot plots showing the distribution of the relative fluorescent values for stool IgG anti–toxin B detection, as measured on a VIDAS platform, in the same 6 cohorts defined in Figure 1. Abbreviations: CDI, Clostridioides difficile infection; IgG, immunoglobulin G; NAAT, nucleic acid amplification test; neg, negative; Tox20, CDI-NAAT or Carrier-NAAT cohorts who had stool toxin A+B concentrations of ≥20 pg/mL.

Figure 7.

Receiver operating curve indicating the accuracy of GCSF concentrations for discriminating between patients with symptomatic CDI (CDI-NAAT, defined by positive stool NAAT result and new-onset diarrhea) and asymptomatic carriers (Carrier-NAAT, defined by positive stool NAAT result and absence of diarrhea). The AUC is 0.842. Abbreviations: AUC, area under the curve; CDI, Clostridioides difficile infection; GCSF, granulocyte colony-stimulating factor; NAAT, nucleic acid amplification test.

DISCUSSION

Despite the widespread availability of highly sensitive assays (such as NAAT) for the identification of toxinogenic C. difficile in stool, the diagnosis of CDI remains challenging and controversial [1, 2, 4]. Furthermore, disease outcomes may vary according to the diagnostic methods used [5, 6]. The presence of toxinogenic C. difficile alone is not sufficient for diagnosis; clinical factors are also required, most notably the presence of significant diarrhea [4]. Symptomless carriage of toxinogenic C. difficile is common, especially in hospital populations [3, 11, 27]. In a recent study we, too, found that 11.5% of hospital inpatients who recently had received antimicrobial therapy were asymptomatic carriers of toxinogenic C. difficile (Carrier-NAAT) [7]. Asymptomatic carriage of toxinogenic C. difficile is not likely to be a bacterial effect as strain types are generally similar to those prevalent in CDI patients [7]. Furthermore, stool concentrations of C. difficile toxins (A and/or B) failed to differentiate CDI-NAAT from Carrier-NAAT patients [7]. Compounding the diagnostic challenge, colonized patients may experience diarrhea from other causes. The current Infectious Diseases Society of America (IDSA)– Society for Healthcare Epidemiology of America (SHEA) guidelines on CDI acknowledge that NAAT-positive patients without CDI may have diarrhea from other etiologies such as chemotherapy, other medication use, dietary intolerances, laxative administration, or other gastrointestinal illnesses such as irritable bowel syndrome (including postinfectious irritable bowel syndrome) or inflammatory bowel disease [4].

In this study we attempt to identify objective measures associated with CDI, where CDI is defined according to the IDSA-SHEA guidelines (ie, either new-onset diarrhea plus NAAT-positive stool [CDI-NAAT cohort] or new-onset diarrhea plus toxin-positive stool [CDI-Tox20 cohort]). CDI-NAAT (and CDI-Tox20) patients were compared with several relevant control groups: 1) Carrier-NAAT patients (colonized, without diarrhea), 2) patients with diarrhea (confirmed using the same definition used for the CDI-NAAT cohort) but who tested NAAT negative on clinical C. difficile testing, and 3) patients with no diarrhea who were NAAT negative (tested to screen for eligibility for the Carrier-NAAT group).

The clinical relevance of both adaptive and innate immune responses to C. difficile and its toxins was introduced briefly above. In this study we were successful in identifying multiple serum markers of inflammation that were elevated in the CDI-NAAT (and CDI-Tox20) cohort compared with each of the other groups. This finding is consistent with toxin-mediated colitis in CDI that is associated with a systemic innate immune response that can be identified using highly sensitive cytokine immunoassays. We also identified adaptive immune responses in the form of increased IgG anti–toxin A and IgA/IgG anti–toxin B responses in blood and in stool, respectively. A small group of analytes (IL-6, GCSF, TNF-α, and IgG anti–toxin A in blood, as well as IgA and IgG anti–toxin B in stool) show excellent promise in objectively separating CDI patients from each of the other groups. These analytes, used singly or in combination, may prove clinically valuable in supporting or refuting a diagnosis of CDI in patients who test positive for C. difficile toxins or for toxinogenic C. difficile—thus addressing an important clinical diagnostic conundrum.

Our study has several strengths. We prospectively enrolled multiple relevant and diverse patient cohorts using carefully constructed and prospectively applied case definitions. These case definitions, in particular the definition of CDI, were based on national IDSA-SHEA guidelines. The study groups were divided into either NAAT-positive or NAAT-negative patients and further divided by the presence or absence of diarrhea (using the IDSA-SHEA definition). We performed NAAT and quantitative toxin assays for toxin A and toxin B for all subjects as well as a broad panel of innate and adaptive immune response markers. We also carefully monitored relevant clinical factors including the use of CDI therapy and compiled an extensive database of clinical and laboratory data for our analyses. Notably, most of the CDI-associated markers that we identified are readily accessible for measurement using standard available assays.

Our study does have several limitations. We lacked access to colonic tissue to evaluate for the presence of colitis, either histologically or by measurement of tissue inflammatory mediators (at the proteome or transcriptome level). This is relevant since one of our main speculations is that C. difficile toxin–induced colitis is the principal source of the inflammatory markers evident in the CDI group and lacking in those without CDI.

Having identified several promising and accessible potential markers for C. difficile toxin–induced colitis, our next steps will be to validate their clinical utility in 2 areas. First, and more importantly, we will attempt to differentiate NAAT-positive patients with CDI from colonized patients with non-CDI diarrhea. Such a study will require us to develop and validate an effective case definition for this latter elusive group based on clinical parameters, pattern of diarrhea, comprehensive stool testing for C. difficile toxins and toxinogenic organisms, and response to or need for anti-CDI therapy. Second, we will evaluate whether the markers of CDI that we have identified can be clinically useful in identifying patients who will go on to develop severe, complicated, or recurrent disease. We will also combine ultrasensitive and quantitative stool toxin measurement with these analyses [7–9]. Our ultimate goal is to provide new, objective measures to clarify, improve, and guide appropriate CDI diagnosis, thereby streamlining and improving patient care.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Acknowledgments. The authors thank all patients who participated in this study, as well as Carolyn Alonso, Javier Villafuerte Gálvez, and Christine Geer.

Financial support. This work was supported by grants from the National Institutes of Health/National Institute of Allergy and Infectious Diseases (grant number 1R01AI116596-01 to N. R. P. and C. P. K.) and Institut Merieux (to N. R. P. and C. P. K.). Calprotectin, Vidas IgA and IgG, and Single Molecule Array assays were performed by bioMerieux.

Potential conflicts of interest. A. B., A. F., A. L., and M. M. are employees of bioMerieux. C. P. K. has acted as a paid consultant to Artugen, Facile Therapeutics, First Light Biosciences, Finch, Matrivax, Merck, Seres Health, and Vedanta. M. M. reports ownership of bioMerieux stock. M. M. and A. F. have a patent pending, “Prediction of the susceptibility of an at risk patient to develop or redevelop Clostridium difficile infection.” N. R. P. has acted as a paid speaker for Singulex. All other authors report no potential conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.