Use of PCR cycle threshold values for toxins A and B quantification in Clostridioides difficile infections

Abstract

Increasingly, tcdA/tcdB qPCR is being used to diagnose Clostridioides difficile infections (CDI). Under the hypothesis that toxin genes’ quantification through Ct values could potentially improve clinical accuracy, this study aimed to assess the linearity of the C. difficile assay on the QIAstat-Dx^® Gastrointestinal Panel 2 (GI2 Panel) and to correlate bacterial to toxins load. Four analytical and thirty-five clinical toxigenic C. difficile isolates were quantified using three validated standard curve qPCR assays targeting adK, tcdA and tcdB genes. Of these, twelve were then tested to characterize the linearity of the C. difficile assay on the QIAstat-Dx^® GI2 Panel. Statistical analysis of the Ct values of adK, tcdA and tcdB obtained from standard curves presented an excellent linear fit (slopes range of 1.008–1.010 ± 0.001). A dynamic range of 1,000–1,000,000 copies/mL with R² ≥ 0.97 was established for the QIAstat-Dx^® GI2 Panel’s C. difficile assay. The correlation among tcdA/tcdB and adK genes allows extrapolation to pathogen concentration. The QIAstat-Dx^® GI2 Panel’s C. difficile assay demonstrated a wide linear range, allowing the accurate quantification of gene-encoding toxins A and B. This, in turn, presents a tool that could be key in establishing the relevance of toxin concentration and, potentially, a Ct cutoff at the time of CDI.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-02526-6.

Introduction

Clostridioides difficile is a common human gut colonizer that is also capable of causing infection (CDI)^1–3. It is known to be the leading cause of nosocomial diarrhea and increasing source of community-acquired diarrhea in both adults and children, and recently it has been classified as an urgent antibiotic resistance-related threat^4–6. Its occurrence is strongly related to microbiota disruption by antibiotic use, becoming a major public health concern with high morbidity and mortality worldwide^7–9. It is therefore associated with longer hospital stays and contributes to healthcare associated costs^7–9.

C. difficile’s pathogenic activity is mainly driven by the production of toxins A and B (TcdA and TcdB)^5,10. It is believed that toxin concentration correlates to CDI severity, where toxin A was first described to be more virulent in animal models^11,12. However, recent studies have shown that TcdB might have a key role in pathogenesis due to its involvement in host damage and inflammatory response^10,13,14. Even though mono-toxin C. difficile have been found in clinical isolates and are believed to be capable of causing CDI, it has been documented that most strains produce both toxins simultaneously^14–17. Genes encoding for both TcdA and TcdB are found in the chromosome in the pathogenicity locus (PaLoc)^17,18, which resembles a mobile genetic element that can be transferred from toxigenic to non-toxigenic strains^19,20. Moreover, there is recent evidence of extrachromosomal circular contigs that are similar to conjugative plasmids carrying tcdB in both toxigenic (PaLoc positive) and non-toxigenic (PaLoc negative) strains^21,22. This suggests that the number of bacterial cells in a specimen might not always correlate to the number of copies of the toxin genes.

The diagnostic approach for CDI remains a controversial topic. Although the gold standard is toxigenic culture (TC), cell cytotoxin assay (CCA) is also considered a reference method³. TC is known to be more sensitive, but CCA is more specific and convenient^23–25. However, they both require technical expertise and time (from 24 h up to 96 h), and are not standardized across testing laboratories^3,26. A simpler and faster alternative commonly used is enzyme immunoassays (EIA)²⁷. C difficile EIA can target the detection of fecal toxins A and B, as well as glutamate dehydrogenase (GDH). While GDH is regarded as a less specific assay, toxin EIA results are considered to correlate with disease, granting that EIA have been proven to be less sensitive^26,28,29. In the last decade, though, PCR (Polymerase Chain Reaction), which usually targets the genes encoding toxins A and B, has gained popularity due to its higher sensitivity, speed, and ease of use^3,30,31.

Current guidelines recommend a two-step algorithm using two qualitative tests, such as a PCR or a GDH EIA plus a toxin EIA to increase sensitivity and specificity at the time of CDI diagnosis^9,32,33. Yet this approach can potentially represent an overburden of clinical settings, causing a delay in treatment or unnecessary prescription, among other consequences^3,34. Therefore, additional research is still needed to find the best diagnostic method to distinguish CDI from colonization in a timely manner. However, several studies advocate for the potential benefits of PCR alone as a diagnostic method rather than a two-step algorithm^30,35–37.

While qualitative (positive/negative) tests show poor predictive value for CDI patient outcome, quantification of toxin genes through quantitative PCR’s (qPCR) Cycle threshold (Ct) values has the potential to improve clinical accuracy. There is evidence suggesting that lower Ct values correspond to greater quantities of the target gene which, at the same time, correlate to a higher pathogen load^25,38,39. Furthermore, it is believed that a higher concentration of bacteria might be consistent with less favorable clinical outcomes^31,40. There are many studies that, based on results that show low Ct values associated with high toxigenic C. difficile concentration in stool and worse clinical outcomes, suggest a Ct cutoff^25,41,42. The implementation of these findings at the time of CDI diagnosis might be informative about the prognosis and beneficial to differentiate causative pathogens from coinfections and from asymptomatic carriage^30,41–43.

The QIAstat-Dx^® GI2 Panel (Cat. 691412, QIAGEN, Hilden, Germany) platform is a fully automated qPCR multiplex syndromic testing device that delivers not only qualitative results but also Ct values for C. difficile (IVDR version only) and 24 other gastrointestinal pathogens⁴⁴. It has been previously demonstrated that the platform’s results are robust and linear, allowing to link Ct values to microorganism load^45,46.

Since linearity is a main feature derived from qPCR Ct values that helps understand trends and patterns, it ultimately defines a range of the PCR within which results are accurate and precise. Therefore, in this study, we aimed to assess the linearity of the C. difficile qPCR assay on the QIAstat-Dx^® GI2 Panel, which targets gene regions encoding toxins A and B. This linearity was then correlated to bacterial load obtained by means of three standard curve qPCR methods to detect and quantify a C. difficile single-copy house-keeping gene and both toxin A and B encoding genes. The results of this study aim to help support Ct value interpretation when using the QIAstat-Dx^® GI2 Panel to diagnose patients suspected of CDI and potentially improve clinical outcomes.

Results

C. difficile standard curve qPCR methods

qPCR assays were developed to amplify C. difficile’s adenylate kinase (adK), tcdA and tcdB genes. When tested against other Clostridioides strains (Table 1) to assess for possible cross-reactions, none of these three assays resulted in a positive amplification, except for tcdB assay when running Paeniclostridium sordellii analytical sample. Nevertheless, all three standard curve assays were validated in a final dynamic range of to copies/mL showing acceptable qPCR efficiencies and linearities (Table 2). All three linear ranges showed a Limit of quantification (LoQ) and a Limit of Detection (LoD) of copies/mL.

Table 1.

Clostridioides strains used for initial qPCR validation, primers’ performance testing and qPCR standard curve quantification.

Clostridioides species	Catalog ID	Supplier
C. difficile	9689	ATCCa
C. difficile	0801619	ZeptoMetrix
C. difficile	BAA-1814	ATCCa
C. difficile	43598	ATCCa
C. difficile (non-toxigenic)	0804105	ZeptoMetrix
Clostridium perfringens	0801585	ZeptoMetrix
Clostridium septicum	0801885	ZeptoMetrix
Clostridium tetani	DSM11744	DSMZb
Clostridium histolyticum	19401	ATCCa
Paeniclostridium sordellii	9714	ATCCa

^aATCC: American Type Culture Collection. ^bDSMZ: Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (German Collection of Microorganisms and Cell Cultures GmbH).

Table 2.

Range for efficiency, slope and linearity obtained during the standard curve method validation of the three assays and during the quantification of commercially supplied and clinical samples.

Assay step	Target gene	Efficiency (%)	Slope	R 2
Standard curve validation	adK	98–100	(−3.3) - (−3.4)	≥ 0.99
	tcdA	97–100	(−3.3) - (−3.4)	≥ 0.99
	tcdB	96–103	(−3.2) - (−3.4)	≥ 0.99
Samples quantification	adK	93–99	(−3.3) - (−3.5)	≥ 0.99
	tcdA	97–99	(−3.3) - (−3.4)	≥ 0.99
	tcdB	94–98	(−3.4) - (−3.5)	≥ 0.99

For repeatability assessment of each assay, plates 1 and 2 resulted in a coefficient of variation (CV%) of 7.83%5.45%, 7.21%7.62%, and 7.63%7.14% between adK, tcdA and tcdB plates respectively. For reproducibility of each assay, CV% between plates 1 and 3 was 13.61%1.01%, 6.17%2.25%, and 11.40%10.34% for adK, tcdA and tcdB, respectively.

Quantification and linearity of the samples

All commercial toxigenic C. difficile (Table 1) and clinical samples resuspended in FecalSwab™ (470 CE, Copan, Brescia, Italy) were quantified by means of the three developed assays: adK, tcdA and tcdB. All qPCR runs resulted in acceptable efficiencies and linearities (Table 2). Final concentration in copies/mL of the samples was obtained from the mean of the triplicates and applying the dilution factor of the purified sample (Table 3).

Table 3.

Analytical and clinical C. difficile samples qPCR standard curve quantification results.

		Copies/mL
Specimen ID	Sample type	adK	tcdA	tcdB
0801619	Commercial	4.05E + 08	2.83E + 08	3.75E + 08
9689	Commercial	3.74E + 09	1.36E + 09	1.88E + 09
43598	Commercial	5.75E + 09	2.11E + 09	2.28E + 09
BAA-1814	Commercial	6.23E + 08	3.83E + 08	5.52E + 08
CD01	Clinical	6.52E + 05	5.89E + 04	1.05E + 05
CD02	Clinical	3.07E + 07	1.40E + 07	1.30E + 07
CD03	Clinical	7.08E + 06	1.43E + 06	2.67E + 06
CD04	Clinical	9.91E + 05	2.26E + 05	4.44E + 05
CD05	Clinical	5.91E + 07	1.94E + 07	2.89E + 07
CD06	Clinical	1.10E + 07	6.01E + 06	8.01E + 06
CD07	Clinical	2.08E + 06	1.84E + 06	1.87E + 06
CD08	Clinical	3.00E + 05	1.24E + 05	2.10E + 05
CD09	Clinical	3.45E + 06	1.45E + 06	2.20E + 06
CD10	Clinical	5.07E + 06	2.51E + 06	3.68E + 06
CD11	Clinical	2.01E + 05	1.00E + 05	1.75E + 05
CD12	Clinical	1.69E + 06	4.96E + 05	8.63E + 05
CD13	Clinical	2.06E + 07	1.07E + 07	1.10E + 07
CD14	Clinical	9.05E + 05	5.59E + 05	1.03E + 06
CD15	Clinical	7.72E + 06	3.27E + 06	5.78E + 06
CD16	Clinical	4.12E + 04	3.92E + 04	3.03E + 04
CD17	Clinical	3.41E + 07	1.97E + 07	2.89E + 07
CD18	Clinical	2.16E + 06	1.04E + 06	1.28E + 06
CD19	Clinical	1.40E + 07	6.40E + 06	5.03E + 06
CD20	Clinical	2.33E + 05	2.40E + 05	1.67E + 05
CD21	Clinical	7.69E + 05	2.66E + 05	4.54E + 05
CD22	Clinical	2.73E + 04	1.10E + 04	1.64E + 04
CD23	Clinical	1.46E + 07	5.56E + 06	7.40E + 06
CD24	Clinical	2.13E + 05	7.24E + 04	1.30E + 05
CD25	Clinical	4.19E + 04	1.34E + 04	2.77E + 04
CD26	Clinical	1.36E + 07	5.67E + 06	9.14E + 06
CD27	Clinical	1.09E + 06	3.33E + 05	6.92E + 05
CD28	Clinical	7.39E + 05	3.52E + 05	5.06E + 05
CD29	Clinical	1.15E + 07	4.71E + 06	7.02E + 06
CD30	Clinical	1.71E + 06	8.55E + 05	1.08E + 06
CD31	Clinical	6.04E + 06	3.95E + 06	4.23E + 06
CD32	Clinical	3.41E + 06	1.42E + 06	2.16E + 06
CD33	Clinical	1.01E + 06	6.59E + 05	5.91E + 05
CD34	Clinical	8.15E + 04	2.69E + 04	4.36E + 04
CD35	Clinical	3.13E + 06	9.93E + 05	1.64E + 06

Regression with 0 intercept of the Ct values of adK with tcdA and tcdB showed an excellent linear fit, and near-ideal slopes of 1.008 ± 0.001 and 1.010 ± 0.001, respectively. In addition, the 1st and 3rd quartile of the residuals of this model are − 0.26 and 0.27 (−0.21 and 0.29), respectively, demonstrating that the data points are in excellent agreement with the fit (Table 3; Fig. 1). These results indicate a correlation between adK and toxin-coding genes concentrations in all samples tested.

Fig. 1.

Deming linear regression of the Ct values obtained by standard curve qPCR quantification of four analytical toxigenic C. difficile samples and thirty-five retrospective clinical samples. Only values obtained from the concentrations included inside the linear range were used. (A) Correlation between Ct values obtained for the house-keeping gene adK and tcdA. (B) Correlation between Ct values obtained for the house-keeping gene adK and tcdB.

All twelve samples (selected considering adequate concentration and availability, two analytical, five clinical specimens resuspended in Para-Pak^® C&S (900612, Meridian Bioscience, Cincinnati, OH) and five clinical samples in FecalSwab™ (470 CE, Copan, Brescia, Italy)) tested in triplicates using the C. difficile assay on the QIAstat-Dx^® GI2 Panel resulted in a linear range that covered from 1,000 to 1,000,000 copies/mL with a R² above 0.97 (Fig. 2). Individual Ct values obtained during linearity tests can be found in Supplementary Material 01.

Fig. 2.

Linearity evaluation using the QIAstat-Dx^® Gastrointestinal Panel 2 (691412, QIAGEN, Hilden, Germany) custom cartridges. (A) Commercially supplied toxigenic C. difficile samples. (B) Clinical samples resuspended in FecalSwab™ (470 CE, Copan, Brescia, Italy). (C) Clinical samples resuspended in Para-Pak^® C&S (900612, Meridian Bioscience, Cincinnati, OH). Ct: cycle threshold. Conc: concentration.

Discussion

In this study, we demonstrated a wide linear range of the C. difficile assay on the QIAstat-Dx^® GI2 Panel, spanning from 1,000 to 1,000,000 copies/mL. These linearity results were obtained by testing two commercially available toxigenic C. difficile samples and ten clinical samples, resuspended in Para-Pak^® C&S and in FecalSwab™, covering all collection methods validated for the QIAstat-Dx^® GI2 Panel. Furthermore, we also assessed the correlation between the quantification of the genes encoding toxins A and B and the quantification of a single-copy house-keeping gene (adK) by means of the development of three reliable qPCR standard curve quantification methods.

Standard curve quantification results and analyses indicated that, despite there being a correlation between concentration of adK and toxin genes, the number of copies for each gene is not the same within one single specimen. This raises uncertainty regarding CDI, such as the possibility of harboring both toxigenic (PaLoc positive) and non-toxigenic (PaLoc negative) strains within one single patient, which would explain the higher concentration of adK gene copies. Moreover, since multiple copies of the PaLoc have not been described, a higher concentration of tcdB compared to tcdA could suggest either the existence of some TcdA-TcdB + organisms, or of multiple copies of TcdB + extrachromosomal molecules, as previously described^16,47,21. Similarly, a higher concentration of tcdA compared to tcdB could suggest the presence of TcdA + TcdB- bacteria, which is in line with previous studies^15,17. Nevertheless, our results indicate a correlation between bacterial and toxin loads, meaning tcdA and tcdB qPCR results could be linked to bacterial load.

C. difficile diagnostic method is still subject to controversy. Despite some studies not being able to find a statistically significant correlation between qPCR results and CDI severity^48,49, many others have indeed reported correlation between Ct values and C. difficile concentration in stool and, therefore, clinical outcomes^{25,34,39,41,42,50}. In this regard, our results support the linearity of Ct values obtained for clinically relevant C. difficile concentrations with the commercially available QIAstat-Dx^® GI2 Panel device. This could facilitate Ct interpretation at the time of diagnosis, allowing for a single test with a dual reporting of the qualitative PCR result plus the Ct value interpretation, rather than a two-step algorithm. This additional information has been postulated to assist in the determination of symptomatic causality and would allow for a faster turn-around time and a higher ease of use, but keeping the high sensitivity characteristic of the qPCR^3,30,42.

Moreover, it has been shown that unformed stool quality (Bristol scale 5–7), collection method and processing, patient’s age and immune status, and C. difficile strain type do not alter median Ct values, suggesting that a Ct cutoff may be applied⁵¹. However, the use of different materials and reagents, or assay conditions discrepancies are factors that might determine Ct variation³. Such variation could be avoided, and test reproducibility could be further enhanced by the use of the fully automated qPCR multiplex syndromic testing platform QIAstat-Dx^® GI2 Panel, integrating all steps from nucleic acid extraction to results’ interpretation. In addition, the QIAstat-Dx^® GI2 Panel, a PCR-based device, could overcome the lack of sensitivity derived from standard CDI diagnostics methodologies, such as EIA. Albeit a positive toxin EIA will most probably indicate a C. difficile active infection, not all infections will be detected by this method^28,29. This gap might, therefore, be filled by the QIAstat-Dx^® GI2 Panel, that accurately quantifies C. difficile toxins in a specimen. This precise toxin quantification, in turn, could potentially pave the way to help establish a Ct cutoff to allow to distinguish disease from colonization.

Nevertheless, our study was limited by the fact that both qPCR standard curve quantification assays developed to amplify C. difficile toxin genes resulted in cross-reaction with P. sordellii, which harbors two genes encoding the lethal (TcsL) and hemorrhagic (TcsH) toxins, homologous to TcdB and TcdA, respectively⁵². Nonetheless, this was mitigated by the assay developed for adK and by the QIAstat-Dx^® GI2 Panel cartridge, both of which do not show cross-reactivity with P. sordellii⁴⁴, and by the fact that this pathogen is rarely found in the human gut, and it would therefore not be expected in stool samples^52–54. Furthermore, the linearity above copies/mL could not be established due to the lack of clinical samples with a higher concentration of the target of interest. Further studies with higher concentrated samples would be needed to overcome this limitation and allow the establishment of a higher upper limit for the linear range.

In conclusion, we were able to demonstrate correlation between bacterial and toxin loads in both analytical and clinical samples tested through the quantification of a C. difficile house-keeping gene, and toxins A and B encoding genes. We also defined a linear range for the C. difficile assay of the QIAstat-Dx^® GI2 Panel, which allows the quantification of toxins A and B. This feature allows the QIAstat-Dx^® GI2 Panel to be a potential key tool in establishing the relevance of toxin concentration and a Ct cutoff at the time of C. difficile infection.

Materials & methods

In this study, an initial design, optimization and validation of three standard curve quantification assays was performed in order to characterize clinical samples subsequently used to assess and establish linearity in the QIAstat-Dx^® GI2 Panel cartridge.

Oligonucleotides design

Three real-time qPCR assays were designed targeting the adK single-copy house-keeping gene for the absolute quantification of C. difficile copies of genomic material⁵⁵, and two more assays to amplify the genes encoding TcdA and TcdB. Out of six designs tested targeting two house-keeping genes, adK and tpi, the adK assay used during this study was chosen based on PCR performance, assessing reproducibility, repeatability, replicability, sensitivity and amplification curves’ shapes, among other parameters. Primers were designed to target a 132 base-pair (bp) size fragment of adK, a 80-bp size fragment of tcdA and a 162-bp size fragment of tcdB using Primer3 tool implemented in Geneious software v2024.0.4 (www.geneious.com). The targeted regions correspond to positions 118,849–118,981 for adK assay, 744,059–744,139 for tcdA and 737,185–737,347 for tcdB assays of the 4.1 Mb size reference C. difficile genome (GenBank sequence ID CP076401). One labeled probe was designed for each assay to bind between primers with an annealing temperature (Ta) above the primer of the same direction to optimize PCR performance (see Table 4). BLASTN analysis (https://blast.ncbi.nlm.nih.gov/Blast.cgi) was run to reduce possible homologies that could produce nonspecific amplifications in the PCR. All six primers were ordered from Integrated DNA Technologies (IDT, Coralville, IA), altogether with the three FAM labeled and double-quenched (ZEN/IBFQ) probes according to supplier recommendations to improve optical performance.

Table 4.

Primers’ and probes’ sequences used for adK, tcdA and tcdB C. difficile genes quantification. All probes were labeled with FAM and double quenching ZEN/IBFQ.

Gene	Oligonucleotide	Sequence (5’−3’)	Length (bp)	Ta (ºC)	Amplicon size (positions in reference genome)
adK	Forward primer	AGAAGAATATGTAAGTCTTGTGGAG	25	61.7	132 bp (118,849–118,981)
	Reverse primer	CTTAGATACAGTTTCTTCATTATCATCAG	29	61.6
	Labeled probe	ACATCACATACACCTTCTACTTTAGGAGGA	30	66.5
tcdA	Forward primer	ACTTASAAGTAGGTTTTATGCCAG	24	62.1–62.5	80 bp (744,059–744,139)
	Reverse primer	ATAGTAAGCTGAYRCATAAGCTC	23	59.3–63.6
	Labeled probe	TGGRCCACTTAAACTTATTGTRGAGCGA	28	66.3–69.9
tcdB	Forward primer	TCTTAGTAATTTAAGTGTAGCAATGAAAGT	30	63.2	162 bp (737,185–737347)
	Reverse primer	ATAGGTAATCCTTCAGATAATGTAGG	26	62.8
	Labeled probe	TTGATACTAATTCAACAACTTTGGCTGCATCTGT	34	68.8

Sample collection and DNA extraction for assays optimization

Four commercial analytical toxigenic C. difficile samples were collected for initial qPCR validation and oligonucleotides’ performance testing and six commercially available samples from the genus Clostridioides were also used for analytical reactivity testing (Table 1). Deoxyribonucleic acid (DNA) extractions for these samples were performed using the automatic purification procedure with the QIAcube system (QIAGEN, Hilden, Germany) and a customized protocol with the reagents of the QIAamp Cador Pathogen Mini Kit (54106, QIAGEN, Hilden, Germany). Correct extraction of all samples was confirmed by qPCR amplification of the internal control added during extraction process, as it is done in the QIAstat-Dx^® GI2 Panel (data not shown).

In addition, three customized commercial double-stranded DNA (dsDNA) fragments (gBlock™: IDT, Coralville, IA) corresponding to positions 118,821–119,021 for adK, 744,017–744,217 for tcdA, and 737,162–737,364 for tcdB of the reference C. difficile genome, along with one C. difficile commercial analytical sample (0801619) were used during the validation of the standard curves (Table 1). The gBlock™ fragments were resuspended in Tris-EDTA buffer 1 × (93283-500ML, Merck KGaA, Darmstadt, Germany) and concentration for each stock was calculated by Nanodrop 2000 C (ThermoFisher Scientific, Waltham, MA) measure using the molecular weight of the sequence, resulting in , , and copies/mL for adK, tcdA and tcdB, respectively.

qPCR assay development

Three standard curve qPCR methods were developed to detect and compare total number of TcdA and TcdB-coding genes copies to total C. difficile bacterial load. Standard curves were validated according to official guidelines and reference publications, accepting values of efficiency E = 90–110%; slope S = (−3.1) - (−3.6), and linearity R² ≥ 0.9^56–59. Each standard curve was constructed using nine consecutive 10-fold dilutions of each of the three dsDNA gBlock™ stocks in 10 replicates by dilution point, starting from copies/mL. For each standard curve, Ct values of those dilution points with linear behavior were used to establish the dynamic range. LoQ was considered as the last detected dilution inside the linear range with a CV ≤ 25%. The linear range for the three standard curves was validated running five of the nine initial dilution points ( to copies/mL) in triplicate. Precision, repeatability and reproducibility of the qPCR assays were evaluated by running a total of three 96-well plates per assay on different days, testing three logarithmic dilutions of two different extractions of sample 0801619 alongside the validated dynamic range of each standard curve. For repeatability of each assay, plates 1 and 2 were run under the same testing conditions on different days. For reproducibility of each assay, plates 1 and 3 were run on different instruments and by different operators on nonconsecutive days.

All qPCR amplifications were conducted following Juanola-Falgarona et al. (2022)⁴⁵ study with minor modifications. A total of 9 µL of template in a final volume of 15 µL was prepared, containing 600 nM of each primer and 300 nM of the labeled probe resuspended in Tris-EDTA buffer 1 × (93283-500ML, Merck KGaA, Darmstadt, Germany), 1.5 mM trehalose solution (T9531-5G, Merck KGaA, Darmstadt, Germany), and 1.5 µL customized Master Mix including 1.2 mM dNTPs, 4 mM MgCl₂, and antibody mediated Phoenix Hot Start Polymerase. Thermal cycles consisted of an initial denaturing step of 30 s at 95ºC, followed by 40 two-step cycles of denaturing at 94ºC for 4 s and annealing-extension step at 60ºC for 21 s. Negative controls were included in all qPCR runs to confirm that there was no cross-contamination. qPCR reactions were run using QuantStudio™ 5 Real-Time PCR System (ThermoFisher Scientific, Waltham, MA) and analyzed in QuantStudio™ 5 Design and Analysis Software v1.5.3., with a threshold of 150,000 for adK assay and 100,000 for tcdA and tcdB assays.

Analytical and clinical samples quantification

A total of four commercial analytical toxigenic C. difficile samples (Table 1) and thirty-five retrospective blinded clinical stool samples collected at Parc Taulí Hospital Universitari (Sabadell, Spain) and internally at QIAGEN STAT-Dx (Barcelona, Spain) following the QIAGEN GmbH standard anonymization procedures were enrolled in this study. These samples were confirmed to be C. difficile positive at source either by QIAstat-Dx^® Gastrointestinal Panel (691411, QIAGEN, Hilden, Germany), QIAstat-Dx^® GI2 Panel (691412, QIAGEN, Hilden, Germany), culture, or alternative C. difficile toxins detection by at-source developed PCR assay. All samples were resuspended in FecalSwab™ (470 CE, Copan, Brescia, Italy), and subsequently quantified using the three validated qPCR assays. Total DNA extraction protocol and qPCR conditions for all samples were as previously described for the standard curve construction. For each assay and 96-well plate, a standard curve consisting of 3 replicates for each dilution within the validated dynamic range was constructed and run alongside triplicates of the samples to be quantified. Final concentration (copies/mL) values were calculated using the linear regression extracted from the validated standard curve.

Using the Ct values of the standard curve qPCR quantification assays as input, the correlation between the Ct values of the house-keeping adK with both toxin genes, tcdA and tcdB was tested using Deming regression with 0 intercept. Only Ct values obtained from concentrations that fell within the linear range were used.

Cartridge linearity

After quantification by means of the standard curve assays, two analytical toxigenic C. difficile samples (9689 and 0801619, Table 1) alongside ten clinical samples were selected to perform linearity testing for C. difficile using the C. difficile assay on the QIAstat-Dx^® GI2 Panel and a QIAstat-Dx^® analyzer system (QIAGEN, Hilden, Germany). The selection of the samples was done based on availability and pathogen concentration, choosing those with the highest bacterial load to obtain a larger dynamic range.

QIAstat-Dx^® GI2 Panel cartridges are composed of 8 reaction chambers (RCs) with one different multiplex Master Mix targeting different pathogens in each of them⁴⁴. For the purpose of this study and to optimize the number of replicates obtained per run, customized cartridges were manufactured. These customized cartridges contained the multiplex Master Mix targeting C. difficile in all RCs except for the one with the cartridge’s internal control. Customized QIAstat-Dx^® GI2 Panel cartridges were used according to QIAstat-Dx^® GI2 Panel cartridges manufacturer recommendations.

All twelve previously quantified samples were prepared in semi-logarithmic serial dilutions starting from stock concentration for the analytical samples, or from the resuspension in the corresponding media for the clinical samples. More specifically, five clinical samples were resuspended in Para-Pak^® C&S (900612, Meridian Bioscience, Cincinnati, OH) and five in FecalSwab™ (470 CE, Copan, Brescia, Italy). Three cartridges by dilution point were run, resulting in a total of 21 data points per dilution per sample. Linear range, linearity, slope and efficiency were characterized using the same parameters as during the standard curve validation. In all cases, dilutions inside the dynamic range were included in the calculations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization, N.C., K.V., C.C.F., L.P. and O.V.; Funding Acquisition F.V.D.; Project Administration D.L. and F.V.D.; Experiments Design and Performance R.S.A, N.C., L.R., S.H., C.C.F., L.P. and O.V.; Statistical analysis R.S.A. and D.L.; Data Accuracy C.C.F., L.P. and O.V.; Data Analysis R.S.A., C.C.F., L.P. and O.V.; Manuscript Writing R.S.A., C.C.F., L.P. and O.V.; all authors edited and reviewed the manuscript.

Data availability

The datasets generated during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

This study was funded by QIAGEN STAT-Dx Life S.L., who also provided support in the form of salaries for most authors: R.S.A., N.C., L.R., S.H., K.V., D.L., F.V.D., C.C.F. and L.P. O.V. declares no conflicts of interest.