Diagnosis of Clostridium difficile: a wrong algorithm and an improvement in immunoassays

Diagnosis of Clostridium difficile: a wrong algorithm and an improvement in immunoassays | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Diagnosis of Clostridium difficile: a wrong algorithm and an improvement in immunoassays Juan Luis Recio This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3327753/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract We compared the 2-step and 3-step algorithms for diagnosis of Clostridium difficile infection (CDI) in clinical cases between October and December 2023. We observed in 3-step algorithm 22 false positives (FP) among the 62 positives (35%), compared to 40 positives without FP in 2-step algorithm. All FP are positive for glutamate dehydrogenase (GDH) and Toxin A and/or B immunochromatography (IC), and negative for toxigenic C.difficile PCR. When processing the samples with a solution of leukocidins and Methicillin Resistant Staphylococcus aureus (MRSA) the FP to GDH are negative, as well as FP from different immunoassays. The cause of FP is an interspecies cross-reaction between human leukocytes and mouse immunoglobulins, a method-dependent error (immunoassay) that increases the probability of FP in the 3-step algorithm. Modified algorithms and sample processing with MRSA and leukocidins allow equivalent results to the 2-step algorithm, with a reduction in the total number of molecular tests required (47-61%). Biological sciences/Microbiology/Clinical microbiology Health sciences/Diseases/Gastrointestinal diseases/Gastroenteritis Figures Figure 1 Figure 2 TEXT Since the appearance of diagnostic algorithms for CDI, these have been modified and adjusted according to scientific evidence and diagnostic needs until reaching those currently used and recognised (1) , the main ones being the 2-step and 3-step algorithms, the latter being considered non-inferior, faster and less costly. When clinical evidence of CDI is found, samples are sent to the laboratory and, if they are adequate, they are analysed to make a diagnosis. In our department, a 3-step diagnostic algorithm was used: -First step: GDH IC (negative is reported, positive at the second step). -Second step: Toxin A and/or B IC (GDH+ and Toxin+ reported positive; GDH+ and Toxin- at the third step). -Step 3: Molecular test by polymerase chain reaction (PCR) for toxigenic C.difficile (Genexpert C.difficile/Epi, CEPHEID) with result reported. The 2-step algorithm does not use Toxin IC, it uses GDH IC and if positive PCR. The immunoassays used were biotical C.difficile GDH card and biotical C.difficile Toxins A+B card, both from Biotical Health. In May 2022, in the microbiology laboratory of Hospital San Cecilio, we observed three doubtful positive cases in a few days. After reviewing the clinical manifestations of the patients and the results of the cultures, we suspected a diagnostic error: Salmonella enterica was isolated and identified in one case and Campylobacter spp. in the other two. Although cases of coinfection are known, three cases in a week is very unusual for a microbiology service that covers 400,000 people, with more than 200 cases investigated per month and a prevalence of 9%. PCR for toxigenic C.difficile was performed on the three doubtful samples and the results were negative. A retrospective analysis of the last month found three other confirmed cases of misdiagnosis: preserved samples were PCR negative for toxigenic C.difficile . Six other suspected cases were classified as "probable" FP because no specimen was recovered: they had other viral or bacterial infections that could explain the clinical manifestations, or responded well to treatments not suitable for CDI. These 12 cases had in common that they did not fit the classic epidemiological criteria for CDI: hospitalisation, surgery and/or previous antibiotic treatment. Our department established special surveillance for cases with these characteristics, checking the positive cases with PCR. Between June and September this special surveillance detected 10 FP cases allowing them to be correctly reported and one probable FP, the sample was lost. Upon review of the 10 FP, one of these did not meet the epidemiological criteria: it had recent previous hospitalisation and antibiotic treatment. At the same time, these clinical samples were frozen to find out where the error was, after ruling out IC batch problems. In October, the commercial company repeated CI and PCR by a different method (Allplex GI Bacterial I Assay, SEEGENE) on 10 of these samples. The PCR results were the same as those obtained previously by Genexpert, negative for CDI, and coincided with the cultures: Allplex GI also detects other pathogens such as Salmonella, Yersinia, Shigella and Campylobacter . Not so the result of the GDH and Toxin IC: only 2 of the 10 samples were still GDH positive, and of these only one was also Toxin positive. Faced with these errors of the 3-step algorithm, the management of the service allowed a change in the diagnostic method from October, although IC Toxin was not stopped, the diagnosis was a 2-step algorithm: GDH, negative is reported, and positive is PCR to decide the diagnosis (regardless of the result of the IC Toxin). Why were samples previously positive and now not positive? Freezing of samples and duration of freezing. The samples that remained positive for GDH were the most recent ones: they had been kept for less time frozen. The interference causing FP was sensitive to freezing. Another clue was provided by a review of the 23 cases detected (16 FP and 7 "probable" FP): 11 cases with positive cultures for enteropathogenic bacteria (6 Campylobacter , 4 Salmonella and 1 Shigella ); and among the remaining cases 3 were diagnosed with inflammatory bowel disease (IBD), 2 were cross-reactive for other IC ( Rotavirus, Adenovirus and Norovirus ) and another case in which a parasite study was requested showed abundant leucocytes under the optical microscope (OM). Most of the samples were inflammatory diarrhoea. What if all these FP were caused by the presence of leukocytes in the stool? Proving this was complex because leukocytes are not observable by OM in all samples and their presence in suspected cases of CDI is normal. To test this hypothesis we needed sample processing to remove leukocytes without altering GDH and Toxin: after processing, repeat IC in true positives (TP) would remain positive and in FP would be negative. Although freezing had guided us, it is not a good experimental model (slow, laborious and unreliable). Heating was not considered because of denaturation and other problems related to sample handling. The use of monoclonal antibodies, biological agents, ionising radiation, acids, bases, salts, etc. were ruled out for various reasons. In a microbiology laboratory there is a simple and safe way to lyse leukocytes: Methicillin Resistant Staphylococcus aureus (MRSA). All S.aureus produce at least 3 types of leukocidins (HlgAB, HlgCB, and LukAB/HG), while the most virulent isolates produce up to 5 types (the above plus Panton-Valentine leukocidin [PVL], and LukED) (2) . These leukocidins are bicomponent water-soluble proteins excreted by the bacteria, with the ability to recognise leukocytes, insert into their membranes and generate pores, which ultimately lyse them. LukAB/HG is slightly different from the others, existing as a stable heterodimer in solution and establishing a reservoir of toxin on the surface of the bacterium. In addition, most of the immunoassays used have been tested with different bacteria, including S.aureus , and cross-reactivity has been ruled out. Since October, all GDH IC positive samples were subjected to: Toxin IC, PCR for toxigenic C.difficile and GDH IC after processing the sample with a suspension of leukocidins and MRSA (see Methods). With this processed sample, Toxin IC was repeated on an ad hoc basis due to lack of resources. RESULTS A total of 660 samples of suspected CDI cases were processed between October and December 2022, a total of 77 samples were GDH positive: 38 Toxin positive and 39 negative. Only 3 of these 77 samples (3.9%) were grossly purulent and/or bloody. - Of the 38 GDH+/Tox+ samples: 16 were positive by PCR and 22 were negative (among the latter 3 were very purulent and/or bloody). - Of the 39 GDH+/Tox- samples: 24 were PCR positive and 15 negative. A total of 40 cases were reported as positive. In contrast, the 3-step algorithm would have resulted in 62 positive cases (38 GDH+/Tox+ plus 24 GDH+/Tox-/PCR+), with an error of 35% (22 FP). Regarding the GDH performed after sample processing with MRSA and leukocidins (MRSA-GDH): -Of the 38 samples +GDH/+Tox: 100% concordance with PCR. -Of the 39 +GDH/+Tox samples: 92.3% concordance. The 24 PCR-positive samples were also positive for MRSA-GDH; but among the 15 PCR-negative samples there were 3 positive for MRSA-GDH. Modified algorithms (figure 1 and 2) with this sample processing obtain the same results as the 2-step algorithm, and with a reduction in total PCR needed between 44-64% (table 1). We have included as a paradox the diagnosis of CDI in a single step, using only MRSA-GDH determination: only 3 FP compared to the 22 FP of the 3-step diagnosis, and without the need to use a single PCR. With 40 TP among 660 samples, we have a prevalence of 6%, down from 7.5-9% in previous months. Among the 22 GDH+/Tox+ FP, a total of 8 cases were patients with hospitalisation, surgery and/or recent previous antibiotic treatment. Estimating 100% sensitivity in the GDH IC and the 2-step algorithm as the gold standard, although it is not correct, we can calculate theoretical data such as the false positive ratio (FPR) that allow us to discuss the results: - GDH IC specificity: 583/(583+37) = 0.9403 - 3-step algorithm specificity: 598/(598+22) = 0,9614 - MRSA-GDH IC specificity: 617/(617+3) = 0.9951 - GDH IC FPR = 0,0597 - 3-step algorithm FPR = 0,0386 DISCUSSION Among the studies of ICD diagnostic methods since 2009, we have not been able to find any that compare the 2-step and 3-step algorithms exactly as we know them today. There are some similar studies (10,16) , which are not comparable for different reasons: low sample size or testing by cell cytoxicity neutralization assay (CCNA) on GDH+/Toxin+ results instead of PCR. Although years ago the CCNA was considered the gold standard, it is a subjective method, not very reproducible and gives results after 24-48 hours (15,17) . These studies are used in other subsequent studies in which 2-step algorithms are evaluated to argue for the use of 3-step algorithms, which would allow similar results to be obtained with a lower PCR cost (11, 12, 13, 14) , although without experimentation. Thus, we found a meta-analysis with a theoretical construction of a 3-step algorithm on 10,000 samples (9) , in which method errors are considered as independent and with a prevalence of 5%, with specificities of 0.87 for the GDH test and 0.96 for the Toxin test: 432 VP and 49 FP (similar total of FP in our study, but with 15 times less sample size). We observed that 3-step algorithms, where the first step was PCR (21) did not present this problem: the total error frequency is reduced as the product of the independent error frequencies of each method (PCR vs. IC). With our results we can think that we use a Toxin IC with very low specificity if we consider as in other studies (9) that the errors of each IC are independent of each other, the total error frequency is given by the multiplication of the individual error frequencies for each method. For the FP, this frequency is the false positive ratio (FPR) that we can calculate with the specificity. As all the FP in the 3-step algorithm are GDH+/Toxin+, we can infer that the FPR of the 3-step algorithm is equal to the product of the FPR of GDH IC by the FPR of Toxin IC, so we can calculate it and infer from it the "theoretical" specificity of the Toxin IC. FPR Toxin IC x FPR GDH IC = FPR3-step algorithm; FPR Toxin IC = FPR 3-step algorithm/ FPR GDH IC FPR Toxin IC = 0.0386/0.0597 = 0.64 → "Theoretical" specificity Toxin IC = 0.36 Although some Toxin IC may be less specific, this one in particular would beat all known IC in the ranking of worst specificity. We know that this is not the problem: first, because experimentally this IC (biotical C.difficile Toxins A+B card) has never had such low specificity values. And second, the errors of these two methods cannot be considered as independent because experimentally we have shown that they are related, even if different antigens are detected. They are very similar methods and share an intrinsic structural error in their design: interference by cross-reactions of human leukocytes with mouse antibodies, which is also the main cause of FP. For this reason, if this error occurs in a GDH IC for a sample, it is much more likely that the same error occurs in the same sample for the Toxin IC: the Toxin IC of the FP were also negative after processing for leukocidins and MRSA, although due to lack of resources this was only performed on some samples. Estimating a Toxin IC specificity range of 0.94-0.84 (FPR = 0.06-0.16), if the method errors were considered independent, the expected FPR of the 3-step algorithm would have been between 0.0035-0.0095; a value 11 to 4 times lower than the real value obtained. If for a sample a method-dependent error occurs in an A IC, the probability of the same method-dependent error occurring in a B IC, instead of decreasing, as they are related and dependent errors, increases, and in the case of GDH and Toxin in the order of 11 to 4 times over the expected frequency. We wonder how it is possible that such an unreliable diagnostic method as the 3-step algorithm (one out of three positives is wrong) has been used for so many years. The answer is complex, and is not only due to considering the GDH and Toxin immunoassay FP as independent and unrelated errors. It is an algorithm with high sensitivity, few false negatives (FN). A large number of samples are processed, with more than 90% being negative, so that any specific findings can be minimised among the total number of samples processed. Moreover, the general consensus on this point has always been diametrically opposed: the suspected error in the use of immunoassays for GDH and Toxin was underdiagnosis, not overdiagnosis (15) . If detecting diagnostic errors in 3 steps is difficult, it is also difficult to associate them with their cause. The use of leukocidin and MRSA in this study makes it possible to see a relationship between the presence of leukocytes in faeces (inflammatory diarrhoea) and FP, but it is impossible without this processing: one of the causes of inflammatory diarrhoea is CDI (3,4,5) . Not all inflammatory diarrhoeas are FP and not all FP are inflammatory diarrhoeas. This interspecies cross-reaction between human leukocytes and mouse immunoglobulin (IG) has been reported before (6) , and does not occur between any leukocyte and any mouse IG, it depends on the individual, the leukocyte and the type of IG. We can observe that leukocidin and MRSA processing also negativises the GDH of Tox- samples (12 out of 39 GDH+/Tox-): in these samples, the leukocytes cross-react with the IG of the GDH IC, but not with the IG of the Toxin IC. We could not observe the effect of leucodidin and MRSA processing in all Toxin+ IC due to lack of resources, but in the FP where we did, Toxin negativised. The result of this difficulty in targeting the problem is in the partial solution we took between June and September: a 3-step algorithm checking positive results with PCR based on clinical/epidemiological data surveillance. Ten FP were detected, although only 9 would have been detected if one of these cases had not been assigned by mistake to the group without hospitalisation, without surgery and without previous antibiotic treatment. The results obtained from October to December reminded us of an old laboratory axiom "where you don't look, you don't find": of the 22 FP, only 14 belonged to this special surveillance group while 8 were cases with hospitalisation, surgery and/or recent prior antibiotic treatment. When we asked ourselves whether the problem was due to a specific type of immunoassay (GDH IC in use) or to the structural design of all the immunoassays, the punctual tests with the Toxin IC indicated that it was a structural design error. We found that this problem was also present in other immunoassays with samples giving FP: in faeces for Rotavirus , Adenovirus , Astrovirus , Norovirus type I, Norovirus type II and Giardia lamblia ICs; and in respiratory samples for Respiratory Syncytial Virus, Influenza A, Influenza B and SARS-Cov-2 ICs. In all of them, the processing of the sample with leukocidins and MRSA negatived the FP result of the immunoassay, coinciding with the result obtained by molecular biology techniques. Although we could also suspect it for Helicobacter pylori IC in stool when negative with leukocidins and MRSA in some samples, we lacked a confirmatory method. Since 2010, the potential FP of the 3-step algorithm has been systematically underestimated. The main problem was the sequential use of two immunoassay methods, similar in performance and also in their intrinsic errors. As these errors are the most important, they have been considered as independent errors, when they are not. By considering the total error frequency as the product of the individual error frequencies of each method, the FP ratio (frequency) of the 3-step algorithm was underestimated: instead of decreasing in proportion to the specificity of both methods and having a frequency of 0.005, the errors committed decrease in a much smaller proportion and are 8 times higher than expected. The use of animals other than mice to obtain IG for immunoassays does not appear to avoid these interspecies cross-reactions, IG from other animals (rabbits, pigs, hamsters, cats, horses) are able to bind human leukocytes (7,8) . Processing samples with leukocidins and MRSA allows for a 92% reduction in GDH FP (from 37 to 3) enabling accurate diagnosis with less use of PCR (44-64% reduction, depending on the algorithm used, fig.1 and 2), and we have indications that its application to other immunoassays will also increase the performance and reliability of these assays. This increase in specificity of immunoassays using leukocidin and MRSA sample processing is an overall improvement for diagnosis, not only qualitatively in terms of accuracy and speed, but also in terms of ease of use and resource savings. As to how feasible it is to undertake a modified algorithm using leukocidin and MRSA sample processing in a microbiology laboratory in the short to medium term: it is unlikely. Although the use of MRSA and leukocidins is straightforward and the use of liquid cultures is less labour intensive than PCR, this processing is not yet commercially available in a kit or alongside immunoassays. The main biases we could not avoid are the use of the 2-step algorithm as the gold standard and not testing leukocidin and MRSA processing in all Toxin-positive IC. Lack of resources prevented us from ascertaining the cause of the 3 FP for GDH IC after leukocidin and MRSA processing. CONCLUSIONS We strongly discourage the use of CDI 3-step diagnostic algorithm. We recommend the use of a 2-step diagnostic algorithm, until a standard leukocidin and MRSA sample processing system is available that allows the use of modified algorithms, which is expected to increase the speed and accuracy of diagnosis at lower cost. The use of this sample processing in immunoassays other than GDH detection will allow faster, more accurate and lower cost diagnostics than current ones, although further research and validation is needed before they can be used in diagnostics. METHODS The IC used were biotical C.difficile GDH card and biotical C.difficile Toxins A+B card, both from Biotical Health, used according to the instructions provided (18,19) . The PCR method used was Genexpert C.difficile/Epi (Cepheid), used according to the instructions provided (20) . Processing of samples with leukocidins and MRSA: A pathogenic clinical strain of Methicillin Resistant Staphylococcus aureus (MRSA) is used. The culture is isolated on blood agar (BA) or blood chocolate agar (BCA), and subcultured on BA, BCA, BHI medium or any other medium that allows MRSA to produce leukocidins, at 37°C for 48h or more (depending on inoculum and total volume, 48 hours to prepare 10 ml in liquid BHI, with 12 more hours of culture more than one litre can be prepared in liquid BHI). As leukocidins are produced only in the stationary growth phase, their concentration will increase over time. Afterwards, they can be stored at room temperature: solid media for 3-4 weeks, until the drying of the colonies makes their handling difficult, and in liquid media for more than 6 months without losing their properties (tested by the author), although their stability is probably even higher. There are two options for the treatment of each sample: (a) From culture on solid medium: resuspend 5-10 colonies of the culture on blood agar in 0,5 ml of the buffer used for the immunoassay (failing this, PBS), to a McFarland > 4. (b) From liquid culture: resuspend 0,5 ml of the culture on BHI (bacteria tend to settle to the bottom) in 0,5 ml of the immunoassay buffer (failing PBS). The volumes and amounts are standard, but should be adjusted depending on the type of sample and immunoassay. In both ways we obtain a buffered solution with MRSA and leukocidins in suspension. The procedure does not differ at this point: -The sample is added, as specified by the procedure of the immunoassay technique to be used: the amounts are usually between 150 to 50 microlitres, to avoid the dilution effect they can be doubled. -Mix and vortex intensively for 30 seconds. -Incubate at 42°C for one hour, or 37°C for one and a half hours. In totally purulent or bloody samples the incubation period should not be less than 3 hours (incubations of more than 6 hours do not negative true positives). After these steps, the immunoassay protocol for detection or quantification of the test analyte is resumed. Declarations CONFLICTS OF INTEREST: Author declare no conflicts of interest. The procedure "Processing of samples with leukocidins and MRSA" is pending a patent application in the name of the author Juan Luis Recio-López. References McDonald LC, Gerding DN, Johnson S, Bakken JS, Carroll KC, Coffin SE, Dubberke ER, Garey KW, Gould CV, Kelly C, Loo V, Shaklee Sammons J, Sandora TJ, Wilcox MH. Clinical Practice Guidelines for Clostridium difficile Infection in Adults and Children: 2017 Update by the Infectious Diseases Society of America (IDSA) and Society for Healthcare Epidemiology of America (SHEA). Clin Infect Dis. 2018 Mar 19;66(7):e1-e48. doi: 10.1093/cid/cix1085. PMID: 29462280; PMCID: PMC6018983. Alonzo F 3rd, Torres VJ. The bicomponent pore-forming leucocidins of Staphylococcus aureus. Microbiol Mol Biol Rev. 2014 Jun;78(2):199-230. doi: 10.1128/MMBR.00055-13. PMID: 24847020; PMCID: PMC4054254. eadstrom PD, Surawicz CM. Chronic diarrhea. Clin Gastroenterol Hepatol. 2005 Aug;3(8):734-7. doi: 10.1016/s1542-3565(05)00298-3. PMID: 16234000. emeth V, Pfleghaar N. Diarrhea. 2021 Nov 29. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2022 Jan–. PMID: 28846339. randa-Michel J, Giannella RA. Acute diarrhea: a practical review. Am J Med. 1999 Jun;106(6):670-6. doi: 10.1016/s0002-9343(99)00128-x. PMID: 10378626; PMCID: PMC7124219. emming AR, Bentlage AEH, de Taeye SW, Bosman GP, Lissenberg-Thunnissen SN, Derksen NIL, Brasser G, Mok JY, van Esch WJE, Howie HL, Zimring JC, Vidarsson G. Cross-reactivity of mouse IgG subclasses to human Fc gamma receptors: Antibody deglycosylation only eliminates IgG2b binding. Mol Immunol. 2020 Nov;127:79-86. doi: 10.1016/j.molimm.2020.08.015. Epub 2020 Sep 15. PMID: 32947169. cCool D, Birshtein BK, Painter RH. Structural requirements of immunoglobulin G for binding to the Fc gamma receptors of the human tumor cell lines U937, HL-60, ML-1, and K562. J Immunol. 1985 Sep;135(3):1975-80. PMID: 3860563. ntonsson A, Johansson PJH. Binding of human and animal immunoglobulins to the IgG Fc receptor induced by human cytomegalovirus. J Gen Virol. 2001 May;82(Pt 5):1137-1145. doi: 10.1099/0022-1317-82-5-1137. PMID: 11297688. robach MJ, Dekkers OM, Wilcox MH, Kuijper EJ. European Society of Clinical Microbiology and Infectious Diseases (ESCMID): data review and recommendations for diagnosing Clostridium difficile-infection (CDI). Clin Microbiol Infect. 2009 Dec;15(12):1053-66. doi: 10.1111/j.1469-0691.2009.03098.x. PMID: 19929972. windells J, Brenwald N, Reading N, Oppenheim B. Evaluation of diagnostic tests for Clostridium difficile infection. J Clin Microbiol. 2010 Feb;48(2):606-8. doi: 10.1128/JCM.01579-09. Epub 2009 Dec 23. PMID: 20032256; PMCID: PMC2815642. artin JS, Monaghan TM, Wilcox MH. Clostridium difficile infection: epidemiology, diagnosis and understanding transmission. Nat Rev Gastroenterol Hepatol. 2016 Apr;13(4):206-16. doi: 10.1038/nrgastro.2016.25. Epub 2016 Mar 9. PMID: 26956066. oldenberg SD, Cliff PR, French GL. Laboratory diagnosis of Clostridium difficile infection. J Clin Microbiol. 2010 Aug;48(8):3048-9. doi: 10.1128/JCM.00223-10. PMID: 20668301; PMCID: PMC2916563. olage CR, Gyorke CE, Kennedy MA, Leslie JL, Chin DL, Wang S, Nguyen HH, Huang B, Tang YW, Lee LW, Kim K, Taylor S, Romano PS, Panacek EA, Goodell PB, Solnick JV, Cohen SH. Overdiagnosis of Clostridium difficile Infection in the Molecular Test Era. JAMA Intern Med. 2015 Nov;175(11):1792-801. doi: 10.1001/jamainternmed.2015.4114. PMID: 26348734; PMCID: PMC4948649. uerrero DM, Chou C, Jury LA, Nerandzic MM, Cadnum JC, Donskey CJ. Clinical and infection control implications of Clostridium difficile infection with negative enzyme immunoassay for toxin. Clin Infect Dis. 2011 Aug 1;53(3):287-90. doi: 10.1093/cid/cir361. PMID: 21765078. eng Z, Ling L, Stratton CW, Li C, Polage CR, Wu B, Tang YW. Advances in the diagnosis and treatment of Clostridium difficile infections. Emerg Microbes Infect. 2018 Feb 7;7(1):15. doi: 10.1038/s41426-017-0019-4. PMID: 29434201; PMCID: PMC5837143. arson AM, Fung AM, Fang FC. Evaluation of tcdB real-time PCR in a three-step diagnostic algorithm for detection of toxigenic Clostridium difficile. J Clin Microbiol. 2010 Jan;48(1):124-30. doi: 10.1128/JCM.00734-09. Epub 2009 Nov 18. PMID: 19923482; PMCID: PMC2812255. harp SE, Ruden LO, Pohl JC, Hatcher PA, Jayne LM, Ivie WM. Evaluation of the C.Diff Quik Chek Complete Assay, a new glutamate dehydrogenase and A/B toxin combination lateral flow assay for use in rapid, simple diagnosis of clostridium difficile disease. J Clin Microbiol. 2010 Jun;48(6):2082-6. doi: 10.1128/JCM.00129-10. Epub 2010 Apr 7. PMID: 20375230; PMCID: PMC2884466. a href="https://biotical.es/wp-content/uploads/2021/03/IFU_RTB25GD_Cdifficile_GDH.pdf">https://biotical.es/wp-content/uploads/2021/03/IFU_RTB25GD_Cdifficile_GDH.pdf a href="https://biotical.es/wp-content/uploads/2021/03/IFU_RTB25CD_Cdifficile_AB.pdf">https://biotical.es/wp-content/uploads/2021/03/IFU_RTB25CD_Cdifficile_AB.pdf a href="https://www.cepheid.com/content/dam/www-cepheid-com/documents/package-insert-files/300-9680-Xpert-C.%20diff-Epi%20US-IVD%20PI%20Rev%20J.pdf">https://www.cepheid.com/content/dam/www-cepheid-com/documents/package-insert-files/300-9680-Xpert-C.%20diff-Epi%20US-IVD%20PI%20Rev%20J.pdf ionne LL, Raymond F, Corbeil J, Longtin J, Gervais P, Longtin Y. Correlation between Clostridium difficile bacterial load, commercial real-time PCR cycle thresholds, and results of diagnostic tests based on enzyme immunoassay and cell culture cytotoxicity assay. J Clin Microbiol. 2013 Nov;51(11):3624-30. doi: 10.1128/JCM.01444-13. Epub 2013 Aug 21. PMID: 23966497; PMCID: PMC3889743. Table Table 1 is available in the Supplementary Files section Additional Declarations Yes there is potential Competing Interest. The procedure "Processing of samples with leukocidins and MRSA" is pending a patent application in the name of the author Juan Luis Recio-López (P202330734). Supplementary Files TABLES.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-3327753","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":231797077,"identity":"fa239841-52c6-4e10-af39-88069a951fbc","order_by":0,"name":"Juan Luis Recio","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAxElEQVRIiWNgGAWjYBACxnYGxgOJ/2wSwLyEAmK0NDMwHHjAlpbAwAbSYkCMNcwMDAcfsB2GaGEgRgtzM++DAwk85/P45bsTPzwwYJDnFztAyGHsBgcSJG4XS7bxbpYAOsxw5uwEQlrYGA4kGNxO3HCMdwNIC5BNlJaEcyAtm3+QoOXAAZCWbSTYktiQnDizLXebRYKBBGG/GLa3MT782WCX2M98dvPNHxU28vzShLQ0oPIl8CsHAXnCSkbBKBgFo2DEAwAy4ESQrAVN1QAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0002-7418-7926","institution":"SAS","correspondingAuthor":true,"prefix":"","firstName":"Juan","middleName":"Luis","lastName":"Recio","suffix":""}],"badges":[],"createdAt":"2023-09-05 12:55:09","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-3327753/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3327753/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":42948663,"identity":"a69f1318-5992-45bd-85e3-bba3804c7d0a","added_by":"auto","created_at":"2023-09-11 15:21:12","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":67359,"visible":true,"origin":"","legend":"\u003cp\u003eModified 3-steps algorithm.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-3327753/v1/7ba6862433d61f1b91620bd7.png"},{"id":42948664,"identity":"d2461d04-f4c7-48b4-a570-d5479dcff285","added_by":"auto","created_at":"2023-09-11 15:21:12","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":65199,"visible":true,"origin":"","legend":"\u003cp\u003eModified 2-steps algorithm.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-3327753/v1/d2550fb2e26816626f465152.png"},{"id":43074821,"identity":"a4cb09be-7284-486f-bb27-d6dbdb9425a2","added_by":"auto","created_at":"2023-09-13 14:38:58","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":271897,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3327753/v1/8ea6464b-c3e6-443a-ace9-142642164d80.pdf"},{"id":42949313,"identity":"fa5309cd-d542-4eb2-a60c-90e8c7feb963","added_by":"auto","created_at":"2023-09-11 15:29:12","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":31001,"visible":true,"origin":"","legend":"","description":"","filename":"TABLES.docx","url":"https://assets-eu.researchsquare.com/files/rs-3327753/v1/f799395ae389d2437612b1dd.docx"}],"financialInterests":"\u003cb\u003eYes\u003c/b\u003e there is potential Competing Interest.\nThe procedure \"Processing of samples with leukocidins and MRSA\" is pending a patent application in the name of the author Juan Luis Recio-López (P202330734).","formattedTitle":"Diagnosis of Clostridium difficile: a wrong algorithm and an improvement in immunoassays","fulltext":[{"header":"TEXT","content":"\u003cp\u003eSince the appearance of diagnostic algorithms for CDI, these have been modified and adjusted according to scientific evidence and diagnostic needs until reaching those currently used and recognised \u003csup\u003e(1)\u003c/sup\u003e, the main ones being the 2-step and 3-step algorithms, the latter being considered non-inferior, faster and less costly. When clinical evidence of CDI is found, samples are sent to the laboratory and, if they are adequate, they are analysed to make a diagnosis. In our department, a 3-step diagnostic algorithm was used:\u003c/p\u003e\n\u003cp\u003e-First step: GDH IC (negative is reported, positive at the second step).\u003c/p\u003e\n\u003cp\u003e-Second step: Toxin A and/or B IC (GDH+ and Toxin+ reported positive; GDH+ and Toxin- at the third step).\u003c/p\u003e\n\u003cp\u003e-Step 3: Molecular test by polymerase chain reaction (PCR) for toxigenic \u003cem\u003eC.difficile\u003c/em\u003e (Genexpert C.difficile/Epi, CEPHEID) with result reported.\u003c/p\u003e\n\u003cp\u003eThe 2-step algorithm does not use Toxin IC, it uses GDH IC and if positive PCR. The immunoassays used were biotical C.difficile GDH card and biotical C.difficile Toxins A+B card, both from Biotical Health.\u003c/p\u003e\n\u003cp\u003eIn May 2022, in the microbiology laboratory of Hospital San Cecilio, we observed three doubtful positive cases in a few days. After reviewing the clinical manifestations of the patients and the results of the cultures, we suspected a diagnostic error: \u003cem\u003eSalmonella enterica\u003c/em\u003e was isolated and identified in one case and \u003cem\u003eCampylobacter\u003c/em\u003e spp. in the other two. Although cases of coinfection are known, three cases in a week is very unusual for a microbiology service that covers 400,000 people, with more than 200 cases investigated per month and a prevalence of 9%. PCR for toxigenic \u003cem\u003eC.difficile\u003c/em\u003e was performed on the three doubtful samples and the results were negative.\u003c/p\u003e\n\u003cp\u003eA retrospective analysis of the last month found three other confirmed cases of misdiagnosis: preserved samples were PCR negative for toxigenic \u003cem\u003eC.difficile\u003c/em\u003e. Six other suspected cases were classified as \"probable\" FP because no specimen was recovered: they had other viral or bacterial infections that could explain the clinical manifestations, or responded well to treatments not suitable for CDI.\u003c/p\u003e\n\u003cp\u003eThese 12 cases had in common that they did not fit the classic epidemiological criteria for CDI: hospitalisation, surgery and/or previous antibiotic treatment. Our department established special surveillance for cases with these characteristics, checking the positive cases with PCR.\u003c/p\u003e\n\u003cp\u003eBetween June and September this special surveillance detected 10 FP cases allowing them to be correctly reported and one probable FP, the sample was lost. Upon review of the 10 FP, one of these did not meet the epidemiological criteria: it had recent previous hospitalisation and antibiotic treatment.\u003c/p\u003e\n\u003cp\u003eAt the same time, these clinical samples were frozen to find out where the error was, after ruling out IC batch problems. In October, the commercial company repeated CI and PCR by a different method (Allplex GI Bacterial I Assay, SEEGENE) on 10 of these samples. The PCR results were the same as those obtained previously by Genexpert, negative for CDI, and coincided with the cultures: Allplex GI also detects other pathogens such as \u003cem\u003eSalmonella, Yersinia, Shigella\u003c/em\u003e and \u003cem\u003eCampylobacter\u003c/em\u003e. Not so the result of the GDH and Toxin IC: only 2 of the 10 samples were still GDH positive, and of these only one was also Toxin positive.\u003c/p\u003e\n\u003cp\u003eFaced with these errors of the 3-step algorithm, the management of the service allowed a change in the diagnostic method from October, although IC Toxin was not stopped, the diagnosis was a 2-step algorithm: GDH, negative is reported, and positive is PCR to decide the diagnosis (regardless of the result of the IC Toxin).\u003c/p\u003e\n\u003cp\u003eWhy were samples previously positive and now not positive? Freezing of samples and duration of freezing. The samples that remained positive for GDH were the most recent ones: they had been kept for less time frozen. The interference causing FP was sensitive to freezing. Another clue was provided by a review of the 23 cases detected (16 FP and 7 \"probable\" FP): 11 cases with positive cultures for enteropathogenic bacteria (6 \u003cem\u003eCampylobacter\u003c/em\u003e, 4 \u003cem\u003eSalmonella\u003c/em\u003e and 1 \u003cem\u003eShigella\u003c/em\u003e); and among the remaining cases 3 were diagnosed with inflammatory bowel disease (IBD), 2 were cross-reactive for other IC (\u003cem\u003eRotavirus, Adenovirus\u003c/em\u003e and \u003cem\u003eNorovirus\u003c/em\u003e) and another case in which a parasite study was requested showed abundant leucocytes under the optical microscope (OM). Most of the samples were inflammatory diarrhoea.\u003c/p\u003e\n\u003cp\u003eWhat if all these FP were caused by the presence of leukocytes in the stool? Proving this was complex because leukocytes are not observable by OM in all samples and their presence in suspected cases of CDI is normal. To test this hypothesis we needed sample processing to remove leukocytes without altering GDH and Toxin: after processing, repeat IC in true positives (TP) would remain positive and in FP would be negative. Although freezing had guided us, it is not a good experimental model (slow, laborious and unreliable). Heating was not considered because of denaturation and other problems related to sample handling. The use of monoclonal antibodies, biological agents, ionising radiation, acids, bases, salts, etc. were ruled out for various reasons.\u003c/p\u003e\n\u003cp\u003eIn a microbiology laboratory there is a simple and safe way to lyse leukocytes: Methicillin Resistant \u003cem\u003eStaphylococcus aureus\u003c/em\u003e (MRSA). All \u003cem\u003eS.aureus\u003c/em\u003e produce at least 3 types of leukocidins (HlgAB, HlgCB, and LukAB/HG), while the most virulent isolates produce up to 5 types (the above plus Panton-Valentine leukocidin [PVL], and LukED) \u003csup\u003e(2)\u003c/sup\u003e. These leukocidins are bicomponent water-soluble proteins excreted by the bacteria, with the ability to recognise leukocytes, insert into their membranes and generate pores, which ultimately lyse them. LukAB/HG is slightly different from the others, existing as a stable heterodimer in solution and establishing a reservoir of toxin on the surface of the bacterium. In addition, most of the immunoassays used have been tested with different bacteria, including \u003cem\u003eS.aureus\u003c/em\u003e, and cross-reactivity has been ruled out.\u003c/p\u003e\n\u003cp\u003eSince October, all GDH IC positive samples were subjected to: Toxin IC, PCR for toxigenic \u003cem\u003eC.difficile\u003c/em\u003e and GDH IC after processing the sample with a suspension of leukocidins and MRSA (see Methods). With this processed sample, Toxin IC was repeated on an ad hoc basis due to lack of resources.\u003c/p\u003e"},{"header":"RESULTS","content":"\u003cp\u003eA total of 660 samples of suspected CDI cases were processed between October and December 2022, a total of 77 samples were GDH positive: 38 Toxin positive and 39 negative. Only 3 of these 77 samples (3.9%) were grossly purulent and/or bloody.\u003c/p\u003e\n\u003cp\u003e- Of the 38 GDH+/Tox+ samples: 16 were positive by PCR and 22 were negative (among the latter 3 were very purulent and/or bloody).\u003c/p\u003e\n\u003cp\u003e- Of the 39 GDH+/Tox- samples: 24 were PCR positive and 15 negative.\u003c/p\u003e\n\u003cp\u003eA total of 40 cases were reported as positive. In contrast, the 3-step algorithm would have resulted in 62 positive cases (38 GDH+/Tox+ plus 24 GDH+/Tox-/PCR+), with an error of 35% (22 FP).\u003c/p\u003e\n\u003cp\u003eRegarding the GDH performed after sample processing with MRSA and leukocidins (MRSA-GDH):\u003c/p\u003e\n\u003cp\u003e-Of the 38 samples +GDH/+Tox: 100% concordance with PCR.\u003c/p\u003e\n\u003cp\u003e-Of the 39 +GDH/+Tox samples: 92.3% concordance. The 24 PCR-positive samples were also positive for MRSA-GDH; but among the 15 PCR-negative samples there were 3 positive for MRSA-GDH.\u003c/p\u003e\n\u003cp\u003eModified algorithms (figure 1 and 2) with this sample processing obtain the same results as the 2-step algorithm, and with a reduction in total PCR needed between 44-64% (table 1). We have included as a paradox the diagnosis of CDI in a single step, using only MRSA-GDH determination: only 3 FP compared to the 22 FP of the 3-step diagnosis, and without the need to use a single PCR.\u003c/p\u003e\n\u003cp\u003eWith 40 TP among 660 samples, we have a prevalence of 6%, down from 7.5-9% in previous months. Among the 22 GDH+/Tox+ FP, a total of 8 cases were patients with hospitalisation, surgery and/or recent previous antibiotic treatment.\u003c/p\u003e\n\u003cp\u003eEstimating 100% sensitivity in the GDH IC and the 2-step algorithm as the gold standard, although it is not correct, we can calculate theoretical data such as the false positive ratio (FPR) that allow us to discuss the results: \u003c/p\u003e\n\u003cp\u003e- GDH IC specificity: 583/(583+37) = 0.9403\u003c/p\u003e\n\u003cp\u003e- 3-step algorithm specificity: 598/(598+22) = 0,9614\u003c/p\u003e\n\u003cp\u003e- MRSA-GDH IC specificity: 617/(617+3) = 0.9951\u003c/p\u003e\n\u003cp\u003e- GDH IC FPR = 0,0597\u003c/p\u003e\n\u003cp\u003e- 3-step algorithm FPR = 0,0386\u003c/p\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eAmong the studies of ICD diagnostic methods since 2009, we have not been able to find any that compare the 2-step and 3-step algorithms exactly as we know them today. There are some similar studies \u003csup\u003e(10,16)\u003c/sup\u003e, which are not comparable for different reasons: low sample size or testing by cell cytoxicity neutralization assay (CCNA) on GDH+/Toxin+ results instead of PCR. Although years ago the CCNA was considered the gold standard, it is a subjective method, not very reproducible and gives results after 24-48 hours \u003csup\u003e(15,17)\u003c/sup\u003e. These studies are used in other subsequent studies in which 2-step algorithms are evaluated to argue for the use of 3-step algorithms, which would allow similar results to be obtained with a lower PCR cost \u003csup\u003e(11, 12, 13, 14)\u003c/sup\u003e, although without experimentation. Thus, we found a meta-analysis with a theoretical construction of a 3-step algorithm on 10,000 samples \u003csup\u003e(9)\u003c/sup\u003e, in which method errors are considered as independent and with a prevalence of 5%, with specificities of 0.87 for the GDH test and 0.96 for the Toxin test: 432 VP and 49 FP (similar total of FP in our study, but with 15 times less sample size). We observed that 3-step algorithms, where the first step was PCR \u003csup\u003e(21)\u003c/sup\u003e did not present this problem: the total error frequency is reduced as the product of the independent error frequencies of each method (PCR vs. IC).\u003c/p\u003e\n\u003cp\u003eWith our results we can think that we use a Toxin IC with very low specificity if we consider as in other studies \u003csup\u003e(9)\u003c/sup\u003e that the errors of each IC are independent of each other, the total error frequency is given by the multiplication of the individual error frequencies for each method. For the FP, this frequency is the false positive ratio (FPR) that we can calculate with the specificity. As all the FP in the 3-step algorithm are GDH+/Toxin+, we can infer that the FPR of the 3-step algorithm is equal to the product of the FPR of GDH IC by the FPR of Toxin IC, so we can calculate it and infer from it the \"theoretical\" specificity of the Toxin IC.\u003c/p\u003e\n\u003cp\u003eFPR Toxin IC x FPR GDH IC = FPR3-step algorithm; FPR Toxin IC = FPR 3-step algorithm/ FPR GDH IC\u003c/p\u003e\n\u003cp\u003eFPR Toxin IC = 0.0386/0.0597 = 0.64 → \"Theoretical\" specificity Toxin IC = 0.36\u003c/p\u003e\n\u003cp\u003eAlthough some Toxin IC may be less specific, this one in particular would beat all known IC in the ranking of worst specificity. We know that this is not the problem: first, because experimentally this IC (biotical C.difficile Toxins A+B card) has never had such low specificity values. And second, the errors of these two methods cannot be considered as independent because experimentally we have shown that they are related, even if different antigens are detected. They are very similar methods and share an intrinsic structural error in their design: interference by cross-reactions of human leukocytes with mouse antibodies, which is also the main cause of FP. For this reason, if this error occurs in a GDH IC for a sample, it is much more likely that the same error occurs in the same sample for the Toxin IC: the Toxin IC of the FP were also negative after processing for leukocidins and MRSA, although due to lack of resources this was only performed on some samples.\u003c/p\u003e\n\u003cp\u003eEstimating a Toxin IC specificity range of 0.94-0.84 (FPR = 0.06-0.16), if the method errors were considered independent, the expected FPR of the 3-step algorithm would have been between 0.0035-0.0095; a value 11 to 4 times lower than the real value obtained. If for a sample a method-dependent error occurs in an A IC, the probability of the same method-dependent error occurring in a B IC, instead of decreasing, as they are related and dependent errors, increases, and in the case of GDH and Toxin in the order of 11 to 4 times over the expected frequency.\u003c/p\u003e\n\u003cp\u003eWe wonder how it is possible that such an unreliable diagnostic method as the 3-step algorithm (one out of three positives is wrong) has been used for so many years. The answer is complex, and is not only due to considering the GDH and Toxin immunoassay FP as independent and unrelated errors. It is an algorithm with high sensitivity, few false negatives (FN). A large number of samples are processed, with more than 90% being negative, so that any specific findings can be minimised among the total number of samples processed. Moreover, the general consensus on this point has always been diametrically opposed: the suspected error in the use of immunoassays for GDH and Toxin was underdiagnosis, not overdiagnosis \u003csup\u003e(15)\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eIf detecting diagnostic errors in 3 steps is difficult, it is also difficult to associate them with their cause. The use of leukocidin and MRSA in this study makes it possible to see a relationship between the presence of leukocytes in faeces (inflammatory diarrhoea) and FP, but it is impossible without this processing: one of the causes of inflammatory diarrhoea is CDI \u003csup\u003e(3,4,5)\u003c/sup\u003e. Not all inflammatory diarrhoeas are FP and not all FP are inflammatory diarrhoeas.\u003c/p\u003e\n\u003cp\u003eThis interspecies cross-reaction between human leukocytes and mouse immunoglobulin (IG) has been reported before \u003csup\u003e(6)\u003c/sup\u003e, and does not occur between any leukocyte and any mouse IG, it depends on the individual, the leukocyte and the type of IG. We can observe that leukocidin and MRSA processing also negativises the GDH of Tox- samples (12 out of 39 GDH+/Tox-): in these samples, the leukocytes cross-react with the IG of the GDH IC, but not with the IG of the Toxin IC. We could not observe the effect of leucodidin and MRSA processing in all Toxin+ IC due to lack of resources, but in the FP where we did, Toxin negativised.\u003c/p\u003e\n\u003cp\u003eThe result of this difficulty in targeting the problem is in the partial solution we took between June and September: a 3-step algorithm checking positive results with PCR based on clinical/epidemiological data surveillance. Ten FP were detected, although only 9 would have been detected if one of these cases had not been assigned by mistake to the group without hospitalisation, without surgery and without previous antibiotic treatment. The results obtained from October to December reminded us of an old laboratory axiom \"where you don't look, you don't find\": of the 22 FP, only 14 belonged to this special surveillance group while 8 were cases with hospitalisation, surgery and/or recent prior antibiotic treatment.\u003c/p\u003e\n\u003cp\u003eWhen we asked ourselves whether the problem was due to a specific type of immunoassay (GDH IC in use) or to the structural design of all the immunoassays, the punctual tests with the Toxin IC indicated that it was a structural design error. We found that this problem was also present in other immunoassays with samples giving FP: in faeces for \u003cem\u003eRotavirus\u003c/em\u003e, \u003cem\u003eAdenovirus\u003c/em\u003e, \u003cem\u003eAstrovirus\u003c/em\u003e, \u003cem\u003eNorovirus\u003c/em\u003e type I, \u003cem\u003eNorovirus\u003c/em\u003e type II and \u003cem\u003eGiardia lamblia\u003c/em\u003e ICs; and in respiratory samples for Respiratory Syncytial Virus, Influenza A, Influenza B and SARS-Cov-2 ICs. In all of them, the processing of the sample with leukocidins and MRSA negatived the FP result of the immunoassay, coinciding with the result obtained by molecular biology techniques. Although we could also suspect it for \u003cem\u003eHelicobacter pylori\u003c/em\u003e IC in stool when negative with leukocidins and MRSA in some samples, we lacked a confirmatory method.\u003c/p\u003e\n\u003cp\u003eSince 2010, the potential FP of the 3-step algorithm has been systematically underestimated. The main problem was the sequential use of two immunoassay methods, similar in performance and also in their intrinsic errors. As these errors are the most important, they have been considered as independent errors, when they are not. By considering the total error frequency as the product of the individual error frequencies of each method, the FP ratio (frequency) of the 3-step algorithm was underestimated: instead of decreasing in proportion to the specificity of both methods and having a frequency of 0.005, the errors committed decrease in a much smaller proportion and are 8 times higher than expected.\u003c/p\u003e\n\u003cp\u003eThe use of animals other than mice to obtain IG for immunoassays does not appear to avoid these interspecies cross-reactions, IG from other animals (rabbits, pigs, hamsters, cats, horses) are able to bind human leukocytes \u003csup\u003e(7,8)\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eProcessing samples with leukocidins and MRSA allows for a 92% reduction in GDH FP (from 37 to 3) enabling accurate diagnosis with less use of PCR (44-64% reduction, depending on the algorithm used, fig.1 and 2), and we have indications that its application to other immunoassays will also increase the performance and reliability of these assays. This increase in specificity of immunoassays using leukocidin and MRSA sample processing is an overall improvement for diagnosis, not only qualitatively in terms of accuracy and speed, but also in terms of ease of use and resource savings. \u003c/p\u003e\n\u003cp\u003eAs to how feasible it is to undertake a modified algorithm using leukocidin and MRSA sample processing in a microbiology laboratory in the short to medium term: it is unlikely. Although the use of MRSA and leukocidins is straightforward and the use of liquid cultures is less labour intensive than PCR, this processing is not yet commercially available in a kit or alongside immunoassays. \u003c/p\u003e\n\u003cp\u003eThe main biases we could not avoid are the use of the 2-step algorithm as the gold standard and not testing leukocidin and MRSA processing in all Toxin-positive IC. Lack of resources prevented us from ascertaining the cause of the 3 FP for GDH IC after leukocidin and MRSA processing.\u003c/p\u003e"},{"header":"CONCLUSIONS","content":"\u003cp\u003eWe strongly discourage the use of CDI 3-step diagnostic algorithm. We recommend the use of a 2-step diagnostic algorithm, until a standard leukocidin and MRSA sample processing system is available that allows the use of modified algorithms, which is expected to increase the speed and accuracy of diagnosis at lower cost.\u003c/p\u003e \u003cp\u003eThe use of this sample processing in immunoassays other than GDH detection will allow faster, more accurate and lower cost diagnostics than current ones, although further research and validation is needed before they can be used in diagnostics.\u003c/p\u003e"},{"header":"METHODS","content":"\u003cp\u003eThe IC used were biotical C.difficile GDH card and biotical C.difficile Toxins A+B card, both from Biotical Health, used according to the instructions provided \u003csup\u003e(18,19)\u003c/sup\u003e. The PCR method used was Genexpert C.difficile/Epi (Cepheid), used according to the instructions provided \u003csup\u003e(20)\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eProcessing of samples with leukocidins and MRSA:\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003eA pathogenic clinical strain of Methicillin Resistant \u003cem\u003eStaphylococcus aureus\u003c/em\u003e (MRSA) is used. The culture is isolated on blood agar (BA) or blood chocolate agar (BCA), and subcultured on BA, BCA, BHI medium or any other medium that allows MRSA to produce leukocidins, at 37°C for 48h or more (depending on inoculum and total volume, 48 hours to prepare 10 ml in liquid BHI, with 12 more hours of culture more than one litre can be prepared in liquid BHI). As leukocidins are produced only in the stationary growth phase, their concentration will increase over time. Afterwards, they can be stored at room temperature: solid media for 3-4 weeks, until the drying of the colonies makes their handling difficult, and in liquid media for more than 6 months without losing their properties (tested by the author), although their stability is probably even higher.\u003c/p\u003e\n\u003cp\u003eThere are two options for the treatment of each sample: \u003c/p\u003e\n\u003cp\u003e(a) From culture on solid medium: resuspend 5-10 colonies of the culture on blood agar in 0,5 ml of the buffer used for the immunoassay (failing this, PBS), to a McFarland \u0026gt; 4.\u003c/p\u003e\n\u003cp\u003e(b) From liquid culture: resuspend 0,5 ml of the culture on BHI (bacteria tend to settle to the bottom) in 0,5 ml of the immunoassay buffer (failing PBS). \u003c/p\u003e\n\u003cp\u003eThe volumes and amounts are standard, but should be adjusted depending on the type of sample and immunoassay. In both ways we obtain a buffered solution with MRSA and leukocidins in suspension. The procedure does not differ at this point: \u003c/p\u003e\n\u003cp\u003e-The sample is added, as specified by the procedure of the immunoassay technique to be used: the amounts are usually between 150 to 50 microlitres, to avoid the dilution effect they can be doubled.\u003c/p\u003e\n\u003cp\u003e-Mix and vortex intensively for 30 seconds. \u003c/p\u003e\n\u003cp\u003e-Incubate at 42°C for one hour, or 37°C for one and a half hours. In totally purulent or bloody samples the incubation period should not be less than 3 hours (incubations of more than 6 hours do not negative true positives). \u003c/p\u003e\n\u003cp\u003eAfter these steps, the immunoassay protocol for detection or quantification of the test analyte is resumed.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eCONFLICTS OF INTEREST:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAuthor declare no conflicts of interest. The procedure \u0026quot;Processing of samples with leukocidins and MRSA\u0026quot; is pending a patent application in the name of the author Juan Luis Recio-L\u0026oacute;pez.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eMcDonald LC, Gerding DN, Johnson S, Bakken JS, Carroll KC, Coffin SE, Dubberke ER, Garey KW, Gould CV, Kelly C, Loo V, Shaklee Sammons J, Sandora TJ, Wilcox MH. Clinical Practice Guidelines for Clostridium difficile Infection in Adults and Children: 2017 Update by the Infectious Diseases Society of America (IDSA) and Society for Healthcare Epidemiology of America (SHEA). Clin Infect Dis. 2018 Mar 19;66(7):e1-e48. doi: 10.1093/cid/cix1085. PMID: 29462280; PMCID: PMC6018983.\u003c/li\u003e\n\u003cli\u003eAlonzo F 3rd, Torres VJ. The bicomponent pore-forming leucocidins of Staphylococcus aureus. Microbiol Mol Biol Rev. 2014 Jun;78(2):199-230. doi: 10.1128/MMBR.00055-13. PMID: 24847020; PMCID: PMC4054254.\u003c/li\u003e\n\u003cli\u003eeadstrom PD, Surawicz CM. Chronic diarrhea. Clin Gastroenterol Hepatol. 2005 Aug;3(8):734-7. doi: 10.1016/s1542-3565(05)00298-3. PMID: 16234000.\u003c/li\u003e\n\u003cli\u003eemeth V, Pfleghaar N. Diarrhea. 2021 Nov 29. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2022 Jan\u0026ndash;. PMID: 28846339.\u003c/li\u003e\n\u003cli\u003eranda-Michel J, Giannella RA. Acute diarrhea: a practical review. Am J Med. 1999 Jun;106(6):670-6. doi: 10.1016/s0002-9343(99)00128-x. PMID: 10378626; PMCID: PMC7124219.\u003c/li\u003e\n\u003cli\u003eemming AR, Bentlage AEH, de Taeye SW, Bosman GP, Lissenberg-Thunnissen SN, Derksen NIL, Brasser G, Mok JY, van Esch WJE, Howie HL, Zimring JC, Vidarsson G. Cross-reactivity of mouse IgG subclasses to human Fc gamma receptors: Antibody deglycosylation only eliminates IgG2b binding. Mol Immunol. 2020 Nov;127:79-86. doi: 10.1016/j.molimm.2020.08.015. Epub 2020 Sep 15. PMID: 32947169.\u003c/li\u003e\n\u003cli\u003ecCool D, Birshtein BK, Painter RH. Structural requirements of immunoglobulin G for binding to the Fc gamma receptors of the human tumor cell lines U937, HL-60, ML-1, and K562. J Immunol. 1985 Sep;135(3):1975-80. PMID: 3860563.\u003c/li\u003e\n\u003cli\u003entonsson A, Johansson PJH. Binding of human and animal immunoglobulins to the IgG Fc receptor induced by human cytomegalovirus. J Gen Virol. 2001 May;82(Pt 5):1137-1145. doi: 10.1099/0022-1317-82-5-1137. PMID: 11297688.\u003c/li\u003e\n\u003cli\u003erobach MJ, Dekkers OM, Wilcox MH, Kuijper EJ. European Society of Clinical Microbiology and Infectious Diseases (ESCMID): data review and recommendations for diagnosing Clostridium difficile-infection (CDI). Clin Microbiol Infect. 2009 Dec;15(12):1053-66. doi: 10.1111/j.1469-0691.2009.03098.x. PMID: 19929972. \u003c/li\u003e\n\u003cli\u003ewindells J, Brenwald N, Reading N, Oppenheim B. Evaluation of diagnostic tests for Clostridium difficile infection. J Clin Microbiol. 2010 Feb;48(2):606-8. doi: 10.1128/JCM.01579-09. Epub 2009 Dec 23. PMID: 20032256; PMCID: PMC2815642. \u003c/li\u003e\n\u003cli\u003eartin JS, Monaghan TM, Wilcox MH. Clostridium difficile infection: epidemiology, diagnosis and understanding transmission. Nat Rev Gastroenterol Hepatol. 2016 Apr;13(4):206-16. doi: 10.1038/nrgastro.2016.25. Epub 2016 Mar 9. PMID: 26956066. \u003c/li\u003e\n\u003cli\u003eoldenberg SD, Cliff PR, French GL. Laboratory diagnosis of Clostridium difficile infection. J Clin Microbiol. 2010 Aug;48(8):3048-9. doi: 10.1128/JCM.00223-10. PMID: 20668301; PMCID: PMC2916563. \u003c/li\u003e\n\u003cli\u003eolage CR, Gyorke CE, Kennedy MA, Leslie JL, Chin DL, Wang S, Nguyen HH, Huang B, Tang YW, Lee LW, Kim K, Taylor S, Romano PS, Panacek EA, Goodell PB, Solnick JV, Cohen SH. Overdiagnosis of Clostridium difficile Infection in the Molecular Test Era. JAMA Intern Med. 2015 Nov;175(11):1792-801. doi: 10.1001/jamainternmed.2015.4114. PMID: 26348734; PMCID: PMC4948649. \u003c/li\u003e\n\u003cli\u003euerrero DM, Chou C, Jury LA, Nerandzic MM, Cadnum JC, Donskey CJ. Clinical and infection control implications of Clostridium difficile infection with negative enzyme immunoassay for toxin. Clin Infect Dis. 2011 Aug 1;53(3):287-90. doi: 10.1093/cid/cir361. PMID: 21765078. \u003c/li\u003e\n\u003cli\u003eeng Z, Ling L, Stratton CW, Li C, Polage CR, Wu B, Tang YW. Advances in the diagnosis and treatment of Clostridium difficile infections. Emerg Microbes Infect. 2018 Feb 7;7(1):15. doi: 10.1038/s41426-017-0019-4. PMID: 29434201; PMCID: PMC5837143. \u003c/li\u003e\n\u003cli\u003earson AM, Fung AM, Fang FC. Evaluation of tcdB real-time PCR in a three-step diagnostic algorithm for detection of toxigenic Clostridium difficile. J Clin Microbiol. 2010 Jan;48(1):124-30. doi: 10.1128/JCM.00734-09. Epub 2009 Nov 18. PMID: 19923482; PMCID: PMC2812255. \u003c/li\u003e\n\u003cli\u003eharp SE, Ruden LO, Pohl JC, Hatcher PA, Jayne LM, Ivie WM. Evaluation of the C.Diff Quik Chek Complete Assay, a new glutamate dehydrogenase and A/B toxin combination lateral flow assay for use in rapid, simple diagnosis of clostridium difficile disease. J Clin Microbiol. 2010 Jun;48(6):2082-6. doi: 10.1128/JCM.00129-10. Epub 2010 Apr 7. PMID: 20375230; PMCID: PMC2884466. \u003c/li\u003e\n\u003cli\u003ea href=\"https://biotical.es/wp-content/uploads/2021/03/IFU_RTB25GD_Cdifficile_GDH.pdf\"\u003ehttps://biotical.es/wp-content/uploads/2021/03/IFU_RTB25GD_Cdifficile_GDH.pdf\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003ea href=\"https://biotical.es/wp-content/uploads/2021/03/IFU_RTB25CD_Cdifficile_AB.pdf\"\u003ehttps://biotical.es/wp-content/uploads/2021/03/IFU_RTB25CD_Cdifficile_AB.pdf\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003ea href=\"https://www.cepheid.com/content/dam/www-cepheid-com/documents/package-insert-files/300-9680-Xpert-C.%20diff-Epi%20US-IVD%20PI%20Rev%20J.pdf\"\u003ehttps://www.cepheid.com/content/dam/www-cepheid-com/documents/package-insert-files/300-9680-Xpert-C.%20diff-Epi%20US-IVD%20PI%20Rev%20J.pdf\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003eionne LL, Raymond F, Corbeil J, Longtin J, Gervais P, Longtin Y. Correlation between Clostridium difficile bacterial load, commercial real-time PCR cycle thresholds, and results of diagnostic tests based on enzyme immunoassay and cell culture cytotoxicity assay. J Clin Microbiol. 2013 Nov;51(11):3624-30. doi: 10.1128/JCM.01444-13. Epub 2013 Aug 21. PMID: 23966497; PMCID: PMC3889743. \u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Table","content":"\u003cp\u003eTable 1 is available in the Supplementary Files section\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-3327753/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3327753/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eWe compared the 2-step and 3-step algorithms for diagnosis of \u003cem\u003eClostridium difficile\u003c/em\u003e infection (CDI) in clinical cases between October and December 2023. We observed in 3-step algorithm 22 false positives (FP) among the 62 positives (35%), compared to 40 positives without FP in 2-step algorithm. All FP are positive for glutamate dehydrogenase (GDH) and Toxin A and/or B immunochromatography (IC), and negative for toxigenic \u003cem\u003eC.difficile\u003c/em\u003e PCR. When processing the samples with a solution of leukocidins and Methicillin Resistant \u003cem\u003eStaphylococcus aureus\u003c/em\u003e (MRSA) the FP to GDH are negative, as well as FP from different immunoassays. The cause of FP is an interspecies cross-reaction between human leukocytes and mouse immunoglobulins, a method-dependent error (immunoassay) that increases the probability of FP in the 3-step algorithm. Modified algorithms and sample processing with MRSA and leukocidins allow equivalent results to the 2-step algorithm, with a reduction in the total number of molecular tests required (47-61%).\u003c/p\u003e","manuscriptTitle":"Diagnosis of Clostridium difficile: a wrong algorithm and an improvement in immunoassays","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2023-09-11 15:21:08","doi":"10.21203/rs.3.rs-3327753/v1","editorialEvents":[{"type":"communityComments","content":1}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"f3d521f8-277b-4def-942d-6b9928f2d384","owner":[],"postedDate":"September 11th, 2023","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":24506411,"name":"Biological sciences/Microbiology/Clinical microbiology"},{"id":24506412,"name":"Health sciences/Diseases/Gastrointestinal diseases/Gastroenteritis"}],"tags":[],"updatedAt":"2024-06-11T05:19:24+00:00","versionOfRecord":[],"versionCreatedAt":"2023-09-11 15:21:08","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-3327753","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3327753","identity":"rs-3327753","version":["v1"]},"buildId":"_2-kVJe1T_tPrBINL-cwx","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}