Clostridioides difficile thrives in an inflamed environment by generating toxins that support prolonged infection, according to a study from North Carolina State University.
The study, published in Nature Communications, showed how C. diff produces toxins that cause inflammation, eliminating competing bacteria and releasing peptides and amino acids that support the growth of C. diff.
“C. diff thrives when other microbes in the gut are absent – which is why it is more prevalent following antibiotic therapy,” corresponding author Casey Theriot, Ph.D., associate professor of infectious disease at North Carolina State University, said. “But when colonizing the gut, C. diff. also produces two large toxins, TcdA and TcdB, which cause inflammation. We wanted to know if these inflammation-causing toxins actually give C. diff a survival benefit – whether the pathogen can exploit an inflamed environment in order to thrive.”
Investigators examined two variants of C. diff in vitro and in an antibiotic-treated mouse model. The variants included a wild type C. diff that produces toxins and a genetically modified variant that does not. They found that the wild type C. diff, associated with toxin production, generated more inflammation and tissue damage than the mutant.
To read the article in its entirety, please click on the following link to be redirected:
Investigators also found changes in the expression of metabolic genes, with C. diff in the inflamed environment expressing more genes related to carbohydrate and amino acid metabolism that sustains growth.
“C. diff’s toxins damage the cells that line the gut,” Theriot said. “These cells contain collagen, which is made up of amino acids and peptides. When collagen is degraded by toxins, C. diff responds by turning on expression of genes that can use these amino acids for growth.”
Inflammation provided a second benefit to C. diff by creating an inhospitable environment for other bacteria that compete for nutrients. Bacteroidaceae were present in control groups that weren’t infected with toxin-producing C. diff, which was consistent with previous studies that found negative associations between C. diff and Bacteroidaceae.
“I always found it interesting that C. diff causes such intense inflammation,” first author Josh Fletcher, Ph.D., a former postdoctoral researcher at North Carolina State University, said. “Our research shows that this inflammation may contribute to the persistence of C. diff in the gut environment, prolonging infection.”
C. diff is the most significant cause of hospital-acquired diarrhea, causing more than 223,900 infections and 12,800 deaths in the US in 2017, according to a recent report.
The disease has two phases, a spore phase, and vegetative phase. Toxins are released during the vegetative phase, causing diarrhea and other symptoms. But the pathogen is often transmitted during the spore phase, during which it is hardy and isn’t susceptible to gastric acids and alcohol-based hand sanitizer, experts explained during a recent discussion of the disease.
Risks for infection include exposure to C. diff spores and antibiotic use. An investigational drug to prevent the disruption of the gut microbiota by antibiotics is among the most recent developments in the fight against a C diff. infection.
Vanderbilt University Medical Center scientists have identified a C. diff protein system that senses and captures heme (part of hemoglobin) to build a protective shield that fends off threats from our immune system and antibiotics. The findings, reported in the journal Cell Host & Microbe, reveal a unique mechanism for C. diff survival in the human gut and suggest novel strategies for weakening its defenses.
In a cruel twist, the bacterium Clostridioides difficile (C. diff) makes us bleed and then uses our blood to defend itself against us.
C. diff the most common cause of health care-associated infections (HAI’s) in the United States causes diarrhea and inflammation of the colon (colitis). Individuals taking antibiotics, which disturb the protective gut microbiota, have increased risk for acquiring a C. diff infection, and 20% of patients suffer recurrent C. diff infections despite treatment.
When C. diff colonizes the gut, it produces toxins that cause tissue damage and inflammation. Blood cells burst, releasing heme, the part of hemoglobin that binds iron and oxygen.
Eric Skaar, Ph.D., MPH, Ernest W. Goodpasture Professor of Pathology, Microbiology and Immunology, and colleagues have studied how bacteria respond to heme, which is both a source of the nutrient iron and a reactive, toxic compound.
“Organisms that experience large amounts of heme have to have ways to deal with heme toxicity,” said Skaar, director of the Vanderbilt Institute for Infection, Immunology and Inflammation (VI4). “We wanted to understand how C. diff deals with heme exposure.”
The investigators demonstrated that C. diff exposed to heme increases expression of a protein system that had not been previously studied. They named the system HsmRA (heme sensing membrane proteins R and A) and showed that HsmR senses heme and deploys HsmA to capture it. They also found that the HsmRA system is genetically conserved in many bacterial species.
The binding of heme in the bacterial membrane by HsmA serves a protective purpose first by simply reducing the concentration of free heme, Skaar explained. The researchers also discovered that HsmA uses heme binding to protect C. diff from oxidative stress, including that produced by neutrophils and macrophages from our immune system to kill bacteria.
“C. diff is using cofactors from our own cells as a shield to protect against our innate immune response,” Skaar said.
Oxidative stress also plays a role in antibiotic action.
“Antibiotics have different molecular targets—they may prevent cell wall synthesis; they may prevent protein translation—but the net result of that stress on the cell is often the massive accumulation of oxidative stress that many believe to be a major contributor to why antibiotics kill bacteria,” Skaar said.
The investigators studied whether the HsmRA system protected C. diff against antibiotics.
“We found a really impressive phenotype with vancomycin and metronidazole, two of the front-line antibiotics used to treat C. diff,” Skaar said. “C. diff that expresses HsmA, when HsmA is bound to heme, is much more resistant to vancomycin and metronidazole.”
They also showed that C. diff strains with inactivated HsmR or HsmA had reduced colonization in a mouse model of relapse C. diff infection.
Skaar said it has not been clear why C. diff produces toxins that cause so much tissue damage.
“It’s interesting to speculate that a benefit of toxin-related damage is that C. diff can capture liberated heme and use it as a shield to protect itself against various insults that cause oxidative stress—that would be immune cells, antibiotics and potentially other bacteria.”
The findings suggest that targeting the HsmA-heme shield might increase the sensitivity of C. diff to antibiotics such as vancomycin and metronidazole. It’s not clear that HsmA, a membrane protein, will be a druggable target, Skaar said.
It might be possible, however, to deprive C. diff of heme building blocks by reducing tissue damage or by administering proteins that bind heme, he said. The researchers will explore whether they can increase the sensitivity of C. diff to antibiotics by co-administering a heme-binding protein during infection in an animal model.
“We’re excited about this as a potentially powerful strategy for treating C. diff,” Skaar said.
In other studies, the researchers will explore if the HsmRA system that is genetically conserved in many different organisms has the same functional role to protect against reactive oxygen species. They are also trying to understand the exact mechanism that HsmA-heme uses to detoxify oxidative stress.
To read article in its entirety, please click on the following link to be redirected:
Koray K. Demir, Data curation, Investigation, Writing – original draft, Writing – review & editing,1Matthew P. Cheng, Supervision, Writing – review & editing,2 and Todd C. Lee, Conceptualization, Formal analysis, Methodology, Software, Supervision, Writing – review & editing2,3,*
Abhishek Deshpande, Editor
Introduction and objective
Diagnostic testing for Clostridioides difficile infection (CDI) by nucleic acid amplification test (NAAT) cannot distinguish between colonization and infection.
A positive NAAT may therefore represent a false positive for infection, since diarrhea due to various aetiologies may occur in hospitalized patients.
Our objective was to help answer the question: “does this medical inpatient with diarrhea have CDI?”
We conducted a retrospective cohort study (n = 248) on the Clinical Teaching Units of the Royal Victoria Hospital (Montréal, Canada). Patients were included if they had a NAAT between January 2014 and September 2015 and their admission diagnosis was not CDI. CDI cases and non-CDI cases were compared, and independent predictors of CDI were determined by logistic regression.
Several factors were independently associated with CDI, including: hemodialysis (OR: 13.5, 95% CI: 2.85–63.8), atrial fibrillation (OR: 3.70, 95% CI: 1.52–9.01), whether the patient received empiric treatment (OR: 3.01, 95% CI: 1.04–8.68), systemic antibiotic therapy prior to testing (OR: 4.23, 95% CI: 1.71–10.5), previous positive NAAT (OR: 3.70, 95% CI: 1.41–9.72), and a leukocyte count of 11×109/L or higher (OR: 3.43, 95% CI: 1.42–8.26). The area under the curve was 0.80.
For patients presenting with hospital-onset diarrhea, various parameters can help differentiate between CDI and other causes. A clinical prediction calculator derived from our cohort (http://individual.utoronto.ca/leet/cdiff.html) might assist clinicians in estimating the risk of CDI for inpatients; those with low pre-test probability may not require immediate testing, treatment, nor prolonged isolation.
Diarrhea is very common among hospitalized patients. There are many causes, including medications, chronic gastrointestinal diseases, as well as acute viral and bacterial infections. Among the latter group, Clostridioides difficile (the preferred name for the former Clostridium difficile  is a leading cause of hospital-onset diarrhea. Since a sudden spike in C. difficile infection (CDI) rates at the turn of the 21st century in both Canadian and American hospitals, it has increased drastically in prevalence across North America [2–6], recently surpassing methicillin-resistant Staphyloccocus aureus (MRSA) as the most common source of nosocomial infection in many American jurisdictions .
Classically, CDI occurs when exposure to antibiotics alters the gut flora, leading to overgrowth of C. difficile and severe diarrhea due to C. difficile toxin. If untreated, infection progresses to pseudomembranous colitis and toxic megacolon, ultimately resulting in shock. In 2011, 500,000 Americans were diagnosed with and 29 000 patients died of CDI, costing an estimated US $4.8 billion in acute treatment alone .
The diagnosis of CDI remains a matter of scientific debate as there is not yet a universally accepted test or testing algorithm. For example, though stool cultures can sensitively detect the presence of C. difficile, they require specific laboratory conditions and expertise, as well as prolonged incubation periods, which have resulted in them falling out of favour as first-line diagnostic tests. A modified toxigenic culture, which not only isolates C. difficile but also confirms the presence of toxin (thereby reducing the possibility of false positive results), is considered the gold standard for diagnosis. However, toxin detection (and the expertise required for these tests) is not yet widely available, so many centres need other rapid tests to rely on.
Since 2009, highly sensitive C. difficile toxin nucleic acid amplification tests (NAATs) have therefore emerged as a test of choice in many institutions . These tests primarily detect and amplify the tcdB gene, which is responsible for the expression of toxin B, and are thought to be very highly sensitive. However, these assays may also detect the presence of unexpressed tcdB genes in patients who do not have disease and may only be asymptomatic carriers of C. difficile. They must therefore be interpreted in the clinical context, mindful of risk factors like recent antibiotic use, to determine if a patient with a positive assay truly has clinical CDI. Ideally, for diagnostic purposes, they should be sent only in patients with high pre-test probability of CDI.
In particular, for hospitalized patients with multiple comorbidities and concomitant medications, it can be difficult to immediately differentiate CDI from another cause of diarrhea when loose stools develop on the ward. Given the associated costs of the overuse of empiric therapy and challenges due to false-positive testing, we believed that clinicians could benefit from an estimate of the pre-test probability of new onset diarrhea for hospitalized patients being caused by CDI prior to requesting a toxin NAAT.
The objective of our study was therefore to identify clinical and laboratory parameters that were ultimately associated with a confirmed diagnosis of CDI, in order to answer: “does this medical inpatient with healthcare-facility onset diarrhea have C. difficile infection?”
Methods and materials
Patient information was obtained from chart review of patients admitted to the medical clinical teaching units (CTU) at the Royal Victoria Hospital in Montréal, Canada. Data was extracted from January 2014 to September 2015. Patients were included in the study if their admission diagnosis was not C. difficile and if they were tested for C. difficile by NAAT during their admission. An episode of CDI was defined as a patient with new diarrhea any time after admission and a positive C. difficile NAAT who received at least ten days of treatment with either oral vancomycin or metronidazole.
Our cases are described as healthcare-facility onset diarrhea with acquisition of C. difficile either in community or at the healthcare institution. At our hospital, in the absence of toxic megacolon or shock, a C. difficile NAAT is sent when there are three or more unformed stools within 24 hours and all cases were confirmed to meet the standard IDSA-SHEA definition  by trained infection control personnel.
Clinical and laboratory data were extracted from patient records. Clinical data notably included medical comorbidities and previous history of positive C. difficile toxin NAAT (dating back to June 2010, which is when the test entered use at our institution). We also recorded laxative prescription within the last 48 hours, number of bowel movements within 24 hours, as well as charting of abdominal tenderness and the quality of stool within 24 hours. Our hospital protocol states that NAAT tests are, in general, not sent for fewer than three bowel movements in 24 hours. Therefore, patients who did not have a specific documented number of bowel movements were assumed to have less than five but at minimum three bowel movements. Additionally, when abdominal tenderness was not explicitly mentioned in the chart, it was assumed to be absent.
Although laxatives were often held after diarrhea developed, this variable was included to determine if laxative prescription at time of diarrhea onset influenced pre-test probability for CDI. Fever within 24 hours of diarrhea was included and assessed by charted vital signs, which at our institution are measured every eight hours on the inpatient unit. Antibiotic exposure within the previous 30 days was also recorded and assessed by prescription history within our hospital network, or mention of recent antibiotic exposure in the patient’s chart. Empiric therapy was defined as the initiation of C. difficile therapy prior to the receipt of NAAT results. At least 10 days of C. difficile directed therapy was considered to be a full course of CDI therapy. The highest white blood cell (WBC) count on the day of sample procurement as well as in the preceding 72 hours was recorded and leukocytosis was defined as a WBC above 11 × 109/L as this is the cut-off in our laboratory. Creatinine levels on the day of testing were also recorded; acute kidney injury was defined as an increase of 26 μmol/mL from a patient’s baseline, excluding patients with end-stage renal disease on dialysis who were considered separately . For both WBC and creatinine, if same-date data were not available, data from within 48 hours before or after the specimen procurement was used.
Ethics approval was granted by the McGill University Health Centre Research Ethics Board who waived the requirement for informed consent given the retrospective nature of our study. After data was extracted, all patient data were de-identified and analyzed in an anonymous fashion.
Univariate comparisons were made using chi-square and multivariate comparisons were made using logistic regression, with a P-value of .05 representing significance. To avoid over-representing patients with multiple admissions and multiple NAAT tests sent over the study period, only the first test result per patient was used in the derivation of the regression model. The initial logistic regression model was selected using backwards selection with a goal of maximizing the c-statistic. We initially started will all covariates in the model and worked backwards by removing those with a p-value above 0.05 until we arrived at the final model. We then attempted to build the model using forwards selection involving all covariates which converged on the same result. We then forced potential confounders such as age (which were not independently associated) into this model to determine if the c-statistic improved. Since it did not, and those covariates were not associated with the outcome of interest, we did not include them in our final selected model. This final model was converted into an interactive Microsoft Excel calculator using the co-efficient for each term in the equation for the post-test probability predicted by regression.
Of the 2537 admissions to the CTU within our specified time frame, 319 NAAT tests for C. difficile were performed. These samples were obtained from 248 unique patients who were admitted in the absence of an admission diagnosis of CDI. 215 of these unique tests were negative (87%), whereas 33 (13%) were positive and were classified as having CDI (see Table 1). 2 (0.8%) patients tested positive but did not receive treatment, were considered asymptomatic carriers by the clinical treating teams, and are therefore not included in Table 1.
Demographics of patients included.
Cases of CDI N = 33 (%)
Controls N = 215 (%)
Median age (years) (IQR)
Coronary Artery Disease
Congestive Heart Failure
Solid Organ Transplant
Solid or Hematologic Cancer
End-Stage Renal Disease on Dialysis
Antibiotic Exposure in Past 30 Days
Five or more bowel movements
Treated Empirically for CDI
Active Laxative Prescription
Proton Pump Inhibitor Use
Previous Positive C. Diff NAAT
Leukocytosis > = 11×109 cells/mL
Median maximal WBC count (x109 cells/mL) (IQR)
Acute Kidney Injury1
1: patients with end-stage renal disease on hemodialysis were excluded from this measurement.
In our multivariable model, a number of factors were independently associated with CDI in medical inpatients including: receipt of hemodialysis (OR: 13.5, 95% CI: 2.85–63.8), atrial fibrillation (OR: 3.70, 95% CI: 1.52–9.01), whether the patient received empiric treatment (OR: 3.01, 95% CI: 1.04–8.68), receipt of systemic antibiotics in the 30 days prior to the test (OR: 4.23, 95% CI: 1.71–10.5), previous positive NAAT (OR: 3.70, 95% CI: 1.41–9.72), and a leukocyte count of 11×109/L or higher (OR: 3.43, 95% CI: 1.42–8.26) (see Table 2). The presence of fever, abdominal tenderness, acute kidney injury, number of bowel movements, description of the stool and laxative use within 48 hours were not found to be independent predictors of CDI. The area under the receiver-operator curve (c-statistic) for the model was 0.80.
Clinical and laboratory factors independently associated with CDI.
Warrants Empiric Therapy
Antibiotic exposure ≤ 30 days
Previous positive C. diff NAAT
WBC Count ≥ 11
We used the results of our study to derive a preliminary clinical prediction calculator (accessible at: http://individual.utoronto.ca/leet/cdiff.html), which converts the coefficients of our regression model into a means by which one could estimate the pre-test CDI probability for individual medical patients. Most non-CDI cases of diarrhea in our cohort were ultimately considered to be related to medication side effects although there were likely other infectious illnesses with diarrhea as a component (influenza and viral gastroenteritis). There were no cases of de novo inflammatory bowel disease diagnosed.
Our data suggest that a number of clinical or laboratory findings are associated with an increased risk of CDI among hospitalized patients with diarrhea. Unsurprisingly, recent antibiotic exposure (OR: 4.23, 95% CI: 1.71–10.5), a previous positive NAAT (OR: 3.70, 95% CI: 1.41–9.72) and leukocytosis (OR: 3.43, 95% CI: 1.42–8.26) were associated with increased risk of CDI in our model. These findings corroborate the literature as being important and independent markers of disease.
In addition, our results support the concept that clinical judgment as shown by the belief that the patient requires empiric therapy while diagnostic testing results are pending is a reasonable predictor that the patient is more likely to have CDI than another cause of diarrhea (OR: 3.01, 95% CI: 1.04–8.68). Although this concept is subjective, it reflects the many complexities of clinical judgment, which are difficult to precisely quantify but which have been shown in other clinical prediction rules to add discriminative power .
Our data also suggests that dialyzed patients with diarrhea were significantly more likely (OR: 13.5, 95% CI: 2.85–63.8) than patients not on dialysis to have CDI. In our analysis, AKI and creatinine levels were not correlated with increased probability of CDI, suggesting that renal injury was not the inherent predisposing risk factor. Rather, this finding is more likely explained by the fact that dialyzed patients, whose frequent weekly visits over a long time period cumulate to significant hospital exposure, are more likely to be exposed to and acquire C. difficile. In addition, although it is possible that these patients’ uniquely thorough medical records may cause misclassification bias, their charts are consolidated upon admission to the inpatient unit, which mitigates this risk. As these patients are also relatively immune suppressed, they may also be more likely to experience disease .
Atrial fibrillation (OR: 3.70, 95% CI: 1.52–9.01) was also found to increase the likelihood of a patient with diarrhea having CDI. This may be a chance association due to an unmeasured covariate within our cohort or represent an increased risk of CDI due to other metabolic and/or pharmacologic differences for these patients.
Our work has several limitations. First, our study population was restricted to the inpatient clinical teaching units of one institution and our sample size was relatively small so our findings should be interpreted with caution. We did not have enough of a sample to provide a derivation and validation subset and our model requires validation elsewhere. However, at the time of our study, these units had the highest rate of inpatient CDI at our institution and as it is a general medical unit the patients represent a heterogeneous and diverse population. Secondly, we have relied on charted comorbidities, vital signs (including temperature), physical findings, and stool characteristics which may have introduced information bias in cases where this information was incorrectly documented. Moreover, as we only had access to data from our own hospital network, we lacked comprehensive data on a full history of healthcare exposures, previous positive C. difficile NAAT assays as well as out-of-centre antibiotic use. Nonetheless, the data available in our chart review was the same information that treating teams used when deciding to test for C. difficile or to treat a positive result. Furthermore, at our institution, admission comorbidities are reviewed at minimum three times (initially by the emergency room physician, the emergency room internal medicine consultant, as well as on admission to the internal medicine unit). They are therefore highly comprehensive at our centre.
Thirdly, although NAAT tests were sent shortly after the onset of diarrhea due to standard operating procedures in our institution, the exact timing relative to symptom onset could not be ascertained due to the retrospective nature of our study It is also worth noting that patients with previous positive NAAT may be more likely to be tested, which may affect our results. Indeed, if CDI colonization status is known (because of prior NAAT results or previous confirmed infection) there could certainly be a bias towards ordering testing in such patients which may bias the results in these cases. A prospective study with structured data collection, assessing information on other risk factors such as the magnitude of health care contact or C. difficile exposure/colonization pressure, would lead to a superior tool.
Finally, we only evaluated patients whose stool was tested and not all patients with diarrhea in the hospital. Therefore, we may have under-sampled the group of patients that the treating team felt had the lowest probability of CDI. It is possible that not all cases of CDI were identified in our cohort, such that false negatives may have been considered CDI-negative controls. However, the NAAT employed has sensitivity close to 100%  and no negative patient was fully treated for CDI.
Our study therefore presents a number of clinical and laboratory findings, outside of the usually well-recognized risk factors, that influence the likelihood of a hospitalized patient with healthcare facility-onset diarrhea having CDI. Importantly, unlike previous work, which compares patients with CDI to those without irrespective of clinical presentation [17, 18], we have attempted to determine who has CDI amongst the more relevant subset of patients with diarrhea. Our results encourage diagnostic stewardship by helping to determine the pre-test probability of CDI in patients with healthcare facility-onset diarrhea, to help clinicians evaluate the utility of a toxin NAAT. To translate our data into an actionable clinical tool, we integrated our results into a model that allows clinicians to enter patient information and determine a likelihood of a patient having CDI. It remains to be seen whether such a rule would be valid in a derivation cohort and if the use of these pre-test probabilities might fit into testing and empiric treatment algorithms.
Our cohort is the first to specifically examine the clinical prediction of CDI in the setting of new onset diarrhea in medical inpatients. For this specific population, certain clinical and laboratory parameters may be useful to differentiate between those with CDI and other aetiologies. If validated in a future cohort, this clinical prediction rule may adjust the pre-test probability of CDI for this patient population and inform diagnostic and management approaches.
The authors would like to thank Dr. Vivian G. Loo for her review of the manuscript, as well as Kendra Bradshaw, MDCM and Alexandre P. Cheng, B.Eng for their assistance in data collection.
The authors received no specific funding for this work.
All relevant data underlying our study’s findings can be found within the manuscript.
1. M100: Performance Standards for Antimicrobial Susceptibility Testing. Clinical and Laboratory Standards Institute, 2018.
2. Loo VG, Poirier L, Miller MA, Oughton M, Libman MD, Michaud S, et al. A predominantly clonal multi-institutional outbreak of Clostridium difficile–associated diarrhea with high morbidity and mortality. New England Journal of Medicine. 2005;353(23):2442–9. 10.1056/NEJMoa051639 [PubMed] [Google Scholar]
3. Muto CA, Pokrywka M, Shutt K, Mendelsohn AB, Nouri K, Posey K, et al. A large outbreak of Clostridium difficile–associated disease with an unexpected proportion of deaths and colectomies at a teaching hospital following increased fluoroquinolone use. Infection Control & Hospital Epidemiology. 2005;26(3):273–80. [PubMed] [Google Scholar]
4. Pépin J, Valiquette L, Cossette B. Mortality attributable to nosocomial Clostridium difficile–associated disease during an epidemic caused by a hypervirulent strain in Quebec. Canadian Medical Association Journal. 2005;173(9):1037–42. 10.1503/cmaj.050978[PMC free article] [PubMed] [Google Scholar]
6. Sailhamer EA, Carson K, Chang Y, Zacharias N, Spaniolas K, Tabbara M, et al. Fulminant Clostridium difficile colitis: patterns of care and predictors of mortality. Archives of surgery. 2009;144(5):433–9. 10.1001/archsurg.2009.51 [PubMed] [Google Scholar]
7. Miller BA, Chen LF, Sexton DJ, Anderson DJ. Comparison of the burdens of hospital-onset, healthcare facility-associated Clostridium difficile Infection and of healthcare-associated infection due to methicillin-resistant Staphylococcus aureus in community hospitals. Infect Control Hosp Epidemiol. 2011;32(4):387–90. 10.1086/659156 . [PubMed] [Google Scholar]
8. Lessa FC, Mu Y, Bamberg WM, Beldavs ZG, Dumyati GK, Dunn JR, et al. Burden of Clostridium difficile infection in the United States. N Engl J Med. 2015;372(9):825–34. 10.1056/NEJMoa1408913 . [PubMed] [Google Scholar]
9. Burnham CA, Carroll KC. Diagnosis of Clostridium difficile infection: an ongoing conundrum for clinicians and for clinical laboratories. Clin Microbiol Rev. 2013;26(3):604–30. 10.1128/CMR.00016-13 ; PubMed Central PMCID: PMCPMC3719497. [PMC free article] [PubMed] [Google Scholar]
12. McDonald LC, Gerding DN, Johnson S, Bakken JS, Carroll KC, Coffin SE, et al. 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). Clinical Infectious Diseases. 2018;66(7):e1–e48. 10.1093/cid/cix1085[PMC free article] [PubMed] [Google Scholar]
13. Lewington A, Kanagasundaram S. Renal Association Clinical Practice Guidelines on acute kidney injury. Nephron Clin Pract. 2011;118Suppl 1:c349–90. 10.1159/000328075 . [PubMed] [Google Scholar]
14. Carrier M, Wells PS, Rodger MA. Excluding pulmonary embolism at the bedside with low pre-test probability and D-dimer: safety and clinical utility of 4 methods to assign pre-test probability. Thromb Res. 2006;117(4):469–74. 10.1016/j.thromres.2005.04.005 . [PubMed] [Google Scholar]
16. de Jong E, de Jong AS, Bartels CJ, van der Rijt-van den Biggelaar C, Melchers WJ, Sturm PD. Clinical and laboratory evaluation of a real-time PCR for Clostridium difficile toxin A and B genes. Eur J Clin Microbiol Infect Dis. 2012;31(9):2219–25. 10.1007/s10096-012-1558-1 ; PubMed Central PMCID: PMCPMC3418502. [PMC free article] [PubMed] [Google Scholar]
17. Pepin J, Saheb N, Coulombe MA, Alary ME, Corriveau MP, Authier S, et al. Emergence of fluoroquinolones as the predominant risk factor for Clostridium difficile-associated diarrhea: a cohort study during an epidemic in Quebec. Clin Infect Dis. 2005;41(9):1254–60. 10.1086/496986 . [PubMed] [Google Scholar]
18. Howell MD, Novack V, Grgurich P, Soulliard D, Novack L, Pencina M, et al. Iatrogenic gastric acid suppression and the risk of nosocomial Clostridium difficile infection. Arch Intern Med. 2010;170(9):784–90. 10.1001/archinternmed.2010.89 . [PubMed] [Google Scholar]