You are here

Prevalence and risk factors of Clostridium difficile infection in patients hospitalized for flare of inflammatory bowel disease: A retrospective assessment

Digestive and Liver Disease, 12, 46, pages 1086 - 1092

Abstract

Background

Recent studies have identified a high frequency ofClostridium difficileinfections in patients with active inflammatory bowel disease.

Aims

To retrospectively assess the determinants and results ofClostridium difficiletesting upon the admission of patients hospitalized with active inflammatory bowel disease in a tertiary care centre and to determine the predicting factors ofClostridium difficileinfections.

Methods

We reviewed all admissions from January 2008 and December 2010 for inflammatory bowel disease flare-ups. A toxigenic culture and a stool cytotoxicity assay were performed for all patients tested forClostridium difficile.

Results

Out of 813 consecutive stays,Clostridium difficilediagnostic assays have been performed in 59% of inpatients. The independent predictive factors for the testing were IBD (ulcerative colitis: OR 2.0, 95% CI 1.5–2.9;p < 0.0001) and colonic involvement at admission (OR 2.2, 95% CI 1.5–3.1,p < 0.0001).Clostridium difficileinfection was present in 7.0% of the inpatients who underwent testing. In a multivariate analysis, the only independent predictor was the intake of nonsteroidal anti-inflammatory drugs within the two months before admission (OR 3.8, 95% CI 1.2–12.3;p = 0.02).

Conclusions

Clostridium difficileinfection is frequently associated with active inflammatory bowel disease. Our study suggests that a recent intake of nonsteroidal anti-inflammatory drugs is a risk factor for inflammatory bowel disease -associatedClostridium difficileinfection.

Keywords: Clostridium difficile infection, Inflammatory bowel disease, Nonsteroidal anti-inflammatory drugs.

1. Introduction

Clostridium difficile(C.difficile) is a Gram-positive anaerobic spore-forming bacterium that produces toxins. It is responsible for a spectrum of clinical presentations, ranging from asymptomatic carriage to clinically significant diarrhoea, fulminant colitis, sepsis and death[1] and [2]. Risk factors forC. difficileinfections (CDI) traditionally include antibiotic use, age, severe comorbidities or contact with healthcare facilities. An alarming trend of increased incidence and severity of CDI over the past several years has been reported worldwide, along with an increased duration of hospitalization and costs[3] and [4]. As an example, the estimated costs attributed to CDI alone are approximately 3.2 billion dollars per year in the United States [5] .

It is particularly difficult to diagnose CDI in patients with inflammatory bowel disease (IBD; Crohn's disease (CD), ulcerative colitis (UC)), because the presentation is similar and consists of abdominal pain and diarrhoea. Indeed, CDI can mimic a flare of the disease, but can also trigger a relapse that could evolve on its own, regardless of the treatment of the infection[6] and [7]. The relationship between CDI and IBD flares has been recognized as an emerging problem. Several retrospective studies on large populations have suggested a rise in the rate of CDI among patients with IBD along with an increase in disease severity. For example, the incidence of CDI among hospitalized patients with IBD increased from 1.8% in 2004 to 4.6% in 2005 [8] . CDI appeared to be three times more frequent in patients with UC than in patients with CD, but the percentage of patients who were tested forC.difficilewas not specified[9], [10], [11], and [12]. CDI was associated with an increased severity of relapse, an increase in the number and length of hospital stays, higher rates of colectomy, and a four-fold increase in mortality[9] and [10]. In these studies, infection in patients with IBD was predominantly community-acquired, in contrast to the general inpatient population where CDI is primarily acquired within the healthcare setting itself[8] and [12].

To date, the risk factors for CDI in patients with IBD are not firmly established. Some authors have tried to identify them in cohorts that may not necessarily be representative of the study population because they have included all hospitalized patients, whereas they should have included only patients who were tested forC. difficile. In our tertiary center,C. difficiletesting upon the admission of a patient for an IBD flare has been gradually introduced as a routine procedure. This allowed us to conduct this study with two objectives: first, to assess the frequency and the determinants ofC. difficiletesting upon the admission of patients hospitalized for a flare of IBD, and second, to identify risk factors for CDI in patients hospitalized for a flare of IBD.

2. Methods

2.1. Study population and variables

Our study population consisted of all patients hospitalized for IBD flares in the Gastroenterology Department of the Saint-Antoine IBD center from 1 January 2008 to 31 December 2010. We used data from the department of medical information and established a list of all the consecutive stays with the following WHO codes: K 50 (CD) and K 51 (UC). In parallel, using hospitalization reports, we identified all patients with an established diagnosis of IBD and symptoms of active IBD (diarrhoea with possible presence of blood/mucus and/or abdominal cramps) and who were tested forC.difficilebetween 1 January 2008 and 31 December 2010. Those two lists were merged to identify patients who were both hospitalized with a flare of IBD and tested forC. difficile. A systematic verification of inclusion criteria was made from reports of hospitalization and/or from medical records. Patients without active IBD at admission were excluded ( Fig. 1 ).

gr1

Fig. 1 Flow chart of case definitions in all hospitalized patients for inflammatory bowel disease flares from 2008 to 2010. IBD, inflammatory bowel disease; CD,Clostridium difficile; CDI, clostridiumdifficileinfection.

Demographic characteristics included gender, age and body mass index (BMI) classified into three groups (BMI <18.5: underweight, BMI between 18.5 and 25: normal weight, BMI >25: overweight and obesity). Comorbidity was adjusted using the Charlson index, a well-validated mortality prognosis index [13] . Other considered risk factors were current smoking status at admission, history of appendectomy or ICD. Medication use included chronic use of proton pump inhibitors (PPIs), anti-inflammatory drugs (NSAIDs) or antibiotics within the two months prior to hospitalization. We recorded disease characteristics including the type of IBD (i.e., CD or UC), anatomic distribution of the disease (isolated small bowel vs. colonic involvement < or >50%, perianal lesions), the duration of disease from diagnosis and any history of a previous surgical resection (segmental or total colectomy, ileocecal resection). We also recorded the type of therapy, including corticosteroid use within the two months prior to hospitalization, which allowed us to classify patients into 3 groups: group 1 = no steroids, group 2 = between 20 and 40 milligrams of prednisolone or budesonide equivalent, group 3 = 40 milligrams or more of prednisolone equivalent. Other therapies included immunomodulators (azathioprine, 6-mercaptopurine), methotrexate and anti-TNF (infliximab, adalimumab, certolizumab). Data on the IBD flare-ups included the date of admission, length of stay, and white blood cell count.

2.2. C. difficile testing

In our tertiary-care center, the diagnosis of CDI systematically involved a stool culture on selective medium and a stool cytotoxicity assay. In the case of a positive culture and a negative stool cytotoxicity assay, a toxigenic culture (in vitro determination of the ability of the isolate to produce toxins) was performed. CDI was defined either by a positive stool toxigenic culture and a positive stool cytotoxicity assay or, in cases of negative stool cytotoxicity assays, a positive toxigenic culture[14] and [15]. In accordance with the definition generally admitted in the literature, CDI was defined as community-acquired if the patient had a positiveC. difficiletest within 48 h following admission and no hospitalization in the two months prior to the current admission. CDI was considered hospital-acquired if patients were hospitalized in the two months prior to the current admission or if their test results were positive during their hospital stay after a negative result from the initial test performed at admission. Finally, if CDI testing had been conducted after three days in the absence of available data at the time of hospital admission, the origin of the CDI was defined as indeterminate.

For patients with CDI, complementary data were recorded, including clinical data (existence of fever, diarrhoea, rectal bleeding), biological data (serum levels of C-reactive protein, albumin), severity of endoscopic criteria (false membranes, ulcerations covering more than half of the circumference, mucosal detachment), and the index of activity (for CD, a Harvey-Bradshaw score higher than 12 defines severe CD [16] , while for UC, a Lichtiger score greater than 10 defined severe acute UC [17] ). We also recorded the following types of therapy and treatments: empiric antibiotics (e.g., fluoroquinolone or metronidazole), CDI treatment (e.g., metronidazole or vancomycin), corticosteroids and maintenance or intensification of immunosuppressive therapy (including methotrexate, azathioprine, 6-mercaptopurine, anti-TNFα). Finally, we recorded any surgical treatment or death.

2.3. Statistical analysis

The data were processed using Statview®software (version 5.0, SAS Institute Inc., Cary, NC). The results of the quantitative variables are expressed as a median and an interquartile range. The qualitative variables are expressed as numbers and percentages. To identify the predicting factors of CDI testing, a univariate analysis was performed using theχ2test and Fisher's exact test, when necessary. The variables were as follows: age, gender, BMI, Charlson index, type of IBD, colonic involvement upon hospital admission, duration of disease, current smoking status, CDI therapy (e.g., corticosteroid equivalent dose equal to or greater than 20 milligrams of prednisolone, immunosuppressants), antibiotics, PPIs and NSAIDs used in the previous two months prior to admission and white blood cell count. For the purpose of the analysis, quantitative variables were dichotomized around the median. A p value less than 0.05 was considered statistically significant. The variables withpvalues <0.35 in the univariate analysis were entered in a multivariate logistic regression analysis using a backward stepwise procedure. To assess the predictors of CDI among the patients who were tested, univariate and multivariate analyses were made from the above variables using the same statistical procedure.

3. Results

3.1. Characteristics of the study population

A total of 813 hospital stays from 1 January 2008 to 31 December 2010 corresponding to 584 patients were analyzed. In all, 104, 23, 16, 3, 3 and 1 patients completed 1, 2, 3, 4, 5, 6 and 7 stays, respectively. Table 1 shows the demographic characteristics of the patients. There were more women than men hospitalized for IBD flares in this study, both in the group with CD and in the group with UC. Approximately two thirds of the patients had CD (n = 538) and one third had UC (n = 77). A majority of the patients had extensive colitis, which was observed in patients with both CD (50, 5%) and in those with UC (72, 4%). Tobacco use was noted in 229 patients with CD (43%) and in 50 with UC (18%). Approximately two thirds of the patients were receiving immunosuppressive treatment at the time of admission, either corticosteroids, immunomodulators or anti-TNF drugs. NSAIDs use, including only non-selective Cox inhibitors, was noted in 27 patients.

Table 1 Characteristics of patients admitted for a flare-up of inflammatory bowel disease.

  Crohn's disease (n = 534) Ulcerative colitis (n = 279)
Gender, n (%)
 Male 236 (44.2) 125 (44.8)
 
Age (years)
 Median (range) 31 (25–43) 33 (24–46)
 
Number of patients, n
 2008 180 88
 2009 199 94
 2010 155 97
 
Body mass index
 Median (range) 20 (18–22) 21 (19–24)
 
Duration of disease (years)
 Median (range) 7 (2–14) 5 (2–10)
 
Extent of disease, n (%)
 Ileal 424 (79) 24 (8.6)
 Colonic <50% 20 (3.7) 63 (22.5)
 Colonic >50% 270 (50.5) 202 (72.4)
 Anoperineal 238 (44.5) 45 (16)
 
Charlson comorbidity index
 Median (range) 2 (1–3) 2 (1–3)
 Smoking, n (%) 229 (43) 50 (18)
 
Inflammatory bowel disease –related medication, n (%)
 1. Corticosteroids >20 mg 90 (16.8) 94 (33.7)
 2. Immunomodulators 152 (28.4) 44 (15.7)
 3. Anti-TNF 1 or 2 or 3 129 (24.1) 67 (24.0)
  371 (69.4) 205 (72.4)

3.2. C. difficile testing rate and predictors for CDI testing

C. difficiletesting within 48 h of admission for patients hospitalized for an IBD flare was performed in 483 (59%) of the 813 patients. By univariate analysis,C. difficiletesting did not vary significantly according to gender or age ( Table 2 ). Predictive factors forC. difficiletesting within 48 h of admission were UC (p < 0.0001) and colonic involvement attributed to IBD on admission (p < 0.0001). By multivariate analysis, the two independent predictors forC. difficiletesting on admission in patients hospitalized for an IBD flare were UC (OR 2, 95% CI 1.5–2.9,p < 0.0001) and colonic involvement (OR 2.2, 95% CI 1.5–3.1,p < 0.0001).

Table 2 Characteristics of patients with or withoutClostridium difficiletesting at admission (univariate analysis).

  Clostridium difficile testing (n = 483) No Clostridium difficile testing (n = 330) p
Age (years)
 Median (range) 32 (25–43) 31 (24–44) 0.74
 
Gender, n (%)
 Male 220 (45.5) 142 (43) 0.69
 
Body mass index, n (%)
 <18.5 114 (23.6) 84 (25.4) 0.72
 18.5–25 289 (59.8) 192 (58.1)  
 >25 80 (16.5) 54 (16.3)  
 
Charlson index
 Median (range) 2 (1–3.25) 2 (1–3) 0.63
 
Inflammatory bowel disease type, n (%)
 Crohn's disease 277 (57.3) 259 (78.4) < 0.0001
 Ulcerative colitis 206 (42.7) 71 (21.6)  
 
Colonic involvement, n (%) 364 (75.3) 191 (57.8) < 0.0001
 
Smoking, n (%) 155 (32.1) 123 (37.3) 0.12
 
History of Clostridium difficile infection, n (%) 14 (3.4) 4 (0.9) 0.11
 
Community context, n (%) 439 (90.7) 301 (91.5) 0.36
 
Maintenance therapy, n (%)
 Corticosteroids a 111 (22.9) 73 (22.1) 0.71
 Immunosuppressants 244 (49.8) 168 (50.9) 0.94
 
Antibiotics, n (%) b 122 (25.2) 69 (20.9) 0.34
 
Proton pump inhibitors, n (%) 54 (11.2) 47 (14.2) 0.21
 
Nonsteroidal anti-inflammatory drugs, n (%) b 19 (3.9) 8 (2.4) 0.18
 
White blood cell count
 Median (range) 8340 (6100–11,270) 8295 (6205–11,580) 0.9

a >20 milligrams prednisolone equivalent.

b In the two months prior to the admission.

pis significant if <0.05.

3.3. Characteristics of CDI and predictors for CDI

CDI was detected in 34 patients, representing 4.1% of all stays over the three years that were included in our study and 7% of stays in whichC. difficiletesting was performed ( Table 3 ). In a univariate analysis, the predictors of CDI for patients who were tested were a BMI <18.5 or >25, NSAID intake within the two months prior to admission, and a high white blood cell count. In a multivariate analysis, the only independent risk factor was NSAID intake within 2 months prior to admission (OR 3.8, 95% CI 1.2–12.3,p = 0.02).

Table 3 Characteristics of patients with or withoutClostridium difficileinfection (univariate analysis).

  Clostridium difficile infection (n = 34) No Clostridium difficile infection (n = 449) p
Age (years)
 Median (range) 33 (22–46) 32 (25–44) 0.85
 
Gender, n (%)
 Male 17 (50) 246 (55.1) 0.58
 
Body mass index, n (%)
 <18.5 11 (32.3) 103 (23) 0.04
 18.5–25 14 (41.2) 275 (61.6)  
 >25 8 (23.5) 57 (12.7)  
 
Charlson index
 Median (range) 3 (2–5) 2 (1–3) 0.85
 
Inflammatory bowel disease type, n (%)
 Crohn's disease 20 (58.8) 258 (57.8) 0.85
 Ulcerative colitis 14 (41.2) 191 (42.2)  
 
Colonic involvement, n (%) 27 (79.4) 337 (75.5) 0.58
 
Duration of disease
 Median (range) 4 (2–6) 6 (2–13) 0.26
 
Smoking, n (%) 11 (32.3) 144 (32.2) 0.97
 
History of Clostridium difficile infection, n (%) 2 (5.8) 16 (3.5) 0.37
 
Maintenance therapy
 Corticosteroids a 8 103 0.93
 Immunomodulators b 8 121  
 Anti-TNF 9 108  
 Total (%) 25 (73.5) 332 (73.9)  
 
Antibiotics, n (%) 12 (35.3) 110 (24.5) 0.16
 
Proton pump inhibitors, n (%) 3 (8.8) 51 (11.3) 0.64
 
Nonsteroidal anti-inflammatory drugs, n (%) 4 (11.7) 15 (3.3) 0.01
 
White blood cell count
 Median (range) 9900 (7200–12,320) 8235 (6088–11,315) 0.0037

a >20 milligrams prednisolone equivalent.

b In the two months prior to the admission, immunomodulators include azathioprine, methotrexate and purinethol.

pis significant if <0.05.

Of the 34 cases of CDI, 19 were community-acquired, 8 were healthcare-associated, and 7 were undetermined infections. Characteristics of patients who developed one of the three types of CDI are detailed in Table 4 . IBD was active by definition in the three groups. Of the 19 community-acquired infections, 11 patients had an endoscopy, three of which revealed pseudomembranes along with three additional endoscopic severity criteria. All patients had been treated for CDI with either oral metronidazole or oral vancomycin. Approximately half of them had received corticosteroids in addition to antibiotics. Background therapy was changed in 6 of 19 patients. In the case of healthcare-associated infections, the activity of IBD met the criteria for severity in the majority of cases. The majority of patients had been diagnosed with UC, and no pseudomembranes were reported on endoscopic examination.

Table 4 Characteristics of patients hospitalized for inflammatory bowel disease flare withClostridium difficileinfection.

  Community- acquired Clostridium difficile infection * (n = 19) Healthcare- associated Clostridium difficile infection ** (n = 8) Indeterminate Clostridium difficile infection *** (n = 7)
Crohn's disease, n (%) 13 (68.4) 3 (37.5) 3 (43)
 
Ulcerative colitis, n (%) 6 (31.5) 5 (62.5) 4 (57)
 
Duration of disease (years)
 Median (range) 4.58 3.70 3.95
 
Severe flare of Crohn's disease **** , n (%) 5 (38.4) 2 (66.7) 1 (33.3)
 
Severe flare of Ulcerative colitis ***** , n (%) 4 (67) 3 (60) 1 (25)
 
Charlson index
 Median (range) 2.5 0.5 3
 
Endoscopic features, n (%) 11 (57.9) 7 (87.5) 4 (57)
 • Pseudomembranes 3 (15.8) 0 0
 • Severity criteria 3 (15.8) 1 (12.5) 1 (14.3)
 
Probabilistic antibiotherapy, n (%) 7 (37) 1 (12.5) 1 (14)
 • Metronidazole 6 1 1
 • Fluoroquinolones 5 1 0
Corticosteroids, n (%) 9 (47) 4 (50) 3 (43)
 
Intensification or maintenance of immunosuppressive drugs, n (%) 6 (31.5) 6 (75) 1 (14)
 • Immunomodulators a 4 3 0
 • Anti-TNF 3 5 1
 
Length of stay (days)
 Median (range) 8 (11.75–29.25) 32.5 (25.25–49.75) 8 (4.75–10.25)
 
Smoking, n (%) 7 (37) 3 (37.5) 1 (14)
 
Clostridium difficile infection treatment, n (%)
 • Metronidazole 13 (68.4) 7 (87.5) 4 (57.1)
 •Vancomycin 6 (31.6) 2 (25) 0

* Clostridium difficile testing conducted within the first forty-eight hours after hospitalization where the patient was not hospitalized in the two months prior to the admission.

** Patient hospitalized in the two months prior to the admission or Clostridium difficile testing was positive during the stay after a negative first test on admission.

*** Clostridium difficile testing conducted after three days without available data on admission.

**** Defined as a Harvey-Bradshaw score > 12.

***** Defined as a Lichtiger score >10.

a Immunomodulators (azathioprine, purinethol, methotrexate).

4. Discussion

Our study shows that CDI occurs in 7% of patients with IBD who are hospitalized for flares and tested forC. difficile. We also found that recent intake of NSAIDS is an independent predictor for CDI.

Presently, there is no consensus methodology for the diagnosis of CDI in a population of patients with IBD. In the general population, the definition encompasses clinical criteria (acute diarrhoea for more than two days), bacteriological criteria (positive detection of toxins in the stool or isolation of a toxigenic strain) and/or endoscopic pseudomembranes [18] . Clinical criteria are difficult to interpret in patients with IBD because diarrhoea is also a symptom of a flare. In addition, colonization withC.difficilewas demonstrated more frequently in patients with quiescent IBD compared to healthy subjects (7% CD, 9% UC, 1% controls) [19] . Various experts claim that there is an association between toxin detection in stool and toxigenic culture in the general population. In our study, we applied this strategy using both complementary techniques to confirm the diagnosis in symptomatic patients. We believe that these methodological conditions were optimal to assess the frequency, impact and risk factors for CDI in patients hospitalized for IBD flares and who were tested for CDI.

In this retrospective study that ended in December 2010, just before generalization of CDI testing in our centre, we found that testing for CDI was performed primarily in patients with UC and in those with colonic involvement at admission; this was most likely influenced by several studies that designated these two characteristics as risk factors for CDI[8] and [12]. In fact, the results of these initial studies were biased because the statistical analysis was not restricted to those patients in whomC.difficiletesting was performed. A review of more recent literature that assessed 42 studies is in agreement with our findings: the incidence of CDI was the same for patients with UC and for those with CD and was the same in patients hospitalized for a flare of the disease and among those who were asymptomatic [20] . In another study, the increase in the prevalence of CDI was similar in patients with UC compared to those with CD and colonic involvement [21] . In our study, UC and colonic involvement were not risk factors for CDI.

We acknowledge that in our retrospective study, only 59% of the patients admitted for active IBD were tested for CDI, which introduced a selection bias of the population tested. Due to this selection bias, our study population cannot be considered as representative of the patients admitted for IBD flare. However, our multivariate analysis for identifying risk factors of CDI was restricted to patients tested for CDI and well-phenotyped. This ‘denominator’ was accurate, in contrast to most of the studies published to date that considered together in the search for risk factors of CDI patients tested and not tested for CDI.

We found that recent use of NSAIDs was an independent predictor for CDI even we remain cautious with this result because of the small number of patients. NSAIDs operate by inhibiting cyclooxygenase (COX), which is an enzyme capable of producing anti-inflammatory prostaglandins [22] . They alter the intestinal barrier, especially by modifying the microcirculation and increasing its permeability [23] . NSAIDs can promote acute diarrhoea in the general population [24] but can also trigger severe acute colitis that may be fatal[25] and [26]. In most cases, NSAIDs trigger or aggravate infectious or inflammatory colitis. Several studies have shown that the use of NSAIDs can also trigger or exacerbate a flare-up of IBD[27], [28], and [29]. Conflicting results have been published that emphasize that NSAID use is not associated with an increased likelihood of active disease in an outpatient population with IBD [30] . However, a prospective trial showed that NSAIDs triggered a flare in approximately 20% of patients with quiescent disease [31] . In accordance with these data, some American scientific societies have considered that the evidence of the deleterious effects of NSAIDs was obvious enough to recommend that patients with CD avoid these drugs [32] .

In our study, antibiotic use within the eight weeks before admission to the hospital was not recognized as an independent risk factor for CDI, contrary to the findings in the general population [33] . Only 35% of patients with CDI had received antibiotic treatment within the two months before admission, which is similar to findings in previous publications [34] . Similarly, a recent study showed that 61% of patients who developed community-acquired CDI had not been exposed to antibiotic treatment within the three months prior to the infection [35] . The difference with the general population is most likely because patients with IBD usually acquire CDI outside of the hospital and are less exposed to antibiotic treatment.

Healthcare-associated CDI were not uncommon in our study population. They mostly involved patients who suffered from UC that was complicated by severe acute colitis in the majority of cases and who were previously treated with corticosteroids. The diagnosis was sometimes made after repetitiveC. difficiletesting, usually if clinical deterioration occurred.

Our study shows that CDI is frequently associated with flare-ups of IBD that require hospitalization. This justifies systematic testing patients with IBD for CDI upon their admission to the hospital. Furthermore, we found that CDI was diagnosed after several previous negative test results during a hospitalization. A change in practice should logically move towards a systematic way to detect CDI upon admission of patients with IBD flares and be repeated in cases of worsening digestive symptoms during the hospital stay. We also show that the use of NSAIDs in the two months prior to admission appears to be the only independent predictor of a combination of CDI and a flare-up of IBD that requires hospitalization; this is an additional reason to recommend that patients with IBD avoid this particular therapy. It is obvious that it is a preliminary result that needs to be validated in a large cohort of patients. Nevertheless, these results suggest that an exploration of how NSAIDs promote the emergence ofC. difficileduring chronic inflammation might be worthwhile.

Conflict of interest

None declared.

References

  • [1] J.G. Bartlett. Clinical practice. Antibiotic-associated diarrhea. New England Journal of Medicine. 2002;346:334-339 Crossref
  • [2] S. Johnson, D.N. Gerding. Clostridium difficile – associated diarrhea. Clinical Infectious Diseases. 1998;26:1027-1034 quiz 35–6
  • [3] L.C. McDonald, M. Owings, D.B. Jernigan. Clostridium difficile infection in patients discharged from US short-stay hospitals, 1996–2003. Emerging Infectious Diseases. 2006;12:409-415
  • [4] F. Barbut, J.C. Petit. Epidemiology of Clostridium difficile-associated infections. Clinical Microbiology and Infection. 2001;7:405-410 Crossref
  • [5] E.R. Dubberke, A.I. Wertheimer. Review of current literature on the economic burden of Clostridium difficile infection. Infection Control and Hospital Epidemiology. 2009;30:57-66 Crossref
  • [6] A.N. Ananthakrishnan. Clostridium difficile infection: epidemiology, risk factors and management. Nature Reviews Gastroenterology and Hepatology. 2011;8:17-26 Crossref
  • [7] H. Yanai, G.C. Nguyen, L. Yun, et al. Practice of gastroenterologists in treating flaring inflammatory bowel disease patients with Clostridium difficile: antibiotics alone or combined antibiotics/immunomodulators. Inflammatory Bowel Diseases. 2011;17:1540-1546 Crossref
  • [8] M. Issa, A. Vijayapal, M.B. Graham, et al. Impact of Clostridium difficile on inflammatory bowel disease. Clinical Gastroenterology and Hepatology. 2007;5:345-351 Crossref
  • [9] A.N. Ananthakrishnan, E.L. McGinley, D.G. Binion. Excess hospitalisation burden associated with Clostridium difficile in patients with inflammatory bowel disease. Gut. 2008;57:205-210 Crossref
  • [10] A.N. Ananthakrishnan, E.L. McGinley, K. Saeian, et al. Temporal trends in disease outcomes related to Clostridium difficile infection in patients with inflammatory bowel disease. Inflammatory Bowel Diseases. 2011;17:976-983 Crossref
  • [11] G.C. Nguyen, G.G. Kaplan, M.L. Harris, et al. A national survey of the prevalence and impact of Clostridium difficile infection among hospitalized inflammatory bowel disease patients. American Journal of Gastroenterology. 2008;103:1443-1450 Crossref
  • [12] J.F. Rodemann, E.R. Dubberke, K.A. Reske, et al. Incidence of Clostridium difficile infection in inflammatory bowel disease. Clinical Gastroenterology and Hepatology. 2007;5:339-344 Crossref
  • [13] L. Fried, J. Bernardini, B. Piraino. Charlson comorbidity index as a predictor of outcomes in incident peritoneal dialysis patients. American Journal of Kidney Diseases. 2001;37:337-342 Crossref
  • [14] S.H. Cohen, D.N. Gerding, S. Johnson, et al. Clinical practice guidelines for Clostridium difficile infection in adults: 2010 update by the society for healthcare epidemiology of America (SHEA) and the infectious diseases society of America (IDSA). Infection Control and Hospital Epidemiology. 2010;31:431-455 Crossref
  • [15] M.J. Crobach, O.M. Dekkers, M.H. Wilcox, et al. European Society of Clinical Microbiology and Infectious Diseases (ESCMID): data review and recommendations for diagnosing Clostridium difficile-infection (CDI). Clinical Microbiology and Infection. 2009;15:1053-1066 Crossref
  • [16] R.F. Harvey, J.M. Bradshaw. A simple index of Crohn's-disease activity. Lancet. 1980;1:514 Crossref
  • [17] S. Lichtiger, D.H. Present, A. Kornbluth, et al. Cyclosporine in severe ulcerative colitis refractory to steroid therapy. New England Journal of Medicine. 1994;330:1841-1845 Crossref
  • [18] D.N. Gerding, S. Johnson, L.R. Peterson, et al. Clostridium difficile-associated diarrhea and colitis. Infection Control and Hospital Epidemiology. 1995;16:459-477 Crossref
  • [19] E.M. Clayton, M.C. Rea, F. Shanahan, et al. The vexed relationship between Clostridium difficile and inflammatory bowel disease: an assessment of carriage in an outpatient setting among patients in remission. American Journal of Gastroenterology. 2009;104:1162-1169 Crossref
  • [20] J.R. Goodhand, W. Alazawi, D.S. Rampton. Systematic review: Clostridium difficile and inflammatory bowel disease. Alimentary Pharmacology and Therapeutics. 2011;33:428-441
  • [21] R. Ricciardi, J.W. Ogilvie Jr., P.L. Roberts, et al. Epidemiology of Clostridium difficile colitis in hospitalized patients with inflammatory bowel diseases. Diseases of the Colon and Rectum. 2009;52:40-45 Crossref
  • [22] J.R. Vane, J.A. Mitchell, I. Appleton, et al. Inducible isoforms of cyclooxygenase and nitric-oxide synthase in inflammation. Proceedings of the National Academy of Sciences of the United States of America. 1994;91:2046-2050 Crossref
  • [23] G. Thiefin, L. Beaugerie. Toxic effects of nonsteroidal antiinflammatory drugs on the small bowel, colon, and rectum. Joint Bone Spine. 2005;72:286-294 Crossref
  • [24] I. Etienney, L. Beaugerie, C. Viboud, et al. Non-steroidal anti-inflammatory drugs as a risk factor for acute diarrhoea: a case crossover study. Gut. 2003;52:260-263 Crossref
  • [25] A.R. Tanner, A.S. Raghunath. Colonic inflammation and nonsteroidal anti-inflammatory drug administration. An assessment of the frequency of the problem. Digestion. 1988;41:116-120 Crossref
  • [26] G.R. Gibson, E.B. Whitacre, C.A. Ricotti. Colitis induced by nonsteroidal anti-inflammatory drugs. Report of four cases and review of the literature. Archives of Internal Medicine. 1992;152:625-632 Crossref
  • [27] J.M. Evans, A.D. McMahon, F.E. Murray, et al. Non-steroidal anti-inflammatory drugs are associated with emergency admission to hospital for colitis due to inflammatory bowel disease. Gut. 1997;40:619-622
  • [28] J.B. Felder, B.I. Korelitz, R. Rajapakse, et al. Effects of nonsteroidal antiinflammatory drugs on inflammatory bowel disease: a case–control study. American Journal of Gastroenterology. 2000;95:1949-1954 Crossref
  • [29] H.J. Kaufmann, H.L. Taubin. Nonsteroidal anti-inflammatory drugs activate quiescent inflammatory bowel disease. Annals of Internal Medicine. 1987;107:513-516 Crossref
  • [30] G.F. Bonner, M. Walczak, L. Kitchen, et al. Tolerance of nonsteroidal antiinflammatory drugs in patients with inflammatory bowel disease. American Journal of Gastroenterology. 2000;95:1946-1948 Crossref
  • [31] K. Takeuchi, S. Smale, P. Premchand, et al. Prevalence and mechanism of nonsteroidal anti-inflammatory drug-induced clinical relapse in patients with inflammatory bowel disease. Clinical Gastroenterology and Hepatology. 2006;4:196-202 Crossref
  • [32] S.B. Hanauer, W. Sandborn. Practice Parameters Committee of the American College of G. Management of Crohn's disease in adults. American Journal of Gastroenterology. 2001;96:635-643 Crossref
  • [33] L.V. McFarland, M.E. Mulligan, R.Y. Kwok, et al. Nosocomial acquisition of Clostridium difficile infection. New England Journal of Medicine. 1989;320:204-210 Crossref
  • [34] P. Bossuyt, J. Verhaegen, G. Van Assche, et al. Increasing incidence of Clostridium difficile-associated diarrhea in inflammatory bowel disease. Journal of Crohn's and Colitis. 2009;3:4-7 Crossref
  • [35] S. Dial, J.A. Delaney, A.N. Barkun, et al. Use of gastric acid-suppressive agents and the risk of community-acquired Clostridium difficile-associated disease. Journal of the American Medical Association. 2005;294:2989-2995 Crossref

Footnotes

a Department of Gastroenterology, APHP, Saint-Antoine Hospital, Paris, France

b Department of Microbiology, APHP, Saint-Antoine Hospital, Paris, France

c National Reference Laboratory for C. difficile, Paris, France

lowast Corresponding author at: Service de Hépatogastroentérologie, Groupe Hospitalier Pitié-Salpétrière, 47-83 boulevard de l’hôpital, 75651 Paris Cedex, France. Tel.: +33 1 42 16 10 34; fax: +33 1 42 16 14 25.