Background
Clostridium difficile is a bacillus of gram-positive, spore-forming anaerobe, and was first identified in the 1970s. It was considered to be the main cause of pseudomembranous colitis [
1].
C. difficile infection (CDI) has become a particular problem for patients with inflammatory bowel disease (IBD). Studies have found a significant increase in the incidence of CDI over the recent few decades. Patients with IBD have an increased risk of poorer outcomes when suffered from CDI, associated with higher frequency of flare-ups, greater morbidity and mortality, poorer response to treatment, higher rates of colectomy, need for more active treatment for IBD and longer duration of hospital stay [
2,
3].
It is difficult to differentiate CDI in IBD patients from IBD flare, because the presentation is similar, which consists of abdominal pain and diarrhea. Many studies have tried to determine the potential risk factors of CDI in patients with IBD. Risk factors for CDI traditionally include age, antibiotic use, severe comorbidity or contact with hospital and other healthcare facilities [
4]. IBD has been identified as an independent risk factor for CDI. Most patients of IBD with CDI have a history of IBD colitis (91%) [
5]. Ulcerative colitis (UC) patients seem to account for the majority of CDI in the IBD population as a whole [
6,
7].
The incidence of UC is higher than that of Crohn’s disease in China, [
8,
9] while the epidemiology and risk factors of CDI in patients with UC are unclear in China. CDI has not been well acknowledged and routinely tested for in IBD patients with a flare-up. In our tertiary center,
C. difficile testing of patients for an IBD flare has been gradually introduced as a routine procedure. This allowed us to conduct this retrospective study to assess the incidence and risk factors for CDI in patients with UC.
Discussion
Patients with UC are at increased risk of acquiring CDI. The incidence of CDI with UC patients has also been rising in recent years [
5,
7]. CDI mimics and exacerbates UC flare, therefore, it is essential that clinicians be alert to identify CDI and its risk factors, as treatment with corticosteroids without appropriate antibiotics may lead to deterioration [
14]. The present study was a retrospective case–control study to determine the incidence rate and risk factors for CDI in patients with UC flare in a tertiary IBD center in China. We found that CDI occurred in 8.92% of patients with active UC who were hospitalized. CMV infection was associated with CDI in patients with UC according to univariate and binary logistic regression analyses. Recent antibiotic exposure (within 3 months), hospitalization (within 1 month) and systemic use of steroids were associated with CDI according to univariate analysis.
Studies from Europe and North America have shown that the incidence of CDI lies between 2.8 and 5.12% among patients with UC [
3]. In China, however, the incidences of CDI in UC patients were reported 19.3 and 7.1% respectively [
15,
16]. The difference in incidence among the studies may be due to the diversity of disease severity, sample size, and testing methods for
C. difficile. In this study, the result was not really incidence in UC patients in general, it was incidence rate detected in active UC patients who was hospitalized in our tertiary hospital. Whether there is a real difference in the incidence of CDI between Asian and Caucasian populations needs further research. The demographic and clinical characteristics factors did not differ between the case and control groups in our study. The results reflect the difficulty in differentiation of CDI from UC flare clinically, and this is a challenge of clinical care, which reminds clinicians to screen for
C. difficile at every flare in patients with UC.
It has been reported that risk factors for CDI in patients with UC are different from those of non-UC patients. Advanced age has been considered as a risk factor for CDI, [
7] while in our study, age and gender, along with the year of stool testing were defined as matched factors and were excluded from the analysis, which could also have excluded relevant confounders. Most patients without IBD (53%) had nosocomial CDI, [
17] while it is reported that most cases of CDI in IBD patients appear in the community [
18,
19]. In the present study, 67.6% of the cases had been admitted during the previous months versus 44.1% of the controls, so hospitalization may still be a risk factor for CDI in patients with UC. The most important risk factor for CDI is antibiotic therapy in the general population. Hensgens et al. found that antibiotic use increased the risk for CDI during therapy and in the 3 months after cessation of antibiotic therapy. The highest risk for CDI was found during the first month after antibiotic use [
20]. However, some researchers have reported that antibiotic use prior to admission is not an independent risk factor for CDI in patients with IBD [
21,
22]. In this study, we found a significant increase in CDI within 3 months after cessation of antibiotic use, while antibiotic exposure within 1 month prior to admission did not increase the risk for CDI. While this result was come from the univariate analyses that might have confounding factors, multivariate analysis didn’t reveal that antibiotic exposure was associated with CDI, which was consistent with some literatures. At the same time, the retrospective nature of the research may cause recall bias because of missing of some data in medical record.
Glucocorticoids and immunomodulators may also be independent risk factors in patients with CDI [
23]. In a large population-based cohort of IBD patients, corticosteroids were associated with a 3-fold increase in the risk of CDI compared with other immunomodulator or biological agents, irrespective of dose and duration [
24]. Maintenance immunomodulators use is independently associated with CDI in IBD [
5]. In this study, we found that systemic use of steroids was risk factor for CDI. Most patients in this study were treated with mesalamine, while immunosuppressive and anti-tumor necrosis factor agents were used in only a few patients, and their role in CDI needs further research.
Our study showed a definite correlation between active CMV infection and CDI in UC patients. CMV colitis has usually been reported co-infected with
C. difficile among immunocompromised patients, who have human immunodeficiency virus infection, organ transplantation, malignant diseases, or IBD especially when receiving immunosuppressive agents [
25]. CMV colitis can mimic pseudomembranous colitis in patients with immunocompromise status [
25]. A history of CMV infection is reported as a significant risk factor for
C. difficile infection [
26]. Cases of concomitant CDI and CMV colitis in immunocompetent patients have also been reported. Coexisting CMV and
C. difficile colitis may be refractory to first-line antibiotics for CDI treatment [
27,
28]. In the present study,
C. difficile and CMV co-infection had more severe colonoscopic features than CDI alone had. 2 cases of CDI recurrence were both co-infected with CMV. Co-infection with CMV may worsen CDI and influence the outcome of CDI in patients with UC.
It is reported that the recurrence rates of CDI were similar in patients treated with metronidazol and vancomycin [
29]. However, 2 patients with recurrent CDI in this study were both con-infected with CMV, which is more serious clinically, and were both treated by metronidazol for CDI when infected the first time. It is reported that severe CDI may be more likely to cause recurrence, [
30] and for severe CDI, vancomycin should be used as a first line [
29]. Our research is a retrospective study, and more reliable conclusions should be clarified by further prospective studies.
Advances in medical management have led to a decreased colectomy rates in patients with UC in recent years, which was 16% within 10 years of diagnosis [
31]. While CDI is associated with an increased risk of colectomy [
32]. A meta-analysis reported that CDI was a significant risk factor for colectomy among patients with IBD, especially those with UC [
33]. In the present study, although there was no significant difference between non-CDI and CDI groups, the rate of colectomy in the CDI group was higher than in the controls (27.3% vs 16.2%). The retrospective nature of our study and different follow-up periods may have caused recall bias.
There were several limitations to our study. First, testing methods for C. difficile mainly consist of nucleic acid amplifications tests (NAATs) for toxin genes or EIAs for toxins in the stools. We used an EIA for the diagnosis of CDI, which detects the presence of CDAB, and has low sensitivity compared to that of NAATs. It may have underestimated the rate of CDI in our UC patients. Second, this was a retrospective study, and not all patients admitted for active UC were tested for CDI especially before year 2012 when Chinese consensus on IBD management was issued, the exact incidence rate needs further research. Third, all patients enrolled were from a tertiary IBD center, who were patients with severe disease and the study results may not represent the overall CDI in Chinese UC patients. Fourth, though we thought to collect as much information as possible to observe the outcomes for CDI, the unequal follow-up window for patients may cause deviation. We considered the outcomes might be related to multiple factors such as UC activity and reaction to treatments. One additional limitation to this study is the lack of data about CMV detected on biopsies. On the other hand, we not only included patients with CMV colitis (diagnosed by tissue IHC), but also included patients with CMV viremia or reactivation (diagnosed by plasma DNA/pp65 or IgM) or so-called CMV infection. A prospective study and CDI test for all patients with UC flare are necessary to assess risk factors in these patients.