CDI has become a major public health problem. In United States, CDI causes about 450,000 infections and 35,000 deaths each year and C. difficile is also the major cause of hospital acquired infections in UK[19, 20]. Aging population and widely use of antibiotics are the current social situation in China, which can increase the risk of CDI. According to one meta-analysis by our early work, the incidence of CDI among hospitalized diarrhea patients is about 14% in mainland China[4]. Among IBD patients, the infection rate of C. difficile has been increasing steadily over the past decades[8, 21]. IBD is proven to be an independent risk factor for CDI. Moreover, IBD populations complicated with CDI usually have worse outcome[22]. In this study, the incidence of CDI among IBD patients was 24.6% (33/134) and was much higher than matched non IBD group. According to our knowldege, the infection rate of C. difficile in patients with IBD most studies reported in western countries is less than 10% and was much lower than this investigation[23, 24]. The method they used to detect CDI was toxigenic culture or an algorithm, including glutamate dehydrogenase (GDH) and C. difficile toxin A or B (CDAB) with nucleic acid amplification testing (NAAT) as confirmatory test, respectively. The method we used in this study was also toxigenic culture. The rates of CDI between our study and thier researches are comparable. However, while camparing the incidence of our study with other data in China, only few studies about the epidemiology of CDI among IBD patiens in China were searched. One study from Peking Union Medical College Hospital reported that thay identified 60 (7.41%) cases of CDI among 810 patients with IBD[25]. However, the method they used to detect CDI is enzyme immunoassay (EIA)-based stool test results for CDAB. It is widely reported that the low sensitivity of this method detecting free toxin in stool ranges from 32%–79%[26–28]. This may leads to the incidence of CDI on the low side. In our study, we also used VIDAS® C. difficile panel (BioMérieux, France) to detect partial diarrhea samples. Among 29 cases of toxigenic C. difficile diarrhea samples detected by toxigenic culture, only 6 cases were positive and 2 cases were suspected for CDAB from diarrhea samples detected by EIA method. The rest test reselts were all negative. The sensitivity of CDAB for CDI detection compared with toxigenic culture is only 27.6% (8/29), even including the suspicious positive cases. Although the number of CDI cases included in this study for CDAB test was small, it could also reflect the issue to some extent. So, the real incidence of CDI in hspitalized diahhrea patients with IBD may be higher in China and our data could fill the gap in this field to a certain extent. In this study, about two thirds of CDI among IBD patients were community acquired infections and nosocomial infections only accounted for 33.3% (11/33). According to the research of Rodemann JF and his colleagues, nearly two-thirds of IBD patients with CDI were community acquired infection, which was consistent with our result[21]. Community acquired infection accounts for the majority of CDI among diarrhea inpatients with IBD.
In addition, no binary toxin positive strains were isolated in this study and all toxigenic strains were both positive for tcdA and tcdB and there was no report about the hypervirulent strain RT027 isolated from IBD patients in China as far as we know. ST54 was the main type of toxigenic C. difficile among diarrhea patients with IBD. After MLST analysis, we discovered that ST54 and ST3 accounted for nearly half of the toxigenic isolations from patients with IBD in this study. Meanwhile, many literatures reported that ST54 was widely distributed in the world and was detected in Japan, India, Chile and other places[29]. It has also been reported that ST54 occupies the second place of CDI in domestic pregnant women[30]. Recently, investigations carried out in Chinese hospitals revealed that ST54 was the predominant type of CDI in diarrhea patients[31, 32]. Also, according to our previous meta-analysis about the domestic molecular epidemiology, the epidemic strain of C. difficile isolated from diarrhea patients were mainly ST3, ST37 and ST54, accounting for 18.1%, 17.2% and 16.7%, respectively. This indicates that ST54 strain plays an important role not only in IBD patients, but also in other populations.
It is widely acknowledeged that broad-spectrum antibiotics, proton pump inhibitors, older age, immunosuppressants and long-term hospitalization are all related to the occurrence and development of CDI[33–35]. However, the incidence of CDI among patients with IBD is still very high even if they have not used antibiotics recently and have not been exposed to the medical environment[36]. Compared with the general population, IBD seems to be an independent risk factor for CDI. This may be caused by the imbalance of gastrointestinal and immune functions, intestinal flora disorders, intestinal epithelial damage and other factors in IBD patients. The differences and associations about the risk factors for CDI between IBD patients and the general population deserve further study. In this study, the independent risk factors for CDI among IBD patients were the recent use of quinolones and the first year after initial diagnosis of IBD. In this study, levofloxacin was the mainly uesd quinolones. Quinolones have been widely recognized as an independent risk factor for CDI and our study confirmed that quinolones remaind a risk factor for CDI among diarrhea patients with IBD. Singh H and his colleagues reported the highest rates of CDI within the first year of IBD diagnosis and shorter duration of IBD was associated with an increased risk of CDI[13]. This may be explained by the higher rates of dysbiosis of gut flora at IBD diagnosis, untreated altered humoral immune responses, or epithelial dysfunction predisposing to CDI. Interestingly, many studies about the independent risk factors for CDI in IBD patients discovered that steroids, biologics and immunomodulators, which have been widely considered as risk factors for CDI, have not been associated with the higher incidence of CDI in IBD patients[24, 37, 38]. In our study, systemic steroids were not associated with CDI susceptibility in patients with IBD. There is no consensus on the independent risk factors for CDI among IBD patients at present. Different studies often come to different conclusions, such as recent use of antibiotics, admission to hospital, PPI and so on[23, 24, 37, 39].
Also, this study did not find that IBD patients complicated with CDI extended the length of hospital stay, which was consistent with the results of multiple single-center studies[8, 40]. One investigation carried out in USA discovered that more than half of the infected IBD patients required hospitalization and 20% required colectomy. However, only several cases in this study accepted colectomy. Two studies from Europe reported the rate of colectomy was about 5%. There is still controversy over whether CDI increases the rate of colectomy among patients with IBD. High quality researches are needed to resolve this issue.
What’s more, almost all infected IBD patients who received CDI treatment were recovered from diarrhea in this study (12/13) while almost half of the untreated CDI patient still had issues with stool (9/20), even 5 cases still had diarrhea when they were at discharge. This indicates the conventional therapy of IBD does not apply to patients complicated with CDI and emphasizes the importance of early diagnosis and targeted treatment against CDI.
Despite the discoveries in this study, there are some limitations should be considered. Firstly, this single-center study was conducted in a tertiary care university teaching hospital and the sample size enrolled in this study was insufficiently large. For example, when analysis the clinical characteristics of IBD patients with or without CDI, cases of some variables are limited. In-depth subgroup analysis about the incidence of CDI among different characteristics of IBD could not be performed. Secondly, not all diarrhea amples used CDAB to detect CDI because this panel was used in our hospital not from the beginning of this study and we could not compare the incidence of CDI with the data from the Beijing, China[25]. Addiotionally, not all diarrhea inpatients sent stool for CDI detection within 48 h after admission and the true proportion of community-acquired CDI maybe higher than the data we received. Last but not least, we did not follow up the enrolled patients because part patients were from other cities and lost to follow up.