Molecular epidemiology of Clostridioides di cile in patients with in ammatory bowel disease in China

Tao Lv Zhejiang University School of Medicine First A liated Hospital Yunbo Chen Zhejiang University School of Medicine First A liated Hospital Lisi Zheng Zhejiang University School of Medicine First A liated Hospital Tao Wu Zhejiang University School of Medicine First A liated Hospital Ping Shen Zhejiang University School of Medicine First A liated Hospital Lanjuan Li (  ljli@zju.edu.cn ) Zhejiang University School of Medicine First A liated Hospital


Introduction
Clostridioides di cile (C. di cile) is one of the leading pathogens of infectious and nosocomial diarrhea and pseudomembranous colitis. [1,2] The incidence and severity of global C. di cile infection (CDI) have steadily increased over the past decades. Furthermore, CDI is no longer restricted to the nosocomial setting. High CDI frequency has been reported in the historically low-risk group and residents that include children who have not been exposed to hospital settings. [3,4] CDI incidence in the general inpatient population in China is approximately 10.0%, which similar to that in developed countries, but differs from the ribotypes, and no outbreak of the BI/NAP/027 strain has been reported. [5][6][7][8][9] However, the number of C. di cile multidrug-resistant (MDR) strains continues to increase, and a much higher MDR rate has been identi ed among the toxigenic isolates obtained from hospitalized patients in China. [5,8] In ammatory bowel disease (IBD), including Crohn's disease (CD) and ulcerative colitis (UC), is a chronic idiopathic disorder that features intermittent in ammation in the gastro-intestinal tract. [10][11][12][13][14] In the early 1980s, IBD was rst found to be accompanied by CDI. [15] Subsequently, several studies suggested that IBD patients carried a higher CDI risk as a result of the intestinal dysbiosis related to IBD, when compared to the general population. [10,11,14] Once CDI complicating IBD, IBD patients tend to have higher recurrent onsets that require repeated hospitalizations, and intensi ed antibiotic treatment and poorer outcomes, leading gastrointestinal surgery and higher mortality, when compared to mono-IBD. [16,17] Retrospective studies have revealed almost a two-fold higher risk for colectomy and a 4-6 fold increase in mortality risk among IBD patients with CDI. [18,19] The CDI incidence in IBD patients varied in different studies, re ecting the differences in both geographical regions and ethnic participants. [16,20] The CDI incidence in IBD patients was moderate in Western countries. [10] Patients hospitalized for IBD are also presented with CDI, which ranged from 3.7-8.0%. [16,17] However, as high as 13.9% has been reported for CDI in Chinese IBD patients. [21,22] The major risk factors for CDI in the general population include the excessive or non-standard use of antibiotics, older age, prior and prolonged hospitalization, and multiple comorbidities. [23] However, CDI may occur in IBD patients, regardless of the younger age and absent antibiotic exposure. CDI in IBD patients was associated with corticosteroid exposure, and CDI in the majority of hospitalized IBD patients was community-acquired. [16,18,24,25] Globally, CDI is typically more common in adults, when compared with children, and the clinical features, which include the outcomes in UC and CD patients with CDI, have been well-studied. [11,16,[23][24][25] However, few studies have investigated CDI in pediatric IBD patients and characterized the molecular features of C. di cile in adult and pediatric IBD patients, especially in China. The present study aimed to investigate the molecular epidemiology of C. di cile in pediatric and adult IBD patients in China.

De nitions
Bacterial isolates Unformed stool samples obtained from adult IBD patients (between June 2011 and June 2015) and pediatric IBD patients (between November 2011 and September 2017), who were suspected of CDI, were sent to the Clinical Microbiology Laboratory at the First A liated Hospital of Zhejiang University (Hangzhou, China). All specimens were cultured on selective media cycloserine-cefoxitin-taurocholate agar (CCFA-TA; Oxoid) supplemented with 7.0% sheep blood at 35 °C for 48 hours. Then, the C. di cile isolates were con rmed by matrix-assisted laser desorption ionization time of ight mass spectrometry (MALDI-TOF MS) analysis using the Bruker Daltonics Micro ex LT system (Bruker Daltonik GmbH, Bremen, Germany).

Detection of toxin genes
Bacterial genomic DNA was extracted using the simpli ed alkaline lysis method. All isolated strains were tested for tcdA, tcdB and binary toxin genes by polymerase chain reaction (PCR), as previously described. [7,8] Multilocus sequence typing (MLST) and phylogenetic analysis MLST was used to genotype all toxigenic isolates, as previously described. [7,8] The allele designations were obtained through the C. di cile PubMLST batch pro le query page (http://pubmlst.org/cdi cile/). The phylogenetic tree was constructed using the MLST sequences (the maximum likelihood method with MEGA X [64-bit] software, http://www.megasoftware.net). Evolutionary distances were calculated using the General Time Reversible model and the bootstrap consensus tree generated from 1,000 replicates.

Antibiotic susceptibility testing and mechanisms of resistance
The antibiotic susceptibility of all toxigenic isolates was analyzed using the E-test strips (bioMérieux, Marcy-l'Étoile, France), which were coated with drug concentrations that ranged within 0.016-256.000 mg/L for metronidazole, vancomycin, clindamycin, erythromycin, tetracycline and linezolid, and 0.002-32.000 mg/L for cipro oxacin, moxi oxacin, levo oxacin and rifampicin. Brie y, these selected isolates were grown on Brucella agar plates that contained 1 mg/L of vitamin K, 5 mg/L of hemin and 5.0% sheep red blood cells. Then, the agar surface was covered with E-test strips and incubated in an anaerobic atmosphere for 24-48 hours at 37 °C. The minimum inhibitory concentrations (MICs) were read at the point at which the zone of complete inhibition intersected with the MIC scale with the Clinical and Laboratory Standards Institute (CLSI) guidelines. The selected resistance breakpoints included 8 mg/L for erythromycin, clindamycin, tetracycline and the uoroquinolones, and 32 mg/L for metronidazole, according to the CLSI interpretative categories approved for anaerobic bacteria. The 4 and 16 mg/L breakpoints were selected for rifampicin and vancomycin, respectively, according to the CLSI interpretive categories approved for Staphylococcus aureus, because no values were provided for anaerobes. The breakpoint for linezolid was set at 4 mg/L. The control isolate was C. di cile ATCC 700057. An isolate that resisted ≥ 3 classes of antibiotics was called multidrug resistance strain. [26] Data analysis All statistical analyses were performed using SPSS 20.0. A P-value of < 0.05 was considered statistically signi cant.

MLST types
A total of 18 sequence types (STs) were classi ed for toxin producing strains (Table 1)

Antimicrobial resistances
The MICs of different antimicrobial agents against C. di cile isolates are presented in Table 2. A 100% resistance to cipro oxacin was observed, followed by 92.7% to levo oxacin, 64.6% to clindamycin, 62.2% to erythromycin, 15.9% to tetracycline, 13.4% to moxi oxacin, and 3.7% to rifampicin. All isolates were sensitive to metronidazole, vancomycin and linezolid, showing that these were within a narrow range (Table 2). A signi cantly higher resistant rate to clindamycin (P = 0.009) was found in isolates obtained from pediatric patients, when compared to adult patients. No heteroresistance was observed. Each of all the tested isolates was resistant to at least one antibiotic. Speci cally, 37.8% of the 82 isolates, in which 26.8% were obtained from adult IBD patients and 11.0% were obtained from pediatric IBD patients, were resistant to only one or two classes of antibiotics, respectively, and as high as 62.2% of these isolates were resistant to ≥ 3 classes of antibiotics or MDR strains. Furthermore, 14.6% of the MDR isolates, in which 12.2% were from adult IBD patients and 2.4% were from pediatric IBD patients, resisted four classes of antibiotics (i.e. clindamycin, erythromycin, tetracycline, levo oxacin and cipro oxacin). Interestingly, the resistant rate to three classes of antibiotics in pediatric IBD patients with was nearly twice as many as that in adult IBD patients (i.e. clindamycin, erythromycin, cipro oxacin and levo oxacin), while the resistance to four classes of antibiotics in adult patients was almost four times higher, when compared to pediatric patients (Table 3) (Table 4). However, no signi cant difference in resistant rate to antibiotic classes by the same multidrugresistant ST strains was observed between these two groups.

Discussion
In the present study, the detected incidence of CDI in adult IBD patients was lower than that reported in China, but comparable to the CDI frequency reported in non-Chinese IBD patients. [16,17,21,22] The different compositions in the general inpatient population in different hospitals may have tilted the CDI frequency one way or the other. It has been suggested that the higher CDI in IBD patients could be partially explained by the increased CDI incidence in the general inpatient population. [11,16] It was also found that the CDI incidence in pediatric IBD patients was higher than that in adult IBD patients. As reported, IBD children had higher intermittent CDI, when compared to IBD adults. [27,28] It is known that 3.0-70.0% of children may carry an asymptomatic C. di cile. [29,30] It was no surprise to detect a higher CDI in IBD children due to the higher CD carriage background in this population. In addition, dynamic changes in microbiome in the gut of children represent another risk factor for CDI in pediatric IBD patients. [27] Previous studies have categorized ST-54, ST-35 and ST-37 as the top three genotypes isolated from inpatients with con rmed hospital-acquired CDI in China. [6,7,9]  McFarland LV et al. [33] suggested that CDI in children was more frequently community-acquired, when compared to adult CDI cases. However, it remains conceivable that CDI in pediatric IBD patients is mainly communityacquired, and that CDI in adult IBD patients is more likely hospital-acquired. However, further studies are required to ascertain the two different transmission sources of C. di cile responsible for CDI in children and adults. In addition, although there was no signi cant difference in the same STs based on the results of the phylogenetic analysis, different ST distributions were noted between adult IBD patients and pediatric IBD patients, but the difference was insigni cant. The differences in transmission source and age may have contributed to the different ST distributions. After all, the molecular epidemiology of C. di cile may be subjected to random, but dynamic changes. [34] Antibiotic use has been considered an important cause for CDI increase. Several classes of antimicrobials have been linked with high CDI risk. For instance, uoroquinolones have been referred as a common cause of CDI. [35] Multiple antibiotics were suggested to facilitate C. di cile spore germination and toxin production. [23] Antibiotic exposure is not essential in CDI occurrence in IBD patients. [11] However, antibiotic pressure may prompt C. di cile to acquire resistance to antibiotics, and confer survival properties and enhanced virulence. The newly acquired abilities by C.
di cile may explain recurrent course and poorer outcomes in IBD patients. [16,17] All isolates in the present study resisted cipro oxacin, and nearly all of these also resisted levo oxacin with the highest MICs (≥ 32 µg/ml − 1 ), resembling the hospital-acquired CDI in China. [5][6][7]9] However, merely 13.4% of isolates in IBD patients exhibited a resistance to moxi oxacin, which was signi cantly lower, when compared with the hospitalacquired CDI in China. The relatively low resistance was likely the result of the infrequent prescription of moxi oxacin for IBD patients. Meanwhile, it was found that the resistance to clindamycin in pediatric IBD patients was signi cantly higher than that in adult IBD patients (75% vs. 56.5%, P = 0.009). This difference likely resulted from the frequent prescription of clindamycin for children, re ecting a fact that a greater number of children catch u with lung in ammation, when compared to adults. In addition, ST-3, ST-35, ST-37 and ST-54 were the prevalent types responsible for clindamycin resistance in pediatric patients. Similar resistant genotypes were identi ed, and these included nearly 70.0% of the isolates that were resistant to clindamycin, including ST-35 (90.9%), ST-54 (89.0%), ST-37 (85.5%) and ST-3 (77.1%), in the previous studies conducted by the investigators. [7,8] In the present study, the MDR strains in adult IBD patients was 52.2%, which was similar to the previously reported 47.1%-58.6% in the inpatient population in China, but signi cantly lower than the reported 75.0% in pediatric IBD patients. [7,8] The MDR strains resistant to the three classes of antibiotics of clindamycin, erythromycin and uoroquinolones accounted for 69.4% of all MDR strains identi ed in pediatric IBD patients, which was signi cantly higher than that in the adult IBD patients in the present study. Thus, clindamycin, uoroquinolones and cephalosporins have been identi ed as the greatest risk for CDI in children, since many children with community-acquired CDI were exposed to multiple classes of antibiotics in the 12 weeks preceding CDI. [28,36] The MDR rate in all STs and RTs in CDI patients in eastern China was high, which was 14.9-19.6% for ST-2/ST-3 and 17.0% for RT001. [5] As shown by the present study, ST-35 is an additional type of toxigenic C. di cile MDR strain in IBD patients in eastern China.

Conclusion
According to the results, the incidence and molecular epidemiology features of CDI in adult IBD patients resembled in the general inpatient population. There were signi cant differences in the distribution of STs and antibiotic resistance rates between adult IBD patients and pediatric IBD patients. C. di cile isolates in pediatric IBD patients had a relatively higher resistance. However, the MDR strains were limited, regardless of the comparable phylogenetic features of C. di cile between the two groups. The high resistance rate accompanied by the dominant MDR strains in pediatric IBD patients call for more effective measures to reduce CDI in children.

Declarations
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.

Availability of data and materials
All data generated or analyzed during this study are included in this published article.

Competing interests
The authors declare that there are no con icts of interest. LGC20C010002.
Author's contributions LjL contributed in the experimental studies, and drafting the work, YbC and PS contributed in the conception of design, LsZ contributed in the acquisition of the data. TL, and YbC contributed to the conception of design and revising the draft, TW contributed to the analysis of the data. All authors read and approved the nal manuscript