Nosocomial transmission of Clostridium difficile ribotype 027 in a Chinese hospital, 2012–2014, traced by whole genome sequencing
- 1.3k Downloads
The rapid spread of Clostridium difficile NAP1/BI/027 (C. difficile 027) has become one of the leading threats of healthcare-associated infections worldwide. However, C. difficile 027 infections have been rarely reported in Asia, particularly in China.
In this study, we identified a rare C. difficile bloodstream infection (BSI) from three isolates of a patient during repeated hospital admission. This finding triggered a retrospective epidemiological study to scan all cases and strains emerged from this ward during the past three years. Using medical personnel interviews, medical record reviews and the genomic epidemiology, two outbreaks in 2012 and 2013–2014 were identified. Through using whole genome sequencing, we succeeded to trace the origin of the BSI strain. Surprisingly, we found the genome sequences were similar to C. difficile 027 strain R20291, indicating the occurrence of a rare C. difficile 027 strain in China. Integrated epidemiological investigation and whole genome sequencing of all strains, we constructed a nosocomial transmission map of these two C. difficile 027 outbreaks and traced the origin of the infection.
By genome sequencing, spatio-temporal analysis and field epidemiology investigation, we can estimate their complex transform network and reveal the possible modes of transmission in this ward. Based on their genetic diversity, we can assume that the toilets, bathroom, and janitor’s equipment room may be contaminated area, which may be suggested to improve infection control measures in the following health care.
KeywordsClostridium difficile NAP1/BI/027 Outbreak Nosocomial infection Whole genome sequencing
Clostridium difficile infection
hospital acquired infection
multi-locus sequence typing
the North American Pulsed-field type one
single nucleotide polymorphism
whole genome sequencing
Clostridium difficile is an anaerobic, gram-positive, spore-forming bacterium responsible for infections ranging from mild diarrhea to pseudomembranous colitis, primarily in elderly patients exposed to antibiotics. Recently, this bacterium has become one of the most frequent microorganisms responsible for healthcare-associated infections in the United States . During 2003, the first outbreak of a strain with “hyperproducer ” of toxin A and B, referred to as the North American Pulsed-field type one (NAP1), restriction-endonuclease analysis type BI, or PCR ribotype 027 (NAP1/BI/027), was reported in North America [2, 3, 4, 5]. Since then, cases have been reported worldwide. Nevertheless, large outbreaks of NAP1/BI/027 are less reported in Asia and Latin America compared to North America and Europe [6, 7]. However, cases of NAP1/BI/027 infection have been already described in Hong Kong and Guangzhou, China [8, 9]. NAP1/BI/027 is characterized by an in vitro overproduction of toxins A and B as well as a binary toxin production, and an 18 bp deletion in tcdC gene . Various epidemiological differences are noted in Asia; for example, the most common circulating ribotypes are 014, 017, and 018 . Moreover, the toxin A-negative, B-positive (A−B+) strain 017 has received wide attention [6, 11]. To the best of our knowledge, no report of C. difficile infection (CDI) outbreaks in China has been published, and even few of case reported so far in the English literature.
In the present study, we report the first NAP1/BI/027 outbreak in a hospital in mainland China, traced by a rare case of bloodstream infection (BSI) using whole genome. This process of identifying outbreak was dramatic by a comprehensive retrospective investigation using temporal and spatial analysis of the cases and strain characterization including PCR-ribotyping, multi-locus sequence typing (MLST) and whole genome sequencing (WGS). The identification of the outbreak immediately provides four key implications: 1) the threat of CDI in China is more serious than that previously believed because of the under diagnosis of CDI in hospitals, and more cases were ignored; 2) the epidemiology of CDI should be carefully investigated to check whether there have been more cases of NAP1/BI/027 infection ; and 3) the infection control strategy should be strengthened. 4) WGS method may be helpful for investigating outbreaks and tracking strain transmission .
Case definition and CDI diagnosis
This study was conducted in the department of Traditional Chinese Medicine respiratory, a hospital of China from March 2012 to April 2014. A CDI case-patient was defined as a patient with 3 or more unformed stools in 24 or fewer consecutive hours and a stool test resulted positive for C. difficile toxin . Stool samples were tested for toxin by enzyme immunoassay (C. difficile toxin A/B, Techlab, Blacksburg, VA, USA), and were anaerobically cultured on cycloserine-cefoxitin-fructose agar. Suspected colonies were identified by agglutination with C. difficile latex reagent for the somatic antigen (Oxoid Ltd, UK).
Retrospectively epidemic investigation
A comprehensive retrospective epidemiological investigation was conducted. First, clinical information from medical records, including hospital stays, admitting diagnoses, antibiotic use, and occurrence of diarrhea were collected. Using these retrospective case searches, in total, 22 isolates were collected. Among 75 patients with diarrhea, 20 were diagnosed as CDI , including one patient with a C. difficile bloodstream infection. Symptom onset occurred after 48 h of hospitalization in 18 patients while 2 patients presented with diarrhea when admission. Second, the architectural structure of the ward and room arrangement were collected by field epidemiology investigation; at the same time, clinicians, nurses, and nurse-assistant were questioned about their behaviors during healthcare, paying particular attention to the overlap of hospital stays, use of the toilet and bathroom, and cleaning procedures.
Validation and typing the strains in two outbreaks
In house-PCR was used to detect tcdA and tcdB, using the primer sets according to Lemee’s study , whereas MLST was used to allele-typing seven housekeeping genes, including adk, atpA, dxr, glyA, recA, sodA and tpi . The isolates were characterized by a capillary gel electrophoresis-based ribotyping method using the NAP1/BI/027 strain UK1 as the reference .
Genetic diversity, whole genome sequencing and bioinformatics analysis
A 500 bp library was constructed and whole genome sequencing was performed using Illumina Hiseq 2000 platform (Illumina, Inc., San Diego, CA, USA) to obtain a 100 bp paired-end read, with 1,463,765,173 bps in average for each sample (Accession number: SRP051490, Additional file 1: Table S2) . The multi-alignment sequences from read data and references (ST1/027: R20291, accession number: NC_013316; ST37: M68, accession number: NC_017175) were obtained by the method that combines alignments from mappings to multiple reference sequences via REALPHY [18, 19], and then the phylogenetic tree was reconstructed via Bayesian evolutionary analysis by BEAST . The genomic single nucleotide polymorphisms (SNPs) between NAP1/BI/027 strains were detected by mapping the C. difficile NAP1/BI/027 reference genome R20291 using SOAP2 software . The sites with quality of <60 were removed. These SNPs were finally used to reconstruct the phylogenetic tree via Bayesian evolutionary analysis by BEAST . The mutation pattern of NAP1/BI/027 was clustered by a hierarchical cluster algorithm using Cluster/Treeview software .
Identification the BSI case of C. difficile NAP1/BI/027 infection
The patient was admitted on August 12, 2013 for community-acquired pneumonia. He was treated by cefepime (2000 mg 2 times per day) and levofloxacin (500 mg 1 time per day) intravenously for 7 days. On August 19, he displayed diarrhea, which progressively worsened. On August 22, his symptoms included fever, cramping, and abdominal discomfort. His stool sample was tested positive for the C. difficile toxin and culture. We simultaneously obtained his C. difficile isolate from an anaerobic blood culture. He was successfully treated with oral metronidazole (400 mg 3 times per day) for 13 days and then sequentially intravenous metronidazole (500 mg 3 times per day) for 4 days, and he was discharged on September 16. He was re-admitted three days later with persistent fever and diarrhea. Stool sample was positive again for toxin and culture but the blood culture was negative. He was treated with oral vancomycin (500 mg 3 times per day) for 12 days, and discharged on October 8. Clinical information and laboratory testing results of this patient (patient ID is P13) have been described in Additional file 1: Table S1.
Retrospective study of clinical information of patients with CDI in the ward
Description of the patients with C. difficile infection
No. of patients (%)
Age range (mean)
No. (%) of patients with nosocomial diarrhea
Length of hospital stay [range, days (mean)]
No. (%) of patients treated before diarrhea with the following
4th generation cephalosporins
3rd generation cephalosporins
Penicillins + β-lactamase inhibitor
Genetic population analysis in outbreaks
Whole genome analysis of NAP1/BI/027 isolates
After mapping the reads of NAP1/BI/027 (ST1) isolates to the reference genome R20291, a total of 62 SNPs were identified in our NAP1/BI/027 strains, including T1476G and G1514A located in rpoB gene associated with fluoroquinolone-resistance (Fig. 3b, Additional file 1: Table S4). All sequences from P13 were tightly clustered together, suggesting that these occurrences of CDI (once for BSI and twice for diarrhea) were related as we assumed (Fig. 2). The route of genetic transfer is clearly from 13A to 13B then to 13C, indicated by the genetic pattern with one SNP for each strain, although strain 13A presented an abnormal mutation. The results of WGS analysis confirmed our subsection that strain 13B causing BSI might have originated from the gut (Fig. 3b). The isolate 13C, responsible for CDI recurrence, was nearest to 13B from his blood. However, this hypothesis ignores the presence of within host diversity. The other possible alternation is that 13A, 13B and 13C may have existed simultaneously in the gut at both time points, and 13B caused BSI at the initial time. Since we cannot obtain multiple colony picks in this retrospective study, we empirically estimate that the gut infection lead to secondary BSI and gut persistence explains the relapse.
Construction of the transmission map of NAP1/BI/027 for this outbreak
To reconstruct a putative transmission map, the clinical epidemiologic data were combined with the genetic data from results of WGS (Fig. 4). First, the temporal and spatial transmission links were constructed. The patients with overlapping time spent in bed were considered as the potential transmission connections (Fig. 4). Second, by conducting the field epidemiological investigation and consulting with heath care workers and patients, we assumed that the toilet and cleaning procedures were potential at high risk of transmission. Indeed, only one commode in the toilet and one mop were available in the ward, increasing the possible risk of the spore transmission among the patients. Last, we constructed the potential connections between the patients by genetic relationship among these strains. By combining all the data, particularly the genetic relationships between isolates and possible contacts between patients, we reconstructed the transmission routes of all NAP1/BI/027 cases to explain the outbreaks. Furthermore, our study indicated a trend between decreasing number of CDI and increasing distance to the janitor’s equipment room and toilet (Fig. 4).
Genome sequences provide the detailed molecular evidence to connect part of the transmission routes. First, phylogenetic analysis showed all NAP1/BI/027 isolates in this ward belonged to the same clone which existed in the ward in three years and caused two outbreaks, and we provided clear clues for explaining these outbreaks. Second, the isolates in these potential transfer routes shared most of the 62 SNPs, resulting in the direct connection among them (Fig. 4). For example, P3 only contains one SNP difference with P1, and this result indicates that these two patients may be infected by one strain or the pathogen may transfer from one to another. Similar results appeared in P7, P4 and P2, P5 and P6. Finally, WGS data also validate that the strains in two outbreaks in consecutive years were only different by 2 stable SNPs, indicating the infected strain may exist in the hospital and transfer again in the following year (Fig. 4). Based on the sequence diversity of our data set, the estimated time to the most recent common ancestor of the C. difficile strain was December 2011 (Additional file 1: Figure S2).
The potential connection of isolates according to time, space distribution, and molecular clues were integrated to define the full map of transmission routes (Fig. 4).
To our knowledge this is the first study investigating outbreaks of healthcare-associated C. difficile NAP1/BI/027 strain infection in China and reconstructing the transmission map by combining retrospective epidemiological analysis with whole-genome sequencing. The outbreak caused by this strain has not been reported in China, especially via bloodstream infection, which has high mortality. We provide evidence that WGS method could be used to tracing the source of infection and predicting the transmission routes.
C. difficile was the most commonly reported pathogen, accounting for 12.1 % of healthcare-associated infections . The increased severity of CDI is primarily attributed to the emergence and rapid spread of the fluoroquinolone-resistant NAP1/BI/027 strain, which remains the most common clone worldwide . Since the first large outbreak described in North America in 2003, the NAP1/BI/027 has become endemic in many hospital from North America, Europe, and more recently in Australia [2, 4, 27]. While, for a long time, lacking of the reports of CDI caused by the strain, makes us confusing about the states of 027 in China. Three major reasons may have contributed to this situation: 1) the major diagnostic method for CDI available in Chinese clinical laboratories is EIA for toxins A/B, and very few hospitals can perform C. difficile culture, resulting in many undetected CDI cases; 2) the lack of anaerobic laboratory equipment and limited CDI knowledge by healthcare staff also contributes to the poor isolation compliance to infection control measures and reporting under diagnosis of CDI in China; and 3) there is no reference center for typing C. difficile strains in China and identification of the different PCR ribotypes requires the collection of reference strains. Therefore, distribution of C. difficile circulating clones including the NAP/BI/027 is not known in China. PCR based sequencing method can detect the pathogen without reference strain by comparing with the public database. Thus, it becomes one of replacement to other method in lots of bacteria. However, not like other hospital acquired infection (HAI) bacteria, such as Mycobacterium tuberculosis and Klebsiella pneumonia , C. difficile is characterized by large genome diversity and instability increasing the difficulties to select proper primers for detection. This characteristic will increase the difficulty to extend the existed molecular typing methods used in US or UK to develop country, and easy to cause clinic false-negative results. So, the fact may be that: NAP1/BI/027 has existed in China before, yet it could not be identified immediately. Whole-genome sequencing can easily distinguish minor differences, providing valuable information for molecular typing, and pathogen or transmission route tracing [12, 29]. Thus, the genome complication would boost the application of WGS in C. difficile research. With the help of WGS, our research will enrich HAI pathogen database, and HAI outbreak tracing clues. Also, our research revealed the potential NAP/BI/027 threat for China will improve the completion of C. difficile NAP/BI/027 global epidemiology distribution [6, 30, 31].
In this study, we investigated CDI outbreaks using medical personnel interview, medical record review and WGS for strain typing. From the spatial analysis of the cases, we assumed that CDI patients are mainly concentrated near to the toilets, bathroom, and janitor’s equipment room (Fig. 4), which may be suggested to improve infection control measures in the following health care. However, we failed to obtain the isolates in environment. We also conducted special consultation with nurse-assistant and determined that the chlorine-containing disinfectant concentration (500 ppm available chlorine) used in the ward was lower than the concentration recommended for CDI infection control (with at least 1,000 ppm available chlorine). The epidemiological investigation revealed the absence of a dedicated area for CDI patient isolation, which was probably the most important contributor to the outbreak. P13, the first patient identified during the 2013–2014 outbreak, was placed in different hospital beds in the days following the emergence of diarrhea. In addition, this patient was neither isolated in a single room nor placed in contact precautions. To the best of our knowledge, person-to-person transmission was the main route of CDI transmission. However, the outbreak occurred during a long period spanning from 2012 to 2014, suggesting the existence of a potential environmental reservoir. For example, the hospitalized periods of patient four and seven had no overlap, and their rooms were quite apart away from each other (Fig. 4), showing that there was no obvious epidemiological connection by conventional methods. However, the WGS analysis revealed that there was only one SNP difference between isolates 4 and 7 from these two patients (Fig. 3b), which provides an important epidemiology clue and help to form epidemiology hypothesis. So, the method of WGS shows obvious potential superior to routine typing methods in epidemiology tracing. For example, we may confirm the transfer route of P13 and the clear route of P16, P17, and P18, which occurred during in the 2013–2014 outbreak. Furthermore, isolated strains from most cases in the outbreaks were sequenced belonging to ST1, and there were also several strains belonging to other STs with sufficient genetic diversity to ST1 in this study. These genetically distinct isolates may indicate diverse alternative sources or more complicated transmission modes, since C. difficile is an ancient organism with diverse strains present in humans, animals, and food. It is also noted that asymptomatic carriers might not be detected, which would have contributed to the transmission. The literature review revealed that approximately 50 % of the molecular resulting clusters can be confirmed by contact tracing documenting transmission links, particularly for a longitudinal time, space, and population span, which are difficult to establish using classic epidemiological links [32, 33]. Therefore, the molecular epidemiology based on WGS should be recommended for HAI investigations, in addition to conventional field epidemiology investigations.
Since we never found NAP1/BI/027 strain in the hospital before, this study is not carried out under a standard surveillance in the hospital, especially lack of pathogen information from environmental monitoring. All clinical trials were drawn back to collect after we identified and validated the strain, the complete transfer route of all strains in hospital is still not very clear. However, to our knowledge, this is the first case to identify C. difficile NAP/BI/027 outbreak by WGS method, which can be extended to other bacteria.
One of the most interesting features of the epidemic NAP1/BI/027 is high level of antibiotic resistance . In the present study, we found the outbreak strain in China is consistent but different to that reported in other countries (Additional file 1: Table S4). For example, the rpoB gene presents the special mutations in T1476G and G1514A compared with R20291, in addition to the previously fluoroquinolone-resistance site reported . In our cases, most patients were from the local community and were mainly admitted for chronic obstructive pulmonary disease and chronic bronchitis. The misuse of antibiotics, including the overuse and inadequate use of antibiotics, actually increased the antibiotic pressure and risk for CDI. In China, before 2012, patients suffering these diseases were usually prescribed antibiotics, particularly cephalosporins and fluoroquinolones. Additional file 1: Figure S1 shows that CDI patients were more likely to be treated with third generation cephalosporins and fluoroquinolones than patients negative for CDI in the ward at the same time. The results are in accordance with the previous publications . Therefore, these observations highlight at least two noteworthy points: 1) primary medical clinicians should be aware of the potential CDI threat before the use of cephalosporins and fluoroquinolones to treat chronic obstructive pulmonary disease and chronic bronchitis of the elderly population and 2) the related family as well as healthcare workers caring for these elderly patients may also have potential C. difficile exposure risk, which should be particularly addressed. As we mentioned above, our cases emerged after the strictest antibiotic regulation has been implemented by the Chinese government in 2012. So, the regulations should be further enforced . In addition, the existence and outbreak of C. difficile in our research may underscore the importance of monitoring antibiotic proper usage and taking preventive action to reduce transmission of such multidrug-resistant organisms in the hospital setting, thereby preventing nosocomial outbreaks. Thus, we propose that C. difficile surveillance must be conducted in health cares in China, Asia and other developing country, as an early warning of antibiotic resistance, allowing time for related detailed antibiotic interventions to be designed and implemented.
By genome sequencing, spatio-temporal analysis and field epidemiology investigation of the C. difficile NAP1/BI/027 outbreaks in 2012 to 2014 in this ward, we can estimate their complex transform network, revealed the possible modes of transmission. Based on their genetic diversity, we can assume that the toilets, bathroom, and janitor’s equipment room may be contaminated area, which may be suggested to improve infection control measures in the following health care.
This work was supported by the National High Technology Research and Development Program (“863” Program, grant no. 2014AA021505 and 2011AA02A109), the National Key Technology R&D Program (grant no. 2012BAH24F05), and the National Natural Science Foundation of China (grant no. 81201322 and 81301402).
Availability of data and materials
The sequencing data described in this study have been deposited in the GenBank database (Accession number: SRP051490), and the sequencing methods and background information of isolates have been described in this article.
CC, YC, YC and HJ designed the project. CC, YC, HJ, PD, HY, YZ, FB and GG wrote the manuscript. HJ, HY, JW, and GH performed the strain isolation and molecular typing. YC, ZY, JZ and QD performed the epidemiological study. YZ, YQ performed the whole genome sequencing. PD, WZ, HW, NH, ZW performed the bioinformatics analysis. All authors have read and approved the final manuscript.
Dr. Hongbing Jia is the Chief of Division of Microbiology, Department of Clinical Laboratory, China-Japan Friendship Hospital. His primary research interests are in clinical microbiology. His group is devoting to special pathogens detection, bacterial resistant mechanisms, bacterial epidemiological research, especially focus on Clostridium difficile.
The authors declare that they have no competing interests.
Consent for publication
Ethics approval and consent to participate
This study was approved by the ethics committee of China-Japan Friendship Hospital.
- 2.Kuijper EJ, Barbut F, Brazier JS, et al. Update of Clostridium difficile infection due to PCR ribotype 027 in Europe, 2008. Euro Surveill 2008;13(31).Google Scholar
- 12.Snitkin ES, Zelazny AM, Thomas PJ, et al. Tracking a hospital outbreak of carbapenem-resistant Klebsiella pneumoniae with whole-genome sequencing. Sci Transl Med. 2012;4(148):148ra116-148ra116. Google Scholar
- 13.Cohen SH, Gerding DN, Johnson S, 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). Infect Control Hosp Epidemiol. 2010;31(5):431–55.CrossRefPubMedGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.