Abstract
Clostridium difficile infection is often considered to result from recent acquisition of a C difficile isolate in a healthcare setting. However, C difficile spores can persist for long periods of time, suggesting a potentially large community environmental reservoir. The objectives of this study were to assess community environmental contamination of toxigenic C difficile and to assess strain distribution in environmental versus clinical isolates.
From 2013 to 2015, we collected community environmental swabs from homes and public areas in Houston, Texas to assess C difficile contamination. All positive isolates were tested for C difficile toxins A and B, ribotyped, and compared with clinical C difficile isolates obtained from hospitalized patients in Houston healthcare settings.
A total of 2538 environmental samples were collected over the study period. These included samples obtained from homes (n = 1079), parks (n = 491), chain stores (n = 225), fast food restaurants (n = 123), other commercial stores (n = 172), and hospitals (n = 448). Overall, 418 environmental isolates grew toxigenic C difficile (16.5%; P < .001) most commonly from parks (24.6%), followed by homes (17.1%), hospitals (16.5%), commercial stores (8.1%), chain stores (7.6%), and fast food restaurants (6.5%). A similar distribution of ribotypes was observed between clinical and environmental isolates with the exception that ribotype 027 was more common in clinical isolates compared with environmental isolates (P < .001).
We identified a high prevalence of toxigenic C difficile from community environs that were similar ribotypes to clinical isolates. These findings suggest that interventions beyond isolation of symptomatic patients should be targeted for prevention of C difficile infection.
Changes in the epidemiology of Clostridium difficile infection (CDI) have been dramatic in the United States, Canada, and Europe [1, 2]. Clostridium difficile infection has been identified as the most common cause of death due to gastroenteritis in the United States with approximately 14000 deaths in 2007 [3]. The most recent study from the Centers for Disease Control and Prevention (2011) increased these estimates to 453000 (95% confidence interval [CI], 397100–508500) cases of CDI annually in the United States with an estimated 29300 deaths (95% CI, 16500–42100) [4]. Building on previous studies that identified CDI in nonhealthcare settings, 34.2% of CDI cases were considered to be community acquired [5].
Historically, CDI has predominantly been thought to result from recent acquisition of a C difficile isolate in a healthcare center. Efforts are primarily directed at prevention of the spread of C difficile spores from symptomatic patients to prevent horizontal transmission of spores to other susceptible patients; however, early isolation of colonized patients has been shown to decrease CDI incidence [6]. Case series have suggested a wider source of potential C difficile contamination, including water, pets, foods, or farm animals [7, 8]. More recently, a large surveillance study in England demonstrated that a minimum of 45% of C difficile strains associated with clinical disease were genetically distinct by whole genome sequencing [9]. The authors concluded that there is a potentially large reservoir of C difficile that was most likely environmental in origin. Given these data, and also that the spores of C difficile are quite hardy and are able to survive for long periods of time in the environment, we hypothesized that the community setting may contain a large burden of C difficile contamination. To test this hypothesis, we conducted a large environmental surveillance study to assess rates of C difficile contamination in homes and public areas within Houston, Texas, a large metropolitan city in the southern United States. The objectives of the study were to assess environmental contamination of toxigenic C difficile and to assess strain distribution and cytotoxicity of environmental compared with clinical strains obtained from hospitalized patients with CDI.
METHODS
Environmental Contamination of Clostridium difficile
To assess C difficile contamination in public areas, common areas of chain merchant stores, fast food restaurants, public parks, and community homes were sampled using presterilized gauze lightly soaked with 0.85% NaCl. To assess C difficile contamination in homes, 3 to 5 household items were collected from a convenience sample of community homes located in various parts of Houston, Texas (2013–2015). Houses were identified by placement of advertisments (ads) in a variety of media outlets throughout Houston. Specifically, ads were sent to community and church groups around the city with a request to disseminate the ads during the meetings. The ads included contact information for the investigators who could then explain the purpose of the study. Interested persons were then given instructions on how to perform swabs, or investigators went to the household to perform the sampling. All households were single family or apartment dwellings. No persons in the homes had CDI within the previous 12 months. Surface samples of approximately 1 square foot from common living spaces (kitchen and bathroom) were swabbed with the presterilized cotton gauze. Shoe bottom samples were collected by swabbing the bottoms of a pair of shoes with the same procedure. Cleaning supplies (mops, brooms, dust pans) were also sampled. All samples were collected in a 50-mL size presterilized tube and transported to the laboratory within 12 hours of collection. For other collection areas (merchant stores, fast food restaurants, and public parks), common use areas that were frequently used by the public were targeted (bathrooms, railings, and playground equipment). To compare environmental contamination in hospital versus nonhospital settings, samples were obtained from bathrooms and patient table from hospital rooms with an admitted patient without CDI. For each batch of 10 swabs, a negative control of 1 swab that was not used but placed in with the rest of the swabs was used to assure no cross-contamination of swabs.
Clinical Isolates From Patients With Clostridium difficile Infection
Clinical strains of C difficile were obtained from a multicenter cohort study in 2 healthcare systems in Houston, Texas (2013–2015) [10]. Seven hospitals served as study sites, including 3 university-affiliated tertiary care centers and 4 community hospitals. In this study, stool specimens from all patients with a stool test positive for C difficile as part of routine clinical care were collected. Enzyme-linked immunoassay or a polymerase chain reaction (PCR) detection of the tcdB gene in unformed stool was the diagnostic methodology used at all study sites during the study period. Clinical microbiology laboratories would only test unformed stool. Stool samples were further analyzed by a centralized research microbiology laboratory for C difficile growth, toxin characterization including cytotoxicity, and ribotyping. This study was approved by the Institutional Review Board at the University of Houston and participating hospitals.
Microbiologic Procedures
Stool and environmental samples were enriched in brain heart infusion broth with 0.05% sodium taurocholate (Sigma Chemicals) and incubated anaerobically at 37°C for up to 5 days. One milliliter of broth culture from each sample was centrifuged to concentrate the cells with the resulting pellet suspended in 100 µL normal saline (0.85% NaCl), plated onto cycloserine cefoxitin fructose agar ([CCFA] Anaerobic Systems, Morgan Hill, CA), and incubated anaerobically at 37°C for 40–48 hours (Forma Anaerobic System, Model 1025/1029). Suspected colonies were tested using latex agglutination reagent (Oxoid, Hampshire, England). Each batch of samples was processed with a positive and a negative control. Presence of toxin genes were assessed using multiplex PCR to detect the presence of toxin A (tcdA), toxin B (tcdB) [11], and the binary toxin (cdtA and cdtB) genes [12]. Cytotoxicity of all C difficile isolates was assessed by the degree of rounding of cells from a Chinese Hamster Ovary cell line (CHO-K1, ATCC CCL-61) as previously described [13].
Clostridium difficile Ribotyping
Fluorescent ribotyping was performed as previously described [14, 15]. In brief, PCR ribotyping primers [16] were synthesized with a fluorescent label (Life Technologies, Inc., Carlsbad, CA) and adjusted to10 pmol/μL. A 25-μL PCR was performed using AmpliTaq Gold DNA polymerase (Applied Biosystems, Foster city, CA) and the following conditions: 95°C (10 minutes); 35 cycles of 95°C (30 seconds), 55°C (30 seconds), and 72°C (1 minute 30 seconds); final extension of 72°C (10 minutes) (Eppendorf vapo protect thermal cycler, Hamburg, Germany). Amplicons were analyzed using an ABI3730xl DNA Analyzer and MapMaker 1000 ROX DNA sizing standard (BioVentures, Inc., Murfreesboro, TN). Reference strains included PCR ribotypes 027, 001, 053–163, 002, 014–020, 017, and 078, respectively. This technique does not distinguish between ribotypes 053 and 163, ribotypes 014 and 020, and ribotypes 078 and 126; therefore, these are reported as combined ribotypes (ie, 053–163, 014–020, and 078–126).
Clostridium difficile Multilocus Variance Analysis
Multilocus variance analysis (MLVA) typing of C difficile isolates ribotype 027 was done using previously published fluorescent MLVA markers [17] using the protocol described previously by Broukhanski et al [18, 19]. In brief, amplicons are diluted in formamide, mixed with LIZ600 deoxyribonucleic acid (DNA) ladder and run on an ABI 3130xl genetic analyzer for fragment analysis. Processing of capillary electrophoresis was done using BioNumerics 7.6 (Applied Maths, Austin, TX) software. Peak files from the genetic analyzer are normalized to the LIZ600 DNA ladder to determine amplicon size. Copy number for each variable number of tandem repeats locus based on amplicon size was determined using previously published guidelines [18].
Analysis
Data were analyzed using SAS, version 9.3 (SAS Institute, Cary, NC). Prevalence of environmental contamination from homes and public space were calculated and compared between sampling sites using Pearson χ2. The MLVA dendrograms were created to assess microbial genetics differences between environmental and clinical strains stratified by ribotype. Risk factors for home C difficile contamination were assessed using Student t test and Pearson χ2 test as appropriate. A P value <.05 was considered significant.
RESULTS
Clostridium difficile Is Highly Prevalent in Environmental Isolates
A total of 2538 environmental samples were collected over the study period. These included samples obtained from homes (n = 1079), parks (n = 491), chain stores (n = 225), fast food restaurants (n = 123), other commercial stores (n = 172), and hospitals (n = 448). Overall, 418 environmental isolates grew toxigenic C difficile (16.5%). Prevalence of positive C difficile isolates were significantly different based on the location of environmental testing (P < .001) (Figure 1). The frequency of C difficile-positive isolates was highest in parks (24.6%), followed by homes (17.1%), commercial stores (8.1%), chain stores (7.6%), and fast food restaurants (6.5%). Home environmental contamination with C difficile is shown in Figure 2. Prevalence of positive C difficile isolates were similar for shoe soles (26.4%), doorsteps (24.7%), and lower for cleaning supplies (13.2%), kitchen areas (8.9%), and restrooms (9.0%). In 448 samples obtained from hospitals, 74 (16.5%) grew toxigenic C difficile.
Community environmental contamination of toxigenic Clostridium difficile.
Home environmental contamination of toxigenic Clostridium difficile.
Ribotypes From Environmental Clostridium difficile Were Similar to Clinical C difficile Isolates
Ribotyping was performed on 344 community environmental toxigenic C difficile isolates and compared with 74 hospital environmental isolates and 615 clinical isolates obtained from hospitalized patients with CDI (Figure 3). All toxigenic ribotypes were toxin positive for toxin A and B by PCR. Ribotypes present in at least 10% or more of samples were 014-020 (21.2%), FP-11 (20.4%), and 002 (12.2%). These results were compared with 615 clinical isolates of C difficile. A similar distribution of ribotypes was observed with one exception. Ribotype 014-020 and FP-11 were present in more than 10% of clinical isolates. However, ribotype 027 was much more common in clinical isolates (24.1%) compared with environmental isolates (4.5%; P < .001). Similar results were observed with hospital environmental isolates in which ribotype 027 was more common in hospital compared with environmental strains. Multilocus variance analysis of ribotype 027 strains showed that there was no genetic distinction between clinical and environmental isolates (Figure 4). All toxigenic isolates identified by PCR (environmental or clinical) were cytotoxic. A representative sample of cytotoxic isolates is shown in Figure 5.
Ribotype distribution of clinical versus environmental Clostridium difficile isolates.
Multilocus variance analysis (MLVA) of environmental and clinical ribotype 027 strains. Environmental strains clustered together with clinical strains.
Cytotoxicity of environmental versus clinical isolates by ribotype.
DISCUSSION
In this large surveillance study, between 6.5% and 24.6% of samples obtained from various environmental settings were positive for C difficile. Prevalence of C difficile was higher in parks and homes and lower in fast food restaurants and merchant stores. Ribotype distribution was similar between environmental and clinical isolates with the exception that the ribotype 027 strain was observed more commonly in hospital clinical and environmental isolates. Strengths of the study include a large sample of environmental samples and comparator clinical isolates obtained in the same geographic location from different hospital types during the same time period. A validated molecular typing method confirmed a similar ribotype distribution between clinical and environmental C difficile isolates and other translational research findings including cytoxicity assessment of identified isolates and MLVA analysis.
These results build considerably upon previous findings of C difficile in the community environs; specifically, animals and food [20]. Using whole genome sequencing, a recent study demonstrated that 45% of all CDI cases in an endemic environment were attributable to a source unique from any other symptomatic case [9]. A possible community environmental reservoir was posited as the source for these unique strains. A smaller study recently examined household contamination of patients with CDI pre- and postfecal microbiota transplantation [21]. In the posttransplant time period, 38% of household environmental samples were positive for C difficile, most commonly the vacuum, toilet, and bathroom sink. In this current study, we demonstrated a high prevalence of C difficile isolates in the community comprised of multiple unique ribotypes. The evolutionary dynamics of C difficile has shown that it is a genetically diverse species with multiple lineages evolving independently to produce disease [22]. Our results agree with these previous findings because all strains contained toxin genes and were toxigenic using the cell cytotoxicity assay. These findings would suggest that the community environment may be a source of colonization for toxigenic C difficile. We have previously shown that half of all patients hospitalized with CDI are diagnosed within 48 hours of admission and up to one third of all cases of C difficile infection occur in the community [4, 23]. These findings suggest that specific patient populations at risk for CDI should be targeted for C difficile prevention efforts rather than solely focus on containment of patients with active disease. This possibility is re-enforced by previous findings that antimicrobial stewardship efforts can significantly reduce rates of CDI [24]. This is a paradigm shift in potential prevention efforts for CDI and might shift focus to patient-specific factors to identify a high-risk population. These theories will require further study.
This study has limitations. This surveillance study was conducted in a large, urban area in the southern United States. Results will need to be tested in different geographies. We used a molecular typing method to identify similar ribotypes between clinical and environmental isolates. We used a more discriminatory method (MLVA) for ribotype 027 isolates, but validation with animal models to further discriminate virulence differences was not performed. Our method for C difficile detection although sensitive is not quantitative. Whether higher concentrations of spores increases risk of infection is an interesting research question for future work. We did not use traditional random naturalistic home sampling in this study but rather a convenience sample of persons who answered an ad we place at randomly selected community and church group meetings. This study was not intended to identify circulation of C difficile isolates between healthcare centers and the community. Likewise, how this spore-forming anaerobe is disseminated throughout the community will require further study.
CONCLUSIONS
In conclusion, we identified a high prevalence of C difficile from community environs that were similar ribotypes to isolates that caused clinical disease. These findings suggest that interventions beyond isolation of symptomatic patients should be targeted for prevention of CDI.
Acknowledgments
Disclaimer. The study sponsor reviewed the final paper but did not have any role in the study design, collection, analysis, and interpretation of data, writing of the report, or the decision to submit the paper for publication
Financial support. This work was funded by a grant from the Texas Department of State Health Services (contract G110483). K. W. G. has received research grants from Merck & Co. and Summit, PLC.
Potential conflicts of interest. All authors: No reported conflicts.All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
References
Comments