Progress 08/15/13 to 03/14/17
Outputs Target Audience:
Nothing Reported
Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?During this reporting period, this project provided training of a geography PhD student in integrating epidemiological data with methods from community ecology. How have the results been disseminated to communities of interest?
Nothing Reported
What do you plan to do during the next reporting period to accomplish the goals?
Nothing Reported
Impacts What was accomplished under these goals?
A total of 3347 S. aureus isolates from human, 913 from meat and 37 from environmental samples were genetically characterized. A subset of isolates were tested for antibiotic susceptibility. S. aureus was quantified in raw meat samples. Morisita indices, non-parametric multidimensional scaling (NMDS) and Permutational Multivariate Analysis of Variance Using Distance Matrices (PERMANOVA) were used to determine the relationship between spa type composition among human samples and spa type composition of meat samples where they report shopping. A quantitative microbial risk assessment (QMRA) was developed to characterize the risk of becoming nasally colonized with MRSA resulting from preparation of contaminated pork product. An exposure assessment model estimated the MRSA dose delivered to the nose via contaminated fingers. Two scenarios were considered: a high-risk scenario, in which auto-inoculation occurred when the fingers were freshly contaminated; and a low-risk scenario, in which auto-inoculation occurred after preparation, when the MRSA on the fingers was subject to die-off and transfer to surfaces. A Beta-Poisson dose-response curve was fit to available human data. Whole genome sequence was carried out on 100 S. aureus isolates resulting from human, meat, and environmental samples. The median MRSA dose was 1 colony forming units (CFU) per preparation event for the low risk scenario and 100 CFU for the high-risk scenario. The resulting median estimates of risk of nasal colonization were 7.2 × 10-3 and 8.7 × 10-5 for the high and low risk scenarios respectively. This represents a non-negligible risk. However, there remains considerable uncertainty in some of the model parameters, mainly MRSA-specific dose response data, concentration of MRSA on retail pork meat, and the fraction of MRSA transfer from fingers to surfaces. We find spa type composition is significantly different between households and meat sampled from their associated grocery stores. spa types found in meat are not significantly different regardless of the store or county in which they were sampled. spa types in people also exhibit high similarity regardless of residential location in urban or rural counties. Such findings suggest meat does not play an important role as a source of S. aureus colonization in shoppers, rather the larger community one lives in is responsible for the genetic diversity of S. aureus strains colonizing individuals. Our research showed that ST9 was a common S. aureus sequence type contaminating meat and infecting humans in the US. Whole genome phylogenetic analyses of the isolates collected in this study and those from previous studies, including those collected directly from livestock, suggest that ST9 has a similar population history as ST398. The findings are an important advancement for our field because they show that the ability to jump among hosts and quickly adapt via loss and gain of phage-encoded colonization factors is not a unique characteristic of ST398. Furthermore, it shows that ST398 is not the only S. aureus clone commonly circulating in livestock and contaminating retail meat products. The intermingled phylogenies for each of the STs suggest that the populations contaminating meat and infecting humans are the result of multiple exchanges between people and either meat or the livestock from which the meat products were derived. However, future analyses that include food-animal isolates are necessary to confirm that the strains on meat are the same as those in the livestock populations.
Publications
- Type:
Journal Articles
Status:
Published
Year Published:
2017
Citation:
Thapaliya D, Forshey BM, Kadariya J, Quick MK, Farina S, OBrien A, et al. Prevalence and molecular characterization of Staphylococcus aureus in commercially available meat over a one-year period in Iowa, USA. Food Microbiology. 2017;65:122-9.
- Type:
Journal Articles
Status:
Published
Year Published:
2014
Citation:
Kadariya J, Smith TC, and Thapaliya D, Staphylococcus aureus and Staphylococcal Food-Borne Disease: An Ongoing Challenge in Public Health, BioMed Research International, vol. 2014, Article ID 827965, 2014. doi:10.1155/2014/827965
- Type:
Journal Articles
Status:
Under Review
Year Published:
2017
Citation:
Carrel, M., Zhao, C., Thapaliya, D., Bitterman, P., Kates, A. E., Hanson, B. M., Smith, T. C. (2017). Assessing the potential for raw meat to influence human colonization with Staphylococcus aureus.
|
Progress 08/15/15 to 08/14/16
Outputs Target Audience:
Nothing Reported
Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?
Nothing Reported
How have the results been disseminated to communities of interest?
Nothing Reported
What do you plan to do during the next reporting period to accomplish the goals?During the next reporting period, we will continue working on analyzing the data gathered during this reporting period; especially in modeling, geospatial analyses and whole genome sequencing. We will perform cluster analysis, to see if certain SA types are found in heavier spatial concentrations than would be expected, and what in the environment, other than meat exposure, is correlated with specific spa types or drug resistant isolates. Since our last update, we have continued to monitor the literature for additional S. aureus/MRSA nasal colonization dose-response data. We identified a number of data sets emanating from studies of pigs; however under closer scrutiny, each of these data sets had limitations that prevented a meaningful dose response analysis. In particular, the dose delivered in each of these studies is difficult to quantify due to housing conditions of the pigs. The greatest challenge that we faced was determining appropriate underlying data to use for model parameters. Specifically, data on kitchen behaviors (e.g., hand washing, cross-contamination events, hand touching rates to fomite and face), initial MRSA loading on meat products, the mass-to-surface area ratio of pork products, and consumption/purchasing data need to be further collected and evaluated. Additionally, the model requires developing a novel mathematical characterization of the meat-to-hand-to-face and the meat-to-fomite-to-hand-to-face routes of exposure. We will iteratively run the model as necessary to finalize our manuscript. We will also carry out MLST on our environmental isolates.
Impacts What was accomplished under these goals?
Using the Geographic Information System (GIS) created in a previous reporting period, combined with the Staphylococcus aureus (SA) sampling data, we have used methods from community ecology (Morisita Index & Nonparametric MultiDimensional Scaling) to determine whether the SA species diversity in stores is significantly similar to species similarity in households that report shopping in those stores. Broadly speaking, there does not appear to be overlap in SA species diversity between households and stores. However, we have found a difference in the distribution of spa types and BURP clusters between the "urban" Johnson County and the "rural" Keokuk County. The spatial plotting of BURP cluster analysis indicated the higher prevalence of BURP clusters in Keokuk County compared to Johnson County. In addition to generally different types in the two counties, livestock-associated (LA) SA types appear in significantly greater numbers in Keokuk County households than Johnson, although they do not spatially cluster within the county. Households living on farms or that own swine or cattle are statistically significantly more likely to have a LA SA colonization, there is no difference observed in households that own turkeys or chickens versus other households. Using the GIS, we calculated the number of pigs within one mile of each participant household; regression indicates a positive and statistically significant relationship between the presence and number of pigs in close residential proximity and the presence of a LA SA type. Currently the genetic diversity of participant households is being compared to that of the meat and environmental samples from grocery stores they shop at (as reported in questionnaires) using a variety of plotting methods. To assess the potential for S. aureus transmission among humans, meat, and the environment, we performed whole genome sequence-based analysis on 100 S. aureus isolates collected from these three sources. Multi locus sequence types (MLSTs) were assigned using a bash script applied to assembled Illumina data. Briefly, Illumina short read sequences were assembled into contigs using the SPADES assembler. Quality of the assembly was determined by the N50 parameter, as well as by mapping reads back to the assembly. Nucleotide-Nucleotide BLAST (Version 2.2.25+) was used to compare the housekeeping gene against each of the assembled genomes. Sequence similarity matches of genes were determined using thresholds of 100% nucleotide identity and 100% coverage of the query sequence length. The script then used the matched genes and MLST profile data to determine the final MLST type. Phyloviz software was used to draw a goeBURST using the in silico predicted MLST types of 98 isolates. The plot was drawn to scale and colored by the Source-type. Novel MLST types were given dummy alleles of >1000. One sample from each MLST was selected as a reference and assembled with Spades. Single nucleotide polymorphisms (SNPs) were called using NASP v1.0 (https://github.com/TGenNorth/NASP) and recombinant regions were removed with Gubbins . The remaining SNPs were used to infer the relationship between the isolates in PAUP with a Maximum Parsimony (MP) tree. MLST analysis revealed that several S. aureus sequence types (STs) were present in multiple sources, and six of these ST were represented by six or more isolates, including ST5, ST8, ST15, ST30, ST45, and ST398 (analyzed previously). To maximize phylogenetic resolution, the isolates from each ST was analyzed separately. Using the NASP software, we identified an average of 1855 SNPs among different STs. Gubbins analysis suggested that only a minority of the SNPs fell within putative recombinant or horizontally-acquired genomic regions. Using SNPs from non-recombinant, core genomic regions for each sequence type, we generated MP trees to infer the phylogenetic relationships among the isolates from various sources. In each tree, meat and human isolates were intermingled, but most isolates, including those from the same sources were separated by many SNPs. There were notable acceptions in the ST5 phylogeny, where multiple isolates were clustered tightly with very few (sometimes zero) SNPs differentiating them. However, each of these clusters were comprised by single-source isolates. The intermingled phylogenies for each of the STs suggest that the populations contaminating meat and infecting humans are the result of multiple exchanges between people and either meat or the livestock from which the meat products were derived. Previous studies suggest that food animals--rather than humans--are the primary source of S. aureus contaminating retail meat. However, future analyses that include food-animal isolates are necessary to confirm that the strains on meat are the same as those in the livestock populations. Regardless of the ultimate source of the S. aureus strains contaminating retail meat, the diversity of the strains and intermingling with human clinical isolates suggests that food may serve as a reservoir for S. aureus with the potential for causing human infections. Previous studies suggest that while ST398 is a common contaminant of retail pork products in Denmark, direct livestock exposure is still the most important source of human exposure and infection in that country. Additional research is required to determine is this is also true for these other STs that may be better adapted to colonizing and infecting humans. The risk of nasal colonization for consumers of MRSA-contaminated retail pork was estimated by incorporating the dose-response model with the exposure assessment model. A probabilistic model was constructed that uses Monte Carlo simulations performed in two dimensions to evaluate uncertainty and variability independently. The probability of colonization during a food preparation event is calculated using the parameterized Beta-Poisson model. We have completed integration of our models and have performed multiple simulations to assess a number of exposure scenarios. Since our last update, we have continued to review availability of data for some key parameters and updated those found into the exposure model. We have also completed evaluating some of the model's assumptions and baseline equations to further increase reliability and plausibility of the current model. Additionally we have recreated the model in R from our original format in Crystal Ball. In adopting the model to the new format we identified a small error in the initial code that once corrected lowered the dose. We also challenged the idea of a steady state concentration on the hands, have added a die-off component on the meat, and ran several scenarios. Lastly, we performed a 2-D analysis to estimate uncertainty and variability. The existing exposure model, though functional, is in the middle stage of development and requires further modification in order to assess different exposure scenarios. A relatively crude hand washing component has been incorporated into the updated model, which yields a non-negligible central tendency risk of 10^-4. We have continued to wrestle conceptually with how the temporal sequence of events (exposure versus control) in an exposure model affects the individual estimates of event risk. By manipulating inputs in the exposure model we are able to assess which parameters have the greatest impact on risk estimates and whether the sequence of events affects the importance of the parameter. We have also been reviewing literature on meal preparation steps, hygiene practices, and cross contamination.
Publications
|
Progress 08/15/14 to 08/14/15
Outputs Target Audience:
Nothing Reported
Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?
Nothing Reported
How have the results been disseminated to communities of interest?
Nothing Reported
What do you plan to do during the next reporting period to accomplish the goals?During the next reporting period, we will be focused on addressing the special aim 2 of this project: Model the spread of S.aureus in the study population, and determine the risk attributable to food in the acquisition of S. aureus. Monitoring of the literature for S. aureus/MRSA nasal colonization dose-response data will continue. In particular, there may be opportunity to utilize data emanating from studies of pigs. We are continuing to review availability of data for some key parameters to be incorporated into the exposure model. We are also currently evaluating some of the model's assumptions and baseline equations to further increase reliability and plausibility of the current model. We plan to investigate the potential for additional routes of exposure that may lead to nasal colonization of people in household settings (e.g., inhalation of aerosolized MRSA). It is unclear at this point whether a functional and useful model will be developed in light of these limitations, but evaluation of the different secondary pathways will at the very least help identify data gaps. This will in turn provide insight into how future studies can be designed to provide this missing information. We will be aiming to provide at least a rudimentary assessment of some of these pathways and/or a detailed discussion on the difficulties associated with this type of effort. We will continue to search for input data (if they exist) and/or adjustment of exposure equations to better characterize exposure mitigating strategies like surface sanitizing and personal hygiene practices. Additionally, we will continue refinement of the exposure model for more accurate risk predictions. We also will execute planned actions to achieve the specific aim 3rd of this project: To determine the origin of food-borne S. aureus in order to examine the importance of contamination on-farm versus post-slaughter. We plan to use whole genomic sequencing approach to carry out this mission. The next step in geospatial analysis is determining whether individual household SA types are associated with types found in environmental and meat samples in close proximity to their house. One challenge we anticipate is determining what, if any, temporal lag should be considered between SA isolates found in the environmental & meat samples and those found in households, in other words, is exposure to SA in the environment or food associated with SA colonization with that type one week subsequently or some other duration of time? We are looking to the literature, primarily on colonization in healthcare settings, for suggestions on what time lag is appropriate or expected. We will also perform cluster analysis, to see if certain SA types are found in heavier spatial concentrations than would be expected, and what in the environment, other than meat exposure, is correlated with specific spa types or drug resistant isolates.
Impacts What was accomplished under these goals?
We finished molecular characterization for all human, meat and environmental samples. The prevalence of S. aureus on meat samples was 27.8% (913/3290). Prevalence was highest in antibiotic-free chicken (53.7%, 66/123) and pork sausage (40.4%, 136/337). Among the positive isolates, 18 were PVL-positive (1.9%) and 41 (4.5%) carried mecA. Phenotypic oxacillin resistance was observed for 3.4% (7/205) of the isolates tested, while 9.8% (20/205) were multi-drug resistant. The number of colony-forming units (CFU) per 10g meat ranged from 2 to 517 (median: 8 CFU per 10g of meat; mean: 28) with the highest bacterial load observed on turkey samples. The prevalence of S. aureus and methicillin-resistant S. aureus (MRSA) in environmental samples were 13.4% (37/276) and 2.5% (7/276) respectively. All positive S. aureus isolates were subjected to spa typing. A total of 132 spa types were detected from 913 contaminated meat samples. Overall, t002 was the most common spa type identified (137; 15.0%), followed by t273 (87; 9.5%), t034 (44; 4.8%), t021 (44; 4.8%), t037 (39; 4.3%), t010 (37; 4.0%), t337 (33; 3.6%), t189 (32; 3.5), and t491 (20; 2.2%). No other spa type constituted more than 2% of S. aureus isolates. spa types were grouped based on genetic proximity to spa types typically associated with specific cluster complexes. CC5-related isolates (e.g. t002) were the most common, followed by CC30-related (e.g. t012) isolates. The overall prevalence of S. aureus among adults and minors was 33.8% (1736/5132) and 30.4% (713/2344) respectively. The overall prevalence of methicillin - susceptible S. aureus among adults and minors was 32.5% (1669/5132) and 29.8% (699/2344) respectively. Likewise, the overall prevalence of MRSA among adults and minors was 1.3% (67/5132) and 0.6% (14/2344). The overall prevalence of the PVL gene was 0.8% (26/3119), and the overall prevalence of mecA gene was 2.6% (81/3119). The overall S. aureus prevalence was higher in Johnson County (37.3%) compared to Keokuk County (31.1%). All positive S. aureus isolates were subjected to spa typing. A total of 125 spa types were detected from 3,119 isolates. The most common spa types were t002 (n=345), t008 (n=248), and t216 (n=180). A quantitative microbial risk assessment (QMRA) model is currently being developed to characterize risk of becoming nasally colonized with MRSA from contaminated retail pork meat. A functional dose-response assessment was developed that specifically models nasal colonization by S. aureus, rather than infection. This is because auto-inoculation during food preparation is considered the primary exposure route of concern and there is availability of published literature on nasal colonization of S. aureus. Two disparate models were developed, one using human data, and another using data from mouse studies. A probabilistic model was constructed that uses Monte Carlo simulations performed in two dimensions to evaluate uncertainty and variability independently. The probability of colonization during a food preparation event is calculated using the parameterized Beta-Poisson model. We updated the human data model to include only data that evaluated presence of MRSA in nostrils after 5 days (the previous model included positives that showed presence of MRSA 24-48 hr after inoculation). Several model parameters previously treated as constants have been switched to distributions that incorporate variability into the model. Additionally, we have incorporated a bacterial die-off parameter that accounts for inactivation that occurs on the fingertips. Including inactivation into our model is supported by data from trials using S. aureus on human fingertips that showed significant inactivation occurred within two to 10 minutes. Our team has initiated an analysis of data gaps with regard to secondary transmission, and has begun to evaluate the complexity of the potential exposures and the temporal integration of control steps (e.g., hand sanitation) and appropriate methods by which to incorporate the uncertainty and variability in estimates of annual risk. We have been also reviewing literature on meal preparation steps, hygiene practices, and cross contamination. We have begun constructing the Geographic Information Systems (GIS) geodatabase that will be used for the proposed geospatial analysis of connections between meat & environmental samples and household participant swabs. All locations of meat and environmental sampling were successfully geocoded (assigned a latitude/longitude coordinate). Ninety-three of the ninety-five household locations were successfully geocoded, the two that were unable to be assigned to a latitude/longitude were due to a PO Box address being listed and a street address listed that does not exist in Iowa (likely due to either data entry issues or to error on the part of participant). The geocode rate (98%) is high for such studies, particularly in rural states where geocoding can be difficult. Also included in the geodatabase are data on hog concentrated animal feeding operations (CAFOs), Census information on socio-demographic characteristics, and land use/land cover data (such as road networks).
Publications
|
Progress 08/15/13 to 08/14/14
Outputs Target Audience: Consumers, Farmers, Policymakers, Researchers, public Health Workers, Veterinarians Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?
Nothing Reported
How have the results been disseminated to communities of interest? We have not disseminated our preliminary results yet. We plan to disseminate our findings via peer-reviewed journal article in near future. What do you plan to do during the next reporting period to accomplish the goals? During the next reporting period, We will be focused on addressing the special aim 2 of this project: Model the spread of S.aureus in the study population, and determine the risk attributable to food in the acquisition of S. aureus. We hypothesized that food has the potential to play a role in the transmission of virulent strains of S. aureus, similar to other environmental sources. Under this aim, we will develop a quantitative microbial risk assessment model following the standard paradigm of : hazard identification, dose-response assessment, exposure assessment, and risk characterization. A Geographic Information System (GIS) and associated methods of spatial analysis will be used to integrate the multiple sources of information hypothesized to be related to the likelihood a person has experienced increased risks of being colonized with S. aureus. We also will execute planned actions to achieve the specific aim 3rd of this project: To determine the origin of food-borne S. aureus in order to examine the importance of contamination on-farm versus post-slaughter. We plan to use whole genomic sequencing approach to carry out this mission.
Impacts What was accomplished under these goals?
During this reporting period the research team members addressed the specific aim 1st of this project: To determine the molecular epidemiology of Staphylococcus aureus isolated from meat, humans, and the environment in Keokuk County and Iowa City, Iowa. In this period, we were focused on the molecular characterization of all human isolates positive for S. aureus. We enrolled 177 adults and 86 minors as a part of a prospective cohort study comprising 95 family units. Throat and nasal swabs from adults, and only nasal swabs from minors were collected on a weekly basis for a period of 1 year. A total of 3,119 isolates were resulted from 10,987 human samples. The overall prevalence of S. aureus among adults and minors was 33.8% (1736/5132) and 30.4% (713/2344) respectively. The overall prevalence of methicillin – susceptible Staphylococcus aureus among adults and minors was 32.5% (1669/5132) and 29.8% (699/2344) respectively. Likewise, the overall prevalence of methicillin-resistance Staphylococcus aureus (MRSA) among adults and minors was 1.3% (67/5132) and 0.6% (14/2344). The overall prevalence of the PVL gene was 0.8% (26/3119), and the overall prevalence of mecA gene was 2.6% (81/3119). The overall S. aureus prevalence was higher in Johnson County (37.3%) compared to Keokuk County (31.1%). All positive S. aureus isolates were subjected to spa typing. A total of 125 spa types were detected from 3,119 isolates. The most common spa types were t002 (n=345), t008 (n=248), and t216 (n=180). Upon the analysis of the spa data using the BURP algorithm, 10 and 7 clusters were observed in Keokuk County and Johnson County respectively. However, the number of spa types per cluster was observed higher in Johnson County. A subset of 725 isolates chosen by selecting all isolates received during the first week of each month was tested for MIC antibiotic susceptibility typing. We observed zero rates of resistance in most of the tested antibiotics. However, the adjusted resistance for erythromycin was 25.9% displaying the highest observed resistance. Quinupristin/dalfopristin had the resistance rate of 17.7%. All other antimicrobials had below 10% observed resistance rates. The study results indicate that majority of the population is capable of being colonized. It further indicates that strong environmental pressure and exposure to the pathogens may have been responsible for human colonization of S. aureus and MRSA.
Publications
- Type:
Journal Articles
Status:
Published
Year Published:
2014
Citation:
Jhalka Kadariya, Tara C. Smith, and Dipendra Thapaliya, Staphylococcus aureus and Staphylococcal Food-Borne Disease: An Ongoing Challenge in Public Health, BioMed Research International, vol. 2014, Article ID 827965, 9 pages, 2014. doi:10.1155/2014/827965
|
|