Evaluation of the health and healthcare system burden due to antimicrobial-resistant Escherichia coli infections in humans: a systematic review and meta-analysis

BACKGROUND

Assessment of the burden of disease due to antimicrobial-resistant Escherichia coli infections facilitates understanding the scale of the problem and potential impacts, and comparison to other diseases, which allows prioritization of research, surveillance, and funding. Using systematic review and meta-analysis methodology, the objectives were to evaluate whether humans with antimicrobial-resistant E. coli infections experience increases in measures of health or healthcare system burden when compared to susceptible E. coli infections.

METHODS

Comprehensive literature searches were performed in four primary and seven grey literature databases. Analytic observational studies of human E. coli infections that assessed the impact of resistance to third/fourth/fifth-generation cephalosporins, resistance to quinolones, and/or multidrug resistance on mortality, treatment failure, length of hospital stay and/or healthcare costs were included. Two researchers independently performed screening, data extraction, and risk of bias assessment. When possible, random effect meta-analyses followed by assessment of the confidence in the cumulative evidence were performed for mortality and length of hospital stay outcomes, and narrative syntheses were performed for treatment failure and healthcare costs.

RESULTS

Literature searches identified 14,759 de-duplicated records and 76 articles were included. Based on 30-day and all-cause mortality meta-analyses, regardless of the type of resistance, there was a significant increase in the odds of dying with resistant E. coli infections compared to susceptible infections. A summary mean difference was not presented for total length of hospital stay meta-analyses due to substantial to considerable heterogeneity. Since small numbers of studies contributed to meta-analyses for bacterium-attributable mortality and post-infection length of hospital stay, the summary results should be considered with caution. Studies contributing results for treatment failure and healthcare costs had considerable variability in definitions and reporting.

CONCLUSIONS

Overall, resistant E. coli infections were associated with significant 30-day and all-cause mortality burden. More research and/or improved reporting are necessary to facilitate quantitative syntheses of bacterium-attributable mortality, length of hospital stay, and hospital costs. Protocol Registration PROSPERO CRD42018111197.

Factors that antimicrobial resistance in food systems: a participatory modelling approach

Background

Antimicrobial resistance (AMR) emerges from a complex web of factors; understanding their dynamics is key to determining sustainable solutions. Thus, we aimed to create a model of the diverse factors influencing AMR in two food systems and model the impacts of interventions.

Methods

We built a causal loop diagram (CLD) of the factors driving AMR in the food chain via 4 participatory workshops (2 in Sweden; 2 in Malaysia) with diverse stakeholders. The CLD became the structure of a compartmental model, which was populated using data from multiple sources (e.g., interviews, surveillance data). Using fuzzy set theory, quantitative and qualitative data were converted to categorical variables. The compartmental model was created in AnyLogic and was used to test how expert-selected solutions (e.g., taxation) might impact AMR under different scenarios.

Results

Factors identified as influencing AMR across Europe and Southeast Asia clustered around key themes: on-farm (e.g., biosecurity); social (e.g., consumer demand); research (e.g., alternatives to antimicrobials [AMs]); economic (e.g., agricultural production levels); policy (e.g., trade agreements); and environment (e.g., waste management). Differences were identified between regions, for example, regulations and standards regarding imports or food safety were more relaxed in Southeast Asia than in Europe. Identified interventions included: AMR education in schools, training diverse stakeholders in AM stewardship, increased on-farm biosecurity measures to limit disease and the need for AMs, and taxing AM-containing products.

Conclusions

Our model captured a range of multi-level, interlinked factors that impact AMR in the European and Southeast Asian food system contexts. Preliminary findings suggest that different principles need to be cultivated (e.g., polycentric governance, cross-sector partnerships) to ensure that interventions addressing AMR are sustainable over time.

Key messages

Our study visually characterized the interlinked factors that impact AMR transmission and emergence in food systems. Our approach provides a tool to model impacts of potential interventions.

A geographic study of West Nile virus in humans, dead corvids and mosquitoes in Ontario using spatial scan statistics with a survival time application

Surveillance of West Nile virus (WNv) in Ontario has included passive reporting of human cases and testing of trapped mosquitoes and dead birds found by the public. The dead bird surveillance programme was limited to testing within a public health unit (PHU) until a small number of birds test positive. These dead corvid and mosquito surveillance programmes have not been compared for their ability to provide early warning in geographic areas where human cases occur each year. Spatial scan statistics were applied to time-to-event survival data based on first cases of WNv in found dead corvids, mosquitoes and humans. Clusters identified using raw data were compared to clusters based on model-adjusted survival times to evaluate whether geographic and sociodemographic factors influenced their distribution. Statistically significant (p < .05) space-time clusters of PHUs with faster time to detection were found using each surveillance data stream. During 2002-2004, the corvid surveillance programme outperformed the mosquito programme in terms of time to WNv detection, while the clusters of first-positive mosquito pools were more spatially similar to first human cases. In 2006, a cluster of first-positive dead corvids was located in northern PHUs and preceded a cluster of early human cases that was identified after controlling for the influence of geographic region and sociodemographic profile.

Cluster Analysis of Campylobacter jejuni Genotypes Isolated from Small and Medium-Sized Mammalian Wildlife and Bovine Livestock from Ontario Farms

Using data collected from a cross-sectional study of 25 farms (eight beef, eight swine and nine dairy) in 2010, we assessed clustering of molecular subtypes of C. jejuni based on a Campylobacter-specific 40 gene comparative genomic fingerprinting assay (CGF40) subtypes, using unweighted pair-group method with arithmetic mean (UPGMA) analysis, and multiple correspondence analysis. Exact logistic regression was used to determine which genes differentiate wildlife and livestock subtypes in our study population. A total of 33 bovine livestock (17 beef and 16 dairy), 26 wildlife (20 raccoon (Procyon lotor), five skunk (Mephitis mephitis) and one mouse (Peromyscus spp.) C. jejuni isolates were subtyped using CGF40. Dendrogram analysis, based on UPGMA, showed distinct branches separating bovine livestock and mammalian wildlife isolates. Furthermore, two-dimensional multiple correspondence analysis was highly concordant with dendrogram analysis showing clear differentiation between livestock and wildlife CGF40 subtypes. Based on multilevel logistic regression models with a random intercept for farm of origin, we found that isolates in general, and raccoons more specifically, were significantly more likely to be part of the wildlife branch. Exact logistic regression conducted gene by gene revealed 15 genes that were predictive of whether an isolate was of wildlife or bovine livestock isolate origin. Both multiple correspondence analysis and exact logistic regression revealed that in most cases, the presence of a particular gene (13 of 15) was associated with an isolate being of livestock rather than wildlife origin. In conclusion, the evidence gained from dendrogram analysis, multiple correspondence analysis and exact logistic regression indicates that mammalian wildlife carry CGF40 subtypes of C. jejuni distinct from those carried by bovine livestock. Future studies focused on source attribution of C. jejuni in human infections will help determine whether wildlife transmit Campylobacter jejuni directly to humans.

Molecular and Statistical Analysis of Campylobacter spp. and Antimicrobial-Resistant Campylobacter Carriage in Wildlife and Livestock from Ontario Farms

The objectives of this study were to (i) compare the carriage of Campylobacter and antimicrobial-resistant Campylobacter among livestock and mammalian wildlife on Ontario farms, and (ii) investigate the potential sharing of Campylobacter subtypes between livestock and wildlife. Using data collected from a cross-sectional study of 25 farms in 2010, we assessed associations, using mixed logistic regression models, between Campylobacter and antimicrobial-resistant Campylobacter carriage and the following explanatory variables: animal species (beef, dairy, swine, raccoon, other), farm type (swine, beef, dairy), type of sample (livestock or wildlife) and Campylobacter species (jejuni, coli, other). Models included a random effect to account for clustering by farm where samples were collected. Samples were subtyped using a Campylobacter-specific 40 gene comparative fingerprinting assay. A total of 92 livestock and 107 wildlife faecal samples were collected, and 72% and 27% tested positive for Campylobacter, respectively. Pooled faecal samples from livestock were significantly more likely to test positive for Campylobacter than wildlife samples. Relative to dairy cattle, pig samples were at significantly increased odds of testing positive for Campylobacter. The odds of isolating Campylobacter jejuni from beef cattle samples were significantly greater compared to dairy cattle and raccoon samples. Fifty unique subtypes of Campylobacter were identified, and only one subtype was found in both wildlife and livestock samples. Livestock Campylobacter isolates were significantly more likely to exhibit antimicrobial resistance (AMR) compared to wildlife Campylobacter isolates. Campylobacter jejuni was more likely to exhibit AMR when compared to C. coli. However, C. jejuni isolates were only resistant to tetracycline, and C.  coli isolates exhibited multidrug resistance patterns. Based on differences in prevalence of Campylobacter spp. and resistant Campylobacter between livestock and wildlife samples, and the lack of similarity in molecular subtypes and AMR patterns, we concluded that the sharing of Campylobacter species between livestock and mammalian wildlife was uncommon.