The objective of this work was to: (a) Describe surveillance and monitoring systems on AMU and AMR in European countries. (b) Describe and compare resistance data of clinical and non-clinical E. coli isolates from livestock in countries. (c) Describe the association of AMU and year with AMR. (d) Provide recommendations for improved “One Health” surveillance at the European level. A literature search was conducted in 2018 to identify relevant peer-reviewed articles and reports on AMU and AMR. It was used for identifying and assessing monitoring and surveillance systems on AMU and AMR in the human and animal sectors and foodborne AMR in six European countries (Germany, France, Spain, The Netherlands, The United Kingdom and Norway). Logistic regression analyses were performed comparing phenotypical data on clinical and non-clinical E. coli isolates of several animal categories from 2014 to 2017 at national and international level. AMU was only included in the national study as we found no approach to overcome the lack of harmonisation on AMU across countries. The Normalized Resistance Interpretation (NRI) method, a statistical approach, was applied to overcome the lack of harmonization on the laboratory methods and procedures in the international study. This work identified overlaps on monitoring and surveillance systems on AMU and AMR in the human and animal sector. A lack of harmonisation was encountered on: (a) Antimicrobial usage for livestock and, therefore, between humans and livestock and (b) AMR in clinical isolates of livestock and, therefore, between clinical and non-clinical isolates within livestock and between isolates of humans and livestock. In both statistical studies, we expected that diseased animals might carry higher resistance levels in bacteria to regular antimicrobial treatments compared to those from healthy animals. However, this was only found in isolates of calves. In contrast, higher resistance levels were mainly found in non-clinical isolates of broilers and turkeys. This phenomenon remains unclear. However, it might be that other factors not analysed in this study such as genetic data, animal age at time of sample collection in clinical isolates and the sample type have a major influence on AMR. Resistance prevalence did not differ between clinical and non-clinical isolates of broilers from Norway in 2016. This might be due to the low number of isolates together with the low resistance prevalence in Norway in isolates from broilers. We also investigated AMR changes over time in both studies. Associations between year and AMR per animal type and antimicrobial were investigated in the study at the national level considering clinical and non-clinical isolates together. However, in the international study, these associations were analysed independently per isolate type. Due to that, results differed between studies. These associations were very limited in the national study. In the international study, the analyses showed in most cases decreasing resistance trends in clinical and/or non-clinical isolates over time for the antimicrobials and animal categories studied in the countries suggesting that measures carried out against AMR including the reduction of AMU in the countries have effective results. We analysed statistically the associations of changes in AMU with changes in AMR of isolates from poultry in the national study considering clinical and non-clinical isolates together. We expected to find a correlation between AMU and AMR. However, this association was only partially shown. This could be partly due to having considered AMR in clinical and non-clinical isolates together as sometimes clinical and non-clinical isolates show different or even opposite trends. In the international study, the association between AMU and AMR was only analysed descriptively across countries. Resistance levels differed between countries for combinations of isolate type – antimicrobial – population. These results were mostly in line with the differences in antimicrobial use between animal categories. Comparability of these data was limited, mainly because of dissimilarities in the granularity of available data, sampling, reporting and in the units of measurement. This work proposes a list of recommendations for improved “One Health” surveillance. A One Health regarding the surveillance and monitoring of AMR and AMU is currently not straightforwardly achievable because of the lack of harmonisation of AMU and AMR surveillance within the livestock sector and also between the human and livestock sectors.
