programmes that results from diagnostic submissions may not reflect the resistance
situation in the animal population, as these types of submissions tend to include
specimens from severe and/or recurrent clinical cases, including therapy failures”.
Franklin et al. ( 2001 ) further explain that because these isolates are likely to
represent biased samples, this type of susceptibility data may not indicate the true
prevalence of resistance within the given animal population. Therefore, caution
should be exercised when interpreting these data. The group emphasised that a
mechanism for mitigating this bias would be to “consider collection of samples
from primary clinical cases not previously treated with antibiotics, or the isolation
of potentially pathogenic bacteria from healthy animals”. For food-producing
animals, the collection of samples at the time of slaughter rather than from clinical
cases can be used to monitor resistance as the animal enters the human food chain.
Such information will help track the global impact of antimicrobial usage in
veterinary species.
Table 3 provides selected examples to highlight the differences in global sur-
veillance systems. All have the endpoint of reporting percentage resistance. The
examples have been confined to surveillance of zoonotic and veterinary pathogens.
For examples of surveillance in human pathogens, all of which use clinical break-
points, see Masterton ( 2008 ).
3.3 Monitoring Programmes: Points to Consider
It is not always appropriate to compare antibiograms of veterinary and human
isolates of the same pathogen (an antibiogram is the in vitro profile of an organism’s
response to a panel of antibiotics, which can be used to determine the sensitivity of an
isolated bacterial strain to different antibiotics). In addition to inconsistent definitions
of resistance, there may be differences in the methodologies used to determine
minimum inhibitory concentration (MIC) values. Furthermore, even when consider-
ing one invariant organism–agent combination, there remain multiple variables that
can influence the outcome of monitoring studies. When ignoring differences in
methodology to determine MIC values, definitions of the term ‘resistance’ or bias
in sampling data (all of which substantially influence the outcome), it is clear that the
most important sources of variation in the interpretation of susceptibility data appear
to be country, host species and the year of isolation. Unfortunately, it is often
impossible to find comparable data for which only one of these parameters differ.
An example of an initiative for generating human–veterinary comparisons
comprises an attempt to combine data from Italy (Moroni et al. 2006 ) with that
generated in Germany (Kaspar 2006 ). The German GERM-Vet monitoring
programme was established for the purpose of determining the current status of
antimicrobial susceptibility of animal pathogens in Germany. This human–
veterinary information gathering initiative introduced scientific rigour to the
surveillance process, addressing many of the variables that could otherwise con-
found interpretation of data (Wallmann et al. 2003 ).
Antimicrobial Drug Resistance 237