An outstanding study from colleagues at Imperial shows that if you use a bucket-load of meropenem (a carbapenem antibiotic) one year, you see an uptick in carbapenem-resistant Enterobacteriaceae (CRE) the next. The figure below plots meropenem use in the previous year with the incidence rate of OXA-48 producing K. pneumoniae CRE.
Figure: The use of meropenem in the previous year plotted against the incidence rate of OXA-48 producing K. pneumoniae
On the one hand this is not very surprising. But the study design is neat and convincing, and the study opens the door to using surveillance of antimicrobial use as a real-time tool for predict outbreaks. How would this work in practice? You set up an “alert level” of carbapenem use, and if this is exceeded, launch an urgent campaign to return carbapenem use to “normal” or risk an outbreak?
The team then evaluate the impact of an antimicrobial stewardship intervention control an outbreak of OXA-48 K. pneumoniae in a renal unit. Whilst the impact of the intervention is clear, I’m not convinced this is all about antimicrobial stewardship. The intervention also included ‘case isolation, screening of contacts, barrier nursing and other infection control interventions’. I accept that the increase in meropenem use immediately prior to the outbreak is convincing evidence that the demonstrated reduction in meropenem use was an important part of bringing the outbreak under control, we can’t rule out the other aspects of the bundle.
The interrupted time series analysis used in this study is a powerful tool, but I would like to see how a “breakpoint model” approach would have looked within this analysis (where the breakpoint is not specified a priori – you allow the model to tell you when the significant change in rate occurred).
Also, was there any correlation between carbapenem use and carbapenem resistance in other Enterobacteriaceae or in non-fermenters, such as A. baumannii? Perhaps (hopefully!) these were too rare to analyse.
So, what drives carbapenem resistance? Carbapenem use, stupid! But of course, it’s not this simple. Antibiotic use provides the selective pressure giving resistant strains the chance to out-compete their sensitive counterparts. However, poor infection control practice will accelerate the problem massively. And will also mean that you will also continue to see problems even with tight antimicrobial restriction.
