Various times, I’ve asked audiences “What percentage of antimicrobial resistance in humans do you think it attributable to antimicrobial use in animals?

  • Answers pretty much range from 0-100%.

The actual number is probably on the low end of that range, but we really don’t know. It’s such a complex system that a simple number can’t be generated.  In fact, we don’t have the data to even get close to an accurate overall estimate.

However, better estimates can be made for certain resistant bacteria, for which more specific data are available. The estimates are still pretty dodgy given current gaps in surveillance, so the numbers have to be taken with a big grain of salt, and we have to take care extrapolating to other bacteria or different geographic ranges. Regardless, the information can be interesting and useful if we are careful not to overinterpret things.

A recent paper in The Lancet Planetary Health (Mughini-Gras et al. 2019) investigated multidrug-resistant E. coli, specifically E. coli that produced extended spectrum beta-lactamases (ESBLs) or that harboured the AmpC gene. These E. coli are resistant to 3rd generation cephalosporins (an very important drug class for treatment of infections in people) and are often resistant to various other antimicrobials as well. The study evaluated data on ESBLs and AmpC E. coli from different Dutch sources, and developed a transmission model to estimate how people were becoming infected (outside of hospitals).

Here are some highlights from the study:

  • People carrying resistant E. coli likely most often (61%) got it from other people, followed by food, animal and environmental sources (in that order).
  • The graph below shows how common resistant E. coli are in various sources (size of the bar to the left of midline) and how important each source is to humans in terms of potential transmission (size of the bar to the right). As you can see, for some sources (e.g. chickens – the birds, not the meat), resistant E. coli are very common but they are not thought to be important sources of transmission to people, while for others, the likelihood of resistant E. coli is low but the sources pose a disproportionately high risk to humans (e.g. raw vegetables). The impact on people varies with the overall amount of exposure and how we handle potentially contaminated sources. For example, even though the rate of contamination of raw vegetables is low, we encounter those very frequently and we often don’t cook them. In contrast, the contamination rate in surface water is high, but we don’t have a lot of direct contact with untreated water.
  • Companion animals came up higher than I would have guessed, being estimated to account for 7% of human infections, most often from dogs (3.9%) (although it reinforces why I’m concerned about ESBLs in dogs and cats, and why we’re studying it).

The over-riding conclusion was that humans are the main source of community-acquired resistant E. coli, but that non-human sources still play important roles. They also concluded that, even though non-human sources accounted for a minority of infections, it would be difficult for these resistant E. coli to be maintained in people without transmission to and from non-human source. So, addressing the problem in people alone will help, but won’t eliminate the problem.

We have to remember that these are just estimates and they may just (or at best) apply to the Netherlands. However, it’s an interesting story and should keep us thinking about the multi-disciplinary (One Health) approach that we need to take to combat antimicrobial resistance.