Last Friday the results of the ESBL Attribution study (ESBLAT) were presented. After considerable media attention for ESBL-producing bacteria on our meat (especially retail chicken meat) and a 84-year old woman being “the first deadly victim of the new chicken-ESBL bacterium” a research consortium was asked to quantify the role of ESBL in animal industry for human health. The “research lab” was the Netherlands: one of the most densely populated countries in the world for both humans and animals, with the highest antibiotic use in the world for animals and the lowest for humans. If anywhere, zoonotic transmission should happen there!
To substantiate the last claims: A single square kilometre of Dutch soil hosts – on average – 400 humans (which is only less than in Taiwan, Bangladesh and South Korea), 1,085 chickens, 290 pigs, 25 calves, 25 sheep and 74 cows. And in 2007 these animals used almost 600.000 kg of antibiotics. At that time close to 100% of the retail chicken meat in our stores harboured ESBL-producing E. coli, and their proportion among invasive E. coli infections had surpassed the 5% level. Several studies demonstrated genetic similarities between ESBL-producing E. coli in animals, on meat and causing human infections. Whom to blame?
In a large cross-sectional prevalence study (n=4,177), in a reasonably representative part of the Dutch population, the prevalence of intestinal ESBL carriage was 4.5%. Out of 86 questions a few risk factors were identified (recent travel Asia/Africa, recent antibiotics, poor kitchen hygiene) but the odds ratios were fairly low. So, no real risk factors.
Where are we exposed to ESBLs? In Asia and Africa, but carriage rapidly declines when back home and transmission to household members seems not sufficient to create a sustained transmission chain. Living close to animal farms was not a risk factor, but we get exposed when swimming or when eating a true Dutch delicacy “filet Americain” (raw beef), albeit that exposure levels are low, see. The highest likelihood for ESBL-carriage stems from working with animals carrying ESBL, such as pigs; either on a farm or in a slaughter house. But that’s only a very small part of the population.
An extensive analysis of all genetic information from ESBL-producing E. coli (i.e., ESBL-genes, plasmids and sequence types) nicely demonstrated clustering in different animal groups, and of humans …. with other humans, see. And if the animal reservoir was an important source for human infections, one would expect that the “natural experiment” of the 70% reduction in antibiotic use (100% reduction of cephalosporins) and a sharp decline in carriage with ESBLs in animals since 2010, would have an effect on the incidence of infections caused by such bacteria in humans. But that effect is not seen until now.
It was, therefore, concluded that under the circumstances tested (the Netherlands) the attribution of ESBL-producing bacteria in animals to human infections is low. Actually, not that different from previous findings in Sweden, see page 48. No media headlines this time. But also no reason to change the successful reduction of antibiotic use in animals. The knowledge gained will help us to reassure our patients that Dutch hospitals are still very, very safe (for antibiotic resistance, at least) and to keep carbapenemase-producing bacteria out of our country.
For all presentations, see.