A high profile article was published earlier this year in Science Translational Medicine, suggesting that Enterococcus faecium can exhibit clinically relevant levels of tolerance to alcohol-based hand hygiene products. The article has generated a huge amount of press coverage and discussion amongst experts. So, I thought it was about time I gave the article a once over. My initial thought was this would be unhelpful extrapolation of low-level tolerance to alcohol gel that wouldn’t be meaningful in a clinical setting. But having read the paper, there’s genuine concern here. Overall though, if true resistance to alcohol gel was going to be a problem, I’m pretty sure it would have reared its ugly head already.

The research team performed a retrospective lab-based study on a series of 139 E. faecium isolates from 1997-2015 in Australia. It’s important to note that the concentration of alcohol used in the lab testing was 23% – a lot lower than is typically used in alcohol gels. And the contact time was 5 minutes – a lot longer than would be the case in clinical practice. This was chosen ‘to provide a discriminating dynamic range among the E. faecium isolates’. Put another way, a usual concentration of alcohol (around 70%) killed the E. faecium dead, and didn’t allow for any differences in tolerance to be examined. The effect of alcohol was expressed as a log reduction in a suspension test. It’s not clear to me why the team didn’t perform more conventional MIC or MBC assays, rather than choose a fixed concentration of alcohol gel. Also, I think that a hard surface test (rather than a suspension test) would have been a more appropriate way to test alcohol susceptibility. Indeed, the levels of reduced susceptibility would almost certainly have been amplified in a hard surface testing, since bacteria attached to surfaces and in biofilms are typically 10s to 1000s of times less susceptible to disinfection.

Methodology quibbles aside, the E. faecium isolates after 2010 were 10-fold less susceptible to alcohol than the older isolates (i.e. a 1-log difference in the amount of E. faecium that survived at the sub-lethal alcohol exposure). At this point, you may be thinking this isn’t sounding hugely clinically relevant. But the authors constructed a clever surface transmission model to show that these moderate differences in alcohol tolerance may translate to differences in transmission in a clinical setting. Mouse cages were seeded with alcohol-susceptible and alcohol-tolerant E. faecium isolates. The cages were then disinfected using 70% alcohol wipes, and the alcohol-tolerant E. faecium isolates were more likely to establish gastrointestinal colonisation. Again, we can challenge the methodology here. The main concern with these findings is tolerance to alcohol gels, and yet the experiment was set up with a surface disinfection model. Alcohol wipes would not be used for this kind of disinfection task in the real world. Also, were the concentrations of bacteria used to seed the cages relevant? Nonetheless, some interesting findings as proof of concept.

Finally, the authors performed some ‘genomic fishing’ to identify genetic mutations linked with alcohol tolerance. They identified a few areas of the genome that may be involved in alcohol tolerance, but validation experiments of these findings were rather equivocal. They did identify some mutations that altered the growth rate in the presence of alcohol, but no ‘smoking gun’ alcohol tolerance gene / genetic determinant. This suggests that the alcohol tolerance pathway is complex and polygenetic.

There’s a parallel here with the issue of chlorhexidine susceptibility in MRSA. I was involved in a similar study looking at chlorhexidine susceptibility of a MRSA over an extended time period. We, too, found that some MRSA isolates were less susceptible to chlorhexidine than other MRSA isolates, and identified some genetic associations. But I am not clear whether or not these subtle reductions in susceptibility are clinically meaningful. Related to this, resistance to biocides is a very different beast to resistance to antibiotics. Resistance to antibiotics tends to be much more absolute, whereas resistance to biocides tends to be more subtle, with reduced susceptibility more common than resistance per se. This probably relates to the multiple molecular targets of biocides that are more difficult for bacteria to circumvent than the single molecular targets of antibiotics. The study on alcohol tolerance took the question a step further with the animal experiments and more sophisticated genomic fishing. But still the question remains whether these findings are clinically meaningful.

The authors conclude that ‘These findings suggest that bacterial adaptation is complicating infection control recommendations, necessitating additional procedures to prevent E. faecium from spreading in hospital settings.’ I think this conclusion pushes the findings of these laboratory-based findings several steps too far.