I’ve had the pleasure of a few days in Geneva enjoying some fine Swiss hospitality, and fine science at ICPIC 2023. Here’s rapid reflection on scaling the risks attached to plasmid-mediated transmission of CPE, and what we can do about it (or not…!).

The session was led by the brilliant Diego Andrey (Switzerland), Kalisvar Marimuthu (Singapore), and Gaud Catho (Switzerland). There’s now compelling and irrefutable evidence that plasmid-mediated transmission of carbapenemase happens in our hospitals:

• Outbreaks of CPE have been shown to be linked by common plasmids in different species.
• A huge genomic analysis of CPE over many years in Singapore illustrates, rather beautifully if frighteningly, that carbapenemase-encoding plasmids are often conserved between species.
• Genomic analysis shows hospital water systems are reservoirs for CPEs.
• Hospital wastewater often has a high level of CPE, even when CPE in patients is rare. Is this due to ‘silent colonisation’ of patients or staff, or contaminated water systems? However, some of the CPEs in these complex ecosystems are not found in humans – it seems that “environmental” Enterobacterales have acquired carbapenemases.
• Simulated studies provide compelling evidence that there’s a route from contaminated drains back to the patient, through tracking the spread of marked microbes.

From a technical point of view, there are some issues and outstanding questions in how we investigate possible plasmid-mediated carbapenemase transmission:

• Short-read sequencing is not suitable because it does not allow genome structure and thus plasmid complete circularization – long read sequencing is required.
• However, there are still some questions about how to define plasmid clonality. Is SNP distance relevant for plasmid evolution in the context of plasmid spread in the same way that it has become accepted for making conclusions about clonal spread? This question is foundational – all the evidence showing that plasmid-mediated carbapenemase transmission is a thing hangs on this inference about plasmid clonality.
• Who is appropriately skilled to do the genomics and bioinformatics here?
• How quickly can the necessary genomic work be done and fed back to frontline IPC teams?

What are the IPC implications?

• I think it’s fair to say we can no longer focus only on “same microbes / same carbapenemase” (i.e. clonal) outbreak definitions, epidemiological associations, and IPC interventions.
• Including “any microbe / same carbapenemase” (i.e. clonal + plasmid-mediated) in, say, and outbreak definition will effectively reduce the specificity of that outbreak definition. This could result in a diversion of already stretched resources, and make things worse not better.
• Classical IPC interventions target clonal transmission. However, this study from Singapore included a large set of CPE WGS showed that around 45% (yes almost half!) of the isolates in the dataset were linked by apparent plasmid-mediated transmission. To reinforce this, clonal transmission reduced in response to national IPC interventions, but plasmid transmission did not, which is what you’d expect because classic IPC interventions do not target plasmid-mediated transmission. Risk factors for clonal transmission = epi contact. No significant associations were risk factors for plasmid-mediated transmission?

Can modified IPC interventions tackle plasmid-mediated CPE transmission?

• The Singapore study found that epidemiological linkage was a significant risk factor for clonal transmission but did not find any significant associations with plasmid-mediated transmission (including epidemiological contact, medical device usage, and community-level variables). Put another way, there’s nothing obvious to “go for” to reduce plasmid-mediated transmission.
• Antimicrobial stewardship is a big factor here – if we make strides forward in reducing the use of key antibiotics, especially carbapenems, then the selective pressure reduces and the risk of CPE in general (whether clonal or plasmid-mediated) recedes (at least slightly).
• Despite no slam-dunk epidemiological evidence, all rounds point to the physical environment as the most important modifiable variable in plasmid-mediated transmission. So, what can we do to tackle the watery parts of our hospitals that we know are contaminated with a microbial soup including CPE:
  • Perhaps the most fail safe is to go entirely water-free, as has been done in some ICU settings where reductions in antibiotic-resistant Gram-negatives and all HCAIs have been demonstrated.
  • If we’re not ready for that rather drastic action, reducing the risk of dispersion from sinks etc can help. In one US study, providing an ultrasonic treatment (to dislodge biofilm) and covering the slop hopper resulted in a rather impressive reduction in CPE.
  • Improved design of sinks could help – and, perhaps, introducing the idea of self-disinfecting sinks, with periodic in-built thermal/chemical disinfection strategies will help.
  • Routine disinfection of sinks and drains. There are chemical, physical (e.g. vibration), and thermal options here. These approaches to sink disinfection has been rather underwhelming so far in terms of impact. However, especially for chemical disinfection, liquid chemicals poured down sinks and drains are unlikely to make much impact. Foam drain disinfection solutions are just coming to a point where they can be evaluated, so more evaluation to do here!
  • Reducing exposure of patients and equipment in the sink splash zone.
  • More education to improve our use of sinks, so that a little more than 4% of the use of hand hygiene sinks is actually for hand hygiene!

So where does this leave us? Plasmid-mediated transmission of CPE seems like an inconvenient yet irrefutable aspect of CPE transmission dynamics. We are at the early stages of working on what to do to address this issue, but it seems that better management of sinks and drains (or even getting rid of them completely) will be a vital part of the solution.