Although there’s some controversy about whether or not we should apply contact precautions (by that I mean single room isolation, enhanced PPE, enhanced disinfection etc) all the time for all organisms, it would be a brave hospital to eschew contact precautions for CPE carriers. And so the question of whether and when we should ‘de-isolate’ patients with known CPE is an FAQ. And so enter a recent study in CMI comparing the spontaneous apparent loss of colonisation with various CPEs, concluding that KPC carbapenemases seem to hang around for longer than NDM carbapenemases, but both almost always last for the duration of a single hospitalisation.

The study was performed in a huge, gargantuan 2,700 bed hospital in Korea. 147 patients were included, 106 with NDM-producing CPE and 41 with KPC-producing CPE. 11% of the NDM carriers spontaneously decolonised during their hospitalisation vs. none of the KPC carriers. Furthermore, 80% of the KPC carriers who were readmitted still carried CPE, compared with about a third of NDM carriers. On the face of it, the study suggests that KPC carriage is somehow more recalcitrant that NDM carriage. However…

• This could be a story about Klebsiella vs. other Enterobacteriaceae masquerading as a story about NDM vs. KPC. Let me explain. 95% of the KPC carriers were K. pneumoniae whereas only 36% for the NDM carriers. The key question is whether KPC carriage last longer than NDM carriage within K. pneumoniae CPE – and the dataset in this study simply isn’t big enough to answer this. I was surprised not to see more discussion on this important aspect throughout the study though.
• The measure of ‘clearance’ was three consecutive negative screens using conventional culture-based methods. We have recently grappled with the issues around carriage, clearance, masking, and acquisition for our study on CPE admission screening, concluding that it can be really tricky to tell whether a CPE is really there or not and that a single or even a series of negative screens can be a treacherous marker of ‘clearance’.
• I wonder how much difference adding a PCR-based method to detect carbapenemases would have made to this study? The feeling is that these are considerably more sensitive for detecting CPE.

This is a helpful piece of work, which adds to a limited information set about CPE clearance. But I don’t buy the key finding that the KPC carbapenemase is more recalcitrant than NDM based on the findings in this study – it seems much more likely to me that K. pneumoniae carriage is more recalcitrant than other Enterobacteriaceae. Still, the study shows that even using less-than-optimal (in terms of sensitivity) culture-based methods to detect CPE, spontaneous decolonisation during a hospital admission is rare indeed. And so the conclusion seems to be that patients should remain under contact precautions for the duration of a hospital admission – and have a high index of suspicion of CPE carriage at the time of their next admission.