I blogged recently about the new ESCMID guidelines on resistant Gram-negative carriage and decolonisation, which supported a “once positive, always positive” approach to CPE carriers due to the lack of effective decolonisation options. A new study suggests that a large majority (75%) of patients who were once identified as CPE carriers no longer had CPE detectable when they were readmitted. This has implications for the management of CPE carriers in hospitals.

The study included all 1194 readmissions of 607 patients who were previously identified as carrying (or infected) with KPC-producing CPE at Manchester University NHS Foundation Trust for about 18 months in 2016/17. Only 17.1% of the 1194 admissions were identified as CPE positive (using PCR direct from the screening specimen) at the time of their readmission. Serial screens were used throughout the stay of the patients (every 72hrs during their stay). A small proportion (8%) of patients who were CPE negative at their admission had a subsequent positive screen. Therefore, overall, just 23% of the previous CPE carriers were identified as “active” CPE carriers. Of the readmissions with CPE detected, 89% were detected within three screens each separated by 72 hours.

The study was performed in Manchester, which has long-standing issues with KPC-producing K. pneumoniae; in fact, in a fairly recent PPS in the Manchester University NHS Foundation Trust, 11% of inpatients carried CPE, most of which was KPC-producing K. pneumoniae. In a setting where 10% of patients carry CPE vs. one where, for example, 1% of patients carry CPE, the decision-making around whether to pursue a “once positive, always positive” approach to managing CPE colonisation is different. If 1 in 10 patients are carriers, there’s no point trying to isolate everybody all of the time – you’d run out of isolation facilities in an instant. In this setting, you need a risk-stratified approach (e.g. only isolate patients with positives from non-screening specimens, or recent positives from surveillance cultures). However, in the 1% prevalence setting, you may still be able to pursue effectively a “once positive, always positive” approach.

In terms of serial screening, the finding that a small proportion of patients were initially screen-negative and subsequently screen-positive is similar to our findings in London. We found that the rate of carriage of resistant Gram-negative bacteria increased sequentially over the course of a hospitalisation – perhaps due to unmasking of pre-existing carriage through various exposures (e.g. antibiotics) or the acquisition of new resistant bacteria. Either way, this argues for regular screening of “high-risk” patients or all patients in “high risk” settings throughout a hospitalisation for the carriage of CPE and other resistant Gram-negative bacteria. (Answers on a postcard please to define “high risk”.)

So, should we apply a “once positive, always positive” approach to managing CPE carriers in acute hospitals? Well, that depends on your prevalence. I can feel a model coming on…!