Addenbrookes hospital in Cambridge (UK) have recently performed a point prevalence survey for antibiotic resistant bacteria. None of 540 patient samples grew CPE, but 130 (24%) grew VRE. So, why no CPE but so much VRE?
All adult inpatients were included as the target population over three days. Around 16% of these patients declined to be screened. (Interestingly, in a comparable study, we managed to get the decline rate down to <5% through a simplification and change in messaging to the patients.) A rectal swab or stool specimen was collected from 577 patients: all 577 were tested for CPE, and no CPE was detected. 540 were sampled for VRE, and 130 (24%) grew VRE.
In Manchester, an arresting point prevalence screen of all inpatients identified 70 CPE positive cases (26 newly identified and 44 previously known) and 592 CPE negative cases, giving a combined prevalence of 11%. The finding that approximately one third of cases were not known to be colonised suggested that these ‘silent carriers’ could be contributing to transmission. This prompted a broadening of the local screening strategy to capture more at-risk patients. However, this study found no CPE in any of the samples tested. This is a surprising finding to me – although in line with the low prevalence of CPE detected on admission to a London hospital a few years ago. The team evaluated the ‘prevalence’ of CPE risk factors, finding that whilst many patients had been hospitalised in the past year (two thirds in fact), only a handful of patients had been hospitalised in London (0.8%), abroad (0.6%), and no patients had been in hospital in the North West in the past year. Perhaps this reinforces that CPE prevalence is patchy currently in the UK, and if you see few patients from abroad and are in a referral network where CPE is rare, you’re unlikely to see much of it?
VRE on the other hand…was pretty much everywhere. Around a quarter of patients overall carried VRE – and VRE was detected on 87% of the 39 wards that were included. The rate of VRE carriage on a ward level exceeded 50% on three wards. As you can see in the table below, VRE was detected across many different specialties – unhelpfully the paper did not supply the denominator to calculate ward-level rates.
|Speciality||Number of patients positive for VRE|
|Hepatobiliary and pancreatic surgery||8|
|Upper gastrointestinal surgery||3|
|Critical care medicine||1|
|Ear, nose and throat||1|
|Renal transplant follow-up||1|
|Small bowel assessment||1|
So why so much VRE? This was not a clonal outbreak affecting the whole hospital: there were 48 different antibiograms amongst the 130 VRE isolates. Is VRE colonisation essentially endemic in the patient population in this hospital (and probably in all comparable hospitals in the country)? These findings would suggest so. Findings from Ireland of a comparable carriage rate of VRE (19%) amongst inpatients add further weight to the interpretation that VRE carriage is likely to be endemic. An effort to contain and reverse the VRE colonisation picture would be a gargantuan effort – and you’d question whether it would be worth the investment.
This is a bit of a Jekyll and Hyde study, with a surprisingly reassuring outcome in terms of CPE carriage, but an equally surprisingly bleak outcome in terms of VRE carriage. The findings are helpful in providing data to configure local screening approaches for AMR bacteria.