The team at Barts Health, one of the largest NHS hospital groups in the country, has published the findings of a point prevalence screen of all inpatients for carbapenemase-producing organism (CPO) carriage. Overall, 30 (3.1%) of the 977 patient tested were carrying 35 different CPOs (all but one of which were CPE). Risk factors for CPO carriage included hospitalisation abroad, any hospitalisation, and overseas travel (especially to India, Pakistan, and Bangladesh). These findings help us to understand an emerging picture of CPO in the UK.
A total of 977 patients were consented into the study, which was 58% of the available inpatients. Overall, 30 (3.1%) of the 977 patient tested were carrying 35 different CPOs. 40% were K. pneumoniae, 31% were E. coli, 17% were E. cloacae, and only 1 was a non-Enterobacterales (Pseudomonas sp). NDM accounted for 60% and OXA-41 for 40%. The number of patients with CRO that did not have a detectable carbapenemase was not reported. Prevalence by medical speciality was not reported, but prevalence was noted to be higher in renal and elderly care and lower in ICU.
The hospital group dealt with 94 CPO infections during the 13 month PPS window. Whilst, as the authors point out, this accounts for only 0.05% of all infections dealt with during this period, it would be interesting to see the data on the ratio of the number of CPO detected by screening and the number detected in clinical cultures.
Some other reflections / limitations:
- The study was conducted over 13 months, so doesn’t provide a point-in-time snapshot of carriage.
- For routine surveillance during this period, the hospital group was using a risk-factor based admission screening approach based on the PHE Toolkit, where patients who have been hospitalised abroad, been an inpatient in the UK hospital where there are reported cases of CPE (isn’t this most of them?!), or have a history of CPE carriage are eligible for screening. It’s not clear how many of the 30 carriers identified through the point prevalence screening would have been detected through routine surveillance screening.
- I was interested to see that the team decided to enrich the swabs before culture – this would have improved the chances of detecting carriers, but probably not feasible for routine workflows.
- 42% of inpatients were not enrolled into the study – it would have been useful to know how many were unable vs. unwilling to provide consent. The methods excluded paediatric patients, and, perhaps more importantly, patients not able to consent for themselves. This seems important, because this group of debilitated patients are also likely to have the most risk factors for antibiotic-resistant organism carriage. When we ran a study of carriage for CPO at Guy’s and St. Thomas’, we had a long and protracted discussion with the ethics committee about why it was so important to include patients who were not able to consent for themselves in the study, so as not to skew the findings. So, it’s interesting to see that carriage in the intensive care unit was low – but I assume that this was in patients in the ICU who were able to consent for themselves? In this same study, we had to work hard on getting our message to patients optimised in order to improve patient enrolment.
- The stats seem rather limited. As far as I can tell, they ran only univarable analysis. It’s odd that a multivariable model was not developed, which would have helped to interpret the relative importance of each variable in predicting CPO carriage.
- It is clear that the team had some issues with data collection, meaning that risk factor data only available for 57% of the patients: ‘For each of the participating wards, the risk factor study sheet was partially or totally incomplete due to resistance and time constraints of the nursing staff.’
The findings fall in the mid-range so far as UK CPO prevalence screens go:
|N patients screened||N CPO carriers||% CPO carriers|
|Barts (this study)||977||30||3.1|
I wonder whether these findings are likely to underestimate true prevalence, mainly due to the exclusion of patients not able to consent for themselves. However, the findings are helpful in providing some information on the prevalence of CPE carriage in this part of London. And demonstrate that by global standards, the carriage of CPE remains a fairly rare event in most parts of the UK (in contrast to some other parts of the world).