Epidemiology of carbapenemase-producing Enterobacterales in England, May 2015–March 2019: national enhanced surveillance findings and approach

Summary

Background

In response to increasing numbers of carbapenemase-producing Enterobacterales (CPE) in England, Public Health England (PHE) launched an electronic reporting system (ERS) for the enhanced surveillance of carbapenemase-producing Gram-negative bacteria. Our study aimed to describe system engagement and the epidemiology of CPE in England.

Methods

Engagement with the ERS was assessed by calculating the proportion of referrals submitted this system. ERS data were extracted and cases defined as patients with CPE isolated from a screening or clinical specimen in England between 1^st May 2015 to 31^st March 2019. Descriptive summary statistics for each variable were prepared.

Results

The ERS processed 12,656 suspected CPE reports. Uptake of the ERS by local microbiology laboratories varied, with approximately 70% of referrals made via the ERS by April 2016; this steadily decreased after March 2018. Six-thousand eight-hundred and fifty-seven cases were included in the analysis. Most cases were from colonised patients (80.6%) rather than infected, and the majority were inpatients in acute hospital settings (87.3%). Carbapenemases were most frequently detected in Klebsiella pneumoniae (39.1%) and Escherichia coli (30.3%). The most frequently identified carbapenemase families were OXA-48-like (45.1%) and KPC (26.4%). Enhanced data variables were poorly completed.

Conclusions

The ERS has provided some insight into the epidemiology of CPE in England. An increasing number of routine diagnostic laboratories have introduced methods to routinely identify acquired carbapenemases and PHE has modified its approach to ensure robust surveillance, which is an essential aspect of an effective response to prevent and control the spread of CPE.

Background

Carbapenems are β-lactam antibiotics that have a broad spectrum of activity against both Gram-positive and Gram-negative bacteria [1]. They have been referred to as “antibiotics of last resort” due to their activity against multidrug-resistant bacteria [2]. Increasing levels of carbapenem resistance amongst Gram-negative bacteria in Europe have been well-documented in recent years [3,4]. Carbapenem resistance due to the emergence of acquired (plasmid-mediated) carbapenemases is of particular concern due to their ability to transfer within and between bacterial species and genera [[5], [6], [7]].

The clinical and public health threat posed by increasing antimicrobial resistance has been recognised by the UK government, exemplified by its inclusion on the national risk register [8] and the publication of a five-year national action plan for 2019 to 2024, which acknowledges difficulties associated with monitoring carbapenemase-producing Gram-negative bacteria and identifies the need for further enhanced surveillance [9].

The Antimicrobial Resistance and Healthcare Associated Infections (AMRHAI) Reference Unit at Public Health England (PHE) has documented year-on-year increases in referrals of carbapenemase-producing Enterobacterales (CPE) [10]; amongst isolates sent for investigation in 2018, 4,028 were identified as carbapenemase producers compared with 72 identified in 2009. In May 2015, in response to the observed increase, PHE launched an electronic reporting system (ERS) for the enhanced surveillance of carbapenemase-producing Gram-negative bacteria [11]. The system was designed to capture data on patient demographics, specimen details and risk factors in an electronic format. The aim of subsequent data analysis was to improve our understanding of the epidemiology of carbapenemase producers in England and inform infection prevention and control practices.

A number of communications were prepared and presented/circulated to potential system users to raise awareness prior to the launch of the ERS. Supporting documents, including a user guide and data dictionary were published online. A service desk was established to support users to register for the system and resolve any issues encountered.

The aims of our study were to assess engagement with the enhanced surveillance programme and to use data captured by the ERS to describe the epidemiology of CPE in England.

Methods

Surveillance system

A description of the ERS has been published previously [11]. The system captured data on patients colonised and/or infected with Gram-negative bacteria suspected of acquired carbapenemase production when isolates were submitted to regional or national reference laboratories for confirmatory testing. The system initially only collected data on isolates referred to AMRHAI or to PHE regional laboratories, but was modified in July 2016 to allow diagnostic laboratories to report cases of CPE identified locally.

ERS variables

The ERS comprised mandatory and optional fields, with the latter used for collecting enhanced data. Mandatory questions were focused on patient demographic information, laboratory details associated with the specimen, the healthcare setting the patient was in at the time of specimen collection and travel history in the previous 12 months (including healthcare received abroad). Following confirmation of patient colonisation or infection with an acquired carbapenemase producer, the referring laboratory was requested to complete the enhanced surveillance questions for inpatients. These fields included admission details, clinical specialty, whether the patient was screened on admission and previous contact with patients infected or colonised with carbapenemase-producing Gram-negative bacteria.

Data extraction and preparation

Data from the ERS were accessed from the main database via an Open Database Connectivity (ODBC) application programming interface established in Microsoft Excel. Relevant data fields were extracted from the main database through the creation of standard views. The dataset was imported into a statistical software package for further manipulation. Incomplete isolate submissions were excluded, and exact duplicates were removed.

Variables for carbapenemase test results were re-coded to align local, regional and national reference laboratory results i.e. unify response type recorded. In the event of discrepancies between local, regional and national reference laboratory results, the national result superseded all other findings. New variables for isolates with multiple carbapenemase mechanisms detected were created.

Analysis

Engagement with the system was assessed by calculating the proportion of referrals received by the AMRHAI Reference Unit via the ERS, with total number of referrals received used as the denominator (diagnostic laboratories were still able to make requests using paper referral forms in addition to the ERS).

Cases were defined as patients with CPE isolated from a screening or clinical specimen in England between 1^st May 2015 to 31^st March 2019. The study period was based on the original specimen date rather than referral date and cases were de-duplicated for each year of surveillance by bacterial species reported, specimen site and resistance mechanism(s) identified. For analytical purposes, surveillance years were defined as follows: May 2015–March 2016 (year one), April 2016–March 2017 (year two), April 2017–March 2018 (year three) and April 2018–March 2019 (year four).

Descriptive summary statistics for each variable were prepared and results from the analysis of enhanced surveillance variables are presented alongside the completeness of each of these variables. Data cleaning, manipulation and analysis were conducted in Stata/SE version 15 (StataCorp., USA.)

Ethical approval

Ethical approval was note required for this study.

Results

Engagement

Twelve-thousand six-hundred and fifty-six isolates of Enterobacterales suspected of producing acquired carbapenemases were referred via the ERS. Uptake of the ERS by local microbiology laboratories across England varied by region, with the highest engagement in laboratories based in the West Midlands (76.2% isolates referred) and the lowest in the East of England (32.8% isolates referred). By April 2016, overall approximately 70% of referrals were made via the system and this remained stable until March 2018 (Figure 1). After March 2018, the proportion of referrals received by the national reference laboratory steadily decreased, with less than 50% of isolates referred via the ERS by March 2019.

Figure 1.

Proportion of isolates referred to the AMRHAI Reference Unit via the ERS and number of CPE positive isolates identified at local, regional and national level, May 2015–March 2019.

Descriptive epidemiology

There were 6,857 cases included in the analysis; 1,914 cases in year one, 3,472 cases in year two, 3,903 cases in year three and 3,367 cases in year four. Patient demographic data are presented in Table 1. Five-thousand nine-hundred and eighty-eight (87.3%) cases were inpatients at National Health Service (NHS) acute trust hospitals at the time of specimen collection; of these, 5,972 (99.6%) records specified the NHS hospital trust where the patient was receiving healthcare at the time of specimen collection. The locations at time of specimen collection for the remaining 12.7% are presented in Table 1.

Table 1.

Descriptive epidemiology for cases of carbapenemase-producing Enterobacterales in England, May 2015–March 2019

Variable	Median (IQR)
Age (in years)	69.4 (54.6–80.1)
	Frequency	Percent (%)
Sex
Female	3164	46.1
Male	3679	53.7
Unknown	14	0.2
Residency
Home (UK resident)	4776	69.7
Nursing/care home	172	2.5
Non-UK resident	89	1.3
Temporary accommodation	9	0.1
Non-health-related institution	8	0.1
Other	32	0.5
Unknown	1771	25.8
Patient location at time of specimen
NHS acute trust (inpatient)	5,988	87.3
GP/walk-in centre	279	4.1
NHS acute trust (outpatient)	256	3.7
NHS acute trust (A&E)	251	3.7
NHS community trust (inpatient)	23	0.3
Nursing/care home	11	0.2
NHS community trust	6	0.1
Independent sector healthcare facility	2	0
Other	19	0.3
Unknown	22	0.3
Overseas travel in previous 12 months
Yes	604	8.8
No	699	10.2
Unknown	5033	73.4
No response given	521	7.6
Healthcare overseas in previous 12 months
Yes	427	6.2
No	726	10.6
Unknown	5186	75.6
No response given	518	7.6

IQR, interquartile range; NHS, National Health Service; A&E, accident and emergency.

Most isolates received were from screening specimens (5,527; 80.6%) rather than clinical specimens (1,318; 19.2%). The ratio of clinical-to-screening specimens remained consistent over time (Figure 2). The nature of the specimen type was unknown for 12 isolates. Of the clinical specimens, the most common specimen types were urine (693; 52.6%), blood (159; 12.0%) and sputum (87; 6.6%). Other specimen types reported included surgical wounds, bone, peritoneal fluid and pus specimens, amongst others.

Figure 2.

Number of CPE confirmed cases confirmed by regional/national laboratories or reported by diagnostic laboratories via the electronic reporting system, May 2015–March 2019.

Overseas travel history was unknown for most cases (81.0%; Table 1). Six-hundred and four (8.8%) cases reported overseas travel in the previous 12 months. Where available, the most frequently reported countries were India (170; 28.1%), Pakistan (74; 12.3%) and Egypt (36; 6.0%). Within Europe, the most commonly reported countries reported were Greece (33; 5.5%), Spain (27, 4.5%) and Romania (14; 2.3%). A similar pattern of reporting was seen in relation to whether cases had received healthcare overseas in the previous 12 months (Table 1).

Enhanced surveillance

Six-thousand and eleven inpatient cases (acute and community hospitals) were eligible for inclusion in the enhanced surveillance. Admission dates were only provided for 1,623 (27.0%) cases. Information on the nature of admission to hospital was provided for 836 cases (13.9%). Of those, most were emergency admissions (615; 75.6%) followed by elective (99, 11.8%) and then inter-hospital transfers (95, 11.4%).

Clinical specialty was provided for 1,150 cases (19.1%). The most commonly reported specialties were general surgery (168; 14.6%), general medicine (152; 13.2%), A&E (97; 8.4%), geriatric medicine (95; 8.3%) and critical care (74; 6.4%).

Information on admission screening for CPE was provided for 1,768 cases (29.4%); 909 (51.4%) had been screened on admission (with 233 (25.6%) reported as being CPE positive), 280 (15.8%) had not and the rest (579; 32.7%) were reported as unknown. Of the 579 cases reported as “unknown”, the specimen type was reported as “screening” for 563 (91.4%).

Due to a technical failure, data on whether a patient had been in contact with a known CPE positive patient within a healthcare setting were not captured by the ERS. Data relating to potential contact with CPE outside the healthcare setting was only reported for a small number of cases (81, 1.3%).

Microbiology

Carbapenemases were most frequently detected in Klebsiella pneumoniae (2,681, 39.1%), Escherichia coli (2,080, 30.3%) and Enterobacter cloacae complex (996, 14.1%). The most frequently identified carbapenemase families were OXA-48-like (3,091, 45.1%), KPC (1,813, 26.4%) and NDM (1,444, 21.1%). Figure 3 summarises the frequency of carbapenemase families across all genera reported.

Figure 3.

Resistance mechanisms identified in isolates submitted to regional/national laboratories or reported by diagnostic laboratories via the electronic reporting system, England, May 2015–March 2019.

Discussion

The enhanced surveillance system for carbapenemase-producing Gram-negative bacteria was launched in May 2015. Though uptake by local laboratories was slow initially, good coverage was achieved and allowed us to gain some insight into the epidemiology of carbapenemase producers in England. Data collected via the ERS has been used in monthly reports to communicate CPE activity in hospital trusts across England. The provision of this information supported recommendations made in the ‘Acute trust toolkit for the early detection, management and control of carbapenemase-producing Enterobacteriaceae’ [12].

Maintaining high levels of engagement with the ERS became more challenging when commercial methods for the identification of acquired carbapenemases became more widely available. However, we found that the majority of records captured by the ERS were from laboratories reporting locally-identified carbapenemase producers on a voluntary basis. We believe that the routine publication of results from the system was an important factor in this continued engagement. We believe it is essential that any national surveillance system should produce timely and informative outputs to enable appropriate action.

A survey conducted by the European Centre for Diseases Surveillance and Control in 2018 reported that 27 EU/EEA countries had established surveillance programmes for carbapenem-resistant Enterobacterales [4]. Systems ranged from the voluntary reporting of individual cases to the mandatory reporting of outbreaks caused by carbapenem-resistant Enterobacterales. In 2018 the US Centre for Disease Control and Prevention launched a programme to detect CPE via a national laboratory network [13]. This scheme focuses on the most prevalent carbapenemases in the United States: KPC, NDM, VIM, IMP and OXA-48-like enzymes. Similar to the ERS, this scheme was based on the referral of isolates to laboratories and therefore may not be representative of the true epidemiology of CPE. Further to this, several states have introduced reporting requirements for CPE [14].

Results from our analysis have revealed that a high proportion of cases were colonised with CPE (80.6%) rather than infected, and most were inpatients in an acute hospital setting (87.3%). These findings likely reflect the screening recommendations made in ‘Acute trust toolkit for the early detection, management and control of carbapenemase-producing Enterobacteriaceae’, which advises acute trusts to screen patients at increased risk of CPE colonisation on admission to hospital [12]. Carbapenemases were most frequently detected in K. pneumoniae and E. coli, with OXA-48-like enzymes identified most frequently across all Enterobacterales (45.1%).

It is important that system evaluations are performed to assess whether systems are meeting their objectives and inform the need for a modified approach where necessary. An evaluation of the ERS found that enhanced surveillance fields were poorly completed [15]. This is an important limitation as it has affected our ability to identify high-risk patient groups and inform the development and implementation of effective prevention and control measures. Similar systems operated by PHE have had more success in capturing enhanced surveillance data, though this information is still only provided for a proportion of cases. For example, data captured as part of the national E. coli bacteraemia mandatory surveillance scheme have been analysed and it was found that a large proportion (41.1%) of cases reported urinary tract infections (UTI) as the primary source of infection [16]. These findings have been used to inform prevention and control efforts, with the specific recommendation that interventions to reduce E. coli bacteraemia should focus on management of UTI in the community setting and comprehensive guidance have been published subsequently [17].

Participation in the enhanced surveillance programme was voluntary and some areas with high CPE prevalence conducting their own local testing for acquired carbapenemases did not participate due to the resource required to report results via the system [15]. The variation in participation across the country limits the representativeness of our findings. However, the finding that KPC and OXA-48-like enzymes were the most frequently identified is consistent with the epidemiology in other northern European countries [18].

Since the launch of the ERS an increasing number of routine diagnostic laboratories have introduced methods to routinely identify acquired carbapenemases. A survey conducted by PHE in 2018 found that nearly 50% of diagnostic laboratories had the capacity to identify and discriminate between the most prevalent carbapenemase families [10]. As a result, engagement with the ERS began to reduce from January 2018. Furthermore, with increased availability of commercial assay for the detection of carbapenemases, the national reference laboratory changed its referral criteria in January 2019 and requested that only acquired carbapenemases producers isolated from sterile sites be submitted.

Conclusions

In light of our study findings and changes to the identification and referral of suspected CPE, PHE has modified its approach to CPE surveillance. The ERS was closed on 30^th April 2019 and modifications to PHE's national microbiology surveillance database (the Second Generation Surveillance System; SGSS) have been made to support a more efficient and streamlined method for diagnostic laboratories to report acquired carbapenemase producers to PHE. This system is based on voluntary reporting of core microbiology data. However, as published in ‘The UK's five-year national action plan’ to prevent and control antimicrobial resistance, we have initiated the process to add acquired carbapenemase-producing Gram-negative bacteria to the Health Protection Regulations, making it a legal requirement for laboratories to report these organisms to PHE [9].

This new approach will facilitate linkage to other datasets, such as Hospital Episode Statistics, and will be vital in improving our understanding of the epidemiology of CPE in England. Furthermore, this approach offers a solution that minimises manual data entry and requires much less resource than case-capture systems. Where electronic data is not available routinely, sentinel surveillance schemes and point prevalence surveys are planned to capture risk factor data.

Representative and robust surveillance is required to improve our understanding of the epidemiology of CPE in England. High-quality surveillance data are essential in allowing us to identify and implement effective prevention and control interventions and reduce the threat posed by CPE.

CRediT authorship contribution statement

Rachel Freeman: Conceptualization, Methodology, Data curation, Formal analysis, Writing - original draft, Visualization. Dean Ironmonger: Conceptualization, Methodology, Data curation, Writing - review & editing. Katie L. Hopkins: Investigation, Resources, Writing - review & editing. Richard Puleston: Conceptualization, Writing - review & editing. Peter Staves: Software, Data curation, Formal analysis. Russell Hope: Methodology, Writing - review & editing. Berit Muller-Pebody: Writing - review & editing. Colin S. Brown: Writing - review & editing. Susan Hopkins: Conceptualization, Methodology, Supervision. Alan P. Johnson: Writing - review & editing. Neil Woodford: Investigation, Resources, Writing - review & editing. Isabel Oliver: Conceptualization, Supervision, Writing - review & editing.

Conflict of interest statement

None.

Funding source

This work was funded as part of Public Health England’s routine service.