Examining the impact and response to an outbreak of carbapenemase-producing Enterobacterales in a neonatal unit in the United Kingdom: An outbreak report

Abstract

Background

Carbapenemase-producing Enterobacterales (CPE) are a group of Gram-negative bacteria causing global concern due to their resistance to carbapenems. In this report, we detail the learning points from a CPE outbreak in a tertiary neonatal unit (NU) in the UK.

Methods

Routine surveillance screening (rectal swabs) of babies on the NU identified a potential cluster of CPE carriage. Samples were sent to a reference laboratory for confirmatory testing. Environmental screening and cot mapping were undertaken to determine movements of babies within the unit. Regular audits of cleaning standards, hand hygiene, and maternal hygiene when expressing breast milk were carried out.

Results

The outbreak lasted 19 weeks. During the outbreak, there were 360 admissions, with 11 babies being colonised with the outbreak strain. Once the outbreak was declared, there were enhanced Infection Prevention and Control (IPC) precautions (including increased environmental and equipment cleaning frequency). CPE screening frequency was increased and cot capacity was reduced. Hand hygiene compliance improved from 92% at the start of the outbreak to 100% by its close. Cleaning standards remained compliant. Maternal hygiene standards varied from 78% to 100%, but no cross-infection links were identified. Environmental screening was negative. No route of cross-infection was identified. Notably, no babies developed invasive CPE infection.

Conclusion

This is the first report of a CPE outbreak in a UK NU. Although no specific mode of cross-transmission was identified and the outbreak’s end cannot be attributed to any single intervention, the bundle of interventions proved successful after a 5-month period.

Background

Enterobacterales are a family of potentially pathogenic Gram-negative bacteria that often reside in the gut. They can translocate into the systemic circulation and cause serious infection (Morales-López et al., 2019). They are easily transmitted through faeco-oral spread, contaminated equipment, or inadequate hand hygiene and may therefore be common sources of hospital and community-acquired infections (Nordmann et al., 2011). Carbapenemase-producing Enterobacterales (CPE) have acquired the ability to produce carbapenemases. These are beta-lactamase enzymes that hydrolyse carbapenem antibiotics. CPE emerged as organisms of concern in the early 21^st century, when Klebsiella pneumoniae-producing KPC-type carbapenemase spread rapidly worldwide (Doi and Paterson, 2015). Further variants of these Enterobacterales were then discovered, including the New Delhi metallo-β-lactamase (NDM)-producing type, which is now endemic in the Indian subcontinent (Lee et al., 2016). Certain carbapenemase genes can be transferred between different species of bacteria in the form of mobile genetic elements, which increases transmission of resistance (Queenan and Bush, 2007). Carbapenems (e.g. meropenem and ertapenem) are bactericidal beta-lactam antibiotics that inhibit bacterial cell wall synthesis. They are often used as ‘last-resort’ antimicrobial agents in cases of multi-drug-resistant bacterial infection (Papp-Wallace et al., 2011). Therefore, the emergence of carbapenemase-producing organisms is of great clinical concern.

Mortality is higher in infections caused by CPE than in infections caused by Enterobacterales susceptible to carbapenems in the adult population (OR 3.39, 95% CI 2.35–4.89) (Borer et al., 2009; Martin et al., 2018; Patel et al., 2008). The burden of disease caused by CPE in children is poorly documented, especially in neonates (Chiotos et al., 2016). Colonisation with CPE predisposes babies to systemic infection and once CPE-colonised, this must be declared in future healthcare encounters so that measures can be implemented to mitigate spread. Since Enterobacterales are gut organisms, decolonisation is not achievable. A neonatal unit (NU) in Pakistan investigating neonatal sepsis due to Klebsiella pneumoniae found that 72% of cases were carbapenemase-producers (Saleem et al., 2013), while another study detailed the high mortality rate after a CPE outbreak at a Nepalese NU (Stoesser et al., 2014). Varying risk factors for transmission have been described in the literature, including maternal colonisation (Seesahai et al., 2021), previous antibiotic exposure (Singh et al., 2018), low birth weight, and invasive devices (Decembrino et al., 2014). However, it has been noted that colonisation risk factors are not completely determined (Almeida et al., 2021). As per the Health Protection Surveillance Centre in Ireland, CPE outbreaks can be defined as two or more confirmed cases of the same carbapenemase-producer (causing infection or colonisation) that are linked epidemiologically in time and place (Health Protection Surveillance Centre, 2019), with the Ontario Public Health Infectious Disease Protocol defining this time as 3 months (Ministry of Health, 2022). Data on CPE outbreaks in a neonatal population in higher-income countries are limited, with no previous reports found for NUs in the United Kingdom (UK).

This report details a CPE outbreak at a regional NU in London, UK. We describe the response to the outbreak, with the hope that this can provide transferable learning.

Methods

Overview

This retrospective report details a CPE outbreak in a tertiary NU at a London Hospital. The outbreak lasted 19 weeks, from 04/10/2018 to 14/02/2019. The last reported case was on 27/01/2019.

Case definition

A confirmed case was a baby with NDM Enterobacter cloacae isolated from clinical/screening specimens, where the organism was identified as belonging to the same outbreak strain by Pulse Field Gel Electrophoresis (PFGE) typing at the reference laboratory.

Possible cases were babies with NDM Enterobacter cloacae isolated from clinical/screening specimens. These isolates were further analysed by PFGE typing at the reference laboratory to determine whether they were the ‘outbreak strain’ or ‘unique’ strains.

Data collection

Data were collected using Badgernet, a NU electronic patient record (EPR), and hospital EPR. Cot mapping was done to determine movements of babies during the outbreak.

Risk factors for transmission were identified through regular audits. These included:

• • An audit assessing maternal practice in expressing breast milk

• • Hand hygiene audits for healthcare professionals and visitors

• • Environmental audits that were undertaken thrice weekly during the outbreak

Population

During the outbreak, there were 360 admissions, of which 11 babies were colonised with the outbreak strain (Table 1). Two further babies were found to have ‘unique’ strains of NDM Enterobacter cloacae on PFGE typing. Babies were admitted either directly post-delivery or transferred from another hospital. All babies were screened at the point of admission to the NU. Any screening results from other hospitals were communicated to the NU prior to admission. All babies in the unit are routinely screened consistently once a week on Sunday night.

Table 1.

Clinical details of babies with CPE colonisation with outbreak strain.

Baby	In born vs. postnatal transfer	Gestational age (weeks = days)	Birth weight (grams)	Sex	Ethnicity	Days admitted to neonatal unit	Day of admission CPE confirmed	Respiratory support on the day of CPE confirmation
1	In born	25+3	850	Male	Asian Indian	95 a	46	CPAP
2	Postnatal transfer	27+6	780	Female	White British	60 b	20	Self-ventilating in air
3	Postnatal transfer	27+6	891	Female	White British	60 b	20	Vapotherm
4	In born	24+1	568	Female	Black Caribbean	167 a	138	Vapotherm
5	In born	25+3	642	Female	Black African	45 a	18	HFOV
6	In born	41+3	3540	Male	White other	11 b	7	Self-ventilating in air
7	In born	36+5	2390	Male	White other	7 b	6	Self-ventilating in air
8	In born	35+7	2520	Male	White British	14 b	8	Self-ventilating in air
9	In born	41+4	3465	Male	White British	18 b	17	Vapotherm
10	Postnatal transfer	31+3	1675	Male	White British	23 a	21	Vapotherm
11	In born	39+2	2640	Male	White British	17 b	5	Vapotherm

Vapotherm – High flow oxygen, HFOV – High Frequency Oscillatory Ventilation, CPAP- Continuous Positive Airway Pressure.

^aTransferred back to local hospital.

^bDischarged home.

Investigations

Laboratory investigations involved analysing rectal swabs or faecal samples taken as part of routine screening in the NU. Routine screening occurred once each week. Samples were cultured on chromogenic culture media. After potential colonies were identified, further identification was carried out using Matrix-assisted laser desorption/ionization-time of flight (MALDI-tof) mass spectrometry analysis, plate-based antimicrobial susceptibility profile, and immunochromatographic lateral flow testing for carbapenemase-specific antigens. All isolates identified as NDM Enterobacter cloacae were sent to the reference laboratory for PFGE typing.

Environmental screening was also done by taking swabs from the sinks of the room where cot mapping linked most outbreak cases. Samples were processed in the same manner as the clinical specimens.

Interventions

Once the outbreak was declared, regular multidisciplinary team (MDT) outbreak meetings were held. Public Health England (PHE) was informed. The outbreak was reported to the London Neonatal Operational Delivery Network (ODN) to highlight potential cot capacity reductions.

Awareness of the outbreak was raised among both staff and visitors. A letter detailing the enhanced infection prevention and control (IPC) precautions was circulated among the NU staff. Parents of admitted babies were informed with a letter that highlighted the IPC measures they needed to do when attending the NU. A high level of transparency was ensured through regular communication at team handovers and huddles across the relevant care areas.

Our NU has multiple routes of patient flow with high levels of movement (Figure 1). Because of the outbreak, a reduction in intensive care (IC) cot capacity was implemented with restrictions on admission flows into the service. Cot capacity was reduced to support dedicated isolation nurseries, increase cot spacing and reduce activity flows of very sick and preterm babies being admitted for intensive care, thus reducing the risk of cross-infection. In addition, although cot capacity was reduced, nurse staffing remained at the usual full capacity model, supporting higher nurse to baby ratios. The unit was closed to all, apart from medically essential admissions. External ex-utero and in-utero admissions were transferred to other units. Any babies that were transferred back to local hospitals had to be admitted to a side room at their receiving unit given the outbreak.

Figure 1.

Flow chart showing admission and discharge routes at the NU.

Any confirmed or possible cases were cared for in a separate CPE cohort room, with a total capacity of six cot spaces.

Enhanced CPE screening was carried out, with twice-weekly rectal screening for multi-resistant Gram-negative rods (MRGNR) for all babies admitted to the NU.

Gloves and gowns were worn when carrying out close personal care for CPE-positive babies, with gloves and aprons worn when providing care for all other babies. All personal protective equipment (PPE), including gloves, gowns, and aprons were single use. Daily PPE and hand hygiene audits were commenced for staff and parents entering the NU.

There was enhanced cleaning of the NU environment, with whole unit cleaning occurring twice daily. Enhanced cleaning consisted of detergent and hot water, dried with single-use paper towels, followed by all surfaces being wiped over with Clinell® wipes. Hydrogen peroxide vapour (HPV) was used for terminal end-cleaning of nurseries. The standards of environmental cleaning were regularly audited.

Ethics

Research Ethics approval was not required as we used fully anonymised routinely collected clinical data and information from outbreak investigation reports.

Results

A detailed timeline of events, along with key interventions, is demonstrated in Figure 2. On 04/10/2018, NDM Enterobacter cloacae was identified in 3 out of 31 babies (Table 1, babies 1, 2, and 3) and a possible outbreak was declared. These babies were isolated on the same day, with screening sent on four other babies sharing the room. Daily screening of the four unaffected babies was carried out over 3 days from the date of the first positive test. The weekly routine screen in the NU was extended to include MRGNR from rectal swabs. This extended screening revealed five additional babies that were outbreak strain positive. Cot capacity in the NU was reduced with restrictions placed on the admission of extremely preterm babies. Environmental swabs were taken from the room where most CPE cases were found, although no CPEs were cultured from these swabs. On 05/12/2018, another baby tested positive. Two further babies tested positive for CPE colonisation on 16/12/2018 and 14/01/2019, but these were found to be two ‘unique’ strains. Another baby tested positive for the outbreak strain on 02/01/2019, with the final case being noted on 27/01/2019. There were no further cases of CPE identified in the subsequent 19-day period. The outbreak was closed on 14/02/2019, as defined by the hospitals Director of IPC (DIPC), and with expert external support from National Health Service England (NHSE).

Figure 2.

Timeline highlighting the key events during the outbreak. Red boxes indicate when CPE outbreak positive cases were found, yellow boxes indicate unique CPE cases and green boxes highlight key interventions performed. Blue boxes are for all other events.

Mapping of cot movements (Figure 3) revealed a link to cot space 14, despite terminal cleaning of the cot space in between admissions. Four cases (babies 3, 4, 7, and 8) had contact with this cot space. Deep cleaning of the room was undertaken using an HPV regime once all babies were discharged or transferred to side rooms. Despite this, a further positive screen occurred. The cot space 14 was very close to the sink in the room that was used for hand washing. A fluorescent dye was used on the sink to observe where the water droplets would fall during use. Extensive droplets of water were observed on the floor and on the outside of the incubator in cot space 14. This finding was discussed with the external HPA Consultant and the MDT outbreak team. Circulation space in ITU room 2 was also tight and reducing cot capacity in the room was felt as valuable, so the cot space was subsequently closed for the duration of the outbreak. Enhanced hand hygiene audits showed initial compliance on 24/10/2018 at 92%, which improved to 100% by 15/02/2022. As all colonised babies were feeding on expressed breast milk, maternal hygiene when expressing breast milk was audited. The hygiene standards varied from 78% to 100%, but no cross-infection links were identified. Environmental audits showed compliance with scores ranging from 98% at the start of the outbreak to 100% by its end.

Figure 3.

Diagram showing all cot movements of the 11 outbreak strain CPE colonised babies during their admission in the NU. Dates at which each baby had a positive screen are noted in the table. Squares indicate the cot the baby was first placed in, while triangles indicate the final cot the baby was placed in during their admission. Circles indicate all other cots the babies were placed in. The time at which each case had a positive screen is indicated by the icon outlined in red with striped centre.

The index case was not determined despite extensive mapping, although it is likely to be either baby 1, 2, or 3. These were the first three babies to have a positive result with the outbreak strain and all three were present in the same room. Similarly, the root cause for the outbreak and reasons for onward transmission were not identified. It was noted that in 2018, there were 18 new nursing staff that came from areas of higher CPE prevalence, such as India and the Philippines. However, staff screening was not recommended by PHE and was therefore not carried out. Other potential contributory factors included that the unit did not have a single standard operating procedure for equipment cleaning that covered all the medical devices used on the unit. Weaknesses in cleaning some multi-use devices, such as ultrasound scanners, were noted and storage space for these devices was thought to be inadequate. However, notably, no babies developed a CPE infection, and the outbreak was limited to colonisation. The epidemic curve shown in Figure 4 demonstrates the distribution of outbreak cases over time.

Figure 4.

Epidemic curve showing the number of babies colonised with the CPE outbreak strain over time.

Discussion

This report describes a CPE outbreak in a large UK NU, detailing the IPC interventions undertaken. The rise of multi-drug-resistant bacteria is a global concern, with CPE posing a particular risk given the role of carbapenems as a last-resort antibiotic. Given the high mortality of CPE infections, especially in the context of neonates, it is critical that colonisation is minimised and progression to infection curtailed. The interventions described in this report are not novel. However, reflecting on the outbreak has allowed for a series of learning points to be developed.

Once the outbreak was declared, an enhanced cleaning programme was implemented. This included the use of HPV for surface disinfection after standard terminal cleaning of the room. Hydrogen peroxide is more oxidising than chlorine and is believed to release oxygen free radicals that damage the bacterial genome (Totaro et al., 2020). Indeed, a study investigating the standards of surface cleaning in an NU found that use of HPV resulted in no colony-forming units (CFUs) after HPV use compared to standard (24 CFUs [EI 0-635]) and enhanced (3 CFUs [EI 0-35]) regimes. In this study, meticillin-resistant Staphylococcus aureus (MRSA) and extended-spectrum beta-lactamase Klebsiella pneumonia that were found after the standard cleaning regime were completely eradicated with the HPV regime (Chiguer et al., 2019). However, in our outbreak, two further babies became colonised after being nursed in cot space 9, which was directly opposite cot space 14 (where there were links with previous cases), despite prior implementation of HPV cleaning. This cot space was close to the handwashing sink, which had been suggested in other outbreaks to be a link for infection spread (Haas and Trezza, 2002). This perhaps suggests a limitation in the effectiveness or execution of these regimes since HPV is not effective on surfaces that are not directly exposed to the HPV.

Throughout the outbreak, audits were conducted at regular intervals to examine compliance with IPC measures, such as enhanced hand hygiene. This helped identify any potential sources of spread, while continuously maintaining staff and visitor awareness. The benefits of audits in preventing healthcare-associated infection have been well documented (Hay, 2006). The regularity with which these audits were conducted allowed for rapid feedback and the timely implementation of any changes, while encouraging constant positive behavioural change.

All cases during this outbreak were colonisations rather than CPE infections. This is similar to a recent Portuguese study (Almeida et al., 2021), where similar measures were applied during the outbreak. These included enhanced IPC measures, systemic microbiological screening, improved hand hygiene, and PPE use. Other literature for CPE outbreaks in NUs worldwide reports infection rather than colonisation, with high mortality rates observed, although the majority were in lower-income settings (Stoesser et al., 2014). In adults, the risk of infection after CPE colonisation has been evaluated with one systematic review suggesting an infection rate of 16.5% (Tischendorf et al., 2016). However, risk factors contributing to the progression from colonisation to infection have not been widely reported. This is something that warrants further investigation given the high mortality rate associated with invasive infection and the difficulties selecting appropriate antimicrobial therapies.

A study investigating carbapenemase-producing Klebsiella pneumonia colonisation in a Vietnamese NU found that 55% of patients that were negative for the bacteria on admission became positive by discharge, suggesting that the colonisation occurred during their stay on the NU (Berglund et al., 2021). The NU is a high turnover environment, and the need to limit spread of multi-drug-resistant bacteria is crucial. This report aims to describe the interventions that were used to help limit the spread of the CPE outbreak strain and to share the subsequent learning. Although no specific mode of cross-transmission was identified and the end of the outbreak cannot be attributed to any single intervention, the bundle of interventions implemented proved successful in limiting further spread after a 5-month period. However, it should be noted that in trying to limit the outbreak, as per guidance from the Department of Health (Anthony et al., 2013), cot capacity had to be reduced. This can have a serious impact on the neonatal service, with potentially wide-ranging effects on neonatal care in the area.

The limitations of this report includes that the root cause and route of cross-transmission for the outbreak were not identified. These limitations stemmed partly from inabilities to identify whether mothers were colonised with CPE or whether there was an unidentified nidus of ongoing environmental contamination despite the enhanced IPC and prolonged enhanced cleaning measures. Maternal colonisation has been suggested as a risk factor for CPE colonisation in the newborn, with one study recommending screening for high-risk mothers, including those that received hospital care in areas of greater CPE prevalence (Seesahai et al., 2021). A report from Italy described a case of CPE transmission from mother-to-child at birth. The affected neonate subsequently developed a bloodstream infection and two other patients on the NU were colonised (Bonfanti et al., 2017). Although screening of mothers of babies admitted to NU for CPE is not currently recommended in the UK, this may need to be considered if cases of CPE rise.

Conclusion

Rates of CPE are increasing, especially within the neonatal population (Ballot et al., 2019). This report describes the response to a CPE outbreak in a UK NU. The combination of enhanced IPC measures, increased environmental cleaning, increased CPE screening, regular audits of practice, and improved staff/maternal education were critical in limiting the outbreak. However, there remains a need for further research into outbreak-limiting interventions given the rise in cases of this multi-drug-resistant bacteria.

ORCID iD

Megha Anil https://orcid.org/0000-0002-7972-476X