The first occurrence of a CTX-M ESBL-producing Escherichia coli outbreak mediated by mother to neonate transmission in an Irish neonatal intensive care unit

The University Maternity Hospital Limerick in Ireland (UMHL) is a tertiary referral centre and includes a NICU with a total of 19 cots. The catchment population of UMHL is approximately 400,000. In the twelve months prior to the outbreak, there were 4905 live births and 909 NICU admissions (of all gestational ages). At the time of this outbreak, the NICU had one intensive care ward consisting of four neonatal intensive care cots, five high dependency cots and two isolation rooms. Two intermediate care rooms, separated from the intensive care ward, contained a further 10 cots. With respect to prevention of nosocomial infection, the NICU intensive care ward provided four washing stations, alcohol hand gels at each bedspace, and a nurse to patient ratio of 1:1 for ICU category cots and 1:2 or 1:3 (depending on staffing levels) for the remainder. A weekly multi-disciplinary NICU ward-round was performed.

Contact tracing of all inpatients who may have been in contact with the index case while in NICU was conducted using urine samples and/or rectal swabs, one sample for each neonate or mother was analysed depending on what was most easily obtained. With consent, the mother of the index case and mothers of subsequently positive neonates were screened for the presence of ESBLs using rectal swabs, high vaginal swabs and mid-stream urine samples. Following a review of related literature, and to be prudent, a decision was made to close the NICU to new admissions from week 12 2013, with exception of emergencies, and visiting was restricted to parents of inpatients only awaiting the results of contact tracing.

Following confirmatory cultures results, all infected or colonised neonates were barrier nursed by personnel wearing disposable gowns and gloves. In addition, a restriction was placed on the prescription of third generation cephalosporins. Due to the availability of only two isolation rooms in the NICU, the index case and one other neonate infected by ESBL-producers were isolated. The remaining colonised neonates were cohorted in incubators in the main NICU ward, with dedicated single-patient equipment. Given physical environmental constraints, it was not possible to increase the space between cots. Neonates who were fit for discharge were cohorted to a single post-natal maternity ward to minimise cross-transmission.

With respect to hygiene, enhanced cleaning of the NICU was instigated in parallel with increased auditing. This involved twice-daily cleaning of affected areas and incubators with detergent. Air sampling and environmental sampling were not performed. An intensive targeted educational programme focussed on standard precautions, particularly hand hygiene compliance and on modes of transmission of ESBL-producing E. coli transmission was provided to all clinical and administrative staff. Screening of staff for carriage of ESBL-producing E. coli was not conducted but instead it was considered of important to put emphasis on zero tolerance to poor compliance with the World Health Organisation’s “5 moments for hand hygiene”. Hand hygiene audits were performed with greater frequency in affected areas, which involved twice weekly observational audits at ward level. The last positive isolate was identified in week 13 2013. The NICU re-opened later in week 13 2013 and weekly multi-disciplinary meetings were held until week 18 2013 to discuss and implement hygiene recommendations, at which point the outbreak was declared over. The last known neonate involved in the outbreak was discharged in week 16 2013.

At the time of the outbreak, the routine ESBL screening policy targeted weekly screening of high-risk neonates identified as such by their managing Consultant Neonatologists (but broadly categorised as premature or in NICU for other than short term stay) and their mothers.

Screening specimens were cultured using ChromID™ ESBL agar (bioMérieux, Marcy l’Ȇtoile, France) and incubated at 37 °C aerobically for 18–24 h. The colour code guide provided by the manufacturer was followed for review of any colonies identified; pink/brown = presumptive E. coli, green/blue = presumptive Klebsiella species, white = other Enterobacteriaceae. All colonies were identified using MALDI-ToF MS (Bruker Daltonics, Bremen, Germany) as described previously [20]. Confirmatory testing was performed by disk diffusion on all organisms that warranted further investigation (Thermo Scientific™ Oxoid™ Disks) using a disc dispenser; cefoxitin 30 μg (FOX), cefapime 30 μg (FED), ceftazidime 30 μg (CAZ) & ceftazidime-clavulanic acid 30/10 μg (CAZCV), cefotaxime 30 μg (CTX) & cefotaxime-clavulanic acid 30/10 μg (CTXCV), Muller Hinton agar, 0.5 McFarland inoculum; 35 +/−2 °C, ambient air, 16–28 h. Following incubation, the zone sizes of the cephalosporin disc to that of a cephalosporin plus clavulanic acid combination disc were compared to determine ESBL status. Criteria for positive ESBL disc confirmatory testing on Enterobacteriaceae: a ≥5 mm increase in zone diameter for either antimicrobial agent tested in combination with clavulanic acid versus its zone when tested alone. Criteria for negative ESBL disc confirmatory testing on Enterobacteriaceae: zone sizes for both cephalosporin and cephalosporin in combination with clavulanic acid that is equal or show no greater difference in diameter than +/− 2 mm. Criterion for inconclusive ESBL disc confirmatory testing: difference between matched discs was >2 but <5 mm. In 2013, our protocols dictated that any inconclusive disk diffusion result, would warrant further confirmatory testing via Etest (bioMérieux, Marcy l’Ȇtoile, France). No isolate from this outbreak required further testing via Etest.

The genetic relationships between the ESBL-producing E. coli isolates were determined by pulsed-field gel electrophoresis (PFGE) of XbaI-digested genomic DNA at the Antimicrobial Resistance and Healthcare Associated Infections (AMRHAI) Reference Unit, Public Health England, London, UK. Electrophoresis was performed on a Bio-Rad CHEF DRII apparatus at 6 V cm⁻¹ for 30 h at 12 °C with ramping times of 5 s to 35 s. CTX-M subtyping was possible on all but one of the isolates subsequent to the outbreak.

Routine breast milk testing involved incubating undiluted samples spread on blood agar (5-10% CO₂) for up to 48 h, blood agar with metronidazole for up to 48 h anaerobically, and MacConkey Agar for up to 48 h aerobically.

Following detection of the index case (neonate two, Table 1), between weeks 12 and 13 2013, following a comprehensive screening seven additional screens of two NICU mothers and five neonates proved positive (Table 1). During this outbreak, infection was defined by clinical and laboratory criteria and requirement for antimicrobial therapy, while colonisation was defined by the absence of relevant symptoms. In total, during the outbreak, 86 ESBL screens from 42 individuals were performed. Of these, specimens from six neonates were positive for ESBL-producing E. coli: two represented infection and four represented colonisation. The mean gestational age was 33 weeks (range 28 to 36 weeks). There were no bacteraemias due to ESBL-producers.

The index case neonate’s mother was informed and agreed to participate in screening. A rectal swab, a mid-stream urine and a sample of expressed breast milk all tested positive for ESBL-producing E. coli, as shown in Table 2, with the same antibiogram as the isolate from her neonate (detailed earlier). A high vaginal swab was negative for the bacterium. This isolation of an ESBL-producing E. coli was the first such result for the patient who had never before had a culture-positive urine test.

One further neonate (neonate six, Table 1) who was in the NICU at the time of the outbreak (Fig. 1) developed urosepsis, eight days after the first detection of an ESBL-producer in the NICU. While two blood specimens proved negative, ESBL-producing E. coli were confirmed from a urine sample and from a rectal swab. An identical antibiogram to that of the index case neonate was noted. The baby was treated with IV meropenem (dosed as per weight). This neonate’s mother was screened whereupon a high vaginal swab and a sample of expressed breast milk cultured negative, but a rectal swab and a mid-stream urine sample were positive for ESBL-producing E. coli demonstrating the same antibiogram as previously found. Again, this second mother had no previous documented urinary tract infections and had never before had a culture-positive urine test.

Of the remaining four neonates, all were clinically stable, underwent weekly surveillance rectal cultures until discharged from the NICU, and did not require treatment with antimicrobials. All mothers of colonised infants underwent screening to determine colonisation or infection and were found to be negative for ESBL-producers.

All ESBL isolates identified during this outbreak had an identical antibiogram regardless of culture source: resistant to ampicillin/amoxicillin, co-amoxiclav, ceftriazone, aztreonam, ciprofloxacin, gentamicin, cefuroxime, piperacillin/tazobactam, tri/sulfamethoxazole, while susceptible to amikacin, chloramphenicol, ertapenem, and meropenem. Following sequencing, all were found to bear CTX-M-15 and OXA-1.

Following PFGE analysis (see Fig. 2), strain designated LIME04ES-3 was found to be shared between the index case, their mother and four other neonates, indicating cross-transmission and was determined to be ST131. A second strain (LIME04ES-4), determined to be ST1284, was found in both the second infected neonate and their mother.