Cases in the UK and Ireland

There was a total of 35 instances of possible neonatal cross-infection which occurred between 1971 and 2012: ten episodes have been described previously as case reports [Reference Storrs and Partridge12–Reference Fullerton21].

Three of these instances occurred in Scotland, two in Wales two in Ireland and the remaining 28 occurred in England. Twenty-three of the instances occurred between 1971 to 1989, six between 1990 and 1998 and the remaining six between 2006–2012. There were no records of neonatal cross-infection identified before 1971 or after 2012.

All 35 instances showed a common pattern (Table 1) of a congenitally infected infant who was recognised as ill at birth or within one day of delivery. Within each episode, there was direct or indirect contact in the same hospital with an apparently healthy second infant (33 episodes), and in two instances two infants (one of which were twins where both became infected [Reference Roberts, Quoraishi and Evans20]). In all instances, the late-onset cases developed meningitis and were diagnosed as ill between 2 and 18 days after contact with the early onset case (50% between 6 and 12 days, mean 7.7 days). In 20 of the 35 instances, the infants were nursed on the same day in the obstetric unit, recovery room or new-born nursery, consequently staff or equipment were common. The most common equipment noted was the use of the same resuscitaire in five instances (Table 1).

Table 1.

Features of 35 episodes of neonatal cross-infection occurring in the UK and Ireland 1971–2015

CSF, cerebrospinal fluid; GA, gastric aspirate; HVS, higher vaginal swab; NR, not recorded; PM, post-mortem; SCBU, special care baby unit.

Seven of the episodes occurred where common contacts extended over time periods longer than one day, the longest being where the same delivery room which was used by what became the late-onset case 10 days after the birth of the early-onset case.

Two of the episodes differed in the likely route of transmission. In the first, the late-onset case was born 9 h before the index case but did not share the same cot or geographical location within the hospital at any time. Both infants were attended by the same medical practitioner who attended the early-onset case who was born by the emergency caesarian section, and the practitioner only examined the late-onset infant prior to initial discharge on the same day as their delivery. The only common factor was the medical practitioner: it was suggested that the most likely mode of transmission was the use of the same stethoscope, although the bacterium was not recovered from this instrument. The late-onset case was readmitted 10 days later when the bacterium was recovered from a CSF sample [Reference Fullerton21]. L. monocytogenes was recovered from the placenta and maternal HVS, but not from the early onset case as specimens were collected after the commencement of antimicrobial treatment [Reference Fullerton21]. In the second instance, the congenitally infected infant died at birth and 5 days later the mother of this infant was nursed in an open postnatal ward and allowed to cuddle the baby in the adjacent bed who developed meningitis at 12 days later [Reference Isaacs and Liberman13].

In all 35 episodes, microbiological confirmation of listeriosis was by the isolation of L. monocytogenes. The bacterium was isolated from 34 of the early onset cases (22 from blood) and all of the resulting late-onset neonatal cases (22 from CSF and 10 from blood). L. monocytogenes was isolated from 11 of the mothers of the congenitally infected infants (HVS, breast fluid and/or blood) but never from the mothers of the late-onset cases. L. monocytogenes was not isolated from any environmental sites associated with any of the instances.

Of the 35 incidents, five were due to L. monocytogenes serogroup 1/2, 29 were due to serogroup 4 and the serogroup was not available in the final incident. In addition to serogrouping, phage typing was used between 1971 and 2001 and amplified fragment length polymorphism analysis between 2002 and 2015: within each individual instance, L. monocytogenes isolates from the early and late-onset cases, and, where available, the mother of the early onset case were of the same type.

WGS was applied to isolates from cases of episodes occurring between 1989 and 2012: one or more isolate was available from eight of the episodes, three were due to L. monocytogenes CC1, three to CC2, one to CC6 and one to CC18. Isolates from the early and late-onset neonatal cases were available from four of the episodes (including that described by Fullerton et al. [Reference Fullerton21]), two were CC1, one was CC2 and one CC6. Analysis of sequence data showed that, within each incident, all were indistinguishable (≤5 SNPs) and therefore indicative of a common source. Sequence accession numbers from the four episodes are SRR16976073 and SRR16976080; SRR16976074 and SRR16976072; SRR17120522 and SRR16941085; SRR16286915 and SRR16286914.

Neonatal listeriosis reported in peer-reviewed literature

Case reports from a further 29 instances were identified from outside of the UK and Ireland occurring between 1936 and 2013. The instances occurred in 11 different countries: ten in France [Reference Dubois and Lefebvre22–Reference Lazarus28], three in Germany [Reference Joppich, Schultze, Joppich and Schultze29–Reference Kachel and Lenard31], three in the USA [Reference Burn9, Reference Florman and Sundararajan32, Reference Nelson33], two in Canada [Reference King, Duncan and Tracey34, Reference Farber35], two in Israel [Reference Levy and Nassau36, Reference Colodner37], four in Sweden [Reference Stormby, Wallengren and Weibera38–Reference Higgins, Kahlmeter and Larsson40], and one each in Costa Rica [Reference Schuchat41], Chile [Reference Garcia42], Italy [Reference Facinelli43], Kuwait [Reference Sethi, Ghafoor and Vandepitte44] and Spain [Reference Tortajada45]. A summary of these 29 episodes is shown in Table 2.

Table 2.

Features of 28 neonatal cross-infection episodes occurring outside the UK and Ireland and reported in the peer-reviewed literature 1936–2013

CSF, cerebrospinal fluid; GA, gastric aspirate; HVS, higher vaginal swab; NR, not recorded; PM, post-mortem.

The 29 episodes showed many similarities to those already described for episodes in the UK and Ireland. For each instance, infants within each episode were nursed within the same hospital. In this series, the late-onset cases were diagnosed as ill between 2 and 48 days after contact with the early onset case (48% between 6 and 12 days, mean 8 days). In 28 of the instances, all infants were born in the same hospital, in the final instance described by Dubois and Lefebvre [Reference Dubois and Lefebvre22], the primary and one of the secondary cases were born at home and subsequent contact occurred in a neonatal hospital ward. In 28 of the instances there was a single early-onset case detected, and in the one instance, two early-onset cases born on the same day in the same delivery unit [Reference Garcia42]. Fourteen of the 28 incidents showed the most common pattern observed in the UK and Ireland of a congenitally infected index case with sepsis which resulted in a further single late-onset neonatal case. In a further 12 episodes two, three or four late-onset cases occurred (Table 2). In the final three episodes, a single early-onset case resulted in five [Reference Sethi, Ghafoor and Vandepitte44], six [Reference Joppich, Schultze, Joppich and Schultze29] or nine [Reference Schuchat41] late-onset cases. Where information was available, all of the early onset cases were ill at birth and 67% of the late-onset cases were diagnosed between 9 and 12 days after contact with the early-onset case. Two of the late-onset cases occurred outside the neonatal period: one at 30 days in a cluster of three [Reference Velin, Dupont and Landragin-Roussel26] and one diagnosed 48 days after delivery [Reference Tortajada45]. Velin et al. [Reference Velin, Dupont and Landragin-Roussel26] reported a cluster of three late-onset cases diagnosed as ill at 7, 14 and 30 days later and who were nursed in the same hospital but in different rooms over a 5-day period with the early-onset case. In the report of Tortajada et al. [Reference Tortajada45], an early-onset case was nursed from birth in the same neonatal ward over a 10-day period as two newborn infants who were diagnosed with meningitis 9 and 48 days later. Staff and a nappy changing surface were common to all these infants.

Microbiological confirmation was obtained by the isolation of L. monocytogenes from all of the early-onset cases (13 from blood and 5 from CSF) except in one instance where a diagnosis of listeriosis was strongly supported by histological examination of the tissue taken at necropsy [Reference Levy and Nassau36]. L. monocytogenes was isolated from maternal sites in six instances [Reference Fessard24, Reference Dickgiesser30, Reference Nelson33, Reference Colodner37, Reference Garcia42, Reference Facinelli43]. From all of the late-onset cases, of the 56 infants where information was available, the bacterium was recovered from CSF in 44, blood in 17 and solely from other sites in 5 cases (Table 2). L. monocytogenes was isolated from non-clinical samples in two instances: a tape measure used in the delivery Unit in one instance [Reference Higgins, Kahlmeter and Larsson40] and from mineral oil used for bathing the infants after delivery in the largest series, see below [Reference Schuchat41].

Characterisation of the L. monocytogenes isolates from the clinical cases was not reported in 16 of the incidents. In the remaining 13, serogrouping was available for ten of the incidents and two were due to serogroup 1/2, the remaining eight were due to serogroup 4. Evidence of the same L. monocytogenes type being recovered from the mothers of the early-onset cases and environmental sites (where available) and both early and late-onset infants within each episode was obtained by: phage-typing [Reference Velin, Dupont and Landragin-Roussel26, Reference Jean27, Reference Nelson33, Reference King, Duncan and Tracey34, Reference Higgins, Kahlmeter and Larsson40, Reference Facinelli43, Reference Sethi, Ghafoor and Vandepitte44]; multilocus enzyme electrophoresis [Reference Farber35, Reference Schuchat41]; random amplified polymorphic DNA [Reference Hof and Lampidis46]; restriction fragment analysis [Reference Farber35, Reference Facinelli43], plasmid analysis [Reference Facinelli43] and pulsed-field gel electrophoresis [Reference Lazarus28, Reference Colodner37, Reference Tortajada45].

Possible routes of transmission were identified with common staff and equipment, particularly the use of the same resuscitaire in four instances (Table 2). In one instance, a common rectal thermometer was used which was not disinfected between infants [Reference Larsson39]: the early-onset case was born shortly before two further infants who were diagnosed as ill with blood in their stool 3 days later. L. monocytogenes was isolated from the blood and CSF of the early-onset case, the faeces of the two late-onset cases, and the faeces of a further two asymptomatic infants who were born on the same and four days later [Reference Nelson33]. The largest series occurred in Costa Rica in 1989 where a single early-onset case resulted in a further nine cases diagnosed as ill three to eight days after delivery [Reference Schuchat41]. All cases were born over a 13-day period in the same delivery room where newborn infants were bathed in mineral oil shortly after birth. There was evidence for contact between the oil and the babies' nose and mouth. The mineral oil was stored in an open container in the delivery room with no additional disinfection agents and was neither cleaned or completely emptied between refilling: the same strain of L. monocytogenes was detected in the index patient's clinical specimens as well as the open oil container in the delivery room.