Article Text

Managing and preventing outbreaks of Gram-negative infections in UK neonatal units
  1. Mark Anthony1,
  2. Alison Bedford-Russell2,
  3. Tracey Cooper3,
  4. Carole Fry4,
  5. Paul T Heath5,
  6. Nigel Kennea6,
  7. Maureen McCartney7,
  8. Bharat Patel8,
  9. Tina Pollard9,
  10. Mike Sharland5,
  11. Peter Wilson10
  1. 1Oxford University Hospitals NHS Trust, Headington, Oxford, UK
  2. 2Neonatal Unit, Birmingham Women's NHS Foundation Trust, Birmingham, UK
  3. 3Department of Infection Prevention and Control, Queen Marys Hospital, Sidcup, Kent, UK
  4. 4Department of Health, Wellington House, London, UK
  5. 5Department of Child Health, St Georges Hospital Medical School, University of London, London, UK
  6. 6Neonatal Unit, St. George's Hospital, London, UK
  7. 7Department of Public Health, Public Health Agency, Belfast, UK
  8. 8Microbiology Services Division, Public Health England, Barts Health NHS Trust, London, UK
  9. 9Neonatal Unit, Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK
  10. 10Department of Clinical Microbiology and Virology, UCLH NHS Foundation Trust, London, UK
  1. Correspondence to Dr Mark Anthony, Department of Neonatology, John Radcliffe Hospital, Oxford OX3 9DU, UK; mark.anthony{at}


De novo guidance on the management of Gram-negative bacteria outbreaks in UK neonatal units was developed in 2012 by a Department of Health, England Antimicrobial Resistance and Healthcare Associated Infection working group. The recommendations included activation of an organisational response and establishing a control team when an outbreak is suspected; screening for the specific organism only during an outbreak; undertaking multidisciplinary reviews of cleaning routines, hand hygiene and Gram-negative bacteria transmission risks; considering deep-cleaning; cohorting colonised and infected babies preferably but not necessarily in isolation cubicles; and considering reducing beds or closing a unit to new admissions as a way of improving spacing and staff:patient ratios until the outbreak is under control. The group advised establishing mechanisms to communicate effectively across the network; informing parents of the outbreak as early as possible, and providing prewritten ‘infection outbreak’ information sheets. For prevention of outbreaks, the group advised meeting national staffing and cot-spacing requirements; following a Water Action Plan; using infection reduction care bundles and benchmarking; and introducing breast milk early and limiting antibiotic use.

  • Infectious Diseases
  • Neonatology

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


This guidance has been produced through the Department of Health, England, by the Neonatal Gram-Negative Infection Sub-group of the Antimicrobial Resistance and Healthcare Associated Infection committee, in response to the general experience of UK neonatal units (NNUs) of Gram-negative blood stream infections and antimicrobial resistance. The review, aimed at neonatologists and infection prevention and control teams interfacing with NNUs, represents an amalgamation of joint expert opinion and consultation with UK neonatal and infection prevention societies, and where available the relevant neonatal literature. Although produced through the Department of Health in England, the recommendations should be applicable to all of the UK. It is divided into three sections: ‘defining’, ‘preventing’ and ‘managing’ outbreaks; and is supported by web-linksw1 w2 etc to relevant Department of Health (DH) and outbreak reports.

Late-onset neonatal sepsis (infection occurring >48 h after birth) in the UK affects 2–3/1000 babies.1 ,2 Gram-negative bacteria (GNB) cause approximately 20–40% of all late-onset sepsis, with Klebsiella responsible for most neonatal outbreaks, followed by Serratia, Enterobacter, Pseudomonas, Escherichia coli and Acinetobacter. There is an associated mortality, and survivors have longer intensive care stays and more neurodevelopmental sequelae. GNB outbreaks also impede the effective functioning of neonatal clinical networks. Approximately 15% of UK NNUs have been investigated for a ‘recent’ infection prevention and control issue, and 12% per year temporarily close for this reason.3w1

Poor hand hygiene, overcrowding, inadequate spacing between cots, low nurse:patient ratios, environmental colonisation, inadequate cleaning of common-use equipment, injudicious use of antibiotics, particularly broad spectrum and prolonged courses, and delaying the introduction of maternal breast milk, all contribute to GNB outbreaks.w2 w3 4–6

Defining an outbreak and triggers for activating a response

Immediate activation of an organisational response is required when an outbreak is suspected. An outbreak can be defined as two or more sterile site isolates of the same species, with the same antibiogram, from different babies (not twins) within the space of 2 weeks. Triggers for activating a response might also be three or more babies colonised with the same GNB (in those NNUs that routinely screen for GNB colonisation), a single case of a rare or never seen GNB, or a single systemic infection with an ESBL-producing or carbapenem-resistant GNB (because these multidrug-resistant organisms confer a higher risk of treatment failure) or Pseudomonas aeruginosa (which is more likely than other GNB to indicate an environmental reservoir). Action-thresholds will be determined by an understanding of the level of activity of each NNU and its background GNB infection rate.

Preventing an outbreak

Outbreak prevention requires an organisation-wide approach with implementation, and assurance of implementation, of best practice across all elements of Health Care Associated infection (HCAI) improvement and infection prevention and control, including exemplary hand hygiene7 with use of alcohol-based gels.

Spread of GNB is exacerbated by inadequate spacing between cots,w2 4 ,5 and high cot-occupancy rates and low nurse:baby ratios promote errors and reduce the time for good infection prevention practices.8 The Neonatal Toolkit for High Quality Neonatal Servicesw6 w7 recommends 1:1 nursing by nurses qualified for intensive care patients, 2:1 for high dependency and 4:1 for low dependency. All NNUs should meet the recommendation for adequate cot spacingw7–9 and optimal staffing ratios for neonatal nurses. Cot spacing needs to be sufficient to provide for two parents, monitors, ventilator and other equipment, as well as space for staff to undertake sterile procedures. There should be no clutter of equipment in the intensive care unit (ITU) and high dependency unit (HDU) areas, facilitated by adequate accommodation for storage and cleaning of equipment (especially ventilators and incubators). There should be sufficient facilities for segregation of infected babies.w2 w7

Environmental contamination, for instance by P aeruginosa in water systems, is an important reservoir of antibiotic-selected organisms,w2 w3 and the hands of staff and multiuse equipment (such as echocardiography and ultrasound machines, ‘cold’ lights, laryngoscopes, breast pumps and stethoscopes) are potential modes of transmission. Every NNU should have a Water Action Planw2 w5 to reduce infections caused by waterborne pathogens, including Pseudomonas, Chrysebacterium and Stenotrophomonas; equipment should be single use (ie, laryngoscope blades) or dedicated to a baby (ie, stethoscopes); and there should be robust cleaning routines for equipment on the NNUw10 with a named responsible person for the cleaning, and evidence that best practice and policies are followed.

NNUs should adopt an infection reduction care bundle, such as the Matching Michigan programme,9 and benchmarking, for instance through the Neonatal Data Analysis Unit,w11 Neonatal Infection Surveillance Network (neonIN)w12 or Vermont-Oxford Network (VON).w13 These should be linked to learning and quality improvement programmes. Cot-side information systems that record in real-time physiological parameters, ventilator changes, drugs administered and laboratory parameters, and which permit easy access to data relevant to the patient's course, including timing of antibiotics, culture results, central line days and trends in inflammatory markers, will also help reduce infections.

Prolonged initial antibiotic use in premature babies is associated with adverse outcomes, including higher nosocomial infection rates and necrotising enterocolitis.10–14 Aim to reduce antibiotic use, give narrow spectrum antibiotics providing they are appropriate for local isolates and resistance patterns, and follow the National Institute for Health and Clinical Excellence CG149 guidance ‘Antibiotics for the prevention and treatment of early-onset neonatal infection’.w14 Establish regular microbiology or paediatric infectious diseases rounds and undertake spot audits of antibiotic use, facilitated by electronic prescribing tools.

Every day of delayed enteral feeding, especially with maternal breast milk, increases the risk of nosocomial infection and necrotising enterocolitis.15 ,16 Where possible introduce maternal breast milk on the first day of life in all preterm babies, facilitated by maternal counselling in the peripartum period. Do not defrost frozen breast milk by placing the container in warm tap water.w2

Colonisation of the intestine or endotracheal secretions (ETS) with GNB usually precede infection.17–19 Currently 21% of UK NNUs undertake routine stool/rectal swabbing for specific and/or gentamicin resistant/ESBL-producing Enterobacteriaceae.3 The rationale is that screening allows for the earliest possible detection of an outbreak, and the earliest instigation of control measures and prescribing of the most appropriate antibiotics if colonised babies become unwell. However, there is currently insufficient evidence on clinical effectiveness to recommend weekly rectal, skin or ETS screening for GNB colonisation in NNUs. Interim guidance in this report is to not routinely undertake screening for GNB colonisation until further evidence emerges. However, depending on microbial epidemiology, some NNUs may decide it is appropriate to continue local screening.

Management of a GNB infection outbreak

In reported outbreaks, control is eventually achieved by improving hand hygiene, reducing the number of cots available (effectively improving cot spacing and nurse:patient ratios), cohorting infected and colonised patients, screening for the specific organism, source identification by environmental screening, deep-cleaning the environment, and equipment cleaning for instance with chlorine-based agents.w1 w2 20 Establishing an infection control team, investigation and instituting precautionary control measures should be considered on suspicion of an outbreak and not await confirmation of the outbreak.

Action when an outbreak occurs

A unit-based reaction to a suspected outbreak should be embedded in an overall organisational response, with robust communication, through the Infection Prevention and Control Team. The team structure will vary depending on local hospital outbreak policy, but may include a lead neonatologist, preferably one with an interest in neonatal infection, paediatric infectious disease/microbiology clinician, the Health Protection Unit, the Director of Infection Prevention and Control and the Chief Executive.w1 w2 The Health Protection Scotland ‘Hospital Outbreak Checklist’w15 is a useful practical guide to early management of a suspected outbreak. Set up an outbreak control team meeting as soon as an outbreak is suspected,w1 w2 and meet regularly to ensure control measures are effective.

After a careful risk assessment, especially of the needs of the Neonatal Network consider closure of the NNU to admissions and avoid transfers to other NNUs. Communicate across the Neonatal Network, so that when one hospital has an outbreak, all hospitals within the network are informed expediently. Inform parents of the outbreak as early as possible and reiterate infection prevention educationw1 w2; provide parents with prewritten ‘infection outbreak’ information sheets; and during an outbreak pay attention to good communication with parents, particularly about their baby being colonised or infected, and inform parents early on when their baby is seriously unwell.w2 Work with the Trust communications department to reach the local community, through a press release, to inform relatives and pregnant women of the outbreak situation.

Reinvigorate attention to hand hygiene during an outbreak partly through re-education, feedback about the outbreak and through frequent hand hygiene audits (aiming for >95% compliance and with immediate feedback on poor performance). As well as hand hygiene, plastic apron and gloves (personal protective equipment) should be worn for all staff contacts with infected and colonised babies. Educate parents again about meticulous hand hygiene.

During an outbreak cohort or isolate infected and colonised babies and, after a risk assessment particularly of the needs of the Neonatal Network, consider reducing bed availability so that British Association of Perinatal Medicine (BAPM) staffing (including taking account of skill mix) and cot spacing standards are met, or closure of the NNU to new admissions. Try to deploy staff so that a separate group of nurses and doctors look after only the colonised or infected patients. Infection prevention and control measures need to be maintained in cohorted babies, so that those colonised/infected with similar, but non-identical strains of the same species, are not cross-transferred between babies.

Ideally, infected and colonised babies should not be moved between NNUs during an outbreak, but decisions should be made on a case-by-case basis taking into account the needs of the Neonatal Network.

Environmental sampling and decontamination

Environmental screening is recommended when a single Pseudomonas infection occurs,w2 but the clinical effectiveness of screening for other GNB is unproven.21

As part of outbreak management, consider screening the environment, including inside incubators, for P aeruginosa, and screen on a case-by-case basis for other GNB. Follow DH advice on water source and sink testing following a single Pseudomonas infection on a NNU—for instance, screen taps and sinks for colonisation in any room that the infected baby has inhabited since birthw2 w3; test any environmental pooled water (underneath sinks, air conditioning units, roof leaks, etc); and screen multiuse equipment and blood gas analysers.

Undertake a thorough multidisciplinary review of the ward (cleaning routines, hand hygiene and GNB transmission risks). Use this review to inform cleaning schedules and frequencies.w10 w16 Consider deep-cleaning the environment, including all equipment.w17

Patient screening

There is some evidence relating to the clinical effectiveness of screening for GNB colonisation as part of outbreak management. Undertake screening during outbreaks for the specific strain responsible for invasive disease, with at least weekly rectal swabs, and ETS in ventilated babies. Screening should be for a defined period that is, 1–2 months or until the outbreak has resolved. A limited period of screening may also be undertaken when there is a trigger for activating a response. Set up a method to feedback screening results regularly to NNU staff.


GNB NNU outbreaks indicate that something in the system (environment, equipment, people or methods) is enabling the transmission of GNB. GNB infections in babies have a high morbidity and mortality and outbreaks can be devastating, within the affected NNU and to the functioning of the Neonatal Network. Avoiding outbreaks requires extremely high baseline standards of infection prevention and control practice. See table 1 for a summary tick list of recommendations.

Table 1

Good practice actions to reduce Gram-negative bacteria (GNB) infections and manage outbreaks

When a GNB outbreak occurs consider reducing cots or closing to new admissions, to improve spacing and nursing ratios, cohort infected and colonised babies preferably in isolation cubicles, but not at the expense of the effectiveness of the Neonatal Network; use personal protective equipment for contact with all infected or colonised patients; and consider rectal swab (and ETS) screening.

Urgent research is required in the following areas:

  • Determine the feasibility and clinical effectiveness of weekly screening of babies to identify colonisation by P aeruginosa and/or antibiotic resistant GNB.

  • Determine the mode of spread of GNB around NNUs using molecular typing, including bacterial whole genome sequencing.

  • Improve neonatal infection diagnosis, that is, molecular diagnostics and the use of combination biomarkers, to more accurately target antibiotics at those babies that are truly infected, and thereby help reduce antibiotic overuse.

  • Explore surveillance systems that include an outbreak ‘early warning’ tool, raising an alert when a pre-set trigger indicates an outbreak has occurred.

  • Explore the clinical effectiveness of environmental measures, such as filters on taps and deep-cleansing agents, on GNB infection reduction.

  • Determine the impact of oral lactoferrin and probiotics on GNB blood stream infections in preterm babies.


The following societies received this guidance for comments:

  • Antimicrobial Resistance and Healthcare Associated Infection Committee


  • British Association of Paediatric Surgeons

  • British Association of Perinatal Medicine

  • British Infection Association

  • British Paediatric Allergy, Immunity & Infection Group

  • Devolved Administration Government Department Health Protection Agencies, UK

  • Healthcare Infection Society

  • Infection Prevention Society

  • MCRN Neonatal Infection Group

  • Royal College of Paediatrics and Child Health


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • MA: lead author

  • Contributors All authors, as members of the Neonatal Infection Sub-Group of the ARHAI (Antimicrobial Resistance and Healthcare Associated Infection) Committee of the DH England, have contributed to the evolution of this review, through an initial meeting in London, and then involvement in two teleconferences and/or commenting on successive drafts.

  • Competing interests All authors are members of ARHAI. No author has declared any competing financial/industry interests.

  • Provenance and peer review Not commissioned; externally peer reviewed.