Article Text

Download PDFPDF

Enhanced surveillance of methicillin-resistant Staphylococcus aureus (MRSA) bacteraemia in children in the UK and Ireland
  1. A P Johnson1,
  2. M Sharland2,
  3. C M Goodall1,
  4. R Blackburn1,
  5. A M Kearns1,
  6. R Gilbert3,
  7. T L Lamagni1,
  8. A Charlett1,
  9. M Ganner1,
  10. R Hill1,
  11. B Cookson1,
  12. D Livermore1,
  13. J Wilson4,
  14. R Cunney5,
  15. A Rossney6,
  16. G Duckworth1
  1. 1Centre for Infections, Health Protection Agency, London, UK
  2. 2Paediatric Infectious Diseases Unit, St George's Hospital, London, UK
  3. 3MRC Centre of Epidemiology for Child Health, Institute of Child Health, London, UK
  4. 4Health Protection Scotland, Glasgow, UK
  5. 5National Disease Surveillance Centre, Dublin, Ireland
  6. 6National MRSA Reference Laboratory, Dublin, Ireland
  1. Correspondence to Dr Alan Johnson, Department of Healthcare-Associated Infections & Antimicrobial Resistance, HPA Centre for Infections, London NW9 5EQ, UK; alan.johnson{at}


Objective To determine the incidence and demographic features of methicillin-resistant Staphylococcus aureus (MRSA) bacteraemia in children in the UK and Ireland and to characterise MRSA isolated from cases.

Design Prospective surveillance study.

Setting Children aged <16 years hospitalised with bacteraemia due to MRSA.

Methods Cases were ascertained by active surveillance involving paediatricians reporting to the British Paediatric Surveillance Unit and by routine laboratory surveillance. Patient characteristics were obtained using questionnaires sent to reporting paediatricians. MRSA isolates were characterised using molecular and phenotypic techniques including antimicrobial susceptibility testing.

Results 265 episodes of MRSA bacteraemia were ascertained, involving 252 children. The overall incidence rate was 1.1 per 100 000 child population per year (95% CI 0.9 to 1.2): 61% of the children were aged <1 year (a rate of 9.7 cases per 100 000 population per year (95% CI 8.2 to 11.4)) and 35% were <1 month. Clinical data were obtained from 115 cases. The clinical presentation varied, with fever present in only 16% of neonates compared with 72% of older children. A history of invasive procedure was common, with 32% having had intravascular lines and 13% having undergone surgery. 62% of patients for whom data were available were receiving high-dependency care (46% in SCBU/NICU and 16% in PICU). Of 93 MRSA isolates studied, 73% belonged to epidemic strains widely associated with nosocomial infection in the UK and Ireland.

Conclusions MRSA bacteraemia in children was relatively uncommon and was predominantly seen in very young children, often those receiving neonatal or paediatric intensive care. Bacteraemia predominantly involved well-documented epidemic strains of MRSA associated with nosocomial infection.

View Full Text

Statistics from

Hospital microbiology laboratories in England, Wales and Northern Ireland voluntarily report cases of bacteraemia to the Health Protection Agency (HPA), often providing data on the anti-microbial susceptibility of the infecting pathogen. In the case of Staphylococcus aureus, there was a dramatic increase in the proportion of bacteraemias due to methicillin-resistant S. aureus (MRSA) during the 1990s,1 with a subsequent levelling and fall between 2000 and 2007. Although MRSA infections are predominantly seen in older patients, data obtained between 1990 and 2001 identified an increase in MRSA bacteraemia in children, with the numbers of cases increasing from <5 (<1% of S. aureus reports received) in 1990 to over 70 in 2000 and 2001 (>10% of S. aureus reports).2 The majority of cases were neonates, although substantial numbers in infants were also reported.

What is already known on this topic

  • ▶. Increasing rates of paediatric methicillin-resistant Staphylococcus aureus (MRSA) bacteraemia have been reported in the UK, but the clinical and microbiological characteristics and risk factors were unknown.

What this study adds

  • ▶. This study has determined that paediatric MRSA bacteraemia is uncommon and largely due to standard NHS hospital strains of MRSA. Children particularly at risk of MRSA bacteraemia are those that are very young and receiving intensive care. To reduce rates still further, there needs to be an enhanced focus on prevention of neonatal healthcare-associated infections.

These trends were viewed with concern by the Specialist Advisory Committee on Antimicrobial Resistance (in particular its paediatric subgroup) whose role was to advise the Chief Medical Officer and government ministers on problems relating to anti-microbial resistance.3 A particular concern was the paucity of data on MRSA infections in children. Although analysis of routine laboratory reporting can identify trends in MRSA bacteraemia, it does not provide sufficient epidemiological information to inform strategies for targeting the problem. In additional, the voluntary nature of the microbiological surveillance system means that there is under-ascertainment of cases. To address these issues, we conducted a 2-year program of enhanced surveillance in the UK and Ireland through the British Paediatric Surveillance Unit (BPSU) and routine laboratory surveillance. In addition, we determined the phenotypic and genotypic traits (including virulence gene profile) and antibiotic susceptibility of MRSA isolates causing bacteraemia in children.


Case definition

Cases were defined as patients aged <16 years of age, who had MRSA isolated from blood culture. Repeat cases occurring within 14 days of the earliest recorded specimen date were counted as a single episode of bacteraemia and merged into a single record.

Case ascertainment

Cases were ascertained over a 2-year period from 1 June 2005 to 31 May 2007 from the sources listed below. Data were collated and de-duplicated to identify cases ascertained from more than one source using a series of case identifiers including date of birth, sex, date of the first positive blood culture, hospital name and the patient's hospital number.

Reporting of cases by paediatricians

Data were collected prospectively by asking paediatricians in the UK and Ireland to report any children with MRSA bacteraemia to the BPSU via the monthly “orange card” reporting system.4 A questionnaire requesting information on the patient's demographics and, clinical presentation at the time of the MRSA-positive blood culture was sent to each reporting paediatrician.

Reporting of cases by hospital microbiologists

Cases of bacteraemia were reported on a voluntary basis by hospital microbiologists in England, Wales and Northern Ireland to the national database (LabBase2) maintained by the HPA.1 Microbiologists in Ireland were requested to report cases directly to the study investigators.

Cases were also ascertained by monitoring the submission of MRSA bacteraemia isolates from children to the national reference laboratories, either for clinical purposes or as part of their participation in the European Antimicrobial Resistance Surveillance System, a Europe-wide surveillance of antimicrobial resistance in a number of pathogens including MRSA.5 Cases were also ascertained from reports of MRSA bacteraemia submitted from hospitals in England following the introduction of the Mandatory MRSA Enhanced Surveillance System (which includes patient demographic data) in October 2005.1

Characterisation of MRSA isolates

Isolates were collected centrally at the HPA Staphylococcal Reference Section (HPA Centre for Infections) for standardised phenotypic and molecular characterisation by phage typing, pulsed-field gel electrophoresis (PFGE) and detection of genes encoding putative virulence factors including Panton–Valentine leucocidin (PVL) as described previously.6 The antibiotic resistance profile of referred isolates was determined by the agar dilution method and interpreted using the susceptibility breakpoints of the British Society for Antimicrobial Chemotherapy.7

Data analysis

Population denominators

Midyear population estimates for 2006 for England, Wales, Scotland and Northern Ireland were provided by the Office of National Statistics.8 Irish population data were taken from the 2006 Census.9

Statistical analyses

Data handling and statistical analyses were performed using STATA V.8.2 (StataCorp LP, College Station, Texas, USA). Using an exact Poisson method, 95% CIs around the estimated incidence rates were calculated.10

Ethical approval

Ethical approval was granted by the Eastern MREC.


Disease incidence

Over the 2-year period, 265 episodes of MRSA bacteraemia involving 252 children were ascertained (nine children had two episodes and two experienced three episodes). This equates to an overall annual rate of 1.1 cases per 100 000 children per year (95% CI 0.9 to 1.2), the rates in the two successive 12-month periods being 1.2 (95% CI 1.0 to 1.4) and 1.0 (95% CI 0.8 to 1.2), respectively. Cases were reported from 104 hospitals, with the majority (97.1%) reporting <10 cases and 53.8% reporting only one case within the 2-year reporting period. There were no significant differences in the incidence rates in the five countries (table 1).

Table 1

Annual rate of MRSA bacteraemia in children <16 years by country, June 2005–May 2007

Demographic and clinical characteristics

Age data were available for 254 (96%) of the 265 cases. As shown in fig 1, the majority were young children, with 154 (61%) aged <1 year (a rate of 9.7 cases per 100 000 population per year (95% CI 8.2 to 11.4)), 90 (35%) aged <1 month, 36 (14%) aged <1 week old and 20 (8%) aged ≤2 days. Fifty-eight per cent of the 245 children whose sex was recorded were male.

Figure 1

Age distribution of children with bacteraemia due to MRSA.

Clinical data derived from questionnaires were available for 115 of the 265 cases (43.4%). Only half of these children had fever, although this varied with age, from 16% (7/43) for children aged <1 month to 72% (47/65) for older children (table 2). Data on hypothermia were not collected. A range of infections in addition to bacteraemia were noted, the most frequent being skin or soft tissue (32.7%), lower respiratory tract (29%) and surgical site or wound infections (12.8%). Data on the aetiology of these infections were not, however, collected. A history of invasive procedure was common, with 32% having had intravenous lines and 13% having undergone surgery. Fifty-eight per cent of patients were reported as having had a sample other than blood positive for MRSA in the week before their first positive blood culture.

Table 2

Clinical features of children diagnosed with MRSA bacteraemia at date of first positive culture

The wards or units where patients were resident on the date of their first MRSA-positive blood culture are also shown in table 2. Sixty-two per cent of patients were receiving high-dependency or intensive care (SCBU/NICU, 47%; PICU, 17%) and among the remainder, more than four times as many patients were based in medical wards (28%) than in surgical wards (6%).

Characterisation of MRSA isolates

Strain profiling and antibiogram data were available for 93 isolates. On the basis of combined phage typing and PFGE analysis, 68 of these isolates (73%) were identified as belonging to one or other of two previously documented epidemic MRSA (EMRSA) strains, designated EMRSA-15 (63 isolates; 67.7%) and EMRSA-16 (five isolates; 5.4%).1 Multiple PFGE subtypes of both these clones were identified (>8 for EMRSA-15 and four for EMRSA-16). A high proportion of the EMRSA-15 and EMRSA-16 isolates were resistant to ciprofloxacin, with slightly fewer also resistant to erythromycin as is typical for these clones.1 All isolates, irrespective of strain type, were susceptible to vancomycin, rifampicin, teicoplanin and linezolid (table 3).

Table 3

Characteristics of MRSA strains causing bacteraemia in children

The other 25 isolates (26.9%) comprised one additional EMRSA strain designated Irish-1, 19 minor MRSA clones and five isolates with the gene encoding PVL (table 3). Of the five patients from whom these isolates were obtained, only one was admitted from home, and all were treated in different hospitals. On the basis of their PFGE profiles, toxin gene complement and susceptibility profiles, these PVL-positive isolates belonged to three different lineages of CA-MRSA, comprising clonal complexes 22 (two isolates), 30 (one isolate) and 80 (two isolates). The latter two lineages have disseminated internationally and are known as the Southwest Pacific and European clones, respectively.


This study found a relatively low overall incidence (1/100 000 children) of MRSA bacteraemia in children in the UK and Ireland. This finding is consistent with data from the enhanced mandatory surveillance of MRSA bacteraemia in England which has shown that MRSA bacteraemia is predominantly seen in adults, with only 2% of the total cases reported involving children.11

Clinical data were obtained for only 43% of cases. This may reflect the nature of the disease that was studied, which was, by definition, diagnosed on the basis of microbiology laboratory investigation. For many of the children, the isolation of MRSA from blood culture may not have been a primary diagnosis but rather just one component of a complex combination of clinical features, with the result that there may have been under reporting by paediatricians. Although attempts were made to elicit further clinical information about cases ascertained through routine microbiology reporting, these were largely unsuccessful.

Although this was not a case-control study, meaning that risk factors cannot be definitively identified, the descriptive epidemiology clearly indicates that the majority of cases comprised very young children and neonates, often those being managed in intensive care units. A high proportion of the patients yielded specimens positive for MRSA in the week preceding onset of MRSA bacteraemia, which is important, as it is well documented that previous colonisation with MRSA is a risk factor for subsequent infection.12 Furthermore, a high proportion of the children for whom data were available had undergone invasive procedures which may have provided a portal of entry for MRSA into the bloodstream. However, it was also notable that 8% of children were aged ≤2 days at the onset of MRSA bacteraemia, raising the possibility of mother to baby transmission. Although maternal MRSA colonisation was not investigated in this study, peripartum maternal colonisation (including vaginal colonisation), while rare, has nonetheless been reported.13 14

Over recent years, there has been considerable interest in the emergence of community-associated MRSA (CA-MRSA) which are phenotypically and genotypically distinct from MRSA strains previously seen in the hospital setting and which were originally reported from children with no apparent epidemiological links to hospitals or other healthcare settings.15 16 To date, CA-MRSA strains characteristically contain genes encoding the PVL toxin which may function as a virulence factor. CA-MRSA have been particularly problematic in the USA, where they have frequently been reported as causing infections in children.17 18 In contrast to the situation in the USA, in the present study, about 73% of isolates available for typing belonged to one or other of the two EMRSA strains that are the predominant strains causing healthcare-associated infections in the UK and Ireland. Only five isolates (5.4% of those tested) carried the PVL genes, and this finding, coupled with analyses of MRSA referred to the HPA Staphylococcus Reference Section,6 support the view that CA-MRSA remain uncommon in the UK. However, an important caveat is that CA-MRSA are commonly associated with skin and soft tissue infections, which was not the focus of the current study.

In summary, this study has characterised the population of children with MRSA bacteraemia. As a high proportion of the cases occurred in either SCBUs or NICUs, further surveillance efforts should be focused in these units, with a view to identifying more fully those children at increased risk of MRSA bacteraemia. With regard to interventions aimed at reducing MRSA bacteraemia in children, one approach may be the application of care bundles, which are currently being developed by the Royal College of Paediatrics and Child Health and the National Patient Safety Agency.19 Although control measures such as screening for MRSA, active decolonisation and enhanced aseptic technique when inserting intravascular lines intuitively seem useful, further active surveillance will be required to assess their effectiveness in preventing infection. Such an approach is a prerequisite for implementation of evidence-based clinical practice.


The authors acknowledge the BPSU, supported by the Department of Health, for facilitating the data collection and the reporting clinicians, particularly those who completed the questionnaires. The authors also thank all the microbiologists who reported cases and submitted isolates. Special thanks must also go to Richard Lynn and Jennifer Ellinghaus at the BPSU. Any views expressed in this paper are those of the investigators and not necessarily those of the BPSU or DH.


View Abstract


  • Funding Department of Health.

  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the Eastern MREC.

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

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.