Outbreak Intervention for Bloodstream Infections Caused by Methicillin Resistant Coagulase-Negative Staphylococci in Neonatal Intensive Care Unit

Coagulase-negative staphylococci (CoNS), which are members of the normal flora of the human skin, have emerged as important nosocomial pathogens during the past few decades [2]. CoNS, particularly Staphylococcus epidermidis and Staphylococcus haemolyticus, may cause severe infections in newborns, including meningitis or bacteraemia but, on the other hand, they are prevalent in the hospital environment and on the hands of health care workers [3]. Multiple drug resistance (MDR) has been widely documented among CoNS and probably reflects the adaptation of their nosocomial clones to the antibiotic selective pressure [4]. Colonization of patients and hospital staff with CoNS usually precedes infection with these organisms and constitutes a reservoir for antibiotic resistance genes [2]. To shed light on the clinical significance of CoNS in infections and provide data for control and epidemiological measures, reliable identification of this organism is crucial [3]. Several publications describing epidemiology of CoNS infections in NICU appeared within the last few years and emphasised increasing importance of these organisms in neonatal sepsis related to catheter use [5,6].


Introduction
Advances in care provided in Neonatal Intensive Care Units (NICU) have improved survival among very-low-birth weight (VLBW) infants, but healthcare-associated infections remain an important cause of morbidity and mortality in this high-risk population [1].
Coagulase-negative staphylococci (CoNS), which are members of the normal flora of the human skin, have emerged as important nosocomial pathogens during the past few decades [2]. CoNS, particularly Staphylococcus epidermidis and Staphylococcus haemolyticus, may cause severe infections in newborns, including meningitis or bacteraemia but, on the other hand, they are prevalent in the hospital environment and on the hands of health care workers [3]. Multiple drug resistance (MDR) has been widely documented among CoNS and probably reflects the adaptation of their nosocomial clones to the antibiotic selective pressure [4]. Colonization of patients and hospital staff with CoNS usually precedes infection with these organisms and constitutes a reservoir for antibiotic resistance genes [2]. To shed light on the clinical significance of CoNS in infections and provide data for control and epidemiological measures, reliable identification of this organism is crucial [3]. Several publications describing epidemiology of CoNS infections in NICU appeared within the last few years and emphasised increasing importance of these organisms in neonatal sepsis related to catheter use [5,6].
The main aim of this study was to investigate the outbreak of CoNS bloodstream infections (BSI) among 18 neonates along with an implementation of preventive strategies.
(holiday period) for 11 weeks in the group of 1016 newborns. In this period of time, neonatologists and the infection control team registered 25 cases of BSI, which accounted for 62% of all infections. As BSI aetiological factors, CoNS were dominant (72%) followed by Enterobacteriaceae (28%). Generally, in the outbreak period, 18 cases of BSI caused by CoNS were registered with a large number of new cases in the 34 th week of 2009, which was associated with a four-day-long public holiday.
To analyze and terminate the outbreak, an independent audit was performed in October 2009 by a team of experts from the Chair of Microbiology, Jagiellonian University Medical College in Krakow. Meetings were held with the nursing staff, neonatologists, nurse practitioners, pharmacy and infection control team to review patient records and establish a plan for decreasing the rate of infections in this high-risk group.
The post-intervention period lasted from January to March 2010 for 13 weeks in the group of 569 newborns. In that period of time, 12 new cases of BSI with CoNS aetiology were registered.
In order to compare the epidemiological situation in those two periods (outbreak period and post-intervention time), an epidemiological rate was applied. The incidence density per 1000 patient days (pds) was calculated according to the following formula: the ratio of new BSI occurrences in a specified quantity of person-time in the population at risk (patients of NICU) expressed as 1000 patient days (pds).

Epidemiological investigation
The inclusion criteria for the study group were the hospitalization in NICU for more than 2 days in the studied periods and occurrence of BSI caused by CoNS confirmed in a routine diagnostic procedure.
Medical and microbiological records were reviewed to identify instances of infection. Case patients of CoNS BSI were defined according to Gastmeier et al. as any neonate in the NICU who had clinical signs and/or symptoms of BSI had CoNS isolated from at least one blood culture [7].
In the outbreak period, 18 newborns with BSI caused by CoNS related to central venous catheterization (CVC) and/or peripheral venous catheterization (PVC) were included into a detailed analysis. In all of the cases, the first signs and symptoms of infection appeared in their third day of hospitalisation or later, and CVC or PVC was employed for at least 24 hours (Table 1).

Bacterial strains
Standard laboratory methods were used to identify the microorganisms that grew from blood cultures.
Polymicrobial CoNS BSI was considered present if there was a record of a blood culture that yielded more than 1 causative organism belonging to the genus Staphylococcus [8].
We had to exclude strains from 4 newborns because relevant isolates were lost in hospital laboratory. Finally, we analysed the phenotype and genotype characteristics of 17 CoNS isolates derived from the blood samples of 14 investigated newborns. The investigated strains included 10 isolates of S. haemolyticus and 7 isolates of S. epidermidis.
Cultures were also performed on environmental samples from NICU, including patient-care and unit devices, e.g. incubators, respirators, as well as staff and parents' uniforms. Moreover, besides these samples, each member of the NICU staff was examined for the presence of dermatitis, or other skin lesions on the hands, and hand cultures were obtained. Environmental samples were collected during the course of the outbreak twice: 4-5.08.2009 and also 18-24.08  environment were included and analysed for their phenotype and genotype characteristics.

Antibiotic susceptibility testing
The resistance phenotypes were determined by disk diffusion method (Oxoid, England) according to CLSI guidelines [9].

Multiplex PCR
The presence of species-specific genes to S. haemolyticus or S. epidermidis and methicillin-resistance gene mecA was confirmed using multiplex PCR amplification according to Pereira et al. [10] with specific primers (

PFGE and calculation of genetic similarity
Isolates were genotyped by pulsed field gel electrophoresis (PFGE) according to GenePath Group1 protocol (Bio-Rad). Bacteria were cultured in 3 mL of Todd-Hewitt broth (Difco Laboratories, USA) at 37°C for 24 h. DNA was digested using SmaI (MBI Fermentas, Latvia) and the fragments were separated in a CHEF-DR III device (Bio-Rad) using the program for Staphylococcus aureus for 24 h according to the manufacturer's instructions (Bio-Rad) with the following parameters, bloc 1: pulses 5-12 sec, 5V/cm, 11 h, 12°C; bloc 2: 20-60 sec, 6V/cm, 13 h, 12°C. The genetic similarity between isolates was calculated using Molecular Analyst (Applied Maths, Belgium) software, making use of the Jaccard coefficient and unweighed pair group method with arithmetic mean (UPGMA). Genotype patterns were compared according to the guidelines of van Belkum et al. [11].

Results
In the outbreak period, 18 newborns with BSI caused by CoNS were included into a detailed analysis, from among 9 (50%) were female and 9 (50%) male, with the average value equal of: gestational age 29.5 weeks, birth-weight 1200 g, infection day 16.2 and length of catheterization 13.3 days (Table 1).
In the study period, the incidence density of BSI in the group of neonates with low and VLBW was 4.5/1000 patient days (pds) from June to September 2009, whereas the incidence density of CoNS BSI was 3.3/1000 pds. Mortality rate in the group of 18 newborns with BSI caused by CoNS was 11.1%.
Average total length of hospitalisation in the outbreak period was 49.2 days, in opposition to average length of stay for patients in 2009 amounting to 5.4 days (95% CI 0.3 -10.2). Moreover, the mean duration of follow-up catheter until the first symptoms of BSI was 14.3 days (95% CI 0.1-4.1).
A large number of new CoNS BSI cases were registered in the 34 th week of 2009 ( Figure. 1), during which a four-day-long public holiday was held and the majority of the permanent unit staff was away on leave. They were then replaced in part by residents and younger nurses on duty.
PFGE analysis revealed that all investigated patients were infected with identical or closely related CoNS clones. Figure 2 shows the genetic similarity among selected strains of S. haemolyticus, 10 strains from patients (P1-P7, P12-P14) and 4 strains from the NICU environment (E1-E4). At a cut off level of 73% similarity, 2 clusters comprising 12 (cluster A) or 2 (cluster B) strains were identified. Figure 3 shows DNA macrorestriction profiles for selected strains of S. epidermidis, 7 from    patients (P8-P14) and 1 from the environment (E1). At a cut off level of 78% similarity, 3 clusters were noticed, comprising 4 (cluster A*), 3 (cluster B*) and 1 (cluster C*) strains with identical profiles.
Out of 21 methicillin-resistant S. haemolyticus strains isolated from the environment, four isolates were admitted to the cluster A (Figure 2), from which two strains, E1SH and E3SH, were isolated from medical personnel's hands, whereas E2SH and E4SH originated from the incubator and staff uniform, respectively. Out of 4 methicillin-resistant S. epidermidis strains isolated from the environment, one strain (E1SE) originating from the respirator was identified as PFGE type A* ( Figure  3).

Termination of the outbreak
The audit performed in October 2009 by an independent expert team from the Chair of Microbiology, Jagiellonian University Medical College in Krakow helped the local unit to establish a plan for decreasing the rate of infections in this high-risk group. The Three Step Plan was developed to address high infection rates. The following procedures, according to CDC recommendations, were standardized: hand hygiene, skin preparation, frequency and hygiene of line changes, blood culture collections and dressing changes [12,13]. All these modifications resulted in a positive outcome both for the infants, decreasing incidence rate of BSI from 4.5/1000 pds to 2.4/1000 pds, and the nursing and medical staff.

Discussion
Newborn babies treated in NICU, particularly prematurely born and newborns with VLBW are prone to developing infections. It is associated both with newborn immune system immaturity and with the need to perform many invasive diagnostic and therapeutic procedures [14][15][16].
There are great difficulties with studying the results of blood cultures from premature infants. Interpretation of blood cultures that are positive for CoNS is particularly difficult because these organisms live on the skin and mucous during the collection process. The neonatologists in our NICU do not obtain blood samples unless a patient has clinical signs and a syndrome compatible with bacteraemia according to Gastmaier et al. [7]. Isolation of CoNS from a single blood culture under these circumstances neither rules out contamination nor establishes a definitive diagnosis. A similar problem was discussed by Sidebottom et al. and Huang et al. These scientists indicated that the newborns with a CoNS-positive single blood culture had, in fact, bacteraemia [17,18].
In the hospital epidemiology, where exposure to the occurrence of infection depends on the length of hospitalization and length of device use, especially in NICU, incidence density is usually used. Incidence associated with BSI in previous study reached 8.5/1000 pds in newborns weighing <1000 g to 4/1000 pds in neonates up to 1500 g [19,20]. In our study, the incidence of BSI was high, despite the fact that the study group consisted of both healthy children and low and VLBW newborns. The high incidence rate was correlated both with CVC and with PVC, which showed important risk factor of BSI at the examined centre. Some studies have found similar association between different types of catheter and BSI [21].
Despite the importance of CoNS as a cause of hospital acquired infections, limited information has been available regarding their reservoirs, distribution, and mode of transmission within the hospital environment. CoNS are members of the human skin flora of adult people but certain epidemic strains can colonize skin for extended periods of time, and transient hand carriage of such strains is thought to be a mode of transmission when no source is identified [22].
Our study demonstrated the occurrence and spread of identical, or closely related, CoNS clones, between newborns. Furthermore, identical CoNS were isolated from the hospital environment, especially from medical staff hands, which suggested that the transmission of CoNS could occur between babies via the staff hands. Thus, molecular studies indicated horizontal spread of selected clones. These findings correlate with Raimundo et al. who add to the increasing evidence that epidemic clones of S. epidermidis may become established in NICU and transmitted between babies [23]. Monsen et al. [4] suggested that clonal spread of CoNS may occur within and even between wards. Similarly, Takei et al. [24] showed the transmission of MRSA/MRCNS in the old NICU of Hospital A and the relocated NICU (new NICU). A homology study of MRSA/MRCNS strains by PFGE DNA restriction patterns identified genotypes that showed similar patterns in the nurses' palms, towels, infant incubators, and room air. Genetic analysis carried out by Hira et al. [25] showed that 90% of the CoNS blood isolates were closely related to the strains found on the personnel's hands and in the hospital environment and they differed from those in the general population. In our study, the detected clones, e.g. S. haemoliticus PFGE type A, were found and remained in the NICU during the study period of almost three months. These findings corroborate results from other reports indicating that certain clones of CoNS may persist over time in certain wards, and that such clones often seem to be ward-specific, reflecting the antibiotic consumption in the paediatric ward [2,4].
Our study indicated high level (21%) of polymicrobial PVPassociated BSI caused by S. haemolyticus and S. epidermidis. Data on polymicrobial BSI in neonates are extremely limited. A high percentage of polymicrobial CVC-associated BSIs among paediatric patients receiving outpatient therapy has been noted in recent years [8]. Polymicrobial BSIs are problematic for clinicians, because appropriate treatment frequently requires multiple broad-spectrum antibiotics [26]. Downes et al. [26] hypothesized that certain risk factors may predispose children to the development of polymicrobial BSIs. Polymicrobial BSIs are relatively common in hospitalized neonates, occurring in approximately 3% to 10% of cases of neonatal BSI [8]. Bizzaro et al. 's [8] study confirmed the association of polymicrobial BSI in infants with the presence of severe underlying conditions. Moreover, the described situation might have been connected to the ability of CoNS to create mixed biofilms on medical devices, from which they can easily spread to the blood of neonates, who are particularly susceptible to this kind of infections.
In most cases, catheter-associated BSI (CABSI) can be prevented. Thus, hospital policies and care procedures should be directed toward the adoption of preventive measures rather than merely the identification and treatment of CABSI. During the past 15 years, there have been major advances in understanding the epidemiology and pathogenesis of CABSI in adult population. The concept described by Sax et al. [27] showed that hand hygiene and educational programs were related to a significant reduction in infection rates.
An important risk factor for BSI in the studied period was inadequate hand hygiene of staff which may be linked to less intensively trained doctors and nurses on duty during holidays. Unfortunately, appropriate action concerning changes in the situation on the ward, including hygiene behavior, were taken in the joint action of microbiologists and neonatologists four months after the first observation of the outbreak. In the light of the results achieved, we are fully aware of the causes of the existing situation and sub-optimal infection control. Th e cooperation between studied NICU staff and the infection control team as well as the expert team permitted to review the patients' records and establish a plan for decreasing the rate of infections in this high-risk group. The modifications resulted in a positive outcome for the infants and the nursing as well as medical staff.
The presented results come from the Polish NICU and are first to describe the procedure associated with epidemic and extinguishing outbreaks by education on hand hygiene. It confirms the fact that, particularly in NICU, rigorous infection control measures such as strict isolation, environmental cleaning, staff education and proper hand hygiene are very important, in particular to terminate an on-going outbreak and prevent new infections. Moreover, the described situation has shown a failure to prepare Polish infection control teams for crisis situations, which an epidemic is.
In summary, the investigated BSI caused by coagulase-negative staphylococci in the Neonatal Intensive Care Unit appeared due to clonal dissemination, therefore, it was potentially preventable by strict adherence to the recommended infection control practices and the implementation of programs aimed at the reduction of unnecessary use of antibiotics.