An outbreak of methicillin resistant Staphylococcus epidermidis among neonates in a hospital in Saudi Arabia

Introduction: Staphylococcus epidermidis is a pathogen associated with nosocomial infection in neonatal intensive care units (NICU). This study investigates an outbreak of methicillin resistant S. epidermidis in an NICU in a hospital in Saudi Arabia. Methodology: A total of 41 isolates identified as Gram-positive cocci were obtained from blood culture, umbilical wound swabs and endotracheal aspirate specimens of neonates, of which 29 were identified as S. epidermidis. Bacterial identification at the species level and determination of antibiotic resistance were performed by MicroScan (Dade Behring, USA). Genotyping was completed using randomly amplified polymorphic DNA (RAPD) and the mecA gene was detected by PCR. Results: All 29 S. epidermidis isolates were found to be resistant to oxacillin and were positive for the mecA gene. The isolates showed several multidrug-resistance patterns; the resistance rates to gentamicin, erythromycin, clindamycin, and trimethoprim/sulfamethoxazole were 89.7%, 86.2%, 75.9% and 72.4%, respectively. All isolates were susceptible to vancomycin, teicoplanin, rifampin, synercid, and ciprofloxacin. Several genotypic and phenotypic patterns were detected among the S. epidermidis isolates: antibiogram typing showed seven different patterns, one of which was shared by 65% of the isolates, whereas the most prevalent RAPD genotype was shared by only five S. epidermidis isolates, and did not correlate with antibiotic resistance phenotype. Conclusion: The diverse clonal origin of tested isolates indicates the presence of multiple S. epidermidis strains among neonates in the NICU setting


Introduction
Staphylococcus epidermidis belongs to the coagulase negative staphylococci (CoNS), a group of Gram-positive cocci that causes a high incidence of bloodstream nosocomial infections [1].S. epidermidis is a major nosocomial pathogen, frequently isolated from the normal skin microbiota of patients and healthy individuals [2,3].
Nosocomial infections caused by S. epidermidis have been reported in neonatal intensive care units (NICU) in hospitals in the Netherlands and in Italy [4,5].Similar studies in Saudi hospitals reported that S. epidermidis caused 36-40% of infections in NICUs [6,7].Moreover, cases of S. epidermidis pediatric bloodstream infections (55%) [8] in Saudi hospitals have been reported.Other studies have reported cases among adults with S. epidermidis bacteremia [9] and postoperative wound infections [10] in Saudi hospitals.
Staphylococcus epidermidis exhibits resistance to methicillin and consequently to all betalactams upon acquisition of the mecA gene.This gene encodes the PBP-2a, a penicillin binding protein that has low affinity for beta-lactams [11].Methicillin resistant S. epidermidis isolates that circulate within hospital settings are mostly resistant to other classes of antibiotics [12] and can act as a reservoir of mobile genetic elements transferred to other S. epidermidis isolates as well as to other Staphylococcus species, such as S. aureus [11].
The geographical clonal dissemination of methicillin-resistant S. epidermidis has been demonstrated, as closely related pulsed-field gel electrophoresis (PFGE) types have been observed among isolates recovered in Iceland, Denmark, Mexico, Uruguay, Greece, and Cape Verde [12].PFGE has also been used to detect the presence of endemic CoNS clones in NICUs in two hospitals in the Netherlands [13].Raimundo et al. [14] identified one epidemic clone of S. epidermidis from an NICU in Australia using a combination of techniques including PFGE, randomly amplified polymorphic DNA (RAPD) and the antibiogram pattern.In a retrospective study between 2000 and 2002 of infants from an NICU in Canada with CoNS bacteremia, S. epidermidis was found to be the most common isolate.In this study, molecular typing was done with both PFGE and RAPD, with 14 RAPD patterns found but no predominant clone was identified [15].Although PFGE is more discriminatory than RAPD [14], the latter technique has been used to type S. epidermidis isolates [14,16,17,18], as it provides results in a more timely manner and is more cost effective [4,14].
S. epidermidis has been found to be responsible for nosocomial infections in the Middle East region [9,17,18], but very few studies have attempted to determine the extent of clonal spread of these isolates in hospital settings using genotyping [17,18].In this study, methicillin-resistant S. epidermidis (MRSE) was found to cause bacteremia in 70% of neonates during a 12-week outbreak in a 29-bed NICU at a tertiary hospital in Saudi Arabia.The possible clinical relatedness of these isolates was assessed using the RAPD technique.A correlation of RAPD and antibiotic resistance patterns was attempted.

Bacterial strains
Samples of blood culture, cerebrospinal fluid (CSF), umbilical wound swabs and endotracheal aspirates were taken from 41 neonates.Blood culture and CSF were collected by the authorized neonatologist physician.Endotracheal aspirates and umbilical swabs were collected using sterile swab collection tubes.Blood culture samples were collected in blood culture bottles (BD BACTEC Peds Plus/F)) and assayed in the BACTEC 9120 system (Becton Dickinson, Sparks, MD, USA).All other samples were plated without delay on the routine culture media plates, i.e., blood agar, MacConkey agar, chocolate agar and Sabouraud's dextrose agar (BD BBL, Franklin Lakes, NJ, USA) and incubated at 37°C for 24 hours.
Forty-one Gram-positive cocci were recovered.The coagulase-negative staphylococci (CoNS) were identified by colony morphology, Gram stain reactions, positive catalase test, and negative coagulase test using the Staphauruex Plus system (Murex Biotech Ltd, Dartford, United Kingdom).Further biochemical identification of CoNS to the species level was performed using MicroScan POS Combo Panel Type PC 1A (Dade Behring, MicroScan, Sacramento, California, USA), and 29 S. epidermidis isolates were identified, of which 24 were isolated from blood culture.The patients were sampled daily for three consecutive days, where the same organism, S. epidermidis was isolated repeatedly.Isolates were stored in glycerol-containing medium at -70°C until further analysis and were subcultured on Tryptic Soy Agar (Difco, Detroit, USA).

RAPD analysis
The DNA template for RAPD analysis was obtained as previously described [20] using Triton X-100 lysis buffer.RAPD PCR was done as previously described [16] using two different primers: ERIC-2 (5'-AAGTAAGTGACTGGGGTGAGCG-3') or primer3 (5'-TTATGTAAAAGGACGGCCAGT-3') [16].RAPD amplified products from the 29 S. epidermidis clinical isolates were analyzed on a 1.3% agarose gel by horizontal electrophoresis, in a Trisacetate-EDTA buffer.Gels were visualized under UV light and photographed with a computer-controlled image analyzer (Whatman, Biometra, USA).The resulting banding patterns were counted, compared, and assigned type names.Patterns differing by one band were considered different types.Differences in band intensity were not taken into consideration [21].The PCR assay was repeated from fresh cultures for each isolate in order to obtain reproducible results.

Medical data of patients
A total of 49 newborn patients were admitted into the NICU of North West Armed Forces hospital in Tabuk, Saudi Arabia, between 5 December 2004 and 26 February 2005.These neonates were born in the hospital and admitted into the NICU due to risk factors associated with the mothers or the babies (Table 1).Of the 49 neonates, 41 were found to have Gram-positive coccal infections.S. epidermidis accounted for 29 of these cases (70%), with other Gram-positive isolates identified as streptococci and S. hemolyticus.Of the 29 neonates infected with S. epidermidis, 18 were male and 11 were female, all but one had low birth weight, and most had risk factors such as congenital heart diseases (Table 1).Mortality was reported for 10% of patients (Table 1).

Antimicrobial susceptibility
The 29 S. epidermidis clinical isolates were tested to determine the MICs of eighteen antibiotics.The isolates showed several multidrug-resistance patterns (Table 2).The high resistance rates to beta-lactams (100%) and gentamicin (89.7%) among the isolates could be due to the increased pressure on ampicillin and gentamicin in the NICU (Table 1).These results are in agreement with those of Villari et al. [5], who reported a 94.6% resistance to both these antibiotics among their S. epidermidis isolates.The authors in that study suggested that increased pressure through use of gentamicin in the NICU might have selected for gentamicin-resistant strains.Antibiotic resistance has been found to be a selective force for certain CoNS strains that circulate within a hospital environment [12].It has been reported that the possession of mecA by CoNS strains also helps such isolates to spread in the NICU [4,21], which may have been the case in this study, as all 29 clinical isolates possessed mecA.
In this study, vancomycin was the drug of choice, prescribed to 21 of 29 cases (Table 1).The results of susceptibility testing demonstrated that all S. epidermidis isolates were not only resistant to oxacillin, but also exhibited a range of multidrugresistant patterns.In an attempt to use the antibiotic resistance profile as an epidemiological marker, seven different patterns were found (Tables 1 and 2).Pattern V was shared by 19 of the isolates (65%) (Table 2).Thus antibiotyping was not a helpful method to discriminate between 65% of S. epidermidis isolates.

Genotyping
To take preventive measures and implement infection control, identification of the source of infection is required.In this study, the use of RAPD was useful in discriminating between S. epidermidis strains tested, confirming what was found in previous studies [17,18,21].Moreover, Khashu et al. [15] found that RAPD was sufficiently discriminatory to detect differences between strains of S. epidermidis.In fact, the comparison between PFGE and RAPD results obtained in their study indicated that the RAPD method concurred with results from PFGE analysis [15].
In this study, RAPD analysis revealed 17 distinct patterns with primer ERIC-2 (Table 1), showing between three and five bands in each pattern of relative size ≤ 1 kb (data not shown).Pattern 1 one was common to six S. epidermidis isolates designated 31, 47, 29, 41, 90, and 85 (Table 1).Each of the RAPD patterns 2 to 7 was found to be shared by two or three isolates (Table 1).A total of 10 isolates had unique RAPD patterns.RAPD analysis using primer3 was less discriminatory, amplifying only two bands of     1).
The RAPD results in Table 1 show that five S. epidermidis isolates (31, 47, 29, 41, and 85) shared identical banding patterns using both primers.In addition, the results of RAPD typing corresponded with antibiogram typing for three of these five isolates (Tables 1, 2). S. epidermidis strains with a particular RAPD genotype that share an identical antibiotype have been reported earlier [12].
In conclusion, RAPD and antibiogram types detected among the S. epidermidis isolates over the examined time period did not indicate any common clones, except for one RAPD genotype shared by five S. epidermidis isolates.The diverse clonal origin of the tested isolates indicates that there was no dissemination of a single S. epidermidis strain among neonates in the NICU setting.
Recommendations were set to limit infection in the NICU and included the use of sterile gowns when entering the NICU, changing gloves when handling different neonates, and hand hygiene using soap and drying with paper towel or rubbing with an alcoholbased product or chlorhexidine gluconate 0.4%, monitored by a surveillance camera placed in front of the dispensing unit.

Table 2 .
Patterns of the antibiogram relative size < 0.5-1.0Kb (data not shown) and resulting in only four different RAPD patterns (Table