Detection of Oxacillin Resistance in Staphylococcus aureus Isolated from the Neonatal and Pediatric Units of a Brazilian Teaching Hospital.

OBJECTIVE
To determine, by phenotypic and genotypic methods, oxacillin susceptibility in Staphylococcus aureus strains isolated from pediatric and neonatal intensive care unit patients seen at the University Hospital of the Botucatu School of Medicine.


METHODS
A total of 100 S. aureus strains isolated from the following materials were studied: 25 blood cultures, 21 secretions, 12 catheters, 3 cannulae and one chest drain from 62 patients in the neonatal unit, and 36 blood cultures, one pleural fluid sample and one peritoneal fluid sample from 38 patients in the pediatric unit. Resistance of the S. aureus isolates to oxacillin was evaluated by the disk diffusion method with oxacillin (1 μg) and cefoxitin (30 μg), agar screening test using Mueller-Hinton agar supplemented with 6 μg/ml oxacillin and 4% NaCl, and detection of the mecA gene by PCR. In addition, the isolates were tested for β-lactamase production using disks impregnated with Nitrocefin and hyperproduction of β-lactamase using amoxicillin (20 μg) and clavulanic acid (10 μg) disks.


RESULTS
Among the 100 S. aureus strains included in the study, 18.0% were resistant to oxacillin, with 16.1% MRSA being detected in the neonatal unit and 21.0% in the pediatric unit. The oxacillin (1 μg) and cefoxitin (30 μg) disk diffusion methods presented 94.4% and 100% sensitivity, respectively, and 98.8% specificity. The screening test showed 100% sensitivity and 98.8% specificity. All isolates produced β-lactamase and one of these strains was considered to be a hyperproducer.


CONCLUSIONS
The 30 μg cefoxitin disk diffusion method presented the best result when compared to the 1 μg oxacillin disk. The sensitivity of the agar screening test was similar to that of the cefoxitin disk diffusion method and higher than that of the oxacillin disk diffusion method. We observed variations in the percentage of oxacillin-resistant isolates during the study period, with a decline over the last years which might be related to improved nosocomial infection control and the rational use of antibiotics.


Introduction
The genus Staphylococcus comprises about 40 species, with Staphylococcus aureus being the main representative and the causative agent of a wide variety of infections. Nosocomial infections are the main cause of morbidity and mortality in pediatric (PICU) and neonatal intensive care units (NICU).
Antimicrobial resistance is easily acquired by these microorganisms because of the high risk of plasmid transfer between strains in the hospital environment and the abusive use of antimicrobial drugs, which represent important factors for the transfer of resistance genes and the selection of multiresistant strains. 1,2 The prevalence of oxacillin-resistant S. aureus (MRSA) in hospitals has increased in most countries. The intrinsic resistance of S. aureus to oxacillin is due to the production of a supplemental penicillin-binding protein (PBP 2' or PBP 2a), which presents low affi nity for semi-synthetic penicillins and is encoded by the chromosomal gene mecA. Although the resistance of mecA gene is present in all the cells of the population with intrinsic resistance, this could be expressed by a small percentage of them, leading to what has been called heterogeneous resistance. Other resistance mechanisms have been described in strains that do not possess the mecA gene, with these strains being called borderline resistant. Borderline resistance is caused by two mechanisms: inactivation of oxacillin due to β-lactamase hyperproduction 3 and modifi ed resistance, known as MOD-SA, due to the production of modifi ed intrinsic PBPs with affi nity for altered oxacillin. 4 The reference methods for the detection of oxacillin resistance in S. aureus recommended by the Clinical and Laboratory Standards Institute (CLSI) 5 include the determination of the minimum inhibitory concentration (MIC) of the drug by the agar or broth dilution method, disk diffusion method and, more recently, the cefoxitin disk diffusion test. 5,6 Since phenotypic methods for the detection of MRSA may yield questionable results, molecular tests for the detection of the mecA gene or its product PBP 2a have been proposed. Detection of the mecA gene by the polymerase chain reaction (PCR) is considered to be the gold standard for the diagnosis of MRSA. The determination of oxacillin resistance in Staphylococcus is important to guide therapy and to prevent unnecessary treatment of the patient with vancomycin, an antibiotic associated with therapeutic complications and whose use may lead to the selection of resistant strains. The prevalence of MRSA presents varies widely, particularly as a function of the size and type of the health institution. Thus, the objective of the present study was to determine oxacillin susceptibility in S. aureus strains isolated from patients hospitalized in the PICU and NICU of the University Hospital of the Botucatu School of Medicine (HC-FMB) using the oxacillin and cefoxitin disk diffusion methods. These methods were compared to mecA gene detection by PCR.

Strains
A total of 100 S. aureus strain isolated from the following materials were studied: 25 blood cultures, 21 secretion, 12 catheters, 3 cannulae and one chest drain from 62 patients hospitalized in the NICU of HC-FMB, and 36 blood cultures, one pleural fluid sample and one peritoneal fluid sample from 38 patients hospitalized in the PICU between 1991 and 2007. The strains were isolated as described by Koneman et al. 7 Detection of oxacillin resistance by the 1 μg oxacillin and 30 μg cefoxitin disk diffusion method and by the agar screening test using Mueller-Hinton agar supplemented with 4% NaCl and 6 μg/mL oxacillin Oxacillin sensitivity was tested by the agar disk diffusion method according to the criteria of the CLSI. 5 The inocula were prepared in brain-heart infusion (BHI) broth and incubated for 4 to 6 h and turbidity was set to 0.5 McFarland standard. The following disks were used: 1 μg oxacillin and 30 μg cefoxitin. Once the density was adjusted, the inoculum was spread with a sterile swab on Mueller-Hinton agar and the disks impregnated with the drug were applied. The plates were incubated for 24 h at 35 °C and antimicrobial activity was evaluated by determining the diameter of the inhibition zone as recommended by the CLSI. 5 The S. aureus ATCC 25923 (negative control) and ATCC 33591 (positive control) reference strains were used in all experiments.
For the detection of MRSA, Mueller-Hinton agar plates containing 6 μg/ml oxacillin and 4% NaCl were used. 8 The inoculum was prepared by previous incubation in BHI broth for 24 h and turbidity was set to 0.5 McFarland standard. After preparation of the inoculum, the strains were seeded in spots on the agar surface with a sterile swab. The plates were incubated for 24 h at 35 °C and the presence of MRSA was defi ned as the growth of at least one colony on the agar surface.
Detection of the mecA gene by PCR Total nucleic acids were extracted from S. aureus strains cultured on blood agar, individually inoculated into BHI broth and incubated for 24 h at 37 °C. DNA was extracted using the GFX kit (Amersham Pharmacia Biotech). Briefl y, staphylococcal cells were fi rst digested with 10 μg/ml lysozyme and 20 μg/ml proteinase K. Next, 500 μl of the extraction solution was added and the mixture was centrifuged at 5000 × g for 1 min. The supernatant was then transferred to a GFX column and centrifuged at 5000 × g for 1 min. The collected fl uid was discarded and 500 μl of the extraction solution was again added to the column. After centrifugation and discarding the collected fl uid, 500 μl of the washing solution was added to the column which was centrifuged at 14,000 × rpm for 3 min. The column was then transferred to a 1.5-ml tube and 200 μl Milli-Q water heated to 70 °C was used for elution. The isolates were centrifuged at 5000 × g for 1 min and the GFX column was discarded. The extracted DNA was stored in a refrigerator at 4 °C.
PCR was carried out in 0.5-ml microcentrifuge tubes in a total volume of 25 μl containing 1 μM of each primer, 2.0 U Taq polymerase, 100 μM deoxyribonucleotide triphosphates, and 150 μg nucleic acid. The following primers were used: mecA1 (forward)-5' AAA ATC GAT GGT AAA GGT TGG 3', and mecA2 (reverse) -5' AGT TCT GCA GTA CCG GAT TTG 3'. The size of the amplifi ed product is 533 bp. PCR was carried out in an appropriate thermocycler using the following parameters as described by Murakami et al. 9 40 cycles of denaturation at 94 o C for 30 s, annealing at 55.5 °C for 30 s and extension at 72 °C for 1 min. After the 40 cycles, the tubes were incubated at 72 °C for 5 min before cooling to 4 °C. International reference strains were used as positive (S. aureus ATCC 33591) and negative controls (S. aureus ATCC 25923) in all tests.
The effi ciency of the amplifi cation reactions was evaluated by electrophoresis on 2% agarose gel prepared in 0.5X TBE buffer and stained with ethidium bromide. The size of the amplified products was compared with the 100-bp standard and the gels were photographed under UV illumination.

Determination of β-lactamase production and β-lactamase hyperproduction
The production of β-lactamase by the S. aureus isolates was tested using disks impregnated with Nitrocefi n (chromogenic cephalosporin, cefi nase, BBL). The disk was moistened with one or two drops of sterile distilled water and placed on the S. aureus colonies previously incubated for 24 h at 35 o C on the Mueller-Hinton plates used for the disk diffusion test with oxacillin (1 μg) and cefoxitin (30 μg), because these drugs stimulate β-lactamase production in Staphylococcus.
A positive reaction was indicated by the development of a red color, whereas the absence of a change in color indicated a negative reaction. For β-lactamase-negative strains, the reaction was reexamined after 1 h according to manufacturer recommendations. For correct analysis of the results, the tested disks were compared with positive (S. aureus ATCC 29213) and negative (S. xylosus ATCC 29979) control strains.
The isolate that was negative for the mecA gene and presented resistance to oxacillin by the phenotypic methods was tested to determine whether it was a hyperproducer of β-lactamase. This strain was tested using disks containing amoxicillin (20 μg) and clavulanic acid (10 μg). The sensitivity breakpoint was the formation of an inhibition zone Ն20 mm after incubation for 24 h at 35 °C. 10

Statistical analysis
Sensitivity and specifi city tests 11 were performed to compare the agar screening method, disk diffusion test with cefoxitin and oxacillin and PCR. The last method is considered to be the gold standard for the detection of intrinsic oxacillin resistance (detection of the mecA gene).
Sensitivity was defi ned as the proportion of PCR-positive S. aureus isolates (detection of the mecA gene) that were resistant to oxacillin using the following phenotypic methods: disk diffusion with cefoxitin and oxacillin, and agar screening test (Mueller-Hinton agar supplemented with 6 μg/mL oxacillin and 4% NaCl).
Specifi city was defi ned as the proportion of PCR-negative S. aureus isolates (no detection of the mecA gene) that were sensitive to oxacillin using the following phenotypic methods: disk diffusion with cefoxitin and oxacillin, and agar screening test (Mueller-Hinton agar supplemented with 6 μg/mL oxacillin and 4% NaCl).

Results
Detection of oxacillin resistance by the 1 μg oxacillin and 30 μg cefoxitin disk diffusion method and by the agar screening test using Mueller-Hinton agar supplemented with 4% NaCl and 6 μg/mL oxacillin Among the 100 S. aureus strains isolated from PICU and NICU patients of HC-FMB, 11 (11%) were found to be resistant to oxacillin in the 1 μg oxacillin disk diffusion test. In seven (7%) isolates, growth of colonies was observed inside the inhibition zone, suggesting heterogenous resistance. Using 30 μg cefoxitin disks, oxacillin resistance was observed in 18 (18%) isolates and one (1%) presented growth of colonies inside the inhibition zone, suggesting heterogeneous resistance (Fig. 1). Nineteen (19%) isolates were found to be resistant by the agar screening method using S. aureus ATCC 33591 and ATCC 25923 as control strains (Fig. 2).
Detection of the mecA gene by PCR PCR revealed the presence of the mecA gene in 18 (18%) isolates. Eleven of these isolates were

Detection of oxacillin resistance in Staphylococcus aureus
Clinical Medicine: Pediatrics 2009:3 resistant in the 1 μg oxacillin and 30 μg cefoxitin disk diffusion tests and one isolate only showed resistance in the 30 μg cefoxitin disk diffusion test. Among the other six PCR-positive isolates, fi ve presented heterogenous resistance by the oxacillin disk diffusion method and resistance by the cefoxitin disk diffusion method. One isolate showed heterogenous resistance with both disk diffusion methods. All 18 mecA-positive isolates were found to be resistant in the agar screening test (Fig. 3 and Table 1).

Determination of β-lactamase production and β-lactamase hyperproduction
The production of β-lactamase was determined using disks impregnated with Nitrocefin. All S. aureus isolates were confi rmed to be producers of β-lactamase.
The isolate that was negative for the mecA gene, was resistant by the 30 μg cefoxitin disk diffusion and screening methods, and showed heterogenous resistance in the 1 μg oxacillin disk diffusion test was studied to determine whether it was a hyperproducer of β-lactamase. This was tested with a disk of amoxicillin (20 μg) and clavulanic acid (1 μg) and was sensitive to this drug, then confi rming the hyperproduction of β-lactamase, that was inhibited by the presence of clavulanic acid.

Statistical analysis
Statistical analysis showed 94.4% sensitivity and 98.8% specifi city of the oxacillin disk diffusion method and 100% sensitivity and 98.8% specifi city of the cefoxitin disk diffusion method. The sensitivity and specifi city of the screening test was 100% and 98.8%, respectively ( Table 2).
Evolution of oxacillin resistance in S. aureus strains isolated from patients seen at HC-FMB Sixty-two samples from NICU patients and 38 from PICU patients were studied. Among the strains isolated from the neonatal unit, 10 (16.1%) were resistant to oxacillin, three of them isolated from blood cultures, 4 from secretions and 3 from catheters. Eight (21%) of the strains isolated from the PICU were MRSA, with 7 strains being isolated from blood cultures and one strain isolated from pleural fl uid (Table 3).
The strains of the present study were isolated between 1991 and 2007. Twenty-eight strains were isolated between 1991 and 1994, and seven (25%) were found to be resistant to oxacillin. Twenty-fi ve strains were isolated between 1995 and 2000, and two (8%) were MRSA. Of the 24 strains isolated between 2001 and 2004, seven (29.2%) were MRSA. Twenty-three strains were isolated between 2005 and 2007, and two (8.7%) were resistant to oxacillin (Fig. 4).  (Table 4 and Fig. 4).

Discussion
The prevalence of oxacillin-resistant S. aureus in hospitals has increased in most countries. Prevalence rates of MRSA may vary, particularly as a function of the size and type of the medical institution. In the present study, 100 S. aureus strains isolated from inpatients of the PICU and NICU of HC-FMB were tested for MRSA.
Although high frequencies of oxacillin-resistant S. aureus have been reported especially in large hospitals and university hospitals, in the present study the percentage of MRSA isolated from inpatients of the PICU and NICU of HC-FMB between 1991 and 2007 was only 18%. Similar results have been reported for PICU and NICU patients with bacteremia in the United Kingdom, with a percentage of MRSA of 15.5%. 12 Furthermore, the prevalence of MRSA observed here was lower than that reported in studies conducted in other countries. In the United States, a prevalence of MRSA infection of 47% was reported in a hospital in Texas in 2003. 13 A study conducted in India found a prevalence of oxacillin-resistant S. aureus of 66% among NICU inpatients. 14 Similar rates have been reported in studies conducted in Japan and Israel, with MRSA infection rates of 52.5% and 60%, respectively, among inpatients of an NICU. 15,16 In the present study, phenotypic tests such as the disk diffusion method and agar screening test were compared with a genotypic method (mecA gene detection by PCR). The cefoxitin disk diffusion method presented 100% sensitivity and 98.8% specifi city and was superior to the oxacillin disk diffusion method (94.4% sensitivity and  According to these authors, compared to the gold standard (mecA gene detection), the disk diffusion method with 30 μg cefoxitin is preferable to the 1 μg oxacillin disk diffusion method for the detection of MRSA.
The mecA gene was detected in all isolates that were resistant in the Mueller-Hinton agar screening test. The sensitivity of the method was 100% and specif icity was 98.8%. The performance of the screening method using Mueller-Hinton agar supplemented with 4% NaCl and 6 μg oxacillin depends on the degree of heterogeneity of the isolates tested, with lower sensitivity (Յ95%) being reported in studies including a larger number of heteroresistant strains and values Ͼ97% in other studies. [22][23][24] In this investigation, the sensitivity of the agar screening test for the detection of MRSA was similar to that of cefoxitin disks and higher than that of the oxacillin disk diffusion method. The high sensitivity of the agar screening method might be explained by the small number of heteroresistant strains. Different results have been reported by Cauwelier et al. 20 who observed 91.7% sensitivity and 100% specifi city. The authors attributed the low sensitivity in the detection of heterogeneous populations to the high variability in the expression of PBP 2a.
Other resistance mechanisms have been described in strains that do not possess the mecA gene, with these strains being called borderline resistant. There are two mechanisms of borderline resistance, one is the inactivation of oxacillin due to hyperproduction of β-lactamase 3 and the other is modifi ed resistance, called MOD-SA, due to the production of modifi ed intrinsic PBPs with affi nity for altered oxacillin. 4 However, these mechanisms are characterized by low levels of resistance (MIC of 8 μg/ml). 24 In the present study, one isolate presented resistance to oxacillin in all phenotypic  tests despite the absence of the mecA gene and was sensitive to the amoxicillin-clavulanic acid disk. This strain was therefore considered to be a hyperproducer of β-lactamase, suggesting borderline resistance.
The MRSA strains detected during the fi rst two study periods (1991 to 2000) were isolated from samples originating from the neonatal unit. This fi nding can be explained by the fact that all strains studied until 1998 exclusively originated from the neonatal unit and strains from the pediatric unit were only included after 1999. Between 2001 and 2004, all oxacillin-resistant strains originated from the pediatric unit, with a decline in the number of MRSA originating from the NICU. In contrast, in the last period the proportion of MRSA originating from the NICU and PICU was the same. Overall analysis of our data showed variations in the percentage of oxacillin-resistant S. aureus during the study period. The percentage of resistant isolates declined between 1995 and 2000, followed by an increase from 2001 to 2004, and then declined  again over the following years. These reductions might be related to improved nosocomial infection control and the rational use of antibiotics.