Clinical features and molecular characteristics of childhood community-associated methicillin-resistant Staphylococcus aureus infection in a medical center in northern Taiwan, 2012

Background Since first reported in 2002, the rate of methicillin-resistant Staphylococcus aureus (MRSA) among childhood community-associated (CA) S. aureus infection in Taiwan increased significantly up to 2005. There have been no reports on this issue since then. Methods We prospectively collected clinical S. aureus isolates from the patients <19 years of age in a university-affiliated hospital in 2012. Only first isolate from each patient was included. The medical records were retrospectively reviewed and the patients were classified as CA or healthcare-associated (HA) by the standard epidemiologic criteria. Isolates as CA-MRSA were further characterized by pulsed-field gel electrophoresis, staphylococcal cassette chromosome (SCCmec) typing, and multilocus sequence typing. Results A total of 409 S. aureus isolates were included, and 260 (63.6%) were MRSA. The proportion of MRSA among all S. aureus isolates in 2012 increased significantly (p < 0.001) compared to that in 2004–2005. Of the 181 CA-MRSA isolates, 86.2% were identified from pus or wound. Nine pulsotypes were identified with two major types (type D, 119 (65.7%); type C, 27 (14.9%). Most of the isolates carried either SCCmec IV (66 isolates, 36%) or VT (112 isolates, 62%). 128 isolates (71%) carried Panton-Valentine leukocidin (PVL) genes. Clonal complex (CC) 59 accounted for 146 isolates (80.7%) of two major pulsotypes, CC45 for 19 isolates, ST30 for 6 isolates and ST8 (USA 300) for 4 isolates. In addition to penicillin (100%), most isolates were resistant to erythromycin (81%) and clindamycin (79.3%). Conclusions Around two-thirds of childhood community-associated S. aureus infections in northern Taiwan were MRSA. Though CC59 is still the prevalent community clone, several new clones emerged in northern Taiwan.

In Taiwan, the predominant strain of CA-MRSA is sequence type (ST) 59/Staphylococcal chromosomal cassette (SCC) mec V T / PVL-positive while HA-MRSA are ST239/SCCmec III/PVL-negative and ST 5/SCCmec II/PVL-negative in 2000s [5][6][7][8][9]. Since first report in 2002 [10], the rate of MRSA amongst childhood CA S. aureus infections increased significantly from 9.8% in 1999-2000 [10] to 55.7% in 2004-2005 [8]. However, there have been no reports published on this issue since 2005. Therefore, we conducted this study to re-evaluate if the epidemiology, clinical manifestations and molecular characteristics of childhood CA-MRSA infections changed in the past decade in Taiwan.

Methods
This study was conducted in Chang Gung Memorial Hospital (CGMH) at Linkou, which is a universityaffiliated teaching hospital in northern Taiwan and provides a range of care, from primary to tertiary care, with 3700 beds. Between 2006 and 2012, all the records of clinical S. aureus isolates, including MRSA, from children less than 19 years of age, excluding the isolates from neonatal units and survey for colonization, were extracted from the dataset of microbiology laboratory of CGMH. For the isolates identified in2008 and 2012, we retrospectively reviewed the medical records of the patients. If there were multiple episodes (isolates) collected from a single patient within the same calendar year, only the first episode (isolate) was included for analysis. We classified MRSA isolates into CA-MRSA and HA-MRSA according to the definitions proposed by Naimi et al. and HA infections were further categorized to hospital-onset (HO) and community-onset (CO). [2] Briefly, hospitalized patients infected by MRSA after 48 h of admission were classified as HO. For those identified within 48 h of admission, the patients with the risk factors including hospitalization, a permanent indwelling catheter, surgery, dialysis, and history of residence in a long-term-care facility within the previous 12 months, were categorized as CO-MRSA infection. In contrast, patients without above risk factors were regarded as community-associated (CA).
MRSA was identified according to Clinical and Laboratories Standards Institute (CLSI) guidelines [11]. Only MRSA isolates identified from CA infections in 2012 were included for further characterization. Antimicrobial susceptibility test was according to CLSI guidelines [11]. The molecular methods included pulsed-field gel electrophoresis (PFGE) with Sma I digestion, Staphylococcal cassette chromosome mec (SCCmec) typing, and detection of the Panton-Valentine leukocidin (PVL) genes. Some isolates of representative PFGE patterns were selected for further characterization by multilocus sequence typing (MLST), and spa typing. One locus difference in the MLST was categorized into the same clonal complex. All the molecular methods were described elsewhere previously [9,[12][13][14][15][16][17]. The results were analyzed by the chi-square test and statistically significance was defined as p < 0.05.

Results
During the study period from 2006 to 2012, the yearly isolate numbers of S. aureus as well as MRSA from pediatric patients, excluding those hospitalized at neonatal units, are provided in Table 1  Among the 260 MRSA isolates identified in 2012, 181 isolates (70%) were recognized as community-associated. Table 2 illustrates the distribution of MRSA stratified by the sources of specimens. CA-MRSA isolates were mainly collected from pus (86.2%) whereas HA-MRSA isolates were mostly collected from pus (48.1%) and sputum (25.3%). Among 181 patients infected by CA-MRSA, 157 (86.8%) presented with skin and soft tissue infection, 13 (7.2%) presented with urinary tract infection, 6 (3.3%) presented with pneumonia, 3 (1.7%) presented with bacteremia, and one (0.6%) each with conjunctivitis, and keratitis, respectively.
The antimicrobial susceptibility results of the 181 CA-MRSA isolates are shown in Table 3. All of the CA-MRSA isolates were susceptible to vancomycin, teicoplanin, doxycycline but resistant to penicillin. There was only one strain resistant to trimethoprim-sulfamethoxazole and fusidic acid. Most isolates were resistant to erythromycin (81%) and clindamycin (79.3%).
The molecular characteristics of 181 CA-MRSA isolates are illustrated in Table 4. There were nine pulsotypes with two major types (type D, 119 (65.7%); type C, 27 (14.9%)). Three SCCmec types were identified and SCCmec type V T accounting for 61.9% of the isolates outnumbered the others. SCCmec type IV accounted for 36.5% of the isolates.70.7% of the CA-MRSA isolates carried PVL genes. All but one isolates of pulsotype D carried PVL genes. For MLST, there were totally 6 sequence types identified. Clonal complex 59(CC59) and clonal complex 45(CC45) were the main MLST types(complexes) and accounted for 80.7% and 10.5% of the isolates, respectively. 29 CA-MRSA isolates were selected for Spa typing and 12 Spa types were identified. Table 5 shows the major clones of CA-MRSA and they are ST59/Pulsotype D/ SCCmec V T /PVL-positive, ST59/Pulsotype C/SCCmec IV/ PVL-negative and ST45/Pulsotype AK/SCCmec IV/PVLnegative.

Discussion
Comparing the results from a previous study [8], which was conducted in a similar design and definition in the  (Table 1). Though still going up since 2005, the rate seemed to reach a plateau up to 2012. These findings correlated with the rising trend of nasal MRSA colonization among pediatric population in northern Taiwan [5][6][7].
The rate of MRSA among community-acquired S. aureus infection varies markedly worldwide, ranges from <1% to >50% in different countries and is higher in children than in adults [3][4][5][6][7]. In some countries such as the United States, Taiwan, Canada, and Australia, CA-MRSA infections in patients is common, while in Europe it is low but increasing. A prospective surveillance study conducted, 2004~2006, in eight Asian countries [18] reported that MRSA accounted for 25.5% of 1463 isolates of CA S. aureus infections, and a rate > 30% was noted in Taiwan, the Philippines, Vietnam and Sri Lanka.
Consistent with those reported previously [1][2][3][4][5][6][7][8][9], most CA-MRSA isolates were identified from pus (86.2%) while the source for HA-MRSA isolates were relatively broader, including sputum (25.3%), pus (48.1%), urine (7.6%), blood 5.1%) and CVP tips (5.1%). Furthermore, among HA-MRSA isolates, CO-MRSA isolates were more likely to be identified from pus (62%) while HO-MRSA isolates were more likely to be identified from sputum (41.4%) and only 24.1% from pus. This origin profiles again indicated that CA-MRSA infections usually presented with skin and soft tissue infections whereas HA-MRSA infections often presented with a diverse spectrum of disease entities. In this study, the clone characterized asST59/SCCmec V T /PVL-positive, named as Taiwan clone, was the predominant clone of CA-MRSA isolates, which is consistent with previous study [8]. The dominance of this clone among CA-MRSA isolates in Taiwan persisted for more than one decade since identified [5][6][7][8][9][19][20][21]. In contrast, sequence type ST239/SCCmec III/ PVL-negative, which was the endemic HA clone and accounted for nearly 5% of all CA-MRSA isolates in the previous study [8], was not identified in this study (Table 5). However, several clones which were rarely reported in Taiwan previously were identified in this study and included ST8/t008/SCCmec IV/PVL-positive (USA 300), clonal complex 45/ SCCmec IV/ PVL-negative and ST 30/t019/ SCCmec IV/ PVL-positive. ST 30 is the major CA-MRSA clone prevailing in southeastern Asian countries, Australia and Japan [6,7]. ST8 (USA 300) is the major CA-MRSA clone prevailing in Canada and USA and has been reported from some Asian countries recently [6,7]. Reviewing the medical records of the patients infected with both clones, we did not identify any obvious travel history or contact history. The issues regarding how did these clones appear and evolve in Taiwan and what is the impact of these clones in Taiwan need further observations. Also, continuing surveillance is needed.
In this study, around 80% of CA-MRSA isolates were resistant to erythromycin and clindamycin. The resistant rates to both antibiotics, though still high, seemed to decrease gradually, compared with those in 2004-2005 (Table 3). Likewise, all the isolates were still susceptible to vancomycin, teicoplanin, linezolid and doxycycline and all but one isolates were susceptible to trimethoprim/sulfamethoxazole and fusidic acid. The antibiotic resistant pattern is correlated with the molecular pattern of the CA-MRSA isolates in different years [5][6][7][8][9].
There are several limitations in this study. First, this study was conducted in a single medical center and the epidemiologic features shown here may not represent the whole perspective in Taiwan. However, our hospital is the largest hospital in Taiwan and the case number in this study was not small, so it still can partly reflect the current status of childhood CA-MRSA in Taiwan. Second, though the isolates were prospectively collected, medical records of the patients were retrospectively reviewed, so some risk factors for MRSA acquisition in the patients may be missed and thus the patients in HA-MRSA group might be misclassified to CA-MRSA group. However, no genetically HA-MRSA isolate was identified from the patients in CA-MRSA group. Third, some of the specimen types e.g. pus, wound swabs and All the isolates were resistant to penicillin while susceptible to tigecycline, linezolid, vancomycin and teicoplanin *p = 0.003 **p = 0.01 ***p = 0.024  Values are given as n (%) PFGE pulsed-field gel electrophoresis, SCCmec staphylococcal chromosomal cassette, PVL Panton-Valentine leukocidin, MLST multilocus sequence typing a A significant difference was found between the community-associated and healthcare-associated isolates in respect to PFGE D, PFGE A, and PFGE F clones (p < 0.001) b A significant difference was found between the community-onset and hospital-onset isolates in terms of the third clone (p = 0.022)