Staphylococcus aureus is one of the pathogens involved most frequently in bacterial infections in children. These infections are of variable severity, ranging from mild forms like superficial cutaneous infections and soft tissue abscesses to severe forms of invasive infection (necrotising pneumonia, pyomyositis, osteomyelitis, sepsis) that carry a high morbidity and may be life-threatening1.

In the late 1990s, there was a marked increase in the prevalence of methicillin-resistant S. aureus (MRSA), reflecting the emergence of resistant strains that spread both inside and outside hospitals, and both nosocomial and community-acquired MRSA infections have since become frequent. Some of the studies in the literature have noted that the increased prevalence of MRSA may be particularly important in the paediatric population2–4.

In Europe, the knowledge of the prevalence of methicillin-resistant Staphylococcus infections throughout the continent is obtained through the European Antimicrobial Resistance Surveillance Network (EARSnet), which started operations in year 19995. This network collects date on episodes of bacteraemia caused by different bacterial species, including S. aureus. Based on the 2019 EARSnet report, the percentage of bacteraemia isolates corresponding to MRSA was 19.2%, which, while lower compared to the 2015 report (25%), continued to be above the European mean (15.5%). The prevalence of MRSA is greater in older individuals (age >65 years: 22.8%), but it is still considerable in the paediatric population (age 0–4 years: 12%; age 5–18 years: 15.7%)6.

The EARSnet focuses on the prevalence of bacteraemia by different species and strains, including MRSA; but there is no system at the European level for epidemiological surveillance of MRSA isolates from samples other than blood.

In Spain, 6 studies have been devoted to the prevalence of Staphylococcus species since 1986 (1986, 1991, 1994, 1996, 2002, 2006) through the analysis in the same single laboratory of all Staphylococcus strains isolated from different types of clinical samples in a single day in every participating hospital, with representation of every region in the country7,8. These studies evinced an increase in the prevalence of MRSA from 1.5% in year 1986 to 31.2% in 2002, followed by stabilization at 29.2% in 2006. The studies included both community-acquired and nosocomial infections.

Data regarding the evolution of antimicrobial drug resistance in bacteria has a substantial influence on the development of strategies for the selection of empirical antibiotherapy. Although the aforementioned nationwide prevalence studies could reflect temporal trends in the prevalence of MRSA, the study universe was the adult population, and therefore the results cannot be extrapolated to the paediatric population.

As regards community-acquired infection by MRSA in children in Spain, the first case series was published in 20069. Later on, other series have been published in different geographical regions. The described infections mainly involved the skin and soft tissues, some were severe (cases of pneumonia and necrotising fasciitis, bacteraemia)10–15. In Spain, only one multicentre study, published in 2009, has assessed the prevalence of MRSA in the paediatric population. This study included all infections by S. aureus in children managed in 4 Spanish hospitals (Barcelona, Madrid, Palma de Mallorca and A Coruña), and found a proportion of community-acquired MRSA infection of 8.8%16.

The primary objective of our study was to determine the proportion of cultures positive for MRSA among the patients managed in paediatric emergency departments (PEDs) in Spain. The secondary objectives were to identify risk factors associated with an increased probability of MRSA isolation and to analyse the morbidity associated with these infections.

We conducted a multicentre retrospective study in 8 PEDs included in the Working Group on Infectious Diseases of the Sociedad Española de Urgencias de Pediatría (SEUP, Spanish Society of Paediatric Emergency Care) (Hospital Infantil Universitario Niño Jesús in Madrid, Hospital Sant Joan de Déu in Barcelona, Corporació Sanitária Parc Taulí in Sabadell, Hospital de Mataró, Hospital de Cabueñes in Gijon, Hospital Virgen del Rocío in Seville, Hospital Río Hortega in Valladolid and Hospital Son Espases in Mallorca) of patients managed between 1/07/2017 and 30/06/2018 in whom S. aureus had been isolated from culture from samples obtained from any site.

The identification of S. aureus isolates was performed at the microbiology laboratories of each hospital. Later on, we reviewed the health records of patients to select those that met the inclusion criteria.

In the period under study, S. aureus was isolated from 403 samples. Table 1 presents the number of episodes with isolation of S. aureus reported by each PED and by autonomous community in Spain.

The mean age of patients was 75.8 months (standard deviation, 59.2), 54.8% were male. Comorbidities were present in 59 children (14.6%), the most frequent of which were: altered immunity (immunodeficiency, steroid therapy at a dose ≥2 mg/kg/day or immunosuppression) (n = 8), severe atopic dermatitis (n = 6), congenital heart disease (n = 6), cancer (n = 6; 3 of these patients also had indwelling central venous catheters); neurogenic bladder (n = 4), cystic fibrosis (n = 3) and preterm birth (n = 3).

Thirty-six patients (8.9%) were using invasive medical devices: 16 osteosynthesis material, 7 gastrostomy tubes; 5 central venous catheters, 2 urinary catheters, 2 ventriculoperitoneal shunts, 1 a urinary catheter and a gastrostomy tube, 1 a cochlear implant and 2 cardiac stents.

The overall proportion of MRSA was 16.6% (67/403; 95% CI, 13−20.2%); Table 1 presents the proportions of MRSA isolates by autonomous community.

Out of the 403 staphylococcus infections, 116 (28.8%) were hospital-associated, and 287 (71.2%) community-associated, with a respective prevalence of MRSA of 18.1% and 16.2% (P > .05).

We proceed to present the results on the anatomical site of the samples, the potential risk factors for methicillin resistance and outcomes of patients with isolation of methicillin-sensitive and methicillin-resistant S. aureus.

Table 2 presents the anatomical sites of the samples from which S. aureus was isolated and the proportion of MRSA isolation by site. The proportion of MRSA isolates was higher in samples obtained from cutaneous abscesses compared to samples obtained from blood, conjunctiva, skin or the ear (P < .05). Fifty percent of reported S. aureus blood culture isolates were associated with osteoarticular infections. The overall prevalence of MRSA in osteoarticular infections was 7.7% (2/16 blood cultures, 0/10 joint drainage cultures).

The proportion of infections caused by MRSA in children born in Spain was 13.3% (95% CI, 9.4−17.2) versus 52% (95% CI, 32−72) in those born outside Spain (P < .01). Methicillin-resistant S. aureus grew in 90% (95% CI, 71.4–108.6) of cultures of patients with a previous history of MRSA infection, compared to 13% (95% CI, 9.6−16.4) of patients without that history (P < .01). We found a greater probability of methicillin resistance in foreign-born children (OR, 7; 95% CI, 3%–16.6 %) and in children with a previous MRSA infection (OR, 59.6; 95% CI, 7.4%–481.2 %). Table 3 presents the comparative analysis of MRSA isolation based on the presence or absence of each of the risk factors under study.

Antibiotherapy was administered to 369 children (91.5%) before culture results were available, in 215 (58.3%) amoxicillin-clavulanic acid and in 20 (4.9%) an antibiotic that covered MRSA (vancomycin, clindamycin or cotrimoxazole).

When it came to patient outcomes in the total sample (403), 167 (41.4%) were admitted to hospital, and 12 (3%) developed complications: poor outcome with need of surgical reintervention in 4 children with postoperative wound infection (with removal of osteosynthesis material in 2 of them), 2 with cutaneous abscesses, 1 with osteomyelitis, 1 with superficial venous thrombosis, 1 with cutaneous necrosis secondary to cellulitis that required surgery, 1 with sepsis in the context of necrotising fasciitis, 1 with osteomyelitis following a cutaneous abscess and 1 with osteomyelitis that progressed with extensive involvement of the femur. The proportion of complications in patients was 3.4% (10/296) in cases in which S. aureus was sensitive to the antibiotic prescribed in the PED and 2.8% (2/73) in the cases in which it was resistant (P > .05).

There were sequelae in 4 children in the sample: genu flexum following osteoarticular infection (n = 2), bone erosion in joint margin (n = 1) and chronic otorrhoea with conductive hearing loss following suppurative otitis media (n = 1). There were no deaths directly associated with the staphylococcus infection. Table 4 presents the comparison of outcomes (admission, complications, sequelae) based on methicillin sensibility.

This is the first multicentre study conducted in Spain that analyses the prevalence of MRSA in cultures of samples collected from paediatric patients managed in PEDs. We found an overall proportion of methicillin resistance of 16.6%, practically double the proportion of MRSA in community-acquired S. aureus infections (8.8%) found in the only multicentre study ever conducted in Spain in the paediatric population, in 200916, although the latter did not only include cultures of samples collected in emergency departments or hospital-associated infections. The MRSA prevalence data reported for Spain by the EARSnet in the 2017 report broken by paediatric age group, which ranged from 14.9% (0–5 years) to 16.4% (5–18 years)20, could be extrapolated to ours. Although the EARSnet only analyses isolates from blood stream infections, this is the only source at the European level that is available to follow temporal trends in the prevalence of methicillin resistance in Europe overall and in each participating country.

Several publications have alerted of the spread of MRSA strains in outpatient and community settings in Spain21–23, a phenomenon that extends to the paediatric population9,16 and corroborated by our study, as we did not find statistically significant differences in the proportion of MRSA between hospital-associated S. aureus infections (18.1%) and community-associated infections (16.2%). In Spain, the COSACO study,24 conducted recently, has contributed relevant and interesting knowledge regarding nasal colonization by MRSA in the Spanish paediatric population. The authors estimated the prevalence of nasal colonization by MRSA at 1.4%, which they considered high compared to the prevalence in European children and adults25–28. The prevalence of nasal colonization by MRSA, a known risk factor for future infection24,29, and the proportion of MRSA out of all S. aureus isolates in the PED setting are not directly comparable. Still, the detection of a high prevalence of nasal carriage of MRSA is consistent with the frequency of methicillin resistance in community-associated staphylococcus infections found in the paediatric population.

The prevalence of methicillin resistance found in our study did not vary significantly between participating autonomous communities, although the sample did not include patients from all autonomous communities and the representation of those that participated was also not optimal, which limits the extrapolation of our results.

Infections by S. aureus most frequently involve skin and subcutaneous tissue7,30,31. In countries with a high prevalence of MRSA, this has become one of the main causal agents involved in these community-acquired infections29,32–36. At the same time, skin and soft tissue infections are the most frequent form of community-acquired MRSA infection29, so it is not surprising that we found the highest prevalence of MRSA in these infections in our sample, with 30% of MRSA isolates corresponding to samples obtained from suppurative cutaneous lesions. This proportion of methicillin resistance was significantly higher compared to the proportion in skin samples in cases of infection that did not require drainage (10.2%). However, we cannot infer that methicillin resistance itself is the factor that promotes suppuration. In fact, there is evidence of an association with the presence of Panton-Valentine leukocidin, independently of the sensitivity or resistance to methicillin12,14,31. In addition, this factor did not affect patient outcomes in infections caused by MRSA, in whom the frequency of complications or sequelae were not higher (Table 4).

The high prevalence of MRSA detected in cutaneous abscess cases does not call for a change in clinical management as regards initiation of empirical antibiotherapy (an antibiotic that covered MRSA was only given to 5% of the patients) but it does call for performance of drainage as early as possible and the use of antibiotics that cover MRSA if the patient is not improving as expected.

On the other hand, invasive infections account for a minority of the total community-associated MRSA infections29. The 12.5% prevalence of MRSA found in blood cultures of peripheral blood draws in our study (4/32) stood out in relation to the proportions reported in previous studies in Spain: 0% (0/57) in the prospective study conducted in the 2007–2010 period in patients aged less than 16 years by Barrado et al in a tertiary care hospital in Madrid37, and 3% in the 2010–2012 period in the retrospective study conducted in Barcelona in patients aged less than 16 years by Cobos-Carrascosa et al.15 The prevalence of MRSA isolated from blood culture in our study (12.5%) was, however, consistent with the frequencies by age group reported by the EARSnet for Spain in 2019, which can be consulted online in the Surveillance Atlas of Infectious Disease (an open-access resource of the European Centre for Disease Prevention and Control available at http://atlas.ecdc.europa.eu/public/index.aspx): prevalence of bacteraemia caused by MRSA of 12% in the 0–4 years age group and 15.7% in the 5–18 years group. Although the prevalence of MRSA isolation from blood culture in our sample was high, this was not associated with poorer outcomes or a higher incidence of complications or sequelae (Table 4).

As regards the risk factors for methicillin resistance, we only found an association with the origin of the patients (higher prevalence in children born outside Spain) and the history of previous isolation of MRSA.

In our study, we found a higher proportion of MRSA in children who were not native Spaniards (52% in children of other nationalities vs 13.3% in Spanish children; P = .000), which was not unexpected, as studies conducted by the European Network for the Surveillance of Imported S. aureus have found a high frequency of skin and soft-tissue infections by S. aureus in travellers that return to Europe (67%), with a documented prevalence of methicillin resistance of 14% in the isolates of these imported cases17. The probability of methicillin resistance in imported S. aureus cases varied between regions and was highest for Latin America17. In our study, we did not analyse the region of origin in children of foreign nationality.

As for the higher prevalence of methicillin resistance found in our study in patients with a previous history of MRSA isolation (90% compared to 13.3% in children without previous MRSA isolation), we ought to note that this is not a surprising finding, as previous MRSA isolation is a thoroughly-documented risk factor for methicillin resistance19. Therefore, it is important to always explore this history in patients that present to the PED with a suspected staphylococcal infection.

The staphylococcus infections managed in PEDs, independently of methicillin susceptibility or resistance, were severe: nearly half the patients required hospital admission and 1 in 30 developed complications. Although methicillin resistance does not make S. aureus more virulent, an aspect that rather appears to be associated with the presence of the Panton-Valentine leukocidin, as noted above12,14,31,38, it is important to document and monitor the prevalence of MRSA in children to know which antibiotics offer the best treatment option in the case of infection of a likely staphylococcal aetiology.

The overall prevalence of MRSA in cultures of samples obtained in PEDs in Spain was significant: 1 in every 6 isolates. The highest frequencies of MRSA corresponded to children born outside of Spain, children with a previous history of MRSA isolation and samples obtained from cutaneous abscess and subcutaneous tissue. In the case of suppurative lesions of skin and soft tissues, it is important to perform early drainage and to consider switching to an antibiotic that covers MRSA if treatment does not achieve an adequate outcome.

Prospective studies in larger samples and with representation of more regions in Spain are required to more accurately estimate the prevalence of MRSA in cultures of samples obtained in Spanish PEDs.

The authors have no conflicts of interest to declare.