Staphylococcus aureus remains the most common cause of invasive bacterial infections in children. Reference Cassat and Thomsen1 Hospital-onset methicillin-resistant S. aureus (MRSA) rates have decreased in the United States, whereas community-onset MRSA infections have decreased to a lesser extent. Reference Kourtis, Hatfield and Baggs2,Reference Klein, Sun, Smith and Laxminarayan3 Although most epidemiologic studies have focused on adults, MRSA rates appear to be declining in children as well. Reference Sutter, Milburn, Chukwuma, Dzialowy, Maranich and Hospenthal4,Reference Khamash, Voskertchian, Tamma, Akinboyo, Carroll and Milstone5 In the United States, pediatric hospitalizations due to S. aureus decreased by 35% between 2009 and 2016, Reference Spaulding, Thurm and Courter6 and pediatric outpatient visits for purulent skin and soft-tissue infections (SSTIs) decreased between 2011 and 2015. Reference Fritz, Shapiro and Hersh7 Although the decrease in pediatric infections due to S. aureus is encouraging, the epidemiology of pediatric S. aureus antibiotic resistance in the community remains incomplete in the United States. Several studies have shown that MRSA susceptibility to clindamycin declined in the 2010s, Reference Sutter, Milburn, Chukwuma, Dzialowy, Maranich and Hospenthal4,Reference Khamash, Voskertchian, Tamma, Akinboyo, Carroll and Milstone5 and concurrently, trimethoprim-sulfamethoxazole prescriptions for SSTIs have increased nationally. Reference Fritz, Shapiro and Hersh7 Pediatric S. aureus infections continue to be a public health problem for children, and clinicians require contemporary S. aureus susceptibility data to guide antibiotic decision making. Most pediatric S. aureus antibiotic data available describe hospital-based trends and antibiotic susceptibilities. Here, we describe recent trends in S. aureus susceptibilities within a multihospital health system with a focus on pediatric community-onset infections and infections managed in the outpatient setting.
Methods
We performed a multihospital, retrospective study of temporal trends in S. aureus antibiotic susceptibility. Using electronic medical record (EMR) data (Epic Systems, Verona, WI), we identified bacterial cultures growing S. aureus obtained from patients aged <18 years between January 1, 2015, and December 31, 2020, from the Johns Hopkins Hospital (JHH) microbiology laboratory. The JHH microbiology laboratory services an urban academic medical center, an urban midlevel teaching hospital, a suburban community hospital, and hospital-affiliated outpatient clinics. We only included the first clinical culture per patient per year. We excluded cultures that likely represented colonization, such as nasal swabs or throat swabs. Invasive infections were defined as positive blood, bone, joint, and central nervous system cultures. Respiratory cultures included tracheal, sputum, bronchoalveolar lavage, pulmonary tissue, and pulmonary abscess cultures. Soft-tissue cultures were defined as skin, abscess, vesicle, and nonsurgical wound cultures. Community-onset infections were defined as cultures obtained within 3 calendar days of hospitalization or from children who were not hospitalized. Infections managed in the outpatient setting were cultures from patients with community-onset infections who were not admitted to a hospital. The Cochran-Armitage test was used to analyze trends in antibiotic susceptibilities. P < .05 was considered statistically significant. Data were analyzed using R software (R Center for Statistical Computing, Vienna, Austria). This study was approved by the Johns Hopkins University Institutional Review Board.
Results
In total, 2,220 patients and 2,387 cultures were identified and met eligibility criteria. Children with a S. aureus culture had a mean age of 6.9 years (SD, ±5.7); 53.0% were male; and 36.6% were white (Table 1). The number of overall clinical cultures per year was stable during 2015–2019 and ranged from 395 to 456. Notably, a 40% decrease occurred in the number of cultures in 2020 (n = 237) compared to 2019 (n = 395). Overall, 1,508 cultures (63.2%) were from soft-tissue sources, and 1,352 (56.6%) were managed in the outpatient setting.
Table 1. Characteristics of Pediatric Patients and Cultures Positive for Community-Onset Staphylococcus aureus Infections, 2015–2020
Note. SD, standard deviation.
Of the 2,387 community-onset S. aureus cultures, 1607 (67.3%) were oxacillin susceptible, and oxacillin susceptibility was 71.5% in 2015 and 70.8% in 2020 (P = .71) (Fig. 1). Within all community-onset cultures, clindamycin susceptibility was stable at ∼75% (P = .27). Trimethoprim-sulfamethoxazole and tetracycline susceptibility remained >90% (P = .21 and P = .50). In cultures of community-onset infections managed in the outpatient setting, oxacillin susceptibility was 70.4% in 2015 and 64.4% in 2020 (P = .38). Oxacillin susceptibility was lowest in SSTIs, with 67.1% oxacillin susceptibility in 2015 and 65.2% in 2020 (P = .75) (Supplementary Fig. 1). Invasive cultures were 75.5% oxacillin susceptible in 2015 and 84% susceptible in 2020 (P = .90).
Fig. 1. Percentage with antibiotic susceptibility with 95% confidence intervals during the study period in (A) community-onset S. aureus cultures; (B) community-onset, outpatient-managed S. aureus cultures, (C) community-onset MSSA cultures; and (D) community-onset MRSA cultures.
Within MRSA cultures, tetracycline susceptibility remained ≥90% (P = .15). Trimethoprim-sulfamethoxazole susceptibility did not change between 2015 and 2020 (P = .90); however, trimethoprim-sulfamethoxazole susceptibility decreased from 90.3% to 84% between 2015 and 2018 and then increased to 99% in 2020. Clindamycin susceptibility in MRSA clinical cultures remained ∼75% throughout the study period (P = .92). Among MSSA cultures, trimethoprim-sulfamethoxazole and tetracycline susceptibility remained ≥90% (P = .84 and P = .07). Clindamycin susceptibility remained ∼75% in MSSA cultures during the study period (P = .21).
Discussion
Within this multihospital health system, MRSA continues to cause 30% of community-onset infections among pediatric patients. MRSA prevalence is particularly high among patients managed in the outpatient setting and those who present with SSTIs. Overall, 84% of invasive infections were MSSA. The stability in methicillin resistance among S. aureus cultures of community-onset infections differs from other epidemiologic studies that have reported decreases in methicillin resistance in pediatric S. aureus infections. Reference Sutter, Milburn, Chukwuma, Dzialowy, Maranich and Hospenthal4–Reference Spaulding, Thurm and Courter6 Most prior studies analyzed trends in hospitalized children, which may account for this difference and would align with our finding of MSSA predominance in invasive infections. The cause of national increases in invasive MSSA is unclear. Studies describing a rise of MSSA within invasive pediatric S. aureus infections have reported differing molecular characteristics, with the lukSF-PV gene detected in 6% Reference Crandall, Kapusta and Killpack8 –30% Reference Hulten, Mason, Lamberth, Forbes, Revell and Kaplan9 of community-onset invasive MSSA isolates. The increase in invasive MSSA infections nationally may be due to the rise of a virulent methicillin-susceptible strain versus community immunity to USA300. Reference Hulten, Mason, Lamberth, Forbes, Revell and Kaplan9 Further immunologic and molecular epidemiology studies are required to understand national trends in invasive MSSA infections.
During the study period, clindamycin susceptibility to MRSA remained relatively low and trimethoprim-sulfamethoxazole and tetracycline susceptibility remained high. During 2015–2019, trimethoprim-sulfamethoxazole susceptibility fell, particularly among MRSA cultures. Surveillance of S. aureus susceptibilities within our health system previously showed a recent decrease in trimethoprim-sulfamethoxazole susceptibility as well. Reference Khamash, Voskertchian, Tamma, Akinboyo, Carroll and Milstone5 Reports have since emerged suggesting that a commercial platform for antibiotic susceptibility testing may overestimate trimethoprim-sulfamethoxazole resistance compared to disc-diffusion testing. Reference Al-Rawahi, Chorlton, Dhaliwal, Golding and Tilley10 In reponse to these reports, the Johns Hopkins Hospital microbiology laboratory transitioned to trimethoprim-sulfamethoxazole disc-diffusion susceptibility testing for S. aureus isolates in 2019. Our results show an increase in trimethoprim-sulfamethoxazole susceptibility in 2019–2020. This increase in susceptibility may be due to the concurrent change in microbiology susceptibility testing. Due to increased clindamycin resistance, there has been a national decline in clindamycin prescriptions and an increase in trimethoprim-sulfamethoxazole prescriptions for SSTIs. Reference Fritz, Shapiro and Hersh7 Our results demonstrate that trimethoprim-sulfamethoxazole susceptibility has remained high despite a national increase in prescriptions; however, this trend requires close monitoring.
This study had several limitations. Culture collection sources were defined by provider-entered orders and may not represent the true source of infection. These data from a multihospital health system may not reflect trends in other geographic regions. Despite these limitations, this study confirms that MRSA continues to cause significant disease burden among pediatric patients within a multihospital health system, particularly those managed in the outpatient setting and those with SSTIs. The divergence between MRSA and MSSA disease severity is consistent with other studies and warrants increased surveillance of community-onset S. aureus infections to identify emerging strains and prevalent virulence factors.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/ash.2022.370
Acknowledgments
Danielle Koontz, MA, MS, contributed to the study design and data collection.
Financial support
This work was supported in part by the National Institutes of Health (grant nos. T32 AI052071 to E.P. and K24 AI141580 to A.M.).
Conflicts of interest
All authors report no conflicts of interest relevant to this article.
Open access
