Declining prevalence of Methicillin-resistant Staphylococcus aureus septic arthritis and osteomyelitis in children: implications for treatment

Abstract

OBJECTIVES To assess clinical characteristics and differences in outcomes between children with Methicillin-resistant Staphylococcus aureus (MRSA) and Methicillin-sensitive Staphylococcus aureus (MSSA) osteomyelitis or septic arthritis and whether initial antibiotic regimen affects patient outcomes.

STUDY DESIGN We analyzed records of children ages 15 days to 18 years admitted between 2009 and 2016 to two tertiary children’s hospitals who were diagnosed with an osteoarticular infection and had a microorganism identified. We compared the relationship of MRSA versus MSSA and initial antibiotic selection on patient outcomes including length of hospitalization and requiring a PICU transfer.

RESULTS A total of 584 patients met inclusion criteria of which 365 (62.5%) had a microbiological diagnosis. MSSA was the most common pathogen identified (45.5%) followed by MRSA (31.2%). The percentage of MRSA cases decreased by 25.4% between 2009 and 2016, whereas MSSA infections increased by 18.6%. Compared to MSSA, patients with MRSA had a higher initial C-reactive protein (median 17.3 mg/dL vs 7.8 mg/dL, p<0.05) and a longer hospitalization (median 8.3 days vs 6.1 days, p<0.05). Patients whose initial antibiotic regimens included vancomycin were more likely to require a PICU transfer (15.0% vs 2.2%, p<0.05) and had a longer hospitalization (median 6.7 days vs 5.5 days, p<0.05) compared to those initiated on clindamycin without vancomycin.

CONCLUSIONS While MRSA was associated with increased severity of osteoarticular infections compared to MSSA, the incidence of MRSA has been declining at our institution. Patients with longer lengths of stay and those requiring a PICU transfer were more likely to be on vancomycin. Clindamycin should be considered in the initial antibiotic regimen for osteomyelitis and septic arthritis with ongoing surveillance of local microbiology and outcomes.

introduction

Acute hematogenous osteomyelitis (AHO) and septic arthritis (SA) are frequent causes of hospitalization in the pediatric population and can have significant morbidity if not quickly recognized and treated appropriately.(1–3) The bacteria most commonly isolated is Staphylococcus aureus, which accounts for greater than 50% of culture positive cases.(1, 4–6)

Over the past two decades, musculoskeletal infections caused by Methicillin-resistant Staphylococcus aureus (MRSA) have been increasing, whereas the rate of Methicillin-sensitive Staphylococcus aureus (MSSA) infections has generally remained steady.(7–10) MRSA osteoarticular infections are frequently associated with higher inflammatory markers,(8, 11, 12) a higher complication rate,(7, 8, 13) longer lengths of hospitalization, (8, 12, 13) and higher associated hospital costs.(13) However, more recently studies have shown a decrease in the proportion of S. aureus infectionscaused by MRSA.(14–17)

The changing epidemiology of these infections in some areas may necessitate a change in empiric antibiotic therapies. The Infectious Diseases Society of America recommends intravenous vancomycin for the treatment of children with acute hematogenous MRSA osteomyelitis and septic arthritis; however they mention that clindamycin can be used as empiric therapy if the clindamycin resistance rate is low (e.g. <10%) and the patient is stable and does not have evidence of ongoing bacteremia or intravascular infection.(18) In a review by Harik and Smeltzer, the authors recommend use of clindamycin if community-acquired rates of MRSA are >10% and clindamycin resistance is ≤25%.(19) McBride et al. found that in children with AHO, hospital readmissions rates and 72-hour improvement in C-reactive protein (CRP) values were similar between patients who received early treatment with MRSA coverage compared to MSSA-only coverage. However, they did not examine differences between initial MRSA drug choices (e.g. clindamycin versus vancomycin) on patient outcomes.(20) Few studies have examined clinical outcomes based on empiric therapy for AHO or SA.

This study aimed to describe the microbiological and clinical characteristics of AHO and SA in a large cohort of children and to compare outcomes between MSSA and MRSA. A secondary aim was to examine the outcomes of children treated initially with vancomycin containing regimens compared to those treated with clindamycin without vancomycin.

methods

Study Design and Participants

A large retrospective cohort study was performed on children ages 15 days to 18 years admitted to two tertiary care children’s hospitals affiliated with Children’s Healthcare of Atlanta between January 1, 2009 and December 31, 2016 with a diagnosis of SA and/or AHO. Details on the overall study design and methodology have been previously reported.(21) In brief, patients were identified through a retrospective search of electronic medical records using International Classification of Diseases, Ninth Revision (ICD–9) and International Classification of Diseases, Tenth Revision (ICD–10) diagnostic codes for osteomyelitis and septic arthritis. The medical records of all patients who met initial inclusion criteria were then manually reviewed. We excluded the following: any patient who was not treated for osteomyelitis and/or septic arthritis, had recent trauma including fractures and foreign body, a post-operative infection or hardware at the site of the infection, had chronic symptoms (>6 weeks in duration) or diagnosed with chronic recurrent multifocal osteomyelitis, were immunocompromised, were a neonate who had never left the hospital, or if they had the majority of their workup performed at an outside hospital. The study protocol was approved by the Children’s Healthcare of Atlanta Institutional Review Board.

Data Collected

Demographic features, type of infection (osteomyelitis, septic arthritis or both), duration of symptoms at presentation, initial white blood cell count, inflammatory markers, fever at presentation, location of infection, microbiological results, and initial antibiotic regimen were retrospectively reviewed and analyzed.

For our primary variable of interest, we analyzed microbiological data from cultures to determine if patients were infected by MRSA, MSSA, other organisms (non-S. aureus),, or if no pathogen was able to be identified. We focused on the differences between MRSA and MSSA since these are the most commonly identified organisms. Positive cultures were identified from samples collected from blood, synovial fluid, bone or wounds/abscesses. Our primary outcomes of interest were length of hospitalization and readmission within 60 days for a problem associated with initial hospitalization for AHO or SA. A secondary outcome of interest was patients requiring a PICU transfer and those found to have a subperiosteal abscess. We also examined clindamycin sensitivity rates for all patients who had an infection with S. Aureus. Clindamycin susceptibility was defined as standard laboratory susceptibility as well as the absence of inducible resistance to clindamycin when this was tested. Isolates with intermediate susceptibility to clindamycin were considered resistant in our study.

A tertiary variable of interest was comparing patients who were initially placed on an empiric antibiotic regimen containing clindamycin (without vancomycin) compared to those who were placed on a regimen containing vancomycin. Clinical features and patient outcomes were then compared for these two antibiotic groups. An exploratory analysis was also done looking at the effect of initial antibiotic discordance for clindamycin-resistant and clindamycin-sensitive S. aureus on complication rate and length of hospitalization. Antibiotic discordance was defined as if a patient’s initial antibiotic regimen did not include another antibiotic to which the organism was sensitive to.

Statistical Analysis

Statistical Analysis was performed using SAS version 9.4 (SAS Institute, Inc, Cary, NC). Descriptive statistics were used to describe the cohorts using medians and interquartile ranges (IQR) for continuous variables and frequencies and percentages for categorical variables. Categorical variables were compared using Chi-squared tests or Fisher’s Exact tests if cell counts were <5. Continuous variables were compared using non-parametric Wilcoxon rank-sums tests. Demographic and clinical characteristics were compared based on (1) pathogen isolated (MRSA versus MSSA), (2) patients who had clindamycin-sensitive S. aureus and those with clindamycin-resistant S. aureus and, (3) stratified by initial antibiotic regimen (those treated with clindamycin without addition of vancomycin versus those treated with regimens that included vancomycin). The length of stay outcome was log-transformed due to non-normality in residuals and presented as a ratio of geometric least-squares (LS) means with 95% confidence intervals. Logistic regression models were run for PICU transfer and readmission within 60 days. Results were presented as odds ratios (OR) with 95% confidence intervals. One-sided Cochran-Armitage trend tests were also performed to compare proportions of MSSA and MRSA isolates during 2009–2016 and to determine changes in clindamycin antibiotic susceptibility patterns. A p-value of <0.05 was considered statistically significant. 

Results

Microbiology of Infections:

A total of 584 patients met inclusion criteria of which a pathogen was isolated in 365 patients (62.5%). The most common pathogen identified was MSSA (45.5%) followed in decreasing frequency by MRSA (31.2%), Group A Streptococci (6.6%), Streptococcus Pneumoniae (3.8%), Group B Streptococci3.0%), other streptococcal species (2.7%) and Kingella Kingae (2.2%). The remaining 5% of isolated pathogens individually accounted for <1.5% of culture positive cases and included Salmonella species, Coagulase negative Staphylococcus, Hemophilus influenza, Enterobacter aerogenes, Fusobacterium species, Bacteroides fragilis, Corynebacterium striatum, Pseudomonas aeruginosa, Citrobacter freundii, Propionibacterium acnes, Neisseria flavescens and Candida parapsilosis. S. aureus was found in 60.5% of patients with AHO, but only 17.6% of patients with SA (p<0.05). Figure 1 displays percentages of MSSA and MRSA culture positive infections from 2009 to 2016. During the 8-year study period, the percentage of culture positive cases due to MSSA increased from 35.9% in 2009 to 54.5% in 2016 (z = +1.65, p<0.05). The reverse trend was seen for MRSA infections; 43.6% of culture positive cases in 2009 were caused by MRSA down to 18.2% of culture positive cases in 2016 (Z = –2.29, p<0.05).

The characteristics of patients with MRSA versus MSSA are shown in Table 1. Patients with MRSA were younger, more likely to be African American, more likely to have a concomitant AHO and SA, and had a significantly higher initial CRP, erythrocyte sedimentation rate (ESR) and WBC count (p<0.05). Patients with MRSA had a longer length of hospitalization, were more likely to require a PICU transfer and were more likely to have a subperiosteal abscess compared to patients with MSSA (p<0.05). However, after adjusting for potential confounders (age, race, infection type, and CRP), compared to patients with MSSA, MRSA patients had a longer length of stay, but there were no statistically significant differences in rates of PICU transfers or subperiosteal abscess between these two groups (See Table 2). Table 3 shows adjusted outcomes between MRSA, MSSA, other organisms and no pathogen identified. The odds of having a subperiosteal abscess were 6.4 times higher for MRSA and 5.1 times for MSSA compared to patients that had no pathogen identified (p<0.05). Of the 29 patients readmitted within 60 days, there were no significant differences in readmission rate based on microorganism.

Antimicrobial Regimen and Susceptibility:

In terms of antibiotic susceptibility, 91.2% (103/113) of MRSA isolates, 85.9% (140/163) of MSSA isolates, and 88.0% (243/276) of S. Aureus isolates combined were sensitive to clindamycin. There was one case of MRSA and three cases of MSSA that did not have sensitivity to clindamycin reported. Of the 33 cases of clindamycin-resistant S. aureus, 63.6% (21/33) exhibited inducible resistance. There was no significant difference in the rates of inducible clindamycin resistance of MSSA compared to MRSA (p = 0.10). During the study period there was a trend of decreased clindamycin sensitivity for MRSA over time (z = –1.94, p = 0.03), but no significant trend was seen for clindamycin sensitivity of MSSA (z = +0.64, p = 0.26) or for overall rates of clindamycin sensitive S. aureus by year (z = –0.70, p = 0.24).

Empiric treatment in our study most commonly included clindamycin (57.4%), ceftriaxone (51.9%), and/or vancomycin (42.1%). Clindamycin and vancomycin were used together in 56 patients (9.6%). Table 4 shows patient characteristics and outcomes based on initial antibiotic regimen. The adjusted analysis suggests that patients on clindamycin (without vancomycin) had a 16% shorter length of stay compared to patients on vancomycin containing regimens, and those on clindamycin were 78% less likely to require a PICU transfer even after adjusting for type of infection, CRP, days of symptoms at presentation, presence of MRSA, presence of subperiosteal abscess/deep vein thrombosis/and/or pyomyositis, and those who received surgical drainage (p<.05) (See Table 5). There were no significant differences in patients with clindamycin-resistant S. aureus in terms of the length of stay or in the percentage of patients that required PICU transfer regardless of their initial antibiotic regimen of vancomycin or clindamycin without another antibiotic that the organism was sensitive to (results not shown).

discussion

In this retrospective cohort study, we describe the bacterial pathogens responsible for osteoarticular infections in children at two large pediatric hospitals. Traditionally, MRSA bone and joint infections are considered more severe leading to greater morbidity than those with MSSA infections.(8, 12, 15) We found MRSA infections were associated with higher inflammatory markers and a longer length of hospitalization compared to MSSA infections, but contrary to many other studies, we did not find any significant differences in the percentage of patients with severe disease sequelae such as a subperiosteal abscess after adjusting for potential confounders. An et al.(22) found no significant differences in inflammatory markers or length of hospitalization between MRSA and MSSA infections. Together these findings could represent an increased virulence of MSSA over time or could relate to geographical variability of bacterial virulence. We did not look for Panton-Valentine leucocidin in our study which is a virulence factor produced by S. aureus and has been associated with a more invasive disease.(23, 24) It is most often produced by MRSA strains but has also been recently detected in some MSSA strains associated with more severe infections and worse outcomes.(15)

Consistent with several recent studies, we found that musculoskeletal infections caused by MSSA in our community have been gradually increasing during our 8-year study period coinciding with a decline in MRSA prevalence. MSSA accounted for 35.9% of isolates in our study in 2009 and increased to 54.5% in 2016. We observed a significant trend of yearly increase in the frequency of AHO and SA caused by MSSA every year except there was a slight decline from 45.3% in 2011 to 44.2% in 2012. In 2016 there were fewer overall cases of AHO and SA included in the study sample compared to prior years. While this could potentially affect the results, cases included were at random, and the overall trend of MSSA cases remained consistent with the previous year’s trends.

In our study, clindamycin and/or vancomycin were frequently used as part of the initial antibiotic regimen for S. aureus coverage. Patients who were initially placed on a regimen containing vancomycin were more likely to require a PICU transfer and had longer length of hospitalization compared to those initially placed on clindamycin even after attempting to control for severity of infection. Historically vancomycin has been recommended in the empiric therapy of invasive musculoskeletal infections to cover MRSA, however more recently, clindamycin has been recommended as a potential alternative in patients who are not critically ill and in areas where clindamycin resistance is low.(18, 19) In 2018 at our institution, guidelines for musculoskeletal infections were developed which recommend clindamycin instead of vancomycin for S. aureus coverage unless the patient was hemodynamically unstable or required intensive care. Clindamycin has demonstrated excellent tissue penetration including in bone(25), is well tolerated(26), and has been effective in treatment of invasive infections caused by MSSA and MRSA.(27) However, there are no randomized studies comparing the efficacy of vancomycin to clindamycin in AHO and SA in pediatric patients.

The majority of MRSA and MSSA isolates at both of our study sites were susceptible to clindamycin, however over the years there was a trend of reduced sensitivity to clindamycin in our MRSA strains but not with the MSSA strains. The opposite trend was seen in a large surveillance study of S. aureus isolates in pediatric patients in the Military Health System, with a decline in MSSA susceptibility to clindamycin from 2005 to 2014, whereas the MRSA susceptibility rates remained stable.(16) Increased use of clindamycin may lead to increased resistance and thus regular surveillance monitoring of antimicrobial susceptibility will be important to inform future treatment guidelines.

We did not find any significant differences in our outcome measures for clindamycin-resistant S. aureus regardless of the initial antibiotic regimen, which gives support that clindamycin may be a safe and effective treatment in these infections while awaiting sensitivity results. However, it should be noted that the number of patients with clindamycin resistance was relatively small which limits conclusions that can be drawn regarding the possible safety of the empiric use of clindamycin. Additional studies comparing the initial antibiotic regimens for MRSA and MSSA musculoskeletal infections are needed to help guide therapy.

Strengths and Limitations:

A key strength of our study was our large sample size with manual chart review which provided a granularity of detail that many other studies have not done. We also recognize several limitations inherent to our retrospective study design. First, our study was conducted at two large institutions in nearby suburban areas and thus our findings may not be generalizable to other locations particularly regarding microbiology and antibiotic susceptibilities. Although the overall sample size was large, small numbers of patients in several subgroup analyses could have resulted in lack of statistical power to detect significant differences. In addition, misclassification of exposures and outcomes serve as a potential source of bias. This was mitigated by using a structured chart review with a pre-defined data collection form. We only examined associations and cannot make any conclusions on causality, and we cannot account for the effects of unmeasured confounders on our observations. In our comparisons of the initial antibiotic regimen, we did not collect data on changes to the initial antibiotic regimen, or the length or route of treatment which could have affected clinical outcomes. While we controlled for various measures of infection severity, it is possible that sicker appearing children were treated initially with a regimen including vancomycin which we were not able to control for in our analyses.

Lastly, we did not look at the timing of PICU transfers or so this outcome measure may not be attributable to the initial antibiotic choice. Despite these limitations, our study provides helpful data that may be used to guide empiric antibiotic choices and inform future studies.

conclusions

The present study adds to the knowledge of the differences and changing epidemiology of MRSA and MSSA bone and joint infections and may help guide initial antibiotic selection. Empiric vancomycin may not be routinely indicated for all children, and clindamycin may be a reasonable alternative in communities with similar local pathogens and resistance patterns. Patients should be closely monitored for antibiotic response and therapy adjusted as appropriate depending on culture results. Ongoing surveillance of the changing trends in MRSA and MSSA osteoarticular infections is necessary and more studies comparing antibiotic regimens will be important to continue to guide optimal treatment recommendations.

references

1.Arnold JC, Bradley JS. Osteoarticular Infections in Children. Infectious disease clinics of North America. 2015;29(3):557–74.

2.Copley LA. Pediatric musculoskeletal infection: trends and antibiotic recommendations. The Journal of the American Academy of Orthopaedic Surgeons. 2009;17(10):618–26.

3.Saavedra-Lozano J, Falup-Pecurariu O, Faust SN, Girschick H, Hartwig N, Kaplan S, et al. Bone and Joint Infections. The Pediatric infectious disease journal. 2017;36(8):788–99.

4.Darville T, Jacobs RF. Management of acute hematogenous osteomyelitis in children. The Pediatric infectious disease journal. 2004;23(3):255–7.

5.Branson J, Vallejo JG, Flores AR, Hulten KG, Mason EO, Kaplan SL, et al. The Contemporary Microbiology and Rates of Concomitant Osteomyelitis in Acute Septic Arthritis. The Pediatric infectious disease journal. 2017;36(3):267–73.

6.Kaplan SL. Recent lessons for the management of bone and joint infections. The Journal of infection. 2014;68 Suppl 1:S51–6.

7.Arnold SR, Elias D, Buckingham SC, Thomas ED, Novais E, Arkader A, et al. Changing patterns of acute hematogenous osteomyelitis and septic arthritis: emergence of community-associated methicillin-resistant Staphylococcus aureus. Journal of pediatric orthopedics. 2006;26(6):703–8.

8.Sarkissian EJ, Gans I, Gunderson MA, Myers SH, Spiegel DA, Flynn JM. Community-acquired Methicillin-resistant Staphylococcus aureus Musculoskeletal Infections: Emerging Trends Over the Past Decade. Journal of pediatric orthopedics. 2016;36(3):323–7.

9.Gerber JS, Coffin SE, Smathers SA, Zaoutis TE. Trends in the incidence of methicillin-resistant Staphylococcus aureus infection in children’s hospitals in the United States. Clinical infectious diseases: an official publication of the Infectious Diseases Society of America. 2009;49(1):65–71.

10.Ju KL, Zurakowski D, Kocher MS. Differentiating between methicillin-resistant and methicillin-sensitive Staphylococcus aureus osteomyelitis in children: an evidence-based clinical prediction algorithm. The Journal of bone and joint surgery American volume. 2011;93(18):1693–701.

11.Dietrich LN, Reid D, Doo D, Fineberg NS, Khoury JG, Gilbert SR. Predicting MSSA in Acute Hematogenous Osteomyelitis in a Setting With MRSA Prevalence. Journal of pediatric orthopedics. 2015;35(4):426–30.

12.Kini AR, Shetty V, Kumar AM, Shetty SM, Shetty A. Community-associated, methicillin-susceptible, and methicillin-resistant Staphylococcus aureus bone and joint infections in children: experience from India. Journal of pediatric orthopedics Part B. 2013;22(2):158–66.

13.Davis WT, Gilbert SR. Comparison of Methicillin-resistant Versus Susceptible Staphylococcus aureus Pediatric Osteomyelitis. Journal of pediatric orthopedics. 2018;38(5):e285-e91.

14.McNeil JC, Forbes AR, Vallejo JG, Flores AR, Hulten KG, Mason EO, et al. Role of Operative or Interventional Radiology-Guided Cultures for Osteomyelitis. Pediatrics. 2016;137(5).

15.Kok EY, Vallejo JG, Sommer LM, Rosas L, Kaplan SL, Hulten KG, et al. Association of Vancomycin MIC and Molecular Characteristics with Clinical Outcomes in Methicillin-Susceptible Staphylococcus aureus Acute Hematogenous Osteoarticular Infections in Children. Antimicrobial agents and chemotherapy. 2018;62(5).

16.Sutter DE, Milburn E, Chukwuma U, Dzialowy N, Maranich AM, Hospenthal DR. Changing Susceptibility of Staphylococcus aureus in a US Pediatric Population. Pediatrics. 2016;137(4).

17.Hulten KG, Mason EO, Lamberth LB, Forbes AR, Revell PA, Kaplan SL. Analysis of Invasive Community-Acquired Methicillin-Susceptible Staphylococcus aureus Infections During a Period of Declining Community Acquired Methicillin-Resistant Staphylococcus aureus Infections at a Large Children’s Hospital. The Pediatric infectious disease journal. 2018;37(3):235–41.

18.Liu C, Bayer A, Cosgrove SE, Daum RS, Fridkin SK, Gorwitz RJ, et al. Clinical practice guidelines by the infectious diseases society of america for the treatment of methicillin-resistant Staphylococcus aureus infections in adults and children. Clinical infectious diseases: an official publication of the Infectious Diseases Society of America. 2011;52(3):e18–55.

19.Harik NS, Smeltzer MS. Management of acute hematogenous osteomyelitis in children. Expert review of anti-infective therapy. 2010;8(2):175–81.

20.McBride S, Thurm C, Gouripeddi R, Stone B, Jaggard P, Shah SS, et al. Comparison of Empiric Antibiotics for Acute Osteomyelitis in Children. Hospital pediatrics. 2018;8(5):280–7.

21.Lansell A VY, Suchdev P, Figueroa J, Kirpalani A. Impact of Antibiotic Pretreatment on Cultures in Children with Osteomyelitis and Septic Arthritis. 2018.

22.An TJ, Benvenuti MA, Mignemi ME, Martus J, Wood JB, Thomsen IP, et al. Similar Clinical Severity and Outcomes for Methicillin-Resistant and Methicillin-Susceptible Staphylococcus aureus Pediatric Musculoskeletal Infections. Open forum infectious diseases. 2017;4(1):ofx013.

23.Bocchini CE, Hulten KG, Mason EO, Jr., Gonzalez BE, Hammerman WA, Kaplan SL. Panton-Valentine leukocidin genes are associated with enhanced inflammatory response and local disease in acute hematogenous Staphylococcus aureus osteomyelitis in children. Pediatrics. 2006;117(2):433–40.

24.Martinez-Aguilar G, Avalos-Mishaan A, Hulten K, Hammerman W, Mason EO, Jr., Kaplan SL. Community-acquired, methicillin-resistant and methicillin-susceptible Staphylococcus aureus musculoskeletal infections in children. The Pediatric infectious disease journal. 2004;23(8):701–6.

25.Panzer JD, Brown DC, Epstein WL, Lipson RL, Mahaffey HW, Atkinson WH. Clindamycin levels in various body tissues and fluids. The Journal of clinical pharmacology and new drugs. 1972;12(7):259–62.

26.Peltola H, Paakkonen M, Kallio P, Kallio MJ. Clindamycin vs. first-generation cephalosporins for acute osteoarticular infections of childhood—a prospective quasi-randomized controlled trial. Clinical microbiology and infection: the official publication of the European Society of Clinical Microbiology and Infectious Diseases. 2012;18(6):582–9.

27.Martinez-Aguilar G, Hammerman WA, Mason EO, Jr., Kaplan SL. Clindamycin treatment of invasive infections caused by community-acquired, methicillin-resistant and methicillin-susceptible Staphylococcus aureus in children. The Pediatric infectious disease journal. 2003;22(7):593–8.

tables

Table 1: Patient characteristics and outcomes based on pathogen isolated

Table 2. Adjusted differences in selected health outcomes comparing Methicillin-resistant Staphylococcus aureus vs Methicillin-susceptible Staphylococcus aureus

Table 3. Adjusted differences in selected health outcomes by isolated pathogen

Table 4: Patient characteristics and outcomes based on initial antibiotic regimen