Osteoarticular Infections Caused by Streptococcus pneumoniae in Children in the Post–Pneumococcal Conjugate Vaccine Era

Abstract

Background: 

Pneumococcal osteoarticular infections (OAIs) are an uncommon manifestation of invasive pneumococcal disease (IPD). We describe the demographic characteristics, hospitalization rate, serotype distribution and antibiotic susceptibility of children with pneumococcal OAI over a 16-year period.

Methods: 

We identified patients ≤18 years old with pneumococcal OAI at 8 children’s hospitals in the United States (2000–2015). Pneumococcal isolates were collected prospectively. Serotyping and antibiotic susceptibility were performed in a central laboratory.

Results: 

We identified 97 (3.3%) patients with pneumococcal OAI out of 2943 patients with IPD. Over 60% of the children were <2 years old. Septic arthritis (56.7%, 55/97) was the most common pneumococcal OAI, followed by osteomyelitis (25.8%, 25/97) and septic arthritis with concomitant osteomyelitis (17.5%, 17/97). Hospitalization for pneumococcal OAI overall decreased from 6.8 [95% confidence interval (CI): 5.2–8.6] to 4.4 (95% CI: 3.0–6.3) per 100,000 admissions from 2000–2009 to 2010–2015 (–35%, P = 0.05). Hospitalization for pneumococcal OAI caused by PCV13 serotypes decreased from 4.6 (95% CI: 3.4–6.2) to 0.9 (95% CI: 0.3–1.9) per 100,000 admissions from 2000–2009 to 2010–2015 (–87%, P < 0.0001). Overall, 12% of isolates had a penicillin minimal inhibitory concentration> 2 µg/mL, 3% a ceftriaxone minimal inhibitory concentration> 1 µg/mL and 15% were clindamycin resistant; these proportions remained unchanged after the introduction of PCV13. Serotypes 19A and 35B were responsible for penicillin and ceftriaxone nonsusceptible isolates in 2010–2015.

Conclusions: 

Pneumococcal OAI represents 3% of all IPD, affecting mainly healthy infants and young children. Hospitalization for pneumococcal OAI caused by PCV13 serotypes dramatically decreased (–87%) after the introduction of PCV13.

Streptococcus pneumoniae is well known for causing pneumonia, bacteremia and meningitis in pediatric patients. Although other manifestations of invasive pneumococcal disease (IPD) such as septic arthritis and osteomyelitis are less common, they have been increasingly recognized in the literature.^1–5Staphylococcus aureus is by far the most frequent causative organism of osteoarticular infections (OAIs) in children and adolescents, followed by Streptococcus pyogenes and Kingella kingae.^6–8 However, approximately 2%–10% of OAI in children are caused by S. pneumoniae,^8–11 and this should be considered when selecting empiric antibiotic therapy. Moreover, pneumococcal OAI represents 3% of all IPD in children.²

A substantial reduction of IPD after the introduction of the 13-valent pneumococcal conjugate vaccine (PCV13) has been documented.^12,13 However, scarce data exist on the impact of PCV13 on pneumococcal OAI in children. The goal of this study was to characterize the hospitalization rate and the clinical and microbiologic characteristics of pneumococcal OAI in children evaluated at 8 children’s hospitals in the United States before and after the introduction of PCV13.

MATERIALS AND METHODS

The US Pediatric Multicenter Pneumococcal Surveillance Study Group consists of investigators from 8 children’s hospitals throughout the country (Houston, TX; Pittsburgh, PA; Little Rock, AR; San Diego, CA; Los Angeles, CA; Chicago, IL; Columbus, OH and Winston-Salem, NC) who have been prospectively identifying children with IPD and collecting their pneumococcal isolates since September 1993.

We identified patients ≤18 years admitted with pneumococcal OAI between January 1, 2000, and December 31, 2015, from our study group database. The study was divided into 2 periods: pre-PCV13 from January 2000 to December 2009 and post-PCV13 from January 2010 to December 2015. The study was approved by the Institutional Review Boards of each of the participating hospitals.

Pneumococcal OAI was defined as the isolation of S. pneumoniae from blood, synovial fluid and/or bone in children with clinical presentation and imaging consistent with septic arthritis and/or osteomyelitis. Clinical information was collected retrospectively from medical records and recorded on a standardized case report form. Concomitant myositis or pyomyositis was also evaluated in patients with pneumococcal OAI who underwent magnetic resonance imaging (MRI). Myositis was defined as skeletal muscle inflammation extending diffusely through ≥1 muscle group without distinct abscess formation, and pyomyositis was defined as inflammation of skeletal muscle with localized abscess formation. Administration of PCV7 and/or PCV13 was documented through the medical records or by contacting the patient’s health care provider.

Pneumococcal isolates were serotyped by the Quellung reaction in the Infectious Disease Research Laboratory at Texas Children’s Hospital, Houston, TX. Minimal inhibitory concentrations (MICs) for penicillin and ceftriaxone were performed by standard microbroth dilution. Clindamycin susceptibility was obtained from patients’ medical records. Susceptibility categories were determined using the 2015 Clinical and Laboratory Standards Institute guidelines (nonmeningeal breakpoints for pneumococci: penicillin: ≤2.0 µg/mL = susceptible, 4.0 µg/mL = intermediate and ≥8.0 µg/mL = resistant; ceftriaxone: ≤1.0 µg/mL = susceptible, 2.0 µg/mL = intermediate and ≥4.0 µg/mL = resistant and clindamycin: ≤0.25 µg/mL = susceptible, 0.5 µg/mL = intermediate and ≥1 µg/mL = resistant).

Descriptive statistics were used to characterize the study population. The χ² test and Fisher exact test were used to analyze dichotomous variables, and Mann-Whitney U test for continuous variables. Annual hospitalization rates per 100,000 admissions with 95% confidence intervals (CIs) were calculated for all pneumococcal OAI and cases caused by PCV13 and non-PCV13 serotypes. A P value ≤ 0.05 was considered significant. IBM SPSS statistics v22.0.0 (IBM Corporation, Armonk, New York) was used.

RESULTS

We identified 97 patients with pneumococcal OAI representing 3.3% of all IPD (2943 patients) from 2000 to 2015. Hospitalization rate for all pneumococcal OAI decreased from 6.8 (95% CI: 5.2–8.6) per 100,000 admissions in 2000–2009 to 4.4 (95% CI: 3.0–6.3) per 100,000 admissions in 2010–2015 (–35%, P = 0.05) (Fig. 1). No differences in age, gender, race/ethnicity or comorbidity were identified between 2000–2009 and 2010–2015 (Table 1). In 2000–2009, 34.3% of patients with pneumococcal OAI had received ≥2 doses of PCV7. In 2010–2015, 43.3% of patients with pneumococcal OAI had received ≥2 doses of PCV13.

TABLE 1.: Demographic Characteristics of Children With Pneumococcal Osteoarticular Infections Before (2000–2009) and After (2010–2015) the Introduction of PCV13

FIGURE 1.: Annual hospitalization rates of pneumococcal OAI per 100,000 admissions from 2000 to 2015. The annual combined admissions among the 8 children’s hospitals were 2000 = 90,097; 2001 = 91,950; 2002 = 94,055; 2003 = 95,809; 2004 = 96,472; 2005 = 99,008; 2006 = 101,116; 2007 = 103,524; 2008 = 105,367; 2009 = 111,969; 2010 = 114,615; 2011 = 118,077; 2012 = 107,737; 2013 = 110,241; 2014 = 111,089 and 2015 = 112,459. Hospitalization rate for pneumococcal OAI decreased 35% from 2000–2009 to 2010–2015 (P = 0.05).

The most common pneumococcal OAI was septic arthritis (56.7%, 55/97), followed by osteomyelitis (25.8%, 25/97) and septic arthritis with concomitant osteomyelitis (17.5%, 17/97). Five patients presented with osteomyelitis of ≥2 bones; of these, 4 had septic arthritis with concomitant osteomyelitis. No statistically significant difference was identified when comparing the proportion of pneumococcal OAI (septic arthritis, osteomyelitis or septic arthritis with concomitant osteomyelitis) before and after the introduction of PCV13 (Table 1). Patients with septic arthritis [12.5 months interquartile range (IQR) (7.6–34.6)] were younger than patients with osteomyelitis [50.6 months IQR (8.0–118.0)], P = 0.05.

The most common joints involved in septic arthritis were hip 35 (48.6%), knee 15 (20.8%), elbow 8 (11.1%), ankle 7 (9.7%), shoulder 5 (6.9%), wrist and sternoclavicular 1 each (1.4% each). The most common bones involved in osteomyelitis were femur 13 (26.5%), humerus 8 (16.3%), tibia 5 (10.2%), calcaneous 5 (10.2%), talus 3 (6.1%), vertebra 3 (6.1%), radius, fibula, ilium and sternum 2 each (4.1% each) and ulna, ischium, scapula and parietal 1 each (2% each). Bacteremia was present in 47.3% (26/55) of patients with septic arthritis, 40% (10/25) of osteomyelitis and 35.3% (6/17) of septic arthritis with concomitant osteomyelitis. Percentages of positive blood, synovial fluid, bone and other cultures are shown in Figure, Supplemental Digital Content 1, https://links.lww.com/INF/C771. MRI was performed in 50 patients; 30 (60%) had myositis, 8 (16%) pyomyositis, 5 (10%) intraosseous abscess and 3 (6%) subperiosteal abscess. No fatalities secondary to pneumococcal infection were reported.

Two isolates were not available for serotyping or antibiotic susceptibility testing. The most common serotypes in 2000–2009 were 19A (27.7%, 18/65) and 14 (12.3%, 8/65). The most common serotypes in 2010–2015 were 33F and 35B (20%, 6/30 each), while 19A accounted for 16.7% (5/30) during the same period (Fig., Supplemental Digital Content 2, https://links.lww.com/INF/C772). Hospitalization rate for pneumococcal OAI caused by PCV13 serotypes decreased from 4.6 (95% CI: 3.4–6.2) per 100,000 admissions in 2000–2009 to 0.9 (95% CI: 0.3–1.9) per 100,000 admissions in 2010–2015 (–87%, P < 0.0001). Hospitalization rate for pneumococcal OAI caused by non-PCV13 serotypes was 2.1 (95% CI: 1.3–3.2) per 100,000 admissions in 2000–2009 and 3.6 (95% CI: 2.3–5.3) per 100,000 admissions in 2010–2015 (+71%, P=0.08) (Fig. 1).

Penicillin, ceftriaxone and clindamycin susceptibility did not change after the introduction of PCV13 (Table 2). In 2010–2015, serotypes 19A (n = 2) and 35B (n = 1) were responsible for all isolates with penicillin MIC> 2 µg/mL and ceftriaxone MIC> 1 µg/mL. Serotypes 19A (n = 3) and 15A (n = 2) were responsible for all clindamycin-resistant isolates in 2010–2015.

TABLE 2.: Antibiotic Nonsusceptibility/Resistance of Isolates From Children With Pneumococcal Osteoarticular Infections Before (2000–2009) and After (2010–2015) the Introduction of PCV13

DISCUSSION

Our study is the largest case series of pediatric pneumococcal OAI assembled to date and the first to characterize the impact of PCV13 on pneumococcal OAI in a pediatric population in the United States. We observed that pneumococcal OAI remained a relatively uncommon presentation of IPD, representing 3.3% of all IPD. Pneumococcal OAI affected mainly healthy children <2 years of age in our study. The introduction of PCV13 substantially reduced by 87% the hospitalization rate of pneumococcal OAI caused by serotypes included in PCV13. Otherwise, no significant changes in clinical presentation or antimicrobial susceptibility among pneumococcal OAI were observed.

After the introduction of PCV13, overall hospitalization per 100,000 admissions for pneumococcal OAI decreased by 35% from 2000–2009 to 2010–2015. A more dramatic decrease of 87% was observed among pneumococcal OAI caused by PCV13 serotypes. As seen with other pneumococcal diseases, 19A was the most common serotype among OAI isolates before the introduction of PCV13. The distribution of other serotypes in this study also reflects the serotype distribution observed among other IPD within our surveillance group and other groups.¹⁴ Penicillin and ceftriaxone susceptibility remained above 85% and 95%, respectively, after the introduction of PCV13. Overall clindamycin susceptibility was 85%, making this antibiotic a reasonable empiric antibiotic choice for OAI in otherwise noncritically ill normal children in whom S. pneumoniae is a possible pathogen in addition to Staphylococcus aureus and S. pyogenes.

We found that 60% of the pediatric population affected by pneumococcal OAI were younger than 2 years of age, which is consistent with 70%–78% reported by other groups.^1,3,5 None of our patients had a concomitant pneumococcal infection outside the musculoskeletal system. In a previous report from our group, a child with pneumococcal OAI was also found to have pneumococcal meningitis.² Others have also noticed that infection sites outside the musculoskeletal system are very unusual in children with pneumococcal OAI.^3,5 However, about half of adults with pneumococcal OAI present with infection at another site, the most common being pneumonia and the second most common being meningitis.⁴

Septic arthritis was the most common presentation of pneumococcal OAI in our study, as reported by others.³ Joints and bones of the lower extremities were most affected. Similar to other groups, we found that the hip and the femur were the most common joint and bone involved, respectively.^2,3,5 In contrast, a pediatric study from France reported the elbow as the most commonly affected joint in pneumococcal septic arthritis.³ In our study, we found 17.5% of pneumococcal septic arthritis with concomitant osteomyelitis. Other groups have reported that 9.3%–23.8% of children with pneumococcal septic arthritis had concomitant osteomyelitis.^1–3

Concomitant pneumococcal pyomyositis was observed in 8 (16%) patients who underwent MRI. S. pneumoniae is not commonly recognized as a causative agent of deep muscle infections; other bacteria such as Staphylococcus aureus and S. pyogenes are more frequently associated with pyomyositis.^15,16 Pneumococcal pyomyositis has been reported mainly in adults with chronic predisposing conditions or as a complication of pneumococcal meningitis, but few cases in children have been documented.^17–19

Even though bacteremia was documented in an important proportion (~40%) of patients with pneumococcal OAI, many more required invasive procedures to obtain samples in which S. pneumoniae was identified. For patients with pneumococcal osteomyelitis, 40% required a bone biopsy to establish the etiologic diagnosis; a similar proportion has been described for other pathogens.²⁰

Some limitations of this study should be considered. First, even though pneumococcal isolates were collected prospectively, clinical information was obtained retrospectively from medical records. Second, the diagnostic approach to OAI has evolved since the early 2000s, thus not all patients included in our study systematically underwent advanced imaging studies and/or surgical procedures. As a result, the proportion of patients with septic arthritis only, osteomyelitis only and concomitant septic arthritis and osteomyelitis may not be precise. Last, it is possible that the sample size in the pre- and post-PCV13 subgroups did not provide enough power to identify statistically significant differences, if they existed.

CONCLUSION

Pneumococcal OAI remains an uncommon presentation of IPD and affects mainly healthy infants and young children. The introduction of PCV13 has had a remarkable impact in decreasing hospitalization for OAI caused by serotypes included in PCV13. However, much of this reduction has been countered by a rise in disease caused by non-PCV13 serotypes. Penicillin, ceftriaxone and clindamycin susceptibility remained high after the introduction of PCV13.