OBJECTIVE. To determine the complications and risk factors for complications associated with using central venous catheters (CVCs) for the treatment of acute hematogenous osteomyelitis (AHO).
METHODS. We conducted a retrospective cohort study of all patients admitted to the Children’s Hospital of Philadelphia between January 1, 2000, and December 31, 2003, with a diagnosis of AHO.
RESULTS. Eighty patients with AHO met inclusion criteria. The median age was 5 years, and 66% of the patients were male. The most commonly affected bones were the femur (25%), tibia (20%), and pelvis (16%). Staphylococcus aureus was the most common organism identified from cultures of bone (67%) and blood (30%). Seventy-five patients (94%) received >2 weeks of intravenous (IV) antibiotic therapy via a CVC and 5 (6%) received <2 weeks of IV antibiotic therapy before conversion to oral therapy for a median of 25 days. None of the patients who switched to oral therapy within 2 weeks was rehospitalized or returned to the emergency department. Of the 75 patients who received >2 weeks of IV therapy, 41% had ≥1 CVC-associated complication. Seventeen patients (23%) had a CVC malfunction or displacement, 8 (11%) had a catheter-associated bloodstream infection, 8 (11%) had fever with negative blood culture results, and 4 (5%) had a local skin infection at the site of catheter insertion. Older age was protective against the development of a CVC-associated complication, whereas the lowest median household income was associated with development of a CVC-associated complication.
CONCLUSIONS. Interventions to reduce CVC-associated complications should be developed and evaluated, particularly for young children and those from families with low household incomes. Clinical trials are needed to evaluate the safety and efficacy of oral antibiotic therapy after a short course of IV therapy as an alternative to prolonged IV therapy.
EACH YEAR IN the United States, 1 of 5000 children <13 years of age is diagnosed as having osteomyelitis, accounting for 1% of all pediatric hospitalizations.1,2 The most common presentation of osteomyelitis among children is acute hematogenous osteomyelitis (AHO), in which infection arises after an episode of bacteremia, resulting in deposition of organisms in the bone. AHO affects primarily young children, with 50% of cases occurring among patients <5 years of age, possibly because of the rich vascularization of rapidly growing bones.3–5Staphylococcus aureus is the pathogen identified in most cases of AHO, followed by Streptococcus pyogenes, Streptococcus pneumoniae, and Salmonella spp among patients with sickle cell disease.6
A common approach to the treatment of AHO has consisted of several days of antibiotic administration through peripheral intravenous (IV) treatment in the hospital, followed by placement of a central venous catheter (CVC) and completion of 4 to 6 weeks of IV therapy at home. Several studies and a systematic review suggested that initiation of therapy with IV antibiotic treatment followed by early conversion to oral therapy may be an effective alternative to longer courses of IV therapy for AHO.7–18 The use of oral therapy is appealing because of the inconvenience, costs, and potential complications of prolonged IV therapy.19 Few studies have evaluated the potential complications of using CVCs in the treatment of AHO or the risk factors for developing a CVC-associated complication. In this study, we determined the complications of using CVCs in the treatment of AHO and the risk factors for developing a CVC-associated complication.
We performed a retrospective cohort study of all patients admitted to the Children’s Hospital of Philadelphia (CHOP), between January 1, 2000, and December 31, 2003, with a diagnosis of AHO. Potential cases were identified in hospital billing data by using discharge International Classification of Diseases, Ninth Revision, codes for osteomyelitis (codes 730.00–730.29 and 003.24). Osteomyelitis was considered acute (versus chronic) if a bone biopsy demonstrated acute infection or, in the absence of a bone biopsy, if onset of symptoms (including fever, bone pain, limp, fatigue, local erythema, and swelling) began within 2 weeks before presentation to the hospital. We excluded patients who were <2 months of age and those who had underlying immunodeficiency (malignancy, sickle cell disease, HIV infection, or chronic immunosuppresion), contiguous infection because of hardware placement, cutaneous ulcers, penetrating trauma, or a history of crush injury. Historically, the standard of care at CHOP for children with AHO has been to administer 4 to 6 weeks of IV antibiotic therapy. Catheter types include peripherally inserted central catheters (PICCs) and Broviac catheters (Bard Access, Salt Lake City, UT). At CHOP, PICCs are placed by interventional radiologists and Broviac catheters are placed in the operating room. Before discharge, all parents receive formal instruction, including a didactic session and handouts, regarding the care and management of CVCs. Translators are available for parents who require teaching in a language other than English.
We used standardized data abstraction forms to review the inpatient and outpatient medical records for all study patients. Outpatient records from the orthopedics and infectious diseases clinics were reviewed for a period of up to 1 year. Records for all return visits to the CHOP emergency department (ED) and inpatient readmissions were reviewed. Data obtained included age, gender, race, duration of symptoms before presentation, and location of infection. Median household income, based on residential zip code, was obtained from US census data. Patients were stratified into 4 socioeconomic groups by using the following categories of median household income: less than $24999, $25000 to $34999, $35000 to $44999, and more than $45000.
All diagnostic testing results were reviewed, including complete blood cell counts, erythrocyte sedimentation rates, C-reactive protein levels, blood culture findings, bone and/or joint culture findings, bone biopsy findings, radiographs, MRI scans, and bone scans. We reviewed the choice of antimicrobial therapy, route of administration, and duration of treatment.
We identified all CVC-associated complications, including CVC malfunction, CVC displacement, catheter-associated bloodstream infection, admission for IV antibiotic therapy to rule out catheter-associated bloodstream infection, and infection at the CVC entry site. Catheter malfunction was defined as mechanical breakage of the catheter or thrombus in the line, necessitating heparin or tissue plasminogen activator infusion. Catheter displacement was defined as partial or complete removal of the catheter. Catheter-associated bloodstream infection was defined as clinical signs of bacteremia, such as fever and chills, in conjunction with isolation of a pathogen from a blood culture. Patients were also considered to have a CVC-associated complication if they presented with fever and clinical signs suggesting infection, necessitating hospital admission and antibiotic administration, even if blood cultures subsequently yielded negative results. Local skin infection was defined as erythema, edema, and/or induration surrounding the site of catheter insertion, necessitating treatment or catheter removal.
We expressed categorical variables as frequencies and percentages. We summarized continuous variables as mean, median, interquartile range (IQR), and SD. We conducted univariate analyses to determine the association between potential risk factors and the development of a complication. We compared categorical variables by using Fisher’s exact test and continuous outcome variables by using the Wilcoxon rank sum test. For all statistical calculations, a 2-tailed P value of <.05 was considered significant. All statistical calculations were performed with a standard program in Stata version 8.0 (Stata Corp, College Station, TX).
Human Subjects Oversight
The study was approved by the Committee for the Protection of Human Subjects at CHOP.
Eighty patients met the inclusion criteria for AHO. Five patients (6%) were diagnosed as having AHO on the basis of bone biopsy results, and 75 were diagnosed on the basis of clinical criteria alone. Demographic features, characteristics of initial hospitalization, and clinical characteristics of infection are presented in Table 1. The median age was 5 years, and 66% of patients were male. The most commonly affected bones were the femur (25%), tibia (20%), and pelvis (16%). We identified a pathogen in blood and/or bone culture for 27 patients (39%), with S aureus being the most commonly identified organism. Two (7%) of the 27 S aureus isolates were found to be methicillin resistant.
Seventy-five patients (94%) received >2 weeks of IV antibiotic therapy via a CVC (median: 31 days; IQR: 27–42 days) and 5 (6%) received <2 weeks of IV antibiotic therapy (median: 5 days; IQR: 5–12 days) and were then transitioned to oral therapy (median: 25 days; IQR: 24–27 days). Among the 75 patients treated with >2 weeks of IV therapy, there were a total of 95 CVCs placed, including 89 PICCs and 6 Broviac lines. Sixty patients (80%) had only 1 catheter inserted, 10 patients (13%) required 2 consecutive catheters, and 5 patients (7%) required 3 consecutive catheters. The median duration of catheter placement was 37 days (IQR: 28–59 days). The majority of patients were treated with one of the following antistaphylococcal antibiotics, depending on the susceptibility of the organism (if available): cefazolin, oxacillin, vancomycin, or clindamycin.
Of the 75 patients who received >2 weeks of IV therapy, 38 (51%) had ≥1 return visit to the ED or readmission for any reason, with 31 patients (41%) having ≥1 CVC-associated complication. CVC-associated complications are summarized in Table 2. The median time from catheter insertion to development of any catheter-associated complication was 17 days (IQR: 7–24 days). There were 9 catheter-associated bloodstream infections among 8 patients. The median time from catheter insertion to development of bloodstream infection was 18 days (IQR: 12–23 days). There was no difference in the duration of in-hospital therapy for patients who did or did not have complications. Organisms identified for patients with catheter-associated bloodstream infections included Pseudomonas spp (n = 5), Escherichia coli (n = 2), Enterobacter (n = 2), Klebsiella (n = 2), and Candida (n = 1). Three patients had >1 organism identified from blood cultures. There was no significant difference in complication rates between the group receiving therapy through a PICC and the group receiving therapy through a Broviac catheter. In addition, there was no significant difference in complication rates among patients receiving therapy 2 times per day, 3 times per day, or 4 times per day.
Among the 38 patients with ≥1 return visit, there were a total of 62 return encounters, 28 (45%) of which resulted in readmission. Forty-five (73%) of the 62 return encounters were CVC associated, including CVC malfunction or displacement (37%), bloodstream infection (15%), fever with negative blood culture results (13%), and local skin infection (8%). None of the patients switched to oral therapy within 2 weeks returned to the CHOP ED or were rehospitalized.
In univariate analyses, older age was protective (odds ratio: 0.90 for each 1 year of age; 95% confidence interval: 0.80–1.0) against the development of a CVC-associated complication, whereas the lowest median household income (odds ratio: 10.6; 95% confidence interval: 1.05–107.1) was associated with the development of a CVC-associated complication. The median age of patients with any catheter-associated complication was 2 years (IQR: 0–6 years), whereas the median age of patients without a catheter-associated complication was 7 years (IQR: 2–11 years; P = .03).
Sixty-five patients (81%) received outpatient follow-up care with orthopedic surgery and/or infectious disease services at CHOP within 1 year after initial hospital discharge. The time to the last documented follow-up visit was variable (median: 34 months; IQR: 26–49 months); therefore, we were unable to document the final outcome of AHO for all children in the cohort. However, 2 patients in the long-term IV therapy group were considered to have treatment failure, because of persistent symptoms and radiographic findings of continued infection, and were readmitted for additional IV therapy.
We found that using CVCs for long-term outpatient treatment of AHO was associated with a significant rate of complications, with 41% of patients having a CVC-associated complication resulting in an ED visit or readmission to the hospital. Young age and low socioeconomic status were associated with development of a CVC-associated complication. Most CVC-associated complications occurred after 2 weeks of catheter placement. We also found that 20% of patients treated with long-term IV therapy had ≥2 CVCs placed during their treatment, with the majority of initial CVCs being removed as a result of catheter malfunction or catheter-associated bloodstream infection.
Previous studies that evaluated the complications of outpatient parenteral antibiotic therapy (OPAT) showed similarly high CVC-associated complication rates. Gomez et al19 retrospectively reviewed data for 229 patients (median age: 8 years) who were treated with OPAT, 74 of whom had acute osteomyelitis. Those authors evaluated the complications of OPAT, compared with long-term inpatient IV antibiotic therapy, and found catheter-associated complications among 29% of patients treated with OPAT. Rehospitalization for any reason was required during 26% of OPAT courses, with 42% of rehospitalizations being attributable to catheter-associated complications.19 In our study, we found higher rates of return encounters and CVC-associated complications than reported by Gomez et al.19 This might be explained by the fact that we included ED visits in our definition of return hospital encounters, whereas Gomez et al19 included only repeat hospital admissions. Therefore, our data might describe more accurately the true burden of complications resulting from long-term CVC therapy. In addition, the median age of patients in our study was lower than the median age of patients in the study by Gomez et al.19 This might be consistent with our observation that younger age was associated independently with an increased risk of developing CVC-associated complications.
To our knowledge, no previous studies have addressed the risk factors for developing CVC-associated complications. In addition to young age, we found that children living in zip codes with lower reported median household income were at increased risk of developing a CVC-associated complication. These findings suggest that parents of younger patients and those with lower socioeconomic status may benefit from increased teaching and nursing supervision.
Several studies7,8,13,15,20 have shown that the majority of patients with AHO can be treated successfully with oral antibiotic therapy after a short course of IV therapy. In our study, we found that the majority of CVC-associated complications occurred after ≥2 weeks of CVC placement, which supports the use of early conversion to oral therapy. However, there have been no randomized, controlled trials evaluating the effectiveness of long-term IV therapy, compared with early conversion to oral therapy. In addition, recent changes in the epidemiological features of S aureus, with the increase in community-acquired, methicillin-resistant S aureus, may influence the success of OPAT and oral therapy. Newer orally administered agents that are active against methicillin-resistant S aureus, such as linezolid, have good oral bioavailability and may have a role in the treatment of AHO in the future.
One of the strengths of our study is the similarity of the demographic features of our population to those in previous population-based reports on AHO among children, which suggests that there was little selection bias in the assembly of our study sample. When compared according to age, gender, race, location of bony involvement, and organism identified, our sample was virtually identical to those described in other studies of AHO.3–5,16,21–24 A possible limitation of our study is the absence of data on patients who might have sought follow-up care at another hospital. The missing follow-up data might also underestimate the failure rates. However, the lack of this information only serves potentially to underestimate our reported complication rate. Another potential limitation is that patients were identified initially through query of hospital billing data according to International Classification of Diseases, Ninth Revision, codes. If cases were coded incorrectly, then we might have missed some patients who might have been eligible for inclusion. Finally, we stratified patients into socioeconomic groups on the basis of median household income according to zip code, as determined from census data. Given the expected individual-level variability of household incomes within a zip code, we might have misclassified the socioeconomic status of patients whose household incomes were inconsistent with the reported median values for their zip codes. However, because this misclassification would bias results to the null hypothesis, the association between low household income and development of a CVC-associated complication actually might be stronger than observed in this study. Finally, because of the small number of patients in the sequential IV to oral therapy group, we were unable to compare the effectiveness of the 2 treatment strategies.
The use of CVCs for long-term outpatient treatment of AHO was associated with significant rates of complications. Young age and low household incomes were associated with increased risks of CVC-associated complications. If the standard of care for the treatment of AHO continues to be long-term outpatient IV therapy, then interventions to reduce CVC-associated complications should be developed, particularly for young children and those from families with low household incomes. Given the frequency of complications with CVCs and the promising results of recent case series demonstrating effective treatment of AHO with early conversion to oral therapy, clinical trials are needed to evaluate the safety and efficacy of this management strategy. and efficacy of this management strategy.
Dr Zaoutis was supported by grant K23 A10629753-01 from the National Institutes of Health and grant U18-HS10399 from the Agency for Healthcare Research and Quality, Centers for Education and Research on Therapeutics. Dr Keren was supported by grant K23 HD043179 from the National Institute of Child Health and Human Development, Bethesda, MD.
- Accepted August 30, 2005.
- Address correspondence to Theoklis E. Zaoutis, MD, MSCE, Division of Infectious Diseases, Children’s Hospital of Philadelphia, CHOP North, Suite 1527, 3535 Market St, Philadelphia, PA 19104. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Gillespie WJ. The epidemiology of acute haematogenous osteomyelitis of childhood. Int J Epidemiol.1985;14 :600– 606
- ↵Peltola H, Unkila-Kallio L, Kallio MJ. Simplified treatment of acute staphylococcal osteomyelitis of childhood: the Finnish Study Group. Pediatrics.1997;99 :846– 850
- Dunkle LM, Brock N. Long-term follow-up of ambulatory management of osteomyelitis. Clin Pediatr (Phila).1982;21 :650– 655
- ↵Kolyvas E, Ahronheim G, Marks MI, Gledhill R, Owen H, Rosenthall L. Oral antibiotic therapy of skeletal infections in children. Pediatrics.1980;65 :867– 871
- ↵Walker SH. Staphylococcal osteomyelitis in children: success with cephaloridine-cephalexin therapy. Clin Pediatr (Phila).1973;12 :98– 100
- Copyright © 2006 by the American Academy of Pediatrics