INTRODUCTION — Infections of the joints may be caused by bacteria, fungi, and viruses. Bacterial joint infections are most common. The term septic arthritis encompasses bacterial arthritis, pyogenic arthritis, suppurative arthritis, purulent arthritis, and pyarthrosis [1,2].

The treatment and outcome of bacterial arthritis in infants and children will be reviewed here. The epidemiology, pathogenesis, microbiology, clinical features, and diagnosis of bacterial arthritis in infants and children are discussed separately. (See "Bacterial arthritis: Epidemiology, pathogenesis, and microbiology in infants and children" and "Bacterial arthritis: Clinical features and diagnosis in infants and children".)

The treatment of arthritis due to Lyme disease is also discussed separately. (See "Treatment of Lyme disease", section on 'Arthritis'.)

OVERVIEW OF TREATMENT — Bacterial arthritis requires prompt recognition and management. Delays in treatment are associated with long-term sequelae. These sequelae can have major lifelong impact if a major weight-bearing joint is involved, particularly when the hip joint is involved in infants.

The goals of treatment include sterilization and decompression of the joint space and removal of inflammatory debris to relieve pain and prevent deformity or functional sequelae [1,3]. Surgical drainage and antimicrobial therapy are the cornerstones of therapy.

Clinical decision making in the management of the child with bacterial arthritis includes the following questions [1]:

• How should the joint be drained and irrigated?
• Which antibiotics should be administered?
• What is the optimal duration of antimicrobial therapy?
• When is it safe to switch from parenteral to oral therapy?

Although drainage and antimicrobial therapy are well-established cornerstones of therapy for bacterial arthritis, there is little evidence from controlled trials to guide decisions about the optimal procedure for drainage or the choice, route, or duration of antimicrobial therapy [1,4].

Indications — Because delayed therapy has been associated with long-term sequelae, treatment of infants and children with suspected bacterial arthritis should begin immediately after blood and synovial fluid cultures are obtained. The synovial fluid white blood cell (WBC) count is used in conjunction with clinical findings to make decisions regarding the need for therapeutic drainage and initiation antimicrobial therapy while awaiting culture results. (See "Bacterial arthritis: Clinical features and diagnosis in infants and children", section on 'Diagnosis'.)

After appropriate cultures have been obtained, empiric antimicrobial therapy should be initiated for infants and children with characteristic clinical and laboratory features of bacterial arthritis, which may include:

• In newborns and young infants: clinical manifestations of sepsis (eg, irritability, poor feeding), pseudoparalysis, evidence of discomfort when being handled, or fever without a focus.
• In older infants and children: fever and monoarticular pain, swelling, and limited range of motion (in older infants and children); clinical manifestations of sepsis (eg, irritability, poor feeding), pseudoparalysis, dislike of being handled, or fever without a focus (in young infants)
• Elevated peripheral WBC count, erythrocyte sedimentation rate (ESR) and/or C-reactive protein (CRP)
• Elevated synovial fluid WBC with a predominance of neutrophils
• Organism identified on Gram stain of synovial fluid

The differential diagnosis also must be carefully considered to identify disorders with clinical features similar to bacterial arthritis that do not require joint drainage or prolonged antibiotic therapy (eg, transient synovitis, Lyme arthritis, systemic onset juvenile rheumatoid arthritis). (See "Bacterial arthritis: Clinical features and diagnosis in infants and children", section on 'Differential diagnosis'.)

General principles — Children with bacterial arthritis should be managed in conjunction with an orthopedic surgeon who is experienced in treating children. Aspiration of the joint should occur as soon as possible when bacterial arthritis is suspected. Joint aspiration provides synovial fluid for analysis and culture and decompresses the joint [5]. (See "Bacterial arthritis: Clinical features and diagnosis in infants and children", section on 'Synovial fluid' and 'Drainage' below.)

Systemic antimicrobial therapy should be administered as soon as possible after joint aspiration. Adequate antibiotic levels in the joint are essential. All antibiotics that have been studied readily enter the joint fluid after systemic administration; intraarticular injection of antibiotics (which may be painful) is unnecessary. The joint fluid concentration averages 30 percent of the serum concentration at the time of peak serum concentration [6-9]. In addition, because the efflux of antibiotic from the joint fluid is slow, the joint fluid concentration may exceed serum concentration immediately before the next dose.

Consultation with an expert in infectious diseases is suggested for children with post-operative infection, chronic joint infection, penetrating injury, unusual pathogens, or antimicrobial allergy [10]. Most children can be treated with sequential antibiotic therapy (parenteral therapy followed by oral therapy) to minimize the length of hospital stay and complications from prolonged vascular access. (See 'Antibiotic therapy' below.)

Adjunctive therapies may include pain control and physical therapy. Some experts recommend routine non-steroidal anti-inflammatory agents which decrease reactive synovitis associated with bacterial arthritis. Administration of dexamethasone for the first four days of therapy also has been explored as a possible adjunctive measure, but adjunctive dexamethasone therapy remains investigational [11]. (See 'Adjunctive therapies' below.)

Septic arthritis of the hip — Infants and children with bacterial arthritis of the hip require prompt surgical drainage and irrigation of the hip in addition to diagnostic aspiration. Delayed drainage increases the likelihood of adverse outcome. Increased intraarticular pressure can compromise the blood supply to the femoral head, resulting in avascular necrosis [12,13]. Additional indications for open surgical drainage versus needle aspiration (or recurrent needle aspiration) are discussed below. (See 'Arthrotomy' below.)

Home care — The duration of antimicrobial therapy for bacterial arthritis is dependent upon the organism isolated or suspected. For S. aureus, treatment is for at least three weeks. Many children receive a portion of this therapy at home. While they are being cared for at home, children with bacterial arthritis may require nursing visits for wound care, physical therapy, and/or a wheelchair or walker [10]. During outpatient therapy, follow-up should occur at least every one to two weeks. (See 'Subsequent outpatient therapy' below and 'Follow-up' below.)

DRAINAGE — No studies have compared drainage to no drainage in human patients with bacterial arthritis. In animal models, antimicrobial treatment without drainage does not prevent joint destruction unless antimicrobial therapy is administered very early in the disease course (within 48 to 72 hours) of inoculation [14-16]. This timeframe is almost never achieved in children with bacterial arthritis in whom the clinical presentation is often delayed.

At the time of presentation, joint infections are essentially closed abscesses. Drainage and lavage are necessary to decompress the joint space and to remove inflammatory debris in order to preserve synovium and collagen matrix [15,17]. Drainage can be accomplished through open surgery (arthrotomy), arthroscopy, or needle aspiration (single or multiple) [1,4,18-20].

Arthrotomy — Arthrotomy (open surgical drainage) facilitates thorough debridement of infected tissue, breakdown of loculations, and irrigation of the joint space [17,21,22]. However, arthrotomy requires general anesthesia and is more invasive than arthroscopy or needle aspiration.

We suggest open surgical drainage for bacterial arthritis of the hip in infants and children [13,23-25]. There are case reports of successful treatment with arthroscopic drainage, but these reports are few in number and generally involved older children [26-28]. Although some authors have reported successful management of bacterial arthritis of the hip in older children with needle aspiration if the duration of symptoms is less than four days [29], we do not advocate this approach.

Decisions regarding the optimal drainage procedure for joints other than the hip should be made on a case-by-case basis. Factors to be considered include the site and extent of joint involvement, duration of symptoms, and the suspected organism [1,17]. Arthrotomy is the preferred procedure for bacterial arthritis of the shoulder in infants and children.

Other indications for arthrotomy in children with bacterial arthritis may include [3,30-33]:

• Exclusion of radiolucent foreign body in children with penetrating injury
• Large amount of fibrin debris or loculations
• Debridement in patients with concomitant osteomyelitis
• Lack of clinical improvement after 48 hours of antibiotic therapy or persistent positive cultures despite appropriate antimicrobial therapy and multiple needle aspirations

Arthroscopy — Arthroscopy is an alternative surgical procedure for drainage of bacterial arthritis, particularly for bacterial arthritis of the knee [34]. As noted above, successful arthroscopic drainage of the hip also has been reported [26-28], but the number of cases successfully managed is small and largely involves older children. Successful management of septic arthritis of the shoulder joint of infants has also been described [35]. When arthroscopic drainage is performed, the surgeon must obtain a full view of the interior of the joint to ensure that no loculated pockets of pyogenic material remain after the procedure [1].

Needle aspiration — For joints other than the hip, needle aspiration (single or multiple) is an alternative to surgical drainage (arthrotomy or arthroscopy) [19,22,36,37]. However, the discomfort and fear associated with repeated needle aspiration in childhood must be considered.

If the joint is thoroughly drained during diagnostic aspiration, a single aspiration may suffice [37]. Repeat aspiration may be necessary if fluid reaccumulates. In observational studies in adult patients, the outcome of bacterial arthritis in patients treated with repeated needle aspiration was better than in those treated with surgical drainage [20,36]. Good outcome was defined by restoration of joint function with minimal or no residual pain.

Aspiration has minimal morbidity and, in the older child, may not require general anesthesia [22]. However, it may provide less satisfactory drainage than arthrotomy or arthroscopy. Because of the frequent need for repeated procedures. This technique is not favored for younger children, who may be difficult to assess and be distressed by repeated aspiration procedures.

ANTIBIOTIC THERAPY — Antimicrobial therapy is essential to the treatment of bacterial arthritis. Antibiotics should be administered as soon as possible after blood and synovial fluid cultures have been obtained [38]. (See "Bacterial arthritis: Clinical features and diagnosis in infants and children".)

Antibiotic therapy for bacterial arthritis has not been evaluated in randomized controlled trials. However, in the preantibiotic era, complications of bacterial arthritis were frequent [1]. In one series of 113 cases, only seven patients (6 percent) had a normal joint after treatment [39]. In the antibiotic era, complications are less frequent (10 to 25 percent) and appear to be related to the duration of symptoms before diagnosis and treatment [1,22]. (See 'Outcome' below.)

Empiric therapy — No randomized controlled studies have evaluated antibiotic regimens for bacterial arthritis. The empiric antimicrobial regimen should target the most common pathogens based on age (table 1), clinical features (table 2), Gram stain, and local bacterial susceptibility patterns [1,3,30,40]. (See "Bacterial arthritis: Epidemiology, pathogenesis, and microbiology in infants and children".)

Commonly used drugs are listed in the table (table 3) [3]. Empiric therapy can be altered when the susceptibility pattern of the causative bacterium is known.

We recommend that empiric therapy for bacterial arthritis in children of all ages include coverage for S. aureus. Depending upon the child's age, Gram stain, and particular clinical circumstances, coverage for additional pathogens may be necessary (table 1 and table 2).

Birth to three months — Empiric therapy for bacterial arthritis in neonates and young infants from birth to three months of age should be directed against Staphylococcus, group B streptococcus, and gram negative bacilli. (See "Bacterial arthritis: Epidemiology, pathogenesis, and microbiology in infants and children".)

Potential regimens include [5]:

• An antistaphylococcal agent (nafcillin, oxacillin, or vancomycin), PLUS
• Gentamicin or cefotaxime

Vancomycin (rather than nafcillin or oxacillin) should be included in the regimen for neonates who have been in the intensive care unit for more than one week because of the risk of nosocomial infection with methicillin-resistant S. aureus (MRSA) or coagulase-negative staphylococci. Drug doses are listed in the table (table 3).

Older than three months — Empiric therapy for bacterial arthritis in children older than three months should be directed toward S. aureus and other gram positive organisms (eg, group A streptococci, Streptococcus pneumoniae) [3,41-43]. Additional coverage for other pathogens (eg, K. kingae, Haemophilus influenzae type b [Hib], Neisseria gonorrhoeae) may be necessary in select populations. (See 'Special situations' below.)

Agents that provide antistaphylococcal and antistreptococcal coverage include [5]:

• Nafcillin (also covers Kingella kingae)
• Clindamycin (also covers MRSA and many penicillin-nonresistant S. pneumoniae, although a large proportion of serotype 19A isolates are resistant to multiple drugs, including clindamycin)
• Vancomycin (also covers MRSA and penicillin nonresistant S. pneumoniae)
• Cefazolin

Drug doses are listed in the table (table 3).

The specific choice of antistaphylococcal agent depends on the local prevalence of community-associated methicillin-resistant S. aureus (CA-MRSA) and the susceptibility of these isolates to clindamycin. (See "Treatment of Staphylococcus aureus bacteremia in children", section on 'Empiric therapy'.)

• We suggest nafcillin if <10 percent of the community S. aureus isolates are methicillin resistant.
• We suggest using either vancomycin or clindamycin when ≥10 percent of community S. aureus isolates are methicillin resistant [2,44].

• - We recommend vancomycin for life-threatening infections; we suggest vancomycin when ≥15 percent of MRSA isolates in the community also are resistant to clindamycin. (See "Overview of antibacterial susceptibility testing", section on 'D test'.)
• - We suggest clindamycin for children who lack signs of sepsis (if clindamycin resistance not a concern, based on local patterns of antimicrobial drug resistance).

Special situations — It may be necessary to add empiric coverage for pathogens other than S. aureus when clinical or laboratory factors suggest a specific pathogen (table 2). (See "Bacterial arthritis: Epidemiology, pathogenesis, and microbiology in infants and children".)

• If gram negative organisms (ie, suspicious for K. kingae) are viewed on Gram stain, coverage for gram negative organisms (eg, a second- or third-generation cephalosporin) should be added to empiric antistaphylococcal therapy. Empiric antistaphylococcal therapy should be continued because Gram stain results are subject to observer misinterpretation.
• Polymicrobial and anaerobic infections must be considered in children with direct inoculation or a history of penetrating trauma to the affected joint. Therapy should include agents directed against all organisms recovered.
• K. kingae is a consideration in children <2 to 3 years of age. Nafcillin, ampicillin, or a second- or third-generation cephalosporin) provide coverage for K. kingae [3,45].
• Hib is a consideration in children <2 years who have not been fully immunized against Hib. Coverage for Hib (eg, cefotaxime, ceftriaxone, cefuroxime) should be added to the empiric regimen for children incompletely immunized against Hib in areas where Hib immunization rates are low, but is not necessary for those in areas with high Hib immunization rates.
• S. pneumoniae is a consideration in children with underlying medical problems (table 4). Penicillin nonsusceptible S. pneumoniae is a consideration in children who received antibiotics within four weeks of diagnosis [46]. Ceftriaxone, cefotaxime, and clindamycin have been used successfully to treat penicillin nonsusceptible bone and joint infections due to S. pneumoniae.
• N. gonorrhoeae is a consideration sexually active adolescents or neonates and is treated with a third-generation cephalosporins (eg, ceftriaxone or cefotaxime) [3,5,47]. (See "Disseminated gonococcal infection" and "Gonococcal infection in the newborn".)

• Salmonella is often the cause of septic arthritis in children with sickle cell disease, and they should receive empiric therapy with a third-generation cephalosporin as part of their initial antibacterial therapy. (See "Overview of the management of sickle cell disease".)

• Enteric gram negative organisms are a consideration in patients with recent gastrointestinal surgery or complex urinary tract anatomy. Empiric gram negative arthritis may be treated with cefotaxime, ceftriaxone or ceftazidime (if concern for Pseudomonas) PLUS an aminoglycoside.
• Organisms that cause infection among injecting drug users (IDU) vary markedly between different communities. Pseudomonas aeruginosa frequently has been reported among IDUs with bacterial arthritis. If P. aeruginosa arthritis is a consideration, the combination of ceftazidime plus either nafcillin (or oxacillin) or vancomycin; this combination also covers both S. aureus and other gram negative bacilli. (See "Pseudomonas aeruginosa skin, soft tissue, and bone infections" and "Treatment of Pseudomonas aeruginosa infections".)

Specific therapy — Specific therapy is based upon culture and susceptibility results. Suggested regimens for the most commonly isolated pathogens are provided in the table (table 5). Consultation with an expert in infectious diseases is suggested for children with an inadequate response to therapy, unusual organisms, and/or drug allergies.

Negative cultures — No organism is identified in 30 to 50 percent of children with clinical bacterial arthritis [8,48-50]. Factors to be considered in the antimicrobial management of such children include age, clinical presentation, and response to empiric therapy.

• In patients who demonstrate clinical and laboratory improvement with empiric therapy, the initial empiric therapy is continued. Patients who are immunocompetent and fully immunized against Hib and S. pneumoniae may be switched to oral therapy if the oral regimen provides the same spectrum of coverage as the initial parenteral empiric regimen. (See 'Oral therapy' below.)

• In patients who do not demonstrate clinical and laboratory improvement with empiric therapy, additional infectious agents and diagnoses other than bacterial arthritis must be considered. An aggressive attempt must be made to isolate unusual pathogens and fastidious organisms such as Kingella kingae. An aspirate of any involved soft tissue or bone biopsy also should be obtained for histopathological staining and culture for bacteria. In addition, appropriate serologic testing and tuberculin skin testing should be performed when exposure to Coccidioides or Brucella species, or to M. tuberculosis is suspected. (See 'Treatment failure' below and "Bacterial arthritis: Clinical features and diagnosis in infants and children", section on 'Differential diagnosis'.)

Total duration — Bacterial arthritis is usually treated for a total of three weeks for S. aureus arthritis, and two weeks for S. pneumoniae or N. meningitidis arthritis. Antimicrobial therapy may be discontinued if the ESR and/or CRP have returned to normal by these time points and there is no radiographic evidence of unsuspected osteomyelitis [3,51]. Longer courses may be necessary for bacterial arthritis of the hip, and for arthritis caused by Enterobacteriaceae or other unusual organisms [51,52].

These treatment durations were based on studies performed in the 1970s. A multicenter open-label non-inferiority trial examined the possibility of using a shorter duration (10 days) of therapy for select patients [37].

The trial was conducted in Finland between 1983 and 2005; 130 children (aged 3 months to 15 years) with culture-positive bacterial arthritis (without adjacent osteomyelitis) were randomly assigned to 10 or 30 days of antimicrobial therapy (clindamycin or first-generation cephalosporin; amoxicillin or ampicillin was added for children aged zero to four years until the causative agent was identified) [37]. Antimicrobial therapy was administered IV for the first two to four days; the remainder of the course was completed orally. Criteria for discontinuation of antimicrobial therapy included clinical improvement and CRP <2 mg/dL, or, for patients who did not meet these criteria at the intended end of therapy, two normal CRP levels. Monitoring included serial CRP and ESR, plain radiographs at admission and on days 10 and 19, and clinical evaluation at 2 weeks, 3 months, and ≥1 year after hospitalization. The following results were noted:

• Percutaneous aspiration was performed in 110 patients; seven of these received needle lavage.
• Surgical arthrotomy was only performed in 7 of the 48 cases of hip joint arthritis.
• 7 of 48 cases (15 percent) of hip joint infection required therapy to be extended beyond the planned duration due to slow resolution of clinical or laboratory abnormalities.
• 5 of 63 patients (8 percent) assigned to short-term therapy required continuation of antibiotics for up to 21 days (four with hip involvement, one with knee involvement) due to slow resolution of clinical or laboratory abnormalities.
• One child in the long-term treatment group had two late-onset osteoarticular infections. No other patients experienced relapse, recrudescence, residual dysfunction, growth disturbance, or other clinically significant sequelae.

As noted by an editorialist [53], there was a marked difference in the microbiology of these cases and those currently seen in the United States. The distribution of pathogens was as follows: methicillin-sensitive S. aureus (MSSA, 58 percent), Hib (18 percent), S pyogenes (12 percent), S. pneumoniae (8 percent), other pathogens (3 percent). There were no cases of MRSA.

In view of the relatively high failure rate of the randomized therapy in septic arthritis of the hip, the absence of cases of MRSA, and the paucity of data from infants (in whom septic arthritis of the hip is frequently associated with sequelae), we do not recommend the use of the short-term regimen described by these authors.

Route — Initial antimicrobial therapy for bacterial arthritis is usually administered parenterally. In neonates (<1 month), we suggest that the entire course of antimicrobial therapy be administered parenterally [3]. Although successful oral therapy for bacterial arthritis in neonates has been described [5], the gastrointestinal absorption of oral antibiotics in neonates is unpredictable [54].

In infants and children older than one month, parenteral therapy is continued at least until clinical and laboratory improvement have been demonstrated, after which the balance of antibiotic therapy can be administered orally [55]. (See 'Response to therapy' below and 'Oral therapy' below.)

Subsequent outpatient therapy — Antibiotic therapy can be completed on an outpatient basis in patients with unequivocal clinical improvement. Clear improvement usually is established 5 to 10 days after initiation of antibiotic therapy. It is indicated by achievement of each of the following conditions:

• The patient has been afebrile for 48 to 72 hours
• Local signs and symptoms of infection are reduced considerably
• WBC has normalized
• CRP and/or ESR has decreased

Intravenous therapy — Outpatient parenteral antimicrobial therapy (OPAT) may be necessary if any of the following apply:

• The patient's isolate is not susceptible to oral antibiotics
• Adherence to an oral regimen appears unlikely
• The patient cannot tolerate the oral agent needed

OPAT also may be necessary for patients with negative cultures if there is not an oral regimen with a spectrum similar to the parenteral regimen that resulted in improvement.

Insertion of a percutaneously inserted central catheter (PICC) facilitates prolonged administration of parenteral antibiotics and can be used for outpatient therapy [10]. However, catheter-related complications occur in 30 to 40 percent of children with osteoarticular infections who are treated with OPAT [56,57]. The risks and benefits of prolonged outpatient antimicrobial therapy must be carefully weighed when considering this option [56].

Oral therapy — Oral antibiotics may be used to complete the course of therapy for bacterial arthritis in infants and children older than one month. As long as the patient has demonstrated unequivocal improvement, successful treatment does not require a minimum duration of parenteral therapy. In randomized and observational studies, treatment of bacterial arthritis with IV antibiotics for short periods (approximately seven days) followed by oral therapy, was as successful as longer courses of parenteral therapy [58-60].

Subsequent oral therapy may be substituted for parenteral treatment provided that certain criteria are met [3,10,12,51,58-61]:

• Clear demonstration of clinical and laboratory improvement (ie, swelling, tenderness, and erythema)
• Decreasing or absent fever
• An oral agent with appropriate coverage is available in a formulation that the child can swallow
• Adherence to the antibiotic and monitoring regimen are assured

• Choice of oral agent — The choice of agent(s) for oral therapy depends upon whether an organism was isolated from the synovial fluid, blood, or other culture. If an organism is isolated, the susceptibility pattern can be used to determine an appropriate oral drug. Additional considerations include the bioavailability and palatability of the oral medication.

When an organism is not isolated, oral therapy should be directed toward the most likely pathogen(s) given the child's age and clinical presentation (table 1 and table 2). The chosen regimen should have a spectrum similar to that provided by the parenteral therapy that was associated with improvement. As examples, a child who improved on cefazolin may be switched to cephalexin; a child who improved on parenteral clindamycin could be switched to oral clindamycin. Consultation with an expert in infectious diseases may be helpful in choosing the optimal oral agent, particularly if the child has an unusual pathogen or allergy to antibiotics.

Antibiotics administered orally for bacterial arthritis are generally given in higher doses than those used for treatment of other infections and recommended in package inserts (table 6) [62,63]. To achieve adequate serum levels, doses of beta-lactams (penicillins and cephalosporins other than cefixime) often can be increased to 200 mg/kg per day without serious adverse side effects. Diarrhea, a frequent complication of high-dose oral beta-lactam therapy, can be mitigated by using the doses indicated in the table (table 6) and adding probenecid for children older than two years of age (40 mg/kg per day in 4 divided doses, maximum of 2 grams per day). (See 'Drug monitoring' below.)

The initial doses of oral therapy should be administered while the child is in the hospital to ensure that the drug is tolerated. The child may be discharged home when he or she is able to tolerate oral antibiotics and it is clear that the caregiver can administer oral therapy as prescribed. (See 'Follow-up' below.)

Drug monitoring

Adequacy of therapy — Expert opinion differs regarding the routine use of drug monitoring for patients on oral therapy for bacterial arthritis. Some have suggested monitoring of serum bactericidal titers [58,64]. However, others suggest that if high-dose therapy with clindamycin or cephalosporin is used, obtaining drug levels is not necessary [10,65].

• Antibiotic levels — Antibiotic levels in children taking oral medications can be used to ensure that the child is able to absorb the medication enterally. Assays are available for most of the antibiotics that are used routinely to treat bacterial arthritis (eg, cephalexin and dicloxacillin). Antibiotic levels may be helpful in select patients if there is a concern regarding adequate absorption or compliance.
• Serum bactericidal titer — The SBT is the highest dilution of the patient's serum that will, after 18 hours of incubation, kill 99.9 percent of an inoculum of the infecting organism. We do not suggest using SBT to monitor drug therapy in children being transitioned to oral therapy for septic arthritis.

Adverse effects — Children who are being treated with oral or parenteral antibiotic therapy for bacterial arthritis require monitoring for potential adverse effects. Adverse effects of high-dose antibiotic therapy may include pancytopenia, leukopenia, impaired liver or renal function, antibiotic associated diarrhea, and pseudomembranous enterocolitis [17].

The monitoring schedule varies depending upon the drug, dose, and route of therapy. We obtain ESR, CRP, complete blood count (CBC), and biochemical profile, including serum aminotransferases weekly initially, while the patient is receiving beta-lactam antibiotics. (See 'Follow-up' below.)

ADJUNCTIVE THERAPIES — Adjunctive therapies for bacterial arthritis in children may include analgesia and physical therapy. Adjunctive therapy with glucocorticoids remains investigational.

Analgesia — Pain management is an important aspect of therapy for bacterial arthritis. Opioid therapy (codeine or morphine) may be necessary; consultation with the pain management service may be helpful. Once they are discharged home, children may receive acetaminophen or ibuprofen for pain control [10].

Physical therapy — Attention must be paid to joint position and rapid mobilization to prevent contractures and promote optimal nutrition to the articular cartilage. Physical therapy may be helpful in children who are reluctant to use the joint [32]. Once discharged home, the child may require a wheelchair or walker and continued physical therapy.

Antiinflammatory agents — The inflammatory response is an important component of the pathogenesis of bacterial arthritis and accounts for at least a portion of the long-term morbidity. Some experts recommend routine non-steroidal anti-inflammatory agents, which decrease reactive synovitis associated with bacterial arthritis.

The use of adjunctive glucocorticoid therapy to decrease inflammation in bacterial arthritis has been evaluated in animal models and in a randomized trial in children, as illustrated below [4,11,66].

• In an experimental model, mice that received glucocorticoid and antistaphylococcal antibiotic three days after inoculation with S. aureus had a decreased incidence and severity of S. aureus arthritis compared to those that received antistaphylococcal antibiotic alone [66].
• In a randomized trial in 123 children (average age of five years) with suspected hematogenous bacterial arthritis, adjunctive dexamethasone during the first four days of therapy appeared to decrease the risk of residual dysfunction 12 months after treatment (2 versus 26 percent) [11]. About 85 percent of the children had hip or knee infection; H. influenzae infection accounted for about 10 percent of cases in both the experimental and control groups.

Although these findings are encouraging, adjunctive glucocorticoid therapy remains controversial and investigational.

RESPONSE TO THERAPY — Clinical and laboratory improvement is indicated by:

• Resolution of fever
• Decreased joint pain, swelling, and erythema
• Increased joint mobility
• Decrease in peripheral WBC count and synovial fluid WBC (if repeated aspirations are performed or a drain is left in place after surgical drainage)
• Decrease in ESR and/or CRP

Monitoring response — The response to therapy can be assessed by serial clinical examination (fever, pain, range of motion), peripheral WBC, ESR and/or CRP, and synovial fluid WBC and culture.

Improvement in joint symptoms usually occurs within two days of drainage and initiation of antimicrobial therapy. Fever usually resolves within three to five days [10,58]. The time required for resolution of symptoms and sterilization of the joint fluid are proportional to the duration of symptoms before initiation of appropriate therapy and the synovial fluid WBC at the time of diagnosis [67-69].

CRP is more useful than ESR for monitoring acute response to treatment. Even with appropriate treatment, the ESR may continue to rise for three to five days, whereas CRP peaks within 36 to 50 hours of onset of infection and quickly falls to normal with appropriate therapy [17]. ESR remains useful for determining the duration of therapy since it normalizes as inflammation resolves.

In joints that initially were drained with needle aspiration, synovial fluid may reaccumulate requiring repeated drainage. Infected knees, in particular, may continue to accumulate fluid for 7 to 10 days. When available (from repeat aspiration or a surgical drain), serial synovial fluid analyses should demonstrate sterilization and a decreasing WBC count. If repeat synovial fluid WBC count remains elevated and the culture remains positive, more definitive joint drainage and/or an alteration in the antimicrobial regimen are warranted.

Treatment failure — Treatment failure may be indicated by: lack of clinical improvement; ESR that continues to rise three to five days after initiation of therapy; persistently elevated ESR, CRP, peripheral WBC, or synovial fluid WBC; or failure to sterilize the synovial fluid in the expected period of one to two days if repeat cultures are obtained.

Patients who do not respond to treatment as expected require reevaluation, adjustment of antimicrobial therapy, and definitive surgical drainage.

Reevaluation may include magnetic resonance imaging for evidence of osteomyelitis with abscess, and review of the history for exposure to unusual organisms (eg, Brucella, M. tuberculosis) or mechanism for direct inoculation, and reconsideration or other conditions in the differential diagnosis. Organisms not covered by the initial empiric regimen (eg, K. kingae in patients initially treated with vancomycin or clindamycin) also must be considered. (See "Bacterial arthritis: Clinical features and diagnosis in infants and children", section on 'Differential diagnosis' and "Bacterial arthritis: Epidemiology, pathogenesis, and microbiology in infants and children", section on 'Microbiology' and 'Drainage' above.)

FOLLOW-UP

During therapy — Patients and families who can comply with the oral regimen and close follow-up may be discharged from the hospital and managed as outpatients once a clear clinical response has been established. Children being treated as outpatients for bacterial arthritis should be seen for follow-up approximately one week after discharge from the hospital, and at one to two week intervals thereafter.

The follow-up visit should include monitoring for continued clinical improvement (including the need to initiate or continue physical therapy) and for complications related to high-dose antibiotic therapy, such as cytopenia, antibiotic associated diarrhea, and pseudomembranous colitis. (See 'Adverse effects' above.)

We suggest that a CBC, ESR, CRP, and biochemical profile, including serum transferases, be obtained at each visit. If patients are not improving as expected after being transitioned to oral therapy, it may be advisable to revert to intravenous therapy. (See 'Adequacy of therapy' above.)

Radiographs — Radiographs should be obtained two to three weeks into the course of treatment for bacterial arthritis to look for bone changes indicative of occult osteomyelitis or osteomyelitis of the epiphysis or metaphysis [70].

Radiographic findings suggestive of osteomyelitis include lytic lesions (picture 1) and periosteal elevation (picture 2), thickening, or new bone formation [5]. Osteomyelitis may require surgical intervention or prolonged antibiotic therapy. Bacterial arthritis with concomitant osteomyelitis is associated with a worse prognosis than bacterial arthritis alone [70]. (See "Evaluation and diagnosis of hematogenous osteomyelitis in children", section on 'imaging' and "Treatment of hematogenous osteomyelitis in children" and 'Prognostic factors' below.)

Long-term follow-up — Long-term follow-up after treatment for bacterial arthritis in infants and children includes monitoring of joint dysfunction and limb-length discrepancy. Severely affected patients should undergo regular imaging after six years of age if progressive limb-length discrepancy occurs [12].

OUTCOME

Complications — Potential complications of bacterial arthritis include [1,5,12,71-73].

• Joint laxity, subluxation, or dislocation
• Joint restriction
• Limb length discrepancy (if the growth plate is involved)
• Avascular necrosis
• Enlargement of the femoral head (coxa magna) in bacterial arthritis of the hip

The reported complication rates vary depending upon the patient population, the involved joint, and the duration of follow-up; prolonged follow-up may be necessary to determine the true extent of injury [5,49,74]. The estimated rate of residual dysfunction is 10 to 25 percent [22]. Even with appropriate management, approximately 40 percent of patients with hip involvement and 10 percent of patients with knee involvement develop significant complications [18,32,33].

Prognostic factors — Factors related to poor outcome have been identified in observational studies. They include [24,48,49,52,75-79]:

• Duration of symptoms before treatment, particularly if >4 to 7 days
• Involvement of the hip
• Involvement of the hip or shoulder with concomitant osteomyelitis
• Young age (ie, less than one year)
• The causative organism (Enterobacteriaceae and S. aureus are associated with more frequent sequelae)

Factors that do not appear to be related to outcome include:

• The mode of drainage provided that adequate drainage is obtained
• The choice of antibiotic provided that it is effective against the infecting organism

SUMMARY AND RECOMMENDATIONS

Overview

• Bacterial arthritis requires prompt recognition and management. Delays in treatment are associated with long-term sequelae. This is especially true when the hip is involved. (See 'Overview' above.)

• The goals of treatment include sterilization and decompression of the joint space and removal of inflammatory debris to relieve pain and prevent deformity or functional sequelae. Drainage of joint fluid and antimicrobial therapy are the cornerstones of therapy. (See 'Overview' above.)

• Because delayed therapy is associated with long-term sequelae, treatment of infants and children with suspected bacterial arthritis should begin immediately after blood and synovial fluid cultures are obtained. (See 'Indications' above.)

Drainage of joint fluid

• We recommend drainage of the joint space in all patients with bacterial arthritis (Grade 1B). Drainage and lavage are necessary to decompress the joint space and to remove inflammatory debris. Drainage can be accomplished through open surgery (arthrotomy), arthroscopy, or needle aspiration. Decisions regarding the optimal drainage procedure for an individual patient should be made on a case-by-case basis, depending upon the site and extent of involvement, duration of symptoms, and the suspected organism. (See 'Drainage' above.)
• We recommend surgical drainage (arthrotomy or arthroscopy) for bacterial arthritis of the hip in infants and children (Grade 1C). We suggest arthrotomy as the procedure of choice (Grade 2C). (See 'Drainage' above.)

• Additional indications for arthrotomy in children with bacterial arthritis may include: suspected radiolucent foreign body in children with penetrating injury; large amount of fibrin debris or loculations; debridement in patients with concomitant osteomyelitis; and lack of clinical improvement or persistent positive cultures despite appropriate antimicrobial therapy and multiple needle aspirations. (See 'Arthrotomy' above.)

• For joints other than the hip, needle aspiration is an alternative to surgical drainage. (See 'Needle aspiration' above.)

Antibiotics

• Antibiotic therapy is necessary to sterilize the joint fluid. Antibiotics should be administered as soon as possible after blood and synovial fluid cultures have been obtained. (See 'Antibiotic therapy' above.)

• Initial antimicrobial therapy for bacterial arthritis is administered parenterally. The empiric regimen should target the most common pathogens in a particular patient. (See 'Empiric therapy' above.)

• We recommend that empiric therapy include coverage for Staphylococcus aureus in all infants and children (Grade 1B). Depending upon the child's age (table 1), Gram stain, and particular clinical circumstances (table 2), empiric coverage for additional pathogens may be necessary. Doses for commonly used drugs are listed in the table (table 3). Empiric therapy can be altered when the susceptibility pattern of the causative bacterium is known. (See 'Empiric therapy' above.)

• Specific therapy is based upon culture and susceptibility results. Suggested agents for the most commonly isolated pathogens are provided in the table (table 5). (See 'Specific therapy' above.)

• Bacterial arthritis is usually treated for a total of three weeks for S. aureus arthritis, and two weeks for S. pneumoniae or N. meningitidis arthritis. Antimicrobial therapy may be discontinued if the erythrocyte sedimentation rate (ESR) and/or C-reactive protein (CRP) have returned to normal by these time points and there is no radiographic evidence of unsuspected osteomyelitis. (See 'Total duration' above.)
• In neonates (<1 month), we suggest that the entire course of antimicrobial therapy be administered parenterally (Grade 2C). In infants and children older than one month, parenteral therapy is continued at least until clinical and laboratory improvement have been demonstrated, after which the balance of antibiotic therapy can be administered orally. In infants an children older than one month, oral therapy may be initiated if the following criteria are met:

• - Clear demonstration of clinical and laboratory improvement
• - Decreasing or absent fever
• - An oral agent with appropriate coverage is available in a formulation that the child can swallow
• - Adherence to the antibiotic and monitoring regimen are assured

Doses for commonly used oral drugs are listed in the table (table 6). (See 'Route' above and 'Oral therapy' above.)

Treatment response

• The adequacy of antibiotic therapy can be assessed by serial clinical examination (fever, pain, range of motion), peripheral white blood cell (WBC) count, ESR and/or CRP, and synovial fluid WBC count and culture. (See 'Monitoring response' above.)

• Patients who do not respond to treatment as expected require reevaluation, adjustment of antimicrobial therapy, and definitive surgical drainage. (See 'Treatment failure' above.)

Follow-up and outcome

• Children being treated as outpatients for bacterial arthritis should be seen for follow-up approximately one week after discharge from the hospital, and at one to two week intervals thereafter. (See 'During therapy' above.)

• The follow-up visit should include monitoring for continued clinical improvement and for complications related to high-dose antibiotic therapy. We suggest that a complete blood count, ESR, CRP, and biochemical profile, including serum aminotransferases, be obtained at each visit. (See 'During therapy' above.)

• Radiographs should be obtained two to three weeks into the course of treatment to look for bone changes indicative of occult osteomyelitis or osteomyelitis of the epiphysis or metaphysis. (See 'Radiographs' above.)

• The estimated rate of residual joint dysfunction is 10 to 25 percent; even with appropriate treatment, approximately 40 percent of patients with hip involvement and 10 percent of patients with knee involvement develop significant complications. Factors related to poor outcome include: duration of symptoms before treatment; involvement of the hip: involvement of the hip or shoulder with concomitant osteomyelitis; and age younger than one year. (See 'Outcome' above.)