Nonvertebral osteomyelitis in adults: Treatment

INTRODUCTION — Treatment of osteomyelitis includes consideration of issues related to debridement, management of infected foreign bodies (if present), antibiotic selection, and duration of therapy; these issues are discussed in the following sections.

General issues related to treatment of osteomyelitis are discussed here. Issues related to clinical manifestations and diagnosis of osteomyelitis are discussed separately. (See "Nonvertebral osteomyelitis in adults: Clinical manifestations and diagnosis".)

Issues related to treatment of vertebral osteomyelitis, osteomyelitis associated with trauma, pelvic and sacral osteomyelitis, and prosthetic joint infection are discussed in detail separately. (See "Vertebral osteomyelitis and discitis in adults" and "Osteomyelitis associated with open fractures in adults" and "Pelvic osteomyelitis and other infections of the bony pelvis in adults" and "Prosthetic joint infection: Treatment".)

Issues related to osteomyelitis in children are presented separately. (See "Hematogenous osteomyelitis in children: Evaluation and diagnosis" and "Hematogenous osteomyelitis in children: Management".)

CLINICAL APPROACH

Nonhematogenous osteomyelitis

Absence of orthopedic hardware — In general, management of osteomyelitis (in the absence of orthopedic hardware) consists of operative debridement followed by antimicrobial therapy for eradication of infection.

There may be a role for antimicrobial therapy alone (in the absence of surgical debridement) for carefully selected patients with diabetic foot osteomyelitis. In observational studies, success rates for treatment of diabetic foot osteomyelitis with antimicrobial therapy alone range from 64 to 82 percent [1-3]. In one small randomized trial including 37 patients treated with antibiotics alone for three months or surgical debridement and antibiotics for ten days, there was no difference in healing between the two groups [4]. However, patients with common comorbid conditions (such as arterial disease, Charcot arthropathy, and poorly controlled diabetes) were excluded. The role of rifampin as adjunctive therapy to reduce the likelihood of amputation in diabetic foot osteomyelitis remains unclear; meta-analyses and large randomized clinical trials are underway [5-7].

Surgical debridement — Surgical debridement for management of osteomyelitis consists of removal of necrotic material and culture of involved tissue and bone. A satisfactory soft tissue envelope overlying the site of infection must be re-established for successful treatment, either via direct closure or flap coverage.

In the setting of dead or injured bone and/or infected fluid collections, surgical debridement (and revascularization, in some cases) is a cornerstone of therapy. Antimicrobial therapy alone is not effective for cure of infected, necrotic bone. In addition, antibiotic penetration into bone may be unreliable in patients with arterial insufficiency or prior scarring associated with trauma [8].

Surgical debridement also allows placement of local antimicrobials, either applied directly to the site of infection or mixed with absorbable (eg, calcium sulfate) or nonabsorbable (eg, polymethyl methacrylate cement) carriers. (See "Management of diabetic foot ulcers" and "Osteomyelitis associated with open fractures in adults".)

Patients who have received antibiotics recently and do not have an acute need for surgical intervention should discontinue antibiotics for at least two weeks prior to debridement to optimize microbiologic diagnosis. However, patients with necrotizing soft tissue infection or with systemic infection secondary to osteomyelitis in whom source control is needed should undergo early surgical debridement.

Antibiotic therapy — The clinical approach to antibiotic therapy is discussed in this section; issues related to selection of antibiotic therapy are discussed below. (See 'Antibiotic therapy' below.)

Residual infected bone present — Patients with residual infected bone that is not amenable to complete removal should be treated with a prolonged duration of intravenous or highly bioavailable oral antibiotic therapy, guided by antimicrobial susceptibility data. (See 'Antibiotic therapy' below.)

The optimal duration of antibiotic therapy for treatment of osteomyelitis with residual infected bone is uncertain. Most experts favor continuing antimicrobial therapy at least until debrided bone has been covered by vascularized soft tissue, which is usually at least six weeks from the last debridement [9]. Further studies are ongoing to define whether a shorter duration of therapy is equally effective [10-12]. Parenteral antimicrobial therapy can be administered on an outpatient basis via a peripherally inserted central catheter or similar device. In some circumstances, highly bioavailable oral antibiotic therapy may be administered.

While on parenteral antimicrobial therapy, patients should have weekly bloodwork for safety monitoring. (See 'Monitoring during treatment' below.)

No residual infected bone — Patients who undergo amputation or complete removal of all involved bone warrant a relatively short course of antibiotic therapy (see 'Antibiotic therapy' below). In the absence of concomitant soft tissue infection, antibiotic therapy may be discontinued as early as two to five days after debridement. When there is evidence of soft tissue infection at the operative site, 10 to 14 days of pathogen-directed parenteral or highly bioavailable oral therapy is reasonable [8].

Presence of orthopedic hardware

Surgical debridement — In general, the approach to management of osteomyelitis associated with orthopedic hardware consists of operative debridement, obtaining bone biopsy for culture, and antimicrobial therapy tailored to culture and susceptibility data [13-15]. The need for hardware to maintain bone stability must be balanced with the potential impact of foreign material on propagation of infection. Additional factors influencing the approach include the microbiology of infection and individual patient circumstances.

Surgical management strategies include debridement with hardware retention or debridement with hardware removal. Hardware retention may be attempted when the stability of the bone and hardware construct would be compromised (such as in the case of fracture fixation hardware with unhealed fracture) or when there is a clear anatomic separation between the osteomyelitis and hardware. In contrast, hardware should be removed if the hardware is no longer needed for bone stability or if adequate debridement of the infected bone cannot be achieved with hardware retention.

Antibiotic therapy — The clinical approach to antibiotic therapy is discussed in this section; issues related to selection of antibiotic therapy are discussed below. (See 'Antibiotic therapy' below.)

Retained hardware — Patients with orthopedic hardware that is not amenable to removal and/or residual involved bone that is not amenable to complete debridement should be treated with a prolonged duration of intravenous antibiotic therapy (see 'Antibiotic therapy' below). The optimal duration is uncertain; most experts favor six weeks of therapy. While on parenteral antimicrobial therapy, patients should have weekly bloodwork for safety monitoring. (See 'Monitoring during treatment' below.)

Following completion of parenteral therapy with resolution of clinical signs of acute infection, patients with retained hardware and/or necrotic bone not amenable to debridement should receive long-term antibiotic suppression with an oral agent, guided by antimicrobial susceptibility data. Regimens for antibiotic suppression are summarized in the table. (See 'Suppressive therapy' below.)

In general, oral antimicrobial suppression should be continued until fractures are united. Once fracture healing is demonstrated radiographically, the timeframe for discontinuation of oral antimicrobial suppression should be determined carefully. Factors influencing the duration of therapy include the microbiology of infection, the duration of infection prior to debridement, the tolerability of the antimicrobial suppression regimen, the status of the orthopedic hardware (if present) at the site of infection, and individual patient circumstances.

Patients who would be unable to tolerate additional surgery in the setting of relapse may warrant antimicrobial suppression for as long as the hardware remains in place. However, the benefit of continuing suppressive treatment for longer than six months is uncertain. In one observational study including 89 patients with retained orthopedic hardware, use of suppressive antibiotics for at least three months after diagnosis was associated with being free of clinical infection (odds ratio 3.5; 95% CI 1.3-9.4), but use of suppressive antibiotic for at least six months after diagnosis was not [16].

Among patients who remain on antimicrobial suppression after fracture union, we discuss with the patient the benefits and adverse effects of ongoing suppression compared with discontinuing suppression. If hardware removal could be performed in the event of infection relapse (such as healed long-bone fracture), the infection appears to be well suppressed, and the patient is comfortable with the small possibility of further surgery, we offer discontinuation of suppression. If this small possibility of surgery is unacceptable, we continue suppression indefinitely.

No retained hardware — Patients with no retained hardware should complete a prolonged duration of intravenous or highly bioavailable oral antibiotic therapy (see 'Antibiotic therapy' below). Most experts favor continuing antimicrobial therapy at least until debrided bone has been covered by vascularized soft tissue, which is usually at least six weeks from the last debridement.

While on parenteral antimicrobial therapy, patients should have weekly bloodwork for safety monitoring. (See 'Monitoring during treatment' below.)

Monitoring during treatment — Laboratory monitoring is needed during prolonged administration of antimicrobial therapy to monitor for adverse drug effects and to assess control of infection.

For patients on parenteral antimicrobial therapy, we obtain weekly complete blood count and chemistries. We obtain serum inflammatory markers (erythrocyte sedimentation rate [ESR] and C-reactive protein [CRP]) at the beginning and end of parenteral therapy and at the time of transition to oral suppressive therapy (if used).

We do not routinely monitor weekly serum inflammatory markers during parenteral antimicrobial therapy. However, if there is clinical suspicion for treatment failure, we use inflammatory markers (in conjunction with clinical examination and radiographic studies such as magnetic resonance imaging or plain radiograph) to guide further management.

For patients on oral suppressive antimicrobial therapy, we obtain a complete blood count, creatinine, and alanine aminotransferase more frequently at the beginning and every 6 to 12 months thereafter. Suppressive antimicrobials with higher risk of adverse reactions (eg, trimethoprim-sulfamethoxazole) may require more frequent monitoring.

Evaluation at the end of treatment — Patients should undergo clinical evaluation at the end of the planned course of therapy. The site of infection should be examined for healing of the wound and soft tissue envelope. Patients should be questioned regarding systemic symptoms of infection as well as pain (type and severity).

We typically perform serum inflammatory markers (ESR and CRP), which are useful to confirm response to antimicrobial therapy as well as to serve as a new baseline for future evaluation.

We do not routinely perform radiographic imaging; frequently, residual inflammatory changes may be mistaken for persistent infection. The decision to pursue radiographic imaging should be guided by clinical suspicion for relapsing infection (as evidenced by worsening symptoms and/or rising inflammatory markers).

In the setting of persistently elevated inflammatory markers after completing a course of antibiotic therapy, the first priority should be to ensure that a thorough and complete debridement has been performed. In addition, the microbiologic diagnosis and susceptibility data should be reviewed.

Persistently elevated inflammatory markers two weeks following completion of antimicrobial therapy (without an alternative explanation) should prompt concern for persistent osteomyelitis [17,18]. Patients with associated symptoms warrant repeat debridement and additional antimicrobial therapy. In the absence of clinical signs consistent with persistent infection, clinical observation may be reasonable.

Hematogenous osteomyelitis — In adults, hematogenous osteomyelitis most commonly involves the vertebral bones. Issues related to vertebral osteomyelitis are discussed separately. (See "Vertebral osteomyelitis and discitis in adults".)

Treatment of nonvertebral hematogenous osteomyelitis consists of parenteral antibiotics; in some circumstances, surgical debridement is also warranted.

In general, patients with infection confined to the medullary canal of the bone may be treated with antibiotics alone. Surgical debridement is warranted in patients with subperiosteal collection or abscess, necrotic bone, and/or in the setting of concomitant joint infection. Depending on the scope of the debridement, bone grafting or other orthopedic reconstruction may be required. A critical component of surgical management is adequate soft tissue coverage.

In all patients, blood cultures should be obtained prior to initiation of antibiotic therapy. If blood cultures are negative, a bone biopsy or aspirate of subperiosteal abscess for culture should be performed.

Empiric antibiotic coverage for treatment of hematogenous osteomyelitis should include activity against methicillin-resistant Staphylococcus aureus and aerobic gram-negative bacilli. An appropriate regimen consists of vancomycin and a third- or fourth-generation cephalosporin. (See 'Antibiotic therapy' below.)

Once an etiologic organism is isolated, the antibiotic regimen should be tailored to susceptibility data [19].

The optimal duration of antibiotic therapy for treatment of hematogenous (nonvertebral) osteomyelitis is uncertain. In general, at least four weeks of parenteral therapy from the last major debridement (if performed) are warranted.

For patients on parenteral antimicrobial therapy, we obtain weekly complete blood count and chemistries. In addition, we obtain serum inflammatory markers (erythrocyte sedimentation rate and C-reactive protein) at the beginning and end of parenteral therapy and at the time of transition to oral suppressive therapy (if used).

ANTIBIOTIC THERAPY — Whenever possible, initiation of antibiotic therapy should be delayed until bone cultures can be obtained. Patients who have received antibiotics recently and do not have an acute need for surgical intervention should discontinue antibiotics for at least two weeks prior to debridement to optimize microbiologic diagnosis.

Suggested antibiotic regimens are outlined in the table (table 1) [20-25]. Antibiotic therapy should be tailored to culture and susceptibility findings when available.

Systemic therapy — Management of complex orthopedic infections usually includes a prolonged course of intravenous (IV) antibiotic therapy; data regarding use of oral antibiotic therapy are emerging.

In one randomized trial including more than 1000 patients with bone or joint infection treated with IV or oral antibiotic therapy for at least six weeks (median duration of therapy 78 and 71 days, respectively), the rates of treatment failure within one year were comparable (14 versus 13 percent) [26]. A variety of orthopedic infections were included, including osteomyelitis with or without an associated implant; surgical debridement occurred in 92 percent of cases. Based on these data, among carefully selected patients who undergo surgical debridement and have infection caused by an organism that is susceptible to oral therapy, a course of oral antimicrobial therapy (preferably with highly bioavailable agents, perhaps with extended duration up 12 weeks) may be a reasonable consideration in some circumstances. Subsequent to the publication of this trial, there is a report of the implementation of the OVIVA trial findings in a single hospital [27]. We believe the generalizability of the findings in the original trial and subsequent studies continue to be limited by the heterogeneous patient population and open-label design, so the results should be interpreted carefully when applied to individual patients [28,29].

Empiric therapy — Empiric treatment of osteomyelitis should consist of antimicrobial therapy with activity against methicillin-resistant S. aureus (MRSA) and gram-negative organisms. Reasonable regimens include vancomycin in combination with a third- or fourth-generation cephalosporin. We avoid combined use of vancomycin with piperacillin-tazobactam, given the risk of nephrotoxicity with this combination [30]. Antibiotic therapy should be tailored to culture and susceptibility data when available.

Definitive therapy

Staphylococci

Staphylococcus aureus — In general, definitive therapy for treatment of osteomyelitis due to S. aureus consists of parenteral antibiotic therapy; regimens are summarized in the table (table 1). In addition, use of adjunctive agents for treatment of S. aureus osteomyelitis may be warranted in some circumstances. (See 'Use of adjunctive agents' below.)

Antibiotics for treatment of osteomyelitis due to methicillin-susceptible S. aureus (MSSA) include oxacillin, nafcillin, and cefazolin (table 1). Use of ceftriaxone for treatment of staphylococcal osteomyelitis is not universally accepted; in our practice, we use it as an alternative regimen for isolates with oxacillin minimum inhibitory concentration <0.5 mcg/mL in the absence of concomitant bacteremia, since once-daily dosing is a convenient outpatient regimen [31,32]. Issues related to duration of antibiotic therapy are discussed above. (See 'Clinical approach' above.)

The antibiotic of choice for treatment of osteomyelitis due to MRSA is vancomycin (table 1); it is the antibiotic agent for which there is the greatest cumulative clinical experience, although there are no controlled trials [33,34]. The Infectious Diseases Society of America (IDSA) guidelines recommend a minimum of eight weeks of antibiotic therapy for treatment of MRSA osteomyelitis [33]; we favor a six-week course of therapy if adequate debridement has been performed. (See 'Clinical approach' above.)

Acceptable alternative agents to vancomycin for treatment of osteomyelitis due to MRSA include daptomycin (and teicoplanin, where available) [35-37]. An observational study including more than 400 patients with osteomyelitis due to gram-positive pathogens noted success rates of 80 percent with daptomycin for treatment of staphylococcal osteomyelitis [37]. The optimal dosing of daptomycin for treatment of osteomyelitis is uncertain; dosing used in the preceding study ranged from 6 to 10 mg/kg/day.

Antibiotic agents that warrant further study for treatment of staphylococcal osteomyelitis include ceftaroline, telavancin, and dalbavancin [38-44]. Ceftaroline has been used successfully for treatment of osteomyelitis but has been associated with increased risk of neutropenia [39,45]. Pending further study, use of ceftaroline should be reserved for patients unable to receive other therapies [39]. Ceftaroline susceptibility must be confirmed; the prevalence of resistance is high in some regions [46]. Use of telavancin for treatment of osteomyelitis should be deferred pending further study to evaluate safety, given increased risk of adverse events observed with its use for treatment of other conditions [38,40,47]. Dalbavancin has the unique advantage of weekly intravenous dosing [41]. It has been used successfully for treatment of osteomyelitis, but should be reserved for use in patients in whom other therapies are not an option and in whom the organism is susceptible, until further data are available [41-44,48].

We do not favor use of trimethoprim-sulfamethoxazole, linezolid, tedizolid, clindamycin, fluoroquinolones, quinupristin-dalfopristin, or tigecycline for definitive therapy of osteomyelitis due to S. aureus. A regimen of trimethoprim-sulfamethoxazole with rifampin is advocated in some treatment guidelines [33] but not others [49]; data to support its use are limited [18,50,51]. Prolonged use of linezolid is limited by bone marrow toxicity, peripheral and optic neuropathy, lactic acidosis, and drug-drug interactions [52-54]. Tedizolid has not been studied sufficiently in osteomyelitis. Prolonged use of clindamycin has been associated with antibiotic-associated diarrhea and Clostridioides difficile infection. Fluoroquinolones have good bone penetration but are associated with inducible resistance to S. aureus when used as monotherapy.

Coagulase-negative staphylococci — Most coagulase-negative staphylococci are methicillin resistant (an exception is Staphylococcus lugdunensis, which is almost universally methicillin susceptible). (See "Staphylococcus lugdunensis".)

Empiric treatment for osteomyelitis due to coagulase-negative staphylococci should consists of an agent with activity against methicillin-resistant staphylococci. If susceptibility testing demonstrates methicillin susceptibility, a beta-lactam with activity against methicillin-susceptible staphylococci should be used (table 1).

In addition, use of adjunctive agents for treatment staphylococcal osteomyelitis may be warranted in some circumstances. (See 'Use of adjunctive agents' below.)

Use of adjunctive agents — Agents that may be used adjunctively in the setting of definitive therapy for treatment of staphylococcal osteomyelitis include rifampin and fusidic acid.

We are in agreement with some experts who favor use of rifampin (in combination with at least one other anti-staphylococcal agent) for its activity against microorganisms in biofilms, given the importance of biofilms in the pathophysiology of staphylococcal osteomyelitis (particularly in the setting of hardware) [33,55,56]. However, others oppose use of adjunctive rifampin given limited evidence for improved outcomes over standard antimicrobial therapy.

Clinical circumstances in which rifampin may be most useful include osteomyelitis in the setting of prosthetic material and osteomyelitis in which therapeutic options are limited. Data may be extrapolated from literature on treatment of prosthetic joint infection to support use of rifampin in the treatment of infections involving other types of orthopedic hardware [57,58]. Use of rifampin must be weighed against the likelihood of toxicity and drug interactions [59]. (See "Prosthetic joint infection: Treatment", section on 'Debridement and retention of prosthesis'.)

We do not use other rifamycins (eg, rifabutin) routinely in the setting of rifampin intolerance. In vitro and in vivo data are promising; clinical trials are warranted [60,61].

Rifampin must be combined with another active antibacterial agent because rapid emergence of resistance in the setting of rifampin monotherapy is common. Rifampin should be initiated several days after surgical debridement; resistance is less likely to emerge when the burden of organisms has been reduced [56,62]. We continue rifampin for the duration of definitive therapy (parenteral therapy or highly bioavailable oral antibiotic therapy). [56]. We do not use rifampin during suppressive therapy. (See "Rifamycins (rifampin, rifabutin, rifapentine)" and "Prosthetic joint infection: Treatment".)

Fusidic acid (not available in the United States) may be used as adjunctive oral therapy in the treatment of chronic osteomyelitis [63-65].

Gram-negative organisms — For treatment of osteomyelitis due to gram-negative organisms, we favor fluoroquinolones (if susceptibility testing confirms sensitivity) since they have high bone penetration with oral administration [25,66]. However, some avoid use of these agents in the setting of fracture, due to possible adverse effects on fracture healing [66].

Other antibiotics for treatment of osteomyelitis due to gram-negative organisms include parenteral beta-lactams and carbapenems. Antibiotic dosing is summarized in the table (table 1).

Among the gram-negative organisms, management of osteomyelitis caused by Pseudomonas aeruginosa may be more difficult than other organisms [67,68]. In one cohort study including 66 patients with osteomyelitis due to P. aeruginosa who underwent surgical debridement, most patients were treated with two weeks of intravenous therapy followed by oral fluoroquinolone therapy (median duration of treatment 45 days); clearance of infection was demonstrated in 94 percent of cases [68]. Most patients received monotherapy; no emergence of resistance was observed.

Streptococci and enterococci — Osteomyelitis due to penicillin-susceptible streptococci may be treated with penicillin; ceftriaxone is also acceptable and is more convenient for outpatient administration (table 1).

For treatment of osteomyelitis due to enterococci, it is unclear whether combination therapy is superior to monotherapy, particularly if thorough debridement is performed [69]. We favor combination therapy in the setting of retained hardware; the regimen of ampicillin and ceftriaxone is generally well tolerated [70]. However, if antibiotic-impregnated material containing gentamicin is implanted at the time of debridement, monotherapy is likely sufficient; in such cases, we favor treatment with intravenous ampicillin or penicillin (table 1).

Suppressive therapy — Long-term antibiotic suppression with an oral agent is warranted for patients with retained hardware and/or necrotic bone not amenable to debridement, following completion of parenteral therapy with resolution of clinical signs of acute infection (table 2). (See 'Retained hardware' above.)

Duration of therapy — Issues related to duration of antibiotic therapy are discussed above. (See 'Clinical approach' above.)

Local antibiotic delivery — Local antimicrobials mixed with absorbable carriers (eg, calcium sulfate beads) or nonabsorbable carriers (eg, polymethyl methacrylate beads or cement) placed in and around the fracture site may be useful as an adjunct to systemic antibiotics for prevention or treatment of infection [71,72].

Data are most robust for polymethylmethacrylate beads; these are typically removed within two to four weeks. Antibiotic-impregnated calcium sulfate beads are absorbable; they have been used as a bone substitute and delivery system for antibiotics repairing bony defects in open fractures caused by combat-related injuries [71,73].

SPECIAL POPULATIONS

Osteomyelitis associated with trauma — Issues related to osteomyelitis associated with trauma are discussed separately. (See "Osteomyelitis associated with open fractures in adults".)

Prosthetic joint infection — Issues related to prosthetic joint infection are discussed separately. (See "Prosthetic joint infection: Treatment".)

Osteomyelitis in sickle cell disease — Issues related to osteomyelitis in sickle cell disease are discussed separately. (See "Acute and chronic bone complications of sickle cell disease", section on 'Osteomyelitis and septic arthritis'.)

ADJUNCTIVE THERAPIES — Hyperbaric oxygen (HBO) has been suggested as an adjunctive therapy in patients with refractory osteomyelitis [74,75]. We do not favor use of HBO for treatment of osteomyelitis, given lack of sufficient data.

Issues related to HBO are discussed further separately. (See "Hyperbaric oxygen therapy".)

OUTCOME — Success rates for treatment of osteomyelitis reported in the literature range from 60 to 90 percent [18]. The success rate can be highly variable given the heterogeneity in completeness of debridement, the possibility of concomitant vascular insufficiency at the site of infection, and other factors.

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Osteomyelitis and prosthetic joint infection in adults" and "Society guideline links: Outpatient parenteral antimicrobial therapy".)

SUMMARY AND RECOMMENDATIONS

●Approach to treatment of nonhematogenous osteomyelitis – In general, management of nonhematogenous osteomyelitis consists of operative debridement for removal of necrotic material and culture of involved tissue and bone, followed by antimicrobial therapy for eradication of infection. (See 'Nonhematogenous osteomyelitis' above.)

•Patients without orthopedic hardware (see 'Absence of orthopedic hardware' above):

-Patients with residual infected bone that is not amenable to complete removal are typically treated with a prolonged duration antibiotic therapy, guided by antimicrobial susceptibility data (table 1). The optimal duration of antibiotic therapy is uncertain; we generally provide at least six weeks from the last debridement. (See 'Residual infected bone present' above.)

-Patients who undergo amputation or complete removal of all involved bone warrant a relatively short course of antibiotic therapy. (See 'No residual infected bone' above.)

•Patients with orthopedic hardware (see 'Presence of orthopedic hardware' above):

-Following debridement, patients should be treated with a prolonged duration of intravenous antibiotic therapy, guided by antimicrobial susceptibility data (table 1). The optimal duration is uncertain; we generally provide six weeks of therapy. (See 'No retained hardware' above.)

-Thereafter, patients with retained hardware and/or necrotic bone not amenable to debridement are often treated with long-term antibiotic suppression with an oral agent, guided by antimicrobial susceptibility data (table 2). (See 'Retained hardware' above.)

●Antibiotic selection for nonhematogenous osteomyelitis

•Empiric antibiotic therapy – We suggest empiric treatment of nonhematogenous osteomyelitis with antibiotics that have activity against methicillin-resistant Staphylococcus aureus (MRSA) and gram-negative organisms (Grade 2C). Reasonable regimens include vancomycin in combination with a third- or fourth-generation cephalosporin. (See 'Empiric therapy' above.)

•Directed therapy – Once a pathogen has been identified from culture, antibiotic therapy should be tailored to the specific organism (table 1). (See 'Definitive therapy' above.)

•Role of rifampin – For patients with methicillin-sensitive or methicillin-resistant S. aureus osteomyelitis and retained hardware, we suggest adjunctive use of rifampin (in combination with at least one other anti-staphylococcal agent) (Grade 2C). Use of rifampin must be weighed against the likelihood of toxicity and drug interactions. (See 'Use of adjunctive agents' above.)

●Hematogenous osteomyelitis – Treatment of hematogenous (nonvertebral) osteomyelitis consists of parenteral antibiotic therapy. In addition, surgical debridement is warranted in the setting of subperiosteal collection and/or in the setting of concomitant joint infection. Empiric antibiotic coverage for treatment of hematogenous osteomyelitis should include activity against MRSA and aerobic gram-negative bacilli. Once an etiologic organism is isolated, the antibiotic regimen should be tailored to susceptibility data. The optimal duration of antibiotic therapy is uncertain; in general, parenteral therapy is continued for at least four weeks from the last major debridement. (See 'Hematogenous osteomyelitis' above.)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Tahaniyat Lalani, MBBS, MHS, who contributed to an earlier version of this topic review.