Vertebral osteomyelitis: Treatment

INTRODUCTION — 

Vertebral osteomyelitis is an infection of any portion of the spinal column. The infection is almost always secondary to preceding or ongoing bacteremia, although blood cultures are not always positive at the time of diagnosis. Treatment of vertebral osteomyelitis consists of prolonged courses of antibiotics; in addition, surgical intervention is necessary in many cases. Without proper treatment, vertebral osteomyelitis can result in significant morbidity, including permanent paralysis.

This topic discusses the treatment of vertebral osteomyelitis. Information on clinical manifestations and diagnosis is found separately. (See "Vertebral osteomyelitis in adults: Clinical manifestations and diagnosis".)

Epidural abscess and psoas abscess, conditions commonly associated with vertebral osteomyelitis, are discussed separately. (See "Spinal epidural abscess" and "Psoas abscess".)

General discussion of the treatment of osteomyelitis can be found separately. (See "Osteomyelitis in the absence of hardware: Approach to treatment in adults".)

ROLE OF SURGERY AND PERCUTANEOUS INTERVENTIONS — 

Many patients require surgical or radiology-guided percutaneous intervention to achieve optimal outcomes.

Surgery — In the setting of vertebral osteomyelitis, surgery may perform three functions:

●Urgent decompression of the spinal canal in patients at risk for permanent neurologic deficits. (See 'Indications for surgery' below.)

●Debridement/debulking of the infection, including evacuation of paravertebral (or epidural) abscess.

●Collection of multiple samples for culture and histopathology. (See "Vertebral osteomyelitis in adults: Clinical manifestations and diagnosis", section on 'Importance of biopsy'.)

Indications for surgery — Urgent consultation should be sought for patients with the following indications for surgery (algorithm 1) [1-3]:

●Presence of neurologic deficits.

●Presence of epidural or paravertebral abscesses (eg, psoas abscess) not amenable to percutaneous drainage.

●Threatened or actual cord compression due to vertebral collapse and/or spinal instability.

In these patients, observational studies suggest that surgical intervention prevents loss of neurologic function [4-8].

We ultimately consult surgery for every patient with vertebral osteomyelitis, but we seek urgent consultation for patients with the above indications. Even in patients without an indication for surgery, evaluation by a spine surgeon allows for expert assessment of spinal stability and provides a baseline surgical assessment in the event that later intervention becomes necessary.

There are no randomized trials evaluating surgical management of vertebral osteomyelitis [2,3].

Placement of hardware — For patients at risk for spinal instability due to infection, hardware placement during surgery is often necessary [3,4,9,10].

Placing hardware in the setting of vertebral osteomyelitis has raised concern regarding fostering persistent infection due to adherence of bacteria to foreign material. Observational data, however, suggest this practice is safe [11,12]; courses of antibiotics are typically administered following placement of hardware. (See 'Duration of therapy' below.)

Some surgeons perform a two-stage procedure in an attempt to decrease the likelihood of infection on newly implanted hardware [13,14]. Such a procedure involves surgical debridement, with temporary stabilization if needed, followed some time later by definitive instrumented fusion. Outcomes appear to be similar in patients who undergo this procedure compared with patients who undergo acute instrumentation in the setting of infection [14].

Percutaneous interventions — Radiology-guided percutaneous procedures include collection of biopsy material and drainage of existing abscesses.

Tissue biopsy — All patients with vertebral osteomyelitis should undergo tissue biopsy for culture and histopathology, unless they have confirmed bacteremia due to an organism commonly associated with spine infection.

Tissue biopsy is obtained either by surgery or radiology-guided sampling. Biopsy is essential to guide antibiotic selection in patients with vertebral osteomyelitis.

(See "Vertebral osteomyelitis in adults: Clinical manifestations and diagnosis", section on 'Importance of biopsy'.)

Abscess drainage — In patients without an indication for surgery, paravertebral abscesses other than epidural abscesses can often be managed by computed tomography (CT)-guided drainage and antibiotic therapy. Smaller abscesses not amenable to drainage are often managed medically. (See "Psoas abscess", section on 'Drainage'.)

For patients with epidural abscesses, the treatment of choice for most patients is decompression of the abscess by a surgeon in the operating room. In carefully selected cases, CT-guided aspiration can be successful. However, all patients with neurologic deficit should undergo surgery; aspiration of an epidural abscess does not decompress the spinal cord. (See 'Indications for surgery' above and "Spinal epidural abscess", section on 'Surgical decompression for most patients'.)

ANTIBIOTIC THERAPY

Empiric therapy

Timing of antibiotics — The timing of initiation of antibiotic therapy in patients with vertebral osteomyelitis depends on the patient's clinical status.

Empiric antibiotics should be administered immediately to patients with the following [2,3]:

●Neurologic deficits

●Sepsis

●Confirmed bacteremia

For patients who do not meet the above criteria and are otherwise stable, we hold antibiotics until cultures are collected. If antibiotics have already been administered, we usually stop them. While guidelines suggest antibiotics be held for 14 days prior to culture collection, we find this to be infeasible in many patients who are in significant pain [2,3]. Delaying biopsy also delays treatment, which may risk disease progression. While the optimal duration off antibiotics is not known, we try to hold antibiotics for up to four days, if possible.

Some studies suggest that prior antibiotic exposure reduces yield from culture [15,16], whereas other studies have found no effect [17-19]. Whether antibiotics should be withheld prior to biopsy and the optimal duration for which antibiotics should be withheld before biopsy is uncertain. While some experts suggest 14 days [2,3], other studies suggest withholding antibiotics for a shorter duration. In one study of 209 patients with suspected vertebral osteomyelitis, withholding antibiotics for at least four days prior to biopsy maximized culture yield [16]. The decision to withhold antibiotic therapy should be made on a case-by-case basis, taking into consideration the patient's clinical condition. (See "Vertebral osteomyelitis in adults: Clinical manifestations and diagnosis", section on 'Importance of biopsy' and "Osteomyelitis in the absence of hardware: Approach to diagnosis in adults", section on 'Timing and technique'.)

Empiric antibioic regimens — For initial empiric treatment of vertebral osteomyelitis, we generally administer antibiotics intravenously. Once culture results return, the decision to continue intravenous therapy or transition to an oral regimen depends on several factors. (See 'Selecting intravenous versus oral therapy' below.)

Empiric regimens cover the most common causes of vertebral osteomyelitis, including staphylococci (including methicillin-resistant Staphylococcus aureus [MRSA]), streptococci, and enteric gram-negative bacilli. (See "Vertebral osteomyelitis in adults: Clinical manifestations and diagnosis", section on 'Microbiology'.)

A patient's prior culture results should be reviewed prior to prescribing empiric therapy. We do not routinely cover organisms such as Pseudomonas spp, multidrug-resistant gram-negative pathogens (eg, extended-spectrum beta-lactamase [ESBL] producers, AmpC producers), or anaerobic bacteria unless there is specific reason to suspect them.

Our preferred regimen is combination therapy:

●Intravenous vancomycin (see table for dosing (table 1))

PLUS one of the following:

•Ceftriaxone (2 g IV once daily), or

•Cefotaxime (2 g IV every six to eight hours), or

•Cefepime (2 g IV every 8 to 12 hours; dosing every 8 hours should be used if Pseudomonas spp is suspected), or

•Ceftazidime (2 g IV every eight hours), or

•Ciprofloxacin (500 to 750 mg orally every 12 hours, or 400 mg IV every 12 hours). The higher dosage should be used if Pseudomonas spp is suspected.

Alternatives to cephalosporins or ciprofloxacin include piperacillin-tazobactam and carbapenems (eg, meropenem).

Pathogen-specific regimens — Once biopsy or blood culture results are available, antibiotics should be tailored to target the identified pathogen.

No randomized controlled studies have compared antibiotic regimens for vertebral osteomyelitis.

Selecting intravenous versus oral therapy — Historical practice favored treatment with prolonged parenteral courses of antimicrobial therapy. However, oral antibiotic administration offers convenience, eliminates risk of catheter-related complications, and may lower costs.

Recent data (mostly observational) has demonstrated that an early switch to oral antimicrobial therapy is equivalent to parenteral antimicrobial therapy in select cases [3,20-24]. One randomized trial of 1054 patients with osteomyelitis of any type found no difference in failure rate with oral therapy (13 percent) versus intravenous (15 percent); 72 patients with vertebral osteomyelitis were included in the trial [20]. In another randomized trial of 359 patients with vertebral osteomyelitis assessing duration of therapy (six weeks versus 12 weeks), most patients received an initial two-week intravenous course before transitioning to oral therapy, and the duration of intravenous therapy was not associated with treatment outcome [25].

We consider oral therapy for patients who meet certain criteria [2,3,20,26]:

●Surgical decompression and debridement, or percutaneous drainage of paravertebral abscess, has been completed in patients with indications for these procedures. (See 'Indications for surgery' above and 'Abscess drainage' above.)

●No undrained epidural abscess or extensive bony destruction. For these patients, we often maintain intravenous therapy for some or all of the treatment course, at least until clinical improvement is demonstrated.

●No infection at another site that requires intravenous therapy (eg, Staphylococcus aureus bacteremia, endocarditis).

For patients with S. aureus bacteremia, the decision to transition to oral therapy is discussed elsewhere. (See 'Staphylococci' below.)

We suggest evaluation for endocarditis for patients whose vertebral osteomyelitis is caused by certain gram-positive organisms (eg, Staphylococcus, Streptococcus, Enterococcus) or in patients with underlying valvular disease, cardiac implantable devices, and/or new-onset heart failure or murmur, even in the absence of bacteremia. If endocarditis is present, longer intravenous courses of therapy may be indicated. (See "Vertebral osteomyelitis in adults: Clinical manifestations and diagnosis", section on 'Evaluation for endocarditis'.)

Detailed information on management of S. aureus bacteremia and on endocarditis is found separately. (See "Clinical approach to Staphylococcus aureus bacteremia in adults" and "Antimicrobial therapy of left-sided native valve endocarditis" and "Right-sided native valve infective endocarditis", section on 'Management'.)

●A highly bioavailable oral option is available to treat the proven pathogen.

●Able to absorb oral medications. Clinicians should check for concomitant medication that may hinder absorption. For example, calcium supplements, multivitamins, antacids, and sucralfate can decrease fluoroquinolone absorption.

●No concerning side effects, allergies, contraindications, or drug-drug interactions that would preclude optimal oral therapy.

●Compliance with oral therapy can be assured or carefully monitored.

Oral antibiotic regimens for treatment of vertebral osteomyelitis are summarized in the table (table 2).

Specific pathogens

Staphylococci — For all staphylococcal infections in the presence of new or retained hardware, we add oral rifampin (300 mg orally twice daily, or 600 mg orally once daily) to the regimen, regardless of whether the regimen is intravenous or oral (table 2). (See 'Hardware-associated' below and "Osteomyelitis in the absence of hardware: Approach to treatment in adults", section on 'Staphylococci' and "Osteomyelitis in the absence of hardware: Approach to treatment in adults", section on 'Role of adjunctive rifampin'.)

●Staphylococcus aureus – We always initiate therapy for S. aureus infections with an intravenous regimen.

The optimal duration of intravenous administration prior to transitioning to oral therapy is uncertain, especially in patients with bacteremia. For patients with vertebral osteomyelitis and S. aureus bacteremia, some experts recommend treating with a full course of intravenous antibiotics. Others treat with intravenous therapy for a minimum of 7 to 14 days before transitioning to oral therapy, if endocarditis has been excluded by thorough evaluation [3]. (See "Clinical approach to Staphylococcus aureus bacteremia in adults".)

•Methicillin-susceptible Staphylococcus aureus (MSSA) – For intravenous therapy, we suggest treatment with an anti-staphylococcal penicillin such as nafcillin or oxacillin (2 g intravenously every four hours) or cefazolin (2 g intravenously every eight hours); flucloxacillin is also an acceptable agent (not available in the United States). We do not use ceftriaxone in the bacteremic patient but would consider it for the non-bacteremic patient in the definitive phase of treatment.

Our preferred oral regimen for susceptible isolates is combination therapy with levofloxacin (750 mg orally once daily) plus rifampin (300 mg orally twice daily, or 600 mg orally once daily) (table 2)

For patients unable to take preferred regimens, regimens used for MRSA are appropriate (see next bullet).

•Methicillin-resistant Staphylococcus aureus (MRSA) – For intravenous therapy, we recommend vancomycin (see dosing in table (table 1)) or daptomycin (6 to 10 mg/kg intravenously once daily).

Our preferred oral regimen for susceptible isolates is combination therapy with levofloxacin (750 mg orally once daily) plus rifampin (300 mg orally twice daily, or 600 mg orally once daily) (table 2).

Intravenous dalbavancin and oritavancin, novel lipoglycopeptide antimicrobials with long half-life, have been administered off-label in a small number of patients with osteomyelitis, including vertebral osteomyelitis [27,28]. Until further data are available, we reserve these agents for patients who are not candidates for oral antimicrobial therapy or in whom other parenteral therapies are not an option [29].

●Coagulase-negative staphylococci – Most coagulase-negative staphylococci are methicillin-resistant and should be treated with a regimen used for MRSA. If the organism tests susceptible to methicillin, one of the regimens for MSSA can be used.

Staphylococcus lugdunensis is unique because it is frequently susceptible to all beta-lactams. Penicillins and cephalosporins are the intravenous agents of choice, if susceptible. (See "Osteomyelitis in the absence of hardware: Approach to treatment in adults", section on 'Staphylococci' and "Staphylococcus lugdunensis", section on 'Treatment'.)

Streptococci — For intravenous treatment, we suggest penicillin G (20 to 24 million units/day by continuous infusion or in six divided daily doses) or ceftriaxone (2 g intravenously every 24 hours). (See "Osteomyelitis in the absence of hardware: Approach to treatment in adults", section on 'Streptococci and enterococci'.)

For susceptible streptococci, our preferred oral regimen is amoxicillin (875 to 1000 mg orally every eight hours) (table 2) [3]. Clindamycin (300 to 450 mg orally every eight hours) is an alternative , if the organism is susceptible.

Gram-negative bacilli — Choice of a specific agent for empiric therapy of gram-negative bacilli depends on the specific organism and results of susceptibility testing.

Our suggested intravenous options are one of the following, depending on culture and susceptibility results:

●Third-generation cephalosporin (ceftriaxone 2 g intravenously daily, ceftazidime 2 g intravenously every eight hours, or cefotaxime 2 g intravenously every six to eight hours). For organisms with potential AmpC production (Enterobacter cloacae complex, Klebsiella aerogenes, Citrobacter freundii), third generation cephalosporins should be avoided in favor of cefepime. (See "Gram-negative bacillary bacteremia in adults", section on 'Regimen choice'.)

●Fourth-generation cephalosporin (cefepime 2 g intravenously every 8 to 12 hours). Dosing every 12 hours is sufficient in most cases, but pseudomonal infection requires dosing every 8 hours. Higher dosing interval may also be considered for organisms capable of AmpC production.

Oral ciprofloxacin (500 mg orally every 12 hours) is the preferred oral antibiotic; for infections due to Pseudomonas spp, the dosage is increased to 750 mg orally every 12 hours (table 2).

Culture-negative cases — If cultures of blood and tissue are negative and the clinical suspicion for vertebral osteomyelitis remains high (on the basis of clinical and radiographic findings), some experts advocate performing a second biopsy. In one retrospective study including 136 patients with vertebral osteomyelitis in the absence of bacteremia, first biopsy was positive in 43 percent of cases and a second biopsy in 40 percent of cases; when combined with blood culture results, this approach led to microbiological diagnosis in nearly 75 percent of cases [30].

Other experts begin empiric therapy after the initial biopsy is obtained, based on other studies that suggest the yield of a second biopsy may be low [31]. Empiric treatment offers potential for earlier symptomatic improvement and can prevent progression of the infection, if active antibiotics are chosen. Empiric treatment in lieu of repeat biopsy should be considered as part of shared decision making with the patient and care team. (See 'Outcomes' below and "Vertebral osteomyelitis in adults: Clinical manifestations and diagnosis", section on 'Importance of biopsy'.)

If repeat biopsy specimens and initial blood cultures are nondiagnostic, we usually initiate empiric therapy as discussed below. If such therapy does not result in objective clinical improvement in three to four weeks, a third percutaneous needle biopsy or an open surgical biopsy may be helpful [1]. (See "Vertebral osteomyelitis in adults: Clinical manifestations and diagnosis", section on 'Importance of biopsy'.)

For treatment of culture-negative infections, we use the same regimens used for empiric therapy, as described above. (See 'Empiric antibioic regimens' above.)

Hardware-associated — Antibiotic selection for vertebral osteomyelitis that developed in the presence of hardware is based on culture data obtained from operative debridement. In patients with post-operative wound infection, appropriately collected samples from the deep wound bed may also be used to guide therapy, but samples obtained during surgical debridement are preferred whenever feasible. (See 'Surgery' above.)

Regimens are the same as those used for infections without hardware except oral rifampin is added if staphylococcal species (S. aureus or coagulase-negative staphylococci) are identified [2,3]. Rifampin is continued as part of combination therapy as long as hardware remains in place and the patient remains on antibiotic therapy and is tolerating rifampin. Rifampin should never be used as monotherapy. For patients on suppressive antibiotics due to retained hardware, the use combination therapy including rifampin is less certain. (See 'Specific pathogens' above.)

Duration of antibiotic therapy in patients with hardware in the setting of vertebral osteomyelitis is often longer than the antibiotic duration in infections without hardware. (See 'Duration of therapy' below.)

Development of vertebral osteomyelitis in the presence of spinal hardware typically reflects a complication of the hardware placement [32]. Most patients have an associated wound or soft tissue infection, although rarely hematogenous seeding of hardware can occur. (See "Vertebral osteomyelitis in adults: Clinical manifestations and diagnosis", section on 'Mechanisms of infection'.)

Duration of therapy — In most cases, we treat patients with vertebral osteomyelitis for six weeks if clinical improvement proceeds as expected [2,3]. (See 'Expected course and monitoring' below.)

In the following situations, we often treat for longer than six weeks:

●Hardware-associated infection – Treatment of hardware-associated infection depends on the whether the hardware was present at the time of infection:

•Initial infection involved hardware – For these infections, duration of therapy depends on whether the infected hardware is surgically removed.

If all infected hardware is removed, typical duration of therapy is six weeks from date of hardware removal.

If infected hardware is retained, the duration of treatment depends on when the infection occurred relative to the date of initial hardware placement [3]:

-If the infection occurred in the early post-operative setting (eg, less than 30 to 90 days from the date of initial hardware placement), antibiotic therapy is typically administered for 6 to 12 weeks from the date of surgical irrigation and debridement of the infected hardware.

-If the infection occurred more than 30 to 90 days from the date of initial hardware placement, chronic antibiotic suppression is often administered following 12 weeks of treatment from the date of irrigation and debridement. Some experts discontinue antibiotics after a minimum of 12 weeks or up to one year of treatment, often contingent on whether bone fusion is achieved. In some cases, removal of hardware can eventually be performed (after bone fusion has been achieved), so that antibiotic suppression can be discontinued.

•Hardware was placed for spine stabilization in setting of native-bone infection – Patients with vertebral osteomyelitis often require surgical spine stabilization with hardware. If infection is adequately debrided at the time of hardware placement, we typically administer antibiotic treatment for 12 weeks following the placement of hardware. There is little evidence to guide antibiotic duration in this setting [3,14,33]. (See 'Placement of hardware' above.)

●Undrained paravertebral or epidural abscess – We typically treat these patients for six to eight weeks, depending on clinical response. We have a low threshold for obtaining a magnetic resonance image (MRI), particularly in patients whose symptoms fail to improve as expected. Some experts treat until MRI reveals resolution of abscess. (See "Spinal epidural abscess", section on 'Duration'.)

●Large phlegmon or extensive bony destruction – For these patients, we often extend therapy for up to 12 weeks, depending on clinical response.

●MRSA infection – MRSA is a risk factor for antibiotic failure in patients with vertebral osteomyelitis. We don't necessarily treat patients with MRSA longer than patients with MSSA infection, but, if clinical response is slower than expected, we consider prolonging antibiotic therapy to a total of eight weeks.

The duration of treatment does not depend on the route of antibiotic administration (intravenous versus oral).

The above approach to duration of therapy is supported by the following studies:

●A randomized trial including 351 patients with vertebral osteomyelitis demonstrated that six weeks of antibiotic treatment was associated with similar efficacy compared with 12 weeks of antibiotic therapy with respect to the proportion of patients cured at one year (>90% cure in both groups) [25]. The number of patients with abscesses in the trial was low (19 percent), and CT-guided drainage was necessary in only 4 percent of cases. A short duration of intravenous therapy was administered (median 14 days) followed by oral therapy; the most common oral regimen was a fluoroquinolone with rifampin (in 73 percent of cases in both groups). The trial did not include patients with culture-negative infections. The most commonly isolated pathogen was S. aureus (145 patients [eight had MRSA]). Older age (>75) and S. aureus infection were risk factors for clinical failure; most diagnoses were based on bacteremia.

●In a retrospective review of 345 patients with vertebral osteomyelitis, outcomes for patients treated with six weeks of therapy were similar to those treated with longer durations, except patients with MRSA infection or an undrained abscess trended toward improved outcomes with eight or more weeks of treatment [34].

ADJUNCTIVE MEASURES — 

Analgesics are generally helpful in managing pain after the diagnosis is established. Clinicians should have a low threshold for referral to a specialist pain service.

In patients with possible spinal instability, a spine specialist should manage the patient's mobility [3]. Bed rest and braces are often utilized until the spine is stabilized.

In the absence of spinal instability, restriction of mobility and use of braces are generally dependent on the patient's clinical and pain status [3]. Some patients may benefit from early mobilization, when directed by the Spine Surgery team. Physical therapists play a critical role in managing the patient's progress.

EXPECTED COURSE AND MONITORING

Clinical course — The main sign of clinical improvement is a gradual, steady reduction in pain. In some patients, neurologic deficits improve as well. Due to permanent damage that may occur prior to onset of treatment and bone remodeling in response to infection treatment, pain and neurologic symptoms may not completely resolve.

The main signs of clinical failure are worsening pain or neurologic symptoms, lack of improvement in pain, or new signs of sepsis. (See 'Refractory cases' below.)

During the early phases of therapy, thorough neurologic examination should be performed on a daily basis [3]. During the course of treatment for vertebral osteomyelitis, the pace of progression can change suddenly. Close observation and patient counseling is essential.

In general, patients whose infections do not involve hardware should be followed for signs of relapse for at least one year; for those with hardware-associated infection, follow-up for two years is warranted [3].

Laboratory monitoring — Studies suggest that erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) may correlate with clinical response to treatment, but in most osteomyelitis studies they did not add any additional value over clinical assessment by a clinician. (See "Osteomyelitis in the absence of hardware: Approach to treatment in adults", section on 'Monitoring and follow-up'.)

●C-reactive protein (CRP) – Some experts follow CRP on a weekly basis, whereas others don’t follow it unless the patient is not improving as expected. Some obtain a CRP at the time of diagnosis and at the time of treatment completion [2,3]. CRP may correlate with clinical response better than ESR because it normalizes more rapidly than the ESR after successful treatment of vertebral osteomyelitis and after uncomplicated spinal fusion surgery [35].

●Erythrocyte sedimentation rate (ESR) – ESR is a less useful inflammatory marker to monitor. In one retrospective study including 44 patients with vertebral osteomyelitis, those without a significant decline in the ESR during the first month of therapy were more likely to fail medical therapy (9 of 18 cases compared with 3 of 26 cases with a >50 percent fall in ESR) [36]. However, a rapid decline is not common. In one study, the ESR either rose or failed to decline during the first two weeks of treatment in 19 of 32 patients who were ultimately cured with medical therapy. Similar findings have been reported by others [37].

Patients on certain antibiotics should have regular laboratory monitoring to assess for side effects; some antibiotics (eg, vancomycin) also require monitoring of drug levels. (See "Outpatient parenteral antimicrobial therapy", section on 'Monitoring' and "Vancomycin: Parenteral dosing, monitoring, and adverse effects in adults", section on 'Approach to dosing and monitoring'.)

Limited role of follow-up imaging — Routine follow-up imaging is not necessary [3].

Imaging studies can be misleading; MRI, CT, and plain films may worsen for several weeks after the initiation of antibiotic therapy despite clinical improvement [38,39]. Furthermore, MRI results often continue to depict findings present on presentation in patients who are cured. In a study of 29 patients who were cured of infection, MRI obtained three and six months after diagnosis revealed persistent disc height loss in all patients and vertebral edema in 15 percent; among patients with abscess, persistence of fluid in the disc space or paravertebral abscess was not uncommon [39]. All epidural abscesses (10 patients) were resolved on MRI by three months.

Indications for follow-up imaging include lack of expected clinical response or clinical worsening on therapy. In patients with clinical worsening, prompt imaging is obtained to evaluate for the presence of an abscess in need of drainage or to detect spinal instability amenable to surgical intervention [2]. Some experts consider follow-up imaging in patients with undrained epidural abscess to document resolution before discontinuing antibiotics, although data supporting this practice are scarce.

REFRACTORY CASES — 

Failure of pain or neurologic deficits to resolve is not necessarily indicative of therapeutic failure. However, failure of pain to improve or worsening pain or neurologic function is concerning for therapeutic failure, as discussed above. (See 'Clinical course' above.)

For patients in whom refractory infection is suspected, we obtain C-reactive protein (CRP) and obtain MRI, surgical referral, or both. In some cases, particularly those that were culture-negative, repeat biopsy (percutaneous or open-surgical biopsy) is appropriate (algorithm 1).

OUTCOMES — 

The most serious complication of vertebral osteomyelitis is neurologic impairment secondary to either abscess formation or bony collapse. Most patients have gradual improvement in back pain after therapy has begun, and the pain typically disappears after bone fusion occurs. However, back pain can persist. The best way to reduce the morbidity and mortality associated with vertebral osteomyelitis is to minimize the time between the onset of symptoms and the initiation of appropriate therapy [2,3,40].

In a 2009 systematic review of over 1000 patients in 14 studies, permanent neurologic deficits occurred in 32 percent and relapse in 32 percent; attributable mortality was 6 percent [41]. In another study including 260 patients with vertebral osteomyelitis, neurologic deficits occurred in 16 percent and persistent back pain in 32 percent [42].

Major depression is a well-recognized complication of chronic back pain (about a third of patients) and is an important potential long-term consequence [43,44].

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

●Overview – Treatment of vertebral osteomyelitis consists of prolonged courses of antibiotics; in addition, surgical intervention is necessary in many cases. Without proper treatment, vertebral osteomyelitis can result in significant morbidity, including permanent paralysis. (See 'Introduction' above.)

●Indications for surgery – Urgent consultation should be sought for in patients with the following indications for surgery (see 'Indications for surgery' above):

•Presence of neurologic deficits.

•Presence of epidural or paravertebral abscesses (eg, psoas abscess) not amenable to percutaneous drainage.

•Threatened or actual cord compression due to vertebral collapse and/or spinal instability.

We ultimately consult surgery for every patient with vertebral osteomyelitis, but we seek urgent consultation for patients with the above indications.

●Empiric antibiotic therapy

•Indications – Empiric antibiotics should be administered immediately to patients with the following (see 'Timing of antibiotics' above):

-Neurologic deficits

-Sepsis

-Confirmed bacteremia

For patients who do not meet the above criteria and are otherwise stable, we hold antibiotics until cultures are collected.

•Regimens – We suggest an intravenous regimen that covers staphylococci (including methicillin-resistant Staphylococcus aureus [MRSA]), streptococci, and enteric gram-negative bacilli (Grade 2C). These are the most common pathogens.

An example regimen is vancomycin (see table for dosing (table 1)) plus cefepime (2 g intravenously every 8 to 12 hours). If Pseudomonas spp is suspected, cefepime dosing every eight hours should be used.

●Pathogen-directed regimens – Once biopsy or blood culture results are available, antibiotics should be tailored to target the identified pathogen. (See 'Pathogen-specific regimens' above.)

•Oral versus intravenous therapy – We suggest oral therapy for patients who meet the following criteria (Grade 2C) (see 'Selecting intravenous versus oral therapy' above):

-Surgical debridement or percutaneous drainage has been performed, if indicated.

-No undrained epidural abscess or extensive bony destruction.

-No infection at another site that required intravenous therapy (eg, S. aureus bacteremia, endocarditis).

-A highly bioavailable oral option is available that the patient can absorb and tolerate and doesn't have concerning drug-drug interactions.

A list of pathogen-specific oral regimens is available in the table (table 2). (See 'Specific pathogens' above.)

For other patient groups, we suggest intravenous treatment (Grade 2C).

●Duration of therapy – For most patients with uncomplicated infection and no hardware, we suggest treating for six weeks rather than longer durations (Grade 2B).

We sometimes use longer durations of treatment for patients with hardware-associated infection, undrained abscess, or extensive bony destruction. (See 'Duration of therapy' above.)

●Follow-up – Patients with vertebral osteomyelitis should be followed closely to ensure clinical response. In patients who improve as expected, there is limited role for routine follow-up imaging, and the value of following inflammatory markers (eg, erythrocyte sedimentation rate [ESR], C-reactive protein [CRP]) is uncertain. (See 'Expected course and monitoring' above.)

ACKNOWLEDGMENT — 

The UpToDate editorial staff acknowledges Malcolm McDonald, MD, who contributed to an earlier version of this topic review.