• Summary
  Click here to print a Quick Reference Guide
• Background
• Clinical Diagnosis
• Investigation
• Treatment
• Non-Antimicrobial Treatment
• Empirical Antimicrobial Treatment
• Directed Antimicrobial Treatment (when microbiology results are known)
• Duration of Treatment
• Switch to oral agent(s)
• Treatment Failure

Key Links

1. Vertebral osteomyelitis and discitis with or without paraspinal involvement.
   • Staphylococcal VO
   • Enterococcal VO
   • Streptococcal VO
   • Enterobacteriaceae [“coliforms”]
   • Pseudomonas aeruginosa
2. Infected spinal instrumentation
3. Epidural abscess

Criteria for use of guideline

• Patients with suspected deep spinal infection – including vertebral osteomyelitis and discitis [VO], deep surgical wound infections involving metal work, epidural drug delivery-related infection,

Diagnosis
History

• Consider deep spinal infection in patients with axial back pain that is present at rest or causing sleep disturbance.
• Assess systemic symptoms of infection: fevers, chills, night sweats, general malaise, anorexia, weight loss.
• No clinical features reliably distinguish mycobacterial infection [TB] from other pathogens but in the former symptoms may be present for many months.
• Determine history of previous or recent spinal surgery, epidural drug delivery or other invasive spinal procedures.
• Determine whether spinal metal work or other prosthetic material is present.

Examination

• Pyrexia may not be present.
• Check for localized spinal tenderness.
• Assess and document any neurological deficit.
• Assess wounds for inflammation/discharge.
• Consider deep seated infection in any patient with a surgical wound infection post spinal surgery.

Standard baseline investigations

• FBC, CRP, U&E and LFT’s.
• 2 sets of Blood cultures at different times.
• Urine sample for culture if dipstick is positive [Link: UTI diagnosis guidelines].
• Plain chest X-ray [haematogenous VO]
• Plain spinal X-rays.
• See main guideline re MRI, other imaging

Non-Antimicrobial Management

• Need for surgery is assessed on a case-by-case basis
• Refer cervical spine infections to neurosurgeons
• Refer thoracic, lumbar and sacral spine infections to orthopaedic spinal surgeons [unless previously operated on by neurosurgeons]

Antimicrobial treatment

• Antimicrobials should NOT be started until after biopsies have been taken unless the patient has severe sepsis or septic shock, or a microbiological diagnosis has already been obtained via blood cultures.
• Empirical therapy should be avoided if possible. If required, discuss with Microbiology on a case-by-case basis; take into consideration recent culture results.
• See full guideline for directed antimicrobial therapy:

Antimicrobial allergy
Discuss with microbiology.

Duration of treatment
see main guideline.

Switch to oral agent[s]
See organism specific recommendation. [Full guideline]

Back to top

Deep spinal infection means those affecting the vertebral body, facet joint, intervertebral disc, spinal canal, epidural space and paravertebral tissues. Vertebral body infection, known as vertebral osteomyelitis [VO] is the most common spinal infection and is usually accompanied by infection of the intervertebral disc (Hadjpavlou et al., 2000). Isolated discitis is uncommon and usually complicates surgery or discography (Hadjpavlou et al., 2000). VO may also involve abscess formation in the paraspinal tissues or epidural space [so called “secondary epidural abscesses”]. Primary epidural abscesses are much less common. Epidural catheter-associated infections may be superficial, involving only the catheter insertion site, or deep, causing an epidural abscess. Such infections may occur while the catheter is in situ or after it has been removed.

Vertebral osteomyelitis and discitis is often a delayed diagnosis and this delay is associated with adverse outcomes (McHenry et al., 2002; Tsiodras & Falagas, 2006). Spinal infections can be misdiagnosed as spinal tumour and empirical therapy is frequently erroneous particularly when mycobacterial infection is diagnosed clinically or conversely when the possibility of mycobacterial infection is not considered (Digby & Kersley, 1979; Kumar et al., 2005). In addition, deep spinal infections are uncommon. Hence the need for a guideline.

Spinal infections can arise:

• via haematogenous spread following a transient bacteraemia or following a bacteraemia from a distant primary focus of infection; or
• by direct inoculation following surgery, epidural drug delivery or trauma; or
• by direct spread from a contiguous focus of infection. e.g. post operative wound infections can spread to involve deep spinal tissues or an infected sacral pressure sore may erode into the sacrum.

Intradural infections such as meningitis and brain abscesses are covered in other LTHT guidelines: Guideline for the management of meningitis in adults, Guideline for the management of brain abscesses in adults.

Uncomplicated wound infections are not covered by this guideline.

The principles of management of deep spinal infection are to:

1. Establish the diagnosis
2. Determine the causative organism[s]
3. Formulate treatment plan [need for surgery, appropriate antimicrobial therapy and planned duration]
4. Monitor response to therapy

1. Establishing a diagnosis requires a low index of suspicion because deep spinal infection can be a difficult diagnosis [see diagnosis section].

2. Most VO episodes are caused by bacteria including staphylococci, streptococci, enterococci and Gram-negative bacilli [e.g. Escherichia coli and other “coliforms”, Pseudomonas aeruginosa] (McHenry et al., 2002). A mixed group of other bacteria are occasionally implicated including mycobacteria, Nocardia spp., Brucella spp., coryneforms, and anaerobes (McHenry et al., 2002). Yeasts and other fungi are occasionally involved (McHenry et al., 2002). Non-tuberculous VO is often termed “pyogenic” VO, but this is a misnomer because both tuberculous and non-tuberculous infection can occur with or without the presence of pus.

A smaller number of different pathogens cause deep spinal infection after surgery. Staphylococci [including meticillin resistant Staphylococcus aureus] and gram negative bacilli being the major causes. Organisms may gain entry to tissues at the time of the operation or invade the operative site following a surgical wound infection. Similar to infections following spinal surgery, epidural catheter-associated infections are mainly caused by staphylococci [including meticillin resistant Staphylococcus aureus] and gram negative bacilli. Spinal infections at a site of previous metalwork implantation have similar microbiology to other surgery-related infections but the presence of infected prosthetic material may require a different surgical approach.

3. A treatment plan will include a course of appropriate antimicrobial therapy [see later section] and may also require surgery [see later section]. Some infections in the presence of spinal metalwork can be eradicated with antimicrobial therapy and appropriate debridement but success is difficult to predict and is more likely if the infected material can be removed.

4. The long courses of antimicrobials required to eradicate these infections require careful monitoring. Regular measurement of clinical response [particularly resolution of rest pain, analgesia requirement], inflammatory markers and adverse effects of antimicrobial therapy re required. Vascular access may become an issue during prolonged courses of intravenous therapy and this will need to be managed.

Back to top

History
Axial back pain is the most common symptom of spinal infection and is present in at least 90% of episodes (Carragee, 1997a; Digby & Kersley, 1979; Sapico & Montgomerie, 1979). Pain caused by VO tends to be constant and present at rest (An & Seldomridge, 2006; Tsiodras & Falagas, 2006). Nocturnal pain with sleep disturbance is highly suggestive of either spinal infection or neoplasia. Consider a diagnosis of deep spinal infection in patients with axial back pain that is present at rest or at night. [Evidence level B]

Pott’s disease [tuberculous infection of the spine], although classically associated with children and young adults, has become a disease of older people in developed countries (Berbari et al., 2005). Systemic symptoms are often absent. Past haematogenous seeding of the spine, with reactivation years later, may mean that evidence of mycobacterial infection elsewhere is lacking. NB. There are no clinical features that reliably distinguish mycobacterial infection from other pathogens but in the former, symptoms may be present for many months (Tsiodras & Falagas, 2006). [Evidence level B]

Constitutional symptoms such as fevers, chills, night sweats, general malaise, anorexia or weight loss may be present but their absence does not rule out deep spinal infection. [Evidence level B]

Determine whether a patient has had previous spinal surgery or epidural drug delivery and whether spinal metal work or other prosthetic material is present. [Evidence level D]

In patients with an epidural catheter in situ deep spinal infection is more likely in patients with a fever, back pain, a systemic inflammatory response, leukocytosis, nocturnal pain or neurological abnormality. [Evidence level D]

Examination
Patients with haematogenous deep spinal infection may not have localized spinal tenderness at the site of infection. [Evidence level B]

Assess all patients for neurological deficit. [Evidence level D].

In patients with an epidural catheter in situ superficial infections at the insertion site are diagnosed on the basis or a purulent exudate or >1cm erythema and tenderness at the site in a patient who is systemically well without any fever, haemodynamic abnormality or back pain. [Evidence level D]

Consider deep seated infection in any patient with a surgical wound infection post spinal instrumentation [Evidence level C].

Careful examination for stigmata of endocarditis (Murmurs, splinter haemorrhages etc) is required in patients with haematogenous VO and discitis. [Evidence level C].

Back to top

The peripheral white blood cell count may be normal or elevated in patients with spinal infection and is therefore not particularly useful (An & Seldomridge, 2006). A full blood count [FBC] is indicated however to determine if the patient is anaemic, to ensure a normal platelet count before any biopsies or surgical intervention and to establish a baseline peripheral white blood cell count since several of the antimicrobials used in treatment can cause leukopaenia.
FBC should be measured at baseline and repeated weekly during therapy. [Evidence level C].

C-reactive protein [CRP] is a sensitive test, being raised in >90% of patients with spinal infection (Carragee et al., 1997; Rankine et al., 2004), but it is non-specific. A raised CRP provides supportive evidence of infection, rather than confirming the diagnosis, and can be used to monitor response to treatment.
CRP should be measured at baseline and weekly during therapy. [Evidence level B.]

Urea and electrolytes [U&E], liver function tests [LFT’s] are necessary baseline tests. Results will influence choice and dose of antimicrobials as well as fluid balance, nutritional support etc.
U&E and LFT’s should be measured at baseline and weekly during therapy [or more frequently if renal function is unstable]. [Evidence level C]

Blood cultures are positive in 33%-71% of cases of vertebral osteomyelitis (Digby & Kersley, 1979; Roblot et al., 2007; Tsiodras & Falagas, 2006). Because bacteria known to be common blood culture contaminants can also cause deep spinal infection [e.g. coagulase negative staphylococci and enterococci], two sets of Blood cultures are recommended prior to starting any antimicrobials. It is important to inoculate the optimal 8-10ml into each blood culture bottle.
Two sets of Blood cultures should be taken at different times in all patients with suspected spinal infection [Evidence level C.]

In patient with possible endocarditis [clickable link to endocarditis guideline], three sets of Blood cultures should be taken at different times during the first 24 hours (Elliott et al., 2004). [Evidence level B].

Urinalysis is indicated at baseline because spinal infections frequently arise from a primary focus in the urinary tract (Sapico & Montgomerie, 1979). [Evidence level B.]

A urine sample should be sent for culture if dipstick is positive for leukocytes or white blood cells. [NB. Sterile pyuria should raise the possibility of mycobacterial infection but there is no need to culture urine for mycobacteria if a spinal biopsy is planned.] [Evidence level B.]

Patients with a purulent exudate at an epidural catheter exit site or site of a previously removed epidural catheter should have a sample of pus sent to microbiology in a sterile container, a pus swab is inferior to a pus sample. [Evidence level B.]

Plain chest X-ray is indicated for all patients with suspected chronic spinal infection because of the possibility of mycobacterial infection. [Evidence level B.]

Plain spinal X-rays are insensitive for the diagnosis of early spinal infection [<2 weeks symptoms] (Berbari et al., 2005). However, plain x-rays can be useful for monitoring progress in terms of bone destruction, development of deformity and bone healing. Plain spinal X-rays should be carried-out in all suspected spinal infections. [Evidence level B.]

Magnetic resonant imaging [MRI] of the spine is the radiological investigation of choice for suspected vertebral osteomyelitis but is neither 100% sensitive nor 100% specific (Tsiodras & Falagas, 2006). 2/20 radiological diagnoses of spinal infection turned out to have malignant lesions after biopsy in a recent series (Rankine et al., 2004). MRI did not exert a significant effect on outcome in one study; the explanation was delay in obtaining MRI (McHenry et al., 2002). MRI can diagnose VO early in the course of infection and may have a positive impact on outcome (Carragee, 1997b). The use of gadolinium contrast is advocated.
MRI with contrast should be carried out as soon as possible after a diagnosis of vertebral osteomyelitis is suspected. [Evidence level B.]

Radiologically-guided biopsy [closed biopsy] of suspected vertebral osteomyelitis is necessary to confirm the diagnosis because of the lack of specificity of clinical and radiological findings and frequent errors when therapy is started empirically (Berbari et al., 2005; Kumar et al., 2005; Rankine et al., 2004). Although there are MRI features considered diagnostic of mycobacterial infection, a false positive rate of 20% has been reported in a small series (Jung et al., 2004) and erroneous diagnoses of spinal tuberculosis can be made (Kumar et al., 2005). In addition, concurrent infection and neoplasia can occur as can mixed infection, such as mycobacteria with other bacterial pathogens. Closed spinal biopsy influenced treatment in 35% of cases in a recent series (Rankine et al., 2004) and may influence more cases if the use of empirical antibiotic therapy for this condition could be reduced.
Radiologically-guided biopsy [closed biopsy] of suspected vertebral osteomyelitis should be undertaken prior to starting antimicrobial chemotherapy in all cases except:

1. where a patient has two or more Blood cultures positive for an organism consistent with VO AND,
2. when there are none of the characteristic MRI features of spinal mycobacterial infection, OR
3. when there is a clear cut indication for surgery [operative samples should be sent for culture], OR
4. when the patient is immunosuppressed (it may not be appropriate to delay empirical therapy, but biopsy if highly recommended).

[Evidence level D]

Sensitivity of open biopsy varies from 40-100% (An & Seldomridge, 2006; Rankine et al., 2004) and a second diagnostic procedure may be required. Sufficient biopsy material should be sent for microscopy, routine culture and mycobacterial culture. Contemporary molecular diagnostic techniques [e.g. 16S ribosomal DNA PCR] may be useful if routine cultures are negative but these require formal evaluation. Adequate biopsies can usually be obtained by closed, percutaneous techniques, typically under radiological guidance. By contrast, open biopsy techniques usually require general anaesthesia. Both techniques have low complication rates associated with the acquisition of tissue but to avoid any risks associated with general anaesthesia, the closed method is preferable.
Open biopsy should be reserved for failed closed biopsy or non-compliance of the patient e.g. the confused, restless patient. [Evidence level C]

Pus found at surgery [not pus swabs] should be sent to microbiology in a sterile container. [Evidence level B].

The biopsy material or pus from the spinal lesion should be sent for microbiology [microscopy, routine culture, mycobacterial culture] and histology in cases of suspected vertebral osteomyelitis. [Evidence level B].

Back to top

Surgical and Other Non-Antimicrobial Management

The indication for surgical intervention fall into distinct categories.

1. Correction of neurological complications such a cord or nerve root compression;
2. Stabilization of painful unstable spinal lesions or deformity resulting from progressive bone destruction;
3. To obtain material for a microbiological diagnosis;
4. To remove pus, infected bone or infected instrumentation when these are preventing, or likely to prevent, a cure with antimicrobial therapy alone.

These indications may co-exist to varying degrees. Such problems may mandate decompression and/or stabilisation. The effect of these complications varies, depending upon the spinal level [and extent] involved. In cases of early diagnosis and treatment, these complications are rare.

Stabilisation may mean nothing more than a period of bed rest or wearing an external orthosis. In some cases, decompression of pus can be via percutaneous guided drain insertion, often at the time of biopsy but neural compression by sequestered bone and/or disc will not be amenable to such treatment. However, the typical delays in diagnosis mean that in practice, many cases require urgent decompression and/or stabilization, by surgical means.

The advice of a spinal surgeon should be sought at an early stage and the role for surgery assessed on a case-by-case basis. In Leeds, cervical infections should be referred to neurosurgeons and thoracic, lumbar and sacral infections to orthopaedic spinal surgeons.
[Evidence level D]

When spinal instrumentation is involved in an infective process the likelihood of relapse is high if the instrumentation is not removed (Benson et al., 1998; Hahn et al., 2005; Ho et al., 2007; Malamo-Lada et al., 1999). In one paediatric series >80% of patients who required three or more irrigation and debridement procedures [with the original instrumentation in place] were not cured until all the metalwork was removed. If spinal fusion is established, removal of infected metalwork may be possible without aggravating deformity. However, removal of instrumentation prior to spinal fusion may lead to progressive deformity. Early post operative infections [within 6 months] can be eradicated in some instances with debridement, irrigation and intravenous antimicrobial therapy.
The decision to remove infected metalwork should be made on a case-by-case basis taking into consideration the time elapsed since the procedure, radiological evidence of spinal fusion and the clinical response to debridement, irrigation and systemic antimicrobial therapy. [Evidence level C]

Back to top

1. Vertebral osteomyelitis and discitis with or without paraspinal involvement.

The choice of antimicrobial agent[s] for treatment of VO will depend on penetration of the agent into spinal tissues, the pathogen involved, evidence from experimental animal models and clinical experience. Most human data come from observational studies and case reports rather than randomized controlled trials. The range of potential pathogens is wide and there is no single agent or simple combination of agents that could be used to empirically cover all of them; obtaining a microbiological diagnosis is therefore highly desirable.

A microbiological diagnosis enables administration of directed antimicrobial therapy, avoiding the need for broad-spectrum, potentially toxic therapeutic combinations or erroneous empirical therapy (Kumar et al., 2005; Rankine et al., 2004). In patients who have been administered antimicrobials before a spinal biopsy, a microbiological diagnosis was made in only 25% of episodes (Rankine et al., 2004), highlighting the need to withhold treatment until appropriate investigations have been carried out.
Antimicrobials should not be started until after biopsies have been taken unless the patient has severe sepsis or septic shock, or a microbiological diagnosis has been obtained via blood cultures. [Evidence level C]

Empirical therapy
Empirical therapy for VO should be avoided whenever possible. Clinical findings, such as possible primary foci of infection, in the patient with severe sepsis or septic shock will influence empirical therapy, hence such patients should be discussed with a Microbiologist on a case-by case basis. If cultures remain negative, it will be necessary to continue the chosen empirical antimicrobial regimen.

Back to top

Staphylococcal VO,.
Recommendations for therapy tend to be extrapolated from animal models of osteomyelitis and treatment of osteomyelitis in general because of the paucity of good quality data specific to VO. Combination therapy has produced markedly better results in animal models of Staphylococcus aureus osteomyelitis. A combination of Vancomycin * and Rifampicin cured 90% of infections in one model (Norden & Shaffer, 1983). Rifampicin also enhanced the activity of cephalothin and Gentamicin sterilizing bone in 90 and 95% respectively (Norden, 1975). Triple therapy with Rifampicin , cephalothin and sisomicin sterilized all bone samples after 14 days (Norden, 1975). Whether these animal models, which predominantly study infection of long bones, are applicable to vertebral infection is unknown.

Nafcillin [very similar to Flucloxacillin ] monotherapy has been compared to nafcillin plus Rifampicin in a small number of patients with chronic Staphylococcus aureus osteomyelitis (Norden et al., 1986). Combination therapy cured 8/10 [80%] while nafcillin monotherapy cured 4/8 [50%]; although not clinically significant, the study was probably underpowered and there was a definite trend in favour of combination therapy (Norden et al., 1986). Combination therapy in spinal infection has not been rigorously investigated.

In terms of newer antimicrobial agents such as Linezolid and Daptomycin , there are limited data: case reports and series describe both clinical success and failure. These agents should only be used in consultation with a Microbiologist (Hayden et al., 2005; Marty et al., 2006; Melzer et al., 2000; Rayner et al., 2004; Till et al., 2002; Vikram et al., 2005).

Suboptimal treatment can result in development of resistance to certain antimicrobials during therapy. Development of resistance to Ciprofloxacin and Rifampicin during therapy can be particularly problematic (Kragsbjerg et al., 2000).

For meticillin-susceptible staphylococci combination of intravenous Flucloxacillin 2g 6-hourly +/- oral Rifampicin * 300-600mg 12-hourly is recommended. Use regimen for meticillin-resistant staphylococci in patients with a genuine penicillin allergy or consider desensitization. [Evidence level D]

For meticillin-resistant staphylococci a combination of intravenous Vancomycin  see dosing guidelines with oral Rifampicin * 600mg 12-hourly is recommended. [Evidence level D]

*Rifampicin should never be used alone for Staphylococcus aureus infections; isolates must be susceptible to Rifampicin ; there should be no historical or radiological features suggestive of tuberculosis. Dose variation because doses need adjustment in patients with severe renal impairment. The addition of Rifampicin is optional, no good quality clinical data supports it addition.

Enterococcal VO
Reports of enterococcal vertebral osteomyelitis exist but it is an uncommon infection (Akermann, 1981; Sandoe et al., 2001; Soda et al., 1996; Summers et al., 2001; Tarr et al., 2004). Most patient have been treated with the combination of a penicillin and aminoglycoside for 4-8 weeks (Tarr et al., 2004). Perhaps unsurprisingly, courses of cefazolin, Ceftazidime , Ofloxacin , Ciprofloxacin and tosufloxacin have all failed to cure a cases of presumed Enterococcus faecalis vertebral osteomyelitis in one case report (Soda et al., 1996; Zamora et al., 1997).
For Amoxicillin-susceptible and high-level gentamicin-susceptible enterococci a combination of intravenous Amoxicillin 2g 4-hourly and Gentamicin 1mg/kg 8-12-hourly is recommended. [Evidence level D]

Treatment of enterococci with other susceptibility profiles requires discussion with a Microbiologist on a case-by-case basis. [Evidence level D]

Streptococcal VO
Evidence to support specific antimicrobial regimens generally comes from in vitro susceptibility data and case reports. Group B streptococcal vertebral osteomyelitis is uncommon but has been successfully managed with a penicillin or cephalosporin (Elhanan & Raz, 1993; Ganapathy & Rissing, 1995; Lischke & McCreight, 1990; Mouly et al., 1999). Group C streptococcal vertebral osteomyelitis is also uncommon but has been successfully treated with penicillin-based regimens (Kumar et al., 2005). Alpha-haemolytic streptococci also cause VO, generally in association with endocarditis.

For penicillin-susceptible streptococci [MIC<0.5] intravenous Benzyl penicillin 1.8-2.4g 4-hourly recommended. [Evidence level D]

Treatment of streptococci with other susceptibility profiles requires discussion with a Microbiologist on a case-by-case basis. [Evidence level D]

Enterobacteriaceae [“coliforms”]
Data from animal models of osteomyelitis caused by coliforms are generally lacking or use agents that are not in common use in the UK e.g. mezlocillin was completely ineffective in an animal model of Morganella morganii osteomyelitis (Norden & Shaffer, 1984). Although these organisms cause a significant proportion of VO cases, specific details of treatment are often lacking from repoted series (McHenry et al., 2002). Escherichia coli spinal infection with a fatal outcome has been reported (Chen et al., 1991; Muckley et al., 2003), in one case a combination of Metronidazole , cefepime, chloramphenicol and netilmicin combined with surgery failed to eradicate infection. A case of Proteus mirabilis VO and discitis was cured with an eight week course of intravenous ampicillin [8g/day] plus Gentamicin [90mg 8-hourly] followed by four months of oral Amoxicillin (Sapico & Montgomerie, 1979). A second case of Proteus mirabilis VO and discitis was cured with subtotal vertebrectomy and three weeks intravenous ampicillin followed by three months of oral ampicillin [6g per day] with oral probenecid [2g per day] (Sapico & Montgomerie, 1979).

“coliforms” [e.g. Escherichia coli, Klebsiella spp., Proteus spp.]
For Amoxicillin-susceptible isolates intravenous Amoxicillin 2g 6-hourly is recommended. Consider adding intravenous Gentamicin 7mg/kg [frequency according to Trust policy] depending on renal function, comorbidity and organism susceptibility. [Evidence level D]

For Amoxicillin-resistant, ciprofloxacin-susceptible isolates oral Ciprofloxacin 500-750mg 12-hourly is recommended. [Evidence level D]

Infections caused by organisms with other susceptibility profiles should be discussed with microbiology on a case-by-case basis. [Evidence level D]

Pseudomonas aeruginosa
Animal models of Pseudomonas aeruginosa osteomyelitis have shown a 6%, 15%, 94% and 94% success rate with 28 days Tobramycin , Aztreonam , Ofloxacin , and Ciprofloxacin therapy, respectively (Norden & Shinners, 1985; Norden & Niederriter, 1987; Norden & Budinsky, 1988). Carbenicillin in combination with sisomicin was more effective than monotherapy with either agent in another experimental study (Norden & Keleti, 1980). A combination of azlocillin and Tobramycin for 28 days cured only 60% of infections but combination therapy was more effective than either agent alone (Norden & Shaffer, 1982). Reduced susceptibility to Ciprofloxacin was seen in 20% [2/10] animals after treatment with Ciprofloxacin for two weeks (Norden & Shinners, 1985). Comparable animal studies suggest that quinolones are the most effective agents in this setting.

Clinical cases have been treated with a variety of regimens. (Akermann, 1981; McGee-Collett & Johnston, 1991).

For ciprofloxacin-susceptible strains, oral Ciprofloxacin 500-750mg 12-hourly is recommended. [Evidence level D]

Mycobacterium tuberculosis
Spinal tuberculosis should be treated with at least six months of the antimicrobial regimens usually used to treat respiratory tuberculosis (Joint Tuberculosis Committee of the British Thoracic, 1998). [Evidence level A]. In practice, treatment courses of a year or more may be required.

Actinomyces spp., Brucella spp. and fungal deep spinal infections are rare occurrences in Leeds and should be treated on a case-by case basis following discussion with microbiology.

2. Infected spinal instrumentation
Animal models of various medical device infections have shown superior cure rates with combination therapy when compared to monotherapy. There is a dearth of published information specific to the appropriate antimicrobial therapy of this infection. Cure of infected spinal instrumentation in Leeds has been achieved by extrapolating from established practice in the treatment of other medical device infections such as prosthetic valve endocarditis and prosthetic joint infection. In the absence of evidence cases need to be managed individually based on the causative organism and its susceptibility.

3. Epidural abscess.
Surgical drainage is key. Antimicrobial regimens as for VO and discitis are appropriate, except that earlier switch to oral therapy and less prolonged courses may be required.

Antimicrobial allergy
contact microbiology for advice if not stated above.

Back to top

Two weeks treatment is usually ineffective in sterilizing bone in animal models. Observational data from patients indicates that short course therapy [<4 weeks] has been associated with treatment failure (Kumar et al., 2005; Sapico & Montgomerie, 1979). Six to eight weeks intravenous antimicrobial therapy [without prolonged oral antimicrobial therapy] is recommended by many authorities (Hadjpavlou et al., 2000; Musher et al., 1976; Osenbach et al., 1990). Such therapy is usually curative provided abscesses have either been drained or seen to resolve radiologically, pain at rest has significantly improved [mechanical pain may persist, depending on instability] and inflammatory markers have markedly improved (Sapico & Montgomerie, 1980). A recent comparative study of 120 episodes of VO showed that relapse was no more likely in patients treated for 6 weeks than those treated for >6 weeks (Roblot et al., 2007). Relapse of infection has been associated with draining sinuses and paravertebral abscesses and more prolonged therapy with surgical intervention may be necessary in such patients (McHenry et al., 2002).

Six weeks treatment is recommended in the first instance. If the patient is afebrile, CRP is returning to normal, paravertebral abscesses have been drained or resolved and rest pain is absent or greatly reduced, then antimicrobials can be stopped. Failure to improve according to these criteria should lead to continuation of treatment, further investigation, such as repeat imaging of the spine, and reconsideration of surgical intervention.

Back to top

There is no evidence to support either a switch to oral antimicrobials during therapy or continuation of oral therapy after an appropriate course of intravenous treatment for the majority of situations (Sapico & Montgomerie, 1980; Sapico & Montgomerie, 1990; Sapico, 1996). In cases where the antimicrobial of choice has high oral bioavailability [such as Ciprofloxacin ], oral therapy may be appropriate. For some low virulence pathogens, such as Eikenella spp. or Kingella spp. conversion to oral therapy may be appropriate after a good clinical response to intravenous therapy. Infections caused by Mycobacterium tuberculosis, Actinomyces and Brucella spp. for which prolonged therapy is required, can be treated with oral therapy (Sapico & Montgomerie, 1980).

Switching to oral antimicrobials should be considered on a case by case basis, but is generally not recommended for Staphylococcus aureus infection or infections involving spinal metalwork. [Evidence level D].

Back to top

Back to top