Infectious Disease Compendium



Infection of bone. IDSA Vertebral osteomyelitis Guidelines.

Epidemiologic Risks

Trauma or hematogenous seeding (the usual triad of IVDA, dialysis and IDDM) are the main ways people acquire bone infections. It can also be chronic or acute.

Bone scans and MRI's are very sensitive, sometime to much so. Many diseases and conditions can lead to a positive MRI. Nearby cellulitis can lead to a false positive MRI as one of many examples. So never order an MRI looking for osteomyelitis in cases of recurrent cellulitis.

A tagged WBC may be of benefit for diagnosing infected fractures (PubMed).

In the adult long bone hematogenous infections are very rare and are called a Brodies abscess. They present with pain and look like a primary bone tumor until, of course, they are biopsied.

The elderly are more likely to have endocarditis, a cancer and have infection due to streptococci (PubMed).

I have NEVER seem cellulitis lead to bone infection but cellulitis will cause a false positive MRI.

Osteomyelitis with sacral pressure ulcers: debride, biopsy and get cultures. Then and only then treat. From an excellent review (PubMed) "We conducted a systematic literature review and did not find evidence of benefit of antibacterial therapy in this setting without concomitant surgical debridement and wound coverage. Furthermore, many patients with chronically exposed bone do not have evidence of osteomyelitis when biopsied, and magnetic resonance imaging may not accurately distinguish osteomyelitis from bone remodeling. The goal of therapy should be local wound care and assessment for the potential of wound closure. If the wound can be closed and osteomyelitis is present on bone biopsy, appropriate antibiotic therapy is reasonable. We find no data to support antibiotic durations of >6 weeks in this setting, and some authors recommend 2 weeks of therapy if the osteomyelitis is limited to cortical bone. If the wound will not be closed, we find no clear evidence supporting a role for antibiotic therapy."


Varies. S. aureus leads the list, but anything can infect bone, especially after trauma. A Brodies abscess is a spontaneous long bone abscess usually due to S. aureus.

Since the microbiology is variable, maybe get cultures? What a concept.

AIDS: Bartonella a common cause of lytic bone lesions as well as MAI (PubMed).

Empiric Therapy

Never, ever, ever, never treat empirically. Down that road lies madness. Therapy should be guided by cultures.

For vertebral osteomyelitis, open biopsy has higher yield than needle biopsy (29 [91%] of 32 vs 32 [53%] of 60 and there was no association of prebiopsy antibiotics with culture results (PubMed).

"Open biopsy of vertebral tissue had a higher diagnostic yield (70%) than fine needle aspirate (41%) or core biopsy (30%). Despite receiving a median of 6 weeks of intravenous antibiotics, only 15% of patients had a full recovery on discharge from index admission. .... Obtaining a microbiological diagnosis is associated with a better outcome (PubMed)."

Debride (No debridement, no cure. Know debridement, know cure) and get cultures, then wait for it. Waaiiitttttt. Then, when you have the infecting organism in your hand (figuratively speaking one hopes) treat with 6 weeks of IV something depending on the organism, and then, if possible or if the labs auger failure (ESR and/or CRP remain elevated), 3-6 months po. No one ever died of osteomyelitis acutely. If you can't get a debridement at least biopsy the damn thing; yield is not unreasonable: about 33%. If cultures are negative, consider sending a 16S ribosome test to U of Washington. Although be careful. U of W will charge you, the doctor, directly and have threatened to send me to collections when payment was slow.

And also consider a second biopsy if the first is negative, it doubles the chance of getting the bug and knowing what to kill. "To optimize microbiological diagnosis in vertebral osteomyelitis, performing a second PNB (after an initial negative biopsy) could lead to a microbiological diagnosis in nearly 80% of patients (PubMed)."

While I am too chicken to try, changing to po something (a quinolone) PLUS rifampin for S. aureus after a 2 weeks IV was successful in one series. In a small series of chronic osteomyelitis followed for 10 years, 8 weeks of oral dicloxacillin plus rifampin has the same cure as iv.

Duration?  6-8-12 weeks have all been used.  6 weeks (PubMed) probably enough for normal hosts with easy to kill bugs and good debridement.  As the confounding factors increase, paranoia tends to lengthen therapy.

S. aureus of the spine: 6 weeks for MSSA. Treat MRSA infection of the spine with 8 weeks and, if possible, debridement (PubMed) (PubMed). Keep the trough > 15. Daptomycin has efficacy equal to vancomycin (PubMed).  For spine, 6 weeks not inferior to 12 (PubMed).

Is IV therapy mandated? Maybe not if you can use oral agents with good bioavailabilty. Most oral therapy (8 to 12 weeks) has the same outcomes as iv (PubMed). IF IF IF there is good debridement (PubMed).

While I am chicken, "Switching to an oral antibiotic regimen after two weeks intravenous treatment may be safe, provided that CRP has decreased and epidural or paravertebral abscesses of significant size have been drained (PubMed).

This reference (PubMed), by the way, has all the bone levels of antibiotics.

Candida osteomyelitis - debride and 6 to 12 months of therapy (PubMed).


- I tend to avoid the quinolones where structural integrity of the bone is important. Quinolones may decrease new bone deposition by 50% (at least in animals). However, their efficacy is probably equal to beta-lactams. I also avoid aminoglycosides as they do not penetrate into bone and only give toxicity instead of efficacy.

- In the spine, bacterial infection starts in the disc and spreads to both adjacent bones. Always. Tumor involves just the bone.

- tuberculosis and fungal spondylodiscitis tends to involve the anterior vertebral body, with a predilection for paraspinal infection, and tends to spare the intervertebral disc. 20% of Candida vertebral osteomyelitis does not have disk involvement.

- Treat MRSA infection of the spine, as well as undrained paravertebral/psoas abscesses and end-stage renal disease go with 8 weeks and, if possible, debridement (Pubmed). And S. aureus in the spine should warrant a careful (as if I would suggest a sloppy) evaluation for endocarditis.

- I rarely get MRI, tagged WBC scans and bone scans when the issue is infection from trauma (and that includes the diabetic foot) as trauma will lead to positive scans. Plain films often take at least 6 weeks before there are changes. Either debride or wait.

- osteomyelitis of the jaw is best treated with SOMETHING OTHER than clindamycin and is one of the few bones that can usually be treated with po..

- Infected hardware around a fracture: the first goal is a good orthopedic result. Wash it out, keep in the hardware, GET CULTURES BEFORE ANTIBIOTICS. Look, it's osteomyelitis. No one dies of osteomyelitis (maybe cellulitis) so have some discipline and hold on the antibiotics until you have a culture, then treat with IV for 6 weeks followed by po until you get good bone healing. At that point, remove the hardware and retreat with 6 weeks iv then po in an attempt to get rid of the infection. With Staph bone/hardware infections, I would treat add rifampin to my anti-staphylococcal antibiotic.

- However, if you have a case of vertebral infection you (or your surgeon) does not need to delay placement of hardware if indicated. As long as there is pathogen-directed prolonged antibiotic therapy (PubMed) putting a CAGE into an infected back is surprisingly safe (Pubmed).

- Diabetes and smoking are the biggest risk factors for failing therapy for osteomyelitis.

- Diabetic foot osteomyelitis: the diagnosis is the cure; if you remove the infected bones, you may save the foot and you do not need long term IV therapy. If you don't remove the infected bone, then the antibiotics will do nothing, and the patient, in the long run, will lose the foot. You do the patient no favor by not being aggressive. I am a hyperbaric oxygen agnostic.

- is there osteomyelitis under that chronic stasis ulcer?  If the platelet count is > 350, you betcha (PubMed). A  sensitivity of 62.5%, specificity 91.7% and a positive predictive value of 88%.

- Curiously, there is an association between chronic osteomyelitis and dementia, increases risk by 1.73 (PubMed).

- According to the Wikipedia, osteomyelitis "has been reported in fossils of the large carnivorous dinosaur Allosaurus fragilis."


Treating bone infection without cultures is way way stupid. Don't do it. Ditto for no debridement. Stupid. Note a pattern here?

Curious Cases

Relevant links to my Medscape blog

Osteo, Cancer and Insurance


Hey. Look at my cool case.

The oldest 730.27. Ever.


Draining Pus

Ribs are back.

Odd Bug in an Odd Place

Better than a header through the windshield

Jawing Away

The Never never

Ramble on

Happy New Year

Another infection without a good solution

Consider The Heel

Save Some Money

Just Another Day at the Hospital

Almost as old as me?

Exceptions to Every Rule

Late Rupture

I Have Returned

The Real Deal

Can is not Should

Rules are made to be broken

Last Update: 01/20/19.