Why your treatment of osteomyelitis is wrong

I believe that osteomyelitis is the most frequently missed and frequently undertreated infection managed by Infectious Disease doctors. A bold statement, I know, and it doesn’t even get into the realm of hardware- and device-related infections – that’s a discussion for another time. In a previous post, I made the case for how and why well-intentioned clinicians can miss bone infections underlying chronic wounds. This time, I want to address the second part of my assertion – I believe our current approach to osteomyelitis treatment is wrong because it neglects to address the pathophysiology of this disease.

The old method

Even when we do recognize the presence of osteomyelitis, I mostly see a “set it and forget it” approach among my colleagues – treat with 4-6 weeks of IV antibiotics (depending on how long you suspect osteomyelitis has been present) and then call it good. When patients return with the same organism in the same place, we might give them one more round of IV therapy with an extra 2 weeks tacked on.

If they come back a third time, it’s hopeless, right? You might as well amputate the affected bone. That approach might work for a diabetic foot, but what about chronic osteomyelitis of the skull? Not so easy to amputate that extremity. People can generally get along pretty well without one foot, but what about their dominant hand?

Challenging the rule of thumb

One of the cases that helped shape my views on osteomyelitis was a young, otherwise healthy patient who suffered an open fracture of his dominant thumb on the job and ended up with MSSA osteomyelitis. I started out with the “correct” treatment, using 6 weeks of IV cefazolin, but by the end of that treatment period, he still had significant swelling, redness and a persistent sinus tract. So we extended to 8 weeks. Then 10 weeks.

At that point, we had the difficult conversation where I told him I wasn’t sure we could salvage his thumb. He was determined, however, and I was willing to venture into poorly-charted territory as long as he understood and accepted the risks. IV therapy was interfering with his day-to-day existence, so we agreed to try oral Bactrim and rifampin based on a few case reports I found on persistent MSSA osteomyelitis.

One month later, his redness and swelling were showing gradual improvement. Another month later, his redness and swelling had essentially resolved, but he had a persistent draining sinus over the fracture site. After we conferred with his hand surgeon, the patient underwent a local exploration in which he was found to have a piece of floating sequestrum over a distal phalanx that was otherwise healing. Once this sequestrum was removed, he healed very quickly, and the drainage stopped.

About another month after surgery, his thumb was nearly normal with only a slight residual deformity left by the fracture, so we stopped his antibiotics at that point. I had him return for two follow up visits over the ensuing months, and each time, I breathed a sigh of relief when his thumb looked stable. He was able to go back to work with no duty restrictions, and four years later, he hasn’t returned with any changes of concern.

Not as hopeless as you might think

I believe a lot of doctors would have told my patient above that there was nothing more they could do at 10 weeks and advised him to go for the amputation. Given the disease process, it’s not an unreasonable approach to take. Having passed that point of hopelessness and ultimately achieving a cure, however, I had to ask myself – how many more limbs could we salvage using longer courses of antibiotics? Maybe our “standard” course just isn’t enough for a lot of people. Subsequent experiences have convinced me that this is absolutely true.

Look back through the medical literature over decades and you’ll find dozens of case reports of seemingly incurable chronic osteomyelitis finally relenting after very long courses of antibiotics – 3 months, 6 months, a year or even longer in some cases. Chronic osteomyelitis is defined by the formation of sequestrum, and I believe this can happen a lot sooner than many of us realize – perhaps after just weeks of infection. Once a bone infection progresses to this stage, it markedly changes the dynamics of treatment.

Why longer courses make sense

Dead bone is problematic, not only for its lack of blood flow, but also because the environment of very low metabolic activity lulls bacteria into their stationary phase. As long as the sequestrum isn’t particularly large, the blood flow issue can be overcome by adequate antibiotic dosing, allowing the drug to diffuse into necrotic bone from adjacent healthy tissues. Dormant bacteria, however, are a bigger problem, since most of our antibiotics only work to kill bacteria while they are growing and dividing.

Osteomyelitis requires very long courses of antibiotics because we are effectively waiting out a small number of senescent bacteria hiding in damaged bone tissue – a similar phenomenon to what we see in hardware infections, where dormant bacteria hide within a layer of biofilm over the prosthetic material. This model fits with what we observe in clinical practice – patients may heal and appear to be doing well on antibiotics, but as soon as we remove the antibiotic pressure, any remaining dormant bacteria start to wake up and reproduce. Most often, this manifests as gradual recrudescence of redness, pain and swelling in the weeks following discontinuation of antibiotics. Left unchecked, the reanimated bacteria start to spill out of the bone and dissect through the soft tissue out to the skin, forming a new draining wound at or near the site of the original wound. Sound familiar?

How long is long enough?

Assuming my theory is correct here (and I hope I’m starting to convince you that it is), how can we know when it’s safe to stop antibiotic therapy? Personally, I’ve settled on two endpoints that appear to produce reliably good outcomes. The first is normalization of C-reactive protein (CRP). I follow sedimentation rate (ESR) as well, but I use it as a secondary marker, since I find that it remains persistently elevated more often than CRP for reasons like advanced age and underlying autoimmune disease. ESR is also less prone to wide fluctuations and tends to track more consistently with clinical progress in the setting of osteomyelitis.

The utility of inflammatory markers

Functionally, I look at CRP first – if it’s consistently trending down week to week, we’re good. If CRP suddenly spikes up but ESR doesn’t change much, I start looking for other causes of the increased CRP – things like a viral illness that week or maybe an outpatient surgical procedure the day before labs were drawn. If both CRP and ESR increase – particularly if it happens twice in a row – I get concerned that osteomyelitis is not being adequately treated.

Why might the labs start to worsen in the middle of an antibiotic course carefully selected from culture data? I would say it’s because new bacteria get introduced through an open wound. A few times, I’ve managed to prove it with an updated wound culture. More often, I find myself empirically broadening the antibiotic regimen, and it works in a majority of cases to get the inflammatory markers trending in the right direction again.

In my earlier days of practice, I went with the status quo on osteomyelitis treatment, stopping therapy at 4-6 weeks regardless of lab markers. After following multiple patients through recrudescence, I started to realize that one thing they usually had in common was a persistently elevated CRP at the end of previous therapy. Once I started continuing antibiotics to normal CRP, I didn’t see nearly as many patients returning in short order with the same infection.

Limitations of ESR and CRP

But what about those cases where patients start out with normal inflammatory markers? My young patient with MSSA osteomyelitis in his thumb fell into this category. Let’s consider for a moment why inflammatory markers might be totally normal in a situation where there is clearly ongoing infection. Inflammation is produced by an immune response to a perceived invader. To me, cases of osteomyelitis starting out with normal inflammatory markers are actually more concerning because normal markers indicate that the immune system hasn’t been able to locate and attack the relevant pathogen. We already discussed how the formation of sequestrum in osteomyelitis limits immune cell access to the causative bacteria, and I believe this is what explains the missing inflammation.

This brings us to my second primary clinical endpoint – wound healing. I have said before that, in my experience, soft tissue will not grow reliably over infected bone. So, when markers start out normal, I think the next most logical goal is to treat until the wound has been filled in with healthy soft tissue. Getting the skin to close over the wound is ideal, especially in diabetic foot infections, which are highly susceptible to seeding with new bacteria. In cases of true chronic osteomyelitis with sizable sequestrum, this might require months of therapy, but I could cite quite a few “hopeless” cases at this point who managed to achieve cure of their osteomyelitis with very prolonged therapy.

Another success story

One recent triumph over osteomyelitis involved a middle-aged diabetic man who I met in the ICU after he nearly died from MRSA bacteremia. He seeded his sternoclavicular joint with MRSA, requiring debridement that left a sizable defect. Imaging showed residual osteomyelitis, and he was planned for additional bone debridement and flap closure of this wound. Unfortunately, his kidneys failed in the interim, and surgery was postponed while he remained on dialysis for weeks after his illness.

Given his near-death experience, a large open wound on his chest and likely residual MRSA osteomyelitis surrounding the SC joint, I did not feel that it was in his best interest to stop antibiotics until the wound was closed. We used oral doxycycline after discharge for simplicity. Three months after his admission, he came to my office to inform me that 1) he had just been able to come off dialysis and 2) he no longer needed surgery because the wound had healed up with doxycycline and local care. We stopped antibiotics at that point, and he did well – just one of many cases I could cite where patients were able to avoid major surgery for osteomyelitis using longer-than-average courses of antibiotics.

Surgery still has its place

That’s not to say that surgery is never necessary in the treatment of osteomyelitis. I’m not sure that my young guy with the thumb infection, for example, would have cleared on oral antibiotics alone without removing the sequestrum. He did not, however, have to lose his entire thumb. My point is that we are often too quick to declare osteomyelitis cases recalcitrant and recommend major, life-changing surgery without trying a longer course of therapy.

Adjust your expectations

Osteomyelitis is not incurable just because it doesn’t resolve by the end of 6 weeks of IV therapy. Of course, there will always be those cases where the sequestrum is too large, arterial flow is too poor or the organism is too virulent to stave off amputation. I just think there are many more patients who could achieve a non-surgical cure by changing our approach to antibiotics from time-based to lab- and symptom-based.

Treatment of orthopedic infections, in general, remains controversial in the field of Infectious Disease. A large part of the problem is that our already meager library of prospective studies on osteomyelitis treatment tend to stick with very regimented antibiotic interventions, assigning patients to 4, 6 or maybe 8 weeks of therapy. We see a lot of variability in outcomes, with reported success rates ranging from 60 to 90% in the absence of orthopedic hardware.

The “IV versus oral” myth

We still haven’t teased out all the factors that predict where patients fall in this range, although I would like to put one myth to rest right now. IV is not better than oral. I know, I know – it just feels better to use IV therapy. Like we’re being more aggressive and improving our patients’ chances. Studied objectively, however, outcomes are almost identical on oral therapy (provided, of course, that the gut works and we’re using an effective oral agent).

I did a literature review and found 25 publications going back 50 years on this issue. The comparison is striking. Overall success rates for IV therapy at an average of 6 weeks of treatment ranged from 60 to 90%. Success rates for oral antibiotics of similar duration? 60 to 91%.


In my experience, one size does not fit all when it comes to treating osteomyelitis. The disease exists on a spectrum from acute, focal infection of otherwise healthy bone tissue to necrosis and breakdown of an entire bone structure after prolonged infection. On the acute side, bacteria are generally behaving as expected and are much more likely to respond to shorter courses of antibiotics. On the chronic side, bacteria lie dormant deep within islands of devascularized bone tissue, impervious to many of our antimicrobials until suitable conditions get them moving again. We can’t expect to cover both ends of the spectrum with a single standardized course of antibiotics, but I believe we can achieve better outcomes by adjusting our endpoints and expectations.

17 views1 comment

Recent Posts

See All