Osteomyelitis

Meet the patient

A 14-year-old boy is on the ward with three days of fever and a right leg he refuses to put weight on. He is exquisitely point-tender over the distal thigh, just above the knee, where the distal femoral metaphysis sits. His X-ray is normal — and the registrar is halfway through writing "growing pains, discharge home" on the round list.[4]

The three questions that will follow this boy, and every osteomyelitis patient you ever meet, are the spine of this whole topic: what is the diagnosis hiding behind a normal film? (image with MRI, not a repeat X-ray), what is the organism? (Staphylococcus aureus until bone culture proves otherwise — never a sinus swab), and will antibiotics alone be enough? (yes for early acute disease, no once bone is dead — then you debride). Hold those three and every section below slots into place.[1]

The one sentence that frames everything

Every osteomyelitis decision orbits one biological fact: dead bone has no blood supply, so no antibiotic and no immune cell can reach it. The sequestrum — the fragment of cortex that has died and separated off — is a permanent reservoir of infection, and the biofilm coating it is near-impenetrable. That single mechanism is why acute disease can be cured medically but chronic disease cannot, why infected implants are so refractory, and why the surgeon's debridement knife is the definitive treatment.[2]

The word itself names the territory: osteo (bone) + myelo (marrow) + itis — the infection lives in the marrow cavity, not just on the cortex. Set it apart from septic arthritis (joint), cellulitis (skin), and a soft-tissue abscess — all of those still have a blood supply and respond to antibiotics or drainage alone. Bone that has died does not.[2]

The clinical skill here is never recognising the florid, draining, X-ray-positive case. It is recognising the early case behind a normal film, choosing MRI over a repeat X-ray, securing a bone biopsy instead of trusting a sinus swab, and knowing when to stop expecting antibiotics to finish a job only surgery can.[1][9]

Three axes — pathogenesis, tempo, anatomic stage

Osteomyelitis is sorted along three independent axes, and an examiner probes all three. Classify by how the bug got in (pathogenesis), by how long it has been there (tempo), and by how much bone is dead (anatomic stage) — the last one decides the operation.[1]

Axis one — pathogenesis (the Lew-Waldvogel classification) sets the empirical antibiotic spectrum and the surgical plan:[2]

• Haematogenous — bacteraemic seeding of richly vascular bone. In children, the long-bone metaphysis (femur, tibia) is the target; in adults, the vertebra. Single organism, almost always Staphylococcus aureus.
• Contiguous focus — spread from adjacent soft tissue: a diabetic ulcer, a surgical wound, an open fracture, a joint replacement. Mixed flora including Gram-negatives and anaerobes; needs debridement.
• Chronic — the disease persists or relapses over months to years; the hallmark is dead separated bone (sequestrum), new periosteal bone (involucrum), a cortical cloaca, a draining sinus tract, and biofilm on necrotic tissue or implants.[1]

Axis two — tempo:[1]

• Acute (within about 2 weeks) — suppurative and neutrophilic; curable medically if treated before necrosis sets in.
• Subacute (weeks to months) — less virulent; the Brodie abscess (an intraosseous cavity with a sclerotic rim) is the classic lesion.
• Chronic (months plus) — defined histopathologically by dead bone and biofilm; it relapses.[1]

Axis three — anatomic stage (Cierny-Mader) crosses an anatomic type with a host class, and it is what the surgeon reads before picking an operation:[3]

[3]

Who gets it — and the organism that always wins

Whatever the host and whatever the textbook pairing, Staphylococcus aureus is the commonest organism at every age and every subtype. Memorise the exotic pairings for the exam — but never let them displace the default. The incidence is rising globally on the back of the diabetes pandemic, ageing populations, and more orthopaedic implants.[2]

[2]

The risk-factor-to-organism pairings (high-yield, name them by the discriminator):[1][9]

The classic trap: the single most tested pairing is sickle cell disease with Salmonella — and the examiner is waiting to see if you also know that even in sickle cell disease, S. aureus is still the commonest single organism. Infarcted bone is the nidus; cover both. The runner-up is IV drug use with Pseudomonas at unusual sites (sternoclavicular, sacroiliac, vertebral) — but again, S. aureus remains in the frame.[2]

Everyone forgets: in the neonate the transphyseal vessels are still patent, so infection crosses into the joint and septic arthritis is common — think Group B strep, E. coli, S. aureus, and expect multifocal disease. The age of the patient rewrites the bug list more than any single feature of the bone.[4]

Why bone dies — the sequestrum cascade

Bone is normally sterile, and the whole disease unfolds from one design flaw: bone sits inside a rigid cortical box that cannot expand. Once infection raises pressure inside that box, it chokes its own blood supply, kills the cortex, and manufactures an avascular sequestrum that antibiotics can never reach. Three routes of entry lead to the same cascade.[2]

The three routes in:[1]

• Haematogenous — organisms lodge in the sluggish, fenestrated hair-loop capillaries of the long-bone metaphysis (children) or the vertebral end-plate (adults, via Batson's vertebral venous plexus). Single organism, usually S. aureus.
• Contiguous spread — from an adjacent soft-tissue infection, diabetic ulcer, surgical wound, or open fracture. Mixed flora including Gram-negatives and anaerobes.
• Direct inoculation — penetrating trauma, surgery, or a foreign body (orthopaedic implant) carries organisms straight into bone.[1]

Why the metaphysis in children? The metaphyseal capillaries make sharp hair-pin loops, with sluggish flow and fenestrated endothelium without a basement membrane — bacteria slow, lodge, and multiply. The physis (growth plate) acts as a barrier: in children older than about 18 months the transphyseal vessels have closed, so infection stays confined to the metaphysis and does not reach the joint — except where the metaphysis is intracapsular (hip, shoulder, knee), where septic arthritis can result. In the neonate the transphyseal vessels are still patent, so joint involvement is common.[4]

The inflammatory cascade — the mechanism that produces every sign:[1]

1. Bacteria adhere to bone matrix via fibronectin and other adhesins, multiply, and form a biofilm (extracellular polymeric slime) on devitalised tissue.
2. Macrophages release IL-1, TNF-alpha and IL-6 → fever, malaise, acute-phase response (raised CRP/ESR).
3. Neutrophils rush in but cannot clear bacteria sheltered within necrotic tissue and biofilm; they release proteolytic enzymes that erode bone (osteolysis).
4. Rising intraosseous pressure — bone is locked in a rigid cortical envelope that cannot expand — compresses the venous and arterial supply, producing venous stasis, thrombosis and ischaemia.
5. Ischaemic cortex dies and detaches as a sequestrum — avascular, so antibiotics and immune cells cannot reach it; it becomes a permanent reservoir of infection.
6. The periosteum is elevated by subperiosteal pus and lays down new bone around the sequestrum = the involucrum.
7. Pus forces its way out through a hole in the cortex = the cloaca, and tracks to the skin as a sinus tract — the pathognomonic feature of chronic osteomyelitis.[1]

The Latin names are viva gold because they describe exactly what the pathologist sees: sequestrum — "something left behind", the dead bone separated off; involucrum — "a covering", the new periosteal bone wrapped around it; cloaca — "a sewer", the hole the pus escapes through. Learn the four together and the chronic disease draws itself.[2]

The biofilm is why chronic osteomyelitis and implant infection refuse to clear: sessile bacteria inside the slime are 10 to 1000 times more resistant to antibiotics and to phagocytosis. That is the rationale for debriding dead tissue, removing infected implants, and adding rifampin — which uniquely penetrates the biofilm and kills adherent staphylococci — to staphylococcal regimens.[7]

Vertebral osteomyelitis runs its own pattern: haematogenous seeding (via Batson's plexus) of the vertebral end-plate, then spread into the adjacent disc and through to the opposing vertebral body — hence the examiner's sign of two adjacent vertebrae and the disc between them. Its two killers are epidural abscess (a neurological emergency) and paraspinal abscess.[2]

Five faces, one disease

Presentation tracks the route, the age, the host, and the chronicity — and a high index of suspicion is essential because systemic signs are frequently absent, especially in contiguous-focus, diabetic, elderly, and immunocompromised patients. The florid febrile child is the easy case; the afebrile diabetic foot and the vaguely-unwell elderly patient are where patients are lost.[1][2]

Acute haematogenous (typically a child): rapid-onset localised bone pain over the long-bone metaphysis (distal femur, proximal tibia, distal tibia, proximal humerus), with tenderness, swelling, warmth and erythema, fever and malaise — and in younger children, refusal to use the limb (pseudoparalysis) and a limp or refusal to bear weight. The pain is often very precisely localised, unlike the diffuse pain of septic arthritis.[4][8]

Vertebral (typically an adult): insidious, progressive back pain — often worse at night and unrelated to activity — with localised spinal tenderness, low-grade or absent fever (up to 50% are afebrile), and a subacute course over weeks. Hunt for a source of bacteraemia (UTI, IV drug use, endocarditis, recent instrumentation, a line). Red flags — radicular pain, limb weakness, sphincter disturbance, saddle anaesthesia — mean epidural abscess or cord compression, a neurological emergency.[2]

Contiguous-focus / diabetic foot: chronic, localised pain, swelling, erythema and warmth over a non-healing wound or ulcer that sits over a bony prominence (the metatarsal heads in a neuropathic foot). Probe-to-bone may be positive. Systemic signs are often absent — the neuropathic, ischaemic foot mounts neither a fever nor a leukocytosis, and neuropathy blunts the pain. Exposed bone in the ulcer base is highly suggestive.[5]

Chronic: relapsing pain over months to years, a draining sinus tract (pathognomonic), low-grade fever, and intermittent discharge of pus or bone fragments. The disease can lie dormant for years and flare after minor trauma. A change in a chronic sinus — new pain, a mass, bleeding — demands biopsy to exclude Marjolin ulcer (squamous cell carcinoma in a chronic sinus tract).[2]

Prosthetic joint / implant-related: persistent joint pain, stiffness, wound drainage, a sinus tract, or radiographic loosening. Classify by timing — early (within 3 months, usually virulent S. aureus or Gram-negative), delayed (3 to 24 months, coagulase-negative staphylococci), or late (over 24 months, usually haematogenous).[7]

Atypical presentations — the ones deliberately tested:[1]

• Elderly / diabetic / immunocompromised: minimal pain, no fever, vague malaise, a non-healing wound, or just unexplained raised inflammatory markers.
• Neonate: may present as sepsis or pseudoparalysis of a limb with no localising signs; disease can be multifocal.
• IV drug user: back pain (vertebral), sternoclavicular or sacroiliac pain — easy to dismiss as musculoskeletal; always screen for endocarditis.[1]

The mimics — exclude malignancy and Charcot

A painful swollen limb or an unexplained spinal lesion is not always osteomyelitis. Distinguish each mimic by its discriminating feature, and reach for biopsy the moment the radiology is atypical or the age is wrong.[1][9]

[2]

What juniors write versus what gets marks: the trainee writes "osteomyelitis" on an atypical lytic lesion in a 16-year-old; the examiner wants osteosarcoma or Ewing sarcoma excluded by tissue biopsy before any antibiotic. Always consider malignancy when the radiology is atypical, the age is wrong for haematogenous disease, or there is no response to appropriate therapy — biopsy is mandatory whenever there is doubt.[2]

The bedside round — probe-to-bone, the sinus, the neuro exam

Examination rarely makes the diagnosis; it detects the source of bacteraemia, the complications, and the emergencies. Run the limb or the spine in a fixed order, and let the findings drive imaging.[1]

Vital signs drive triage. Fever, tachycardia, hypotension, tachypnoea and altered mentation indicate sepsis — apply the Surviving Sepsis hour-1 bundle.[1]

Focused limb examination:[1]

• Localised point tenderness over the affected bone — often very precisely localised in acute haematogenous disease (unlike the diffuse tenderness of septic arthritis or cellulitis).
• Warmth, swelling, erythema over the site; fluctuance suggests subperiosteal pus or an abscess.
• Sinus tract with discharge (chronic disease) — note its location and the character of the discharge.
• Distal neurovascular status — pulses, capillary refill, sensation, motor power. Neurovascular compromise is a surgical emergency.
• Examine the joint above and below — coexisting septic arthritis is common in children, especially at the hip.
• Refusal to bear weight / pseudoparalysis in infants and non-verbal children — a key sign.[1]

Diabetic foot — the probe-to-bone test: with sterile technique, pass a blunt metal probe through the wound to its base; if hard, gritty bone is felt, the test is positive. In high-prevalence diabetic foot clinics a positive test carries a high positive predictive value (around 90%), and sensitivity is highest when bone is visible in the ulcer base. Combine with MRI for confirmation.[5]

Spine examination: point tenderness over the affected spinous processes, then a full neurological examination — motor power, sensation (including the perineal and saddle area), reflexes, anal tone. Saddle anaesthesia, urinary retention, or bilateral leg weakness = cauda equina or epidural abscess — emergency MRI and surgical referral.[1]

Search for the source of bacteraemia: skin and soft-tissue infection, an indwelling vascular device, endocarditis (auscultate for a murmur — its association with vertebral osteomyelitis is strong), a urinary tract infection, dental infection, IV drug use injection sites.[1]

MRI not the X-ray, bone biopsy not the swab

Investigations do three jobs — confirm infection, identify the organism, and map the anatomy for surgery — and the plain X-ray is normal for the first 1 to 2 weeks, so a normal early film never excludes the disease. The two recurring mistakes are repeating the X-ray instead of moving to MRI, and culturing the sinus tract instead of the bone.[1][2]

Bloods: FBC (leukocytosis); CRP — the best marker for monitoring response, it should fall steadily with effective therapy; ESR — slow, may stay elevated for weeks, useful as a baseline; U&E, LFTs, glucose, lactate (sepsis), coagulation. Blood cultures BEFORE antibiotics — positive in about 40% of acute haematogenous disease and 70 to 90% of vertebral osteomyelitis. In suspected endocarditis-associated vertebral disease, add echocardiography.[2]

Imaging — climb the ladder:[1]

Microbiological diagnosis — the gold standard is deep bone biopsy culture. Bone is taken at surgical debridement or by percutaneous (often CT-guided) biopsy — never by swabbing the sinus tract. Send for aerobic and anaerobic culture, and for mycobacteria and fungi if risk factors dictate. Histology confirms the diagnosis (acute: neutrophils in marrow; chronic: plasma cells, lymphocytes, necrotic bone) and excludes malignancy.[2][5]

Probe-to-bone (diabetic foot) is a bedside test, described above. The IWGDF/IDSA 2023 guideline combines clinical findings (probe-to-bone, exposed bone) with imaging (MRI) and, where feasible, bone culture before committing a patient to prolonged antibiotics.[5]

Investigations by scenario: vertebral disease — MRI plus CT-guided biopsy plus blood cultures plus echocardiography; paediatric — bloods, blood culture, ultrasound of the adjacent joint for an effusion, MRI under sedation; prosthetic joint — joint aspiration (synovial WCC, culture), blood cultures, PET-CT.[1]

[1]

Cultures first, but not in sepsis — ABCDE

Resuscitation is the Surviving Sepsis hour-1 bundle when the patient is septic, and a culture-first approach when they are stable — the two are in deliberate tension, and you weigh them against the clinical urgency. There is no room for delay in sepsis.[1]

ABCDE first. If the patient is septic or systemically unwell, apply the Surviving Sepsis hour-1 bundle:[1]

• Oxygen to target SpO2 94 to 98% (or 88 to 92% in COPD or CO2-retention risk).
• Blood cultures BEFORE antibiotics (where feasible) — at least two sets, plus local (bone or joint) cultures.
• Empirical IV antibiotics within 1 hour of recognising sepsis.
• Lactate; balanced crystalloid 30 mL/kg if hypotensive or lactate over 2 mmol/L; noradrenaline for fluid-refractory shock; reassess fluid responsiveness before further boluses.
• Analgesia, and splint or elevate the affected limb in the position of comfort.[1]

Spinal epidural abscess or cord compression with a new neurological deficit is a separate, time-critical emergency. It needs urgent MRI of the whole spine and neurosurgical or orthopaedic referral for decompression within 24 to 48 hours to preserve neurological function. Dexamethasone is sometimes given for cord oedema but is secondary to surgical decompression.[2]

Debride the dead bone — the antibiotic ladder and the 4-to-6-week fork

Osteomyelitis is a co-management disease: orthopaedic surgery for source control, infectious diseases for culture-directed antibiotics — and the unifying rule is that dead bone cannot be cured by antibiotics. Cover S. aureus first, broaden by scenario, then narrow to culture, and add rifampin for the staphylococcal biofilm.[2][3]

Empirical antibiotic choice — cover Staphylococcus aureus and tailor to the scenario:[1][9]

Definitive therapy — narrow to culture sensitivities once bone culture returns. For staphylococcal chronic osteomyelitis, add rifampin 600 to 900 mg PO once daily to a beta-lactam or a fluoroquinolone (for example ciprofloxacin 750 mg PO 12-hourly plus rifampin) for at least 6 weeks — rifampin uniquely penetrates the biofilm and kills adherent staphylococci. The evidence base has been challenged (the Renz 2021 controversy), but the overall evidence and current guidelines still support its use in staphylococcal implant infection.[7]

Duration:[1]

• Acute osteomyelitis: 4 to 6 weeks total (commonly 2 weeks IV, then an oral step-down once afebrile, improving, and CRP is falling).
• Chronic osteomyelitis: 6 weeks minimum, often longer (up to 6 months in some); guided by clinical response and the CRP trend.
• Vertebral osteomyelitis: typically 6 weeks.
• Paediatric acute haematogenous: the Peltola / Pääkkönen data support a shortened IV course (3 to 4 days) and an early oral switch — a total of 3 to 4 weeks is sufficient in uncomplicated cases.[4][8]

IV-to-oral step-down criteria: haemodynamically stable, afebrile for 24 to 48 hours, CRP falling, able to swallow and absorb, and a bioavailable oral agent available (ciprofloxacin plus rifampin, clindamycin, linezolid, doxycycline). OPAT (outpatient parenteral antimicrobial therapy) is a well-validated option for completing the IV course at home. Honest caveat: the Cochrane review found insufficient high-quality RCT evidence on antibiotic choice and duration for chronic disease, so practice here is largely expert- and guideline-based, not trial-driven.[6]

Surgical management — the principles of source control, applied by Cierny-Mader stage:[3]

1. Drain the abscess (subperiosteal, intraosseous, paraspinal).
2. Debride all necrotic bone back to bleeding, healthy bone — the paprika sign, punctate bleeding from viable cortical bone, marks adequate debridement.
3. Remove infected hardware if possible — the biofilm on it is incurable.
4. Manage the dead space — an antibiotic-laden polymethylmethacrylate (PMMA) cement spacer (often loaded with vancomycin plus gentamicin), or a biodegradable carrier; later replaced with bone graft and/or a vascularised soft-tissue or muscle flap.
5. Stabilise any pathological fracture or instability.
6. Provide soft-tissue cover — often by plastic surgery (a free or rotational flap), particularly for an exposed tibia and the diabetic foot.
7. Revascularise if the limb is ischaemic (diabetic or vascular disease) — antibiotics cannot reach an ischaemic limb.[1]

Chronic disease refractory to standard therapy: consider hyperbaric oxygen therapy as an adjunct (controversial; not first-line), and long-term suppressive antibiotics if the patient is not fit for surgery (Host C).[3]

Prosthetic joint infection — the Zimmerli / IDSA pathway:[7]

• Acute (under 4 weeks, mobile implant): debridement, antibiotics and implant retention (DAIR) plus rifampin for staphylococcal disease, for 6 weeks, then suppressive therapy.
• Chronic (over 4 weeks, loose implant): two-stage exchange — remove the implant, place an antibiotic-loaded cement spacer, give 4 to 6 weeks of antibiotics, then reimplant. One-stage exchange in selected centres with good soft tissue.
• Not operable: lifelong suppressive antibiotics.[1]

The subtypes that bite

Each subtype rewrites the bug, the site, and the operation — name them by their discriminating feature and the marks follow.[2]

• Acute haematogenous in children — long-bone metaphysis (distal femur, proximal or distal tibia, proximal humerus); S. aureus; often medical (4 to 6 weeks of antibiotics) if there is no abscess or sequestrum, with surgery reserved for subperiosteal pus, a sequestrum, or failure to improve in 48 to 72 hours. The Peltola / Pääkkönen shortened-IV data underpin the modern approach.[4][8]
• Vertebral / discitis (adults) — back pain, two adjacent vertebrae and the disc, end-plate oedema; risk of epidural or paraspinal abscess; check for endocarditis; 6 weeks of targeted antibiotics; surgery only for instability, cord compression, or failure of medical therapy.[2]
• Contiguous-focus (post-trauma, post-surgical) — mixed flora including Gram-negatives and anaerobes; surgical debridement is essential; remove infected hardware if possible.[3]
• Diabetic foot — multifactorial (neuropathy, vascular disease, a pressure ulcer over a bony prominence); mixed flora; probe-to-bone, MRI; combined surgical (resection or amputation) plus prolonged antibiotics; revascularise if ischaemic; a multidisciplinary foot team; offloading (a total contact cast). The IWGDF/IDSA 2023 guideline is the framework.[5]
• Chronic — sequestrum, involucrum, cloaca, sinus tract, biofilm; surgical; the risk of Marjolin ulcer.[2][3]
• Brodie abscess (subacute) — an intraosseous cavity with a sclerotic rim; chronic local pain; surgical curettage.[2]
• Prosthetic joint — biofilm; DAIR versus one- or two-stage exchange; rifampin for staphylococci.[7]
• Salmonella in sickle cell disease — the classic exam answer; but S. aureus is still commonest; infarcted bone is the nidus.[2]
• Neonatal — Group B streptococcus, E. coli, S. aureus; transphyseal vessels are patent, so joint involvement (septic arthritis) is common; often multifocal.[4]
• Tuberculous (Pott disease of the spine) — chronic and destructive; a paraspinal cold abscess; MRI; tissue diagnosis (AFB, culture, PCR); standard anti-tubercular therapy.

When bone infection kills you later — complications and drug toxicity

Complications separate by mechanism, and the mechanism is what makes each one examinable — chronic relapse from inadequate debridement, growth arrest from physeal damage, cord compression from vertebral extension, and the long-latency malignancy in a sinus tract.[2]

Disease-related: chronic relapsing infection (inadequate debridement), septic arthritis (extension through physis or cortex), joint destruction and growth arrest or limb-length discrepancy in children (physeal damage), pathological fracture through weakened or lytic bone, epidural abscess, cord compression, cauda equina (vertebral), sepsis, septic shock, metastatic infection (endocarditis, brain abscess), amputation (diabetic foot), and death.[2]

Marjolin ulcer — squamous cell carcinoma arising in a chronic draining sinus tract, typically after 10 to 30 years of chronic osteomyelitis. Any change in a chronic sinus — new pain, a mass, bleeding, increased discharge — mandates biopsy. The lesion is often aggressive; wide local excision and staging follow.[2]

Secondary amyloidosis (from long-standing chronic suppuration) and renal disease from prolonged nephrotoxic antibiotics (vancomycin, aminoglycosides) are the long-term complications examiners reach for after the acute ones.[2]

Antibiotic-related complications — name the drug and its signature toxicity:[7]

• Vancomycin — nephrotoxicity (monitor troughs and renal function).
• Aminoglycosides — nephro- and ototoxicity.
• Rifampin — hepatotoxicity and potent CYP3A4 induction (it slashes the efficacy of oral contraceptives, warfarin, DOACs, and antiretrovirals — warn the patient and adjust).
• Linezolid — myelosuppression, peripheral and optic neuropathy, and serotonin syndrome (it is a reversible MAO inhibitor).
• Fluoroquinolones — tendinopathy and rupture, and QT prolongation.[7]

The classic pitfalls — the recurring trainee errors:[1]

• Accepting a normal early X-ray as excluding osteomyelitis — it does not; image with MRI.
• Trusting a sinus tract swab as the organism — it is contaminated; obtain bone biopsy culture.
• Treating chronic osteomyelitis with antibiotics alone — dead bone cannot be sterilised; debride.
• Forgetting to add rifampin for staphylococcal implant or biofilm infection.
• Missing an epidural abscess in vertebral disease with new leg weakness.
• Missing a Marjolin ulcer in a chronic sinus that has changed.
• Not screening for endocarditis in vertebral osteomyelitis.
• Not addressing ischaemia before expecting antibiotics to work in the diabetic foot.[1]

Prognosis and the CRP that tells you when to re-debride

Outcome is set by organism virulence, host fitness, the site, and — above all — the speed of diagnosis and the adequacy of surgical source control. S. aureus is the most virulent; diabetes, immunosuppression, and vascular disease all worsen the outlook; the vertebra carries the highest mortality.[1][2]

• Acute haematogenous in children: excellent with prompt therapy — a cure rate over 90%; chronicity 5 to 15%.[4]
• Vertebral: mortality 2 to 12%; neurological recovery depends on the speed of decompression for those with epidural extension.[2]
• Diabetic foot: high recurrence and amputation (10 to 30% if ischaemic or inadequately treated); 5-year mortality approaches 50%, comparable to many cancers.[5]

The prognostic marker that earns its keep is the CRP trend. A falling CRP with clinical improvement (afebrile, less pain, less drainage) means response; persistent fever, a rising CRP, or ongoing drainage means inadequate source control and mandates re-debridement — do not "just extend the antibiotics".[1]

Disposition: IV antibiotics (often via OPAT) for 2 to 6 weeks, an oral step-down to complete a 4 to 6 week (acute) or longer (chronic) course, and long-term follow-up — recurrence can surface years later, especially in diabetics and after implant retention. Multidisciplinary follow-up spans orthopaedics, infectious diseases, diabetology, podiatry, and plastic surgery as indicated.[1]

Special populations

Examiners use these scenarios heavily on NEET-PG and INICET — each one changes the bug, the dose, or the imaging threshold.[1]

• Children — weight-based dosing: cefazolin 50 to 100 mg/kg/day IV in 3 divided doses, or flucloxacillin 25 to 50 mg/kg IV 6-hourly; clindamycin or vancomycin if MRSA. Remember the physeal barrier and the risk of coexisting septic arthritis (especially at the paediatric hip, where the metaphysis is intracapsular). The Kocher criteria distinguish septic arthritis from transient synovitis in the limping child.[4][8]
• Diabetes mellitus — a low threshold to image (MRI); probe-to-bone; address ischaemia — revascularisation may be required before antibiotics can work; a multidisciplinary foot team; offloading (a total contact cast); combined surgical and medical therapy.[5]
• Pregnancy — avoid tetracyclines (fetal bone and teeth), fluoroquinolones (relative caution), and aminoglycosides (fetal ototoxicity); beta-lactams are safe; use weight-based dosing with the increased GFR. Surgical decisions balance maternal benefit against fetal risk.
• Elderly — atypical presentation; vertebral disease is common; keep malignancy in the differential; lower the threshold for imaging; beware drug toxicity with declining renal function (dose-adjust vancomycin and beta-lactams).[1]
• Immunocompromised (neutropenia, transplant, HIV) — broader empirical cover including Pseudomonas, fungi (Aspergillus, Candida), and atypical mycobacteria; a low threshold for biopsy.
• Sickle cell disease — Salmonella classically; but S. aureus is still commonest; infarcted bone is the nidus; cover both.
• Anticoagulated — balance surgical bleeding risk against thromboprophylaxis; bridging may be required; chronic infection itself is prothrombotic.[1]
• IV drug users — Pseudomonas; unusual sites (sternoclavicular, sacroiliac, vertebral); screen for endocarditis and blood-borne viruses; harm reduction.

The evidence that built the framework

Osteomyelitis practice is built more from landmark reviews and guidelines than from large RCTs — and the honest answer to "what is the evidence?" names both the pillars and the gaps.[2]

The landmarks, and what each changed:[2]

• Lew and Waldvogel, Lancet 2004 — the definitive modern review defining the pathogenic classification (haematogenous versus contiguous versus chronic) that still structures practice.[2]
• Cierny-Mader, Clin Orthop 2003 — the anatomic-host staging system that guides surgical decision-making in adult chronic osteomyelitis.[3]
• Peltola and Pääkkönen, NEJM 2014 — established that short IV courses (3 to 4 days) and an early oral switch are safe in uncomplicated paediatric acute osteomyelitis, transforming paediatric practice away from prolonged IV lines.[4]
• IWGDF/IDSA 2023 guideline — the global standard for diabetic foot infection diagnosis and treatment; it endorses probe-to-bone, MRI, and a limited duration of antibiotics after surgical resection (no more than 1 week of antibiotics after a clean amputation margin).[5]
• Conterno, Cochrane 2013 — found insufficient high-quality RCT evidence on antibiotic choice and duration for chronic osteomyelitis; practice is therefore largely expert- and guideline-based, not RCT-driven.[6]
• Renz and Zimmerli, 2021 — a critical appraisal of the rifampin in biofilm evidence; it concluded that the evidence overall supports rifampin in staphylococcal implant infections but flagged methodological concerns and the need for better trials.[7]

Where the evidence is genuinely weak: the optimal duration of antibiotics for chronic osteomyelitis; rifampin in non-staphylococcal biofilm; hyperbaric oxygen as an adjunct; one- versus two-stage exchange for prosthetic joint infection; and surgical versus medical management of uncomplicated vertebral osteomyelitis.[6]

Mnemonics and the numbers

Three mnemonics carry the chronic disease, the imaging ladder, and the trap list — learn them as scenes, not letters.[2]

[2] [1] [1]

How osteomyelitis patients come to harm (the preventable list)

Every harm in this disease is a failure of one of five habits — image early, culture the bone, debride the dead, screen the source, and revascularise the ischaemic limb. Name each preventable death and you have named the standard of care.[1]

• Discharging the limping child because the first X-ray is normal — the X-ray lags 1 to 2 weeks; the harm is a missed acute osteomyelitis that progresses to septic arthritis and growth arrest.[4]
• Trusting a sinus tract swab and treating the wrong organism for weeks, while the real pathogen smoulders in dead bone.[2]
• Treating chronic osteomyelitis with antibiotics alone — dead bone cannot be sterilised; the disease relapses, sometimes years later.[3]
• Missing an epidural abscess in a vertebral patient who develops new leg weakness — a delay past 24 to 48 hours costs neurological function permanently.[2]
• Missing a Marjolin ulcer in a chronic sinus that has changed — squamous cell carcinoma after 10 to 30 years, missed until it is unresectable.[2]
• Not screening for endocarditis in vertebral osteomyelitis — the metastatic infection is the real killer.[2]
• Not revascularising the ischaemic diabetic foot before expecting antibiotics to work — no blood supply means no antibiotic delivery, and the foot is lost.[5]

Ward-round test — three stems, thirty seconds each

Each stem is a pattern the viva repeats — answer in your head, then reveal the model answer.[1]

Stem 1 — the limping boy with a normal X-ray (answer)

A 14-year-old boy has three days of fever and progressive, severe pain over the distal thigh, just above the knee, with localised tenderness, warmth, and a limp. His X-ray is normal. What is the diagnosis, the next best imaging test, and the empirical antibiotic? Model: This is acute haematogenous osteomyelitis of the distal femoral metaphysis — the normal X-ray does not exclude it, because films lag 1 to 2 weeks. MRI is the next best test (sensitivity and specificity about 90%). Send blood cultures and CRP first, then start flucloxacillin or cefazolin IV, adding vancomycin if there is MRSA risk. In an uncomplicated child, the Peltola/Pääkkönen data allow a short IV course (3 to 4 days) and an early oral switch, totalling 3 to 4 weeks.[4][8]

Stem 2 — the diabetic foot with probe-to-bone positive (answer)

A 60-year-old diabetic has a chronic plantar ulcer over a metatarsal head that will not heal. You pass a sterile blunt probe through the wound and hit hard, gritty bone. What is the diagnosis, the confirmatory test, and the management principle the IWGDF/IDSA guideline adds? Model: Diabetic foot osteomyelitis — a positive probe-to-bone test in a high-prevalence clinic carries a positive predictive value around 90%. Confirm with MRI, and obtain bone culture where feasible (not a sinus swab). Management is combined surgical and medical — resection or amputation of infected bone plus culture-directed antibiotics; revascularise if the limb is ischaemic, provide offloading (a total contact cast), and run a multidisciplinary foot team. The IWGDF/IDSA 2023 point: after a clean amputation margin, no more than 1 week of antibiotics is needed.[5]

Stem 3 — the IV drug user with back pain and new leg weakness (answer)

A 34-year-old intravenous drug user has had three weeks of worsening lower-back pain. Overnight he develops bilateral leg weakness and cannot void. Temperature 38.2, CRP 180. What is the diagnosis, the time-critical action, and the organism cover? Model: This is vertebral osteomyelitis complicated by a spinal epidural abscess with cord compression — a neurological emergency. The time-critical action is urgent MRI of the whole spine and neurosurgical or orthopaedic referral for decompression within 24 to 48 hours; dexamethasone is secondary to surgery. Cover S. aureus and Pseudomonas (piperacillin-tazobactam or ceftazidime, plus vancomycin for MRSA), then narrow to bone or blood culture. Always screen for endocarditis with echocardiography and further blood cultures — the metastatic infection is the real threat.[2]

The mantra, and the viva-honesty line

The mantra: Staph is always the default, MRI not the X-ray, bone biopsy not the swab, debride the dead bone, add rifampin for the biofilm.[1][2]

References

1. [1]Hatzenbuehler J, Pulling TJ. Diagnosis and management of osteomyelitis Am Fam Physician, 2011.PMID 22046943
2. [2]Lew DP, Waldvogel FA. Osteomyelitis Lancet, 2004.PMID 15276398
3. [3]Cierny G 3rd, Mader JT, Penninck JJ. A clinical staging system for adult osteomyelitis Clin Orthop Relat Res, 2003.PMID 12966271
4. [4]Pääkkönen M, Peltola H Acute osteomyelitis in children N Engl J Med, 2014.PMID 24693913
5. [5]Senneville E, Albalawi Z, van Asten SA, et al. IWGDF/IDSA guidelines on the diagnosis and treatment of diabetes-related foot infections (IWGDF/IDSA 2023) Diabetes Metab Res Rev, 2024.PMID 37779323
6. [6]Conterno LO, Turchi MD. Antibiotics for treating chronic osteomyelitis in adults Cochrane Database Syst Rev, 2013.PMID 24014191
7. [7]Renz N, Zimmerli W. Controversy about the Role of Rifampin in Biofilm Infections: Is It Justified? Antibiotics (Basel), 2021.PMID 33562821
8. [8]Shapiro K, Carrillo-Marquez MA, Arnold SR Diagnosis and Management of Acute Osteoarticular Infections: Summary of New Guidelines Pediatr Rev, 2025.PMID 40306703
9. [9]Carek PJ, Dickerson LM, Sack JL. Diagnosis and management of osteomyelitis Am Fam Physician, 2001.PMID 11430456