Fracture Healing

Fracture healing is the regeneration of bone. Other tissues heal by scar formation; bone regenerates its normal structure, through a complex cascade of cellular and biochemical events that restores its integrity.

Stability chooses the pathway. The strain at the fracture decides between primary and secondary healing. Under absolute stability (strain less than 2%), as when a compression plate eliminates micromotion, the bone behaves as if it were intact and remodels across the fracture with cutting cones. Under relative stability (strain 2-10%), as when an intramedullary nail allows micromotion, the body lays down cartilage callus to stiffen the gap.

Primary (direct) healing is healing without callus. It needs absolute stability and an anatomical reduction, and it takes one of two forms depending on the gap:

• Contact healing, gap less than 0.01mm: cutting cones cross the fracture directly, osteoclasts boring tunnels that osteoblasts fill with osteoid, healing by Haversian remodelling
• Gap healing, gap less than 1mm: woven bone fills the gap first as a scaffold, and cutting cones then remodel it

Secondary (indirect) healing is the natural form, healing by callus. It needs relative stability and vitality, and it requires micromotion. It runs through four phases:

1. Inflammation (days): haematoma forms and mesenchymal stem cells are recruited
2. Soft callus (weeks): cartilage, rich in type II collagen, bridges the gap by enchondral ossification
3. Hard callus (months): the callus mineralises into woven bone, type I collagen
4. Remodelling (years): woven bone is replaced by lamellar bone according to Wolff's law

View larger

Distraction osteogenesis is a third pathway: intramembranous ossification under tension stress (Ilizarov). It depends on four conditions:

• Stability
• Latency of 7 days
• Rate of 1mm/day
• Rhythm of 4 x 0.25mm

These pathways exist on a spectrum.

The cells. Three cells build, sense and remove bone.

1. Osteoblasts ("build") derive from mesenchymal stem cells. They secrete osteoid (type I collagen), regulate its mineralisation and express RANK-L, through which they control osteoclasts.
2. Osteocytes ("sense") are mechanosensors trapped in lacunae deep within the mineralised matrix, communicating through canaliculi. They secrete sclerostin, which inhibits bone formation; mechanical loading inhibits sclerostin and so allows bone to form.
3. Osteoclasts ("chew") are multinucleated giant cells of the monocyte/macrophage (haematopoietic) lineage. They resorb bone with acid (HCl) and proteases (cathepsin K), are activated by RANK-L and are inhibited by OPG.

The matrix. The inorganic 65% is hydroxyapatite, calcium phosphate crystals that give compressive strength. The organic 35% is mostly type I collagen (90%), a triple helix that gives tensile strength, with non-collagenous proteins: osteocalcin (a marker of turnover), osteopontin (cell attachment) and the BMPs (growth factors).

The blood supply reverses after fracture. Normal flow is centrifugal, from marrow to cortex to periosteum, inside to outside. After fracture it becomes centripetal, periosteum to cortex to marrow, and the periosteal supply becomes dominant once the medullary supply is disrupted, for example by reaming. That is why the soft-tissue envelope and periosteum are preserved at surgery.

Biophysics. Stress on collagen generates electric potentials, the piezoelectric effect. The compression side becomes electronegative and stimulates osteoblasts to form bone; the tension side becomes electropositive and stimulates osteoclasts to resorb it. Fluid flow in the canaliculi generates streaming potentials that stimulate osteocytes.

Perren's strain theory. A tissue can only exist in a gap if it can withstand the strain there. Strain is the change in gap divided by the original gap. Bone ruptures at 2% strain, while cartilage and granulation tissue tolerate more than 10%. To form bone directly, as in primary healing, strain must be kept below 2%. Callus healing works in steps: granulation tissue first stiffens the gap, reducing strain enough for cartilage, and the cartilage then allows bone.

The haematoma is not inert debris. It is the first regenerative signal, and preserving it is a surgical goal rather than an afterthought.

Why the haematoma matters. It forms within minutes and provides both a fibrin scaffold and a reservoir of platelet-derived signalling molecules (PDGF, TGF-beta). Experimentally, excising or preventing the early haematoma impairs union, whereas a transplanted haematoma can itself induce bone. That is the biological reason to preserve it with closed or indirect reduction and to avoid over-stripping the fracture ends.

The acute pro-inflammatory burst (first 24-72 hours). Neutrophils and then macrophages infiltrate and release a coordinated cytokine wave: TNF-alpha, IL-1, IL-6, IL-11 and IL-18. These cytokines are chemotactic for mesenchymal stem cells, promote angiogenesis and prime the periosteal and marrow progenitors that will build callus.

Macrophages are obligatory. Depleting them in models markedly impairs callus formation. The switch from pro-inflammatory (M1) to reparative (M2) macrophages marks the transition out of the inflammatory phase.

Angiogenesis precedes osteogenesis. The injured, hypoxic fracture site stabilises HIF-1-alpha, which drives VEGF and new-vessel ingrowth. No vessels, no bone: revascularisation must precede mineralisation, which is why the periosteal blood supply and gentle soft-tissue handling dominate the outcome.

The COX-2 link. Prostaglandins generated by COX-2 during this phase are required for progenitor differentiation and for the conversion of soft to hard callus. That is the mechanistic basis for caution with NSAIDs: blunting the inflammatory signal can blunt healing. The clinical NSAID evidence is under Factors Affecting Healing and Controversies; the dedicated nsaids-in-orthopaedics topic holds the full drug discussion.

History. Resolution of pain is the first sign of union. Ask about function, whether the patient can weight bear and has returned to the activities of daily living, and about the risk factors: smoking, diabetes, steroid use and NSAIDs.

Examination. Clinical union means no tenderness at the fracture site and no abnormal mobility, with painless stressing of the fracture. Assess the soft-tissue envelope as well: healed wounds and skin grafts.

Radiographs and the RUST score. The Radiographic Union Scale for Tibial fractures (RUST) scores the four cortices seen on the AP and lateral films: anterior, posterior, medial and lateral. Each cortex scores:

• 1 point: fracture line visible, no callus (unhealed)
• 2 points: fracture line visible, callus present (healing)
• 3 points: fracture line invisible, bridging callus (healed)

The total runs from 4 (unhealed) to 12 (fully united). A score greater than 10 usually correlates with mechanical stability and the ability to weight bear without pain. The general radiographic signs of union are bridging bone on 3 of 4 cortices and blurring of the fracture line, but radiographs lag behind the clinical signs.

Investigating a suspected non-union. Each test answers a different question:

• CT: the gold standard for assessing bridging, and able to distinguish bony union from fibrous non-union
• MRI: infection (osteomyelitis) versus sterile non-union
• Laboratory tests: vitamin D, calcium, phosphate, PTH, and ESR/CRP for infection

Casting. A cast holds a reduction by three-point moulding: pressure at the apex of the curve, with counter-pressure at the proximal and distal ends.

Functional bracing (Sarmiento). The brace leaves the joints free to move, and hydrostatic containment of the soft tissues stabilises the fracture. The micromotion it allows stimulates callus, so it heals by secondary healing. It is commonly used for humeral and tibial shaft fractures.

The Diamond Concept. Union needs four things, remembered as CSSM:

• Cells: osteogenic cells (osteoblasts, MSCs)
• Scaffold: an osteoconductive matrix (bone graft, collagen)
• Signals: osteoinductive factors (BMPs, VEGF)
• Mechanical stability: fixation

When the biology is poor, as in an atrophic non-union, it is augmented with graft or growth factors.

Autograft is the patient's own bone, from the iliac crest (the gold standard) or harvested from the femur with the reamer irrigator aspirator (RIA). It is osteogenic, osteoinductive and osteoconductive, providing cells, signals and scaffold. The price is donor-site morbidity: pain, infection and nerve injury.

Allograft is cadaveric bone from a bone bank, as structural struts or morcellised chips. It is strictly osteoconductive, with no cells and minimal signals. Its drawbacks are disease transmission (rare, 1 in 1 million), slower incorporation and a minor immune reaction.

Synthetics (ceramics, TCP) are osteoconductive scaffolds. They are brittle and expensive and can cause seroma.

Growth factors supply the signals:

• BMP-2 (Infuse): a potent osteoinductor that recruits MSCs to differentiate into osteoblasts, indicated for acute open tibial fractures. The risks are swelling (airway compromise in the cervical spine), heterotopic ossification and cost.
• BMP-7 (OP-1): historically used for non-unions
• PRP/BMAC: bone marrow aspirate concentrate, rich in MSCs and growth factors (PDGF, TGF-beta)

The osteocyte's sclerostin brake, which loading switches off, and the PTH/Wnt axis make a druggable target for accelerating healing. This is systemic treatment, distinct from the local osteoinductive BMP grafts under Biologic Adjuvants (and in the dedicated growth-factors-bone-healing topic).

Teriparatide (recombinant PTH 1-34). Intermittent, once-daily PTH is anabolic: it favours osteoblast survival and enhances callus formation, the opposite of the catabolic effect of the continuously elevated PTH of hyperparathyroidism. It is used off-label to support healing in selected fragility fractures and established non-unions, for example pelvic and vertebral insufficiency fractures and atypical femoral fractures related to bisphosphonates.

Not licensed for fracture healing. Supportive evidence exists but is limited and heterogeneous, so teriparatide stays an adjunct in high-risk hosts rather than routine care. It is avoided where a bone tumour, skeletal metastasis, Paget disease or prior skeletal irradiation is a concern.

Romosozumab (anti-sclerostin antibody). A monoclonal antibody that neutralises sclerostin, giving a dual effect of increased formation and reduced resorption, licensed for osteoporosis. It is mechanistically attractive for healing, but dedicated fracture-healing acceleration trials (tibial and hip) have not met their primary endpoints, so it is not established for speeding union.

Do not confuse these with anti-resorptives. Bisphosphonates and denosumab are anti-resorptive, not pro-healing, and do not accelerate union. The link between bisphosphonates and atypical femoral fractures is a separate problem from augmenting repair.

Optimising the host is as important as the surgery, and the fracture environment dictates the biology.

Planning for union covers the patient, the implant and the biology. Stop smoking, optimise diabetes and correct nutrition. Choose a load-sharing implant (nail) or a load-bearing one (plate) according to the fracture pattern and the soft tissues. Preserve the soft tissues to preserve the biology, which bears on the choice between open and closed reduction.

Ancient splinting. Splinting and casting go back to the Egyptians.

The AO era (1950s). The AO Foundation, founded by Mueller, Allgower and colleagues, emphasised anatomical reduction and rigid fixation. The result was primary healing, but a high non-union rate from biological stripping.

Biological fixation (1990s). Practice shifted to minimally invasive plate osteosynthesis (MIPO) and IM nails, putting biology (the blood supply) before anatomy and accepting secondary healing by callus.

The future.

• Gene therapy: viral vectors delivering BMP-2 genes to local cells
• 3D printing: custom scaffolds matching the defect, seeded with MSCs
• LIPUS (low-intensity pulsed ultrasound): mechanical vibration stimulates integrins
• PEMF (pulsed electromagnetic fields): induces electrical currents (streaming potentials) to stimulate calcification; contraindicated in patients with pacemakers

Implant stiffness sets the strain. The stiffness of the implant dictates the strain at the fracture site. An implant that is too stiff, such as a thick steel plate, takes all the load, and the bone beneath it senses no load (strain less than 2%) and no need for strength. The result is stress shielding: bone resorption and porosis under the plate, Wolff's law in reverse. Titanium, being less stiff and closer to bone, allows some load transfer and reduces stress shielding. Carbon fibre has a modulus closer still to bone and, being radiolucent, is used in oncology.

Delayed union is healing that takes longer than expected for the specific fracture and host (typically 3-6 months) but is still progressing. Conservative options are functional bracing, low-intensity pulsed ultrasound (LIPUS, whose evidence is set out under Controversies) and nutrition. The surgical option is dynamisation, removing locking screws to increase load and strain.

Non-union is failure to heal by 6-9 months, or no progression on radiographs for 3 consecutive months (FDA definition). The type decides the treatment:

• Septic: infection until proven otherwise; check CRP/ESR and obtain an MRI
• Hypertrophic: good biology, poor stability, the "elephant foot"; it needs stability
• Atrophic: poor biology, the "pencil tip"; it needs biology and stability
• Oligotrophic: intermediate

Work-up asks about smoking, diabetes and NSAIDs, examines for mobility at the fracture site and a sinus, and uses the imaging and laboratory tests under Investigations.

Malunion is union in a non-anatomical position, measured as shortening, rotation and angulation. Tolerance depends on the bone:

• Humeral shaft: tolerates a large deformity (20° angulation, 3cm shortening)
• Forearm: zero tolerance, because pronation and supination are lost
• Tibia: tolerates minimal varus or valgus (less than 5°)

Synostosis is fusion between two adjacent bones, radius to ulna or tibia to fibula. It follows disruption of the interosseous membrane with high energy and a single-incision approach, and in the forearm it costs rotation.

Early phase (0-2 weeks). Elevate to reduce oedema, start early range of motion if the fixation allows, to prevent stiffness and stimulate blood flow, and check the wound.

Middle phase (2-6 weeks). Proprioceptive weight bearing begins, a Wolff's law stimulus, and radiographs check alignment and maintenance of the reduction.

Late phase (6-12 weeks). Strengthening restores muscle mass. Full weight bearing follows once hard callus is visible on 3 of 4 cortices.

Time to union. On average, upper-limb fractures unite in 6-8 weeks and lower-limb fractures in 12-16 weeks. The tibia is the slowest bone to heal because of its poor blood supply, and 16-20 weeks is common for open fractures.

Non-union rates

• Tibia: 10-15%, especially when open; the most common long-bone non-union
• Femur: 1-2% with an IM nail, a success story of modern orthopaedics
• Clavicle: 5-10% with conservative treatment
• Scaphoid: high, because of its retrograde blood supply

Predictors of a poor outcome

• Smoking, consistently associated with non-union, though the size of the effect is not well established
• Open fracture (Gustilo III)
• Infection
• Prolonged NSAID use

Children are not just small adults. Their thick, vascular periosteum has high osteogenic potential, and their fractures heal twice as fast.

Remodelling. A child's fracture can correct angulation in the plane of motion but cannot correct rotation. Femoral fractures also stimulate growth through hyperaemia, and a 1-2cm leg-length discrepancy from overgrowth is common.

Physeal injury. Injuries to the physis are described by the Salter-Harris classification, and a growth arrest (physeal bar) leads to deformity.

Global Epidemiology

• Tibial shaft fracture is the most common long-bone fracture; nonunion affects roughly 5-15% (highest with open injuries and segmental bone loss).
• Smoking, diabetes, NSAID exposure, and high-energy/open injury are consistent risk factors for impaired union worldwide.

Side-by-Side Society Guidance

Registry & Trial Signals

• Large RCTs (SPRINT, TRUST, BESTT) anchor practice: ream closed tibial nails, LIPUS adds no benefit for fresh fractures, BMP-2 reduces reoperation in open tibiae.
• National trauma/arthroplasty registries (NJR, AOANJRR, Swedish/Norwegian) track reoperation and nonunion as quality metrics rather than implant survival for diaphyseal fixation.

High- vs Limited-Resource Practice

• High-resource: routine IM nailing, CT for suspected nonunion, ortho-plastic flap cover, selective BMP/autograft.
• Limited-resource: external fixation and functional bracing (Sarmiento) remain mainstays; autograft (iliac crest) preferred over costly BMP/synthetics; clinical and plain-radiograph union assessment.

Smoking Cessation (Universal Principle)

• Offer behavioural support plus pharmacotherapy (nicotine replacement, varenicline, or bupropion where available).
• For elective osteotomy/fusion, aim for cessation before surgery; cotinine testing can confirm abstinence.

NSAIDs. Animal data and a meta-analysis (OR ~2.1 overall) suggest impaired union, but the effect is dose- and duration-dependent and absent in children. Short low-dose courses for simple fractures are likely safe, and prolonged use is best avoided in high-risk cases. True equipoise persists.

Bone stimulators (LIPUS, PEMF). The TRUST RCT showed no benefit for fresh tibial fractures. Their role, if any, is debated for established non-union, where high-level evidence is lacking.

BMP. It reduces reoperation in open tibiae, but cost, swelling, heterotopic ossification and off-label spinal complications have narrowed enthusiasm, and it is reserved for selected high-risk or non-union cases.

Reaming in open fractures. SPRINT showed benefit only in closed tibial fractures, and whether to ream open fractures remains debated, especially in contaminated wounds.

Defining union. No universally accepted radiographic threshold exists. RUST improves reliability but CT is often needed, and clinical and radiographic union frequently disagree.

Biologic adjuncts (PRP/BMAC, synthetics). Widely marketed but supported mainly by low-level evidence; routine use is not established.

The Tibial Shaft (The "Unforgiving Bone")

• Anatomy: One third is subcutaneous (anteromedial surface). Poor blood supply.
• Healing: Slow (16-20 weeks).
• Management: IM Nail (standard) vs Plate (distal/proximal).
• Pearl: Reamed nailing creates a "bonfire" of growth factors and autograft.

The Scaphoid (Retrograde Flow)

• Anatomy: Blood enters distal pole. Proximal pole depends on intraosseous flow.
• Risk: Proximal pole fracture → Avascular Necrosis (AVN).
• Healing: Compression screw (Herbert screw) buried in bone.

The Clavicle (S-Shaped Strut)

• Anatomy: Membranous bone formation (unique).
• Malunion: Shortening greater than 2cm affects shoulder biomechanics (scapular dyskinesis).
• Non-Union: Atrophic common in smokers.

The Femoral Neck (Intracapsular)

• Anatomy: Synovial fluid washes away hematoma (no clot = no callus).
• Result: Must heal by Primary Healing (Compression).
• Risk: AVN due to medial circumflex artery damage.

The Humeral Shaft

• Healing: Typically 8-12 weeks.
• Acceptance: Accepts up to 20 degrees anterior angulation and 30 degrees varus/valgus due to shoulder range of motion compensation.
• Nerve: Radial nerve palsy (Holstein-Lewis fracture in distal third). 90% resolve spontaneously. Watch and wait for 3 months if closed injury.

The Distal Radius (Colles)

• Healing: 6 weeks for bony union.
• Remodeling: Limited in adults. Malunion (shortening) leads to ulnar impaction syndrome.
• Management: Volar locking plate allows early motion.

The Ankle (Weber B)

• Mortise: 1mm shift = 42% decrease in contact area.
• Healing: 6 weeks NWB (Syndesmosis) or 2 weeks NWB then boot (stable).
• Risk: Post-traumatic arthritis if articular step greater than 2mm.

The Pelvis (Ring)

• Healing: Cancellous bone heals fast (6-8 weeks).
• Weight Bearing: Depends on posterior ring stability (Sacrum/SIJ).
• Complication: Venous Thromboembolism (highest risk in orthopaedics).

The Fifth Metatarsal (Jones Fracture)

• Zone 2: Metaphyseal-Diaphyseal junction.
• Physics: Adductor longus traction creates tension side failure.
• Healing: Poor blood supply. High non-union rate.
• Management: Screw fixation often required for athletes.

The Next Frontier The future of fracture healing lies in targeted biological intervention.

• Personalized Medicine: Genetic profiling for non-union risk.
• Smart Implants: Sensors embedded in plates to measure strain and notify patients when to weight bear.
• Bio-printing: 3D printed vascularized bone grafts.