Bone Healing and Fracture Biology

Bone is the only tissue that heals by regeneration rather than by scar. What bridges a united fracture is bone, with the original structure and mechanical properties restored, and that capacity rests on the osteogenic cells held in periosteum and marrow, the osteoconductive matrix of bone itself, and a mechanical environment that guides remodelling (Wolff law).

Why the biology decides the operation. Choosing between absolute stability with a plate and relative stability with a nail is a choice of healing pathway, made before the incision. The same biology governs what else the patient needs - smoking cessation, adequate stability, and biological augmentation such as BMP in high-risk cases (smoking, diabetes, revision surgery) - and it is why delayed union and nonunion are looked for early, so that intervention is timely.

The numbers. Between 90 and 95% of closed fractures unite with appropriate treatment, and nonunion runs at 5-10% overall. The rate is bone-specific:

• Scaphoid 5-10%
• Femoral neck 10-30%
• Tibial shaft 5-15%

Who fails to unite. Smoking is present in 30-40% of nonunions, diabetes prevalence is increasing, and the population is ageing. Nonunion treatment consumes significant healthcare resources, and costs the patient prolonged disability and lost productivity.

Primary (direct) healing

Primary (direct) healing occurs only where the surgeon creates the conditions for it: fragments anatomically reduced and rigidly fixed with absolute stability, so that osteoclasts and osteoblasts cross the fracture directly and no intermediate callus forms.

What absolute stability requires:

• Anatomical reduction, with direct bone-to-bone contact
• Interfragmentary strain under 2%
• Compression between the fragments, from lag screws or compression plates
• Rigid fixation that prevents any motion

Cutting cones. Osteoclasts assemble on one side of the fracture, begin resorbing bone, and tunnel across the fracture line in a cone-shaped front. Osteoblasts follow behind them, depositing lamellar bone, so that each tunnel becomes a new osteon - a Haversian system crossing the fracture and bridging the gap directly. Multiple cutting cones do this, and together they restore continuity.

Where primary healing is the aim. A lag screw across a simple oblique or spiral fracture compresses it into absolute stability, which is the everyday application in the ankle malleoli, the forearm bones and the scaphoid. Dynamic compression plating does the same for the forearm and clavicle, compressing the near cortex while the far cortex is loaded in tension. Articular fractures are fixed this way for a different reason: the joint surface must be reduced anatomically, lag screws through the plate supply compression, and primary healing restores articular congruity.

What it buys and what it costs. The bone is restored anatomically, the implant supplies mechanical stability from the day of surgery, there is no bulky callus to interfere with the soft tissues, and a periarticular fracture can be moved early. Against that:

• It requires open reduction and internal fixation, with the surgical trauma that entails
• Extensive soft tissue dissection may impair the blood supply
• Early strength depends entirely on the implant
• Stress shielding by the plate may lead to bone resorption
• Biological healing is slow, taking months for full strength

Secondary (indirect) healing

Secondary (indirect) healing is the natural biological response to a fracture held with relative stability, where controlled interfragmentary motion is allowed. Callus forms around and between the fragments and bridges the gap by enchondral ossification, through a cartilage intermediate, in four distinct phases. It does not require anatomical reduction and is more biologically robust than primary healing.

View larger

The cells and the signals

View larger

Where secondary healing is expected

Intramedullary nailing of a femoral or tibial shaft fracture gives relative stability with controlled motion, and the large callus that follows provides biological bridging, with faster clinical union than plating. External fixation is used for open fractures, infected nonunions and distraction osteogenesis, and allows stability and compression to be adjusted, with the callus monitored and modulated as healing proceeds. Cast immobilisation treats most non-displaced or minimally displaced fractures along the same natural pathway, the callus supplying stability as it matures.

View larger

Both ossification routes run in the same callus. Intramembranous ossification occurs subperiosteally at the fracture peripheries, away from the gap, where the environment is relatively stable and well vascularised: osteoblasts lay down woven bone directly, with no cartilage intermediate, forming the peripheral hard callus. Enchondral ossification occurs in the central fracture gap, which is more mobile and hypoxic, where MSCs form a cartilage (soft callus) intermediate that calcifies and is then replaced by bone after vascular invasion. A single healing fracture therefore shows intramembranous bone at the edges and enchondral bone bridging the gap, and the proportions shift with the local strain and oxygen tension.

The blood supply reverses. Bone has a dual supply: the nutrient (medullary, endosteal) artery supplies the inner two-thirds of the cortex by centrifugal, outward flow, and the periosteal supply feeds the outer one-third by centripetal, inward flow. A fracture disrupts the medullary supply, so the periosteal and extraosseous supply becomes dominant during early healing, and the medullary supply is re-established as healing progresses. This is the biological reason to preserve the soft-tissue envelope and avoid periosteal stripping, and why reamed nailing transiently injures the endosteal supply while the periosteal supply compensates.

View larger

Strain and the mechanical environment

Perren's interfragmentary strain theory. Strain is the change in the fracture gap divided by the original gap, ΔL / L, where ΔL is the change in gap width under load and L is the original gap width. Different tissues tolerate different amounts of strain, so the gap and the stiffness of the construct together decide what tissue can form in it.

The surgeon sets the strain. A compression plate holds the fracture in the range where bone forms directly and no callus appears; an intramedullary nail leaves enough motion for a cartilage intermediate and callus; fixation inadequate for the pattern leaves the fracture in the fibrous range, and it does not unite.

Absolute stability is achieved by compression across the fracture, requires anatomical reduction, and produces primary healing with no callus. It is delivered by lag screws, compression plates, and an external fixator in compression mode.

Relative stability allows controlled interfragmentary motion, does not require anatomical reduction, and produces secondary healing with callus. It is delivered by:

• Intramedullary nails, locked or unlocked
• Bridge plating
• An external fixator in neutral or dynamic mode
• Cast immobilisation

Getting the mechanics right. The principles that govern the construct:

1. Match the stability to the healing goal - articular fractures need anatomical reduction and primary healing, while secondary healing is acceptable for diaphyseal fractures
2. Minimise the gap - large gaps delay or prevent healing
3. Preserve the biology - minimise periosteal stripping, and use indirect reduction techniques where possible
4. Correct length and alignment - malreduction impairs healing as well as function

The diamond concept

The Diamond Concept (Giannoudis et al., 2007) describes the four elements a fracture needs in order to heal: osteogenic cells, an osteoconductive scaffold, osteoinductive growth factors, and an appropriate mechanical environment. Deficiency in any one of them increases the risk of delayed union or nonunion, which is what makes it useful at the bedside - a nonunion is analysed by asking which element is missing.

Using the diamond on a nonunion. Ask which element is deficient: cells (radiation, smoking, excessive stripping), scaffold (a large gap, bone loss), growth factors (avascular tissue, diabetes), or mechanics (inadequate fixation, persistent motion). Then treat every deficient element rather than the most convenient one. An atrophic nonunion needs cells, scaffold and growth factors - bone graft, with BMP or DBM - alongside stable fixation; in a hypertrophic nonunion the mechanics are the likely problem, and the answer is compression and a reduced gap.

Differential diagnosis: impaired fracture healing

A fracture that is not progressing must be characterised precisely, because each pattern has a different deficient diamond element and therefore a different treatment. Delayed union and the nonunion subtypes are the core differential.

What impairs healing

Patient factors. Systemic conditions and drugs change how a fracture heals, and each carries its own management step.

Local factors. The fracture itself and its soft-tissue envelope carry risks of their own.

NSAIDs remain controversial. Animal studies show impaired healing, COX-2 being important for fracture healing, and human studies show mixed results, some showing delayed union and others no effect; the meta-analysis cited below found more delayed union and nonunion in adults, with no significant effect in children or with low-dose, short-duration exposure. The current consensus follows that split: avoid high-dose, long-term NSAIDs (especially COX-2 selective agents) in high-risk fractures, while short-term use of 1-2 weeks for pain is likely acceptable, with paracetamol, short-term opioids or regional anaesthesia as the alternatives.

Other drugs. Bisphosphonates reduce bone remodelling and may delay the remodelling phase, but they do not prevent healing and are continued for osteoporosis. Cytotoxic chemotherapy impairs cell proliferation, so elective surgery is delayed until after it where possible. Warfarin impairs the vitamin K-dependent proteins including osteocalcin, of unclear clinical significance, and is continued for thromboembolic disease.

Optimising the patient. Before surgery:

• Smoking cessation, six to eight weeks beforehand where the operation can be planned
• Glucose control in diabetes
• Nutritional supplementation with protein, vitamin D and calcium
• The lowest steroid dose the underlying disease allows, with biological augmentation considered where it cannot be reduced

At operation, minimise periosteal stripping, aim for stability that is adequate but not excessive, fill bone gaps with graft, and augment biologically with BMP or bone graft in high-risk cases. Afterwards:

• Avoid NSAIDs, or keep them short-term only
• Allow early weight-bearing if the fixation permits, since loading stimulates healing
• Monitor healing clinically and radiographically
• Intervene early where delayed union is suspected

Augmenting and rescuing healing

The diamond concept says what to optimise; a viva will also ask how to augment or rescue a slow-healing fracture actively. Three categories are high-yield.

Teriparatide (recombinant PTH 1-34). Intermittent low-dose PTH is anabolic, a net stimulus to osteoblasts, unlike the catabolic effect of continuous PTH. Its off-label uses are delayed union and nonunion (especially osteoporotic), pelvic and sacral insufficiency fractures, stress fractures and atypical femoral fractures, where impaired remodelling is the problem. Some randomised data, for instance in the distal radius, suggest faster healing, but the overall evidence is mixed, so it is an adjunct rather than a guaranteed therapy. Avoid it in the skeletally immature, in Paget disease, after prior skeletal radiation, or with an unexplained high alkaline phosphatase (the osteosarcoma signal in rodents), and it is typically limited to a lifetime course of about two years.

Biophysical stimulation. Low-intensity pulsed ultrasound (LIPUS) and pulsed electromagnetic fields or electrical stimulation (capacitive, inductive or direct current) are non-invasive adjuncts proposed for delayed union and nonunion. The evidence is mixed - the large TRUST randomised trial showed LIPUS did not improve tibial fracture healing - so they are selective adjuncts of uncertain benefit, never a substitute for correcting the deficient diamond element.

Surgical rescue. Exchange reamed intramedullary nailing suits a diaphyseal, especially hypertrophic, nonunion, the reaming providing autograft while the new nail is larger and stiffer. The Masquelet induced-membrane technique handles segmental defects in two stages: a cement spacer induces a vascular membrane, which is later opened and packed with bone graft. Bone transport, or distraction osteogenesis in the Ilizarov sense, is for large segmental defects.

Global Epidemiology of Fractures and Impaired Healing

Fractures are among the most common large-organ traumatic injuries worldwide, and approximately 10% fail to heal normally (Einhorn and Gerstenfeld, Nat Rev Rheumatol 2015, PMID 25266456). Global Burden of Disease modelling shows lower-limb fractures (patella, tibia/fibula, ankle) carry the highest age-standardised prevalence of all fracture sites, with falls the dominant mechanism and a projected continuing rise in fall-related fractures to 2035 (GBD 2019 analysis, PMID 39543488).

Guidance and Principles Across Jurisdictions

Bone-healing biology is governed by AO principles rather than disease-specific guidelines; the table below summarises how the major bodies frame stability and biology, with the underpinning evidence level.

Registry and Surveillance Evidence

There is no dedicated international fracture-healing registry; population-level signal derives from trauma registries and Global Burden of Disease surveillance (PMID 39543488). Where biologics or implants are involved, device and adverse-event surveillance (e.g. national regulator pharmacovigilance) informs practice — for example, post-marketing signals shaped the cautious, lowest-effective-dose use of rhBMP-2/rhBMP-7 in spine and long-bone surgery.

Practice Variation

• Osteobiologic use varies markedly by funding and regulation. rhBMP use is constrained by cost in publicly funded systems; autograft remains the global default for biological augmentation.
• Smoking-cessation pathways differ but the principle (cessation around the time of surgery; structured support) is universal given the OR 2.32 nonunion signal (PMID 24740664).
• NSAID policy varies between centres given dose- and duration-dependent harm in adults but a reassuring profile in children and short courses (PMID 30260913).
• Underserved and remote populations carry compounded risk. Higher diabetes prevalence and barriers to surgery and follow-up access (including telehealth reach) worsen healing outcomes in Indigenous, remote-rural, and low-resource settings, making equitable access and culturally appropriate planning part of optimising healing globally.