1. Introduction

Fracture nonunion is considered a challenging complication in orthopedic trauma, often resulting in prolonged patient morbidity and substantial healthcare costs. Despite advancements in surgical techniques and technology, the incidence of long bone fracture nonunion has remained significant; roughly 5%–10% of all fractures (10%–15% of surgically managed fractures) fail to unite within expected timeframes [1, 2]. Tibial shaft fractures, in particular, have the highest nonunion rates among long bones (the reported incidence ranges from around 1% in low-risk cases up to 19% in severe cases), while the nonunion rate of femoral fractures is between 10% and 12%, depending on injury severity [1]. These statistics underscore the tibia and femur as frequent sites of nonunion, often due to the tibia’s subcutaneous location and tenuous blood supply, as well as the high-energy mechanisms associated with tibial and femoral fractures.

Nonunions impose a heavy burden on both patients and the healthcare system. They lead to prolonged disability, repeated surgeries, and psychosocial distress. Economically, the cost of treating a nonunion is dramatically higher than that of an uncomplicated fracture. One analysis showed that the average total cost of care for patients with a nonunion was nearly three times higher than for patients whose fractures healed normally ($102,000 vs. $35,000 in one study) [1]. In the United States, an open tibia fracture nonunion incurs a median cost of $25,000, roughly double that of a similar fracture that achieves union [2]. Beyond direct costs, nonunion patients experience extended loss of work income (often between 6 and 12 months or more of work absence) and higher utilization of healthcare resources, including increased use of analgesics (opioids/nonsteroidal anti-inflammatory drugs [NSAIDs]) for chronic pain.

The impact on patient quality of life is profound. Persistently unhealed fractures cause chronic pain, functional limitations, and psychological distress [2]. Studies have found that patients with tibial or femoral nonunion report significantly worse physical and mental health scores than those who healed normally, with higher rates of depression, reliance on assistive devices, and earlier retirement due to disability [1]. These considerations highlight that fracture nonunion is not only an orthopedic problem confined to bone healing but a complex condition affecting the whole patient [3]. In light of this, it is crucial that the orthopedic surgeon understands that nonunion is rarely a simple or single-factor process but rather a complex, multifaceted condition and that evidence-based tools are available to promote bone healing.

2. Definition and classification

Definition: Fracture nonunion has long been difficult to define with precision, as it represents a failure of normal healing rather than a distinct time-bound event. Traditionally, the U.S. Food and Drug Administration (FDA) definition (circa 1998) is cited: a nonunion is “established” by nine months post-fracture with no visible progressive healing for at least three consecutive months [1, 2]. In clinical practice, many surgeons diagnose and treat a “nonunion” earlier, often around six months post-fracture, if there are no signs of healing progress on serial radiographs and the clinical picture suggests the fracture will not heal without intervention [1]. The Danish Orthopedic Trauma Society has proposed a more pragmatic definition: “a fracture that will not heal without further intervention.” This highlights that the essence of nonunion is the biological inability to heal spontaneously, rather than an absolute time frame. However, this pragmatic definition still requires the clinician to judge when healing has ceased. In practice, a combination of chronological and clinical criteria is used: if a fracture shows no radiographic bridging and clinically persistent pain or instability by around six months (sooner in the absence of any healing progress), it is generally considered a nonunion, especially if serial X-rays over a two- to three-month period show no callus progression and maturation. It is important to distinguish a true nonunion from a “delayed union,” which implies slow healing that may still complete given more time; typically, delayed union is suspected by three to six months when healing is not on track [2].

Despite attempts at consensus, there is no single universally accepted definition of nonunion that captures both the time element and the biological aspect [2]. Various scoring systems (e.g., the Radiographic Union Scale for Tibial fractures (RUST)) and definitions (e.g., FDA, Medicare, etc.) exist, which can complicate diagnosis and research inclusion criteria.

Classification: Once a nonunion is identified, classification helps guide treatment and communicate the underlying pathology. The most commonly used classification is based on the radiographic appearance and inferred biology of the nonunion, as described by Weber and Cech [4]. This scheme broadly divides nonunions into three categories [4]:

Hypertrophic (Hypervascular) nonunion: Abundant callus formation is seen on X-ray, but the fracture fails to bridge. This “elephant foot” or hypertrophic callus indicates that the biology (blood supply, osteogenesis) is robust; the primary problem is usually mechanical instability. In hypertrophic nonunions, the body is producing callus in response to motion, but the excessive interfragmentary movement prevents consolidation. These are viable nonunions with a “favorable biological response” but inadequate stability [4].
Atrophic (Avascular) nonunion: Minimal or no callus is visible. The fracture ends may be sclerotic or atrophic (“pencil tip” appearance) with a gap. This reflects an unfavorable biological environment and a poor healing response, often due to insufficient blood supply, extensive cortical devitalization, or large bone loss. Atrophic nonunions are nonviable; the biology is quiescent and needs enhancement in addition to stability. These cases often require aggressive measures to stimulate healing, such as debridement of sclerotic bone and the addition of grafts or growth factors [4].
Oligotrophic nonunion: An intermediate category characterized by little to moderate callus that does not progress to union. It is sometimes considered a subset of viable nonunion, where there is some biological activity, but perhaps not in the optimal location or amount to achieve union. Oligotrophic nonunions often result from fracture fragments that are not apposed (distraction or a gap) or from inadequate reduction, which leads to fibrous tissue rather than mineralized callus bridging the gap. They “bridge the gap” between hypertrophic and atrophic cases in terms of biology; the healing response is present but not effective enough [4].

Another crucial practical classification distinction is between septic and aseptic nonunion. An aseptic nonunion has no evidence of infection, whereas a septic (infected) nonunion indicates that the fracture site is harboring an infection (chronic osteomyelitis), which can inhibit healing. This distinction is critical because the presence of infection radically alters the treatment approach, requiring debridement, antibiotics, and often staged reconstruction. Any nonunion classification system must account for infection status. For example, the Cierny–Mader classification, originally developed for osteomyelitis, is sometimes applied to infected nonunions, taking into account the host’s physiological status and the local bony involvement (medullary, superficial, segmental, etc.).

Host Factors Classification: There have also been attempts to classify nonunions by host and injury factors. One such approach is the “Nonunion Risk Score” or Calori classification, which assigns points for risk factors (like smoking, diabetes, soft tissue damage, etc.) to stratify the difficulty of healing. While not a formal part of nonunion nomenclature, considering host factors is important; for instance, the Nonunion Scoring System (NUSS) helps predict which fractures are at high risk for nonunion and might benefit from early intervention [2].

It should be noted that traditional radiographic classifications have limitations. They are based on static X-ray appearances and may not capture issues like infection or systemic factors. Mills et al. [5] emphasized that radiographic classification alone can be misleading and should be used in conjunction with a thorough clinical and etiological assessment. In Mills’ series of 100 nonunions, only 31% had a single identifiable cause; the majority (69%) had two or more contributing factors. Moreover, 5% of ostensibly “aseptic” nonunions in that series turned out to have unexpected infections on culture. These findings highlight that nonunions are often multifactorial – a combination of mechanical, biological, and patient factors – and that careful evaluation, including microbiology, is necessary rather than relying solely on the X-ray appearance [5].

3. Pathophysiology

Why do fractures go on to nonunion? The pathophysiology of nonunion can be understood as a failure in one (or more) of the fundamental requirements for bone healing. The “Diamond Concept” of fracture healing describes four essential components: (1) a mechanically stable environment, (2) osteogenic cells, (3) an osteoconductive scaffold, and (4) osteoinductive growth factors [6]. If any of these components is insufficient, the fracture may not heal. Broadly, the causes of nonunion are often categorized into mechanical factors, biological factors, infectious factors, and host (systemic) factors [5]. These roughly correspond to problems with stability, local biology (including blood supply), the presence of infection, and patient-related issues, respectively.

Mechanical pathophysiology: Bone healing is highly sensitive to the mechanical environment. According to Perren’s strain theory, if interfragmentary strain (the change in gap length under loading, divided by gap length) exceeds approximately 2%, bony bridging is inhibited; high strain leads to fibrous tissue or cartilage instead of bone [5]. In an adequately stabilized fracture, interfragmentary strain drops as callus forms. However, if a fracture has inadequate stabilization (e.g., insufficient fixation rigidity or an excessive gap), the motion at the fracture site remains too great for mineralized bone to form. Instead, the fracture ends may develop abundant callus that never fully bridges (resulting in a hypertrophic nonunion), or in cases of a large gap, even callus formation is thwarted and a fibrovascular nonunion forms. A classical example is a comminuted fracture fixed with a plate that is too small or with insufficient fixation points, leading to micromotion that stimulates callus but not union. Large fracture gaps are particularly problematic: Experimental data show that even with rigid fixation, a critical size gap can fail to heal because the initial bridging tissue cannot span the distance [5].
Biological pathophysiology: The flip side is an environment where stability might be adequate, but the biology is compromised. Bone healing requires a viable osteogenic response, an adequate population of progenitor cells that can form bone, and a supportive blood supply to nurture those cells. If the fracture fragments are devascularized, the bone ends may simply not generate callus. Early vascularity is crucial; insufficient revascularization of the fracture site predisposes it to an atrophic nonunion [5]. Causes of biological failure include poor local blood supply, extensive bone loss (no scaffold or cells present in the gap), or extinction of cellular activity due to factors like radiation, infection, or excessive periosteal injury [7]. The fracture ends can become “inert,” effectively walled off by sclerotic bone or scar tissue. The radiographic atrophic appearance correlates with this scenario. Additionally, certain medications (e.g., chronic high-dose steroids or prolonged use of NSAIDs, which interfere with inflammation and bone formation) can suppress the biological healing response. If biology is the limiting factor, simply adding stability (e.g., re-fixation) without addressing the biological deficit (through bone grafting or stimulatory factors) may not succeed. This is why atrophic nonunions usually require both debridement of poorly vascular bone and the addition of autograft or other adjuncts to reinitiate healing.
Infection: The presence of infection profoundly alters the healing environment. Osteomyelitis at the fracture site leads to persistent inflammation, osteolysis, and often instability as hardware loosens in infected tissue [5]. Bacteria form biofilms on implants and devitalized bone, interfering with osteogenesis. Even if a fracture was initially adequately stabilized, an indolent infection can cause what would otherwise be a healthy callus to halt and resorb. Infected nonunions often have a mix of mechanical and biological issues: Infection induces bone resorption and may create a gap, while the body’s attempt to clear the infection can divert resources away from regenerative processes. The result is often a recalcitrant nonunion with sequestra (dead bone) and drainage. Notably, infection can present subtly; for example, a fracture that simply fails to heal (so-called “occult infection” nonunion) without obvious signs of infection. In one series, all patients who had been previously treated for infection and then believed to be aseptic were found to have positive cultures at revision, indicating that any history of infection should raise suspicion in a nonunion [5]. Given the high stakes, routine deep cultures are recommended in nonunion surgery to rule out infection, even in cases that appear aseptic. Infection pathophysiology in nonunion mandates a two-front battle: Eradication of bacteria and promotion of bone healing.
Systemic/Host factors: The patient’s overall health and habits critically influence bone healing. Smoking is one of the most well-documented risk factors for delayed healing and nonunion; nicotine and other tobacco metabolites cause vasoconstriction and impair osteoblast function, roughly doubling the risk of nonunion in long bone fractures. Studies consider heavy smoking (e.g., > 1 pack/day or ≥ 10 cigarettes/day) a significant risk factor [8]. Diabetes mellitus, especially if poorly controlled, impairs wound and bone healing by reducing cellular function and blood flow. Other chronic diseases, such as renal failure or anemia, can compromise healing as well. Nutritional status (e.g., vitamin D levels, protein intake) matters: malnutrition or deficiencies can reduce the substrate needed for bone repair. Certain medications, like chronic steroids (which inhibit osteoblasts and prolong inflammation) or NSAIDs (which, when used for prolonged periods early in fracture healing, have been shown in animal models to impair callus formation), can predispose to nonunion [5]. Age is a factor too; older patients heal more slowly due to lower osteogenic cell populations and comorbidities. Additionally, the concept of “biological aging” of a fracture – a long delay in intervention (fracture fragments left with motion for many months) – can lead to a state where the healing responses have burned out and scar tissue has filled the gap, making later healing attempts more difficult. Host immune status is important in infected cases; immunosuppressed or ill patients may not contain infection or mount a healing response well. In summary, systemic factors such as smoking, diabetes, vascular disease, alcoholism, and advanced age are recognized contributors to nonunion risk [2]. Often, these act in concert with local factors, for example, a smoker with a badly comminuted open tibia has a compounded risk.

In summary, the pathogenesis of tibial and femoral nonunions can be traced to failures in mechanics (too much or too little stability) and/or biology (inadequate blood supply or cells), often compounded by infection or systemic issues. A hypertrophic nonunion indicates that biology is intact but stability was insufficient; an atrophic nonunion indicates biological failure (and often also a mechanical issue, such as a gap or poor vascularity). These are two ends of a spectrum. The surgeon must interpret the nonunion’s appearance and history to determine the dominant causes. Understanding these mechanisms underpins the rational selection of treatments. For example, a purely hypertrophic nonunion might heal with improved stabilization alone, whereas an atrophic nonunion will require a biological boost (graft or growth factors) in addition to stabilization. The “Diamond” concept reminds us that bone healing is achieved when stability, cells, scaffold, and signals are all present; the pathophysiology of nonunion lies in the absence of one or more of these critical factors.

4. Nonoperative management of nonunion

By the time a fracture meets the criteria for nonunion, surgical intervention is the definitive treatment in the vast majority of cases, especially for tibial and femoral shaft nonunions. These are weight-bearing bones where a stable union is critical for function; thus, leaving them ununited is usually not acceptable long-term. That said, there are some limited scenarios and adjunct measures in which nonoperative or minimally invasive strategies might be employed, either in an attempt to avoid surgery or as a temporizing measure. It should be emphasized that nonoperative approaches are generally more successful in delayed unions or very early nonunions, as opposed to established, refractory nonunions. Once a nonunion is well-established (especially if the fragments are sclerotic or there is significant deformity), purely non-surgical management has a low likelihood of success.

Prolonged immobilization and functional bracing: The simplest nonoperative management involves giving the fracture more time with continued immobilization (cast or bracing) and protected weight-bearing. Historically, before modern surgical methods, some fractures that had not united after several months were treated with extended casting or functional braces, allowing partial weight-bearing to stimulate healing (Wolff’s law states that mechanical stimulation encourages bone formation). For instance, Sarmiento’s functional brace for tibial fractures can sometimes help a slowly healing tibia achieve union by allowing controlled motion and weight-bearing in a brace. In modern practice, if a fracture shows some progress but is slow (delayed union), one might continue immobilization beyond the typical timeframe rather than operate immediately. However, simply keeping a truly stagnant fracture in a cast indefinitely is usually futile. There are cases, often anecdotal, of very late unions (e.g., a fracture finally unites after 18 months), but relying on that is not advisable if no progress is evident by six to nine months. An exception might be in a patient who is non-ambulatory or has low demand; for example, in a very elderly patient with minimal pain, one might treat a fibrous nonunion conservatively if the function is acceptable.

Dynamization of existing fixation: This is a minimally invasive “semi-operative” strategy particularly applicable to fractures treated with intramedullary nails. Nail dynamization involves removing one or more locking screws from the nail (typically the static interlocking screw that prevents collapse) to allow the fracture ends to come closer and introduce some axial micromotion. The rationale is that controlled compression and slight motion can stimulate callus formation if the fracture was too rigid or if there was a small gap preventing contact. Dynamization is usually attempted in around three to four months if a nailed fracture shows delayed healing but some callus presence. In the context of nonunion, dynamization is more likely to help in a hypertrophic or oligotrophic scenario where biology is present but perhaps the fracture ends are held apart or too rigidly fixed. It is generally not useful for atrophic nonunions (where biology is absent) or when there is a large gap. Additionally, dynamization only helps if the nail was initially statically locked; if the fracture was already dynamized or if all screws are already out (as in an unlocked nail), then this option is not available. Historical data on nail dynamization success are mixed. Older series reported union achieved in roughly 50%–70% of cases when dynamization was done for delayed union [9]. A recent review notes that success rates for nail dynamization in established tibial/femoral shaft nonunions are on the order of 55%–65% [10], meaning nearly half of cases will still require further surgery. Therefore, dynamization is generally considered in the earlier phase of a developing nonunion (delayed union) as a first step, but if significant healing does not occur in the next two to three months post-dynamization, more invasive measures (such as exchange nailing) are indicated [10]. One must also ensure that dynamization will actually transmit load across the fracture. For example, if a fracture has a large gap or the nail has a bend, simply removing screws may not lead to compression, and in such cases, dynamization could even create too much instability. Additionally, there is a risk: if the fracture still does not heal, removing screws can lead to hardware failure (excessive motion causing nail breakage or deformity) [9].

External bone stimulation: Two primary modalities exist – pulsed electromagnetic field (PEMF) stimulation and low-intensity pulsed ultrasound (LIPUS). These modalities are noninvasive treatments applied externally to the limb that aim to stimulate bone healing through biophysical signals. They are among the few FDA-approved nonoperative treatments specifically indicated for aiding nonunion healing.

PEMF: The patient uses coils or a device that emits electromagnetic fields around the nonunion site for several hours a day. Clinical evidence has shown that PEMF can induce healing in a significant proportion of nonunions, with reported success rates ranging from 73% to 85% in various series. In practice, PEMF is often prescribed for patients who are not good surgical candidates or who refuse surgery. It can also be used as an adjunct to surgery. We often request a PEMF device for a patient with a recalcitrant tibial nonunion as a supplement to surgical revision, especially if the biology is poor [11].
LIPUS: This is an ultrasound device (e.g., Exogen) applied daily that delivers low-intensity ultrasound waves to the fracture. LIPUS has been shown in some studies to accelerate the healing of fresh fractures and some nonunions by promoting callus formation and angiogenesis [12]. It has the advantage of being easy to use at home for 20 minutes daily. Like PEMF, it is often employed in scenarios where surgery is being delayed or avoided. For instance, if a patient’s fracture shows minimal healing at four months, adding LIPUS might convert it to union by six to seven months, obviating surgery. Some meta-analyses suggest that LIPUS can significantly shorten the time to union in nonunions/delayed unions [12]. It is considered a safe, cost-effective modality with essentially no systemic risks, though patient compliance is needed.

It is important to note that nonoperative adjuncts like bone stimulators are more likely to succeed in aiding a healing response that is slow, as in delayed nonunion, rather than creating one from nothing. For example, a hypertrophic tibial nonunion with lots of callus might benefit from ultrasound or PEMF to “push it over the edge.” In contrast, a totally atrophic, mobile nonunion with a big gap is unlikely to unite with a stimulator alone. Nonetheless, in certain patients (e.g., those who cannot undergo surgery due to medical issues), a course of PEMF or LIPUS is worth trying, as it may induce enough bone formation to achieve union or at least a pain-free fibrous union.

Pharmacological adjuncts: There is ongoing research into systemic therapies that might promote fracture healing. One example is teriparatide (recombinant PTH 1–34), an anabolic osteoporosis drug that has shown some promise in case reports for enhancing difficult fracture healing. While not standard, some surgeons have used off-label teriparatide daily injections for a few months to encourage callus formation in a delayed union. Results are anecdotal but biologically plausible, as PTH can stimulate osteoblasts [13]. Similarly, optimizing the patient’s biology with nutritional supplementation (vitamin D, calcium, protein) is sensible if deficiencies are found. Good endocrinological health (thyroid, parathyroid, etc., in balance) is important – rarely, an undiagnosed metabolic bone disorder could be contributing, so in refractory cases, checking for osteoporosis or other metabolic issues can be considered.

In summary, nonoperative management of nonunion is limited. Extended immobilization, functional bracing, and delaying surgery might be attempted in the hope of late union, but success is far from guaranteed. Biophysical stimulation (PEMF, LIPUS) can be valuable adjunct therapies, having shown the ability to induce union in a substantial fraction of cases without surgical intervention. For certain patients, these modalities offer a chance to avoid another operation. Nail dynamization is a minor intervention that can help some delayed unions consolidate without a full reoperation, particularly in well-aligned fractures lacking only compression. Ultimately, if these measures do not yield clear signs of healing within a couple of months, the surgeon should proceed to operative management rather than prolong the patient’s disability. Keeping the patient engaged in weight-bearing (when safe) and physical therapy for joint motion is also important during any nonoperative trial, because maintaining limb function will aid recovery once union is achieved (either spontaneously or after later surgery).

5. Surgical treatment strategies

When nonoperative measures are insufficient or inappropriate, surgical intervention is required to achieve union. Indeed, for most tibial and femoral nonunions, surgery is the definitive solution. The goals of surgery in a nonunion are twofold: (1) provide a stable mechanical environment (usually by internal or external fixation) and (2) enhance the biological environment (often by debridement of fibrous tissue and nonviable bone, and addition of bone graft or other osteogenic material) [10]. These principles apply broadly, but the optimal surgical strategy can differ between the tibia and femur due to anatomical and biomechanical considerations. Furthermore, the strategy depends on the type of nonunion (hypertrophic vs. atrophic) and whether infection is present. This section will focus on aseptic diaphyseal nonunions of the tibia and femur, discussing the typical surgical options and the evidence behind them – notably the contemporary debate of exchange intramedullary nailing versus augmentative plating for nonunions after initial intramedullary nailing.

5.1 Tibial nonunion

Initial considerations: The tibial shaft is commonly stabilized with an intramedullary nail for acute fractures. Thus, many tibial nonunions present as an unhealed fracture with a nail in place. Others may occur after plate fixation or external fixation. Deciding on the surgical approach involves analyzing why the tibia did not heal: Was it primarily a stability issue (suggesting the need for improved fixation), a biology issue (suggesting the need for bone grafting), or both? Key factors in tibial nonunion management include alignment (any malalignment should be corrected during surgery), the condition of the soft tissues (sometimes residual wounds or the need for flap coverage), and the presence of any infection. Assuming an aseptic tibial nonunion, the classic options are Exchange reamed nailing, augmentative plate fixation (with or without removing the nail), and in some cases, external fixation if needed (particularly for very complex cases or if multiple attempts have failed).

Exchange nailing: This has long been considered the gold standard for aseptic tibial shaft nonunions after a nail [9, 14]. The technique involves removing the current nail, freshening the medullary canal with reamers (usually reaming to a larger diameter, which provides internal bone graft from the reamings and increases endosteal blood flow), and inserting a new, larger-diameter intramedullary nail [10]. Typically, the new nail is 1–2 mm larger in diameter to increase stability (“up-nailing”). Union rates with exchange nailing of the tibia are generally high, reported in the range of 76% to 96% success in various series [14]. These high rates presume good technique, including thorough reaming to remove fibrous tissue, achieving a tight fit of the nail, and locking it appropriately (and sometimes dynamizing it later, if needed). Exchange nailing addresses both mechanics (a stiffer nail and fresh locking screws improve stability) and biology (reaming debris acts as autograft and reaming also works by inducing a new inflammatory response in the canal). Many surgeons also perform cortical bone grafting at the nonunion site during exchange nailing: After reaming and before inserting the new nail, they may take some of the reamed bone material (often collected by the reamer–irrigator–aspirator (RIA) device, if used) and pack it into the fracture gap [15]. Even without deliberate graft packing, the reamings that are forced into the fracture site can have a positive effect (this is sometimes called the “internal bone graft” of exchange nailing). A critical part of exchange nailing is achieving good fracture apposition; if there is a gap, one might consider doing an open exposure to remove interposed fibrous tissue and ensure the bone ends contact. In some cases, especially if the original nonunion had a deformity or misalignment, the surgeon will formally open the nonunion, debride it, correct alignment, and then insert the new nail (this is sometimes called an “opened exchange nailing”). If alignment is acceptable and there is some fibrous tissue, others will perform it closed, letting the reamer clear the tissue.

Augmentative plating (“Plate-over-Nail” technique): This approach has gained popularity in recent years as an alternative to re-nailing. In an augmentative plating procedure, the original nail is typically left in place (if it is well-positioned), and a plate is applied to the fracture site to provide additional stability. The plate is usually a compression or locking plate applied percutaneously or through a limited open approach, spanning the nonunion. The idea is to augment the existing fixation: the nail controls the alignment and takes some load, while the plate adds rotational and bending stiffness, significantly improving overall stability [10]. During this procedure, the nonunion site is often exposed via a small incision to allow for decortication (scraping the bone ends to bleeding bone) and bone grafting (usually autologous cancellous graft from the iliac crest or another source, placed around the nonunion) before the plate is fixed. Essentially, augmentative plating addresses both stability (very rigid due to the nail + plate construct) and biology (by adding bone graft and freshening the bone ends). An advantage of this method is that one does not disturb the intramedullary canal significantly (as opposed to exchange nailing, which involves re-reaming the canal and potential weakening of the bone or risk of a new fracture during nail insertion). Plate-over-nail can be particularly useful if the nonunion has an element of axial malalignment that can be corrected with the plate by shifting one fragment relative to the nail slightly (though large deformities typically require nail removal or osteotomy).

Clinical outcomes for augmentative plating of tibial nonunions have been very encouraging. Several studies report high union rates comparable to or exceeding those of exchange nailing. For instance, in the study by Birjandinejad et al. [16], combined femoral and tibial nonunions were treated with plate-over-nail and achieved union in 100% of femur cases and 84.6% of tibia cases.

However, it is important to select cases properly for augmentative plating. If the original nail is badly malpositioned (e.g., significant malalignment or if the nail is broken), one might not want to keep it; in those cases, exchange nailing or nail removal and plating anew would be chosen. Augmentative plating works best when the nail is in an acceptable position but union did not occur (perhaps due to inadequate stability or biological reasons). By leaving the nail, one maintains alignment and also avoids having to fill the canal after nail removal (which would leave a stress-riser). The plate is encouraged to be placed using the minimally invasive percutaneous osteosynthesis (MIPO) technique, preserving blood supply.

Exchange nailing versus augmentative plating in tibia: For a long time, exchange nailing has been the go-to solution for tibial nonunions post-nailing [10]. However, recently, more surgeons are using plate-over-nail, especially in tricky cases (for example, a tibia that has failed one exchange nailing already or a tibia with a short end-segment that a nail alone struggles to stabilize). The question arises: which is better as a first-line revision for aseptic tibial nonunion after nailing? High-level evidence specifically in tibia is limited; no large randomized trial focusing on tibial nonunions alone has been published to date. There is a shortage of randomized controlled trials (RCTs) and meta-analyses specifically in tibia.

It’s worth mentioning that one can also perform “exchange nailing plus augmentative plating” in extreme cases, though that typically means removing the old nail, inserting a new nail, and also adding a plate. That’s rarely necessary initially but has been done in recalcitrant cases or in some difficult femur nonunions.

Plating after failed plate fixation: Not all tibial nonunions result from nails. If a tibial fracture was initially plated and progressed to nonunion, the revision options are re-plating (often with a larger, stronger plate and graft), switching to an intramedullary nail (if the fracture pattern allows it), or using external fixation (such as an Ilizarov frame). Re-plating with a compression plate and adding bone graft is a common approach for atrophic tibial nonunions that were initially plated. Switching to a nail can be performed if the alignment is acceptable and the soft tissues can accommodate a nail (for example, if an open wound has healed and there is sufficient bone stock to ream a nail). Classic work by Wiss and Stetson [17] demonstrated the success of intramedullary nailing for tibial nonunions after failed plating. Conversely, compression plating has been used after failures of other methods, including external fixation [10]. Essentially, the surgeon must determine which method will provide the best stability for the specific nonunion configuration.

External fixation: In tibial nonunions with significant bone loss or infection, circular external fixation (Ilizarov or Taylor Spatial Frame) can be used to both stabilize and gradually achieve union (sometimes via bone transport or compression techniques). In aseptic cases with intact bone length, external fixators are less commonly chosen if internal fixation is feasible, because patient comfort and joint motion are generally better with internal fixation. However, if the soft-tissue condition is poor (risk of wound complications from internal hardware) or if multiple surgeries have failed, a period of external fixation can sometimes get a tibia to heal by stimulating callus (mechanical stimulation from the frame can be modulated) and by allowing compression or distraction osteogenesis as needed [10]. For an aseptic tibial nonunion with no bone loss, normally you would not jump to an Ilizarov unless internal methods are contraindicated or have failed.

Bone grafting and adjuncts: Virtually all tibial nonunion surgeries (except perhaps the most hypertrophic ones) include some form of bone graft or bone graft substitute application [18]. For the tibia, autologous cancellous bone graft from the iliac crest is commonly added at the nonunion site, especially if atrophic. The RIA technique can harvest a large volume of bone graft from the femur or tibia, which is a popular method to avoid an extensive iliac crest harvest. This can be done concurrently if one is exchanging a nail (ream extra to collect graft). In augmentative plating, one typically opens the nonunion, resects any scar tissue or sclerotic bone (until the “paprika sign” of punctate bone bleeding is achieved), packs the area with bone graft, and then applies the plate [10]. The combination of decortication and grafting is often termed the classic “Müller’s technique” for atrophic nonunions – decortication (opening the bone ends to bleeding bone) and stable fixation with graft yields a high union rate.

Tibial nonunion outcome: With appropriate surgical management, union rates for aseptic tibial nonunions are high, often above 90% after one well-planned intervention [14]. For example, exchange nailing achieves a success rate of 90%, as noted, and augmentative plating yields similarly high results. The time to union after revision is typically in the range of four to six months for plating or nailing, as reported in many series [19]. If the first revision fails, further surgery – sometimes using a different method or adding additional biological enhancements such as bone morphogenetic proteins (BMPs) or cell therapy – can salvage most remaining cases. In summary, for tibial nonunion: Exchange intramedullary nailing has been the workhorse for aseptic tibial shaft nonunions, boasting high union rates (76%–96% in the literature). Augmentative plating (plate-over-nail) is an increasingly popular alternative that provides excellent stability and high union rates, and may lead to faster healing in some cases [16]. The choice of method may depend on the specific scenario; augmentative plating can be especially useful if nail removal is undesirable or if additional stability is needed beyond what a nail can achieve (e.g., distal tibia). Both methods often incorporate bone grafting to maximize success. Good surgical technique – including adequate fixation, alignment, decortication, and infection vigilance – is crucial. In recalcitrant cases or those with bone loss, techniques such as bone transport with external fixation or the Masquelet induced-membrane two-stage grafting come into play [10]. However, for most routine tibial nonunions without infection, a single-stage internal fixation (either exchange nail or nail + plate) with grafting offers an excellent chance of union and return to function.

5.2 Femoral nonunion

Management principles for femoral shaft nonunions are similar to those for the tibia in terms of requiring stability and biology, but the femur has some unique aspects. The femoral shaft is almost always treated primarily with an intramedullary nail (except in certain cases, such as periprosthetic fractures or fractures that are very proximal or distal, requiring plate fixation). Therefore, as with the tibia, a common scenario is a femoral shaft fracture initially treated with a nail that failed to unite. Femoral nonunions can be particularly challenging because the femur experiences high mechanical loads, and patients often expect a pain-free, full weight-bearing function. A nonunion in the femur is very limiting, as patients usually cannot walk without assistive devices or a limp.

Options for aseptic femoral shaft nonunion after nailing: The two main competing strategies are Exchange Nailing versus Augmentative Plating (Plate-over-Nail), analogous to the tibia. Historically, exchange nailing was standard for the femur as well and indeed has a solid success rate reported (often in the 75–95% range) [14]. However, the femur is a larger bone and experiences more torsional stress, and it has been observed that some femoral nonunions (especially when there is a bit of a gap or if the nail is of relatively smaller diameter) might benefit from the added rigidity of a plate.

Exchange nailing (femur): This involves removing the initial nail, reaming the canal larger (often significantly larger, because femoral canals can accommodate large nails), and inserting a new nail with a bigger diameter (and sometimes a different locking configuration if needed). Reaming the femur not only helps insert a larger nail but also promotes a healing response through endosteal bleeding and autografting with reamings. Exchange nailing is classically effective for most femoral shaft nonunions unless specific issues, such as a very distal fracture or a segmental bone defect, are present. In earlier series, initial exchange nailing healed roughly three-quarters of femoral nonunions, with overall union rates approaching or exceeding 90%–100% after additional interventions, especially when a systematic protocol (larger nail, static locking, metabolic optimization, dynamization when needed) was used [14, Swanson et al., 2015]. Exchange nailing in the femur might have a slightly lower first-pass success rate than in the tibia in some older reports, possibly because some femoral nonunions had issues that a nail alone could not fix (such as a large gap requiring bone grafting). In principle, though, a properly performed exchange nailing with up-sizing and possibly static locking (with later dynamization if needed) works for many femoral nonunions.

Augmentative plating (femur): In recent years, augmentative plating of femoral nonunions has been studied extensively, and evidence suggests it may outperform exchange nailing in many cases. The technique is similar: Leave the original nail in place, open the nonunion minimally to clean out fibrous tissue, add bone graft, and apply a strong plate (usually a broad 4.5 mm locking compression plate) bridging the nonunion [20]. By essentially “double plating” the bone (nail inside plus plate outside), the construct becomes extremely stiff, particularly in torsion and bending. The exchange nail construct allows the most rotational motion at the nonunion, which is undesirable for healing. Their conclusion: A robust augmentative plate (4.5 mm) is biomechanically superior to exchange nailing in stabilizing a femoral shaft nonunion. This biomechanical evidence aligns with clinical observations that plating over a nail tends to stop the interfragmentary motion that might be hindering union, especially in cases where a nail alone is somewhat flexible (such as in a wide canal or comminuted fracture with a gap).

More compelling is the clinical evidence from a prospective randomized controlled trial published in 2024 by Motififard et al. [21], which directly compared exchange nailing versus augmentative plating (with autologous bone grafting) in aseptic femoral shaft nonunions. This multicenter RCT provides high-quality data. The key results at 12 months post-revision were as follows:

Time to union was significantly shorter in the augmentative plating group (average 5.4 ± 1.3 months) compared to the exchange nailing group (7.4 ± 2.0 months). This was a highly significant difference (p < 0.001). Essentially, patients healed about two months faster with the plate strategy.
Union rate at 12 months: 100% of patients in the plating + graft group achieved union, versus 89.7% in the exchange nailing group. So, while exchange nailing still had a respectable 89.7% union rate, plating achieved virtually 100% union in this trial.
Clinical outcomes: The plating group had better outcomes on multiple fronts. Operative time was slightly shorter (99 minutes vs. 106 minutes on average), and blood loss was less (about 515 mL vs. 548 mL) in plating compared to nailing, with both differences being statistically significant. Pain relief was greater in the plating group; VAS pain scores dropped from 6 pre-op to 2.6 post-op, whereas in the nailing group, pain went from 5.7 to 3.7. Knee range of motion improved more in the plating group as well. Finally, complication rates were notably different: only approximately 3.6% in plating compared to 17.2% in nailing.

The authors of that RCT concluded that augmentative plating with autologous bone graft is superior to exchange nailing for aseptic femoral shaft nonunions, leading to faster healing, higher union rates, better functional outcomes, and fewer complications [21]. They went so far as to recommend augmentative plating with graft as the preferred treatment option for these cases.

This Level 1 evidence, combined with supportive biomechanical data and prior retrospective studies, is shifting the paradigm in femoral nonunions. Augmentative plating is increasingly seen as the go–to solution, particularly for femur nonunions that might be less likely to respond to a second nail (for example, when the original nail was already large or when the nonunion is at the mid-isthmus, where a nail might still allow rotation). A 2024 meta-analysis of long-bone lower-extremity nonunions after IM nailing [22] found that augmentative plating had higher odds of union (OR ≈ 8.6), shorter union times, less blood loss, and fewer complications than exchange nailing. Subgroup analysis showed shorter union times for augmentative plating in both isthmal and non-isthmal nonunions.

When might exchange nailing still be favored? There are instances where exchange nailing could be the logical choice in the femur:

If the patient’s femoral canal was initially reamed to a small size and a substantially larger nail can now be used, exchange nailing might significantly improve stability.
If the nonunion has minimal callus and you want the reaming effect to boost biology (though one could ream even with plating by using an RIA graft).
If there is an infection concern, one might perform an exchange with an antibiotic-coated nail (but that’s an infected scenario).
Patient factors – some might have a strong preference to avoid a plate.
Surgeon familiarity – not all trauma surgeons are equally comfortable with submuscular plating of the femur.

Other considerations in femoral nonunion surgery:

Distal femur nonunions: If the nonunion is near the distal metaphysis, a plate might be clearly superior because nails in the distal femur have limited distal fixation, whereas a locking plate can capture distal fragments more effectively.
Complex femur nonunions: If multiple surgeries have been performed or the bone is very osteoporotic, a plate combined with a nail is sometimes used for additional support.
Bone grafting: As with the tibia, one nearly always bone grafts atrophic femoral nonunions. In the RCT, the plating group received autograft by protocol, whereas the nailing group presumably did not receive any graft (which might partially explain the higher union rate in the plating group) [21]. This raises an interesting point: Should we also bone graft when performing an exchange nail in the femur? A separate study by Davulcu et al. [23] examined femoral exchange nailing with versus without added bone graft and found no significant difference in union rate or healing time [23]. They concluded that using larger nails is key, and grafting did not significantly change outcomes in their series.
Compression versus bridge plating: In augmentative plating of the femur, typically one uses a bridge plating technique (spanning the nonunion with a long plate and placing screws only far from the fracture, especially if there is any gap, to avoid disturbing the blood supply). However, since the nail is already in situ providing some alignment, some surgeons will perform a so-called “plate-nailing” technique, where they actually compress the nonunion with the plate (if the fracture ends are in contact) to produce axial compression while the nail maintains alignment. Modern locking plates can also be used in compression mode if needed. In hypertrophic nonunions, performing compression with the plate (once fibrous tissue is cleared) can promote direct healing.
Antibiotic-coated nails: For aseptic nonunions, these are not needed, but it is worth noting that there are intramedullary nails with antibiotic coating (e.g., gentamicin-coated nails) that are mainly used for infected cases, though sometimes prophylactically in high-risk scenarios [10].

Outcomes in femoral nonunion surgery: As illustrated by the RCT, excellent outcomes can be achieved. Augmentative plating has shown nearly 100% union with good functional restoration [21]. Exchange nailing, while slightly less successful in direct comparison, still heals the majority of cases (≈ 85%–90% union). When initial revision fails, subsequent attempts usually succeed. It is rare for a femoral shaft nonunion to persist after multiple properly performed interventions unless infection or massive bone loss is an issue.

Summary for femur: The augmentative plating with nail in situ and bone graft approach is emerging as the preferred strategy for aseptic femoral shaft nonunions, given strong evidence of faster and more reliable healing [21]. Exchange nailing remains a viable and historically proven option, especially in cases where plating is less ideal, but evidence now tilts toward plating for most scenarios. Both methods aim to create a stable construct and are often combined with biological enhancement (reaming autograft or iliac crest graft). The specific surgical plan should be individualized. As a final note, patient positioning and approach for femur plating versus nailing differ – exchange nailing is done either supine or lateral on a traction table, usually, while plating often requires a lateral decubitus or supine position with a bump and a fairly large exposure (though sometimes MIPO). With the improved outcomes shown, many orthopedic trauma specialists now recommend augmentative plating (with autograft) as the first-line treatment for a femoral shaft nonunion post-IM nailing, reserving exchange nailing for select cases.

6. Role of bone grafting in nonunion treatment

Bone grafting is a cornerstone of surgical management for atrophic and oligotrophic nonunions, and even in hypertrophic cases, it is often used to speed healing. The fundamental principle is to supply the nonunion site with the biological components that may be lacking: osteogenic cells, osteoinductive factors, and an osteoconductive scaffold. Different types of bone grafts and bone-graft substitutes provide these to varying degrees. Here, we review the main options: autograft (including iliac crest and RIA graft), allograft, and synthetic bone graft extenders or substitutes. We will also touch on specialized grafting techniques, such as vascularized grafts.

Autogenous bone graft (autograft): This remains the gold standard for promoting bone union [24]. Autograft is typically harvested from the patient’s own bone, most commonly the iliac crest. It contains living osteogenic cells (osteoblasts and progenitors), osteoinductive growth factors (like BMPs stored in the bone matrix), and provides an osteoconductive scaffold (the bone mineral itself). When placed into a nonunion site, autograft can directly participate in new bone formation and stimulate the host bone to heal. There are two main forms: Cancellous autograft (the spongy medullary bone, which has a high cell and factor content and revascularizes quickly) and cortical autograft (structural pieces of cortical bone, which provide mechanical support but revascularize more slowly). For most nonunions without large segmental defects, cancellous autograft is used because it incorporates and induces healing more rapidly.

The iliac crest (pelvis) is the most common donor site because it has a substantial volume of cancellous bone that is easily accessible. Typically, a surgeon can harvest 30–60 cc of graft from the anterior iliac crest (or more from the posterior crest if needed). Iliac crest bone graft (ICBG) has a very high success record; one reference calls it the “benchmark for the treatment of nonunion regardless of age” [10]. Even in elderly patients, the iliac crest graft still possesses the necessary biology to aid union. Although older patients have fewer stem cells, they still benefit from the graft’s osteoconductive and osteoinductive properties.

The downside of autograft is donor site morbidity. Harvesting from the iliac crest can cause significant pain, risk of bleeding, risk of infection, and, rarely, complications like hernia or fracture at the pelvis. Patients often report pain at the crest that can last weeks or months; chronic donor site pain is reported in a minority of patients. Because of these issues, surgeons have sought alternative sources and methods to minimize donor site morbidity.

One such alternative is the RIA system (see Figures 1–5), which allows harvesting bone graft from the intramedullary canals (usually the femur, sometimes the tibia) by reaming under suction. The RIA sucks out a paste-like mixture of reamings (which contain bone marrow, small bone fragments, and blood) into a filter trap. This material can yield a large volume of graft (up to 100 cc) rich in marrow elements. RIA graft is essentially intramedullary autograft; it contains marrow MSCs and growth factors, but in a more liquid/cancellous form. Studies have shown RIA graft to have similar union promotion as iliac crest graft for long bone nonunions [25]. The advantage is that there is no separate incision in the pelvis; however, using RIA means you have to access a long bone’s canal (which is easy if you are already performing an exchange nailing, for instance). RIA is especially useful in cases with large defects or when you need a lot of graft (such as in the Masquelet technique second stage, where you might need 50–100 cc). It also avoids donor site morbidity associated with iliac crest harvest. There are some risks with RIA too; reaming can cause cortical perforation or major bleeding (the RIA flushes fluid, risking fat embolism or hemorrhage if not careful). But overall, it is a well-established technique now.

Figure 1.
A 71-year-old female involved in a motor vehicle accident, sustaining multiple traumas including multiple bilateral rib fractures, an open right tibial shaft fracture (Gustilo–Anderson type III), a closed left distal third tibial shaft fracture,, and a closed right mid-shaft femoral fracture.

Figure 2.
Sixteen weeks postoperatively, radiographs of the left distal third tibial shaft fracture treated with a circular external fixator (with the foot frame removed at eight weeks postoperatively) demonstrated an absence of visible callus formation and no progression.

Figure 3.
Five months postoperatively, radiographs demonstrated atrophic nonunion of the distal third of the tibial shaft.

Figure 4.
Intraoperative images showing autologous bone graft harvested using the reamer–irrigator–aspirator (RIA) system, performed retrogradely through the ipsilateral femur, followed by placement of the RIA bone graft into the defect site.

Figure 5.
Nine months postoperatively, the tibial nonunion had achieved complete radiographic union with a good functional outcome; the patient was able to bear full weight without pain.

Another form of autograft is autologous bone marrow aspirate (BMA) or concentrate (BMAC). It involves injecting marrow (with or without concentration) into a nonunion site. While not a “graft” in the structural sense, it is a cellular graft. It has shown good results in some cases and can be used alone or to augment other grafts.

For large structural deficits (>2–3 cm of bone loss), vascularized autografts can be employed. A common example is the free vascularized fibula graft, where a segment of fibula (with its artery and vein) is transplanted to the defect and reconnected via microsurgery to blood vessels near the defect. This provides a living cortical graft with its own blood supply, which can heal into the host bone over time and hypertrophy. Vascularized grafts are complex and usually reserved for scenarios such as atrophic nonunions with massive bone loss or cases where the blood supply of the bed is poor (e.g., after radiation or in a very large defect following tumor resection or infection debridement). In the context of this chapter, for a run-of-the-mill post-traumatic nonunion with a small gap, a vascularized graft is seldom needed; it is more commonly used for limb salvage in extreme cases (such as a 10 cm femoral defect). When blood supply is severely compromised, a vascularized graft can succeed where a non-vascularized graft would fail. Vascularized grafts (e.g., fibula) have shown good success but require specialized surgical expertise. Orthopedic references point out that vascularized bone grafts can achieve union in situations of poor blood supply and are the graft of choice in such cases [26]. However, they carry donor site morbidity as well (losing a fibula can slightly affect ankle stability, though usually not significantly if one fibula remains).

Allograft: Allograft refers to bone graft obtained from a donor (cadaver bone, typically processed and stored in a bone bank). It comes in many forms: cortical shafts, corticocancellous chips, morselized cancellous bone, or demineralized bone matrix (DBM; more on that below). Allograft provides a scaffold (osteoconductive), and if not heavily processed, it retains some collagen and proteins that may be mildly osteoinductive. However, it contains no living cells (unless it is a fresh transplant, which is rarely used due to immune reaction). The main advantages of allograft are its availability in large quantities and the avoidance of donor site surgery for the patient. Large structural allografts (like a femoral segment) can be used to bridge massive defects in limb reconstruction. Morselized allograft (freeze-dried chips) is often used as a filler or extender with autograft, for example, mixing autograft 1:1 with cancellous allograft to increase volume.

The downsides include slower incorporation (allograft bone must be slowly remodeled by the host, which can take a long time for large pieces) and the potential for immune response or graft resorption. Disease transmission is extremely low but theoretically possible (modern processing and screening make it very safe, albeit not zero risk). Allograft is often used when autograft is insufficient; for example, in a 5 cm defect after debridement of an infected tibia, one might perform the Masquelet technique and then fill the induced membrane with a mixture of autograft and allograft to achieve sufficient volume. Another scenario: if a patient cannot tolerate autograft harvest (for instance, due to severe osteoporosis or because they have already had their crest harvested previously), one might use allograft chips combined with orthobiologics.

Demineralized bone matrix (DBM): This is a form of allograft that is processed through acid extraction of minerals from cadaver bone, leaving behind the organic collagen matrix and growth factors (notably BMPs) embedded within it. DBM is available as putties or strips. It is primarily osteoinductive (due to BMP content) and osteoconductive (collagen scaffold), but not osteogenic (as it contains no cells). DBM has mild bone-healing capabilities on its own but is usually used as a graft extender (mixed with autograft to spread the autograft further) or as an adjunct to provide some BMP stimulus. Some studies show that DBM alone in a nonunion has limited success, but when combined with autograft, it may enhance the result [25].

Synthetic bone substitutes: These include materials like calcium phosphate ceramics (e.g., hydroxyapatite [HA], tricalcium phosphate), calcium sulfate (plaster of Paris-type materials), bioactive glass, and newer composite scaffolds. Synthetics are osteoconductive scaffolds; they provide a framework for bone growth but usually lack cells or strong inductive factors (unless combined with something like BMP). They come in forms like granules, blocks, or injectable cements. For example, an osteoconductive calcium phosphate cement can be used to fill a bone defect; it hardens, provides immediate structural support, and over time is resorbed and replaced by bone. In nonunion surgery, synthetics can be used to fill residual gaps or augment autograft. One common strategy is to mix synthetic granules with BMA; the synthetic acts as a scaffold, and the marrow brings cells and signals.

Calcium sulfate (e.g., “bone plaster”) is sometimes used to fill small defects; it dissolves relatively quickly (within weeks to months) and is fully replaced by bone if everything goes well, but it does not provide long-lasting support. It is more often used as a carrier for antibiotics in infected sites (e.g., Stimulan beads with antibiotics).

Calcium phosphates and HA are more slowly resorbed, providing a scaffold for a longer duration. Some products are even preloaded with collagen or other growth factors. The lack of osteoinductive and osteogenic properties, as well as the high cost of synthetic grafts, has been noted as a concern [10]. They do not contain living cells and typically do not have significant amounts of BMP unless specifically engineered to include it. Therefore, they are usually used as adjuncts rather than as the primary grafts for addressing challenging nonunions.

Graft placement techniques: How the graft is placed also matters. Traditionally, for a stiff atrophic nonunion, surgeons perform an “open decortication and grafting,” the nonunion site is fully exposed, all fibrous tissue and sclerosis are removed, and multiple holes are drilled or the cortical bone is scored to induce bleeding (paprika sign) [27]. The area is then packed circumferentially with autograft (sometimes mixed with allograft or DBM) and stabilized with a plate or nail. This creates a local environment rich in osteogenic material.

Special mention – Bone grafting in the presence of infection: In infected nonunions, bone grafting is usually deferred until the infection is controlled (stage 2 of Masquelet or after a period of antibiotic nails, etc.). There are also specialized graft-like techniques, such as the Masquelet-induced membrane technique, where an antibiotic cement spacer is placed in a defect for six to eight weeks to induce a vascularized membrane. The spacer is then removed, and the space is filled with autograft (often combined with allograft/DBM to bulk it) inside that membrane [26]. This technique has shown good success in treating segmental bone loss of up to 25 cm in some series.

To sum up, the role of bone grafting in nonunions is to re-establish the biology needed for bone healing. Autologous bone (iliac crest or RIA reamings) is the most potent option and is indicated in most atrophic nonunions or any case with bone loss or poor healing potential. It accelerates union and improves union rates significantly [10]. Allografts and synthetic grafts are important adjuncts, particularly when volume is needed or to avoid additional morbidity, but they are usually combined with autograft or orthobiologics for the best results. Ultimately, successful bone grafting requires a well-prepared bed (debridement of scar and sclerotic bone to ensure vascular contact) and stable fixation so that the graft is not extruded and the new vessels can penetrate. When appropriately applied, bone grafts (especially autograft) dramatically improve healing in difficult nonunions, making the difference between persistent failure and successful union in many cases [24].

7. Infection and complex nonunions

Nonunions complicated by infection (septic nonunions) or those with large bone defects and other complexities represent the most challenging subset of cases. They require a tailored approach that addresses both the underlying infection and the lack of bone healing, often necessitating multiple stages of treatment and advanced reconstruction techniques. Infected and complex nonunions typically arise from high-energy open fractures, cases with previous hardware infections, or multiple failed surgeries where biology is severely compromised. Here, we outline strategies for managing these difficult situations.

7.1 Infected nonunions

An infected nonunion means there is persistent osteomyelitis at the fracture site. The dual problem is as given below:

Infection control, eradicating bacteria and infected tissue;
Fracture stabilization and bone healing, achieving union once the site is clean.

These often conflict because infection control might require the removal of hardware and debridement (which destabilizes the bone), and conversely, attempting to stabilize with hardware can introduce surfaces for bacteria to cling to. Thus, careful planning is needed, often in stages.

General principles: As one review aptly stated, infected nonunions present “two interrelated orthopedic problems: deep bone infection and a failure of fracture healing” [28]. The treatment plan must explicitly address both. This usually involves adequate debridement of all infected and nonviable tissues (sequestra, infected bone, scar); culture of tissues to guide antibiotic therapy; stability of the bone (often achieved with external fixation initially, since internal hardware in an infected field is risky until infection is controlled); reconstruction of bone loss (via bone graft, bone transport, or other techniques) once infection is quiescent; and systemic and local antibiotics for a sufficient duration.

Classically, infected nonunions have been treated in stages:

Stage 1: Eradicate infection. Remove any existing hardware (since it harbors biofilm), perform aggressive surgical debridement of the nonunion site (remove all necrotic bone until bleeding healthy bone – the Paprika sign is achieved), collect multiple deep cultures [27], and initiate targeted antibiotic therapy. After debridement, one must maintain stability of the limb; often, an external fixator (e.g., an Ilizarov ring or a unilateral frame) is applied to stabilize the bone after debridement. In some cases, an antibiotic-impregnated cement spacer or an antibiotic cement nail (for intramedullary infection) is placed to deliver a high local antibiotic concentration and maintain some structural support.

Stage 2: Once the infection is under control (typically after a course of 6–12 weeks of antibiotics and confirmation that cultures have become negative and systemic signs have subsided), proceed to definitive reconstruction. This could involve conversion to internal fixation with bone grafting if the infection is eradicated, or bone transport (distraction osteogenesis) using an external fixator to fill any defects, or the Masquelet technique if a large segment has been removed [26].

However, with modern techniques, some cases can be managed in a single stage; for instance, if the infection is low-grade and bone loss is minimal, one might ream out the infection, re-insert a new nail with an antibiotic coating, and concurrently graft. The success of this approach can be good in select cases, but careful patient selection is essential.

Specific strategies:

Antibiotic-coated nails: Intramedullary nails with antibiotic coatings (gentamicin-coated) are now available [29]. These allow one-stage reaming, nail insertion, local antibiotic release, and stability, which can lead to eventual union while treating infection. They are primarily used in femoral or tibial infections. Studies show good infection control and union with these, often obviating the need for a stage with an external fixator.

Ilizarov external fixation (bone transport): For infected nonunions with bone loss, the Ilizarov method is a time-tested solution. After debridement, a circular external fixator is applied. One technique is bone transport, where an osteotomy is created at a healthy bone site (e.g., proximal tibia), and that segment is slowly moved (distracted) across the defect toward the distal segment, generating new bone in the wake [30]. This simultaneously addresses bone loss and avoids large grafts. It is particularly useful if more than 4–5 cm of bone is missing. It also keeps the site open, so if infection flares, additional debridement is possible. The downside is patient discomfort and the long duration in a frame (several months to a year).

Masquelet-induced membrane technique: This two-stage technique has gained popularity for both aseptic and septic segmental bone defects. Stage 1: Place a PMMA antibiotic cement spacer in the bone defect after thorough debridement [26]. This spacer maintains the space and releases antibiotics. The body forms a vascularized membrane around the spacer in four to eight weeks. Stage 2: Remove the spacer and fill the cavity with bone graft (often an autograft + allograft mixture) inside the induced membrane, and stabilize it (either internally or externally). The membrane provides a friendly environment rich in growth factors (like a pseudo-periosteum) that helps the graft consolidate. This technique is effective for defects often up to 25 cm, though large defects may need repeat grafting.

Acute limb shortening and re-lengthening: In some cases, surgeons excise the infected nonunion (removing all compromised bone) and then acutely shorten the limb to bring the bone ends together, addressing the defect by overlap. This approach resolves the gap problem and can eliminate the infection in one step, albeit at the expense of limb length. Subsequently, they gradually lengthen the limb at a different site using an external fixator (similar to performing distraction osteogenesis at the proximal bone). This method addresses both infection and union in one aggressive step; it is particularly considered when soft tissue coverage is problematic or when a very large resection is required (such as acute docking followed by lengthening) [31].

Flap coverage: Many infected nonunions follow open fractures where soft tissue has been damaged. Ensuring good soft tissue coverage is crucial. If a tibial nonunion is open or has scarred skin, bringing in a muscle flap (vascularized muscle or free flap) can help deliver blood supply and antibiotics to the area, aiding in the fight against infection. Flaps (such as a free fibula flap or free muscle flap) can be part of the reconstruction, such as combining a fibula flap for structural support and infection control [32].

Suppression and union-first strategy: There is an approach wherein, if a nonunion is well-fixed and not grossly purulent (just a low-grade infection), one strategy is to suppress the infection with antibiotics (without removing hardware) until union occurs, then remove the hardware and treat osteomyelitis. This is somewhat controversial, but in certain cases (like a minimally symptomatic infection, stable hardware, and good progression of healing), a surgeon might keep the hardware in place, give the patient long-term suppressive antibiotics, allow the fracture to unite, and then remove all hardware and clean out any remaining focus of infection. Some references call this “treat the fracture, then the infection definitively” [33]. This is one strategy out of several. It carries the risk that the infection could worsen but has the benefit of achieving stability and union without destabilizing the construct early.

Outcomes: With comprehensive management, union rates in infected nonunions can still be high (80%–95%), but it often requires multiple surgeries and a prolonged course. The resolution of infection is also high if protocols are followed (most series report > 85% infection eradication, although it sometimes requires more than one debridement). It demands a multidisciplinary approach: collaboration with infectious disease specialists for tailored antibiotic therapy (often six or more weeks of IV antibiotics followed by some oral antibiotics), plastic surgeons for soft tissue coverage, and more. Patience and patient compliance (with external fixator care, avoiding smoking, and maintaining proper nutrition) are critical. Each case is unique; some might be resolved in three months, while others may take 18 months and numerous operations.

7.2 Complex nonunions (large defects, deformities, etc.)

This category often overlaps with infected cases but also includes aseptic cases with significant bone loss or those that have failed multiple surgeries and have poor biology. Examples include a 5 cm segment missing from the femur after a tumor resection (not infected but a complex nonunion scenario), a tibia that has undergone three prior surgeries and is now atrophic, or a nonunion with concurrent issues such as joint contractures or limb length discrepancy.

Large bone defects (aseptic): The strategies mirror those for infection, minus the infection concern. The Masquelet technique is very popular for aseptic defects as well and has shown good success (some studies report ~80%–90% union in large defects) [26]. Bone transport using Ilizarov frames is a classic approach, yielding union in the majority of cases, though the frame time is long (the “rule of thumb”: time in frame = 1 month per cm of bone transport plus 2 months, roughly) [30]. Vascularized fibula grafts: For very large defects (>6 cm), some surgeons prefer a free fibula transfer, often combined with a plate for immediate stability. Vascularized fibula can hypertrophy over time and provide a living solution to a large defect, with union at host bone interfaces in ~80% initially, increasing to ~95% after any necessary reoperations [34]. Combinations: For example, plating the bone to bridge a defect with a massive allograft segment and augmenting it with BMP and autograft inside, then hoping it consolidates. Large allografts can incorporate slowly or fracture, but in selected cases (such as a long femoral gap), an allograft segment is one option, often used in tumor reconstructions [35].

Cages or structural scaffolds are employed to bridge large segmental defects, restore anatomical continuity, and provide a stable environment for bone graft incorporation [2]. The treatment of bone loss and nonunion using a cage and autograft is grounded in the principles of mechanical stabilization (via cage or internal fixation) and biological augmentation (using autogenous ICBG), often supplemented by bioregenerative adjuncts (see Figures 6–9).

Figure 6.
A 19-year-old male was involved in high-energy motor vehicle accident sustaining multiple traumas including an open comminuted distal tibial fracture (right ankle), classified as Gustilo–Anderson IIIC. Initial management included monoplanar external fixation, wound debridement, and vascular repair of the anterior and posterior tibial arteries in addition to fasciotomy.

Figure 7.
One month later, the patient underwent circular external fixation (TL-HEX) with resection of 6 cm of distal tibia and placement of vancomycin-loaded cement spacer.

Figure 8.
Five months later, conversion to internal fixation with intramedullary nailing and cage support was performed using an autologous iliac crest bone graft to fill the bone loss.

Figure 9.
Four years later, the bone fully healed with good functional outcomes and no signs of infection.

Nonunions with malalignment: Sometimes a nonunion is present alongside a malunion (poor alignment). The revision should address the deformity while also achieving union. Techniques involve performing an osteotomy to realign the bone, followed by fixation with a bone graft. External fixators (e.g., Taylor Spatial Frame) can gradually correct the deformity and compress the nonunion [36], or internal methods can be used with a carefully planned osteotomy. Correcting alignment is crucial because a malaligned nonunion, even if forced to heal, could result in poor function and an increased risk of re-fracture.

Host optimization: Complex nonunions often require addressing patient factors aggressively, for example, metabolic workup (vitamin D, thyroid function, etc.), cessation of NSAIDs, strict no smoking (possibly even using medical therapy like Chantix to ensure quitting), good nutrition with sufficient protein and possibly supplements (vitamin D, calcium). If the patient has endocrine issues or is on steroids, involvement of an endocrinologist to manage those could improve outcomes [6]. Some centers have “fracture healing clinics” to optimize host physiology.

Adjuvant therapies: In extremely difficult cases, surgeons might throw the kitchen sink; for example, revision with plating, autograft, BMP-2, and then post-op PEMF, and maybe teriparatide for six months. These combined approaches follow the diamond concept to maximize success [6]. It is not always clear which element made the difference, but the goal is to give the bone every chance to heal.

Functional impact and realistic goals: Managing complex and infected nonunions can be a marathon. It is critical to communicate with the patient about the timeline, the potential need for multiple procedures, and the expected outcomes. The psychological toll is significant – many such patients have been dealing with a nonunion for a year or more, often with chronic pain and repeated hospitalizations [1]. Psychological support and counseling are important (and often underappreciated) components of care. The surgeon should work with a multidisciplinary team, including physical therapists to maintain joint motion and muscle strength during treatment, plastic surgeons for soft tissue management, and mental health professionals for psychological support [33].

In summary, infected and complex nonunions require an individualized, often staged approach that eradicates infection, provides stability, and reconstructs bone and soft tissues. Successful outcomes are attainable, with reported union rates often exceeding 85% after comprehensive treatment, but achieving these outcomes requires significant time and resources. This underscores the importance of specialized centers (limb reconstruction centers) for these cases, as evidence shows that such multidisciplinary teams improve outcomes with fewer complications [30]. Patience and perseverance from both the surgeon and the patient are key, as is the willingness to adapt the plan if one method fails (having a plan B, C, or D as needed).

8. Summary and recommendations

Fracture nonunion of the tibia or femur is a complex but treatable condition. Through advances in both surgical techniques and adjuvant therapies, orthopedic surgeons today can achieve union in the vast majority of cases, even those that historically might have resulted in amputation or permanent disability. This chapter has detailed the multifactorial nature of nonunions and the array of tools available to address them. Here, we distill key points and provide practical recommendations:

Early identification and classification: Distinguish delayed healing from nonunion early. If a tibial or femoral fracture shows no progress by four to six months, begin evaluation for causes of nonunion. Use a pragmatic definition: Any fracture that will not heal without intervention should be considered a nonunion. Classify the nonunion by radiographic appearance (hypertrophic, oligotrophic, atrophic) and by whether infection is present. This guides treatment: Hypertrophic typically means stability is the issue, atrophic means biology is lacking, and infection changes the entire algorithm. Always investigate for occult infection in “aseptic” nonunions (get cultures intra-op), a routine advocated by Mills et al. [5].
Address all contributing factors (mechanical and biological): Successful nonunion treatment requires correcting mechanical instability and enhancing the biological environment. This “dual requirement” is encapsulated in the Diamond Concept [6]. In practice:
- Mechanical: Provide adequate stability through internal fixation (plate or nail) or external fixation. In tibial nonunions, both exchange nailing and plate-over-nail yield high union rates; choose based on patient factors and surgeon expertise. Meta-analytic data suggest that augmentative plating can offer higher union odds and shorter healing times in long-bone nonunions overall [22]. For most femoral nonunions, we recommend augmentative compression plating with autologous bone graft as the first-line revision strategy, given its faster union and higher success rates.
- Biological: Atrophic nonunions need aggressive biological enhancement. Perform thorough decortication of the nonunion site, removing sclerotic bone and interposed fibrous tissue until punctate bleeding bone is seen (paprika sign) [4]. Then, add autologous bone graft (iliac crest or RIA reamings) liberally, as it remains the gold standard osteogenic material [24]. Consider adjuvants like BMP-2 or BMP-7 in refractory cases or when autograft is limited; they have been shown to significantly improve union rates and can substitute for autograft in certain scenarios [2].
Infection control: If infection is present or suspected, prioritize its management. Our recommendation is a staged approach: Stage 1 – debride thoroughly and eradicate infection using external fixation or an antibiotic-coated nail for provisional stability, and administer targeted antibiotics (guided by cultures) for at least six to eight weeks [37]. Stage 2 – reconstruct after confirming infection clearance (normalized CRP, good soft tissue healing). Reconstruction may involve internal fixation and grafting or bone transport/Masquelet technique for large defects [26]. Alternatively, consider one-stage strategies in select cases using antibiotic-coated implants if tissue conditions allow – but be prepared to convert to a multi-stage approach if any doubt remains. Always obtain multiple deep cultures at the time of nonunion surgery (even if considered aseptic) to detect latent infection [38].
Use evidence-based fixation strategies: For femoral shaft nonunions after nailing, the latest Level I evidence clearly supports augmentative plating with autograft over exchange nailing [21]. We therefore advise augmentative locked plating (4.5 mm plate) applied over a retained nail with concurrent autograft for femoral nonunions, especially those with any rotational instability or those that have already undergone one failed nailing. The biomechanical superiority and clinical success (100% union in RCT) make it the preferred option. Either approach is acceptable; if one method fails, the other can serve as a backup. Locking plates spanning the nonunion combined with intramedullary stability (either nail in situ or fibular graft) provide excellent outcomes and should be part of the armamentarium.
Adjunctive stimulation: Consider employing bone stimulators as adjuncts, particularly in situations where additional surgery is risky or as a complement post-surgery. We recommend LIPUS (20 minutes daily) for delayed unions or as an adjunct after revision fixation of an atrophic nonunion, given that LIPUS is safe, although unsupported by strong evidence [12]. Likewise, PEMF therapy (eight hours daily) is an effective adjunct or standalone treatment, with healing rates of approximately 75%–85% in long bone nonunions [11]. For example, in a patient who is a poor surgical candidate or refuses surgery, a trial of PEMF is warranted and often successful in achieving union over several months. Even after surgical fixation, there is minimal downside to adding a PEMF device to “hedge” against another nonunion, especially in biologically tenuous cases. Ensure patient compliance with these devices, as their effectiveness depends on regular use.
Plan for bone loss and complex reconstruction: In cases of significant bone loss or segmental defects (common in high-energy tibial injuries), anticipate the need for advanced techniques:
- For defects measuring 1–5 cm, the Masquelet-induced membrane technique is a reliable two-stage option. We recommend this method for managing contained defects: Stage one involves an antibiotic spacer, and stage two utilizes an autograft/allograft filler with stable fixation. This approach has demonstrated high union rates, even in infected scenarios [26].
- If the defect is larger or if limb length is a concern, use Ilizarov external fixation for bone transport. Although lengthy, it can restore both length and continuity without a graft (especially useful if infection precludes large grafting) [30].
- Consider vascularized bone grafts (e.g., free fibula flap) for very challenging defects or when the regional blood supply is poor. These grafts bring live bone and blood supply and have excellent success in experienced hands for bridging massive gaps [34].
- Always involve plastic or reconstructive surgeons early if soft tissue coverage is an issue – a well-vascularized soft tissue envelope is paramount for both infection control and bone healing. A muscle flap can be as important as the bone work in some cases [6].
Psychosocial support: Do not overlook the psychological impact. Nonunion patients often endure chronic pain, loss of function, and social role changes (e.g., inability to work, financial stress). Depression and anxiety are common. We recommend incorporating mental health evaluation and support as part of nonunion care [1]. This might involve counseling, support groups (some find solace in hearing other patients’ journeys), or appropriate medications if needed. As the clinician, maintain a positive yet realistic outlook with the patient. Reassure them that union is achievable (e.g., “With the plan we have, there’s over a 90% chance your bone will heal,” is encouraging) but also prepare them for the commitment and possible hurdles. Setting expectations about treatment duration and potential setbacks helps patients cope better and adhere to the plan.
Multidisciplinary approach: Whenever possible, treat nonunions in specialized centers or teams that include orthopedic trauma/reconstructive surgeons, plastic surgeons, infectious disease specialists, physiatrists, and so on. The evidence and our experience suggest that a multidisciplinary, comprehensive approach yields the best outcomes in complex nonunions [33]. This ensures each aspect (bone, soft tissue, infection, rehabilitation, psychological) is optimally managed. Regular team meetings to review progress can be useful for complex cases.
In conclusion, successful management of tibial and femoral nonunions hinges on identifying and correcting all factors inhibiting healing, from providing rigid fixation and viable bone grafts, to controlling infection and optimizing patient health. The combination of modern fixation methods (particularly augmentative plating for long bone nonunions), advanced bone grafting and orthobiologic adjuncts, and dedicated patient care has made even the most stubborn nonunions curable in most instances. By following evidence-based strategies and tailoring them to each patient’s unique situation, orthopedic surgeons can expect high union rates and restored function. Always remember the aphorism: “The only failure in nonunion treatment is failure to address the cause of the nonunion.” By thoroughly addressing mechanical stability, biological vitality, and host factors, the cause(s) of nonunion are removed, and successful bone healing will follow.

Conflict of Interest

The authors declare no conflict of interest.

Xuejiao Hu

Your cart