Nonunion, Delayed Union, and Malunion

Bone is unique in its tremendous ability for repair via regeneration rather than scarring. Indeed, fracture healing typically proceeds in an orderly fashion provided the necessary systemic factors and local environment are present [1]. In fact, bone healing often occurs without veterinary intervention, although the convalescent and long-term morbidity may not be ideal. It is, thus, the veterinarian’s role to determine when and how to intervene so that patient morbidity is minimal and functional outcome is maximized. Simplistically, the veterinarian must provide the fracture zone with biologic and mechanical environments that are conducive to the natural reparative process. Failure to provide either of these prerequisites predictably results in failure to achieve bone union and restore limb function.

Union

Bony union refers to a healed fracture. Although it sounds simple enough, considerable “gray area” exists as to when union is achieved depending on the measurement tool being used. “Radiographic union” refers to a subjective assessment of radiographs where the evaluator is satisfied that adequate amounts of new reparative bone have been deposited to achieve bridging of the fracture. “Histologic union” and “mechanical union,” where the healed bone is indistinguishable from the bone prior to injury, may take many months to years. A clinician often refers to “clinical union” where he or she is convinced by the radiographs, an appropriate healing time period, and the patient’s use of the limb that functional healing has occurred.

Delayed Union

A delayed union is defined as slower than anticipated healing compared with similar fractures treated with similar methods in similar patients. This term is vague because numerous variables exist, including patient age, systemic health, degree of soft tissue injury, presence of concurrent injuries or disease, patient size, fracture configuration and location, treatment method, and surgeon expertise, that may all affect the healing rate [2]. When delayed union is diagnosed, the veterinarian must determine whether treatment intervention is indicated. This determination is based on serial evaluations of radiographic appearance, patient comfort, and limb function. Delayed union is a diagnosis of a dynamic process that may either progress to complete union or culminate in nonunion. Because delayed union can be a transitional state between injury and nonunion, the distinction between delayed union and nonunion may be subtle in some instances.

Nonunion

Delayed union may culminate in nonunion. A nonunion fracture is one in which progression of healing has ceased and union is unlikely without further intervention [2,3]. Nonunions are classified as viable or nonviable based on the biologic activity detected at the fracture site. This distinction directs subsequent treatment strategies [2,3].

Terminology

Viable Nonunions

Viable nonunion (also called reactive or vascular or biologically active nonunions) are biologically active at the fracture site and have evidence of bone reaction and callus formation [2-4]. Because radiography is the traditional instrument used for assessment of fracture healing, viable nonunions are further subclassified based on the degree of radiographic callus present (Fig. 91-1). Hypertrophic nonunions, often called “elephant’s foot nonunion” have an abundant amount of radiographic callus and are often caused by inadequate fracture stabilization in a highly vascularized region of a young animal. Moderately hypertrophic nonunions are characterized by lesser degrees of radiographic callus formation, often referred to as “horse hoof callus,” and are often associated with fracture instability in a bone with less soft tissue attachment and/or in a more skeletally mature animal. Oligotrophic nonunion, although considered biologically active, produces little callus and can be difficult to distinguish from nonviable nonunion. Nuclear scintigraphy may demonstrate biologic activity of the fracture zone where it is not evident on radiographs.5,6 When scintigraphy is not readily available, it is clinically advisable to err toward classification as nonviable nonunion as these fractures are treated more comprehensively.

Figure 91-1A. Types of viable nonunions. (A) Hypertrophic (“elephant’s foot”) have an abundant amount of radiographic callus; (B) Moderately hypertrophic (“horse-hoof”) have a moderate amount of radiographic callus; and (C) Oligotrophic have minimal to no detectable radiographic callus (making them easy to confuse with nonviable nonunions).

Nonviable Nonunions

Nonviable nonunions (also called biologically inactive nonunions) have a severe interruption of blood supply and lack callus production and osteogenic potential. Nonviable nonunions are subdivided into four categories (Fig. 91-2) [2,3]. Dystrophic nonunion is nonunion in which the blood supply to an intermediate fragment is compromised. The fragment heals to one of the main fracture segments, but not to the other. The area of nonunion is biologically inactive as evidenced by a lack of callus, a persistent fracture gap, and rounded, sclerotic fracture ends.

Figure 91-2. Types of nonviable nonunions. (A) Dystrophic nonunion has healing of an intermediate fragment to one of the main segments, but not the other; (B) Necrotic nonunions have one or more isolated, nonhealing, avascular bone fragments (“sequestra”) that appear as radiodense fragments with sharp margins; (C) Defect nonunions have a large, nonhealing fracture gap; (D) Atrophic nonunions have defect filled with scar tissue and partial resorption of adjacent fractured bone ends.

Necrotic nonunions have one or more avascular bone fragments within the fracture zone. These avascular bone fragments, called sequestra, have no union with any surrounding bone and have sharp margins and sclerosis on radiographs. Necrotic nonunions in dogs and cats are nearly always the result of excessive elevation of soft tissues from a bone fragment combined with bacterial contamination and inadequate stabilization. Conversely, necrotic nonunion seldom occurs when closed or “open but do not touch” treatment strategies are employed to highly comminuted fractures. Defect nonunions occur when there is a large gap at the fracture site as a result of the original trauma or from bone fragment loss owing to sequestration, resorption, or surgical removal. Regardless of the specific cause of the defect nonunion, a critically sized defect exists and callus cannot bridge the gap. Atrophic nonunion is the end result of the preceding types of nonviable nonunion. The bony defect is filled with scar tissue and the fractured bones are partially resorbed. Poor limb use leads to osteoporosis and muscle atrophy.

A pseudoarthrosis may occur with time as a result of motion at the site of the nonunion [2,3]. The clinical conditions are sclerosis of the bone ends, formation of fibrocartilage between the bone ends, and a fibrous capsule filled with serum. Periarticular fibrosis may cause stiffness of the adjacent joint such that much of the motion is concentrated in the pseudoarthrosis, especially if it is near a joint.

Etiology

A complex array of local and systemic factors is necessary to achieve predictable bone healing [1,7,8]. Systemic factors include the patient’s nutritional state, hormonal balance, age, general health, and medications. Local factors include the conditions within the fracture zone such as health of the soft tissues, size of the fracture gap, stability, contamination/infection, and foreign materials.

The local environment of the fracture zone must be suitably stable (mechanical environment) and viable (biologic environment) in order for bone healing to occur. Although many local and systemic factors may contribute to the development of nonunion, more often than not, an imbalance in the mechanical and/or biologic conditions of the fracture zone is the major contributor to nonunion.

Poor blood supply to the fracture zone is a common contributing cause of delayed union or nonunion. Blood supply to healthy, intact diaphyseal long bone is primarily from centrifugal flow outward from the intramedullary space [9]. The outer one third of bone is supplied by periosteal vasculature arising from the soft tissues surrounding the bone. When a bone is fractured, the normal blood supply is disrupted and the initial vascular supply for bone healing comes almost entirely from the surrounding extraosseous tissues. Being a viscoelastic material, bone is stiffer and capable of absorbing larger amounts of energy prior to failure when it is more rapidly loaded [10]. Once the failure point is reached, however, the bone fractures and releases this absorbed energy to the surrounding soft tissue envelope, disrupting the blood supply to the fracture zone. Such high energy absorbed to failure often produces a highly comminuted fracture pattern. A surgeon who attempts to anatomically reconstruct the highly comminuted fracture pattern often compounds the vascular compromise to the injured bone [11]. During the extensive manipulation of each fracture fragment, the soft-tissue envelope is further disrupted. If cerclage wire is used in these efforts, it must be used properly because improperly used wire typically loosens and cleaves off any developing extraosseous blood supply [12]. Because even the most delicately performed open approach to the fracture zone induces some injury to the fracture zone’s blood supply, the surgeon should contemplate the biologic impact of the treatment plan before starting surgery. Poor blood supply may also be inherent to the region of the fracture. Whereas the femur has a strong muscular attachment along its caudal margin at the linea aspera, the distal tibia has minimal surrounding muscular envelope. Indeed, fractures of the tibia accounted for approximately 60% of all appendicular nonunions in cats in one study [13]. Toy and miniature breeds of dogs have decreased vascular density at the distal diaphyseal-metaphyseal junction of the radius as compared with large breeds of dogs [14].

Fracture zone instability is another common contributing cause of delayed or nonunion in small animal patients. Strain is a measure of fracture zone motion. Osteoblastic cells have low strain tolerance, consequently direct bone union cannot progress in the presence of excessive fracture zone motion [15]. Callus and fibrocartilage are more strain tolerant and, logically, formation of these tissues is an intermediate step of callus bone healing. However, if formation of these tissues does not effectively reduce fracture gap strain down to tolerable levels, osteoblastic activity cannot occur and a nonunion develops. Indeed, persistence of fibrocartilage in the fracture zone produces a typical histologic appearance of a nonunion [16]. It is important for the surgeon to understand the forces acting on a fracture and the ability of the selected fixation to constrain those forces. A relatively common nonunion scenario arises when a single large Steinmann pin is used to stabilize a transverse, slightly interdigitating, femoral diaphyseal fracture [17]. What appears to be adequate rotational stability intraoperatively is soon discovered to be inadequate as the patient’s recovery continues. The torsional moments induced during weight-bearing and muscular contraction exceed the torsional stability provided by the interdigitation of the fracture ends and the negligible frictional interaction of the smooth IM pin surface against the endosteal bone surface. The highly vascularized femoral diaphysis attempts to heal by developing exuberant periosteal callus, but is incapable of reducing strain levels down to a level conducive to osteoblastic activity. Predictably, a hypertrophic viable nonunion develops.

Conversely, fracture zone hyperstability may also contribute to nonunion [2,3]. Although the exact mechanical environment required for optimal bone healing is not known, axial micromotion appears to be beneficial in stimulating bone healing [18,19]. The desired mechanical environment is apparently dynamic, with rigid stability desired early in healing followed by controlled axial micromotion in later stages. It may well be that hyper-rigid fixations may delay or halt healing in these later stages. Indeed, one study suggested that the use of oversized, rigid type II external skeletal fixation contributed to the development of nonunion in cats [13].

Large fracture gaps, which may result from inadequate fracture reduction or bone loss owing to the original trauma, excision at surgery, or removal of a sequestrum, may contribute to the development of nonunion. Recent emphasis in the treatment of highly comminuted fractures has been on the use of biologic strategies in order to maintain the viability of all bone fragments within the fracture zone [11,20]. These approaches certainly do not advocate surgical removal of vascularized bone fragments, and the development of sequestra appears rare when such strategies are employed in dogs and cats. However, one must be sure that, in the shift of emphasis away from anatomic reduction of these comminuted fractures, large fracture gaps do not remain. Surrounding soft tissues may become interposed within the fracture zone and impede callus formation. There is a species-specific critical-size bone defect greater than which bone healing is not likely. Key’s hypothesis states that a long-bone defect greater than 1.5 times the diaphyseal diameter exceeds the regenerative capacity of bone in skeletally mature dogs, thereby causing nonunion [21,22]. Indeed, a 21-mm long defect in the femur stabilized with a bone plate induced atrophic nonunion in dogs.23 Key’s hypothesis appears to overestimate the regenerative capacity of bone in the adult cat.22 When large fracture gaps are unavoidable, appropriate autogenous or allogeneic bone grafting is indicated to prevent soft-tissue interposition and provide some biologic stimulus for healing within the gap [8,11,24-29]. Alternatively, an osteotomy can be performed adjacent to the fracture gap, and bone transport combined with callus distraction used to fill the large fracture gap [30-33].

Infection within the regional soft tissues or the bone itself is a relatively uncommon cause of delayed bone healing in dogs.34 Nonetheless, appropriate aseptic surgical technique must be employed to minimize the incidence of surgical infection. Particular care should be given to the treatment of open fractures as they have a greater incidence of delayed and nonunion [35]. When adequately stabilized, bone can heal in the presence of infection, but healing may be delayed by bone lysis and/or implant loosening [2,36]. Neovascularization of the fracture zone, critical for bone healing, may be impeded by the presence of infection.

In addition to the basic clinical prerequisites for predictable bone healing (fracture zone stability and viability), a complex cascade of events must take place. Local and systemic mediators stimulate the differentiation of precursor mesenchymal cells into new fibroblasts, chondroblasts, osteoclasts, and osteoblasts needed for neovascularization, soft callus, and bone formation [1,4]. One study found fewer committed mesenchymal stem cells in the early stages of experimental nonunions and suggested that the differentiation of mesenchymal stem cells is inhibited in states of chronic nonunion [37].

Although alterations in the local environment of the fracture zone are more frequently recognized contributors, systemic factors may also delay fracture healing. Hyperparathyroidism, whether primary or secondary to renal or nutritional causes, delays bone healing as a result of calcium and phosphorus imbalance. Hyperadrenocorticism, whether primary or iatrogenic, may alter calcium absorption and deposition enough to delay callus formation. Bone healing may be delayed by administration of corticosteroids, some antineoplastic agents, and possibly some fluoroquinolones and NSAIDs [-44]. Elderly patients exhibit delayed bone healing as compared with younger, especially skeletally immature, patients.

Diagnosis

Patient signalment and fracture location influence the likelihood of a delayed or nonunion. A study of 2825 fractures in dogs showed a 3.4% incidence of nonunion. The incidence of nonunion was highest in the radius/ulna (40.6%) and femur (38.5%), whereas the humerus (12.5%) and tibia (4.2%) developed nonunion less commonly [34]. A study of 422 fractures in cats showed a 4.3% incidence of nonunions [13]. This fact highlights that, contrary to popular opinion, cats are every bit as prone to major complications of fracture healing as are dogs. Nonunion in cats was most common in fractures of the tibia/fibula (61.1%), radius/ulna (16.7%), and proximal ulna with intact radius (16.7%). Older cats, heavier cats, cats with open fractures, cats with comminuted fractures and cats with fractures stabilized with type II external skeletal fixation were more likely to develop nonunion [13].

Although radiography is the principal assessment tool for bone healing, one should not lose sight of the importance of physical examination and history taking when making patient assessments and therapeutic decisions. Instances often arise where the radiographic progression of healing is slow, but patient function and comfort are optimal. In such instances, surgical intervention may not be warranted, particularly if fracture healing is progressive, albeit slower than anticipated. Conversely, some patients exhibit lameness, muscle atrophy, pain on fracture-site palpation, stiffness, and even gross fracture instability. These findings are usually associated with some loss of the original fixation through implant migration, implant breakage and/or ensuing fracture instability. Surgical site infection is usually associated with draining tracts originating from the fracture zone, swelling and pain on deep palpation of the fracture zone, or limb manipulation.

Radiographs made at 4- to 8-week intervals are commonly used to assess fracture healing, implant stability, and maintenance of bony apposition and alignment. Making a definitive diagnosis of delayed union is difficult. In delayed union, radiographs indicate progressive healing, but at a slower rate than initially expected. Radiographs may show early bone resorption of the fracture ends, periosteal bone reaction, endosteal and periosteal callus formation, and finally, bridging callus. The radiographic diagnosis of nonunion is based on the lack of progressive fracture healing during a period of 3 or more months and the persistence of a fracture gap. The characteristic radiographic appearance of a nonunion is varied as previously described for each classification. Viable nonunions have variable amounts of callus, but no callus bridging of the fracture zone. Nonviable nonunions lack callus formation. The distinction between viable and nonviable nonunion is critical as this directs the therapy. The distinction between an oligotrophic, viable nonunion and a nonviable nonunion can be particularly challenging. Alternative imaging modalities such as ultrasonography, nuclear scintigraphy, dual x-ray absorptiometry, computed tomography, and magnetic resonance imaging have each been used to assess fracture healing and may have roles in distinguishing viable from nonviable nonunions [5,45-51]. However, as with traditional radiography, each imaging modality has its limitations. Recent investigations have evaluated the predictive value of serum biomarkers in making an early diagnosis of nonunion [52]. In the absence of definitive distinction between viable and nonviable nonunions, clinicians should assume nonviability as this will direct a more aggressive and comprehensive course of action.

Treatment

Identification and treatment of all factors contributing to nonunion are the keys to effective management. Treatment of nonunion typically consists of removal of all loose implants, anaerobic and aerobic bacterial culture (of the fracture zone and/or removed implants/sequestra), and application of a fixation capable of providing long-term stability to the fracture zone. Sequestra, if present, should be removed. Nonviable nonunions require additional specific interventions to stimulate the biologic activity of the fracture zone. These traditionally include reestablishing the intramedullary canal of the main bone fragments, decortication/shingling of the regional main bone segments, and autogenous cancellous bone grafting [2-4]. One study described successful treatment of nonviable nonunions by limited en bloc ostectomy of the nonunion site and compression plate fixation [53]. In this report, autogenous cancellous bone grafting was only used if significant bony defect remained after ostectomy. Other modifications of the traditional treatment strategies for nonviable nonunions include use of demineralized bone matrix allografts, autogeneic and allogeneic mesenchymal stem cells, human recombinant bone morphogenetic protein – 2 (rhBMP-2), other growth factors, autologous bone marrow injections, pulsed electromagnetic fields, low-intensity pulsed ultrasound, extracorporeal shock-wave therapy, and hyperbaric oxygen administration [24,25,54-71].

Malunion

A malunion fracture is a fracture that has healed in a nonanatomic position. Some degree of malunion is present in most instances of closed-fracture stabilization. Malunion may also be present when biologic strategies for spatial alignment are employed for treatment of highly comminuted fractures. Malunion can occur following anatomic reduction if implants fail prior to complete bone union. Mild degrees of malunion often are tolerated with no apparent short-term or long-term patient morbidity.

Terminology

Malunions are described by their displacement (Fig. 91-3). Overriding malunions occur when the major bone segments have slid past one another, but normal axial and rotational alignment has been maintained. Minor overriding is usually well tolerated, especially in the pelvic limb, and surgical correction may not be warranted. One study showed that dogs can adapt to 23.5% loss of femoral length without inducing significant patient disability [72]. Angular malunions occur when deviation from normal axial alignment is significant. Angular malunion may be in the mediolateral plane (varus or valgus deviation) or the craniocaudal plane (recurvatum or antecurvatum deviation) or any plane in between. Minor recurvatum or antecurvatum is relatively well tolerated, but minor varus or valgus deviation can cause significant locomotor disability. Femoral varus malunion may cause medial patellar luxation and valgus malunion may cause lateral patellar luxation. Rotational malunions occur when torsional malalignment is present. Minor torsional malunion is often well tolerated, but excessive torsion can alter a patient’s gait and place abnormal stresses on adjacent joints. Excessive femoral torsional malunion may cause patellar luxation or contribute to coxofemoral osteoarthritis.

Etiology

Gravitational, ground reaction forces and muscular pull on the bone segments determine the deviation that occurs with malunion formation. Collapse of the pelvic canal is relatively common when ilial body fractures are treated non-surgically or with insufficient internal fixation [73-75]. External torsional deformity of the proximal femur is common when the pull of the external hip rotator muscles (iliopsoas, gemelli, internal obturator, and quadratus femoris muscles) is not neutralized by intramedullary pin fixation or the rotational displacement is not recognized at the time of fracture reduction and rigid stabilization [17]. After fracture healing, patients tend to hold the hip in a normal position such that the distal limb is internally rotated. Valgus and external torsional deviation of the distal limb is common when applying closed fixation to fractures of the antebrachium and crus [76,77]. Hanging limb positioning has been advocated for surgical fixation of diaphyseal fractures of the radius/ulna and tibia/fibula. Using a large pointed reduction forcep anchored to the malleoli or talus seems to reduce the tendency for cranial angulation of the distal segment in tibial diaphyseal fractures.

Figure 91-3. Malunions are described by their displacement. (A) Over-riding malunion; (B) Angular malunion; (C) Rotational malunion.

Diagnosis and Treatment Plannning

Careful gait observation, physical examination, and radiographic assessment are needed prior to surgical corrective osteotomy procedures. Mild alterations in limb position during radiography can induce artifactual limb alignment measurements on radiographs because radiographs reduce a three-dimensional structure to a two-dimensional image [78,79]. Placement of the affected bone parallel to and as close to the radiographic cassette as possible should reduce radiographic artifact [78]. Care should also be taken to place the limb in neutral rotation and include the joints proximal and distal to the affected bone. It is advisable to perform repeated radiographs with the limb held in the proper position to be certain that malalignment measurements are repeatable. It is essential that the surgeon think three-dimensionally when studying the radiographs. Alternatively, computed tomography including three-dimensional reconstructions can be used [79-82]. In fact, three-dimensional models can be made commercially from these CT studies [83].

Treatment

Surgery is indicated when the patient does not have acceptable function. The goals of surgical correction of malunion are to improve limb function by restoring more normal alignment and length, thereby correcting abnormal stresses in adjacent joints. If only minor improvements in alignment, limb length, and most importantly, limb function are anticipated, then the morbidity, risks, and costs of the procedure may preclude surgical treatment. When surgical treatment for malunion is indicated, a corrective osteotomy is usually performed at the point of maximal deformity. Ideally, the osteotomy would allow maximal bony contact at the osteotomy and preserve or restore normal limb length and limb alignment. Different forms of osteotomies can be performed, each with their own advantages, disadvantages, and indications. Commonly performed osteotomies include transverse, closing wedge, opening wedge, oblique, step, and dome osteotomies (Fig. 91-4).

Figure 91-4 A-B. Corrective osteotomies. (A) Transverse; (B) Closing wedge.

Figure 91-4 C-D. Corrective osteotomies. (C) Opening wedge; (D) Oblique.

Figure 91-4 E-F. Corrective osteotomies. (E) Step; (F) Dome.

Transverse osteotomies are indicated for correction of simple torsional malunions. Closing and opening wedge osteotomies are used for correction of angular malunions. Closing wedge osteotomies maximize bony contact, but may result in some loss of limb length. Opening wedge osteotomies sacrifice bony contact at the osteotomy in favor of increasing limb length. Sliding of oblique osteotomies can be used to correct translational deformity and to achieve some lengthening (or shortening) of the limb. Step osteotomies are relatively complex and are, on rare occasions, used to obtain acute limb lengthening. Dome osteotomies are used to correct angular deformity while maximizing bony contact and preserving limb length. Dome osteotomies have been described for correction of malunions of the femur, radius and ulna [84,85]. Conceptually, a three-dimensional dome osteotomy would function similar to a ball-in-socket joint and would easily allow for three-dimensional corrections in limb alignment prior to application of the fixation. In reality, most dome osteotomies are made in two dimensions such that three-dimensional corrections sacrifice some bony contact at the osteotomy.

Osteotomies can be used for acute limb alignment correction with static fixation or for slow progressive limb alignment correction using dynamic fixation methods. Static fixation is most often obtained with external fixators, bone plates, or interlocking nails. External fixators offer several advantages for corrective osteotomies: (1) placement of the fixation pins parallel to the adjacent joints prior to osteotomy helps the surgeon visualize when the joints are properly aligned; (2) complex contouring of a bone plate is not required; (3) fixation pins can be placed into a relatively small bone segment [76,86,87]. Locking screw/plate designs may simplify the use of bone plate stabilization of corrective osteotomies because the interlock between the screw head and the plate obviates the need for precise plate contouring.88 Bone plates and interlocking nails can be of limited use when the osteotomy leaves a relatively small bone segment. On occasion, cross-pin fixation can be used in young dogs in osteotomies performed in the metaphyseal region. Rarely, coaptation has been used with dome osteotomies [84].

If significant limb lengthening and angular correction are required, progressive callus distraction (callotasis) is indicated [89-91]. Often correction of angular malalignment can be performed acutely, but restoration of large discrepancies in limb length must be performed gradually. Callus distraction involves stabilizing an osteotomy with a specially designed external fixator that utilizes linear motors to induce distraction. Ring fixators with linear motors are commonly employed for such osteotomy distractions and provide the surgeon with several key advantages: (1) the rings placed around the limb allow the surgeon to correct complex three-dimensional malalignment with relative ease; (2) corrections to alignment can be made after the fixation is applied; and (3) the use of small-diameter, tensioned fixation wires allows the system to be applied to small bone segments. After allowing a lag period for initial callus formation, the osteotomy is gradually distracted at a rate of approximately 1 mm per day. Once the desired limb length is achieved, daily distractions cease and the static fixation (1 month per cm distraction) allows regenerate bone in the osteotomy gap to consolidate.