Indirect fracture healing regenerates bone through a cartilage callus scaffold (Fig. 2.5) [13]. It still requires a relatively stable environment, but it does not require rigid stability or anatomical reduction. Rather, micromotion, to an extent, stimulates the healing response. Indirect healing is the predominant mechanism in most fractures treated by nonoperative means. It is also achieved by interventions that allow for relative stability. These include intramedullary nailing of long bone fractures (Fig. 2.6), external fixation (Fig. 2.7), bridge plating (Fig. 2.8), and splinting, bracing, or casting.

Three fundamental phases of indirect healing have been described [21]: inflammatory, reparative, and remodeling. Trauma initiates the acute inflammatory phase, and, through the release of mediators, cytokines, and growth factors, recruits progenitor cells responsible for initiating repair. In the reparative phase, progenitor cells lay down cartilaginous and bony callus, facilitate neoangiogenesis, and replace callus with woven bone. The remodeling phase replaces the woven bone with a mature lamellar bone structure.

Following injury, hematoma occupies the fracture site. Fracture hematoma serves two key functions. It provides a physical scaffold for subsequent occupation by progenitor cells, granulation tissue, and ultimately callus. Furthermore, the hematoma itself contains progenitor cells, cytokines, and growth factors that directly participate in the healing process [22, 23]. Recent studies have identified higher levels of factors and signaling molecules in fracture hematoma. These include macrophage colony-stimulating factor (M-CSF) , transforming growth factor-beta (TGF-β), and interleukins (IL), all of which have important roles in stimulating fracture healing (Table 2.1) [24–27].

The initial inflammatory response occurs immediately after injury and lasts several days. The response is marked by infiltration of macrophages, platelets, polymorphonuclear leukocytes, and lymphocytes into the fracture site. These secrete proinflammatory cytokines including interleukins (IL-1, IL-6), platelet-derived growth factor (PDGF) , and tumor necrosis factor-alpha (TNF-α) . These factors recruit other inflammatory cells, promote angiogenesis, recruit progenitor stem cells, and induce their differentiation.

The recruitment of MSCs is an essential component of fracture healing. MSCs reside throughout the body, including the periosteum, bone marrow , trabecular bone, muscle, and systemic circulation [28]. Periosteal- and bone marrow-derived MSCs were traditionally thought to be the primary sources of progenitor cells in early fracture repair [29]. However, current data suggests that other sources of MSCs, namely from muscle and systemic circulation, may also contribute to the progenitor cell population [28, 30].

Inflammation at the time of injury releases a number of chemokines, growth factors, and signals to recruit MSCs and other inflammatory cells. In the early phase, TNF-α, IL-1, and IL-6 play key roles in chemotaxis, mesenchymal stem cell (MSC) recruitment, and osteogenic and chondrogenic differentiation [14]. Peak levels of IL-1 and IL-6 are reached within the first 24 h, and then decline precipitously after 72 h. IL-1 and IL-6 contribute to chemotaxis of other inflammatory cells and of MSCs and promote angiogenesis via vascular endothelial growth factor (VEGF) production [31]. TNF-α and IL-6 promote recruitment and differentiation of muscle-derived stromal cells. TNA-α, at low concentrations, also stimulates chondrogenic and osteogenic differentiation [32–34] (see Table 2.1). In vivo injection of TNF-α accelerates fracture healing and callus mineralization [32]. Conversely, the absence of TNF-α signaling appears to delay both chondrogenic differentiation and endochondral resorption [14, 24, 34].

Emerging evidence has also supported the role of stromal cell-derived factor (SDF-1) in skeletal repair. SDF-1 is a potent chemoattractant expressed at sites of injury to recruit MSCs from both circulating and local sources. Kitaori demonstrated that SDF-1 expression is upregulated in periosteum at the fracture site and recruits MSCs that participated in the healing process. Additionally, blocking the function of SDF-1 significantly reduced bone formation, indicating SDF-1 has a crucial role in fracture healing [35].

By this time, the fracture hematoma has been converted to granulation tissue, containing inflammatory cytokines and growth factors that stimulate MSC differentiation, proliferation, and production of extracellular matrix. The formation of cartilaginous callus marks the initial attempts at achieving fracture union. The result is a calcified cartilaginous bridge that both provides stability and creates a template for further remodeling.

Cartilaginous callus formation is driven by growth factors, chondrocytes, fibroblasts, and mechanical stimulation across the fracture site. TGF-β and IGF-1 play primary roles in this stage of chondrogenesis and endochondral bone formation, stimulating the recruitment, proliferation, and differentiation of MSCs. BMPs also promote chondrogenesis. Several days after fracture, chondrocytes derived from MSCs proliferate and synthesize collagen. Starting from the periosteum and the fractured ends, chondrogenesis progresses by appositional replacement of adjacent granulation tissue with cartilage matrix [29]. Fibroblasts produce fibrous tissue in areas with limited cartilage production. Micromotion across the fracture stimulates callus formation, and increased callus formation provides more mechanical stability to the fracture. When sufficient callus and stability have been attained, roughly 2 weeks after fracture, chondrocytes undergo hypertrophic differentiation. Proliferation ceases. Collagen synthesis is downregulated. Hypertrophic chondrocytes release vesicular stores containing calcium, proteases, and phosphatases into the surrounding matrix. As the collagen matrix is degraded, released phosphate ions bind with calcium to promote cartilage calcification. These calcium and phosphate deposits become the nidus for hydroxyapatite crystal formation [8].

At the same time, intramembranous ossification occurs in areas of low strain, beneath the periosteum, and directly adjacent to the fractured cortices. Within 24 h following injury, MSCs from the bone marrow differentiate into osteoblastic phenotypes. Proliferation and differentiation peak at day 7–10. Woven bone is formed in these regions without a cartilage scaffold.

Fracture healing begins in a relatively hypoxic environment; injury to vessels, periosteum, and soft tissue compromises local blood supply [22]. Early cartilage callus can form in this hypoxic environment. However, as healing progresses, subsequent callus remodeling and bone formation require adequate oxygen delivery. Failure to do so leads to delayed healing. Revascularization is thus critical for progressive healing and bone formation [9, 11, 12, 36–38].

Two main molecular pathways regulate this process: an angiopoietin-dependent pathway and a VEGF-dependent pathway. Angiopoietins promote formation of larger vessels and collateral vessels off existing vessels. VEGF promotes endothelial cell differentiation, proliferation, and neoangiogenesis, and it mediates the principal vascularization pathway [11, 24].

Inflammatory cytokines from early fracture healing, particularly TNF-α, induce expression of angiopoietin, allowing for early vascular ingrowth from existing periosteal vessels [9, 33]. However, the primary vascularization process is driven by VEGF. Following calcification of cartilage callus, osteoblasts and hypertrophic chondrocytes housed in callus express high levels of VEGF, stimulating neoangiogenesis into the avascular chondral matrix [36, 38, 39]. Concurrently, matrix metalloproteinases (MMPs) degrade calcified cartilage to facilitate ingrowth of new vessels [40].

With the onset of neoangiogenesis, the next event is characterized by the transition from soft callus to hard callus: the removal of calcified cartilage and its replacement with woven bone matrix. This process is mediated by MMPs, BMPs, osteoclasts, chondroclasts, and osteoblasts [36, 40, 41].

Osteoclasts have historically been considered the key cell type in soft callus resorption. However, more recent evidence suggests that resorption is nonspecific and mediated by multiple cell lines, including osteoclasts and chondroclasts alike, and by MMP expression [40, 41]. This has been supported by findings that impaired osteoclast function does not necessarily impair healing. In an osteoclast-deficient osteopetrosis mouse model, there was no difference in callus remodeling or union rates compared with control mice [42].

Cartilage callus is removed and subsequently replaced by woven bone. Mature osteoblasts secrete osteoid, a combination of type I collagen, osteocalcin , and chondroitin sulfate. Collagen fibrils are randomly oriented, producing an irregular structure known as woven bone [41].

Remodeling is characterized by woven bone resorption followed by lamellar bone formation. Osteoclasts are multinucleated polarized cells that attach to mineralized surfaces. At sites of attachment, osteoclasts form ruffled borders, effectively increasing surface area through which lysosomal enzymes and hydrogen ions are secreted. Enzymes degrade the organic collagen components, while the acidic milieu demineralizes the bone matrix. The erosive pits left by the osteoclasts are termed “Howship’s lacuna.” Following resorption, osteoblasts form new bone within these lacunae. This process progresses along the length of hard callus, layer upon layer, replacing woven bone with lamellar bone [43, 44].

Activation and regulation of remodeling depends on intimate coupling between osteoblasts and osteoclasts. Osteoblasts initiate remodeling by producing factors to stimulate osteoclastogenesis and osteoclast function. The principle cytokines secreted by osteoblasts are M-CSF, receptor-activated NF-κβ ligand (RANKL), and osteoprotegerin (OPG) . M-CSF and RANKL are essential for osteoclast formation. Osteoblasts express RANKL on their cell membranes, whereas mononuclear osteoclast progenitors express the complementary receptor, RANK. Upon contact, RANKL interacts with RANK to induce fusion of osteoclast progenitors and thus produce mature multinucleated osteoclasts. Alternatively, osteoblasts can also secrete OPG, which acts as a decoy by binding RANK and consequently disrupts RANKL–RANK interactions. By modulating RANKL and OPG expression, osteoblasts can tightly regulate osteoclast activation. Osteoblasts express and secrete M-CSF, which induces osteoclast precursor proliferation and differentiation. Additionally, M-CSF upregulates the expression of RANK on osteoclast precursors [43–45].

Diaphyseal bone healing hinges on the interrelationship between biomechanics and biology. Early in the healing process, the mechanical environment determines the biologic response, whether healing will proceed by direct or indirect means. In stable situations, healing proceeds directly to osteogenesis. In unstable conditions, healing begins with chondrogenesis. The same holds true for metaphyseal healing. Under rigidly stable conditions, newly formed bone bridges the fracture gap with minimal chondrogenic tissue, similar to direct healing. Under more flexible conditions, bone intermixed with islands of chondrogenic tissue forms across the gap, analogous secondary healing. Interestingly, both situations do not generate a significant amount of external callus [47]. Whereas progenitor cells need to be recruited in diaphyseal healing, the metaphysis houses a large reservoir of precursor cells, obviating the need for a large periosteal reaction and MSC recruitment [48].

The relationship between mechanics and biology is well established in skeletal physiology. Wolff’s law stipulates that bone structurally adapts to its loading conditions. Likewise, biomechanics plays a central role in skeletal repair. Following injury, the mechanical environment influences the biologic healing response. This response in turn attempts to restore skeletal integrity. Understanding how biomechanical factors affect healing is therefore fundamental to fracture treatment. The existing body of literature has identified three mechanical parameters that impact fracture healing: interfragmentary strain, gap size, and hydrostatic pressure. The degree to which these parameters affect healing, and the timing at which they are applied, will be discussed in this section.

The accurate assessment of fracture union is often a difficult undertaking, but nonetheless fundamental to clinical practice and research. Nonunions can be a source of significant disability, and its early diagnosis and treatment is paramount to improving patients’ quality of life and return to function [55]. The definition of nonunion provided by the United States Food and Drug Administration (FDA) requires a minimum of at least nine months to elapse since the initial injury and no signs of healing for the final three months. Yet, there are no standardized methods of assessing fracture union, and there still remains considerable variability among clinicians and researchers alike [56, 57]. However, advances in imaging techniques, improved knowledge about the biology and biomechanics of fracture healing, and new scoring systems are refining our ability to assess fracture healing.