➢ Segmental long-bone defects are rare injuries, making study of their treatment and outcomes difficult.
➢ The true critical-sized defect (defined as a defect that will not heal during the lifetime of the animal or patient) is unknown, and translation of animal models and studies to clinical practice remains challenging because of the limitations associated with each model.
➢ There is a lack of optimally designed clinical trials evaluating different aspects of segmental bone defects after traumatic injury.
➢ Free vascularized fibular grafting procedures are technically demanding and are associated with a lower union rate when compared with other treatment alternatives.
➢ The induced membrane technique takes advantage of the creation of a vascularized envelope with osteogenic factors that can improve union rates when staged bone-grafting is performed.
➢ Bone transportation has proved to be a reliable option for the treatment of segmental defects; however, patient comfort, the difficult initial learning process associated with the technique, and the demands of postoperative frame management have likely contributed to its limited use.
Long-bone defects are one of the most challenging problems associated with limb reconstruction following high-energy extremity trauma. Without reconstitution of structural integrity, amputation is the likely result. The current orthopaedic literature on segmental bone loss is limited to case reports, case series, and observational studies, making it difficult to establish sound conclusions on the basis of the levels of evidence provided. Ten-year data from a fracture registry indicated that 0.4% of all fractures that were treated at a Level-I orthopaedic trauma unit were associated with bone loss1. Data from the same registry also indicated that the most common site of bone loss was the tibia (68%), followed by the femur (22%), with the remaining fractures occurring evenly at different sites. An additional retrospective review of tibial bone defects indicated that 3.6% (twenty-six) of 725 fractures had >3 cm of bone loss, with the distal part of the tibia having a higher proportion of bone loss (6%; seven of 123) in comparison with the shaft (4%; fourteen of 351) and the proximal part of the tibia (2%; five of 251)2.
Segmental bone defects represent a difficult clinical challenge in the reconstruction of injured limbs. The earliest treatments for these problems included amputation or shortening, whereas more recent treatments have included bone lengthening, bone-grafting, and distraction osteogenesis3-10. In addition, previous authors have reported spontaneous healing of femoral defects. Hinsche et al., in a case series of four patients, reported that segmental femoral bone defects measuring 6 to 15 cm filled spontaneously in an average of eight months11. The same group also reported the case of a patient with a 14-cm diaphyseal tibial defect that spontaneously healed12. However, such outcomes are rare, particularly for diaphyseal tibial segmental defects larger than 2 cm1.
Combat operations in the last decade have served to emphasize the critical gaps in the treatment of segmental long-bone defects. Although precise data on the severity of segmental bone loss are not available, surgeons from military centers have described a variety of techniques for treating such injuries, including the use of combinations of internal or external fixation, autograft bone, recombinant proteins, and synthetic substitutes13,14. Currently, there is little evidence to guide clinicians in the optimum treatment techniques that expedite healing and avoid complications in patients with bone loss.
There remains no well-accepted definition for the so-called critical-sized bone defect. Critical-sized bone defects are those that will not heal during the lifetime of the animal or patient15. Preclinical animal models for bone defects have had several limitations, including the lack of soft-tissue damage seen in association with clinically occurring fractures or defects, variability within animal species, and the lack of comorbidities (such as obesity, diabetes, or osteoporosis) that may negatively influence fracture-healing16. These factors limit the development of an adequate preclinical segmental bone defect model and its generalizability to the human species.
Currently, there are no validated treatment algorithms for segmental defect nonunions. For defects up to 10 cm in length, satisfactory results may be achieved with cancellous autografts. For larger deficits, the use of bone transport17, vascularized autograft10, or nonvascularized autograft18 has been moderately successful. Attempts to augment healing with use of novel agents such as bone morphogenetic proteins (BMPs) and other bone-graft substitutes have been reported19. However, despite these advances in knowledge, technique, and materials, the treatment of large segmental bone defects remains a problem. The current article provides an overview of the current treatment alternatives for segmental bone defects. In addition, it attempts to identify current gaps in our knowledge regarding the treatment of these injuries as well as promising areas of research in the foreseeable future.
Use of Levels of Evidence in the Assessment of Scientific Information
The Levels of Evidence guideline for this journal was established in January 200320. This guideline provides levels of evidence that are based on a hierarchical rating system for the classification of study quality. A five-level rating system was developed for four different study types: therapeutic, prognostic, diagnostic, and economic or decision-modeling. Recommendations for or against different aspects of treatment of segmental bone defects were graded with the use of these guidelines. The reasoning pertaining to each of the treatment recommendations is explained in each section below. A summarized review of this information and details on grades of recommendation are shown in Table IV.
Free Fibular Grafts
Various configurations for the free vascularized fibular graft have been described. This variability allows for defect filling to be planned on a case-by-case basis and involves the performance of various osteotomies. The naturally straight configuration is the one that is most frequently used as it takes advantage of dual vascularity (endosteal and periosteal). The endosteal blood supply is composed of the nutrient artery, which enters the middle third of the diaphysis via the nutrient foramen and then divides into an ascending branch and a descending branch. Eight to nine periosteal branches compose the periosteal blood supply, mostly in the middle third of the diaphysis21.
With the advancement of medical devices and improvements in microsurgical techniques, the free vascularized fibular graft has become an additional alternative for the osseous reconstruction of segmental long-bone defects measuring ≥6 cm. The procedure, which consists of graft harvesting, recipient site preparation, and final defect reconstruction, has been routinely used in both the upper and lower extremities22. Several reports have demonstrated its utility for the treatment of tibial defects since it was popularized by Taylor et al.23-27. Other vascularized sites have been used for harvest, including iliac crest and rib, but these have not been shown to provide the strength, decrease in donor-site morbidity, or morphologic characteristics of a fibular graft21,28.
The advantages of this method of reconstruction include a shorter duration of external fixation when compared with bone transportation as well as simultaneous soft-tissue coverage. An additional advantage is the tendency for hypertrophy to occur in response to microscopic stress fractures29. Conversely, patients with increased hypertrophy have shown a decreased prevalence of stress fractures30. El-Gammal et al. reported a significant positive correlation between the time to graft union (the start of partial weight-bearing) and the time to significant graft hypertrophy (defined as >30% of original fibular width) (r = 0.9, p = 0.01)31. Younger patients demonstrated a higher percentage of hypertrophy compared with older individuals, with the amount of hypertrophy being dependent on the mechanical loading of the graft.
The disadvantages of vascularized free fibular grafts are the occurrence of a combination of recipient and donor-site complications, including infection, stress fractures, blood loss, venous thromboembolism, the requirement of microsurgery, and amputation32,33. Lin et al., in a retrospective study, evaluated the outcomes for ninety patients undergoing ninety-seven composite vascularized bone transfers for the treatment of lower extremity osseous defects28. Sixty-four of these procedures were performed by means of fibular osteocutaneous flaps for the treatment of defects averaging 11.5 cm in length. Only three fibular transplantation procedures failed, allowing for a flap survival rate of 95.3% (sixty-one of sixty-four). Complications in the sixty-one patients with surviving flaps included skin paddle partial loss (11.5%), post-transplantation osteomyelitis (13.1%), stress fracture (13.1%), nonunion (3.3%), and malunion (4.9%), with an overall rate of donor-site morbidity of 21.3%. Despite the considerable morbidity associated with free vascularized fibular grafts, the rate of return to weight-bearing and adequate functionality remained high (96.7%). Table I shows the times to radiographic union, union rates, and reported complications for several studies in which this technique was used. The grade of recommendation for the use of free fibular grafts in the treatment of segmental bone defects is B (Table IV).
Induced Membrane (Masquelet) Technique
Masquelet et al. reported on a case series of thirty-one two-stage reconstruction procedures that were performed for the treatment of large diaphyseal defects ranging from 5 to 25 cm after debridement34. The first stage consisted of radical debridement, soft-tissue repair with flaps when needed, and insertion of an acrylic (polymethylmethacrylate) cement spacer and placement of an external fixator. The second stage, in which the cement was removed but the membrane covering the spacer was left in place, was performed six to eight weeks later. The defect was then filled with cancellous morselized autograft, and demineralized bone was added in cases in which the defect was too large for the amount of autograft available. Immediate complications included three cases of free flap failure that were successfully reconstructed as well as two early stress fractures that healed with immobilization. The average time to full weight-bearing was 8.5 months. The concept of an induced membrane was developed during the evaluation of the first patients in this series. The authors did not debride this membrane because of its excessive bleeding and described the mechanical and biological roles of the cement spacer in the treatment of segmental long bone defects.
This technique has been increasingly adopted by orthopaedic trauma surgeons and has shown consistently successful results when used for the treatment of large segmental long bone defects (Table II). There has been notable variability in terms of the type of bone graft used for the induced membranes (Masquelet) technique. The studies shown in Table II involved the use of different bone-graft materials for reconstruction of the segmental defect, including autologous bone from the anterior and posterior iliac crest, fibula, or rib; bone obtained following the use of a Reamer-Irrigator-Aspirator system (RIA; DePuy Synthes, West Chester, Pennsylvania); and allograft. In addition, demineralized bone matrix, recombinant human BMP-2 (rhBMP-2), calcium sulfate (CaSO4), and/or platelet-rich concentrate were used by several authors and in different arrangements. The current literature on this topic is underpowered because of the low number of patients per series as well as the variability in treatment. Additional research is needed to determine whether there are significant differences in union rates when controlling for graft materials used in conjunction with this technique.
The French Society of Orthopaedic Surgery and Traumatology has performed the largest retrospective study to date on the treatment of posttraumatic bone defects with the induced membrane technique35. A total of eighty-four defects were included, of which 57% were ≥5 cm and 70% involved the lower limb. The investigators identified a 50% infection rate, a 90% union rate, an average of 6.1 interventions (involving a combination of irrigation and debridement, flap coverage, and osseous reconstruction procedures) until successful treatment, and a 9.5% failure rate. This study reinforces the high union rate, in spite of frequent complications requiring repeat operations, of the induced membrane technique in the management of segmental long bone defects.
Apard et al. evaluated the use of this two-stage reconstruction technique in a study of twelve patients with tibial segmental bone loss that was treated with nailing36. The authors reported bone union in all patients but one. Four patients developed a deep infection: two were managed with nail change followed by antibiotic therapy, and the other two were managed medically. The average time to weight-bearing was four months. The authors concluded that the use of a nail does not increase the infection rate and may allow for faster weight-bearing as well as eliminate all pin-track infections and the discomfort associated with a long-term external fixator. They attributed the late infections to the use of antibiotic-impregnated cement spacers, which may have masked insufficient debridement. In other words, the use of the antibiotic spacer does not compensate for an ill-performed debridement. Donegan et al. also reported success in association with the use of this technique and pointed out that, as this procedure becomes more widely applied, standardization of treatment of segmental bone defects may be possible37.
Experiments have demonstrated that the main property of induced membranes is to regenerate, contain, and avoid resorption of the cancellous bone38,39. This property is believed to be due to the high concentrations of vascular endothelial growth factor (VEGF), transforming growth factor beta (TGF-β), and BMP-2 that are found in the membrane38-40. Aho et al. provided additional histologic characterization of this membrane in a study in which vascularization was found to be greatest in one-month-old samples, with a >60% reduction in vascularization two months later41. The highest concentrations of VEGF, interleukin-6 (IL-6), and type-I collagen (Col-1) were noted at four weeks, with the levels at eight weeks being substantially (>40%) lower. In addition, the investigators showed that the induced membrane has the capability to differentiate toward calcified tissue, allowing for the formation of bone by means of endochondral ossification, and also noted the presence of mature lamellar bone in the induced membrane. These findings may redirect the timing of bone-graft procedures from the more common period of six to ten weeks to a sooner period of four to six weeks, possibly reducing time to weight-bearing and overall recovery. The grade of recommendation for the use of the Masquelet (induced membrane) technique for the treatment of segmental bone defects is B (Table IV).
The first reports of bone transportation date back to the 1950s with the development of the Ilizarov method (named after Gavriil A. Ilizarov) for the treatment of long tubular bone defects42. Briefly, the Ilizarov method involves the performance of a cortical osteotomy through healthy bone at some distance from the defect site. The segmental fragment is pulled through the limb with wires that are attached to an external ring fixator until the defect is closed by newly formed bone, often referred to as the “regenerate.”43 Transportation usually begins after a latency period of five to ten days following the corticotomy44,45. The rate of transport is typically 1 mm/day and is usually achieved by making four 0.25-mm adjustments of the device at evenly spaced time intervals throughout the day. This distraction rate minimizes the risk of premature consolidation while not exceeding the rate of vascular ingrowth, thus promoting bone growth. After the distraction is complete, the frame must remain in place to allow consolidation of the regenerate for a period of time that historically has been described as two to three times the length of time required for distraction. Several authors have used an external fixator index to describe the time required in the frame for each centimeter of bone defect that was regenerated. The external fixator index for bone transportation has been reported to be approximately 1.4 to 2.1 months per centimeter45-47. These data are useful when counseling patients with regard to the length of time that they are likely to wear the frame.
The five etiologies of segmental bone defects in the first reported case series of patients managed with this technique included congenital pseudarthrosis (n = 7), closed fractures with nonunion (n = 2), pseudarthrosis following femoral lengthening (n = 1), open comminuted fractures without infection (n = 4), and open comminuted fractures complicated by osteomyelitis (n = 7)48. The defect length (defined as the distance between the two fragments added to the length of anatomic shortening) ranged from 4 to 19 cm. Correction of the anatomic shortening as well as elimination of the bone defect was seen in 90.5% (nineteen) of twenty-one cases. The main benefit provided by this technique is that a segmental bone defect of any size can be reconstructed while maintaining the original limb length. In addition, donor-site morbidity is minimized as bone-grafting may not be needed, although it is sometimes used to achieve union of the docking site during bone transport. A common problem with this technique involves nonunion of the docking site, which is the site where the transported bone end meets the intact, stationary bone end. This problem often occurs as a result of the length of time required for distraction, with fibrocartilage or soft tissues preventing cortical apposition. Union of the docking site can be accomplished with various methods, including the application of gradual compression to the fixator at a rate of 0.25 mm/day or bone-grafting of the docking site with or without compression plating, among others43,44,47,49.
The difficulties associated with deformities following treatment in the frame and the difficult initial learning process associated with the application of a ringed fixator for bone transport and deformity correction led to the development of the Taylor spatial frame (TSF; Smith & Nephew Richards, Memphis, Tennessee). This device allows for correction in multiple planes, oblique angulation, translation, and rotation, by adjusting the strut lengths of a single frame50. In at least one study, lower complication rates and generally good results were reported when this method was compared with the traditional Ilizarov method51.
The clinical results associated with bone transportation have been variable in the literature. However, most series have shown high union rates. Chaddha et al. evaluated the functional outcomes associated with this method in a study of twenty-five patients with massive posttraumatic bone defects involving the femur (three patients) or tibia (twenty-two patients)52. The average bone defect length was 8.9 cm (range, 5 to 17 cm). While pin-track infections were common (prevalence, 80%) and only five patients returned to their original job, union was achieved in 92% (twenty-three) of the twenty-five patients.
In 2011, Liodakis et al. compared union rates and outcomes in a study of thirty-nine posttraumatic tibial bone transport procedures that were performed with or without the use of an intramedullary nail53. Union was achieved in nineteen (90%) of the twenty-one patients who were managed with an external ring fixator and thirteen (72%) of the eighteen patients who were managed with a monorail fixator and an intramedullary nail. There was no significant difference between the two groups in terms of any of the Short Form-36 (SF-36) health survey categories. However, a significant reduction in the duration of external fixation was noted in the intramedullary nail group as compared with the ring fixator group (mean and standard deviation, 5.9 ± 1.5 compared with 15.9 ± 6.1 months; p < 0.0001). Also, a lower rate of deformity of >5° was observed in the intramedullary nail group (6% compared with 33%; p = 0.049). Although the results of this technique are similar to those of the original Ilizarov method, there is a substantial risk of the spread of organisms throughout the intramedullary canal, with the additional risk of the conversion of a minor pin-track infection into osteomyelitis. Table III shows the times to radiographic union, union rates, and reported complications for several studies involving this technique. The grade of recommendation for the use of the bone transportation technique for the treatment of segmental bone defects is B (Table IV).
Antibiotic cement is commonly used for the treatment of segmental defects that require delayed bone-grafting. Antibiotic beads have spaces between them that were shown, in a basic-science model, to allow bacterial growth54. A single piece of antibiotic cement eliminates this concern by creating a uniform space with a surrounding membrane. A large block of cement does not elute as much local antibiotics as a large number of beads does, but it occupies space that could otherwise harbor bacteria and creates space for bone graft that will be contained by the induced membrane. Basic-science data have shown that antibiotic spacers decrease bacterial colonization and create an environment that is optimal for reconstruction with the use of bone graft54. Orthopaedic trauma surgeons commonly use antibiotic cement for the treatment of traumatic segmental defects, with 65.4% (248) of 379 surgeons reporting its use prior to bone-grafting55. Current practice, basic-science data, and clinical data56 allow for a grade-B recommendation regarding the use of antibiotic cement spacers for the treatment of segmental defects (Table IV).
Christian et al. performed reconstruction, without free fibular transfer, for the treatment of eight large diaphyseal defects5. The authors used antibiotic-impregnated beads followed by large amounts of autologous cancellous bone graft once the soft-tissue flap had healed. One patient had an infection, which resolved after debridement and antibiotic therapy. Ristiniemi et al. used a similar technique for the treatment of twenty-three fractures with bone loss in the proximal, diaphyseal, and distal aspects of the tibia2. Antibiotic-impregnated beads were implanted first, followed by autologous bone-grafting approximately eight weeks later. The average healing time for all tibiae was forty weeks. The distal part of the tibia showed an average healing time of thirty weeks, which was seven weeks shorter than that for the proximal part of the tibia and sixteen weeks shorter than for the tibial shaft. This time difference could be attributed to the larger defects seen in the proximal part of the tibia (62 mm) and the diaphyseal aspect of the tibia (54 mm) as compared with the distal part of the tibia (43 mm). The authors concluded that although the method was effective for all parts of the tibia, longer or shorter healing times may be expected, depending on the location of the defect.
Iliac crest bone graft is the current so-called gold-standard material for the treatment of bone defects of <5 cm57. Previous authors have reported high rates of acute and chronic severe pain at the harvest site, ranging from 16.5% to 18.7% after one year58,59. However, a recent prospective observational study evaluating the prevalence of pain and complications associated with the use of iliac crest bone graft showed that the rate of infection was 3% and the rate of donor-site pain was 2% at twelve months postoperatively. The authors concluded that more-than-moderate persistent pain at the iliac crest donor site was rare and that the use of iliac crest bone graft did not impair functional health and well-being compared with alternative methods for the treatment of nonunion of long-bone fractures57. Other commonly used autogenous bone-graft harvest sites include the proximal part of the tibia and the distal part of the femur, but the amount of cancellous bone available at these sites is less than what can be harvested from either the anterior or posterior iliac crest.
The RIA system is an alternative method for harvesting cancellous autograft from the entire length of a long bone. The RIA system was developed as a way to decrease the release of fat emboli into the systemic bloodstream at the time of the reaming process during intramedullary nailing of femoral fractures60-62. However, the large volumes of graft material obtained (30 to 90 cm3, the high percentage of stem cells, growth factors, and good handling characteristics have made the RIA system a viable option for bone graft harvest at the time of treatment of nonunion and segmental long-bone defects63-68. Stafford and Norris evaluated the use of this system for the treatment of twenty-seven segmental defect nonunions involving the tibia (nineteen) or femur (eight), with radiographic union as the primary end point63. They established a standard treatment protocol consisting of aggressive soft-tissue and osseous debridement with subsequent use of antibiotic spacers and early wound closure. Six to eight weeks later, they harvested bone from the ipsilateral femur and, if additional grafting was required, from the contralateral femur. The average total amount of bone extracted was 47 cm3. These results were biased by the use of allograft and BMP in 85% of the cases. The authors reported a 90% union rate at one year following a single bone-graft procedure and reported no donor-site complications. Among the remaining 10% of patients, one underwent additional bone-grafting, one had a below-the-knee amputation, and one was lost to follow-up. Additional studies have demonstrated similar results, reiterating the heightened biological activity of intramedullary bone and its usefulness for the treatment of segmental long-bone defects66. However, to date, there have been no published prospective, randomized studies that have compared union rates among all bone graft materials.
Autograft has a finite volume and is associated with a real risk (albeit low) of secondary infection or pain at the harvest site57. An ideal alternative to autograft has not been developed. The segmental defect model is similar to a nonunion model. In a retrospective series of long-bone nonunions, BMP-2 mixed with allograft demonstrated equal union and complication rates when compared with autologous iliac crest bone graft69. Patients in the autograft group had a one-hour-longer operative time, a one-day-longer hospital stay, and a 223-mL increase in the average blood loss. A multicenter randomized trial comparing the use of rhBMP-2/allograft with autologous bone graft for the treatment of tibial fractures with cortical defects showed comparable times to healing between treatment groups (six months) as well as lower mean estimated blood loss in the rhBMP-2/allograft group70. A prospective randomized trial comparing autograft with BMP-2/calcium was initiated but was aborted because of low enrollment71. Grades of recommendation in the current literature regarding filling materials for segmental bone defects are shown in Table IV.
Other bone-graft extenders and substitutes include allograft, demineralized bone matrix, calcium phosphates, and calcium sulfates, but these are not typically used independent of an osteogenic substance (i.e., autograft or BMP-2) for the treatment of segmental defects. Additionally, component variability between manufacturers as well as between lots of demineralized bone matrix limit the design of clinical trials and the generalizability of the results. Basic-science research continues to be performed to identify the ideal bone graft substitute than can be placed early after definitive fixation, allow for coverage that would induce bone growth, and have the microstructure for conduction of bone growth and vascular ingrowth while having the mechanical strength to allow early weight-bearing.
Segmental long-bone defects are rare injuries, making study of their treatment and outcomes difficult. In addition, the true critical-sized defect is unknown and translation of animal models and studies to clinical practice remains challenging because of the limitations associated with each model. Currently, there are multiple different protocols for the treatment of segmental long-bone defects1,72. There is controversy because of the lack of optimally designed clinical trials evaluating different aspects of segmental bone defects after a traumatic injury, making both the identification of injuries that require intervention and the optimal treatment methods largely unknown. Future efforts should be directed toward basic-science investigations with improved clinical translation as well as toward large multicenter trials that can take advantage of large numbers, both to better define the problem of the critical-sized defect and to clarify the success of current options for the treatment of segmental long-bone defects.
Source of Funding: No external funds were received in support of this study.
Investigation performed at the Department of Orthopedics, Orthopedic Trauma Institute, Vanderbilt University Medical Center, Nashville, Tennessee
Disclosure: None of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of any aspect of this work. One or more of the authors, or his or her institution, has had a financial relationship, in the thirty-six months prior to submission of this work, with an entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. No author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.
- Copyright © 2014 by The Journal of Bone and Joint Surgery, Incorporated