Review of soft tissue coverage options in distraction osteogenesis of the extremity

Aim: To review the choices of soft tissue coverage in distraction osteogenesis of the extremity. Methods: A PubMed literature search yielded 14 articles included for systematic review. Data were extracted from each article if available (sample size, patient age, surgical indications, type of flap, use of additional modalities, method of bone osteogenesis, postoperative events, follow-up, satisfaction, weight-bearing status, and success rate). Unpaired t-tests were performed to compare complication rates. A retrospective review of three cases was also conducted. Results: Fourteen articles discussed 145 patients with a mean age of 33.4 years and 146 extremity injuries followed over 3.3 years on average. Indications included chronic osteomyelitis or nonunion (58.2%) and acute trauma (41.8%). Average time from injury was 1.1 years. Ilizarov frame was used in 12 articles. Free flaps (88.0%) or rotational flaps (12.0%) were used, with muscle flaps (96.7%) being most common. Most extremities received free latissimus dorsi or rectus abdominis flaps. Bone grafts and antibiotic beads were often used in conjunction. Although complications and reoperations were not uncommon (up to 30%), 98.8% of patients on average were ultimately weight bearing and all articles reported > 91% success rate. Additionally, the rates of any complication were not statistically different between “fix and flap” protocol and flap or frame first. Lastly, a three-patient case series is presented. Conclusion: Bone transport with soft tissue reconstruction remains an excellent choice for patients with large bony defects or who are unable to undergo autologous bone grafting. Not one surgical approach to limb salvage is superior, and decision should be made on a case by case basis between the surgeon and the patient. Page 2 of 18 Stoneburner et al. Plast Aesthet Res 2020;7:13 I http://dx.doi.org/10.20517/2347-9264.2019.028


INTRODUCTION
High-energy lower extremity trauma with massive bone and soft tissue loss poses a challenge for orthopedic and reconstructive surgeons. Higher Gustilo-Anderson fracture type, Orthopaedic Trauma Association Open Fracture classification, and limb salvage index score are associated with a poor prognosis, including high risk of infection, nonunion, and amputation [1][2][3] . With improvements in surgical techniques and protocol-driven traumatic lower extremity management at tertiary care centers, limb salvage rates have significantly increased, reaching up to 90% in some reports [4][5][6] .
Early aggressive radical debridement, skeletal stabilization and soft tissue coverage is the critical first step in lower extremity salvage [7] . Depending on the extent of trauma and the size and nature of the defect, there are many reconstructive options available to the orthopedic and plastic surgeon. When the bony injury is "critical sized" -that is, too large to heal without additional intervention -additional steps must be taken to achieve bony union.
The choice of reconstructive technique is often determined by size. The induced membrane technique is a viable option for small and large defects, although results may differ based on anatomic location. Larger defects may require vascularized transfer from the fibula or iliac crest [7][8][9] . Some patients may be poor candidates for vascularized bone graft, the defect may exceed the size of the available graft or there may be concerns about size mismatch between vascularized bone and the defect location. In these cases, bone transport using distraction osteogenesis can be utilized. Bone transport in combination with free soft tissue transfer is an effective treatment for nonunions, segmental defects, and osteomyelitis. This systematic review summarizes and discusses reconstructive options available for traumatic extremity injuries when bony transport and free tissue transfer are required for limb salvage. A series of case examples in combination with the senior authors' surgical techniques at a single academic institution is described.

Bone deficit treatment options
Critical bone loss is devastating for patients; orthopedic and plastic surgeons must be prepared to treat these patients with a multidisciplinary approach. It most commonly occurs in patients following trauma, infection, and oncologic resection. Critical bony defects, by definition, necessitate additional intervention, as the defect is too large to undergo the normal physiologic healing process. The inability for bone to spontaneously regenerate is often largely controlled by the presence or absence of periosteal blood supply, as well as a viable docking site for reduction [10] . Defects greater than 2.5 cm are usually considered critical, although there is poor evidence and thus some disagreement between experts with this definition [11][12][13] . In these situations, primary bone grafting is also insufficient to obtain union. More advanced treatment options include the induced membrane technique and bone transport through distraction osteogenesis.

Induced membrane techniques
The induced membrane technique, often eponymously referred to as the Masquelet technique, is a twostage technique involving placement of a foreign body within the defect that results in a biologicallyactive membrane around the defect, which can then provide biology for subsequent bone formation using graft [14,15] . The first stage entails radical debridement and skeletal stabilization [15,16] . A spacer, typically made from polymethylmethacrylate cement, is placed. A pseudomembrane will form around the spacer. This spacer is biologically active and contains growth factors involved in fracture healing and angiogenesis. In the second stage, the previous incision is reopened, the pseudomembrane is incised, bone graft is placed into the defect, and the membrane is closed around the graft.

Autologous bone grafting
The graft used in this technique is typically autogenous bone graft. Autologous graft is histocompatible, osteoconductive, osteoinductive, and osteogenic [17,18] . This provides a structure through which bone, host capillaries, and mesenchymal stem cells can have in-growth. The source of autogenous bone is determined by many factors including location of bone defect, desired function of bone graft, size of bone defect, associated complications, and expected graft site. Locations and methods for graft include the distal tibia, calcaneus, proximal tibia, iliac crest, and via the Reamer-Irrigator-Aspirator system. The optimal timing for placement for the bone graft is debated, with studies suggesting peak osteo-and angiogenic properties occur at 4 weeks, with a subsequent decline [19][20][21] . In situations in which the induced membrane technique is contraindicated, the orthopedic surgeon's best option for attaining bony union is by means of bone transport.

Distraction osteogenesis
Indications for bone transport via distraction osteogenesis include massive bone loss (greater than 5 cm), inadequate soft-tissue coverage, infection, and the need for prolonged stability. However, even with defects this large, other methods along with bone transport should be considered. Of paramount importance, the surgeon must be competent in managing such patients throughout the entirety of their care, as it has been suggested that the success of bone transport is highly dependent upon surgeon experience [22] . Distraction osteogenesis (DO) was described by Ilizarov. The classic method utilizes thin-wire circular frames to provide stability and gradually distract at the osteotomy site. This method of DO was developed in the 1950s and has been a popular method for long bone expansion in Europe and the United States since the 1980s as a treatment modality for critical bone loss. All distraction osteogenesis is dependent on three things: a low-energy corticotomy, bony stability, and distraction with a specific rate and rhythm [23,24] . After the bone defect is defined, circular external fixation with thin K-wires are percutaneously fixated to healthy bone both proximal and distal to the site of bone loss. They are held under tension via fixation bolts to the rings surrounding the affected limb fragments. These rings are then joined using threaded rods or adjustors. The use of multiple fixation rings for each segment allows for increased control of the bone segments, allowing for an increased number of connecting rods and greater ability for bone gap repair [22] . Each case will progress through three phases: latency, in which the healing process begins; distraction, during which time the DO focus is distracted at a specific rate and rhythm; and consolidation, in which the regenerate ossifies [25] .
Ilizarov's method of DO can reliably achieve bony union in limbs that would otherwise not be salvageable. Time of distraction is variable and dependent on defect size, location, and host factors. The maximum rate of distraction is 1 mm/day, and consolidation typically requires twice as long as the distraction phase [26][27][28][29] The published rates of successful limb reconstruction approach 100% in most publications [30][31][32][33][34] . Complication of DO include pin-site infections, stretching of neurovascular structures, soft-tissue scarring, poor regenerate, and non-union of the docking site [30,35] . Classic Ilizarov DO is still widely utilized, but alternative techniques such as cable transport and all-internal techniques using motorized medullary nails have proven successful. These techniques offer similar outcomes with decreased time with external fixation, allowing for improved patient comfort, decreased site infection risk, and sooner time to rehabilitation.
One potential alternative to classical Ilizarov DO is utilization of Taylor Spatial Frames. This unique frame has dual rungs with six struts allowing for telescoping [36,37] , thus providing many of the same benefits to Ilizarov DO but with the ability to provide correction in six planes (coronal angulation/translation, sagittal angulation/translation, rotation, and shortening). In addition, this system has the ability to utilize computer accuracy [36][37][38] . Studies have shown this to be as effective as, but not superior to, classical Ilizarov DO in overall clinical outcomes [38][39][40][41] , while some advantages include use for residual deformity following Ilizarov DO, lower rate of return to the OR, and the ability to correct in all planes [36] .
Finally, the pull wire system is a newer technique that can induce bone transport with utilizing both internal and external fixation through medullary nail placement [42] . This dual distraction method provides similar rates of healing to classical Ilizarov DO, while limiting the risks associated with long-term external fixator use [42,43] . Rozbruch et al. [33] (2008) also suggested that there is expedited bone healing and a decreased risk of refracture of the site of bone transport.

Soft tissue coverage in lower extremity trauma
In many cases of lower extremity trauma, free tissue transfer becomes necessary in an attempt at limb salvage. While successful flap reconstruction may be achieved, complication rates are relatively high in lower extremity reconstruction for a variety of reasons, which may include trauma-induced edema, preexisting vascular conditions, other patient comorbidities, or poor patient compliance during postoperative recovery.
In a meta-analysis conducted by Xiong and colleagues on free flap reconstruction of lower extremity defects, total flap loss and the rate of thrombosis were both found to be about 6% [44] . Overall, 26.1% of the flap losses were due to venous thrombosis, whereas 10.1% were due to arterial thrombosis. Minor complications such as hematoma, partial necrosis, infection, and wound dehiscence occurred at rates between 4.0% and 8.0% [44] .
Given the heterogeneity of lower extremity trauma, several different reconstructive options can be utilized by the plastic surgeon, depending on the size of the defect, the structures involved, and the comfort level of the surgeon.
Historically, muscle flaps were believed to reduce infection rates in contaminated wounds. However, more recently, free fasciocutaneous flaps have proven to be comparable to muscle flaps in terms of success rates, infection rates, and bony union in lower extremity reconstruction [4,[45][46][47] . Additionally, fasciocutaneous flaps are thought to be simpler to re-elevate for subsequent orthopedic procedures, require fewer secondary skin graft procedures, and result in lower donor site morbidity [46][47][48][49] .
Regarding free versus pedicled flaps in reconstruction of the distal third of the lower limb, the same review by Bekara et al. [50] concluded that, while partial flap necrosis is higher in pedicled-propeller flaps, wound dehiscence and infection rates are higher in free flaps. Failure and overall complication rates were similar in both groups [50] .
In addition to the flap that is used, the reconstructive surgeon must also select the appropriate recipient vessel and anastomotic technique. While studies have shown no difference in complication rates between end-to-end and end-to-side arterial anastomoses, end-to-side anastomoses are generally the preferred choice for extremities at risk for vascular insufficiency in order to maintain perfusion to the distal leg and foot [51,52] .
While it was previously thought that only vessels proximal to the injury could be used as recipient vessels for free flap coverage, it has since been shown that there is no difference in outcomes or reoperation rates when flaps are anastomosed to vessels distal to the injury [53,54] .

Bone transport with simultaneous soft tissue reconstruction
It is now well-established that early soft tissue coverage of complex lower extremity trauma leads to successful limb salvage with improved flap success rates and lower rates of infection [55][56][57][58] . Thus, the current practice at most institutions is simultaneous bony fixation and soft tissue coverage. This so-called "fix and flap" protocol, consisting of radical debridement and skeletal stabilization with immediate or very early free flap coverage, has resulted in faster union times with lower infection rates [59,60] .
To overcome segmental bone loss and manage later consequences of lower extremity trauma such as malunion or nonunion, bone transport is often used. By combining free tissue transfer with bone transport, large segmental and soft tissue defects can be simultaneously treated with great success. Flap coverage combined with bone transport allows for better limb length restoration by maintaining length from the start; improving vascularity, which is important for fracture healing; and facilitating bone grafting or other subsequent procedures at the docking site [61] .
While classically, the injured limb with critical bony defect was initially treated by debridement and resection and shortening of bone to allow for primary soft tissue closure, the more recent practice of combining free tissue transfer with bone transport allows for maintenance of the limb's original length and avoids the frequent complications associated with the traditional compression-distraction technique. The free flap provides healthy vascularized soft tissue, under which distraction osteogenesis can then take place [7] .
Few studies have described the effects of distraction on the transferred free tissue and its anastomotic pedicle. Jupiter et al. [62] reported that both the free tissue and the native tissue show equal magnitude of stretch and lengthening without any scar dehiscence after bone transport despite their different tensile strength and mechanical properties. Many studies have shown that, with major vessel repair in lower extremity trauma, the anastomotic site tolerates initial distraction process as early as 2-3 days after surgery [62][63][64] . However, the outcome of the microvascular anastomosis in these cases has not yet been reported. It has been demonstrated, however, that immediate distraction osteogenesis with a recent free flap has not been found to compromise the flap [61] .
Finally, while placement of pins through the free flap has also been noted to be safe, caution must be taken when placing the pins so that their anticipated path does not pierce the pedicle [62,65] . Careful planning and collaboration between the orthopedic surgeon and reconstructive surgeon are needed for these cases to ensure safe distraction against the pedicle and microvascular anastomosis.
Another potential option for patients needing both soft tissue coverage and assistance with bony growth is a medial femoral condyle vascularized graft. First described in the 1990s, this method utilizes the highly vascularized periosteum and either the medial superior genicular or more commonly the descending genicular artery, due to its length and ease of identification [66][67][68] . This method usually follows failure of conventional therapies and has been proven to be efficacious for osteomyelitis, avascular necrosis (AVN), and nonunions. Specifically, it has been used for the humerus and ulna [69][70][71][72] , as well as tibial and femoral defects [70,73,74] . Notably, this methodology has been well studied in the treatment of scaphoid AVN [66,68,75] , with studies reporting up to 100% of patients achieving union [76,77] and level III evidence that it is superior to 1,2-intercompartmental supraretinacular artery graft [68,77,78] . Finally, these studies have reported limited to no complications at both donor and recipient sites [73,77,79,80] .

Literature review
A systematic review of the literature was performed using PubMed database in adherence with PRISMA guidelines. Combinations of the following search terms were used: "'bone transport' AND 'free tissue transfer'", "'Ilizarov' AND 'free tissue transfer'", and "'distraction osteogenesis' AND 'soft tissue coverage'". No limits were placed on any of the search queries. All articles were independently reviewed by two authors. Following the primary search, titles and abstracts were either included or excluded based on predefined eligibility criteria. Inclusion criteria included bone transport as research theme, involvement of soft tissue coverage, relevance to limb salvage, and general population as the sample. Articles were limited to English language and review articles, letters, and comments. Subsequently, evaluation of the full text of selected articles was similarly performed and their bibliographies were assessed for additional articles to include.
Articles that described patient outcomes, details of bone transport, and type of soft tissue coverage were included. Studies were excluded if they had inadequate data, did not involve a combination of bone transport and soft tissue coverage, or were anecdotal in nature.
The following data from each article were extracted: sample size, mean patient age, indications for intervention, type of flap, use of additional modalities (antibiotic beads or bone substitute, vein or bone graft, and hardware), details of bone osteogenesis (average bone and soft tissue defect, time-to and rate of distraction, and total distraction), postoperative events (complications, total or partial flap loss, skin graft complication, hematoma, nonunion, malunion, infection, recurrent osteomyelitis, broken fixation wires, flap depression, bone exposure, fracture, limb-length discrepancy, second flap surgery or reoperation, and amputation), average follow-up, satisfaction, weight bearing status, and success rate.
Unpaired t-tests were utilized to assess if complication rates were statistically significant between the simultaneous placement of frame and flap and the placement of the flap before or after fixation. Fisher's exact test was used to compare differences in complication rates in patients with acute and chronic wounds. Statistical significance was determined at P < 0.05.

Case series
A retrospective review was conducted of patients undergoing bone transport with simultaneous soft tissue coverage performed at the Keck Hospital of the University of Southern California between 2016 and 2019. Three non-consecutive cases performed by a single plastic and orthopedic surgeon are described below.

RESULTS
Our initial literature search identified 91 unique articles, of which 13 full text-articles were assessed for inclusion. Six full-text articles from the search were ultimately included, added to eight articles selected from additional search. These 14 articles  Table 2]. No studies included patients with tumors or masses. The average time from injury was 1.1 years (range of 18 days to 3 years) in 8 of the 14 articles. An additional 2 of the 14 articles reported two separate times from injury: 1 [7] or 3 [81] weeks for patients with acute trauma and 10 [81] or 18 [7] months for patients with chronic osteomyelitis or nonunion.
All 14 articles described a surgical approach. In 11 of the 14 articles, the authors used a classic Ilizarov frame; in one, they used either an Ilizarov frame or segmental transport external fixator; and, in one, they used a hexapod frame. Only 1 of the 14 used anterior double-stacking external fixation. Some patients   Table 3]. With respect to soft tissue reconstruction, four of the 14 articles quantified the soft tissue defect. The average soft tissue defect was 125 cm 2 (range of 88-219 cm 2 ).
All 14 articles discussed the general type(s) of flaps used. Soft tissue reconstruction was most commonly performed with free flaps (88.0%) as opposed to rotational flaps (12.0%). Of these, most were muscle flaps (96.7%) with fewer being fasciocutaneous flaps (3.3%). However, only 13 of the 14 articles specified the type of flap used for each procedure. The most frequently used flaps were free latissimus dorsi and rectus abdominis, followed by rotational gastrocnemius [ Table 4]. Seven of the 14 articles discussed type of anastomosis. End-to-side anastomosis (55.2%) was performed slightly more often than end-to-end anastomosis (44.8%), and, while a large range of recipient vessels was used, posterior tibial (53.2%) and anterior tibial (24.2%) were the most common [ Table 5].
With regard to bone reconstruction and transport, bone defect was mentioned in nine articles, whereas total length of distraction was mentioned in four of the 14 articles. The average bone defect was 8.9 cm (range of 6.0-12.5 cm) and average total distraction was 6.1 cm (range of 4.3-10 cm). The time to distraction was discussed in five of the 14 articles, and rate of distraction was discussed in six of the 14 articles. The average time to distraction was 24.1 days (range of 7-73.5 days) and the average rate of distraction was 0.96 mm/day (range of 0.75-1 mm/day).
As discussed in 11 of the 14 articles, additional techniques were utilized for some extremities. These techniques included bone graft (62.3%), antibiotic beads (48.6%), antibiotic bone substitute (13.0%), vein    Complications were further assessed based on timing of soft tissue coverage, which occurred simultaneously with bone transport frame as well as before or after bone transport. Starting with simultaneous placement of flap and frame, five of the 14 articles used this approach for 50 patients with a mean age of 34.4 years in 50 extremities [ Table 6 and Figure 2]. Three extremities needed reoperation; however, none required amputation or a second flap surgery. These articles reported infections requiring readmission or i.v. antibiotics (n = 5, 10%), nonunions (n = 4, 8.0%), malunions (n = 3, 5.8%), partial flap loss (n = 1, 2%), total flap loss (n = 1, 2%), fracture (n = 1, 2%), and limb-length discrepancy (n = 1, 2%) in several extremities. In contrast, seven of the 14 articles documented complications after soft tissue reconstruction with flaps either before or after application of frame used for bone transport [ Table 5 and Figure 2]. This method included 82 extremities of 81 patients with a mean age of 32.9 years. A higher number of extremities required amputations (n = 2, 2.4%), reoperation (n = 25, 30.5%), and second flap surgery (n = 7, 8.5%). Similar to the simultaneous group, these articles reported total flap losses (n = 6, 7.3%), partial flap losses (n = 5, 6.3%), nonunions (n = 7, 8.5%), malunions (n = 6, 7.3%), limb-length discrepancies (n = 7, 8.5%), infections requiring readmission or i.v. antibiotics (n = 3, 3.7%), and fracture (n = 1, 1.2%) in a number of extremities. Additionally, this group reported flap irregularities or depression (n = 8, 9.7%), bone exposures (n = 2, 2.4%), skin graft complication (n = 1, 1.2%), hematoma (n = 1, 1.2%), and broken fixation wires (n = 1, 1.2%). Of note, there was no statistical significance between the rates of any complication between the 2 groups.  one of which is a 52-year-old male with a compound comminuted fracture of the distal leg, a resultant 14-cm tibial bony defect, and a large soft tissue defect. He received a free rectus abdominis muscle flap, which failed secondary to venous thrombosis about 55 days after injury. He then had a second and successful free rectus abdominis flap. Months later, he underwent bone transport, resulting in full weight bearing capacity 18 months after injury. Hutson et al. [88] denoted two flap failures out of 19 extremities undergoing reconstruction with flaps first. The methods utilized were two latissimus dorsi free flaps to recipient posterior tibial arteries anastomosed in either end-to-side or end-to-end fashion, both resulting in flap failure. The patients were reported to later undergo bone distraction, although further soft tissue rearrangement, if any, was not mentioned. Lastly, Chim et al. [89] reported on 28 extremities undergoing flap reconstruction before or after fixation, three of which failed. One patient had successful bone transport and was managed with distraction lengthening of soft tissue, whereas the other two had concurrent failure of distraction and were managed with amputations. One amputation occurred in a 34-year-old patient with a Gustilo II infected bony nonunion after a motor vehicle collision and an associated wound size of 50 cm 2 , who had a failed gracilis free flap to recipient posterior tibial artery with end-to-side anastomosis and venous commitantes in end-to-end fashion. The other amputation was on a 57-year-old patient with a Gustilo IIIB acute bone loss defect due to a crush injury and an associated wound size of 700 cm 2 , which was managed with a latissimus dorsi free flap through end-to-side anastomosis of posterior tibial artery and end-to-end anastomosis to anterior tibial venous commitantes.
Four studies with a total of 32 patients described the management of acute injuries. Regarding union, there was one nonunion and three malunions, while no patients suffered a refracture. Regarding the flap, there were 5 losses (three total and two partial) and 9 flap irregularities. Sixteen patients required a secondary reoperation, and seven of them were for the flap specifically. One patient suffered from a deep infection. There were five studies that characterized their wounds as chronic. However, Lowenberg et al. [61] (1996) reported the lone study to clarify their complications and describe them. Of their 23 patients, six suffered from a complication. Ones of note include two with nonunion, one with malunion, and one with partial loss of a flap, in deep infection. Additionally, there was an increased rate of second surgery (P = 0.049), as five patents needed a second surgery -one for a flap replacement specifically following initial failure.
Despite significant complications and reoperations, 98.8% (range of 92.9% to 100%) of patients averaged from 6 articles were ultimately weight bearing, and the success rate was 98.4% (range of 92.8% to 100%) averaged across 11 articles. One of the 14 articles reported a satisfaction rate of 100% from the 34 patients involved.

Cases
Case 1: free anterolateral thigh flap with taylor spatial frame first A 69-year-old male with a history of obesity and hypertension was involved in a motor vehicle collision 40 years prior that resulted in an open tibia fracture that was treated with plate fixation at that time. He recovered from that surgery and did well without complication until 2 years prior to presentation when he was found to have purulent drainage around the plate, which was treated at an outside facility with serial debridement and skin grafting. He was ultimately referred to our center at Keck Hospital and was found to have chronic osteomyelitis of the tibia. After thorough debridement of bone and soft tissue, he was left with 17.5-cm bony defect of the tibia and a wound that was approximately 100 cm 2 . The bony defect was temporized with antibiotic spacer and culture-directed intravenous antibiotics until cultures were negative, at which point the patient was placed in a hexapod frame. Five days thereafter, he received free flap coverage of his large wound with an anterolateral thigh (ALT) fasciocutaneous flap anastomosed in end-to-side fashion into the posterior tibial vessels. Bony transport was initiated once he had completed 6 weeks of i.v. antibiotics and he could be scheduled for surgery for corticotomy, which was 98 days after his initial external fixator placement. He was distracted at a rate of approximately 0.5-1 mm per day. After two subsequent surgical revisions of his multiplanar external fixator, he underwent removal of the external fixator 505 days after his initial ex-fix placement, and he underwent a Masquelet procedure to complete bony union. He did suffer equinovarus deformity of the foot, for which he received tendon lengthening and ankle-spanning external fixation, as well as arthrofibrosis of his knee, for which he underwent two lysis of adhesions and quadricepsplasty. Currently, he is weightbearing as tolerated and proceeding to follow with physical therapy. At his most recent follow-up four years after presentation, he had achieved bony union but developed an infection for which he is currently being treated with oral antibiotics and is scheduled to undergo surgical irrigation and debridement.

Case 2: free latissimus dorsi and rotational gastrocnemius flap with delayed NuVasive precise frame
A 45-year-old man sustained a left Gustilo-Anderson IIIB tibia/fibular shaft fractures with an associated distal fibular fracture, for which he underwent debridement and open reduction external fixation at an outside facility. He was transferred to our services when it was determined that he would need soft tissue coverage of his extremity. He was taken to the operating room for debridement of necrotic fibula stripped of periosteum and antibiotic spacer placement. His bony defect was approximately 8 cm and his soft tissue wounds were approximately 75 cm 2 . Five days after his initial debridement at our institution, the patient received both a free latissimus myocutaneous flap anastomosed in end-to-end fashion to the anterior tibial vessels as well as a rotational gastrocnemius myocutaneous flap for coverage of a more posterior wound. Split thickness skin grafting was employed to cover his remaining non-critical wounds. He proceeded with physical therapy and was ultimately weight bearing; however, it thereafter became apparent that he had nonunion of his tibia fracture, likely due to the extensive zone of injury. As such, 276 days after this initial treatment, he was taken to the operating room for placement of motorized magnetic transport nail (Nuvasive, Aliso Viejo, CA) for bone transport. He started distraction 11 days after placement of bone transport nail. He went to the operating room 77 days later for preparation of the docking site as well as bone grafting from the iliac crest. He has completed transport, is fully weight bearing, and participating in physical therapy. Case 3: free latissimus dorsi with ilizarov frame first A 31-year-old man sustained a right distal tibia/fibula fracture from a motorcycle crash, treated at that time at an outside facility with debridement and stabilization with an external fixator, which two weeks later was converted to internal fixation. At the time of definitive fixation, the wounds were not able to be closed and ultimately a rotational flap was employed for closure. He thereafter developed infection and remained non-weightbearing. He was told that amputation was his only option, which he refused, and approximately 6 months after his initial fixation, he presented to our clinic to discuss limb salvage options. He was found at that time to have osteomyelitis and nonunion of his tibial fracture. His wounds had many sinus tracts, which were draining purulent effluent. He was taken to the operating room for debridement, hardware removal, and antibiotic-impregnated nail placement at the diaphyseal tibial fracture. He subsequently underwent two more debridements as well as removal of internal antibiotic coated nail and placement of multiplanar ringed external fixator. He was ultimately left with a bony defect of approximately 11.5 cm and soft tissue defect of approximately 500 cm 2 , at which point he was determined prepared for soft tissue coverage. One week after external fixation, he underwent free latissimus dorsi muscle flap with microvascular anastomosis in end-to-end fashion to the posterior tibial vessels as well as skin grafting for coverage. He underwent corticotomy in preparation for bony transport 42 days after his free flap, and subsequently initiated transport 55 days after his free flap. He received one additional surgery for elevation of the flap out of the docking site and simultaneous bone grafting to the docking site. Since that time, he has been distracting and has reached equal limb length. He has been weight bearing and working with physical therapy [ Table 7].

DISCUSSION
This systematic review evaluates the choices for soft tissue coverage in distraction osteogenesis for upper and lower extremities, and presents both indicators of success (weight bearing status and satisfaction) and complication rates to better inform the surgeon's decision-making process. The indications for surgical intervention include chronic nonunion, osteomyelitis, and, less commonly, acute trauma. Fewer articles detailed reconstruction after acute trauma, which was   1 years), which ranged from within weeks to up to three years. Two cases in the present study describe treatment within days to a few weeks, whereas the other indicates over 40 years since initial injury, which better aligns with what is seen in our review of the literature. Additionally, the present study includes a single article that reports flap and frame repair of upper extremity injuries [61] . Although there is a paucity of literature discussing the use of soft tissue coverage and distraction osteogenesis in upper extremity reconstruction, the study was included to bring light to the innovative, diverse applications of this technique.
The current gold standard for managing large soft tissue defects that cannot be closed directly is the "fix and flap" protocol, which has several recognized advantages [4,51,59,60] . However, with respect to the results of our literature review, only about one third of the extremities described underwent simultaneous flap and frame placement. Similarly, this simultaneous approach is taken in only 1 of the 3 cases presented. While this is partially due to the inclusion of older studies that were published prior to the acceptance of modern "fix and flap" approach [7,9,41,61,62,65,[84][85][86]88,89] , it is also important to note that the gold standard typically applies to acute traumatic injuries, whereas many patients undergoing bone transport are patients with chronic extremity issues.
With respect to reconstruction of bone, the present review measured the average bone defect to be 8.9 cm, which required intervention to restore limb length. The mainstay of distraction osteogenesis is the application of the Ilizarov method for external fixation, although nowadays novel methods of complementary internal fixation have been innovated with improved patient comfort, decreased infection rates, and quicker recovery [43,[91][92][93][94][95] . The present review identified that the majority of articles employed external fixation consistent with the Ilizarov method. The remaining articles used variations of the Ilizarov method, except one that used anterior double-stacking external fixation. However, for many reasons, an array of surgical techniques may be used to tailor reconstructive approach to the individual's unique presentation.
While bone defects smaller than 8 cm can be successfully closed using bone grafting [7,9,96] , larger defects may require a vascularized bone graft [97] . Common vascularized bone grafts that have been described include contralateral fibula and iliac crest [8] . These operations, however, carry significant donor site morbidity and are also limited by the size of the donor site [7,77] . For patients with large defects, the anatomy of the iliac crest makes it unsuitable [27] , and harvesting a contralateral fibula in a patient with unilateral lower extremity trauma can be problematic due donor site complications [98] and insufficient pedicle length [99] . In certain circumstances, specifically in previously infected femoral shafts, efficacy has been shown for vascularized fibular grafts [100] . As seen in the present review, 62.3% of extremities were managed with both distraction osteogenesis and bone grafting. Two of the three cases discussed similarly received bone grafts.
When comparing bone grafting with soft tissue coverage against resection and bone transport, results have been shown to be similar, with significantly less limb length discrepancy in the bone transport group [101] . Although distraction osteogenesis in combination with free flap reconstruction has clearly been proven to be a useful treatment modality, the procedure is time consuming and can produce a multitude of challenging complications if not approached with great care.
Some of the common complications associated with this procedure are bone exposure [94] , nonunion [7] , flap necrosis [102] , and downward depression of the flap [90] . While some of the problems can be resolved with conservative care or minor revisions, there are others with more serious consequences. As found in the present review, flap failure may result from venous thrombosis or vasculature disruption due to acute or chronic pathologies [81,82,88,89] . The articles described these losses in patients with large bony and soft tissue defects undergoing a number of different free flaps (latissimus dorsi, rectus abdominis, and gracilis) through both end-to-end and end-to-side anastomoses to either anterior or posterior tibial vessels, consistent with the most frequently used surgical techniques. While two patients eventually underwent amputations, this adverse event was seen in only 1.4% of all extremities. Likewise, there was no statistically significant difference in rates of flap failure or amputation between the extremities that were managed according to the "fix and flap protocol" versus the placement of a flap before or after the frame. While the difference in outcomes between the most commonly used free flaps, the rectus abdominis and latissimus dorsi flaps, would be beneficial to evaluate, this was not possible given the nature of the studies. Nearly all of the case series that were included in our review reported multiple different types of flaps that were used in their patients, however the results were not separated according to flap type.
Looking further into patient complications, the present study indicates that injuries managed in the acute setting were significantly more likely to undergo secondary surgical repair, compared to those who were chronic. While this may be due to more controlled surgical management of chronic wounds predating the beginning of the study time, the variability in injuries, patient morbidities, surgeon expertise, and postoperative care limits the value of this finding.
An additional issue that can arise during the distraction process is the necessity to revise pin position or flap configuration. In cases of flap necrosis, however, distraction is often delayed or stopped altogether, necessitating additional procedures such as bone grafting [102] . Thus, meticulous planning of the flap in addition to careful postoperative care is crucial for successful reconstruction.
Deciding the reconstructive method for soft tissue coverage can be challenging, as the options to choose from are vast. In the articles reviewed, free flaps and muscle flaps were more commonly utilized than rotational flaps and fasciocutaneous flaps, respectively. As for specific flaps, rectus abdominis flaps were used for the greatest number of extremities, whereas free latissimus dorsi flaps were cited in the greatest number of distinct articles. Microsurgical technique varied, with end-to-end anastomosis being used almost as often as end-to-side anastomosis. Recipient vessels were most commonly the anterior and posterior tibial arteries, although a diverse set of choices exist. Given the comparable success rates with all of these different techniques, flap choice is often left to the discretion of the surgeon.
Aside from infection and fracture, the present review found that simultaneous placement of flap and frame, or the "fix and flap" protocol, had fewer reported adverse events overall [ Figure 1]. These results align with the well-documented faster union times but differ from reported lower infection rates in the current literature [59,60] , given that our review actually found a higher rate of infection in the simultaneous placement of flap and frame. Two of the cases presented required additional surgeries, one for knee arthrofibrosis and foot deformity and the other for flap elevation, bone grafting, and docking.
Other factors must be taken into consideration when attempting to mitigate the risk of flap loss, with one of which being rate of bone transport. While Jupiter et al. [62] concluded that the free tissue and the native tissue undergo equal amounts of stretch and lengthening, Horas et al. [103] noted a difference in speed between bone transport and soft-tissue movement, which could potentially jeopardize the vascular pedicle secondary. This risk increases with the amount of transport needed. Interestingly, all of the papers assessed in the review performed distraction at a rate between 0.75 and 1 mm per day, indicating that there is agreement that this range of rates produces optimal results.
Although limb salvage and reconstruction through soft tissue coverage and distraction osteogenesis comes with substantial risk of complications and reoperations, outcomes remain strong. Over 98% of patients were ultimately weight bearing, and article authors on average reported rates of success and patient satisfaction as greater than 98%. All three of the case examples outlined in the present study are fully weight bearing as tolerated and working with physical therapy.
There are several limitations to the present review. Due to the paucity of literature evaluating the outcomes after utilizing soft tissue coverage and bone transport after critical bone loss for limb salvage, this systematic review was limited to case reports and case series. Thus, the quality of studies included is a limitation. In addition, the studies included are very heterogeneous with the parameters discussed and outcomes evaluated, which led to an even smaller group of articles coinciding with any one parameter. Given the degree of inconsistency in the way results were reported in each article, many studies had to be excluded, leading to a lower power. Lastly, reoperation for cancellous bone grafting for nonunion was frequently performed across articles; however, it remained unclear in the literature if this was an expected complication that needs anticipated surgery vs. an actual complication. These events were included under the category of bone grafting.
In conclusion, the reconstruction of extremities with critical bone loss, whether due to acute trauma, nonunion, or chronic osteomyelitis, remains a significant challenge for surgeons. Our findings suggest that no one surgical approach is superior, and the treatment algorithm ultimately remains a decision to be made between the surgeon and the patient. In patients with bone defects too extensive for autologous bone grafting or in patients who are not candidates for this approach, a combination of bone transport with soft tissue reconstruction remains an excellent choice for satisfactory functional outcomes.