|Home | About | Journals | Submit | Contact Us | Français|
Early vascularized soft tissue closure has long been recognized to be essential in achieving eventual infection free union. The question of whether muscle or fasciocutaneous tissue is superior in terms of promoting fracture healing remains unresolved. Here we review the experimental and clinical evidence for the different tissue types and advocate that the biological role of flaps should be included as a key consideration during flap selection.
Open tibial fractures are severe injuries, largely affecting young men of working age, and take on average 43 weeks to unite, with 13% developing non-union in the best centres. There is, therefore, an urgent need to enhance the process of bone repair in these patients. There have been numerous innovations in the techniques used for fracture stabilization as well as biological therapy, such as bone morphogenetic proteins (BMPs). Improvement in the care pathway, through a multidisciplinary and integrated orthoplastic approach, has also led to significant improvements in patient outcomes[3-6]. These refinements have reduced the mean union time to 26 weeks.
Considerations when planning soft tissue coverage include the size and location of the defect as well as donor site morbidity. An area which has not featured prominently in determining flap choice thus far is the potential biological role the flap may play in the fracture repair process. However, there is a growing body of experimental evidence that demonstrates that the biological characteristics of the tissues in a flap can significantly influence fracture healing, thereby potentially reducing union time and the rate of delayed or non-union.
The role of soft tissue reconstruction in open fractures is not limited to wound coverage to prevent wound desiccation and infection. Soft tissues also contribute to fracture repair by serving as a local source of stem or osteoprogenitor cells, growth factors and vascular supply[7-10].
A key role of soft tissue flaps in lower limb trauma is to serve as source of vascular supply to bone ends that have been stripped of periosteum and undergone disruption of the endosteum. There is evidence that muscle contributes greater vascularity to a defect than fasciocutaneous tissue[12-16]. A study using a canine model to compare the blood flow at the musculocutaneous and fasciocutaneous flap/wound interfaces with no underlying fracture showed that whilst there was an initial increase in muscle blood flow in the first 24 hours, the deep surface of the fasciocutaneous flap underwent a slower and steadier increase in blood flow over the experimental period of 6 days to exceed that of muscle by this time point, yet there was greater evidence of healing in the muscle group . Using a murine tibial fracture model, Harry et al. found that at all time points the vascular density was greater in fasciocutaneous tissue in apposition with a periosteally stripped fracture than muscle, and in spite of this, fracture repair was more rapid in the muscle group[19, 20]. These observations suggest that while vascularity is essential for wound healing, including bone repair, other biological factors become limiting, once an adequate blood supply threshold has been met.
Fracture repair requires the recruitment of osteoprogenitor cells. Mesenchymal stem cells (MSCs) are, by definition, multipotent and can therefore serve as a source of osteoprogenitor cells. MSCs may originate from a variety of tissues including the bone marrow, periosteum, dermis, adipose tissue and muscle, as well as blood vessels and the circulation. In closed fractures, the main sources of osteoprogenitor cells are thought to be the bone marrow and periosteum[21-25]. However, high energy open fractures of long bones are characterized by loss of the periosteum and bone marrow, especially following insertion of an intramedullary rod. Under these circumstances the main osteoprogenitor cells must originate from the local soft tissues or the circulation[10, 26].
It is well established that muscle provides a suitable environment for osteogenesis, although damaged muscle is less effective. In 1965, Urist found that new bone formed readily when decalcified bone was implanted into muscle and deduced that the inductor cells were derived from the host bed. Furthermore, purified BMPs injected into muscle are capable of inducing ectopic bone formation[29, 30]. Using a mouse model, Zacks et al. found that muscle (but not liver tissue) demonstrated a significant osteogenic effect. Extraskeletal ossification observed in patients with fibrodysplasia ossificans progressiva and heterotopic ossification following either orthopaedic surgery or blast injuries tend to occur in muscle[33, 34].
Both fasciocutaneous tissue and muscle are rich reservoirs of MSCs[9, 10]. However, the characteristics, including the osteogenic potential, of MSCs vary depending on their tissue origin. For example, human stromal cells derived from muscle exhibit a significantly greater potential for osteogenesis than those from fasiocutaneous tissue, including both skin and adipose, and are equivalent to those from bone marrow. Using a critical sized rat femoral diaphyseal defect model, muscle was found to be more effective in promoting bone repair than adipose tissue. Muscle-derived stem cells can be recruited from muscle and stimulated to undergo osteogenic differentiation by proinflammatory cytokines, especially TNF-α, released at the site of injury.
Muscle also provides a bone anabolic environment through the expression of members of the transforming growth factor-β (TGF-β) superfamily of growth and differentiation factors, including the BMPs. The reciprocal relationship between muscle and bone mass is well described, particularly the strong association between sarcopenia (age-related loss of muscle mass) and osteopenia. Muscle and bone are believed to be mutually regulating via physical forces and cytokine control. Indeed, recent evidence indicates that muscle serves as an endocrine organ that releases trophic factors, known as myokines, which have been identified as key regulators of the muscle and bone mass. Further observations suggest that intact muscle supports bone repair via the release of bone anabolics, including IGF-1, IL-6, BDNF and FGF-2[37-39] while severely injured muscle, such as following military trauma, impairs this process through the release catabolic myokines, including myostatin (GDF-8)[40-42]. Therefore, the net effect on bone is dependent on the balance of these factors.
Soft tissue flaps are believed to possess an anti-bacterial property that is independent of vascularity. Chang and Mathes used a canine model to compare the anti-microbial properties of different tissues during wound healing. Chambers inoculated with bacteria were inserted beneath random pattern flaps raised on the flanks with no underlying fracture. Muscle was found to be superior in eliminating bacteria from the wound bed. In a separate study, they compared bacterial growth within the wound fluid at interface of musculocutaneous and fasciocutaneous flaps and found that despite a higher blood flow and tissue oxygen tension in the fasciocutaneous group, muscle exhibited a greater ability to reduce the bacterial count[17, 18]. Moreover, histological examination revealed greater evidence of wound repair, including increased collagen deposition, at the muscle interface.
Recent evidence suggests that the presence of muscle is an important contributor to bone healing[9, 10, 44]. For example, the size of fracture callus is greater adjacent to muscle and muscle coverage accelerates fracture repair in murine models[19, 46].
Schemitsch et al.[11, 47-50] compared cutaneous and muscle tissues in a series of studies using a canine open tibial fracture model. A devascularized segment of tibia was covered with either transposed tibialis muscle and the skin incision closed (muscle flap group) or skin closed directly following excision of the underlying fascia (skin group), and fracture healing was assessed. There was a significant increase in the bone blood flow and rate of union in the muscle flap group compared to the skin group[11, 48]. Muscle flaps were also found to significantly increase cortical porosity, enveloping callus and intracortical new bone formation. Notably, there was no direct correlation between the soft tissue blood flow and the indices of bone repair, and resting muscle blood flow was found to be higher in the control limb using the microsphere technique. Subsequent investigation of flap perfusion showed no difference in extraosseous soft tissue perfusion at the fracture site between the different groups. However, this model does not emulate the clinical scenario as fascia beneath the anterior skin was excised in both groups, and only one-third of the circumference of the osteotomised tibial segment was in contact with the soft tissue flap, with the posterior segment in direct apposition with intact periosteum and musculature in both groups.
Our group developed a murine tibial fracture model to emulate the high-energy injuries encountered in clinical practice. One third of the circumference of the fracture was permitted direct contact with either muscle of fasciocutaneous tissue by excluding the remainder with polytetrafluoroethylene. At 28 days following fracture, there was greater healing in the experimental muscle coverage group compared to skin and fascia alone with almost 50% more mineralized bone content and a three-fold stronger union in the muscle group compared to fasciocutaneous group despite a higher vascular density in the fasciocutaneous tissue compared to the muscle at all time points[19, 20].
In a series of studies, Utvag et al.[27, 46, 51, 52] examined the effect of separating muscle from the fracture site in the long bones of the lower limb in rodents. Interposition of an impermeable membrane between periosteum and muscle resulted in impaired healing in a rat femoral model. However, a delay of 2 weeks in insertion of the impermeable membrane did not have any detrimental effect, indicating that early direct contact of muscle with the fracture site enhances fracture healing. Excision of the anterolateral compartment muscles in a rat tibial fracture model also resulted in delayed healing. This effect was abolished when the muscle defect was corrected by transposition of the gluteal muscle. Furthermore, isolating a tibial fracture in a rat model with nitrocellulose membranes with pore sizes ranging from 3 to 50kDa still resulted in impaired healing, confirming that direct contact of muscle with the fracture site, likely the cellular component, is an important factor in the healing of diaphyseal fractures.
Most of the relevant clinical evidence comprises descriptive retrospective observational case series (Table 1) and all studies are categorized as Level 4 evidence according to the Oxford Centre for Evidence-based Medicine. Few of these specifically compared muscle with fasciocutaneous flaps and those that did were severely limited by the lack of power and case heterogeneity, including a wide variety of patients with clinical indications ranging from open fractures to burns or contour deficits. There were insufficient details in the publications to allow us to separate the flaps used to cover open fractures. Furthermore, the outcome measures differed considerably between studies, for example, not all studies reported time to fracture union, rates of deep infection or even flap survival. Therefore, the currently the published literature precludes amalgamation of data from different studies and hence any meaningful meta-analysis or systematic review that can provide guidance for the use of different flap options in the management of open fractures of the lower limb.
It has been observed that open fractures of bones not surrounded by muscle, such as the tibia, unite slowly and that healing of open bone defects is accelerated when a muscle flap is used to cover the wound. Furthermore, intact muscle appears to be more effective at promoting bone repair than injured muscle. In a retrospective review of 84 consecutive patients with severe open tibial fractures, which included 79 grade IIIB and five Gustilo grade IIIC fractures, Gopal et al. presented their ‘fix and flap’ approach comprising early effective debridement, skeletal stabilization and subsequent obliteration of the dead space with a well-vascularized muscle flap. Their longer-term outcome of 34 severe open tibial fractures, including 30 graded as Gustilo grade IIIB, showed a mean union time of 41 weeks, and rates of limb salvage and amputation compared favourably with other series.
Other authors have also commented that muscle provides superior coverage of open tibial fractures[55, 57-60]. Georgiadis et al. highlighted the ability of muscle flaps to reduce both healing time and deep infection while Small and Mollan retrospectively reviewed 168 open tibial fractures treated over a 15-year period and found a lower necrosis rate in local muscle flaps (13.3%) and free tissue transfer (most were muscle only latissimus dorsi and rectus abdominis flaps; 10%) compared to fasciocutaneous flaps (21.2%).
Fasciocutaneous flaps are popular and have been used successfully in large clinical series to reconstruct open tibial defects[62-68]. Local fasciocutaneous flaps are reliable for lower limb reconstruction, as demonstrated by Ponten in his study of 23 cases. They offered significant advantages, including simplicity, availability and versatility, replacing ‘like with like’ without sacrificing muscle function[62, 63, 65, 70]. However, in a series of 100 consecutive local fasciocutaneous flaps, which included 67 to the lower extremity, Hallock reported that 15% required further surgical intervention, with the majority in lower limb wounds and attributed to peripheral vascular insufficiency. Although the majority of patients requiring vascularized tissue had been subject to trauma, it was not clear that all patients had fractures. The coverage of contaminated wounds was highlighted, with short-term healing achieved, suggesting that local fasciocutaneous flaps could be used to cover previously infected fractures.
The major advantage of local fasciocutaneous flaps is their relative simplicity of procedure. However, in patients with high-energy injuries, they may be susceptible to tip necrosis. Erdmann et al. published their experience of pedicled fasciocutaneous flaps in lower limb trauma. Over a five-year period, they used distally-based, islanded fasciocutaneous flaps to reconstruct open tibial fractures to cover the distal one-third of the leg, ankle, heel or foot in 61 patients, with 25 fractures graded as Gustilo IIIB. The overall complication rate was 7.6%, which included five patients with Gustilo IIIB fractures suffering complete flap loss and four patients developing chronic osteomyelitis that led to non-union. Thus, the complication rate for coverage of Gustio IIIB fractures with distally-based islanded fasciocutaneous flaps reached 20%. The mean time to fracture healing was 5.9 months. In a prospective multicentre study involving high energy lower limb trauma, rotational flaps, including fasciocutaneous tissue and muscle, were compared to free muscle flaps in 195 limbs in 190 patients. In patients with the most severe grade of osseous injury, wound complications including infection, necrosis or flap loss, were significantly higher in the rotational flap group (44% compared to 23%), and furthermore, these were 4.3 times more likely to require operative intervention.
Fasciocutaneous flaps have been found to be useful in chronic osteomyelitis of the lower limb by Hong et al.. Over a three-year period, they treated 28 consecutive patients with surgical debridement and reconstruction using free anterolateral thigh perforator flaps, although six of these fasciocutaneous flaps were combined with a segment of vastus lateralis muscle. The well-contoured soft tissue flaps allowed effective resurfacing at the level of the ankle, permitting normal footwear, and unlike the muscle flaps, the elasticity of the skin flaps permitted easy re-exploration for secondary bone grafting procedures, with tension-free closure. Although lacking long-term follow-up, they felt that with adequate debridement and obliteration of dead space, the anterolateral thigh perforator flap was a time-efficient, functional, aesthetic and safe procedure that provided successful coverage for chronic infection.
More recently, the sural artery flap has gained popularity. However, in a multicentre review of 70 flaps, Baumeister et al., found that up to 36% developed necrosis, and this was most likely to occur in patients with comorbidities, including diabetes mellitus, venous insufficiency and peripheral arterial disease. This is the sub group that is erroneously considered by some surgeons to be unsuitable for free flaps.
In a retrospective review over an 18-year period, Hallock assessed the role of muscle and fascia flaps in lower extremity trauma. Details of flap coverage in 160 limbs in 155 patients, of which 60 were local muscle, 50 local fascial and 74 free muscle and fascial flaps, were reported. Flap selection was not randomly assigned, but based on clinical need. Complications were related to the severity of the injury, with 39% associated with free flap transfer, whereas local muscle and local fascia flaps had similar morbidities of 27% and 30%, respectively.
Donor site morbidity is often a factor in flap selection. In a retrospective review, the same author compared the relative donor site morbidity of muscle and fascial flaps. In total, 147 local muscle/musculocutaneous and 122 fascia/fasciocutaneous flaps were used to reconstruct all regions of the body. These included a total of 45 muscle and 72 fasciocutaneous flaps for the lower limb, although it was not clear whether all these patients had exposed fractures. Overall, donor site complications were equivalent at 14% for each group while major complications, including nerve injury, failed graft, necrosis or ulceration, were infrequent in both. Most difficulties, however, were encountered below the knee with fasciocutaneous flap donor sites, where no local muscle option was available, and the skin grafted donor sites were described as cosmetically unappealing.
Finally, a retrospective review of patients with open tibial fractures treated with either free muscle or facsciocutaneous flaps showed that similar numbers went on to achieve bony union and were able to walk unaided at two years. The authors found that muscle conformed better to complex defects but fasciocutaneous flaps better tolerated secondary surgical procedures.
Meticulous wound debridement removes any non-viable soft tissue including muscle that may serve as a nidus of infection and a source of catabolic myokines to inhibit bone repair. From the available data and our own experience, we suggest that fasciocutaneous flaps may be superior to muscle for coverage of rapidly uniting metaphyseal fractures, particularly around the ankle, thereby avoiding skin grafts, which might be susceptible to minor trauma. However, muscle in direct apposition with diaphyseal fractures would aid healing. While muscle flaps covered with skin grafts are aesthetically unappealing and can be difficult to elevate for secondary procedures such as bone grafting, an alternative which retains the biological benefits of muscle apposition is to use chimeric flaps, such as a free anterolateral thigh flap that includes a segment of vastus lateralis (Fig. 1). The plasticity of muscle also helps to obliterate the dead space, thereby reducing potential complications associated with hematoma formation. In summary, thorough wound debridement and early flap coverage of open fractures achieves infection-free union and the biological contribution of the constituent tissues should be taken into consideration during flap selection.
Financial Disclosures and Products Page Project No F-09-23N was supported by the AO Foundation. JKC is in receipt of a Wellcome Trust Research Training Fellowship and a Royal College of Surgeons of England Research Fellowship. The Kennedy Institute of Rheumatology receives a core grant from ARUK (Registered Charity No. 207711).
Author participation JN, JKC conceived the idea of the manuscript.
JKC, JN, GW and LH co-wrote and edited the manuscript.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.