Prefabrication and free transfer of a tissue engineered composite flap – An experimental model in the rat

1 Introduction

In the 1960’s, technical advances led to a new surgical field which enabled reconstructive surgery enrichment that now is an essential component of everyday clinical life: i.e. microsurgery. Microsurgery can lead to excellent results in transplantation, replantation and flap surgery, but it still is not fully understood and thus remains a contemporary issue with respect to optimization, modification and innovation. To date, the irreplaceable workhorse of microsurgical research is the animal model. Experimental flap models, mostly performed in rats, are suitable for investigations involvingphysiology, reinnervation, and the design and function of flaps since they provide conditions that precisely resemble clinical situations. The number of experimental model designs in the literature correlates with the variety of tissue flaps, for example those containing skin [1], muscle [12], bone [3] or nerv. Additionally, a variety of so-called composite flaps have been described, which include the transfer of more than one tissue structure, such as musculocutaneous-, osteomuscular- [4], neurocutaneous- [6], fasciocutaneous-, neuromusculocutaneous- or osteomusculocutaneous flaps. Since these techniques are limited by the anatomy of the vascular system, surgeons began to prefabricate flaps [7]: The blood supply of a defined tissue is delivered to an autologous, microsurgically-generated pedicle in order to subsequently transfer it as a free flap [18]. Engineered tissue of specific dimensions may be handled in a comparable manner due to the following fact: in vitro generated tissue constructs cannot be nourished sufficiently via diffusion [8, 13] and lack an intrinsic vascular system. The result is cell necrosis after implantation, primarily affecting the predisposed central areas of a seeded scaffold. Tissue Engineers solved this drawback by purpose fully implanting the scaffolds with contact to pedicles [23] in order to generate an intrinsic vascular system due to the physiological onset of neovascularization [11]. Consequently, the scaffold is equipped with a pedicle that allows the option of subsequent free transfer. Further investigations have shown that certain vessel configurations, such as the arteriovenous loop, have greater potential for neovascularization than an axial vessel bundle [25]. Consequently, today there are numerous models described which routinely combine tissue engineering with the technique of flap prefabrication [14, 24]. Our workgroup focuses on the field of cartilage tissue engineering, especially targeting external ear reconstruction [21]. Even though good results are currently achieved in our hospital using the technique of ear reconstruction with autologous rib cartilage [20], we are working towards the following objective: replacement of the presently used rib cartilage framework by a tissue engineered scaffold which has been seeded with in vitro expanded chondrocytes of the patient. An aspect of external ear reconstruction via cartilage tissue engineering that has been neglected thus far is the soft tissue situation. Patients often present with disadvantageous soft tissue conditions such as a deficiency of local skin, which is inadequate for coverage of an ear reconstruction framework. Previous manipulation, especially as a result of past operations, trauma or tumour defects, not only hinder such an intricate surgical outcome, as is the case with the outer ear, but also increases the risk of wound healing after surgical intervention. Presently, this local handicap for ear reconstruction may be minimized by additional reconstruction such as raising a superficial temporal fascia flap and skin grafts for adequate coverage [16]. However, frequently a disadvantageous soft tissue situation is the main contraindication for surgical intervention leading to prosthetic device.

The aim of this study was to establish an experimental model for prefabricated composite flaps containing a biomaterial construct and local full thickness skin coverage. The clinically relevant free transfer of the generated composite graft is of additional interest. Functional integrity of the transferred composite flap (skin coverage as monitor) at the recipient site was also investigated.

2 Methods

A rat model was established in order to investigate complex reactions within tissue engineered composite flap prefabrication that simulate clinical conditions.

3 Results

In a total of 8 rats, the step 1 procedure (generation of an arteriovenous loop and implantation of polyurethane scaffolds) was performed successfully. Postoperatively, all animals were monitored with a hand-held Doppler after 24 hours and 7 days: the sensor was placed on top of the implant including the microvascular loop after having detected the inguinal pulse of the contralateral side as reference. In all 8 animals a viable loop perfusion was observed. Due to wound healing disorders, 2 animals had to be excluded from the experiment. The remaining animals underwent the step 2 procedure. In 6 rats the tissue engineered composite construct was successfully transplanted as a free flap (Table 1). Postoperatively, the pulsating of the pedicles was persistently palpable. Prior to sacrifice, the cervical region was reopened in order to check on the composite grafts loop perfusion. One animal showed thrombosis in the distal loop ending and was excluded from the experiment. 5 animals showed good perfusion of the construct pedicles. Regarding the transplanted skin island perfusion, the following tendency was noted in terms of colour: after a maximum of 45 minutes of ischemia during successful flap anastomosis, no obvious reperfusion was noticeable. 6, 12 and 24 hours postoperatively, the prefabricated skin island remained pale with cyanotic margins; 48 hours postoperatively it showed extensive necrosis that persisted until the last control was taken 72 hours post transfer. Compared to the prefabricated skin islands, monitoring of the skin islands attached to the control-scaffolds never regained their preoperative, native colour. 24 hours after the step 2 operation, they too presented pale with the caudal half being notably darker. The subsequent colour monitoring showed continuous darkening towards necrosis, especially in the caudal half of the island. Colour monitoring of the full thickness skin graft that had been transferred from the cervico-nuchal region to the former prefabricated scaffold groin site showed the following: there was no major change in colour during the observation period (72 hours); its complexion constantly appeared pale with a red border along the caudal sector. In contrast to the skin islands attached to scaffolds, the solely transplanted skin showed remarkable contraction after 48 hours (ca. 1/3 loss of original surface), and 72 hours (ca. 1/2 of its original surface) postoperatively.

4 Discussion

In 6 animals a tissue engineered prefabricated composite construct was transferred as a free flap; thus the establishment of a clinically relevant model for the surgical transfer of prefabricated flaps was achieved. The microsurgical setup “femoral artery to common carotid artery” had a caliber discrepancy of approximately 1.5 and represents a common clinical situation. However, investigations into the functional integrity of the transferred composite flap showed that viability could not be achieved by the experimental setup used: the skin covering the scaffold was initially incised as a random-pattern skin flap with cranial basis. After the prefabrication period of 6 weeks, the skin island was assumed to gain its major blood supply by neovascularization from the arteriovenous loop. Although free microsurgical transfer was successful, the skin island appeared nonviable and finally necrotic. A variety of factors must be taken into account, excluding loss of the skin island due to missing pedicle perfusion: a pulse was palpable at all times. Furthermore, additional reopening of the cervical site prior to the animal's sacrifice verified viable loop conditions. In contrast, our expectation concerning the vital status and colouring of the skin island controls was confirmed. While the full thickness skin graft without underlying scaffold remained pale, the skin graft attached to the sole scaffold suffered from undersupply and turned necrotic. Considering the fact that tissue can only be supplied over a short distance (1–2 mm) by diffusion [8, 13], this effect was to be expected. In addition, to provide sufficient oxygen to mitochondria inside the cell, the minimum distance from the cell to the closest capillaries in many metabolically active tissues is rarely greater than 40 to 200 μm [5]. Therefore, the full thick skin supply may hardly be nourished solely by diffusion through the 2 mm thick scaffold. Hence, the experimental setup of this study is based on the principle of neovascularization: only when a sufficient number of vessels originating from the loop have perforated the porous scaffold, can the overlying skin be supplied exclusively by the loop after transfer. Regarding the study results, the question arises as to whether survival of the prefabricated skin islands failed due to insufficient neovascularization. One has to consider that the prefabrication period of 6 weeks was too short to bridge a scaffold of 2 mm thickness with newly developed vessels. Data concerning the time range necessary for successful harvesting of a prefabricated skin flap vary from 5 days (16) to 12 weeks (17). However comparable data on the neovascularization of tissue engineered composite flaps is very limited in the current literature. The concept of tissue engineered composite grafts, including full thickness skin coverage as performed in this investigation, is a new aspect in science, making the described rat model an innovation. The only comparable work was done by Walton and Brown scheduled 6 weeks to prefabricate a polytetrafluorethylene scaffold of similar dimension (2 mm thick, 2×2 cm wide) in the rabbit ear. However, Walton and Brown used an axial vessel bundle that today is known to be less effective for neovascularization compared to arteriovenous loops [25]. Moreover, unlike in our studies, Walton and Brown designed the composite graft consisting of scaffold plus split thickness skin, which was not put on the scaffold until after a prefabrication period of 6 weeks. By then the scaffold was placed subcutaneously. Split thickness skin viability was achieved 100% in 2 out of 8 implants and free flap transfer was performed after a prefabrication period of a total of 7 weeks. One of the two transplants soon suffered necrosis due to pedicle kinking. The other scaffold exposed to the surface within 4 weeks after transfer [26]. A similar concept was used by Hirasé et al. The major vessel bundle of a rabbit's pinna was dissected, ligated and, in contrast to our study, positioned between an autologous cartilage sheet and a full thickness skin sheet. Skin and cartilage had both initially been incised as random pattern flaps with a one-sided basis. After a prefabrication period of 3 weeks, the workgroup accomplished a free microvascular transfer of the composite flap to the contralateral side [10]. However, there is also a chance that 6 weeks exceeded the prefabrication period needed in our experiment. The rat is known for its surprising wound healing mechanisms. Therefore, the random pattern skin flap may have regained its original superficial blood supply due to accelerated healing despite the initial incision and therefore never reached the desired dependency on the underlying loop. One method of resolution might be using an additional delay with sole incision of the skin island prior to the free transfer. In this way, the necessary dependence of the full thickness skin island on the prefabricated loop through the scaffold might be guaranteed [9]. Nevertheless, the purpose of future experiments will be to examine how the factor of time influences the integration of a skin island into a tissue engineered construct. It also has to be taken into account whether the tissue passage “scaffold to skin” represents an aggravative condition for the neovascularization processes. Here, particularly inconstant initial contact between scaffolds and skin coverage may play a decisive role. However, discontinuous scaffold/skin contact is a concomitant phenomenon within postoperative mobilization and movement of the animals. Nevertheless, to minimize this handicap more attentive fixation (tissue glue or additional sutures) could be attempted in future experiments. Though the free flap ischemic period never extended the general valid critical time range of 60 minutes as determined by Scholz and Evans [19], dissecting the main vascular structures of a flap causes microvascular turbulence and influences the flap’s ischemic tolerance in a negative way. In the experiment described above, the prefabricated pedicle proved to be remarkably more sensitive to contact, and embedded more easily into scar tissue during the step 2 operation compared to its native appearance during the step 1 operation. Assumed that the dissection of the more sensitively prefabricated vessel had an impact on the construct’s microcirculation, does this mean that the tolerable ischemic period in this case was drastically shortened? If so, it may be advantageous to revert to strategies that have proven to sustainably optimize the vitality of transplanted flaps: for example, precondition a flap to the upcoming ischemic status [28] or performing a delay: in the context of the described study, additional incision of the skin islands prior to transfer might reinforce dependence on the underlying loop [22].

The skin color change noted in this study conforms to the criteria of venous congestion. Venous congestion is considered a typical and dreaded complication occurring within free flap transfer. Often its cause remains unclear, especially when anastomoses remain open [27]. Comparable to our experiment, Wayne Morrison et al. reported three cases with venous congestion occurring after transfer of a prefabricated skin flap. All flaps could only be preserved by additional treatment, such as the application of leeches/hirudo medicinalis or generation of a super-charge anastomosis [15]. However, Morrison et al. did not discuss potential causes for venous congestion in flap prefabrication. They simply pointed out that flap prefabrication involves a certain amount of unpredictability as far as vascularisation is concerned and therefore needs to remain a focus of future experimental investigations until it can be more successfully integrated into the standard repertoire of the clinical routine.