Poor wound-healing and skin necrosis are potentially devastating complications after total knee arthroplasty1,2. Necrosis of the soft tissues about the knee can rapidly lead to deep infection at the site of the arthroplasty; the need to replace the prosthesis; or, in extreme cases, amputation3. Wound dehiscence, necrosis at the edge of longitudinal skin incisions, or total exposure of the prosthesis are very often due to poor vascularization of the skin secondary to the presence of multiple scars on the lateral and medial surfaces of the knee from prior operations3-5. While some authors have recommended free tissue transfers or pedicled fasciocutaneous flaps, such as an anterolateral thigh flap6,7, for the treatment of periprosthetic infections of the knee8, local muscle flaps are considered the most useful for reconstruction in cases of an exposed prosthesis. Primary soft-tissue coverage with a medial or lateral gastrocnemius transposition flap has become the primary reconstructive method of choice5,9,10. The proximally based gracilis and semimembranosus muscles have been used for supplemental coverage after detachment of their tendon distally11.
The gracilis muscle has been used both as a free flap and as a proximally based pedicled flap, secondary to the adequacy of its proximal vascular pedicle and minimal associated morbidity12,13. A distally based gracilis muscle flap nourished by distal vascular pedicles has been proposed to cover areas of proximal tendon exposure in the knee14.
In cases where a lateral or medial gastrocnemius flap is not available, or to treat large defects of the soft tissues in combination with the use of a gastrocnemius flap, we proposed the use of an extended reversed gracilis flap based on secondary pedicles (the GReSP flap) for the reconstruction of large defects of the knee. The purpose of this report is to describe anatomical features of the gracilis muscle on the basis of radiographic studies and dissections of cadaver thighs. We also present the preliminary clinical results in three patients in whom severe soft-tissue complications of total knee arthroplasty were treated with the GReSP flap.
Anatomical Study
Ten thighs of ten fresh cadavers (five male and five female) were studied. The external iliac artery was selectively injected with 200 mL of acrylic resin (Batson's #17 anatomical corrosion compound; Polysciences, Warrington, Pennsylvania), 1 cm above the inguinal ligament. The adductor region of the thigh was exposed through three skin incisions15. The skin was then reflected, and the superficial and deep fascia were removed, thus exposing the gracilis muscle and isolating its vascular pedicles. After polymerization of the resin (thirty minutes), the gracilis muscle and its vascular pedicles were sampled (Fig. 1).
Radiographic Study
Computed tomographic angiograms of fifty patients (forty male and ten female) were randomly selected from the archive of a diagnostic center. The patients had undergone computed tomography examination because of vascular abnormalities involving the abdominal aorta. Computed tomography images were obtained with a sixteen-slice multidetector computed tomography scanner (LightSpeed 16; GE Healthcare Milwaukee, Wisconsin) (2.5-mm slice thickness, 27.5-mm/rotation speed, 120 kV, and 300 mA). Analysis and post-processing of computed tomography scans was carried out on an Aquarius Workstation (version 3.6.2.3; TeraRecon, San Mateo, California).
Computed tomographic angiograms of both lower limbs were analyzed, with the investigator focusing on the anatomical features of the gracilis muscle (the length of the muscle and its muscle belly—i.e., the length without the tendons) and its secondary vascular pedicles.
We recorded the number, origin, course, caliber, and distance from the pubis to the entry point into the muscle of these vascular pedicles. Only secondary pedicles identified on computed tomographic angiograms and clearly directed toward the gracilis muscle were examined (Fig. 2). Results are expressed as mean values (and standard deviation) and ranges.
A linear regression test was used to look for a correlation between the caliber of the proximal secondary pedicle and the number of secondary pedicles and between the caliber of the proximal secondary pedicle and the length of the part of the gracilis muscle distal to the entry point of this pedicle (measured as the difference between the length of the whole muscle and the distance from the pubis to the entry point of the proximal secondary pedicle). P < 0.05 was considered significant.
Clinical Study
On the basis of the anatomical features, an extended reversed gracilis flap based on its secondary pedicles was used for three patients with severe soft-tissue complications of total knee arthroplasty. A preliminary computed tomographic angiogram was obtained to assess the adequacy of the secondary pedicle—i.e., the caliber of the proximal secondary pedicle had to be >2 mm and its distance from the ischiopubic attachment (pivot point) had to be adequate for the arc of rotation of the flap. All of the patients provided informed consent for the surgical procedure. The demographic data and clinical histories of the patients are presented in Table I.
GReSP-Flap Surgical Technique
The patient was positioned supine, with the thigh abducted. The position of the gracilis muscle was identified by palpation, and the location of the main and secondary vascular pedicles, identified on the preoperative computed tomographic angiogram, were marked. A 10 to 15-cm sagittal skin incision was made at the anterior margin of the muscle, from the pubis to the location of the proximal secondary vascular pedicle, exposing the superficial aspect of the muscle. Particular attention was paid to the saphenous vein and nerve, which were protected (Fig. 3). Correspondence between the position of the vascular pedicles in the surgical field and that seen on the preoperative computed tomographic angiogram was verified. The muscular dissection was carried up to the entry point of the proximal secondary pedicle. The proximal tendon of the gracilis muscle was then transected to provide the maximal length for the muscle flap. The flap was then checked for perfusion, and the main vascular pedicle was clamped in order to verify that the blood supply to the proximal part of the muscle was sufficient when nourished only by the secondary pedicles. Then, the main neurovascular pedicles were ligated and transected. The muscle flap was reversed 180° and tunneled under the residual skin of the anterior aspect of the distal part of the thigh, in order to reach the exposed prosthesis. The proximal secondary pedicle was dissected as needed up to its origin from the superficial femoral artery, in order to achieve better mobility of the flap. The perfusion of the muscle was then evaluated intraoperatively with Doppler ultrasound, and the flap was sutured in place and covered with a split-thickness skin graft.
Source of Funding
There was no external source of funding for this study.
The gracilis muscle was identified in all patients—i.e., both in the cadaver dissections and on the computed tomographic angiograms. The muscle had a mean length of 41 ± 2.1 cm (range, 37 to 45 cm) in the series as a whole, 42 ± 1.8 cm in males, and 39 ± 1.6 cm in females. The muscle belly had a fusiform shape, sagittally oriented in the medial aspect of the thigh. It had a mean length of 30 ± 2.1 cm (range, 27 to 34 cm) in the series, 31 ± 1.8 cm in males, and 28 ± 1.5 cm in females; a mean anteroposterior dimension of 44 ± 1 mm (43 ± 18 mm in males and 45 ± 15 mm in females), and a mean mediolateral dimension of 11 ± 2 mm (10 ± 18 mm in males and 13 ± 15 mm in females).
The anatomical dissections showed that the main pedicle contained one artery, two venae comitantes, and a branch of the obturator nerve, whereas the secondary pedicles had only one artery and venae comitantes, without a nerve (Fig. 1).
On the computed tomographic angiograms, each patient was seen to have at least one vascular pedicle distal to the main one. These secondary pedicles varied in number, with males having one to four (mean, 1.8) and females having one or two (mean, 1.5). These pedicles originated from the superficial femoral artery or the popliteal artery and reached the muscle by passing between the sartorius and the adductor longus muscle. The most proximal secondary vascular pedicle was often the largest, with a mean caliber of 2.0 mm (2.11 in males and 1.75 in females) (Fig. 2). A significant positive association (p = 0.001; r2 = 0.49; slope [and standard error of the mean] = 1.1 ± 0.3 other pedicles/mm of proximal secondary pedicle caliber) was found between the caliber of this proximal secondary pedicle and the number of other secondary pedicles. An association also was documented between the caliber of this proximal secondary pedicle and the length of the part of the gracilis muscle distal to the entry point of this pedicle (p = 0.05; r2 = 0.23; slope [and standard error of the mean] = 59 ± 28 mm of gracilis muscle/mm of proximal secondary pedicle caliber).
The distance between the entry point of the pedicle into the gracilis muscle and the pubis was a mean of 21 ± 3.6 cm (range, 16 to 28 cm) in the series as a whole, 22 ± 3.9 cm in males, and 18 ± 2.4 cm in females. When this point was considered in relation to the muscle length, it was found to be at the midpoint (as determined by dividing the entry point by the muscular length) and to be located between one-third and two-thirds of the muscle length in 95% of the cases.
Other distal secondary pedicles, when present, were smaller, arose from the superficial femoral artery or popliteal artery, and reached the distal part of the muscle belly, or its distal tendon, at a mean distance of 26 ± 2.7 cm (27 ± 0.9 cm in males and 26 ± 1.4 cm in females) from the pubis.
Clinical Study (Table I)
Case 1 (Fig. 4)
Seven days after delayed reimplantation of a knee prosthesis and contemporary reconstruction of the capsule and patellar tendon with use of an artificial polyethylene terephthalate tendon implant, the patient developed a large necrotic area of skin over the anterior surface of the knee. Surgical debridement completely exposed the prosthetic tendon and prosthetic implants.
After extensive debridement of the prosthesis, a medial gastrocnemius muscle flap nourished by the ipsilateral sural vascular pedicle was transposed to cover the distal portion of the defect. Because of the insufficient coverage by the medial gastrocnemius flap, a GReSP flap was developed and tunneled under the anteromedial skin of the thigh to reconstruct the proximal half of the defect. A skin graft was applied to the muscle surface, and the knee was immobilized with a rigid orthosis.
Systemic antibiotic therapy was administered for two weeks after the surgical procedure. The skin graft survived completely, with the flap being fully viable. Evaluation at twenty-two months revealed stability of the reconstruction without evidence of infection.
Case 2
Ten days after delayed reimplantation of a knee prosthesis, the paramedian skin incision developed a central dehiscence, with necrosis at the skin edges. Surgical debridement exposed a partially necrotic patellofemoral tendon and the prosthetic implant on the anterior aspect of the knee.
The gracilis muscle flap was mobilized and tunneled under the anteromedial skin of the thigh. The flap survived completely, and the patient was discharged after two weeks to begin physiotherapy. Systemic antibiotic therapy was administered for two weeks after the surgical procedure. At twenty-four months, complete soft-tissue healing had occurred.
Case 3
Six months after delayed reimplantation of a knee prosthesis, secondary stiffness of the knee joint prompted surgical revision to improve the range of motion. The patient developed a large area of necrotic skin over the anterior aspect of the knee. An irregular, oval 4 × 10-cm ulcer, oriented along the axis of the limb, was present on the anterolateral surface of the knee, exposing the patellar tendon and the total knee prosthesis. Following debridement, a GReSP flap was used to cover the defect, and the knee was immobilized with a rigid orthosis. Systemic antibiotic therapy was administered for two weeks after the surgical procedure. At twenty-four months, the soft tissues were healed without evidence of infection.
Skin necrosis following total knee arthroplasty is a serious complication and one of the most difficult challenges for surgeons who perform knee joint reconstructions. Early reconstructive intervention is the best way to reduce the risk of deep infection, removal of the prosthetic implant, or amputation16-18. The presence of multiple scars over the knee surface and related poor skin vascularization often necessitate the use of a pedicled muscle flap from the leg19,20 to cover such lesions, instead of using local fasciocutaneous flaps or tissue expanders21. The gracilis muscle has often been used as a free and proximally based pedicled flap20,22-25. We believe that the use of the gracilis muscle as a distally based flap nourished by secondary vascular pedicles can be an effective alternative, or addition, to a medial or lateral gastrocnemius flap.
We are not aware of any reports regarding the use of the gracilis muscle based solely on its distal pedicles for the treatment of extensive soft-tissue defects of the knee. Recently, Cavadas et al.14 presented the preliminary clinical results of the use of a reversed gracilis muscle flap in two patients with small defects of the quadriceps tendon, but the gracilis muscle reached only the suprapatellar region.
Only a few studies have focused on the secondary vascular pedicles originating from the superficial femoral artery or the popliteal artery26. The number of these pedicles has been reported to range between one and five27,28. In our study, the number ranged from one to four, with a mean of 1.8. Taylor et al.25 reported that the mean caliber of the proximal secondary vascular pedicle was 1.1 mm, whereas in our study it was 2.0 mm. This difference may be due to the fact that the study by Taylor et al. was based only on cadaver dissection, while our study included living patients who had undergone computed tomographic angiography.
The positive association between the caliber of the proximal secondary pedicle and the number of secondary pedicles confirms the importance of this pedicle in nourishing the gracilis muscle, especially its distal portion. A significant association between the caliber of this artery and the length of the muscle belly distal to the entry point of the pedicle was also found.
It should be noted that both Taylor et al.25 and Cavadas et al.14 demonstrated extensive intramuscular connections between the dominant and secondary pedicles through angiographic study of the gracilis muscle. In particular, Cavadas et al. showed that the proximal secondary pedicle usually divides into two branches, one ascending branch that anastomoses to descending branches of the dominant pedicle and a lesser descending branch that anastomoses to the distal secondary pedicle, with consequent retrograde flow. These observations suggest the importance of the secondary pedicles in nourishing not only the distal part of the muscle, but also the proximal part, which is classically considered to be perfused only by the main vascular pedicle. The mean 2-mm caliber of the secondary pedicles and the good vascularization of the muscle demonstrated intraoperatively after clamping of the principal pedicle in all three of our patients confirmed the feasibility of using the gracilis as a distally based pedicled flap.
In the study by Taylor et al.25, the mean entry point of the proximal secondary vascular pedicle into the muscle was 21.8 cm from the ischiopubic branch. In our study, the pedicle was located at a mean of 21 ± 3.6 cm from the proximal insertion of the muscle, and a more distal secondary vascular pedicle, which was usually present, was found at a mean of 26 ± 2.7 cm. The distance of the proximal secondary pedicle from the ischiopubic branch represents the length of the muscle belly that is suitable for rotation and transposition as a muscular flap—i.e., it represents its pivot point. It is important to determine this distance preoperatively with computed tomographic angiography when a large, or very distal, defect must be covered, in order to ensure adequate length of the pedicle flap. To enhance rotation of the muscle, we suggest careful release of the secondary pedicle as far as the superficial femoral artery.
Multidetector computed tomographic angiography has emerged as a noninvasive operator-independent choice for evaluating the location and size of small vessels29-33. Our radiographic study shows that computed tomographic angiography is a useful, noninvasive preoperative tool, which can supply detailed information on each patient. These findings assist the surgeon in selecting patients for whom this kind of flap is suitable, providing a detailed description of the patient's vascular anatomy, and minimizing the time of operative dissection.
From these studies, we can conclude that the GReSP flap may be considered as an alternative, or an addition, to a gastrocnemius flap for soft-tissue reconstruction over an exposed total knee prosthesis in patients with suitable vascular anatomy as evaluated with computed tomographic angiography.
Note: The authors are grateful to Gloria Sarasin and Anna Rambaldo for their skillful technical assistance. The authors also thank Euganea Medica for the radiographic study.