Vascularized Bone Graft from the Iliac Crest for the Treatment of Nonunion of the Proximal Part of the Scaphoid with an Avascular Fragment*

GABL, MARKUS M.D.†; REINHART, CLAUDIA M.D.†; LUTZ, MARTIN M.D.†; BODNER, GERD M.D.‡; RUDISCH, ANSGAR M.D.‡; HUSSL, HERIBERT UNIV.PROF., M.D.§; PECHLANER, SIGURD UNIV.DOZ., M.D.†, INNSBRUCK, AUSTRIA

Journal of Bone & Joint Surgery - American Volume:
Article
Abstract

Background: It was hypothesized that nonunion of the proximal third of the scaphoid associated with avascular necrosis could be treated successfully with a free vascularized bone graft obtained from the iliac crest. Methods: Fifteen patients who had a nonunion of the proximal part of the scaphoid that had been present for an average of two years and three months (range, nine months to seven years) were managed with use of a free vascularized bone graft obtained from the iliac crest. Avascularity of the scaphoid, as assessed on preoperative radiographs, was characterized by loss of trabecular structure, collapse of subchondral bone, and formation of bone cysts. The results of the procedure were assessed in terms of osseous union, pain, active motion of the wrist, and osteoarthritis. Postoperatively, vascularity of the scaphoid was evaluated with use of magnetic resonance imaging and color Doppler ultrasonography. The average duration of follow-up was six years and one month (range, two years and one month to eight years and one month). Results: Preoperatively, one patient had had pain with any movement of the wrist and fourteen had had pain after strenuous manual labor or sports activity. The average pain score, derived with use of a 10-point visual analog scale, was 2.4 points (range, 1.0 to 6.7 points). Postoperatively, union was achieved in twelve patients; six were pain-free, and six had occasional pain during strenuous manual labor or sports activity, or both. The average pain score for these twelve patients was 1.1 points (range, 0.0 to 4.2 points) on the visual analog scale. Preoperatively, osteoarthritis was limited to the region between the radial styloid process and the distal part of the scaphoid in fourteen patients and to the radioscaphoid region in one patient. Postoperatively, the degree of osteoarthritis remained unchanged in seven of the twelve patients who had union and progressed to the radioscaphoid region in five. Vascularity, as seen on the imaging studies, was restored in all twelve patients who had union. The nonunion persisted in three patients, all of whom had progressive osteoarthritis leading to carpal collapse. Conclusions: The index procedure was successful in twelve of the fifteen patients who had a symptomatic nonunion of the proximal part of the scaphoid associated with avascular necrosis and osteoarthritis that was limited to the radioscaphoid joint.

Author Information

†University Hospital of Traumatology, Anichstrasse 35, A-6020 Innsbruck, Austria. E-mail address for Dr. Gabl: m.gabl@uibk.ac.at.

‡University Hospital of Radiology, Anichstrasse 35, A-6020, Innsbruck, Austria.

§University Hospital of Plastic and Reconstructive Surgery, Anichstrasse 35, A-6020, Innsbruck, Austria. E-mail address for Dr. Hussl: heribert.hussl@uibk.ac.at.

Article Outline

The ideal treatment of nonunion of the scaphoid remains unresolved and controversial. Displaced fracture fragments or soft-tissue interposition can prevent union of acute fractures of the proximal third of the scaphoid by interrupting the somewhat tenuous blood supply3,6.

Osteoarthritis may become progressively worse in association with long-standing nonunion of the scaphoid. Watson and Ryu52 reviewed the radiographs of 210 wrists and noted that the early changes of osteoarthritis were localized to the region between the radial styloid process and the distal part of the scaphoid and that the changes progressed in the midcarpal joint, leaving the proximal fragment unaffected. This sequence of changes has been reported in association with untreated symptomatic and asymptomatic nonunions of the scaphoid16,27,33.

Avascularity of the proximal fragment appears to be associated with the development of osteoarthritis. Ruby et al.43 observed the sequence of arthritic changes in thirty-one patients who had a nonunion of the scaphoid and noted that the one patient in whom osteoarthritis developed within four years after the injury also had avascular necrosis. Inoue and Sakuma27 reported that avascular necrosis developed earlier in patients who had a nonunion of the proximal third of the scaphoid than it did in those who had a nonunion of the middle third.

The rate of osseous union seems to depend on the stability and vascularity of the proximal fragment. Alnot et al.1, in a study of seven patients who had a nonunion of the proximal pole, reported that union was obtained with use of dorsal screw fixation when the proximal fragment was viable. Herbert22 reported similar results when the proximal fragment was viable.

The unfavorable results20,25,39 that have been reported in association with the use of autogenous bone grafts for the treatment of nonunions of the scaphoid that are complicated by an avascular proximal fragment led to the development of alternative methods, including replacement with a prosthesis54 or an allograft11 as well as salvage procedures such as resection arthroplasty and arthrodesis of the wrist10,22. Some investigators have suggested that vascularization may be enhanced with implantation of a vascular pedicle along with an autogenous bone graft18,24 or with use of a free vascularized bone graft8,12,21,30,35,38,53.

On the basis of an unreported observation by Russe and one of us (S. P.) that avascularity was the main cause of failure of Russe inlay bone-grafting, one of us (S. P.) and colleagues38 developed a technique in which a free vascularized bone graft from the iliac crest is used to repair an avascular nonunion of the scaphoid. We report the results of this procedure in fifteen patients.

Back to Top | Article Outline

Materials and Methods

Between 1984 and 1998, a free vascularized bone graft was used to restore function of the wrist in 157 patients who had a variety of conditions. In fifteen patients, the index procedure was performed because of a symptomatic nonunion of the proximal third of the scaphoid with an avascular fragment. The average age of the thirteen male and two female patients was thirty years (range, seventeen to fifty-four years), and the nonunion had been present for an average of two years and three months (range, nine months to seven years). Nine patients worked in road or railroad construction, forestry, glass manufacturing, or farming; the remaining six had a sedentary job in administration or management. The average duration of follow-up after the index procedure was six years and one month (range, two years and one month to eight years and one month).

All patients had a history of a dorsiflexion injury of the wrist. Ten patients had sustained the fracture in a simple fall; four, during a sports activity; and one, in a traffic accident involving a motorbike. The dominant wrist was affected in eleven patients.

Ten patients did not seek medical treatment immediately after the injury and were managed unsuccessfully with various nonoperative measures. In one of these patients, the fracture was missed on the initial radiographs. Four other patients were managed soon after the injury with use of a below-the-elbow cast that incorporated the thumb; the cast was worn for twelve to fourteen weeks. The remaining patient was managed with a Russe inlay bone-grafting procedure that failed to lead to union.

Before the index procedure, all patients had intermittent pain that was localized to the radial aspect of the wrist. One patient had pain with any movement of the wrist; the remaining fourteen patients had pain only after strenuous manual labor or sports activity. (Younger patients had pain primarily during sports activities such as gripping a ski pole, snowboarding, riding downhill on a mountain bicycle, or weight-lifting, whereas older patients had pain chiefly during strenuous manual labor.) In all patients, the pain was sporadic and was present only during the day.

Standardized radiographs were made in the posteroanterior, lateral, and oblique projections, as described by Böhler6 and by Schernberg45. Avascularity was characterized by loss of trabecular structure, collapse of subchondral bone, and formation of bone cysts. Patients who had evidence of irregular areas of patchy sclerosis in the proximal fragment that were suggestive of partial necrosis were excluded from the study.

Osteoarthritis was classified according to three grades. Grade 1 indicated slight narrowing of the joint space; grade 2, marked narrowing of the joint space and formation of osteophytes; and grade 3, complete loss of the radiocarpal joint space and formation of osteophytes as well as subchondral bone cysts. Progression of osteoarthritis was defined on the basis of the criteria proposed by Watson and Ryu52. In stage 1, there are degenerative changes between the radial styloid process and the distal part of the scaphoid; in stage 2, the changes extend proximally to the radioscaphoid joint; and in stage 3, there are changes between the capitate and the scaphoid and between the capitate and the lunate.

Back to Top | Article Outline
Operative Procedure

Two of us (H. H. and S. P.) performed dissection simultaneously (one performed it at the iliac crest and the other, at the wrist), which reduced the duration of the procedure to four and one-half hours. The procedure was performed with the patient under general anesthesia and in the supine position. The bone graft was removed through an incision made parallel to the inguinal ligament, beginning four centimeters medial to the anterior superior iliac spine and extending ten centimeters along the iliac crest (Fig. 1-A). The deep circumflex iliac vessels were identified beneath the internal oblique muscle, and the three main branches (the ramus profundus, ramus ascendens, and ramus horizontalis) were isolated carefully. These fine vascular branches were identified in the fascia with use of loupe magnification and were traced along their course to the iliac crest. A three to four-centimeter-wide fascial layer that included these branches was dissected and removed along with a tricortical segment of bone from the iliac crest. The operative time for one of us (H. H.) to obtain the vascularized pedicle bone graft was one and one-half hours. The bone graft was approximately sixteen millimeters long and five millimeters wide, with a height of four millimeters proximally and ten millimeters distally. The vascularized pedicle was approximately five centimeters long. The dimensions and shape of the graft varied on the basis of anatomical differences among individuals. The vascularity of the graft was confirmed by the presence of bleeding from the cut ends.

The scaphoid was exposed, under tourniquet control and with use of loupe magnification, through a volar incision that began at the tubercle of the scaphoid, extended proximally along the tendon of the flexor carpi radialis, and ended three centimeters proximal to the most proximal palmar wrist crease. The radial artery and its branches were protected during incision of the wrist capsule, with care being taken to leave the palmar ligaments intact. The fibrous nonunion was debrided, and the necrotic bone of the proximal pole was excised. The proximal pole was excavated, leaving the entire articular cartilage and a thin layer of subchondral bone undisturbed (Fig. 1-B). The distal part of the scaphoid was curetted, with care being taken to remove only the amount that was necessary in order to obtain a vascularized tissue bed for the graft.

The graft, which was shaped to fit the contours of the cavity in the scaphoid, was placed with the cortical surface on the palmar aspect to maintain reduction and alignment of the fragments and was transfixed with a one-millimeter-diameter pin. The deep circumflex iliac bundle was anastomosed end-to-side to the radial artery and end-to-end to the venae comitantes (Fig. 1-C). We believe that the wide capillary network in the fascial layer, which was removed along with the bone graft, provided sufficient arterial-to-venous flow to keep the vessels from clotting postoperatively.

Back to Top | Article Outline
Postoperative Regimen

To monitor arterial flow in the pedicle, Doppler ultrasonography was performed postoperatively every three hours for the first two days and then daily for five days.

The extremity was immobilized in an above-the-elbow cast that incorporated the thumb. The cast was exchanged at four weeks for a below-the-elbow cast, which was worn for an additional eight weeks. The cast was changed once during this eight-week period. Patients were allowed to use a computer keyboard and to perform nonstrenuous activities. The pin was removed at twelve weeks, with use of local anesthesia, to avoid irritation of the skin when physiotherapy was started. Radial deviation and sagittal flexion and extension were restricted if radiolucent zones, indicative of poor bone-healing, were seen on radiographs, particularly around the distal part of the graft. Patients were allowed to use the involved hand for daily activities when radiographs showed bone trabeculae bridging the two fragments. They were instructed to avoid strenuous manual labor and sports activity for six months.

The parameters that were evaluated included pain, active range of motion of the wrist, grip strength, and radiographic evidence of union and osteoarthritis. Pain was assessed with use of a visual analog scale that ranged from 1 point (no pain) to 10 points (severe pain). In addition, patients were asked whether the pain was present during the day or at night, how long it lasted, and whether it was related to activity. Active range of motion of the wrist in the sagittal and frontal planes, rotation of the forearm, and grip strength were measured by a physiotherapist with use of a BTE primus device (Baltimore Therapeutic Equipment, Baltimore, Maryland). Similar measurements were performed on the uninvolved wrist.

Radiographs were made every four weeks for the first six months to assess union and to determine the degree of motion that could be permitted. Motion of the wrist was permitted when there was no evidence of a gap at the site of the graft. Radiographs also were made one year postoperatively and at the latest follow-up evaluation. Union was considered to be present when there was no gap at the site of the former nonunion and bridging trabeculae were seen on all radiographs. Progression of osteoarthritis was assessed with use of the staging system of Watson and Ryu52. The osseous vascular pattern was investigated with magnetic resonance imaging performed with use of gadopentetate dimeglumine and was described as either showing or not showing enhancement. The examination was performed with the use of a 1.5-tesla superconductive Magnetom unit (GBS2; Siemens, Erlangen, Germany).

Color Doppler ultrasonography was performed by a radiologist (G. B.) to evaluate arterial and venous flow in the vascular pedicle. The examination was performed with use of a ten-megahertz linear-array probe (HDI 3000 or HDI 5000; Advanced Technologies Laboratories, Bothell, Washington) at a Doppler frequency of six hertz. A silicone pad was used as an interface between the probe and the site being evaluated. The radial artery and the pedicle were assessed on longitudinal and transverse views according to color flow and the Doppler spectral wave. The vascular pathway was considered to be open when the spectral wave showed systolic and diastolic flow and a resistive index of less than 0.8, and it was considered to be closed when there was only systolic flow and a high resistive index of 1.0.

Back to Top | Article Outline

Results

Avascular necrosis was confirmed intraoperatively by the absence of punctate bleeding points after débridement of the proximal fragment of the scaphoid20. Histopathological examination of spicules of bone obtained from several sites in the proximal fragment revealed fibrocytic infiltration and bone-remodeling suggestive of avascular necrosis7.

Union was achieved in twelve patients (Figs. 2-A and 2-B and Table II). Postoperatively, six of the twelve patients were pain-free and the other six had intermittent pain during strenuous manual labor or sports activity. The average pain score for these twelve patients was 1.1 points (range, 0.0 to 4.2 points) on the visual analog scale.

In the twelve patients who had union, extension of the wrist averaged 58 degrees (range, 40 to 75 degrees); flexion, 47 degrees (range, 35 to 50 degrees); radial deviation, 19 degrees (range, 10 to 30 degrees); and ulnar deviation, 34 degrees (range, 25 to 50 degrees). Grip strength averaged 107 kilograms-force (range, seventy to 120 kilograms-force). The osteoarthritis remained unchanged at stage 1 in six of the twelve patients and unchanged at stage 2 in one patient; in the remaining five patients, the osteoarthritis progressed from stage 1 to stage 2 (Tables I and II).

Magnetic resonance images that were made at the latest follow-up evaluation showed enhancement of the united bone in all twelve patients who had union, although the density of the contrast agent was decreased compared with that of the adjacent carpal bones (Fig. 2-C). Color Doppler ultrasonography revealed open flow in the vascular pedicle. New-bone formation was observed on the radiopalmar aspect of the wrist, in the region of the periosteal layer of the vascularized graft, in eleven patients.

Three patients had a persistent pseudarthrosis. In the first patient (Case 1), the postoperative position of the vascularized graft within the proximal fragment was incorrect; the fragment became displaced, and the vascularized graft was resorbed. In the second patient (Case 11), who initially had a comminuted fracture of the scaphoid, the vascularized graft consolidated distally but the proximal segment remained ununited. In the third patient (Case 12), the proximal fragment was not well reduced and became displaced (Figs. 3-A and 3-B). None of these three patients had a reoperation. All three had pain during loading of the wrist. In one of these patients the pain was unrelated to activity, and in the other two it was related to strenuous manual labor or sports activity, or both. The pain often lasted for several days and occurred intermittently. These three patients had a pain score of 1.8, 5.4, and 5.6 points, according to the visual analog scale.

In the three patients who had a persistent nonunion, extension of the wrist averaged 28 degrees (20, 30, and 35 degrees); flexion, 33 degrees (30, 30, and 40 degrees); radial deviation, 13 degrees (10, 10, and 20 degrees); and ulnar deviation, 30 degrees (20, 30, and 40 degrees). Grip strength averaged seventy-three kilograms-force (fifty, seventy, and 100 kilograms-force). The osteoarthritis in these three patients progressed from stage 1 to stage 3 (carpal collapse). Magnetic resonance imaging studies showed no evidence of enhancement in two patients (Cases 1 and 12) and demonstrated patchy enhancement in the third (Case 11), in whom the excavated proximal fragment and the graft could not be placed in a stable position. Ultrasonography showed the pedicle to be closed in two patients (Cases 1 and 12) and open in one (Case 11).

There were no instances of infection or late pain at the donor site over the anterior iliac crest. One patient had a hypertrophic scar. All patients had wrist pain during the first three weeks after the procedure. There were no neurological problems.

The vascular anastomosis failed postoperatively in two patients (Cases 1 and 12), possibly because the bridging fascial layer between the bone and the vessels was too long and there were only a few bleeding points on the graft. We believe that the fascial bridge did not include a sufficient number of feeding vessels and that a shorter capillary distance between the bone and the blood vessels will provide a better blood supply. There was no evidence of cancellous bone-bridging in these two patients, and the persistent nonunion led to carpal malalignment, subluxation of the graft, and progression to stage-3 osteoarthritis (carpal collapse)52.

Back to Top | Article Outline

Discussion

Fractures of the scaphoid often are unrecognized after an injury of the wrist and may remain ununited. However, nonunion can occur even if the fracture is recognized promptly and is treated appropriately. Cooney et al.14 suggested that nonunion may be related to a number of factors, including a delay in diagnosis or treatment, the tenuous blood supply of the scaphoid, displacement and angulation of the fracture fragments, and carpal instability.

Mack et al.32, in a study of forty-six patients (forty-seven wrists), described the natural history of an untreated symptomatic nonunion of the scaphoid. The duration of the nonunion ranged from five to fifty-three years. Those authors observed that the osteoarthritis initially was confined to the scaphoid; it then extended to the radioscaphoid joint and, finally, to the entire wrist. In most patients, the osteoarthritis was debilitating. Ruby et al.43 described the natural history of an untreated nonunion of the scaphoid in a study of fifty-five patients (fifty-six wrists) and noted that osteoarthritis developed in thirty-one (97 percent) of thirty-two patients who had been injured at least five years earlier. The one patient in whom osteoarthritis had developed less than four years after the injury also had avascular necrosis. Those authors also noted that, when dorsal intercalated-segment instability was present, osteoarthritis developed between the scaphoid and the capitate and between the scaphoid and the lunate in addition to the region between the distal pole of the scaphoid and the radius. Inoue and Sakuma27 reviewed the cases of 102 patients (104 symptomatic nonunions) and noted osteoarthritis in all patients who had a nonunion of more than ten years' duration. Those authors found no relationship between pain and the severity of the osteoarthritis, but there was a relationship between the duration of the nonunion and reduced grip strength or decreased motion of the wrist. Lindström and Nyström31 reviewed the cases of thirty-three patients who had had an asymptomatic nonunion for ten to seventeen years. All patients had evidence of osteoarthritis, and the authors concluded that osteoarthritic changes are progressive even in asymptomatic patients.

The Russe inlay bone-grafting procedure is considered to be a reliable method for the treatment of a symptomatic nonunion of the scaphoid. Russe44 reported union in twenty of twenty-two patients who had been managed with bone graft from the iliac crest. The palmar approach minimized damage to the blood supply and also facilitated correction of any flexion deformity at the fracture site. Russe inlay bone-grafting has resulted in consistently high rates of union. Jiranek et al.28 reported union after this procedure in twenty-one of twenty-six wrists that were followed for seven to eighteen years; Stark et al.49, in 147 of 151 wrists that were followed for one to ten years; and Cooney et al.13, in thirty-eight of forty-four wrists that were followed for twelve to 163 months. Hooning Van Duyvenbode et al.23 reported the clinical and radiographic findings for seventy-four patients who were managed with a Matti-Russe procedure. The duration of follow-up ranged from twenty-two to thirty-four years. There was satisfactory relief of pain and stiffness, but some patients continued to have limited motion and grip strength as well as osteoarthritis. One of us (S. P.) and colleagues39 suggested that the rate of union was related to the duration of the nonunion. Those authors reported union in fifty-three of sixty-one wrists that had had a nonunion for twelve months or less, in nineteen of twenty-two that had had a nonunion for thirteen to twenty-four months, in fourteen of eighteen that had had a nonunion for twenty-five to forty-eight months, and in only six of eleven that had had a nonunion for forty-nine months or more.

Amadio et al.2 described a so-called flexion or humpback deformity of the scaphoid as seen on a lateral tomogram and found that an intrascaphoid angle of more than 45 degrees was associated with an increased risk of pain and osteoarthritis. They noted that union was an insufficient criterion of success in the treatment of scaphoid fractures. Fernandez17 suggested that an anterior bicortical wedge-shaped graft could be used to correct this deformity and to prevent osteoarthritis. That author reported that union was obtained in all six of his patients, who were managed with resection of the pseudarthrosis, insertion of an iliac-crest graft, and internal (pin) fixation. The average duration of follow-up was five years (range, 2.5 to eleven years). Four nonunions were located at the waist of the scaphoid; one, in the distal third; and one, in the proximal third. Two nonunions were associated with avascular necrosis in the proximal fragment.

Screw fixation has been used to increase the stability of bone fragments and bone grafts. Warren-Smith and Barton51 compared the results of Russe bone-grafting with those of Herbert-screw fixation in fifty patients who had a nonunion of the scaphoid. Union was achieved in seventeen of the twenty-eight patients who had been managed with Russe bone-grafting and in all twenty-two patients who had been managed with a wedge-shaped bone graft and fixation with a Herbert screw. The high rate of failure associated with the Russe method was believed to be related to the fact that the two small corticocancellous grafts were not inserted with their cortices facing outward as recommended by Russe and also to the use of a more rigid definition of union. Nakamura et al.37 reported union in all eight of their patients, who had a symptomatic nonunion with dorsal intercalated-segment instability that had been treated with open reduction, insertion of an anterior wedge-shaped bone graft, and fixation with a Herbert screw. Cooney et al.14 reported union in fifteen of twenty-one patients in whom an unstable fracture of the scaphoid had been treated with a wedge-shaped bone graft and fixation with a Herbert screw.

Russe44 stated that his procedure is relatively contraindicated for patients in whom the proximal pole of the scaphoid is totally avascular. One of us (S. P.) and Beck40 noted a persistent nonunion in twenty of eighty-six patients who had been managed with the Russe-I procedure, three of eighteen who had been managed with the Russe-II procedure, three of eighteen who had been managed with Herbert-screw fixation, and three of thirty-five who had been managed with a free vascularized iliac-crest graft. The nonunion was thought to be due to avascularity if it had been present for more than four years and to decreased stability after débridement especially if it was located at the proximal pole. Cooney et al.14 reported resorption of the wedge-shaped bone graft in four of twenty-one patients; all four of these patients had associated avascular necrosis. Richards and Regan42 reported a lack of punctate bleeding intraoperatively in two of ten patients who had a persistent nonunion after bone-grafting and Herbert-screw fixation. Smith and Cooney48 evaluated nineteen patients who had a persistent nonunion after bone-grafting and identified only four who had avascular necrosis and could be managed with a vascularized graft. Green20 reported a persistent nonunion in all five of his patients in whom the proximal fragment was totally avascular.

Viability and stability of the fragments are essential for obtaining union of a pseudarthrosis. The current practice is to treat avascular necrosis by increasing the vascularity of the bone. One of us (H. H.) and colleagues26, in a study of ninety-four rabbits, found that cartilage and skin can be revascularized with use of a muscular-vascular pedicle. Burchardt9 reported that cancellous and cortical bone grafts have three histological differences: (1) cancellous bone grafts are revascularized more rapidly and completely than are cortical bone grafts; (2) cancellous bone is replaced by a process of creeping substitution in which appositional bone formation is followed by resorption, whereas cortical bone is replaced by a process in which osteoclastic activity along the haversian systems is followed by osteoblastic activity, resulting in the apposition of new bone; and (3) cancellous bone grafts usually are completely reconstituted, whereas cortical bone grafts remain an admixture of necrotic and viable bone after the catabolic and anabolic stages of repair have been completed. It is uncertain whether enough osteogenic bone substance survives after a nonvascularized bone-grafting procedure. Berggren et al.4, in a study of twenty-six mongrel dogs, investigated the effect of prolonged ischemia on the survival of osteocytes and osteoblasts in composite grafts that were revascularized by microvascular anastomoses and found that the cells survived for as long as twenty-five hours. Therefore, a poorly vascularized bed can be reconstituted with use of a bone graft composed of living cells.

Laboratory studies of vascularized and nonvascularized grafts have demonstrated that the former are associated with earlier union as well as with greater strength and stiffness between six weeks and six months postoperatively. Fujimaki and Suda19, in an experimental study of animals and a clinical study of thirty-two patients, evaluated the hypertrophy of vascularized bone grafts and noted that the presence of reactive bone formation just beneath the periosteum depended on the vascularity of the graft. However, the exact mechanism of bone formation was not described.

Shaffer et al.47, in a controlled experiment involving forty-five beagle dogs, found that bone repair occurred more rapidly in association with vascularized grafts than it did in association with nonvascularized grafts, although the patterns of bone repair were similar. A four-centimeter-long vascularized ulnar autogenous graft was removed along with a one-millimeter vascular bundle from one forelimb, and a nonvascularized ulnar graft was removed from the contralateral forelimb. The bones were reimplanted and stabilized with bone sutures. The dogs were given fluorochrome markers, and serial analyses were performed at one week, six weeks, three months, six months, and one year. Torsional strength and stiffness were assessed with a servohydraulic testing machine, and microradiographs of transverse sections were analyzed. The strength and stiffness of the vascularized bone grafts were greater than those of the nonvascularized bone grafts after six weeks, suggesting rapid repair.

Moore et al.36 compared the biomechanical behavior of vascularized and nonvascularized grafts in nine dogs in which autogenous rib grafts were used to reconstruct ulnar defects. Dell et al.15 evaluated the characteristics of vascularized and nonvascularized fibular grafts in forty-nine adult dogs at two, six, twelve, and twenty-four weeks. Bone repair was assessed with use of tetracycline-labeling. The nonvascularized grafts healed by means of peripheral and internal resorption followed by callus formation and remodeling, whereas the vascularized grafts did not show extensive resorption, callus formation, or remodeling. Strength was assessed by measuring torsional load, and porosity was determined with microradiographs. The vascularized grafts initially were stronger, but after six weeks they were equal in strength to the nonvascularized grafts. Dell et al. concluded that vascularized grafts might be more effective when there is an infection at the site of a nonunion, an irradiated or scarred soft-tissue bed, a congenital pseudarthrosis, or a large segmental defect. There are currently no laboratory data, to our knowledge, on the behavior of vascularized grafts in small bones.

Different methods of obtaining vascularized grafts have been described. Chacha12 reported on two patients in whom a bone graft with a vascular pedicle from the pronator quadratus was used to treat a nonunion of the scaphoid that was complicated by avascular necrosis, but he did not comment on the results because of the very short duration of follow-up. Chacha noted that the pedicle was too short to reach the waist of the scaphoid and that the muscle bulk limited adequate visualization of the reduction. Rath et al.41, in an anatomical study, were able to obtain a bone graft with a six-centimeter-long pedicle, with a retrograde blood supply, from the pronator quadratus; however, they presented no clinical data. Kawai and Yamamoto29 used a modification of the method of Rath et al.41 to treat eight symptomatic nonunions that had been present for four months to twelve years and four months. Union was achieved in all patients between the seventh and tenth weeks. However, that report did not include the criteria for union or the details of the preoperative evaluation of avascularity. Mathoulin and Haerle35 reported on seventeen patients who were managed with a vascularized bone graft with a pedicle from the palmar carpal artery. Ten patients had had failure of a previous operation. Union was achieved in all seventeen patients at an average of sixty days (range, forty-five to ninety days) postoperatively. The average duration of follow-up was sixteen months (range, twelve to thirty-six months). The use of a small muscular pedicle ensured continuity of the feeding vascular bundle.

Zaidemberg et al.53 reported on eleven patients in whom a vascularized bone graft from the radial aspect of the distal part of the radius was used to treat a long-standing symptomatic nonunion of the scaphoid that had been present for five months to eleven years. The vascular pedicle was based on the recurrent branch of the radial artery. The fracture had been undiagnosed in five patients, and a Russe bone-grafting procedure had been unsuccessful in another five patients. Union was achieved in all eleven patients within five to eight weeks after the procedure; however, the criteria that were used to define union were not described. Those authors recommended this procedure for the treatment of nonunion, cystic degeneration of the scaphoid, or a limited carpal fusion on the radial side of the wrist. Khan et al.30 thought that bone grafts from the distal part of the radius would create a bone defect close to the articular load-bearing area of the wrist, increasing the risk of fracture. Those authors also suggested that bone flaps raised close to the wrist cannot provide a vascular pedicle of adequate length.

Guimberteau and Panconi21 reported on eight patients in whom a recurrent nonunion of the scaphoid was treated with a vascularized cortical and cancellous bone graft from the ulna. The vascular pedicle was obtained from the ulnar artery. The length of the graft was three centimeters, and the width was equal to one-third of the diameter of the ulnar diaphysis. The nonunion, which had been present for five to twelve years, was located at the scaphoid waist in six patients and at the proximal pole in two. All eight patients had had a failure of a previous bone-grafting procedure. Union was achieved in all eight patients within four to six months after the index procedure. The average duration of follow-up ranged from one year and two months to two years and eight months. The disadvantages of this procedure include a long operating time (approximately three hours) and the need to reconstruct the ulnar artery with a venous graft. In addition, the procedure is associated with a risk of fracture of the ulna.

Brunelli et al.8, in an anatomical study of twenty cadavera, described a bone graft that was obtained from the distal part of the second metacarpal. The vascularity of the graft was derived from the deep artery of the first web space on the lateral aspect of the metacarpal. The authors reported that seven patients had been managed successfully with such a graft, but no clinical data were provided. Mathoulin and Brunelli34 reported on fifteen patients in whom a vascularized bone graft from the index metacarpal was used to treat a nonunion of the scaphoid that had failed to respond to other operative procedures. The nonunion was at the scaphoid waist in thirteen patients, in the proximal segment in one, and in the distal segment in one. Union was achieved in fourteen of the fifteen patients; the time to union ranged from two to ten months. Mathoulin and Brunelli reported that it was difficult to obtain the graft and to position it within the scaphoid when the nonunion was in the proximal or the distal segment and suggested that this procedure should be limited to wrists without evidence of osteoarthritis. Khan et al.30, in an experimental study on thirty-one fresh cadaveric hands that had been injected with a mixture of lead oxide and gelatin, suggested that a vascularized bone graft can be readily obtained from the second metacarpal, which offers a long pedicle based on the second dorsal metacarpal artery. No clinical data were provided.

Hori et al.24, in an animal study, found active proliferation of blood vessels and new-bone formation after transplantation of an artery, a vein, and a vascular bundle into intact bone, isolated bone segments, necrotized bone, and bone grafts. There was proliferation of new blood vessels in all specimens, although half of the necrotized bone remained necrotic. The procedure also was performed in nine patients who had Kienböck disease and in one who had a nonunion of the scaphoid with intraoperative evidence of avascularity. The latter patient, who had had failure of an earlier bone-grafting procedure, had radiographic evidence of union and remodeling within three months after the index operation. Fernandez and Eggli18 reported on eleven patients who were managed with inlay bone-grafting, internal fixation, and implantation of a vascular pedicle because of a nonunion of the scaphoid and avascular necrosis. The absence of the blood supply to the proximal pole was noted on preoperative radiographs as well as intraoperatively. The nonunion, which had been present for six months to two years and nine months, was located at the proximal aspect of the scaphoid in eight patients and at the scaphoid waist in three. Union was achieved in ten patients at an average of ten weeks (range, eight to twelve weeks) postoperatively. The duration of follow-up ranged from two years and five months to eleven years.

It is difficult to assess avascularity in clinical practice. Berggren et al.5, in an experimental study of nine dogs, noted that preservation of the periosteal blood supply alone can result in survival of the bone graft even in a poorly vascularized tissue bed. The viability of vascularized bone grafts was confirmed with light microscopy, microangiography, and scintigraphy. Green20 concluded that the absence of intraoperative punctate bleeding points on the cancellous surface indicates avascularity of the proximal pole and may explain the failure of bone-grafting procedures.

Magnetic resonance imaging appears to be useful for diagnosing avascular necrosis. Brody et al.7 evaluated the results of devascularization of the femur in ten dogs. Magnetic resonance images initially showed homogeneous high signal intensity, corresponding with fat necrosis. Later, there were areas of lower signal intensity, corresponding with fat necrosis and inflammatory reaction; patchy low signal intensity, corresponding with fibrocytic infiltration; and, finally, homogeneous low and intermediate signal intensity, corresponding with organized fibrocytes and new-bone formation. Vande Berg et al.50, in a study of sixty-nine patients who had avascular necrosis or bone-marrow edema of the femoral head, reported that contrast-medium-enhanced magnetic resonance imaging studies were diagnostic in that they demonstrated hypervascularity in areas of bone-marrow edema and hypovascularity in areas of avascular necrosis.

Schnitzler et al.46 compared tetracycline-labeled biopsy specimens that had been obtained from the iliac crest and the distal part of the radius in eighteen patients. Bone structure, static and dynamic turnover, and marrow cellularity were evaluated histomorphometrically. Bone from the distal part of the radius showed thinner cortices, lower volume, thinner trabeculae, greater separation between trabeculae, decreased thickness of the wall, decreased cellularity of the marrow, and poor cellular quality. Those authors concluded that bone from the distal part of the radius was structurally inferior and had lower turnover than bone from the iliac crest.

The dense structure of the cancellous bone from the iliac crest made it possible for us to shape the graft so that it had a cortical proximal portion that would fit the cartilaginous shell of the excavated proximal scaphoid fragment. This was done to ensure that the entire cancellous and subchondral surface of the proximal pole fragment was in contact with the bone graft. The high cellular quality facilitated union, and remodeling of cancellous bone was more rapid and complete than that of cortical bone, which remained partially avascular5,9.

In the present study, the shape and structure of the grafts that were used and the donor sites that were selected were based on the data provided by previous studies. Local ischemia in a poorly vascularized tissue bed is known to be detrimental to the viability of grafted bone cells4; therefore, vascularity of the graft must be maintained. To obtain union, we used bone graft with a vascular pedicle derived from the circumflex iliac vascular bundle and the fascial and periosteal capillary system, maintaining an intraosseous vascular bed with cancellous bone, nutrient foramina, and haversian systems.

The viability of the vascular anastomosis and pedicle were assessed with color Doppler ultrasonography, and the vascular pathways within the bone were evaluated with gadopentetate-dimeglumine-enhanced magnetic resonance imaging. These noninvasive methods were used to obtain as much information as possible about the vascularized graft, although we are aware that the microcapillary system and pathways entering into bone cannot be clearly identified. A viable vascular pathway was present in all patients who had union. However, similar to the limitations noted in other studies of vascularized bone grafts8,18,21,24,35, it was not possible to determine whether the bone graft or the local tissue provided the blood supply. Berggren et al.4 noted that osteocytes could survive twenty-five hours of ischemia. We believe that it is essential to maintain vascularity in the early phase of bone-healing in order to bridge the ischemic period, especially when the tissue bed is poorly vascularized.

The patients who had union did not report pain, but motion of the wrist and grip strength were decreased compared with the uninvolved side. Osteoarthritis remained limited to the radioscaphoid joint, and there was no carpal collapse. We recommend this procedure for younger patients who have an avascular nonunion of the proximal part of the scaphoid of more than two years' duration and osteoarthritis that is limited to the radioscaphoid joint.

*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.

Investigation performed at University Hospital of Traumatology, Innsbruck

1. Alnot, J. Y.; Bellan, N.; Oberlin, C.; and De Cheveigné, C.: Les fractures et pseudarthroses polaires proximales du scaphoïde carpien ostéosynthèse parvissage de proximal à distal. Ann. chir. main, 7: 101-108, 1988.
2. Amadio, P. C.; Berquist, T. H.; Smith, D. K.; Ilstrup, D. M.; Cooney, W. P., III; and Linscheid, R. L.: Scaphoid malunion. J. Hand Surg., 14A: 679-687, 1989.
3. Amadio, P. C., and Taleisnik, J.: Fractures of the carpal bones. In Operative Hand Surgery, edited by D. P. Green. Ed. 3, pp. 799-831. New York, Churchill Livingstone, 1993.
4. Berggren, A.; Weiland, A. J.; and Dorfman, H.: The effect of prolonged ischemia time on osteocyte and osteoblast survival in composite bone grafts revascularized by microvascular anastomoses. Plast. and Reconstr. Surg., 69: 290-298, 1982.
5. Berggren, A.; Weiland, A. J.; Östrup, L. T.; and Dorfman, H.: Microvascular free bone transfer with revascularization of the medullary and periosteal circulation or the periosteal circulation alone. A comparative experimental study. J. Bone and Joint Surg., 64-A: 73-87, Jan. 1982.
6. Böhler, L.: Kahnbeinbrüche. In Die Technik der Knochenbruchbehandlung, pp. 2730-2757. Edited by L. Böhler. Munich, Wilhelm Maudrich, 1977.
7. Brody, A. S.; Strong, M.; Babikian, G.; Sweet, D. E.; Seidel, F. G.; and Kuhn, J. P.: Avascular necrosis: early MR imaging and histologic findings in a canine model. AJR: Am. J. Roentgenol., 157: 341-345, 1991.
8. Brunelli, F.; Mathoulin, C.; and Saffar, P.: Description d'un greffon osseux vascularisé prélevé au niveau de la tête du deuxième metacarpien. Ann. chir. main, 11: 40-45, 1992.
9. Burchardt, H.: The biology of bone graft repair. Clin. Orthop., 174: 28-42, 1983.
10. Carrozzella, J. C.; Stern, P. J.; and Murdock, P. A.: The fate of failed bone graft surgery for scaphoid nonunions. J. Hand Surg., 14A: 800-806, 1989.
11. Carter, P. R.; Malinin, T. I.; Abbey, P. A.; and Sommerkamp, T. G.: The scaphoid allograft: a new operation for treatment of the very proximal scaphoid nonunion or for the necrotic, fragmented scaphoid proximal pole. J. Hand Surg., 14A: 1-12, 1989.
12. Chacha, P. B.: Vascularized pedicular bone grafts. Internat. Orthop., 8: 117-138, 1984.
13. Cooney, W. P.; Dobyns, J. H.; and Linscheid, R. L.: Nonunion of the scaphoid: analysis of the results from bone grafting. J. Hand Surg., 5: 343-354, 1980.
14. Cooney, W. P.; Linscheid, R. L.; Dobyns, J. H.; and Wood, M. B.: Scaphoid nonunion: role of anterior interpositional bone grafts. J. Hand Surg., 13A: 635-650, 1988.
15. Dell, P. C.; Burchardt, H.; and Glowczewskie, F. P., Jr.: A roentgenographic, biomechanical, and histological evaluation of vascularized and non-vascularized segmental fibular canine autografts. J. Bone and Joint Surg., 67-A: 105-112, Jan. 1985.
16. Düppe, H.; Johnell, O.; Lundborg, G.; Karlsson, M.; and Redlund-Johnell, I.: Long-term results of fracture of the scaphoid. A follow-up study of more than thirty years. J. Bone and Joint Surg., 76-A: 249-252, Feb. 1994.
17. Fernandez, D. L.: A technique for anterior wedge-shaped grafts for scaphoid nonunions with carpal instability. J. Hand Surg., 9A: 733-737, 1984.
18. Fernandez, D. L., and Eggli, S.: Non-union of the scaphoid. Revascularization of the proximal pole with implantation of a vascular bundle and bone-grafting. J. Bone and Joint Surg., 77-A: 883-893, June 1995.
19. Fujimaki, A., and Suda, H.: Experimental study and clinical observations on hypertrophy of vascularized bone grafts. Microsurgery, 15: 726-732, 1994.
20. Green, D. P.: The effect of avascular necrosis on Russe bone grafting for scaphoid nonunion. J. Hand Surg., 10A: 597-605, 1985.
21. Guimberteau, J. C., and Panconi, B.: Recalcitrant non-union of the scaphoid treated with a vascularized bone graft based on the ulnar artery. J. Bone and Joint Surg., 72-A: 88-97, Jan. 1990.
22. Herbert, T. J.: The Fractured Scaphoid, pp. 91-151. St. Louis, Quality Medical, 1990.
23. Hooning Van Duyvenbode, J. F. F.; Keijser, L. C. M.; Hauet, E. J.; Obermann, W. R.; and Rozing, P. M.: Pseudarthrosis of the scaphoid treated by the Matti-Russe operation. A long-term review of 77 cases. J. Bone and Joint Surg., 73-B(4): 603-606, 1991.
24. Hori, Y.; Tamai, S.; Okuda, H.; Sakamoto, H.; Takita, T.; and Masuhara, K.: Blood vessel transplantation to bone. J. Hand Surg., 4: 23-33, 1979.
25. Hull, W. J.; House, J. H.; Gustillo, R. B.; Kleven, L.; and Thompson, W.: The surgical approach and source of bone graft for symptomatic nonunion of the scaphoid. Clin. Orthop., 115: 241-247, 1976.
26. Hussl, H.; Russell, R. C.; Zook, E. G.; and Eriksson, E.: Experimental evaluation of tissue revascularization using a transferred muscular-vascular pedicle. Ann. Plast. Surg., 17: 299-305, 1986.
27. Inoue, G., and Sakuma, M.: The natural history of scaphoid non-union. Radiographical and clinical analysis in 102 cases. Arch. Orthop. and Trauma Surg., 115: 1-4, 1996.
28. Jiranek, W. A.; Ruby, L. K.; Millender, L. B.; Bankoff, M. S.; and Newberg, A. H.: Long-term results after Russe bone-grafting: the effect of malunion of the scaphoid. J. Bone and Joint Surg., 74-A: 1217-1228, Sept. 1992.
29. Kawai, H., and Yamamoto, K.: Pronator quadratus pedicled bone graft for old scaphoid fractures. J. Bone and Joint Surg., 70-B(5): 829-831, 1988.
30. Khan, K.; Riaz, M.; and Small, J. O.: The use of the second metacarpal artery for vascularized bone graft. An anatomical study. J. Hand Surg., 23-B: 308-310, 1998.
31. Lindström, G., and Nyström, A.: Natural history of scaphoid non-union, with special reference to “asymptomatic” cases. J. Hand Surg., 17-B: 697-700, 1992.
32. Mack, G. R.; Bosse, M. J.; Gelberman, R. H.; and Yu, E.: The natural history of scaphoid non-union. J. Bone and Joint Surg., 66-A: 504-509, April 1984.
33. Martini, A. K.: Der Spontanverlauf der Kahnbeinpseudarthrose. Orthopäde, 23: 249-254, 1994.
34. Mathoulin, C., and Brunelli, F.: Further experience with the index metacarpal vascularized bone graft. J. Hand Surg., 23-B: 311-317, 1998.
35. Mathoulin, C., and Haerle, M.: Vascularized bone graft from the palmar carpal artery for treatment of scaphoid nonunion. J. Hand Surg., 23-B: 318-323, 1998.
36. Moore, J. B.; Mazur, J. M.; Zehr, D.; Davis, P. K.; and Zook, E. G.: A biomechanical comparison of vascularized and conventional autogenous bone grafts. Plast. and Reconstr. Surg., 73: 382-386, 1984.
37. Nakamura, R.; Hori, M.; Horii, E.; and Miura, T.: Reduction of the scaphoid fracture with DISI alignment. J. Hand Surg., 12A: 1000-1005, 1987.
38. Pechlaner, S.; Hussl, H.; and Künzel, K. H.: Alternative Operationsmethode bei Kahnbeinpseudarthrosen Prospektive Studie. Handchirurgie, 19: 302-305, 1987.
39. Pechlaner, S.; Lohmannn, H.; Buck-Gramcko, D.; and Martin, L.: Zur Problematik der Kahnbeinpseudarthrose-Erfahrung an 240 Fällen. Handchirurgie, 19: 306-309, 1987.
40. Pechlaner, S., and Beck, E.: Sanierende Operationsverfahren bei Skaphoidpseudarthrose. Unfallchirurg, 93: 150-156, 1990.
41. Rath, S.; Hung, L. K.; and Leung, P. C.: Vascular anatomy of the pronator quadratus muscle-bone flap: a justification for its use with a distally based blood supply. J. Hand Surg., 15A: 630-636, 1990.
42. Richards, R. R., and Regan, W. D.: Treatment of scaphoid nonunion by radical curettage, trapezoidal iliac crest bone graft, and internal fixation with a Herbert screw. Clin. Orthop., 262: 148-158, 1991.
43. Ruby, L. K.; Stinson, J.; and Belsky, M. R.: The natural history of scaphoid non-union. A review of fifty-five cases. J. Bone and Joint Surg., 67-A: 428-432, March 1985.
44. Russe, O.: Fracture of the carpal navicular. Diagnosis, non-operative treatment, and operative treatment. J. Bone and Joint Surg., 42-A: 759-768, July 1960.
45. Schernberg, F.: Classification des fractures du scaphoïde carpien. Etude anatomo-radiologique des traits. Rev. chir. orthop., 74: 693-695, 1988.
46. Schnitzler, C. M.; Biddulph, S. L.; Mesquita, J. M.; and Gear, K. A.: Bone structure and turnover in the distal radius and iliac crest: a histomorphometric study. J. Bone and Min. Res., 11: 1761-1768, 1996.
47. Shaffer, J.W.; Field, G. A.; Goldberg, V. M.; and Davy, D. T.: Fate of vascularized and nonvascularized autografts. Clin. Orthop., 197: 32-43, 1985.
48. Smith, B. S., and Cooney, W. P.: Revision of failed bone grafting for nonunion of the scaphoid. Treatment options and results. Clin. Orthop., 327: 98-109, 1996.
49. Stark, H. H.; Rickard, T. A.; Zemel, N. P.; and Ashworth, C. R.: Treatment of ununited fractures of the scaphoid by iliac bone grafts and Kirschner-wire fixation. J. Bone and Joint Surg., 70-A: 982-991, Aug. 1988.
50. Vande Berg, B. E.; Malghem, J. J.; Labaisse, M. A.; Noel, H. M.; and Maldague, B. E.: MR imaging of avascular necrosis and transient marrow edema of the femoral head. Radiographics, 13: 501-520, 1993.
51. Warren-Smith, C. D., and Barton, N. J.: Non-union of the scaphoid: Russe graft vs Herbert screw. J. Hand Surg., 13-B: 83-86, 1988.
52. Watson, H. K., and Ryu, J.: Evolution of arthritis of the wrist. Clin. Orthop., 202: 57-67, 1986.
53. Zaidemberg, C.; Siebert, J.; and Angrigiani, C.: A new vascularized bone graft for scaphoid nonunion. J. Hand Surg., 16A: 474- 478, 1991.
54. Zemel, N. P.; Stark, H. H.; Ashworth, C. R.; Rickard, T. A.; and Anderson, D. R.: Treatment of selected patients with an ununited fracture of the proximal part of the scaphoid by excision of the fragment and insertion of a carved silicone-rubber spacer. J. Bone and Joint Surg., 66-A: 510-517, April 1984.
Copyright 1999 by The Journal of Bone and Joint Surgery, Incorporated